Full publication of results initially presented in abstracts

Abstract

Background

Abstracts of presentations at scientific meetings are usually available only in conference proceedings. If subsequent full publication of results reported in these abstracts is based on the magnitude or direction of the results, publication bias may result. Publication bias creates problems for those conducting systematic reviews or relying on the published literature for evidence about health and social care.

Objectives

To systematically review reports of studies that have examined the proportion of meeting abstracts and other summaries that are subsequently published in full, the time between meeting presentation and full publication, and factors associated with full publication.

Search methods

We searched MEDLINE, Embase, the Cochrane Library, Science Citation Index, reference lists, and author files. The most recent search was done in February 2016 for this substantial update to our earlier Cochrane Methodology Review (published in 2007).

Selection criteria

We included reports of methodology research that examined the proportion of biomedical results initially presented as abstracts or in summary form that were subsequently published. Searches for full publications had to be at least two years after meeting presentation.

Data collection and analysis

Two review authors extracted data and assessed risk of bias. We calculated the proportion of abstracts published in full using a random‐effects model. Dichotomous variables were analyzed using risk ratio (RR), with multivariable models taking into account various characteristics of the reports. We assessed time to publication using Kaplan‐Meier survival analyses.

Main results

Combining data from 425 reports (307,028 abstracts) resulted in an overall full publication proportion of 37.3% (95% confidence interval (CI), 35.3% to 39.3%) with varying lengths of follow‐up. This is significantly lower than that found in our 2007 review (44.5%. 95% CI, 43.9% to 45.1%). Using a survival analyses to estimate the proportion of abstracts that would be published in full by 10 years produced proportions of 46.4% for all studies; 68.7% for randomized and controlled trials and 44.9% for other studies. Three hundred and fifty‐three reports were at high risk of bias on one or more items, but only 32 reports were considered at high risk of bias overall.

Forty‐five reports (15,783 abstracts) with 'positive' results (defined as any 'significant' result) showed an association with full publication (RR = 1.31; 95% CI 1.23 to 1.40), as did 'positive' results defined as a result favoring the experimental treatment (RR =1.17; 95% CI 1.07 to 1.28) in 34 reports (8794 abstracts). Results emanating from randomized or controlled trials showed the same pattern for both definitions (RR = 1.21; 95% CI 1.10 to 1.32 (15 reports and 2616 abstracts) and RR = 1.17; 95% CI, 1.04 to 1.32 (13 reports and 2307 abstracts), respectively.

Other factors associated with full publication include oral presentation (RR = 1.46; 95% CI 1.40 to 1.52; studied in 143 reports with 115,910 abstracts); acceptance for meeting presentation (RR = 1.65; 95% CI 1.48 to 1.85; 22 reports with 22,319 abstracts); randomized trial design (RR = 1.51; 95% CI 1.36 to 1.67; 47 reports with 28,928 abstracts); and basic research (RR = 0.78; 95% CI 0.74 to 0.82; 92 reports with 97,372 abstracts). Abstracts originating at an academic setting were associated with full publication (RR = 1.60; 95% CI 1.34 to 1.92; 34 reports with 16,913 abstracts), as were those considered to be of higher quality (RR = 1.46; 95% CI 1.23 to 1.73; 12 reports with 3364 abstracts), or having high impact (RR = 1.60; 95% CI 1.41 to 1.82; 11 reports with 6982 abstracts). Sensitivity analyses excluding reports that were abstracts themselves or classified as having a high risk of bias did not change these findings in any important way.

In considering the reports of the methodology research that we included in this review, we found that reports published in English or from a native English‐speaking country found significantly higher proportions of studies published in full, but that there was no association with year of report publication. The findings correspond to a proportion of abstracts published in full of 31.9% for all reports, 40.5% for reports in English, 42.9% for reports from native English‐speaking countries, and 52.2% for both these covariates combined.

Authors' conclusions

More than half of results from abstracts, and almost a third of randomized trial results initially presented as abstracts fail to be published in full and this problem does not appear to be decreasing over time. Publication bias is present in that 'positive' results were more frequently published than 'not positive' results. Reports of methodology research written in English showed that a higher proportion of abstracts had been published in full, as did those from native English‐speaking countries, suggesting that studies from non‐native English‐speaking countries may be underrepresented in the scientific literature. After the considerable work involved in adding in the more than 300 additional studies found by the February 2016 searches, we chose not to update the search again because additional searches are unlikely to change these overall conclusions in any important way.

Plain language summary

Full publication of results initially presented in abstracts

Key message

Two important factors increase the probability that a study described in an abstract will subsequently be published in full, (1) the presence of 'positive' or statistically significant results in the abstract and (2) whether the team examining subsequent full publication were from an English‐speaking country or wrote their report in English. The consequence is that systematic reviews relying on fully published research may provide inaccurate or biased findings because of an over‐reliance on studies with positive results or from English‐speaking countries.

Our question

We reviewed the evidence about how often studies submitted as abstracts at a scientific meeting are published in full, usually as a journal article. We found 425 relevant reports, involving 307,028 abstracts.

Background

Investigators prepare and submit abstracts for presentation at scientific meetings. Abstracts selected for presentation are usually collated as conference proceedings, but these are not easily found. Thus, it is important to know whether the work submitted and presented is later published as a journal article, which can easily be identified and contains more study information than the abstract. It is also important to know if the publication of the study depends on the size or direction of results or other factors. If so, systematic reviews relying on the published literature for evidence about health and social care will have incomplete or unbalanced information, leading to inaccurate or biased estimates of the effects of the interventions studied.

Study characteristics

We included 425 research reports described in 551 articles, which had studied the subsequent full publication of 307,028 abstracts from a variety of biomedical and social sciences. Fifty‐four reports included data from abstracts describing randomized or controlled trials. Of the 425 reports, 376 were published in English, and 49 in other languages.

Key results

1. Less than half of all studies, and about two‐thirds of randomized trials, initially presented as summaries or abstracts at meetings, are published as journal articles in the 10 years after presentation.

2. Studies with positive results are more likely to be published.

3. Studies with larger sample sizes are more likely to be published.

4. Studies with abstracts presented orally are more likely to be published than those presented as posters.

5. Studies accepted for presentation at a meeting are more likely to be published than those not accepted.

6. Studies describing basic science are more likely to be published that those describing clinical research.

7. Studies describing randomized trials are more likely to be published than those describing other types of studies.

8. Studies that took place in multiple centers are more likely to be published than those at a single center.

9. Studies classified as ‘high quality’ are more likely to be published than ‘low quality’ studies.

10. Studies with authors from an academic setting are more likely to be published than those with authors from other settings.

11. Studies considered by the report authors to have a high impact are more likely to be published than other studies.

12. Studies with funding source reported are more likely to be published than those not reporting funding.

13. Studies originating in North America or Europe are more likely to be published than those originating elsewhere.

14. Studies from English‐speaking countries are more likely to be published than studies originating elsewhere.

Quality of the evidence

We have confidence in our findings. We considered five criteria to constitute a risk of bias in the included reports, including methods to identify and match full publications to abstracts, and methods to determine whether a factor was associated with full publication. Overall, 7.5% (32/425) of the reports were scored as having an overall high risk of bias, 83.1% (353/425) had at least one criterion at high risk of bias, and 6.1% (26/425) had all criteria at low risk of bias.

Search Date

Our search updated our 2007 review and is current to February 2016. After the considerable work involved in including more than 300 additional studies from the February 2016 searches, we chose not to update the search again because additional searches are unlikely to change our overall conclusions in any important way.

Background

Results of many types of scientific research are presented at professional meetings and summarized in abstracts. Most abstracts are available only in proceedings or journal supplements, although abstracts of many biomedical conferences have been indexed in the bibliographic database Embase from 2009 forward. This means that only some results presented in abstract form are available to those without access to conference proceedings and even though these are increasingly available online (such as on the website of the society hosting the meeting), the abstracts might still need to be checked through individually.

Publication of a trial is more common when its results are 'positive' or 'significant' (Dickersin 1987; Dickersin 1992; Dickersin 1993; Easterbrook 1991; Hopewell 2009; Schmucker 2014; Simes 1986). This publication bias can be divided into two main time periods: from the analysis of the trial to publication of a meeting abstract, and from publication of a meeting abstract to subsequent full publication (von Elm 2003a).

For systematic reviewers, there is difficulty not only in finding studies that are only reported as abstracts, but also in correctly identifying the study design and methods used as described within an abstract (Scherer 2012; Scherer 2013; Scherer 2015). Even for abstracts describing studies with randomized treatment assignment, it is not always possible to tell whether the results are truly those of a randomized trial (Scherer 1994; Scherer 2012). Also, it might not be possible to include results from abstracts in meta‐analyses because of insufficient information describing the outcome or results (Mayo‐Wilson 2017).

The reliability of results presented in abstract form is also questionable. Abstracts may present preliminary results of an ongoing study and may differ from those eventually published in full. Authors, sample sizes, and estimates of treatment effects reported in abstracts have been shown to differ from those presented in subsequent peer‐reviewed publications (Chokkalingam 1998; Hopewell 2003b; Rosmarakis 2005; Saldanha 2016; Weintraub 1987).

For these reasons, attempts to prepare unbiased systematic reviews of a body of evidence may be thwarted by the existence of unpublished studies or by incomplete or imprecise information provided in abstracts. It is important that authors of systematic reviews take the extra time and effort to contact the authors of abstracts to verify information presented in their abstract. If this is not done, then conclusions drawn from the review may be biased or imprecise. On the other hand, the paucity of information within an abstract makes inclusion of study data in a systematic review difficult without the accompanying detail from a full publication or entry in a trials registry. The conflicting information between abstract and full journal report presents yet another problem.

Determining the proportion of studies that were initially presented in abstract form but never published in full is important in assessing the size and nature of the problem. The previous version of this Cochrane Methodology Review found that less than half of all studies initially presented as abstracts were published in full, confirming that this is a major problem (Scherer 2007). That systematic review also found evidence of publication bias in that full publication was associated with 'significant' or 'positive' results The estimated publication proportion using survival analyses at nine years was 52.6% for all studies and 63.1% for randomized or controlled trials. This earlier review included 79 reports, and findings from all medical specialties were considered together. Since then, a substantial number of reports on the subsequent publication of study results have been published, allowing the proportion of abstracts that are published by medical specialty to be considered separately, as reported for oral health by Hua and colleagues (Hua 2016). These newly published reports also considered other factors impacting the proportion of abstracts published, such as academic affiliation. Lastly, because multiple factors may contribute to the probability that an abstract is published, it is important to consider these together using a multivariable model. Therefore, this version of our review updates and extends the earlier Cochrane Methodology Review.

Objectives

• To determine the proportion of abstracts of any study design that are subsequently published in full.

• To determine the proportion of abstracts describing results of randomized or controlled trials that are subsequently published in full.

• To determine the publication rate by time, i.e. the distribution of mean and/or median times to publication and the cumulative publication rate by time.

• To determine the association between full publication of results initially presented in abstracts and characteristics of these abstracts and the studies they report on.

• To determine the association between full publication of results initially presented in abstracts and characteristics of the reports examining full publication of abstracts.

Methods

Criteria for considering studies for this review

Types of studies

We included all reports that examined the proportion of studies that were initially presented in abstract or summary form, which go on to be published in full. We will refer to the papers considered for this systematic review as 'reports', which analyzed abstracts describing various 'studies'.

Eligible reports included information on the following:

• the number of abstracts identified;

• the number or proportion of abstracts followed by full publication of the studies; and

• follow‐up of at least 24 months to assess full publication.

Types of data

Biomedical research studies.

Types of methods

The following study, abstract and report characteristics were examined for association with publication:

• 'positive' or 'significant' results;

• sample size equal to or above the median or mean of all presented studies, or as dichotomized by the report author;

• oral versus poster presentations;

• acceptance (versus rejection) for presentation at a scientific meeting;

• clinical research versus basic science;

• study design;

• multi‐center versus single‐center studies;

• study quality, as defined by the report author;

• academic affiliation versus other type of affiliation;

• other author characteristics;

• perceived potential impact of the abstract results, as defined by the report author;

• funding source;

• country of origin (North American versus European versus rest of the world) of the abstract as classified in the report;

• report written in English versus other languages; and

• correspondence address for the report being in a native English‐speaking country versus elsewhere (reports from Australia, Canada, New Zealand, the UK and the USA were considered native English‐speaking countries).

We did not impose our definitions for 'positive' results, 'clinical research', or 'basic science', but used the definitions used by the authors of the included reports, and recognize that abstracts describing clinical research most likely include many types of study designs, covering a spectrum that includes case studies and randomized trials. We also used the report authors' classification of whether an abstract described a randomized or controlled trial.

Types of outcome measures

Outcomes were subsequent full publication of study results described in the abstract and the time interval between presentation at meetings and subsequent full publication.

Search methods for identification of studies

Electronic searches

We searched MEDLINE, Embase, the Cochrane Library, Web of Science, reference lists, and author files. For the full search strategy, see appendices (Appendix 1; Appendix 2; Appendix 3). Science Citation Index and Google Scholar "cited by" function was searched for articles that cited included identified reports. We ran these searches in February 2016 and have agreed with the Cochrane Methodology Review Group that the search should not be updated further at this time. This is because the very large number of reports identified in the February 2016 search and the stability of our overall conclusions means that updating our searches would be of much less relevance than it would be for Cochrane Reviews focusing on clinical questions.

Searching other sources

References listed in each identified report were reviewed for inclusion (Horsley 2011), and additional reports were found through author files or by word of mouth.

Data collection and analysis

Two review authors independently screened each title and abstract obtained from the electronic or reference searches, classifying each for inclusion as 'Yes', 'No', or 'Maybe'. Full‐text articles were obtained for all reports classified as 'Yes' or 'Maybe'. Two review authors independently compared each full‐length article with the eligibility criteria for the review. We resolved disagreements on classification of eligibility through discussion and consensus.

Data extraction and management

We included reports if abstracts or other research summaries were followed for at least 24 months after presentation and the proportion of abstracts that were published was presented or could be calculated from reported results or from personal communication with the author. For reports that followed abstracts for varying lengths of time, only data from follow‐up of abstracts for 24 months or more were included. If an explicit length of follow‐up was not given, the number of follow‐up months was calculated. If a month (e.g. for the date of the original meeting) was not given or could not be extracted from additional sources (e.g. meeting or relevant society websites), we used the middle of the year. If authors of reports determined the proportion of abstracts that were subsequently published in full using survival analysis, we included the reports because follow‐up of less than 24 months was taken into account in determining the proportion of abstracts that were published in full. We assumed that the criterion for a minimum follow‐up of 24 months had been met if the length of time between the original meeting and the publication of the report was at least three years. If the time between the meeting and the publication of the report was only two years or less, we assumed that this inclusion criterion could not have been met.

At least two review authors extracted information related to publication from each report. This included total number of abstracts, number or proportion of abstracts subsequently published in full, proportion of abstracts published in full by six‐month time intervals since presentation, median or mean time to publication, median or mean sample size, and proportion of abstracts subsequently published by the risk factors indicated above. We also extracted information related to full publication and study characteristics for abstracts describing randomized or controlled trials separately. We asked the corresponding authors for data that were collected but not published. One reminder email was sent to authors not initially responding. We resolved disagreements by discussion and consensus.

Assessment of risk of bias in included studies

Two review authors independently assessed the risk of bias; conflicts were resolved by discussion. For this update to our review, we modified the 'Risk of bias' assessment developed by Schmucker and colleagues (Schmucker 2014), who studied reports looking at publication of trials from inception. The following five items were considered to constitute a risk of bias in the included reports.

• Methods used to sample abstracts

  • Reports with a low risk of bias included all studies presented at a meeting, a random sample of studies, a systematic subset of studies, or a sample based on a specific study design.

  • Reports with a high risk of bias included only abstracts based on 'selected' questionnaires, if author contact was used to identify publications, or other methods subject to selection bias were used.

  • Reports with an unclear risk of bias did not report the sampling method.

• Length of follow‐up time between presentation at the meeting and date of the search for publications.

  • Reports with a low risk of bias had a follow‐up time of 48 months or longer.

  • Reports with a high risk of bias had a follow‐up time of less than 48 months.

  • Reports with an unclear risk of bias did not report the date of the search for full publications and it could not be calculated to allow classification as either high or low risk of bias.

  • If a report included meetings held in multiple years, we considered the length of follow‐up between the majority of the meetings and the search for subsequent publications.

• Methods used to identify full publications

  • Reports with a low risk of bias searched at least two electronic databases for publications, or searched at least one database followed by author contact, or contacted authors with a response rate of at least 80%. For this criterion, we considered MEDLINE, PubMed, and PreMEDLINE to be the same database, because they differ only by when citations are added to the database.

  • Reports with a high risk of bias searched only one database, or had a response rate of less than 80% when they contacted authors.

  • Reports with an unclear risk of bias did not report the search methods, or did not report the response rate with author contact.

• Methods used to match an abstract with a full publication

  • Reports with a low risk of bias matched the abstract with the full publication using two or more criteria (e.g. author name, sample size, contents, etc.) when full publications were identified using electronic databases, or used author contact to identify or confirm publication match.

  • Reports with a high risk of bias matched the abstract with the full publication using only one criterion.

  • Reports with an unclear risk of bias did not report the matching criteria.

• Adjustment for factors possibly related to full publication

  • Reports with a low risk of bias reported on two or more factors possibly associated with publication, adjusted using multivariable analyses, either for proportion published or for time to publication.

  • Reports with a high risk of bias reported on only one factor associated with publication, or on two or more factors possibly associated with publication without using multivariable analyses.

  • Reports with an unclear risk of bias reported on two or more factors possibly associated with publication, but the methods used for analyses were not reported.

  • Reports that did not report on any factors possibly associated with publication were classified as not applicable.

We classified reports as having high overall risk of bias if they had been classified as high risk of bias for three or more of the risk items described above, or for two or more items if only four items could be assessed.

Data synthesis

Estimates of proportion of abstracts published

We estimated the overall proportion of abstracts subsequently published as journal articles for all included reports using a random‐effects meta‐analysis based on logit‐transformed proportions (Borenstein 2010). The between‐study variance was estimated using the Paule‐Mandel method (Paule 1982), and the 95% confidence interval (95% CI) for the overall proportion was calculated using the Hartung‐Knapp method (Hartung 2001). We also calculated the proportion of published abstracts describing various subgroups, including study design, medical specialty, source of abstract (meeting, register, etc.), and type of sample (oral, poster, submitted, etc.). We did not attempt to separate the same abstracts in the few instances where a single meeting was reported in more than one report, but included all abstracts in the analyses.

Time to publication was analyzed by using survival analyses to allow for differing lengths of follow‐up. We pooled reported numbers of full publications at six‐month intervals since presentation. Abstracts that had not been published were censored at the point in time when follow‐up for that report was ended, because subsequent publication could not be ascertained. We did not analyze time to publication by looking at mean or median time to publication because those values are dependent on the length of follow‐up.

We combined reports that examined the association of the same factor with subsequent full publication. Strengths of association of factors possibly associated with publication are expressed as risk ratios (RR) with 95% CIs using a random‐effects model. We tested for heterogeneity by using Chi² test or I² statistic and looked for sources of heterogeneity when statistical heterogeneity was present.

We evaluated year of report publication to assess differences over time using meta‐regression and a mixed model with the publication year centered at year 2000 and a logit transformed scale. We also conducted a mixed‐effect multivariable analysis of the association of subsequent full publication with characteristics of the reports. Characteristics of the reports that we considered for inclusion in the model included: year report was published, report in English language or not; funding reported or not; report from a native English‐speaking country or not; and type of sample (submitted abstract, oral presentation only, poster presentation only, or mixed). However, we found that including the latter covariate in the model resulted in loss of information. With several P values for this covariate, it was impossible to interpret, especially when using a different reference group than 'oral only'. Therefore, we used a likelihood‐ratio test to evaluate the overall effect of the covariate 'type of sample'.

Analyses not available in RevMan 5 were conducted using R package meta, version 4.9‐0 (Schwarzer 2007).

Sensitivity analysis

We performed sensitivity analyses, firstly, excluding reports that were themselves published only as abstracts and, secondly, excluding reports that had a high overall risk of bias.

Results

Description of studies

Results of the search

From the 6872 reports we identified in electronic searching and other sources, we retrieved 626 full‐text articles (Figure 1). Of these, 75 were excluded. The majority of excluded reports (41 of 75) investigated other aspects related to publication such as the process for selecting abstracts for presentation at a meeting, comparison of results represented in abstracts with those in a full‐length report, or evaluating the quality of reporting of abstracts. Sixteen reports were excluded because they were editorials or commentaries (Alfaro 2015; Castro 2009; de Carvalho 2013; de Castro 2011; Floch 2002; Hayes 2011; Humberto 2011; Knobloch 2009; Leder 1985; Mallory 1999; Olsson 2006; Stewart 2013; Tannenbaum 1978; Tintinalli 2001; von Elm 2001; Wein 2005); and 11 because no relevant numeric data were given (Covington 2012; Garvey 1971; Gidding 1992; Godkin 1993; Jourbert 2004; Kain 2010; Koren 1986; Kurien 2014, Singer 1999; Taran 2008; Timmer 2001b). Four reports reported a follow‐up period of less than 24 months between the meeting and search for full publications (Cuellar 2005; De Giorgi 2014; Garvey 1970; Huber 2001); and two reports did not describe a biomedical meeting topic (Aleixandre‐Benavent 2009; Fennewald 2005).

1.

Study flow diagram.

Included reports

In this updated review, we have included 425 eligible biomedical research reports, described in 551 articles. The total number of abstracts followed in all reports was 307,028 and ranged from nine to 27,208 (median = 383) in individual reports. Sixty reports were only presented in abstract format, and 17 were letters to the editor. One report was an unpublished manuscript provided by the author (Halpern 2002), and the remainder were available in one or more full‐length articles. Most reports investigated studies presented at meetings, but seven reported on abstracts located in a specialized register or other abstract collections ((Chalmers 1990a; Cheng 1998; Delamere 2005; Koene 1994; Lensen 2015; Siegfried 2007; Tieman 2010). Specialized registers included the Oxford Database of Perinatal Trials (Chalmers 1990a), all abstracts published in the Nederlands Tijdschrift voor Geneeskunde (Koene 1994), abstracts within Cochrane specialized registers (Cheng 1998; Delamere 2005; Lensen 2015), abstracts within AIDSearch (Siegfried 2007), or abstracts within CareSearch, a database of palliative care grey literature (Tieman 2010). Two reports studied abstracts presented at various meetings by a specific group of authors (Dirk 1996; Morrison 1994) rather than from a specific meeting.

Overall report characteristics

Of the 416 reports that examined meeting abstracts, 18 looked at abstracts that were submitted to a meeting (Camacho 2005; Canosa 2011; Cauchy 2014; Dhaliwal 2008; Duchini 1997; Fede 2010; Feldman 2015; Hashkes 2003; Jörgens 2014; Leles 2006; Moorthi 2013; Saeed 2011; Toro‐Polo 2012; Varghese 2011; Winnik 2012; Wong 2006; Yamaguchi 1990; Yoon 2010), 16 examined both submitted and presented abstracts (Bernstein 1983; Callaham 1998; Carroll 2003; Cifuentes 2007; De Bellefeuille 1992; de Meijer 2015; Eloubeidi 2001; Goldman 1980; Goldman 1982; Jackson 2000; McCormick 1985; Ohlsson 1999; Papoutsis 2015; Schibilisky 2014; Timmer 2002; Todd 1997), and the remaining 382 examined only those abstracts that had been accepted and presented at a meeting.

Among all reports, 54 only included abstracts presented orally and 10 only those presented as posters. Four only looked at abstracts describing results presented in a plenary session (Cohen 2012; Kaifi 2013; O' Connor 2015; Riordan 2000).

Fifty‐four reports included data from abstracts describing randomized or controlled trials, of which 26 included only randomized trials (Autorino 2010; Chalmers 1990a; Cheng 1998; Curry 2003; Das 2013; Delamere 2005; Diezel 1999; Evers 2000; Herbison 2004; Hopewell 2003; Klassen 2002; Kottachchi 2010; Krzyzanowska 2003; Lensen 2015; Moorthi 2013; Ohlsson 1999; Ospina 2006; Polyzos 2011; Saldanha 2016; Scherer 1994; Siegfried 2007; Tam 2008; Tam 2011; Toma 2006; Turpen 2010; Zaretsky 2002), with another six including both randomized and controlled trials (Akbari‐Kamrani 2008; Blackwell 2009; Camacho 2005; De Sio 2012; Hoeg 2009; Vecchi 2009), and one included randomized trials and observational studies (Jones 2014). Other investigators limited the study design of the abstracts they examined to diagnostic accuracy studies (Brazzelli 2009); surveys (Joe 2015); qualitative studies (Petticrew 2008); cost‐effectiveness studies (Chan 2013), or systematic reviews (Hopewell 2015). Ten investigators limited their report to a subset of abstracts by authors from a specific country or region that had been presented at a multi‐national meeting. These authors included those from Brazil (Arap 2014; Fernandes 2008; Pinheiro 2009; Saad 2008); India and the UK (Arora 2012); Malaysia (Chan 2002); Korea (Ha 2008); China (Lin 2011); Africa and North America (Siegfried 2007); or Serbia (Vuckovic‐Dekic 2001). One author looked only at abstracts from English‐speaking countries (Sivan 2010). The geographic location of local or national meetings in a large number of reports typically resulted in a higher proportion of studies from local or national authors, although abstract authors from other countries were usually not excluded. Eight investigators looked at abstracts presented by residents or medical students, or abstracts from meetings aimed at resident presentations (Ahlers‐Schmidt 2009; Duthie 2012; McKelvey 2010; Mowla 2006; O'Dell 2012; Olmos‐de‐Aguilera 2013; Susarla 2015; Toro‐Polo 2012).

Three hundred and seventy‐six reports were published in English, and 264 were authored by individuals from native English‐speaking countries. Of 49 reports not written in the English language, 15 were in Spanish, seven in French, six in German, six in Portuguese, six in Turkish, four in Korean, two in Dutch, two in Chinese and one in Japanese. Funding information was reported in 90 reports with 19 authors reporting having received funds from government, 18 from a foundation, one from industry and 42 reporting having received no funds.

Abstracts represented work in many different medical specialties as well as basic and other related sciences (see Characteristics of included studies and Proportion of Abstracts Published (Analysis 1.1, and Table 1). Within broad classification by medical specialty, there were nine reports in anesthesiology; 14 in cardiology; seven in dermatology; 24 in emergency medicine; four in endocrinology/nutrition; 18 in gastroenterology; 16 in general medicine/primary care; 18 in gynecology/obstetrics; four in hematology; nine in human reproduction and development; six in infectious disease/immunology; two in nephrology; six in neurology; 21 in oncology; nine in ophthalmology/optometry; 15 in oral health; 37 in orthopedics/orthopedic surgery; 15 in otolaryngology/head and neck surgery; three in pathology; 13 in pediatrics; nine in physical and rehabilitation medicine; six in psychiatry; 16 in radiology; seven in rheumatology; 49 in surgery; 27 in urology, and 17 in other clinical specialties. There were also reports from allied health fields, including two in nursing, 14 in pharmacology, and three in sociology. Five reports covered medical decision making/health policy and seven medical education/library science. Lastly, there were 13 reports in other non‐clinical specialties, such as biochemistry or clinical chemistry, food safety research, or forensic science. These latter reports were included because they addressed human health in some way.

1.1. Analysis.

Comparison 1 Proportion of abstracts published, Outcome 1 Overall proportion of abstracts published.

Overall proportion of abstracts published
Study	Topic	Number of Abstracts	Number of full publications	Percent full publications	Limited by design	Limited by presentation
Abes 2004	Surgery ‐ pediatric surgery	776	156	20.1	No	No
Abicht 2012	Orthopedics/orthopedic surgery ‐ foot and ankle	825	198	24	No	Posters only
Abuzeid 2013	Cardiology ‐ cardiovascular research	3565	858	24.1	No	No
Abzug 2014	Surgery	1127	518	46	No	No
Acevedo 2014	Oncology ‐ hematology	1972	353	17.9	No	No
Adhikari 2011	Emergency medicine ‐ ultrasound	1479	333	22.5	No	No
Aggarwal 2012	Rheumatology	271	19	7	No	No
Ahlers‐Schmidt 2009	General medicine/primary care	67	23	34.3	No	Posters only
Akbari‐Kamrani 2008	Surgery ‐ laser medicine and surgery	198	87	43.9	Yes (clinical trials)	No
Al‐Qaoud 2013	Urology	439	248	56.5	No	Oral only
Allart 2013	Physical and rehabilitation medicine	231	49	21.2	No	No
Allart 2015	Physical and rehabilitation medicine	779	169	21.7	No	No
Alonso‐Arroyo 2014	Cardiology	300	115	38.3	No	Oral only
Alpi 2011	Medical education/library science	62	23	37.1	No	No
Amarilyo 2013	Rheumatology	2149	1269	59.1	No	No
Amirhamzeh 2012	Orthopedics/orthopedic surgery ‐ pediatrics	762	386	50.7	No	No
Arap 2014	Urology	195	100	51.3	No	No
Arora 2012	Oncology‐ pediatrics	479	139	29	No	No
Arrive 2001	Radiology	456	39	8.6	No	Oral only
Arrive 2004	Radiology	1897	635	33.5	No	Oral only
Autorino 2006	Urology ‐ endourology	1100	234	21.3	No	No
Autorino 2007	Urology	1406	666	47.4	No	No
Autorino 2008	Urology	1877	415	22.1	No	No
Autorino 2010	Urology ‐ endourology	94	45	47.9	Yes (RCTs)	No
Bagheri 2005	Otolaryngology head and neck surgery ‐ oral and maxillofacial surgery	446	155	34.8	No	No
Bakkum 2014	Physical and rehabilitation medicine	776	249	32.1	No	No
Bakkum 2015	Ophthalmology/optometry ‐ optometry	518	108	20.8	No	No
Balasubramanian 2006	Surgery	241	136	56.4	No	No
Beker‐Acay 2015	Radiology	3192	300	9.4	No	No
Bello 2013	Urology	75	18	24	No	Oral only
Bergoli 2011	Oral health ‐ oral research	140	37	26.4	No	No
Bergqvist 2008	Surgery ‐ vascular surgery	699	553	79.1	No	No
Bernstein 1983	Gastroenterology ‐ hepatology	82 (presented) 177 (submitted)	56 (presented) 106 (submitted)	68.3 (presented) 59.9 (submitted)	No	No
Berookhim 2013	Human reproduction and development ‐ sexual and reproductive medicine	208	107	51.4	No	No
Bhandari 2002	Orthopedics/orthopedic surgery	465	159	34.2	No	No
Bhasin 2007	Other clinical specialties ‐vascular research	106	63	59.4	No	Oral only
Bird 1999	Other non‐clinical specialties ‐ marine mammology	425	249	58.6	No	No
Blackwell 2009	Gynecology/obstetrics ‐ maternal‐fetal medicine	90	50	55.6	Yes (controlled trials)	No
Bolac 2009	Dermatology	540	299	55.4	No	Oral only
Boldt 1999	Anesthesiology	566	233	41.2	No	No
Bonitz 2011	Human reproduction and development ‐ male reproduction	284	107	37.7	No	No
Bowrey 1999	Surgery	496	233	47	No	No
Brace 2010	Other non‐clinical specialties ‐ veterinary vaccine	154	11	7.1	No	No
Brazzelli 2009	Neurology ‐ stroke	160	121	75.6	Yes (diagnostic accuracy studies)	No
Brost 2005	Gynecology/obstetrics ‐ maternal‐fetal medicine	288	208	72.2	No	Oral only
Buchan 2011	Ophthalmology/optometry ‐ ophthalmology	179	64	35.8	No	No
Burden 2014	Gastroenterology	112	37	33	No	No
Bydder 2004	Radiology	481	168	34.9	No	No
Byerly 2000	Pharmacology	716	126	17.6	No	No
Callaham 1998	Emergency medicine	179 (presented) 492 (submitted)	111 (presented) 214 (submitted)	62.0 (presented) 43.4 (submitted)	No	No
Camacho 2005	Oncology	275 (submitted)	185 (submitted)	67.3 (submitted)	Yes (phase I)	No
Canosa 2011	Pediatrics	140 (submitted)	16 (submitted)	11.4 (submitted)	No	No
Carroll 2003	Pediatrics	454 presented) 614 (submitted)	243 (submitted) 210 (presented)	39.6 (submitted) 46.3 (presented)	No	No
Cartwright 2007a	Gynecology/obstetrics ‐ urogynecology	130	77	59.2	No	Oral only
Cartwright 2007b	Gynecology/obstetrics ‐ urogynecology	116	64	55.2	No	No
Castagnetti 2013	Urology ‐ pediatrics	1194	564	47.2	No	No
Castaldi 2015	Other non‐clinical specialties ‐ public health	621	146	23.5	No	No
Castillo 2000	Anesthesiology	491	84	17.1	No	No
Castillo 2002	Anesthesiology	472	199	42.2	No	No
Cauchy 2014	Gastroenterology ‐ digestive and hepatobiliary surgery	897 (submitted)	334 (submitted)	37.2 (submitted)	No	No
Chalmers 1990a	Pediatrics ‐ perinatology	176	64	36.4	Yes (RCTs)	No
Chan 2002	Pediatrics	105	37	35.2	No	No
Chan 2008	Medical decision making/health policy ‐ health technology	257	122	47.5	No	No
Chan 2011	Emergency medicine	389	129	33.2	No	No
Chan 2013	Oncology	164	65	39.6	Yes (cost‐effectiveness studies)	No
Chand 2008	Cardiology ‐ cardiothoracic surgery	2172	648	29.8	No	No
Chapman 2012	Medical decision making/health policy ‐ systematic reviews, health technology	908	312	34.4	No	N
Cheng 1998	Other clinical specialties ‐ cystic fibrosis	178	57	32	Yes (RCTs)	No
Chung 2012	Urology	1005	421	41.9	No	No
Chung 2012a	Surgery ‐ plastic and reconstructive surgery	1176	455	38.7	No	No
Ciesla 2001	Pathology ‐ cytopathology	257	116	45.1	No	No
Cifuentes 2007	Gastroenterology ‐ hepatology	766 (presented) 1533 (submitted)	366 (presented) 572 (submitted)	47.2 (presented) 37.3 (submitted)	No	No
Clemency 2014	Emergency medicine	635	281	44.3	No	No
Cobos 1998	Cardiology	120	26	21.7	No	No
Cohen 2012	Oncology ‐ gynecologic cancer	378	309	81.7	No	Plenary only
Cohen 2012a	Otolarygnology head and neck surgery	2463	1152	46.8	No	No
Collet 1993	Oral health ‐ dentistry	747	94	12.6	No	No
Collet 1997	Oral health ‐ odontology	506	61	12.1	No	No
Collet 2006	Oral health ‐ dental research	1671	489	29.3	No	No
Collier 2010	Otolaryngology head and neck surgery ‐ oral and maxillofacial surgery	623	147	23.6	No	No
Cornu 2012	Urology	226	96	42.5	No	Oral only
Corry 1990	Oral health ‐ dental research	275	63	22.9	No	No
Costa 2009	Oncology ‐ breast cancer	1620	721	44.5	No	No
Craig 2001	Orthopedics/orthopedic surgery	1005	495	49.3	No	No
Cromer 1998	Pediatrics ‐ adolescent medicine	128	58	45.3	No	Oral only
Curry 2003	Surgery ‐ pediatric surgery	9	1	11.1	Yes (RCTs)	No
Czorlich 2016	Surgery ‐ neurosurgery	433	175	40.4	No	No
Dahllof 2008	Oral health ‐ pediatric dentistry	771	204	26.5	No	No
Daluiski 1998	Orthopedics/orthopedic surgery	888	463	52.1	No	No
Dangouloff‐Ros 2015	Radiology	744	298	40.1	No	Oral only
Daruwalla 2015	Orthopedics/orthopedic surgery	443	125	28.2	No	No
Das 2013	Neurology ‐ epilepsy	52	33	62.7	Yes (RCTs)	No
Davies 2002	Pediatrics	172	78	45.3	No	No
de Andrade 2011	Emergency medicine ‐ trauma	347	10	2.9	No	No
De Bellefeuille 1992	Oncology	81 (presented) 197 (submitted)	63 (presented) 115 (submitted)	77.8 (presented) 58.4 (submitted)	No	No
de Meijer 2015	Surgery	1305 (presented) 2174 (submitted)	870 (presented) 1358 (submitted)	66.7 (presented) 62.5 (submitted)	No	No
de Oliveira 2012	Neurology ‐ pain	348	31	8.9	No	No
De Sio 2012	Urology	3110	453	14.6	Yes (controlled trials)	No
Del Rio Moro 2010	Nursing ‐ cardiology nursing	610	76	12.5	No	No
Delamere 2005	Dermatology	30	13	43.3	Yes (RCTs)	No
Delgado Gallego 2007	General medicine/primary care	528	57	10.8	No	No
DeMola 2009	Orthopedics/orthopedic surgery	558	321	57.5	No	No
Dhaliwal 2008	Ophthalmology/optometry ‐ ophthalmology	200 (submitted)	33 (submitted)	16.5 (submitted)	No	No
Dicembrino 2014	Pediatrics	232	136	58.6	No	No
Diezel 1999	Psychiatry	95	44	46.3	Yes (RCTs)	No
Dirk 1996	Anesthesiology	147	80	54.4	No	N/A
Donegan 2010	Orthopedics/orthopedic surgery	756	367	48.5	No	No
Dooley 2003	Pharmacology ‐ hospital pharmacology	187	21	11.2	No	No
Doğan 2012	Otolarygnology head and neck surgery	218	61	28.0	No	No
Dressler 2015	Radiology	402	112	27.9	No	No
Drury 2012	Surgery ‐ cardiovascular surgery	909	606	66.7	No	No
Duchini 1997	Gastroenterology	2512 (submitted)	1229 (submitted)	48.9	No	No
Dudley 1978	Surgery	51	29	56.9	No	No
Dumville 2008	Other clinical specialties ‐wound management	467	57	12.2	No	No
Durinka 2014	Surgery ‐ organ transplantation	1938	244	12.6	No	No
Durinka 2014b	Surgery	335	64	19.1	No	Posters only
Duthie 2012	Psychiatry	20	10	50	No	No
Dyson 2006	Other non‐clinical specialties ‐ veterinary anesthesiology	283	208	73.5	No	No
Eck 2005	Other clinical specialties ‐ sports medicine	358	133	37.2	No	No
Ejnisman 2013	Orthopedics/orthopedic surgery	653	174	26.6	No	No
Elder 1994	General medicine/primary care	475	226	47.6	No	No
Elliot 2016	Surgery ‐ neurologic surgery	754	383	50.8	No	No
Eloubeidi 2001	Gastroenterology ‐ endoscopy	247 (presented) 451 (submitted)	80 (presented) 113 (submitted)	32.4 (presented) 25.1 (submitted)	No	No
Ensom 1998	Pharmacology ‐ hospital pharmacology	363	89	24.5	No	No
Erdağ 2014	Otolarygnology head and neck surgery	1454	319	21.9	No	No
Ersoy 2015a	Gynecology/obstetrics	243	84	34.5	No	No
Ersoy 2015b	Gynecology/obstetrics	161	46	28.6	No	No
Evans 2015	Pharmacology	1496	109	7.3	No	No
Evers 2000	Human reproduction and development	151	79	52.3	Yes (RCTs)	No
Facione 2007	Urology	239	60	25.1	No	No
Fede 2010	General medicine/primary care	408 (submitted)	138 (submitted)	33.8 (submitted)	No	Oral only
Feldman 2015	Hematology ‐ apheresis	1152 (submitted)	196 (submitted)	17.0 (submitted)	No	No
Fernandes 2003	Surgery	1133	29	2.6	No	No
Fernandes 2008	Other clinical specialties ‐ sports medicine	263	58	22.1	No	No
Fernandez 2011	Physical medicine and rehabilitation ‐ cardiac rehabilitation	279	17	6.1	No	No
Fesperman 2008	Urology	1195	400	33.5	No	No
Forlin 2013	Orthopedics/orthopedic surgery	267	58	21.7	No	No
Fosbol 2012	Cardiology	27208	8335	30.6	No	No
Freeman 2012	Oncology	90	70	77.8	No	No
Frost 2015	Orthopedics/orthopedic surgery ‐ spine	764	339	44.4	No	Oral only
Galang 2011	Oral health ‐ dental education	370	71	19.2	No	Posters only
Gandhi 2004	Gynecology/obstetrics	265	190	71.7	No	No
Garcia‐Covarrubias 2002	Other clinical specialties ‐ undersea and hyperbaric medicine	503	86	17.1	No	No
Garcia‐Muret 2009	Dermatology and venereology	1471	200	13.6	No	No
Gaundong Mbethe 2008	General medicine/primary care	303	82	27.1	No	No
Gavazza 1996	Surgery ‐ hand	376	165	43.9	No	No
Gilbert 2004	Gynecology/obstetrics ‐ maternal‐fetal medicine	176	120	68.2	No	Oral only
Ginzburg 2014	Gynecology/obstetrics ‐ urogynecology	140	80	57.1	No	Oral only
Glick 2006	Surgery ‐ organ transplantation	1147	607	52.9	No	No
Goldman 1980	Cardiology	207 (presented) 276 (submitted)	113 (presented) 137 (submitted)	54.6 (presented) 49.6 (submitted)	No	No
Goldman 1982	Hematology ‐ nephrology	202 (presented) 303 (submitted)	127 (presented) 171 (submitted)	62.9 (presented) 56.4 (submitted)	No	No
Gorman 1990	Other non‐clinical specialties ‐ toxicology	269	134	49.8	No	No
Gourtaud 2009	Urology	443	153	34.5	No	No
Greenberg 2008	Medical decision making/health policy	239	64	26.8	No	No
Gregory 2012	Surgery ‐ plastic surgery	888	400	45	No	Oral only
Grzeskowiak 2014	Pharmacology ‐ pediatrics and perinatal health	96	10	10.4	No	No
Guryel 2006	Orthopedics/orthopedic surgery	415	137	33.0	No	No
Ha 2008	Radiology	1097	301	27.4	No	No
Hackett 2014	Surgery ‐ liver transplantation	2345	913	38.9	No	No
Hajji 2016	Gynecology/obstetrics ‐ urogynecology	270	110	40.7	No	Oral only
Halpern 2002	Gynecology/obstetrics ‐ obstetrical anesthesia	145	51	35.2	No	No
Hamlet 1997	Orthopedics/orthopedic surgery	1465	668	45.6	No	No
Hanchanale 2014	General medicine/primary care ‐ palliative care	538	230	42.7	No	No
Harel 2011	Nephrology	300	127	42.3	No	Posters only
Harris 2006	Orthopedics/orthopedic surgery	200	62	31.0	No	Oral only
Harris 2007	Orthopedics/orthopedic surgery	318	175	55.0	No	No
Harshavardhana 2009	Orthopedics/orthopedic surgery	278	203	73.0	No	Oral only
Harshavardhana 2009a	Orthopedics/orthopedic surgery ‐ scoliosis	1063	560	52.7	No	No
Harvey 2010	Medical education/library science	442	122	27.6	No	No
Hashkes 2003	Rheumatology	331 (submitted)	134 (submitted)	40.5 (submitted)	No	No
Herbison 2004	Urology ‐ incontinence	82	41	50	Yes (RCTs)	No
Hernandez‐Garcia 2011	Medical decision making/health policy ‐ quality in healthcare	325	15	4.6	No	Oral only
Hernández‐Jiménez 2011	Endocrinology/nutrition	1045	186	17.8	No	No
Herron 1993	Emergency medicine ‐ air medical transport	160	72	45	No	No
Hoag 2006	Urology	1584	875	55.2	No	No
Hoeg 2009	Oncology	559	361	64.6	Yes (phase II trials)	No
Hogan 2009	Dermatology	117	54	46.2	No	No
Hopewell 2001	Medical decision making/health policy ‐ systematic reviews, health technology	91	39	42.9	No	No
Hopewell 2003	General medicine/primary care	962	589	61.2	Yes (RCTs)	No
Hopewell 2015	Other clinical specialties ‐ rheumatology, endocrinology ‐ diabetes, cardiology, psychiatry, anesthesiology, oncology, hematology, thoracic medicine, infectious diseases	197	105	53.3	Yes (systematic reviews)	No
Hopper 2009	Gastroenterology	4096	1618	39.5	No	No
Housri 2008	Surgery	1200	722	60.2	No	No
Hussein 2014	Other clinical specialties ‐ vascular medicine	367	239	65.0	No	No
Izadpanah 2014	Surgery ‐ plastic surgery	128	73	57.0	No	No
Jackson 2000	Orthopedics/orthopedic surgery ‐ pediatric orthopedics	349 (presented) 777 (submitted)	184 (presented) 348 (submitted)	52.7 (presented) 44.8 (submitted)	No	No
Jamjoom 2014	Surgery ‐ neurological surgery	273	35	12.8	No	Oral only
Jamjoom 2015	Surgery ‐ neurological surgery	494	181	36.6	No	No
Jara‐Tracchia 2010	Oral health ‐ dental research	4137	1156	27.9	No	No
Jasko 2003	Oncology ‐ musculoskeletal tumor	336	137	40.8	No	Oral only
Joe 2015	Emergency medicine ‐ burns	162	27	16.7	Yes (surveys only)	No
Jones 2008	Infectious disease/immunology ‐ HIV, AIDS	139	47	33.8	No	No
Jones 2014	Urology	127	43	33.9	Yes (RCTs and observational studies)	No
Juzych 1991	Ophthalmology/optometry ‐ ophthalmology	175	105	60	No	No
Juzych 1993	Ophthalmology/optometry ‐ ophthalmology	327	206	63	No	No
Jörgens 2014	Endocrinology/nutrition ‐ diabetes	493 (submitted)	209 (submitted)	42.4 (submitted)	No	No
Kabay 2005	Surgery	2118	120	5.7	No	No
Kaifi 2013	Surgery	76	60	78.9	No	Plenary (oral) only
Kalkan 2015	Emergency medicine	1721	437	25.4	No	No
Kalyoncu 2011	Rheumatology	799	173	21.6	No	No
Kaya 2010	Human reproduction and development ‐ sexual medicine	208	106	51.0	No	No
Kaya Mutlu 2013	Physical and rehabilitation medicine ‐ physical therapy	181	37	20.4	No	Oral only
Kearney 2012	Surgery	1005	371	36.9	No	No
Kim 1998	Orthopedics/orthopedic surgery	357	156	43.7	No	Oral only
Kim 2011	Orthopedics/orthopedic surgery ‐ spine	288	167	58.0	No	Oral only
Kinsella 2015	Orthopedics/orthopedic surgery	444	298	67.1	No	No
Kiroff 2001	Surgery	302	165	54.6	No	No
Klappenbach 2011	Surgery	200	22	11.0	No	No
Klassen 2002	Pediatrics	447	264	60.1	Yes (RCTs)	No
Kleine‐Konig 2014	Orthopedics/orthopedic surgery	646	237	36.7	No	No
Kleweno 2008	Other clinical specialities ‐ sports medicine	165	98	59.4	No	Oral only
Koene 1994	General medicine/primary care	803	385	47.9	No	No
Korn 2000	Emergency medicine	1637	629	38.4	No	No
Kottachchi 2010	Gastroenterology ‐ inflammatory bowel disease	82	64	78.0	Yes (RCTs)	No
Koçak 2014	Physical and rehabilitation medicine ‐ physical therapy	188	47	25.0	No	No
Krzyzanowska 2003	Oncology	510	415	81.4	Yes (RCTs)	No
Kumar 1995	Gastroenterology ‐ endoscopy	275	76	27.6	No	No
Kunadian 2015	Cardiology	2119	1401	66.1	No	No
Kwong 2007	Emergency medicine ‐ trauma	278	112	40.3	No	Oral only
Landry 1996	Emergency medicine ‐ trauma, burn	168	44	26.2	No	No
Larian 2001	Otolaryngology head and neck surgery	839	270	32.2	No	No
Lau 2016	Otolarygnology head and neck surgery	460	259	56.3	No	Oral only
Lee 2012	Oral health ‐ dental research	346	128	37	No	No
Leles 2006	Oral health ‐ odontology	775 (submitted)	116 (submitted)	15.0 (submitted)	No	No
Lensen 2015	Human reproduction and development ‐ subfertility	230	117	50.9	Yes (RCTs)	No
Levett 2000	Other clinical specialties ‐ health research	790	263	33.3	No	No
Li 2004	Emergency medicine	2054	781	38.0	No	No
Lim 2013	Physical and rehabilitation medicine	1027	317	30.9	No	No
Lin 2011	Cardiology ‐ arrhythmias	25	10	40.0	No	No
Liu 1996	Cardiology ‐ circulation, basic biomedical science, gastroenterology, neurology	400	141	35.3	No	No
Livas 2014	Oral health ‐ orthodontics	590	308	52.2	No	No
Lloyd 2006	Orthopedics/orthopedic surgery ‐ hip and knee surgery	292	168	57.5	No	Oral only
Loevy 1997	Oral health ‐ pediatric dentistry	189	87	46.0	No	No
Macdonald 2012	Surgery ‐ pediatric surgery	862	302	35	No	Oral only
Macmillan 2007	Emergency medicine	404	124	30.7	No	No
Maguire 2014	Human reproduction and development ‐ reproductive medicine	577	335	58	No	No
Maleck 1998a	Emergency medicine ‐ prehospital care	98	10	10.2	No	No
Maleck 1998b	Emergency medicine	109	11	10.1	No	No
Malicki 2014	Medical decision making/health policy ‐ peer review and biomedical publication	504	305	61.5	No	No
Manuck 2015	Gynecology/obstetrics ‐ maternal‐fetal medicine	3281	1780	54.3	No	No
Martinez 2008	Other non‐clinical specialties ‐ biochemistry	388	42	10.8	No	No
Marx 1999	Radiology ‐ neuroradiology	527	194	36.8	No	No
Maxwell 1981	Nursing ‐ oncology nursing	65	28	43.1	No	No
McCormick 1985	Pediatrics	355 (presented) 1238 (submitted)	172 (presented) 330 (submitted)	48.5 (presented) 26.7 (submitted)	No	No
McCue 2005	Surgery ‐ colon and rectal surgery	672	400	59.5	No	No
McKelvey 2010	Pharmacology ‐ residents	272	43	15.8	No	No
McKinley 2010	Dermatology	115	64	55.7	No	No
McLennan 2008	Gynecology/obstetricts ‐ urogynecology	92	68	73.9	No	Oral only
Meininger 2011	Anesthesiology	2207	813	36.8	No	No
Meissner 2014	Oncology ‐ radiation oncology	172	88	51.2	No	Oral only
Menditto 2015	Emergency medicine	298	43	14.4	No	No
Meral 2016	Surgery	1368	559	40.9	No	No
Meranze 1982	Anesthesiology	379	122	32.2	No	No
Micieli 2012	Ophthalmology/optometry ‐ ophthalmology	508	158	31.1	No	No
Miguel‐Dasit 2006a	Radiology	2992	464	15.5	No	No
Miguel‐Dasit 2006b	Radiology	1020	479	47.0	No	Oral only
Miguel‐Dasit 2007	Radiology	991	449	45.3	No	Oral only
Mihok 2013	Orthopedics ‐ knee surgery	602	200	33.2	No	No
Miller 2015	Endocrinology/nutrition	648	165	25.5	No	No
Mily 2008	General medicine/primary care ‐ hypertension	128	34	26.6	No	No
Mittal 2011	Pediatrics	75	28	37.3	Yes (award‐winning)	No
Moar 2013	Otolarngology head and neck surgery ‐ craniofacial medicine	318	67	21.1	No	Oral only
Montane 2007	Pharmacology	248	71	28.6	No	No
Moorthi 2013	Nephrology	73 (submitted)	39 (submitted)	53.4 (submitted)	Yes (RCTs only)	No
Morgan 2005	Dermatology	131	67	51.1	No	No
Morrison 1994	Gynecology/obstetrics	61 with >24 months follow‐up)	36 with >24 months follow‐up)	59.0 (with >24 months follow‐up)	No	No
Mowla 2006	General medicine/primary care	890	98	11.0	No	No
Muffly 2014	Gynecology/obstetrics ‐ urogynecology	438	239	54.5	No	No
Murrey 1999	Orthopedics/orthopedic surgery	764	377	49.3	No	No
Mutlu 2015	Psychiatriy ‐ pediatrics and adolescents	214	54	25.2	No	Posters only
Nader 2009	Other clinical specialties ‐ geographic medicine	125	34	27.2	No	No
Nasir 2012	Surgery ‐ pediatric surgery	128	52	40.6	No	Oral only
Nasir 2013	Surgery ‐ pediatric surgery	164	49	29.9	No	No
Ng 2004	Urology	4302	1627	37.8	No	No
Nguyen 1998	Emergency medicine ‐ orthopedic trauma	490	292	59.6	No	Oral only
Nqwena 2007	Surgery	160	47	29.4	No	Oral only
O' Connor 2015	Surgery	298	168	56.4	No	Plenary (oral) only
O'Dell 2012	Pharmacology	446	19	4.3	No	No
O'Kelly 2015	Urology	322	144	44.7	No	No
O'Neill 2014	Orthopedics/orthopedic surgery	203	66	32.5	No	No
Odunsi 2015	Neurology	627	244	38.9	No	No
Ogilvie 2014	Otolaryngology head and neck surgery	392	198	50.5	No	Oral only
Ohlsson 1999	Pediatrics ‐neonatology	107 (presented) 141 (submitted)	62 (presented) 73 (submitted)	57.9 (presented) 51.8 (submitted)	Yes (RCTs)	No
Ohtori 2013	Orthopedics/orthopedic surgery	3205	1198	37.4	No	No
Okafor 2015	Orthopedics ‐ cervical spine	321	211	65.7	No	Oral only
Olive 2004	Rheumatology	249	52	20.9	No	No
Oliveira 2009	Urology	313	122	39.0	No	Oral only
Oliver 2003	Surgery ‐ plastic surgery	370	118	31.9	No	No
Olmos‐de‐Aguilera 2013	General medicine/primary care	59	11	18.6	No	No
Olson 2012	Pharmacology ‐ residents	270	17	6.3	No	No
Ospina 2006	Emergency medicine	383	194	50.7	Yes (clinical trials and RCTs)	No
Ozel 2007	Gynecology/obstetrics	116	63	54.3	No	No
Ozkösem 2013	Other non‐clinical specialties‐ free radical biology and medicine	943	398	42.2	No	No
Ozkösem 2013a	Human reproduction and development ‐ reproductive biology	1546	703	45.5	No	No
Paes 2015	Pathology	701	63	9.0	No	No
Papagikos 2005	Oncology ‐ radiation oncology	802	452	56.4	No	Oral only
Papoutsis 2015	Cardiology	5535 (presented) 8411 (submitted)	1425 (presented) 1909 (submitted)	25.7 (presented) 22.7 (submitted)	No	No
Papp 2011	Medical education/library science	201	66	32.8	No	No
Parkar 2013	Radiology	137	62	45.3	No	No
Patel 2011	Surgery ‐ neurosurgery	3827	1243	32.5	No	No
Paulson 2011	Oncology ‐ blood and marrow transplantation	501	217	43.3	No	No
Payne 1999	Endocrinology/nutrition ‐ diabetes	186	87	46.8	No	No
Peng 2006	Otolarygnology head and neck surgery	473	237	50.1	No	No
Pereyra‐Elias 2011	Gastroenterology	417	34	8.2	No	No
Perez 2004	Rheumatology	79	63	79.7	No	No
Perron 2011	Sociology ‐ social work	121	52	43.0	No	No
Petticrew 1999	Sociology	77	39	50.6	No	Oral only
Petticrew 2008	Sociology	224	99	44.2	Yes (qualitative studies)	Oral only
Pinheiro 2009	Oncology	154	26	16.9	No	No
Polyzos 2011	Human reproduction and development ‐ reproductive medicine	155	89	57.4	Yes (RCTs)	No
Post 2013	General medicine/primary care	1329	457	34.4	No	No
Prendergast 2013	Gastroenterology	938	165	17.6	No	No
Preston 2005	Emergency medicine ‐ orthopedic trauma	815	476	58.4	No	No
Prohaska 2013	Pharmacology	1941	384	19.8	No	No
Propst 2015	Gynecology/obstetrics ‐ gynecologic surgery	867	483	55.7	No	No
Rabenda 2015	Orthopedics/orthopedic surgery ‐ musculoskeletal disorders	619	191	30.9	No	No
Ramsey 1995	Hematology ‐ transfusion medicine, blood bank	249	89	35.7	No	No
Rao 2006	Urology	449	142	31.6	No	No
Ravn 2014	Emergency medicine	63 (with >24 months follow‐up)	17 (with >24 months follow‐up)	27.0 (with >24 months follow‐up)	No	No
Richling 2014	Human reproduction and development ‐ autism and intellectual disabilities	880	77	8.8	No	No
Riessland 2004	Pharmacology	347	182	52.4	No	No
Riordan 2000	Pediatrics	88	55	62.5	No	Plenary (oral) only
Rodriguez 2012	Otolaryngology head and neck surgery ‐ oral and maxillofacial surgery	311	85	27.3	No	No
Rollin 2009	Other clinical specialties ‐occupational health	318	105	33.0	No	No
Rosmarakis 2005	Infectious disease/immunology	190	68	35.8	No	No
Roukis 2011	Surgery ‐ foot and ankle	206	139	67.5	No	Oral only
Roy 2001	Otolarygnology head and neck surgery	660	456	69.1	No	No
Rubin 2014	Gastroenterology ‐ endoscopy	847	408	48.2	No	No
Saad 2008	Oncology	154	26	16.9	No	No
Saeed 2011	Oncology ‐ blood and marrow transplantation	141 (submitted)	43 (submitted)	30.5 (submitted)	No	No
Salami 2013	Oncology ‐ pediatric hematology	184	114	62.0	No	Undefined subgroup
Saldanha 2016	Ophthalmology/optometry ‐ ophthalmology	513	230	44.8	Yes (RCTs)	No
Sanders 2001	Gastroenterology	255	178	69.8	No	No
Sanossian 2006	Neurology ‐ stroke	353	220	62.3	No	No
Sanz Arrufat 2003	Pharmacology	638	92	14.4	No	No
Sawatsky 2015	Medical education/library science	144	64	44.4	No	No
Scherer 1994	Ophthalmology/optometry ‐ ophthalmology	93	61	65.6	Yes (RCTs)	No
Schibilisky 2014	Surgery ‐ heart and lung transplantation	529 (presented) 777 (submitted)	189 (presented) 259 (submitted)	35.3 (presented) 33.3 (submitted)	No	No
Schnatz 2008	Gynecology/obstetrics ‐ menopause	661	253	38.3	No	No
Schoenfeld 2012	Orthopedics/orthopedic surgery	770	352	45.7	No	No
Scholey 2005	Oral health ‐ dentistry	546	252	46.2	No	No
Schulte 2012a	Orthopedics/orthopedic surgery ‐ orthopedic and trauma surgery	1100	392	35.6	No	No
Schulte 2012b	Orthopedics/orthopedic surgery‐ spine surgery	839	317	37.8	No	No
Schwartz 1992	Emergency medicine ‐ air medical services	312	56	17.9	No	Oral only
Seaton 1981	Other clinical specialties ‐ speech, language and communication	583	174	29.8	No	No
Seaton 1983	Other clinical specialties ‐ speech, language and communication	625	202	32.3	No	No
Sebel 2001	Anesthesiology	581	277	47.7	No	No
Secil 2005	Radiology	4413	521	11.8	No	No
Secil 2007	Radiology ‐ gastrointestinal and abdominal radiology	276	109	39.5	No	Oral only
Shelmerdine 2015	Radiology‐ pediatric radiology	363	57	15.7	No	Posters only
Siegfried 2007	Infectious disease/immunology ‐ HIV, AIDS	127	86	67.7	Yes (RCTs)	No
Simons 2007	Infectious disease/immunology ‐ allergy	113	73	64.6	No	No
Singh 2015	Pediatrics ‐ neonatology	102	14	13.7	No	No
Sinno 2011	Surgery ‐ plastic surgery	138	78	56.5	No	No
Sivan 2010	Physical and rehabilitation medicine	251	86	34.3	No	No
Smart 2013	Otolaryngology head and neck surgery ‐ oral and maxillofacial surgery	122	90	73.8	No	No
Smith 2007	Urology	1683	740	44.0	No	No
Smith 2011a	Physical and rehabilitation medicine ‐ physical therapy	823	209	25.4	No	No
Smith 2014	Medical education/library science ‐ pediatrics	138	47	34.0	No	No
Smollin 2006	Other non‐clinical specialties ‐ toxicology	237	57	24.1	No	No
Snedeker 2010	Other non‐clinical specialties ‐ food safety research	149	70	47.0	No	No
Song 2010	Pathology	4824	1725	35.8	No	No
St John 2013	Gastroenterology	327	112	34.3	No	No
Stolk 2002	Pharmacology ‐ pharmacoepidemiology	1216	319	26.2	No	No
Stranges 2015	Pharmacology	655	76	11.6	No	No
Stöcker 2009	General medicine/primary care	368	200	54.3	No	No
Sullivan 2014	Urology	227	95	41.9	No	No
Sun 2011	Ophthalmology/optometry ‐ ophthalmology	2080	881	42.4	No	No
Susarla 2015	Surgery ‐ plastic and reconstructive surgery	78	50	64.1	No	No
Tam 2008	Oncology	74	55	74.3	Yes (phase III)	No
Tam 2011	Oncology	709	643	90.7	Yes (phase III and RCTs)	No
Tambuscio 2010	Other non‐clinical specialties ‐ forensic science	623	102	16.4	No	No
Tan 2010	Emergency medicine	54	10	18.5	No	Oral only
Theman 2014	Surgery ‐ hand	719	393	54.7	No	Oral only
Tieman 2010	General medicine/primary care ‐ palliative care	1338	213	15.9	No	No
Timmer 2001	Gastroenterology ‐ pancreatology	594	341	57.4	No	No
Timmer 2002	Gastroenterology	541 (presented) 836 (submitted)	292 (presented) 392 (submitted)	54.0 (presented) 46.9 (submitted)	No	No
Todd 1997	Otolarygnology head and neck surgery	53 (presented) 118 (submitted)	35 (presented) 43 (submitted)	66.0 (presented) 36.4 (submitted)	No	No
Toma 2006	Cardiology	186	148	80.0	Yes (RCTs)	No
Toro‐Polo 2012	General medicine/primary care	532 (submitted)	52 (submitted)	9.8 (submitted)	No	No
Turpen 2010	Urology	126	79	62.7	Yes (RCTs)	No
Tyagi 2013	Anesthesiology	726	99	13.6	No	No
ul Haq 2011	Orthopedics/orthopedic surgery ‐ trauma and orthopedics	179	65	36.3	No	No
Uysal 2008	Other non‐clinical specialties ‐ clinical chemistry	900	125	13.9	No	Posters only
Uzun 2013	Other clinical specialties ‐ undersea and hyperbaric medicine	179	62	34.6	No	No
Valderrama‐Zurian 2009	Psychiatry ‐ drug addiction	689	254	36.9	No	No
Van Balken 2005	Urology	61	29	47.5	No	No
van der Steen 2004	Surgery ‐ plastic surgery	306	195	63.7	No	Oral only
Van Royen 2010	General medicine/primary care	251	113	45.0	No	No
Varghese 2011	Orthopedics/orthopedic surgery ‐ pediatrics	1191 (submitted)	598 (submitted)	50.2 (submitted)	No	No
Vecchi 2009	Psychiatry ‐ drug and alcohol addiction	581	359	61.8	Yes (clinical trials and RCTs)	No
Vo 2005	Psychiatry ‐ mental health	287	54	18.8	No	No
Vogt 2008	Other clinical specialties ‐ sports medicine	36	20	55.6	No	No
von Hardenberg 2013	Urology	732	386	52.7	No	No
Vuckovic‐Dekic 2001	Oncology	63	42	66.7	No	No
Wai 2006	Orthopedics/orthopedic surgery ‐ lumbar spine	270	146	54.1	No	Oral only
Walby 2001	Emergency medicine	207	73	35.3	No	No
Walsh 2013	Medical education/library science	449	156	34.7	No	No
Wang 1999	Orthopedics/orthopedic surgery ‐ spine	1188	517	43.5	No	No
Waters 2011	Infectious disease/immunology ‐ HIV	201	107	53.2	No	Oral only
Weale 2006	Surgery	768	246	32.0	No	No
Weitz 2005	Gastroenterology	392	100	25.5	No	No
Whitehouse 2009	Orthopedics/orthopedic surgery	1138	252	22.1	No	Oral only
Wieser 2015	Other non‐clinical specialties ‐ veterinary anesthesiology	607	384	63.3	No	No
Wilson 2015	Neurology ‐ dementia	250	97	38.8	No	No
Winnik 2012	Cardiology	1002 (submitted)	293 (submitted)	29.2 (submitted)	No	No
Wong 2006	Surgery	67 (submitted)	28 (submitted)	41.8 (submitted)	No	No
Wong 2009	Cardiology ‐ electrophysiology	790	377	47.7	No	No
Wong 2013	Cardiology ‐ electrophysiology	4848	2406	49.6	No	No
Yalcinkaya 2013	Emergency medicine ‐ orthopedics and traumatology	770	227	29.5	No	No
Yamaguchi 1990	Infectious disease/immunology ‐ bacteriology, internal medicine	599 (submitted)	488 (submitted)	81.5 (submitted)	No	No
Yentis 1993	Anesthesiology	215	108	50.2	No	No
Yilmaz 2013	Rheumatology	1732	601	34.7	No	No
Yolcu 2015	Otolaryngology head and neck surgery ‐ oral and maxillofacial surgery	1322	246	18.6	No	No
Yoo 2002	Other clinical specialties ‐ sports medicine	333	188	56.5	No	No
Yoon 2007	Surgery ‐ hip and knee	154	73	47.4	No	Posters only
Yoon 2010	Gastroenterology ‐ pancreato‐biliary	475 (submitted)	95 (submitted)	20.0 (submitted)	No	No
Yoon 2012a	Other clinical specialties ‐ sports medicine	154	76	49.4	No	Oral only
Yoon 2012b	Urology	614	183	29.8	No	No
Yoshida 2008	Surgery ‐ vascular surgery and angiology	1108	70	6.3	No	No
Yuan 2011	Oral health ‐ dental education	369	71	19.2	No	Posters only
Yun 1999	Radiology	2457	614	25.0	No	No
Zamakhshary 2006	Surgery ‐ pediatric surgery	183	118	64.5	No	No
Zaretsky 2002	Hematology	85	69	81.2	Yes (phase III trials)	No
Zheng 2007	Oral health ‐ stomatology	523	330	63.1	No	No
Özyurt 2012	Dermatology	1023	135	13.2	No	No

1. Proportion of abstracts published in full by specialty and presentation status, with numbers of contributing reports.

Reports	Total No., RR (95% CI)	Submitted No., RR (95% CI)	Presented at plenary No., RR (95% CI)	Presented orally No., RR (95% CI)	Presented as poster No., RR (95% CI)	Presented or accepted, presentation not specified No., RR (95% CI)	Other (author group or register) No., RR (95% CI)
All	N = 425, RR = 37.3 (95% CI 35.3 to 39.3)	N = 34, RR = 39.3 ((95% CI 32.3 to 46.8)	N = 4, RR = 70.8 ((95% CI 47.8 to 86.6)	N = 54, RR = 43.1 ((95% CI 37.9 to 48.6)	N = 10, RR = 24.6 ((95% CI 17.7 to 33.0)	N = 314, RR = 36.0 ((95% CI 33.7 to 38.3)	N = 9, RR = 44.1 ((95% CI 31.3 to 57.8)
Specialty		No. reports	No. reports	No. reports	No. reports	No. reports	No. reports
Anesthesiology	9, 35.7 (24.5 to 48.6)	0	0	0	0	8	1
Cardiology	14, 39.7 (30.0 to 50.4)	2	0	1	0	11	0
Dermatology	7, 36.9 (19.6 to 58.4)	0	0	1	0	5	1
Emergency medicine	24, 28.4 (21.5 to 36.5)	1	0	4	0	19	0
Endocrinology/nutrition	4, 31.7 (14.1 to 56.9)	1	0	0	0	3	0
Gastroenterology	18, 39.7 (30.1 to 50.2)	6	0	0	1	11	0
General medicine/primary care	16, 30.1 (21.5 to 40.5)	1	0	1	1	11	2
Gynecology/obstetrics	18, 53.8 (46.7 to 60.8)	0	0	5	0	12	1
Hematology	4, 46.8 (10.2 to 87.2)	1	0	0	0	3	0
Human reproduction & development	9, 43.9 (29.3 to 59.7)	0	0	0	0	8	1
Infectious disease/immunology	6, 57.1 (35.7 to 76.1)	1	0	1	0	3	1
Medical decision‐ making/health policy	5, 26.9 (7.9 to 61.0)	0	0	1	0	4	0
Medical education/library science	7, 38.6 (28.9 to 49.2)	0	0	0	0	7	0
Nephrology	2, 46.7 (5.2 to 93.3)	1	0	0	1	0	0
Neurology	6, 45.6 (18.9 to 75.1)	0	0	0	0	6	0
Nursing	2, 24.4 (0.0 to 100.0)	0	0	0	0	2	0
Oncology	21, 53.5 (41.4 to 65.2)	3	1	3	0	14	0
Ophthalmology/optometry	9, 41.0 (27.3 to 56.2)	0	0	0	0	9	0
Oral health	15, 28.5 (20.8 to 37.7)	0	0	0	2	13	0
Orthopedics/orthopedic surgery	37, 43.3 (38.9 to 47.9)	2	0	9	1	25	0
Otolaryngology head & neck surgery	15, 38.3 (28.8 to 48.8)	1	0	3	0	11	0
Pathology	3, 26.38 (2.1 to 85.4)	0	0	0	0	3	0
Pediatrics	13, 38.9 (28.5 to 50.4)	4	1	1	0	6	1
Pharmacology	14, 15.4 (10.2 to 22.6)	0	0	0	0	14	0
Physical and rehabilitation medicine	9, 23.0 (15.9 to 32.1)	0	0	1	0	8	0
Psychiatry	6, 38.3 (22.4 to 57.2)	0	0	0	1	5	0
Radiology	16, 26.7 (19.6 to 35.2)	0	0	6	1	9	0
Rheumatology	7, 34.5 (14.0 to 62.9)	1	0	0	0	6	0
Sociology	3, 45.0 (37.0 to 53.3)	0	0	2	0	1	0
Surgery	49, 39.2 (32.3 to 46.4)	4	2	8	1	34	0
Urology	27, 39.14 (34.1 to 44.4)	0	0	4	0	23	0
Other clinical specialties	17, 38.2 (29.6 to 47.6)	0	0	3	0	13	1
Other non‐clinical specialties	13, 32.4 (19.8 to 48.2)	0	0	0	1	12	0

CI: confidence interval
 RR: risk ratio

One hundred and eighty‐one reports calculated a cumulative publication proportion by time for a total of 253 meetings. Two‐hundred and sixty‐seven reports included a median or mean time to publication. Forty‐five reports examined the association between 'positive' results and publication in which 'positive' was defined as statistically significant results and 34 in which it was defined by direction of the results presented in the abstract. Fifteen reports examined the association with publication using both definitions (Akbari‐Kamrani 2008; Harris 2006; Harris 2007; Izadpanah 2014; Klassen 2002; Krzyzanowska 2003; Lensen 2015; Mily 2008; Ospina 2006; Polyzos 2011; Salami 2013; Saldanha 2016; Sinno 2011; Tam 2008; Toma 2006). Twenty‐three reports examined the association between sample size and publication, with 13 looking specifically at the proportion published with sample size above or below the median. The association between type of presentation (oral or poster) and subsequent publication was examined in 141 reports, and the type of research (clinical or basic science) was examined in 68 reports. The proportion of abstracts accepted for presentation at meetings with subsequent publication was compared with that for rejected abstracts in 22 reports. Fifty‐four reports compared the subsequent publication of randomized or controlled trials with that of other study designs. Authors of 15 reports scored the quality of abstracts and examined the association of such quality with subsequent full publication. Forty‐four reports compared the publication of abstracts from multi‐ versus single‐centered studies. However, the definition of multi‐center status varied from more than one center to more than three centers. Eight reports compared the proportion of abstracts published that had been written in the English language versus non‐English language abstracts. The association between country of origin of the abstract and publication was compared in 52 reports, defined as North America, Europe, or the rest of the world. However, not all reports included all three geographical categories.

The association between funding source and full publication was evaluated in 12 studies and the association between authors with an academic versus other affiliation in 28 reports. Six reports examined the association of gender and 10 reports the association of expected impact with full publication. Impact was defined in various ways, including award winning, late breaking, or highly clinically important. Some associations were investigated in reports that only looked at abstracts describing randomized trials. Fifteen reports examined the association of significant results and 13 of results in a positive direction with subsequent full publication. Other factors analyzed in samples of abstracts describing randomized trials were sample size in six reports; oral versus poster presentation in nine reports, multi‐center versus single‐center status in seven reports, higher versus lower quality in three reports, and industry versus other or no funding in 12 reports.

Risk of bias in included studies

We evaluated five aspects contributing to risk of bias in each report as described above in the Methods.

Methods used to sample abstracts

Almost all (416 of 425; 97.9%) reports were at a low risk of bias for the methods used to sample abstracts, including either an entire set of abstracts presented at, or submitted to, a meeting (297 reports), a set of abstracts systematically or randomly selected from all abstracts (28 reports), a specified subset of abstracts (70 reports), abstracts that covered a specific topic (27 reports), or some combination of these. Four reports had insufficient information to assess the risk of bias on this item (Amarilyo 2013; Arrive 2004; Chan 2008; Smith 2014), usually because the sample was not clearly defined. These were scored as unclear. Five reports were considered at high risk of bias in cases where the report authors collected information directly from abstract authors. In three of these instances, the response rate was less than 80% (Dicembrino 2014; Nader 2009; Yamaguchi 1990) and in one case, the report author selected which responses to use in their report (Seaton 1983). One report was considered to be at high risk of bias because there was substantial variation in the sample size as described in the report (Kunadian 2015).

Length of follow‐up between presentation at the meeting and date of the search for publications

Based on our previous results (Scherer 2007), we chose a cut‐off of 48 months as consideration for a low risk of bias related to the length of the interval between the date of the relevant meeting and the search for subsequent full publications. Typically, the date of search was reported either as the month and year of an electronic database search or the date when abstract authors were directly contacted about subsequent publications. A majority of reports (266 of 425, 62.6%) had a follow‐up of 48 or more months and so were considered at low risk of bias. The date of the search was not reported in 105 reports, in which case the risk of bias was classified as unclear. The remaining 54 reports were at high risk of bias because the length of follow‐up was less than 48 months. No report had follow‐up less than 24 months, in keeping with our inclusion criteria.

Methods used to identify full publications

Overall, 199 of 425 (46.8%) reports were classified as having a low risk of bias for identification of subsequent full publications, seven were unclear, and 219 were judged to be at high risk of bias because they searched a single database or had a response rate to abstract author queries of less than 80%. Of the 410 reports that included a search of one or more databases, 241 searched a single database, 79 searched two databases, 27 searched three, and the remainder searched four or more (range, one to 11, median 1). The majority of reports (407 of 425; 95.8%) looked for full publications by searching the US National Library of Medicine's MEDLINE database, either as PubMed or an earlier version of this database (e.g. MEDLARS, Index Medicus). The authors of 53 reports searched Embase, 59 searched Google/Google Scholar, and 27 searched the Cochrane Library or the Cochrane Central Register of Controlled Trials. Sixty‐nine reports used questionnaires sent directly to the abstract authors to obtain information about subsequent publication. Of these, 60 also searched electronic databases and were considered at low risk of bias. Of the six reports that only contacted the authors and did not search any database, two had a response rate of less than 80% (Dicembrino 2014; Seaton 1983) and one did not report the response rate (Dirk 1996). These were classified as high and unclear risk of bias, respectively. Other sources that authors used to search for publications included handsearching of journals (Castaldi 2015; Chan 2002; Chapman 2012; Dooley 2003; Evers 2000; Maxwell 1981; Schwartz 1992; Stöcker 2009; van der Steen 2004; Walby 2001), publication lists or curriculum vitae of abstract authors (Cornu 2012; de Oliveira 2012; Fede 2010; Halpern 2002; Perron 2011; Richling 2014), websites (Dressler 2015; Fernandes 2008), and 'personal contact' (Nasir 2012).

Methods used to match an abstract with a full publication

Overall, 194 of 425 (45.6%) reports were at low risk of bias, indicating that the investigators used two or more criteria to match the abstract to the publication or had contacted the abstract author. Match criteria often included names of one or more authors; 47 reports matched by the first author, and 72 by all authors. Other criteria used for matching included title (95 reports), content (85 reports), methods (70 reports), results (58 reports), sample size (42 reports), and keywords (10 reports). Almost half of the reports (211 of 425; 49.6%) did not describe any matching criteria and the risk of bias was thus unclear. The remaining 20 reports (4.7%) were judged as of high risk of bias because only a single criterion was used to match an abstract with a full publication.

Adjustment for factors possibly related to full publication

Fifty‐nine reports did not look at any factors that might be related to full publication and so were classified as not applicable for this item. Of the remaining 366 (86.1%) reports that evaluated factors associated with full publication, 99 used some kind of multivariable model to assess the impact of two or more factors on either the proportion of abstracts published or the time to publication. They were considered to have a low risk of bias. Two hundred and sixty‐one reports did not use any type of multivariable model, but used stratified analyses, simple t tests or Fisher tests to evaluate the association between individual factors and full publication. They were considered to have a high risk of bias. Six reports did not mention the analytical method used to determine the association and were classified as unclear risk of bias.

Overall risk of bias

Overall, 32 reports were scored as having a high risk of bias because three or more of the five items were classified as high risk, or two or more of four items were classified as high risk for reports where the fifth criterion was not applicable (Arrive 2001; Beker‐Acay 2015; Ciesla 2001; Czorlich 2016; de Oliveira 2012; Dicembrino 2014; Dudley 1978; Facione 2007; Ginzburg 2014; Goldman 1980; Gorman 1990; Hopewell 2015; Koene 1994; Korn 2000; Kumar 1995; Levett 2000; Lloyd 2006; Meissner 2014; Meranze 1982; O'Dell 2012; Ohtori 2013; Papagikos 2005; Parkar 2013; Rabenda 2015; Rao 2006; Riessland 2004; Rodriguez 2012; Seaton 1981; Seaton 1983; ul Haq 2011; Yamaguchi 1990; Yilmaz 2013) (Figure 2; Figure 3). Among the remaining reports, 353 (83.1% of 425 reports) had at least one criterion scored as high risk of bias. Twenty‐six reports (6.1%) had all criteria scored as low risk of bias (Bello 2013; Bergoli 2011; Callaham 1998; Castaldi 2015; Dumville 2008; Elder 1994; Elliot 2016; Eloubeidi 2001; Glick 2006; Harel 2011; Jones 2008; Livas 2014; Ospina 2006; Petticrew 2008; Richling 2014; Rollin 2009; Smart 2013; Snedeker 2010; Stolk 2002; Susarla 2015; Toma 2006; Toro‐Polo 2012; Varghese 2011; Winnik 2012; Wong 2009; Zheng 2007.)

2.

'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Effect of methods

Proportion of abstracts published as full publications

The number or proportion of results presented in abstracts and subsequently published in full was reported for all included reports (in keeping with our inclusion criteria). The proportion published within each report ranged from 2.6% to 90.7% (median 38.4%). A random‐effects meta‐analysis showed the overall proportion of abstracts published in full was 37.3% (95% CI, 35.3% to 39.3%), but with very high heterogeneity (I² = 98.5%). This represents subsequent full publication of results from 107,959 studies out of 307,028 presented at biomedical meetings. The proportion of abstracts published did not differ by whether the report was a full publication (37.0%; 95% CI, 34.8% to 39.2%), abstract (38.2%; 95% CI, 27.4% to 50.3%), or letter (38.2%; 95% CI, 27.4% to 50.3%), but heterogeneity remained high for all three report formats. The result of our sensitivity analysis excluding 32 reports at high risk of bias showed similar results.

The pooled proportion of abstracts published in full in the 26 reports that looked separately at abstracts describing randomized trials was 59.8% (95% CI, 52.1% to 67.0%). The median proportion was 55.4% (range, 11.1% to 90.7%). This represents subsequent publication of results from 3592 randomized trials out of 5763 presented at biomedical meetings. In the other seven reports that examined both randomized and controlled trials, the proportion of abstracts subsequently published in full was 47.6% (95% CI, 28.4% to 67.5%). None of these 33 reports was at high risk of bias. For the one report on abstracts of diagnostic accuracy studies (Brazzelli 2009), the proportion of published abstracts was 75.6% (95% CI, 69.0% to 82.2%); for the one on systematic reviews (Hopewell 2015), it was 53.3% (95% CI, 46.3% to 60.3%); for the one on cost‐effectiveness studies (Chan 2013), it was 39.6% (95% CI, 32.2% to 47.1%); and for the one on qualitative studies (Petticrew 2008), it was 44.2% (95% CI, 37.7% to 50.7%).

The proportion of abstracts reaching full publication for 34 reports that focused on abstracts submitted to a meeting was 39.3% (95% CI, 32.3% to 46.8%). It was similar to the pooled estimate in the 314 reports that included abstracts that had been presented at a meeting without specifying the type of presentation (plenary, oral, or poster) (36.0%; 95% CI, 33.7% to 38.3%). Nine reports obtained abstracts from a register or a specific author group (e.g. all members of a specific department) and found a slightly higher proportion of fully published abstracts (44.1%; 95% CI, 31.3% to 57.8%). For the four reports specifying type of presentation, we found that abstracts presented at plenary sessions had the highest proportion of abstracts subsequently published in full (70.8%; 95% CI, 47.8% to 86.6%). In 54 reports that only included abstracts presented orally, the proportion of abstracts reaching full publication was 43.1% (95% CI, 37.9% to 48.6%), which is higher than in the 10 reports that only included poster presentations (24.6%; 95% CI, 17.7% to 33.0%). Table 1 shows the proportion of abstracts subsequently published by biomedical specialty and the number of reports contributing to that specialty by type of presentation. The proportion of abstracts published in full by specialty ranged from 15.4% (pharmacology) to 57.1% (infectious disease/immunology).

When we examined the language of the report, we found that the proportion of abstracts subsequently published in full in the 376 reports that were written in English was 39.1% (95% CI, 37.0% to 41.2%). The proportion was, in general, lower in those reports not in English language. For the 15 reports in Spanish it was 15.6% (95% CI, 11.2% to 21.3%); seven reports in French 31.4% (95% CI, 17.8% to 49.1%); six reports in German 42.7%; (95% CI, 29.2% to 57.3%); six reports in Portuguese (9.9% (95% CI, 4.2% to 21.5%); six reports in Turkish 21.6% (95% CI, 16.0% to 28.5%); four reports in Korean 35.3% (95% CI, 13.6% to 65.4%); two reports in Dutch 47.9% (95% CI, 46.6% to 49.2%); two reports in Chinese 53.7% (95% CI, 0.3% to 99.8%); and one report in Japanese 81.5% (95% CI, 78.2% to 84.4%). We also compared the 164 reports in which the correspondence address was in native English‐speaking countries with the 161 reports where it was not. A random‐effects model showed that the proportion of abstracts published in full for reports from native English‐speaking countries was 42.7% (95% CI, 40.4% to 45.1%) as opposed to 29.0% (95% CI, 26.2% to 32.1%) for reports from non‐native English‐speaking countries. The reports from non‐native English‐speaking countries typically evaluated abstracts from national meetings, abstracts presented in national sections of international meetings (Collet 1993; Collet 1997; Collet 2006; Jara‐Tracchia 2010), or abstracts by authors from the same country or region as the report (Acevedo 2014; Arap 2014; Chan 2002; Fernandes 2008; Ha 2008; Lin 2011; Pinheiro 2009; Saad 2008; Vuckovic‐Dekic 2001).

Reports were published from 1978 to 2016, with a larger proportion of reports appearing in more recent years. We evaluated time of report publication in a mixed model and found a significant effect of time, with an intercept of ‐0.4036 (95% CI, ‐0.5283 to ‐0.2790) and a yearly decrease of ‐0.0155 (95% CI, ‐0.0275 to ‐0.0034). These values translate to publication probabilities of 43.82%, 40.04%, and 36.39% for reports published in years 1990, 2000, and 2010, respectively.

Cumulative proportion of abstracts published over time

Data on cumulative time to publication was reported in 181 reports. Using survival analysis, the estimated cumulative publication rate for all studies was 46.4% after 120 months (Figure 4). The annualized rate of publication was highest during the first three to four years following presentation at a meeting, and decreased during each subsequent year. After 10 years, the estimated publication rate for abstracts describing results of randomized or controlled trials in 18 reports was 68.7%, compared to 44.9% for abstracts describing other types of study design in 163 reports (Figure 5).

4.

Kaplan‐Meier survival curve showing proportion of abstracts published in full

5.

Kaplan‐Meier survival curve showing proportion of abstracts describing trials and all other abstracts published in full

Factors associated with publication

'Positive' versus not 'positive' results

Overall, 64 reports examined whether 'positive' results were associated with full publication. In the previous version of this review (Scherer 2007), we investigated individual authors' definition of 'positive' results, and identified two different definitions. The first definition described 'positive' results as those showing statistically significant results or a definite preference for either treatment arm compared to neutral results. Pooled results from the 15,783 abstracts in 45 reports using this definition also show an association between 'positive' results and full publication (risk ratio (RR) = 1.31; 95% CI 1.23 to 1.40), and this was unchanged after excluding one report at high risk of bias. We did not pool results because some studies examined both definitions. The second definition described 'positive' results as those showing a statistically significant result in the direction of, or a stated preference for, the experimental intervention compared with the control intervention. In this updated review, the pooled results from 8794 abstracts in 34 reports using this definition show an association between positive results and full publication (RR =1.17; 95% CI 1.07 to 1.28). See Analysis 3.1 for forest plots for both definitions. Heterogeneity was high in both meta‐analyses with an I² of 66% and 69%, respectively. No reports contributing to this meta‐analysis were at high risk of bias. One report (Mily 2008) that examined publication using both definitions was an abstract itself. A sensitivity analysis excluding it did not change the pooled estimate. An additional eight reports that examined whether the presence of positive results was associated with full publication did not provide data that could be included in a meta‐analysis (Buchan 2011; Cartwright 2007a; Dumville 2008; Moorthi 2013; Snedeker 2010; Wai 2006; Wong 2013; Zaretsky 2002). Using either univariate or multivariable analyses, three of these reported an association between full publication and positive results (Dumville 2008; Snedeker 2010; Wong 2013); three found no association (Cartwright 2007a; Moorthi 2013; Zaretsky 2002), and the remaining two simply reported that the proportion of abstracts reaching full publication was higher than that of those not reporting positive results.

3.1. Analysis.

Comparison 3 'Positive' versus 'not positive' results, Outcome 1 Publication by 'positive' versus 'not positive' results.

In 20 of the reports that looked at the influence of 'positive' results, only randomized or controlled trials were included (Analysis 3.2). We again looked at the subgroups using both definitions of 'positive' results. We found an association with full publication in 2616 abstracts in 15 reports using the first definition (RR = 1.21; 95% CI, 1.10 to 1.32) and in 2307 abstracts in 13 reports using the second definition (RR = 1.17; 95% CI, 1.04 to 1.32) (Analysis 3.2). In both meta‐analyses, heterogeneity remained moderately high (I² of 47% and 56%, respectively). Eight of the 20 reports examined the association of 'positive' results with full publication using both definitions, but did not find any striking differences (Akbari‐Kamrani 2008; Klassen 2002; Krzyzanowska 2003; Ospina 2006; Polyzos 2011; Saldanha 2016; Tam 2008; Toma 2006). None of the reports contributing to these meta‐analyses were at high risk of bias.

3.2. Analysis.

Comparison 3 'Positive' versus 'not positive' results, Outcome 2 Publication by 'positive' versus 'not positive results,' randomized or controlled clinical trials.

Sample size

In the previous version of this review (Scherer 2007), we combined the results of reports that compared the full publication of abstracts with a study sample size equal to or above the median of all presented studies with those below the median. However, we found that authors of more recent reports frequently reported on sample size using an alternative numeric threshold (e.g. less than 200 participants versus 200 or more participants). We thus broadened our definition regarding sample size to "larger versus smaller", using the report authors' definitions. The pooled results from 8689 abstracts in 23 reports showed that larger versus smaller sample size failed to be associated with full publication (RR = 1.09; 95% CI, 0.99 to 1.19; Analysis 4.1) with an I² of 65%; no reports were at high risk of bias. However, the six reports (727 abstracts) that only examined abstracts describing randomized or controlled trials showed an association of larger versus smaller sample size with full publication (RR = 1.25; 95% CI, 1.08 to 1.46; Analysis 4.2) and showed no heterogeneity (I² =0%). The five reports that examined this factor but did not report data that could be included in a meta‐analysis showed mixed results; three showing an association (DeMola 2009; Schibilisky 2014; Zaretsky 2002), one not finding an association (Sanders 2001), and one only reporting that a higher percentage of abstracts with a sample size above the median were published in full compared with those below the median (Yilmaz 2013).

4.1. Analysis.

Comparison 4 Sample size, Outcome 1 Publication by larger versus smaller sample size.

4.2. Analysis.

Comparison 4 Sample size, Outcome 2 Publication by larger versus smaller sample size; randomized or controlled clinical trials.

Oral versus poster presentation

Results from 115,910 abstracts in 143 reports showed that abstracts presented orally rather than as posters were associated with full publication (RR = 1.46; 95% CI, 1.40 to 1.52; Analysis 5.1), although heterogeneity was high with an I² statistic of 83%. Excluding 14 reports that were abstracts (Abes 2004; Cartwright 2007b; Harshavardhana 2009a; Kumar 1995; Lensen 2015; Maleck 1998a; McCue 2005; Menditto 2015; Mily 2008; Ng 2004; Odunsi 2015; Ozel 2007; Schibilisky 2014; Walsh 2013) did not change this association (R = 1.45; 95% CI, 1.39 to 1.52) or heterogeneity (I² = 83%). These results are consistent with the meta‐analyses presented above of the proportion of abstracts published in full when authors only included abstracts presented orally (43.1%) or as posters (24.6%). Excluding reports at high risk of bias changed neither the pooled estimate nor the level of heterogeneity appreciably (RR = 1.46; 95% CI, 1.40 to 1.52; I² = 82%). The association of full publication with oral (versus poster) presentation in nine reports that examined abstracts of randomized or controlled trials was slightly less (RR = 1.32, 95% CI, 1.13 to 1.54 (I² = 70%); Analysis 5.2). Of 13 reports that looked at this factor but did not report data that could be included in the meta‐analysis, seven used statistical testing and found that abstracts presented orally were more likely to be published than those presented as posters (Abuzeid 2013; Amirhamzeh 2012; Daruwalla 2015; Glick 2006; Moorthi 2013; O'Kelly 2015; Snedeker 2010, while one did not (Goldman 1980). All reports that simply reported the proportion of abstracts that went on to full publication showed that a higher proportion of abstracts that had been presented orally were published in full as compared with those presented as posters (Aggarwal 2012; Buchan 2011; Ramsey 1995; Tambuscio 2010; Tieman 2010).

5.1. Analysis.

Comparison 5 Oral versus poster presentations, Outcome 1 Publication by oral versus poster presentations.

5.2. Analysis.

Comparison 5 Oral versus poster presentations, Outcome 2 Publication by oral versus poster presentations, randomized or controlled clinical trials.

Acceptance versus rejection

Twenty‐two reports (22,319 abstracts) evaluated associations between acceptance (versus rejection) of abstracts for presentation at a meeting and full publication. The pooled results from these 22 reports showed a strong association between acceptance of abstracts for presentation at a meeting and full publication (RR = 1.65; 95% CI, 1.48 to 1.85; Analysis 6.1), but with high heterogeneity (I² = 87%). Excluding four reports which are abstracts themselves (Bernstein 1983; Jörgens 2014; Ohlsson 1999; Todd 1997) resulted in a similar estimate (RR = 1.61; CI, 1.43 to 1.82), as did excluding one report at high risk of bias (RR = 1.66; 95% CI, 1.47 to 1.86).

6.1. Analysis.

Comparison 6 Accepted abstracts versus rejected abstracts, Outcome 1 Publication by acceptance versus rejection for presentation.

Clinical versus basic science

Pooled results from 97,372 abstracts in 92 reports showed that clinical research was not published in full as often as basic science (RR = 0.78; 95% CI, 0.74 to 0.82; (Analysis 7.1), but again there was high heterogeneity (I²= 84%), possibly because report authors may have used differing definitions for clinical research or basic science. Excluding nine reports which were themselves abstracts (Bernstein 1983; Jamjoom 2014; Kaya 2010; Korn 2000; Menditto 2015; Ng 2004; Ozkösem 2013a; Schibilisky 2014; St John 2013) did not change this association (RR = 0.78; 95% CI, 0.74 to 0.82). The sensitivity analysis excluding four reports at high risk of bias also did not change this association (RR = 0.77; 95% CI, 0.73 to 0.81). Reports that investigated this factor but did not provide data that could be included in a meta‐analysis had mixed results, with three reports showing a statistically significant association (Glick 2006; Kunadian 2015; Ozkösem 2013), and three not (Al‐Qaoud 2013; Gilbert 2004; Muffly 2014). A further four reports that evaluated this factor only reported percentages of abstracts that went on to full publication with all showing that abstracts describing basic science were published in full more often than those describing clinical research (Buchan 2011; Duchini 1997; Sullivan 2014; Wai 2006).

7.1. Analysis.

Comparison 7 Clinical versus basic research, Outcome 1 Publication by clinical research versus basic science.

Study design

With regard to study design, the pooled results from 28,928 abstracts in 47 reports showed that abstracts describing randomized trials were published in full more often than abstracts describing other types of clinical research (RR = 1.51; 95% CI, 1.36 to 1.67; Analysis 8.1) with high heterogeneity (I² = 81%). This value was unchanged when four reports published as abstracts (Menditto 2015; Mily 2008; Ozkösem 2013a; Schibilisky 2014) were excluded (RR = 1.49; 95% CI, 1.34 to 1.65). This analysis did not include the reports that combined randomized and controlled trials or the three reports that compared the proportion of publication of abstracts describing randomized versus controlled trials (Akbari‐Kamrani 2008; Blackwell 2009; Vecchi 2009). Of three reports that examined this factor but without reporting data that could be used in a meta‐analysis, two found a significant association of randomized trial design with full publication compared with other study designs (Sun 2011; Wong 2013), and one found no association (Muffly 2014).

8.1. Analysis.

Comparison 8 Study design, Outcome 1 Publication by randomized trial versus other study designs.

Thirty reports (14,938 abstracts) compared the proportion of abstracts published when the study design was described as 'prospective' versus 'retrospective'. These reports showed that prospective studies are more likely to be published in full than retrospective studies (RR = 1.17; 95% CI, 1.06 to 1.30; Analysis 8.2), but with high heterogeneity (I² = 74%).There was no difference in this finding when a single report (Korn 2000) that was itself an abstract was excluded nor when three reports at high risk of bias were excluded (RR = 1.14; 95% CI, 1.02 to 1.28). One report that could not be included in the meta‐analysis but which did statistical testing showed that a higher proportion of abstracts describing prospective studies resulted in full publication compared with abstracts describing retrospective studies (Gilbert 2004), and two showed no effect (Gandhi 2004; Muffly 2014). Reports that only included the relative percentage of abstracts that were published in full had mixed results. One of these reports showed a higher proportion of abstracts describing prospective studies (Gourtaud 2009), and two showed a higher percentage of abstracts describing retrospective studies (Autorino 2006; Buchan 2011) published in full.

8.2. Analysis.

Comparison 8 Study design, Outcome 2 'Prospective' versus 'retrospective'.

As a surrogate for study design, seven authors reported the numbers of abstracts published in full by 'level of evidence' (Castagnetti 2013; Forlin 2013; Izadpanah 2014; Kiroff 2001; Kleine‐Konig 2014; Schulte 2012a; Schulte 2012b). Compared with abstract results classified as evidence levels III, IV, and V, those with results classified at lower evidence levels (I and II) had a higher proportion of abstracts leading to full publication (RR =1.50; 95% CI, 1.20 to 1.86; Analysis 8.3; I² = 82%). In one report (Sinno 2011), there was no statistically significant difference in the proportion of abstracts published in full by level of evidence (odds ratio (OR) = 0.10, 95% CI, 0.0 to 5.93).

8.3. Analysis.

Comparison 8 Study design, Outcome 3 Level of evidence.

Multi‐centered versus single center

The pooled results from 44 reports (22,780 abstracts) found that abstracts describing studies with multiple centers rather than a single center were more likely to lead to full publication (RR = 1.32; 95% CI, 1.21 to 1.44) (Analysis 9.1), but with high heterogeneity (I² = 82%). This result was unchanged when five reports that were abstracts (Brost 2005; Menditto 2015; Odunsi 2015; Ozkösem 2013a; Walsh 2013) were excluded (RR = 1.34; 95% CI, 1.24 to 1.46). It was also unchanged after excluding two reports at high risk of bias (RR = 1.33; 95% CI, 1.21 to 1.46). We found a somewhat stronger association in 1539 abstracts described in seven reports that examined publication of abstracts describing randomized or controlled trials (RR = 1.47; 95% CI, 1.09 to 1.98; Analysis 9.2) but again with high heterogeneity (I² = 88%). Five reports presented results of statistical testing that could not be included in the meta‐analysis, with two reports showing an association between multi‐center status and full publication (Fesperman 2008; Zaretsky 2002) and three showing no association (Balasubramanian 2006; Cartwright 2007a; Moorthi 2013).

9.1. Analysis.

Comparison 9 Multi‐centered versus single center, Outcome 1 Multi‐centered versus single center.

9.2. Analysis.

Comparison 9 Multi‐centered versus single center, Outcome 2 Multi‐centered versus single center, randomized or controlled clinical trials.

Higher versus lower 'quality'

Twelve reports (3364 abstracts) scored abstracts by quality and compared the proportion achieving full publication in those having higher versus lower quality. Definitions or methods to determine quality varied across reports from those that used some kind of scale (Callaham 1998; Castaldi 2015; Chalmers 1990a; Jörgens 2014; Menditto 2015; Sawatsky 2015; Scherer 1994; Timmer 2002), to others that used a single methodological criterion for assessment (e.g. blinding or allocation concealment) (Brazzelli 2009; Herbison 2004; Mily 2008; Sinno 2011). Pooling all results, we found that higher (versus lower) quality abstracts were more likely to be fully published (RR = 1.46; 95% CI, 1.23 to 1.73; Analysis 10.1), but had high heterogeneity (I²=72%). Removing three reports (Jörgens 2014; Menditto 2015; Mily 2008) that were abstracts attenuated this association (RR = 1.29; 95% CI, 1.14 to 1.47). There was no association between quality and the probability of publication in three reports Chalmers 1990a; Herbison 2004; Scherer 1994) that examined abstracts describing randomized or controlled trials (RR = 1.09; 95% CI, 0.84 to 1.42; Analysis 10.2) and this analysis had an I² of 0%.

10.1. Analysis.

Comparison 10 Higher versus lower quality, Outcome 1 Higher versus lower quality.

10.2. Analysis.

Comparison 10 Higher versus lower quality, Outcome 2 Higher versus lower quality, randomized or controlled clinical trials.

Abstract author characteristics

Thirty‐four reports (16,913 abstracts) examined the association of abstracts prepared at a university or academic center versus those prepared elsewhere and found an association with full publication (RR =1.60; 95% CI, 1.34 to 1.92; Analysis 11.1) with high heterogeneity (I² = 90%). Seven of these reports were abstracts themselves (Feldman 2015; Jamjoom 2014; Kaya 2010; Menditto 2015; Ozel 2007; Schibilisky 2014; St John 2013), and excluding them from the analyses decreased the effect slightly (RR = 1.48; 95% CI, 1.21 to 1.88), but excluding two reports at high risk of bias did not change the association (RR = 1.58; 95% CI, 1.31 to 1.89). Five other reports evaluated this factor statistically but did not present data that could be included in a meta‐analysis. Among these five, three found no association (Gandhi 2004; Gilbert 2004; Muffly 2014), and two showed that a higher percentage of abstracts of academic origin went on to full publication compared with abstracts originating elsewhere (Buchan 2011; Montane 2007).

11.1. Analysis.

Comparison 11 Abstract author characteristics, Outcome 1 Academic center versus non‐academic center.

Six reports (10,935 abstracts) (Canosa 2011; Castaldi 2015; Cohen 2012a; Fesperman 2008; Papoutsis 2015; Winnik 2012) examined the proportion of abstracts published in full when the primary author was female (as compared with male). Pooling the data showed no association with full publication (RR = 0.89; 95% CI, 0.77 to 1.03; Analysis 11.2), but again had high heterogeneity (I² = 62%). All of these reports were full‐length articles and none were at high risk of bias. Other characteristics of abstract authors examined in these reports included professional status, either comparing trainees (e.g. residents) and faculty or some comparison of faculty rank (Gavazza 1996; Nasir 2012; Susarla 2015), or whether a physician (McKelvey 2010), individual holding a PhD degree (Elliot 2016), or member of the society organizing the meeting (Bolac 2009) was an author. Five reports evaluated the number of abstract authors, with most finding a higher proportion of abstracts published in full when there was a higher number of authors (Buchan 2011; Grzeskowiak 2014; Kim 1998; Micieli 2012; Uzun 2013). Lastly, one report showed no difference between abstracts with an author who disclosed at least one conflict of interest versus those with no author with a disclosed conflict of interest (Saldanha 2016).

11.2. Analysis.

Comparison 11 Abstract author characteristics, Outcome 2 Female versus male.

Abstract impact

Eleven reports classified abstracts as 'high versus low impact' and investigated the association of this distinction with full publication. Abstracts that were considered to have a high impact were placed into three subgroups: 'award‐winning' (Ensom 1998; Fede 2010; Frost 2015; Kalyoncu 2011; Manuck 2015; Okafor 2015; Peng 2006); considered to be of significant clinical importance by the report authors (Klassen 2002), and those that were high impact for another reason (e.g. assessed at the meeting as 'top ranking' (Burden 2014), having a 'high profile' (Simons 2007) or late‐breaking (Toma 2006)). In these 11 reports, of 6982 abstracts, those considered as having a high impact were more likely to lead to full publication compared with those not considered to have a high impact (RR = 1.60; 95% CI, 1.41 to 1.82; Analysis 12.1). These findings had high heterogeneity (I² = 74%), but were consistent across all three subgroups. The only exception was a single study in the subgroup defined by clinical importance, which was not statistically significant. The overall association was only slightly weaker when one report (Burden 2014) that was an abstract was not included in the meta‐analysis (RR = 1.52; 95% CI, 1.37 to 1.68).

12.1. Analysis.

Comparison 12 Impact, Outcome 1 High versus low importance or impact.

Funding

Thirteen reports (5832 abstracts) examined the association of funding with full publication. The proportion of abstracts going on to be published in full was higher for those with funding of any type reported compared with those without funding or with funding not reported (RR = 1.48; 95% CI, 1.27 to 1.73; Analysis 13.1; I² = 77%). Abstracts whose authors reported funding from industry (versus other sources of funding or no funding) showed a possible association with full publication in 12 reports (3938 abstracts) (RR = 1.18; 95% CI, 1.00 to 1.40; Analysis 13.2; I² = 80%). None of the reports contributing to these analyses were themselves abstracts. In contrast to the findings of our meta‐analysis, a statistically significant association between industry funding and full publication was reported in two reports whose data could not be included in the meta‐analysis (Glick 2006; Moorthi 2013).

13.1. Analysis.

Comparison 13 Funding, Outcome 1 Funding reported versus no funding or funding not reported.

13.2. Analysis.

Comparison 13 Funding, Outcome 2 Industry funding versus other type or no funding.

North America versus Europe versus other origin

Some reports examined the full publication of abstracts by country of origin, comparing those originating in either North America or Europe versus those from the rest of the world. Using two‐way comparisons, there was an association between full publication and country of origin. For 19,984 abstracts described in 52 reports, those from North America were more likely to be published than those from the rest of the world (RR = 1.13; 95% CI, 1.04 to 1.22; Analysis 14.1). Among 21,281 abstracts described in 44 reports, those from Europe were more likely to be published when compared with the rest of the world (RR = 1.24; 95% CI, 1.11 to 1.39; Analysis 14.2). There was no association of full publication with country of origin when comparing North America versus Europe (RR = 1.04; 95% CI, 0.97 to 1.13; Analysis 14.3) in 41 reports (20,669 abstracts). These results showed high heterogeneity (I² = 70%, 83%, and 72%, respectively). Results were unchanged when reports that were abstracts themselves were not included in each of the analyses or when reports at high risk of bias were excluded.

14.1. Analysis.

Comparison 14 North America versus Europe versus other origin, Outcome 1 North America versus Other.

14.2. Analysis.

Comparison 14 North America versus Europe versus other origin, Outcome 2 Europe versus Other.

14.3. Analysis.

Comparison 14 North America versus Europe versus other origin, Outcome 3 North America versus Europe.

Some reports looked at the full publication of abstracts originating in a specific country versus elsewhere. Twenty‐four reports compared the full publication of 45,931 abstracts by origin in the USA versus elsewhere, finding a small effect (RR = 1.05; 95% CI, 1.01 to 1.09; Analysis 14.4; I² = 84%). Other reports that compared national versus other origin of the abstracts included the UK (RR = 0.90; 95% CI, 0.67 to 1.21; Analysis 14.5; three reports, 1070 abstracts, I² = 5%), Australia/New Zealand (RR = 0.67; 95% CI, 0.54 to 0.84; Analysis 14.6; two reports,1262 abstracts, I² = 0%), and Brazil (RR = 1.26; 95% CI, 0.22 to 7.02;Analysis 14.7; two reports, 966 abstracts, I² = 70%). Single studies compared the subsequent publication of results from abstracts originating in France (Cauchy 2014), India (Dhaliwal 2008), Saudi Arabia (Jamjoom 2014), Quebec (Al‐Qaoud 2013), Ireland (O' Connor 2015) or Spain (Hernandez‐Garcia 2011) versus elsewhere with mixed results. Results from 4825 abstracts in eight reports sought, but did not find, an association between English language versus non‐English language abstracts and full publication (RR = 1.16; 95% CI, 0.96 to 1.41; Analysis 15.1), with high heterogeneity (I² = 80%). None of these reports were abstracts or at high risk of bias. The meetings covered in these reports were all either international or European based, apart from one meeting held in Malaysia (Chan 2002).

14.4. Analysis.

Comparison 14 North America versus Europe versus other origin, Outcome 4 USA versus Other.

14.5. Analysis.

Comparison 14 North America versus Europe versus other origin, Outcome 5 UK versus Other.

14.6. Analysis.

Comparison 14 North America versus Europe versus other origin, Outcome 6 Australia/New Zealand versus Other.

14.7. Analysis.

Comparison 14 North America versus Europe versus other origin, Outcome 7 Brazil versus Other.

15.1. Analysis.

Comparison 15 English language vs non‐English language, Outcome 1 English language versus non‐English language.

Multivariable models

To evaluate whether any characteristics of the reports influenced the estimated proportion of subsequent full publication, we conducted a mixed‐effect multivariable analysis. Covariates included the publication year of the report, report in English language or not; report from a native English‐speaking country (Australia, Canada, New Zealand, the UK and the USA) or not, type of abstract sample (oral versus poster versus submitted versus accepted versus other); and funding reported or not. The intercept of the model was set to year 2000, report not in English and not from a native‐English‐speaking country, sample of orally presented abstracts, and no reported funding. Type of sample was removed from the model due to uninterpretable results. Covariables that remained significant in the final model included English language (versus not) (0.33; 95% CI, 0.04 to 0.61); and native English‐speaking country (0.47; 95% CI, 0.28 to 0.66). For this model, the intercept corresponded to a proportion of abstracts published in full of 31.9%. The probability of full publication for the other characteristics was 40.5% for English language reports, 42.9% for reports from native English‐speaking countries, and 52.2% when both were included. Evaluation of 'type of sample' using a likelihood‐ratio test showed that it also had an influence on the probability of publication ( P = 0.00002).

Discussion

Summary of main results

Overall proportion of abstracts published in full

We found that less than half of all studies first presented as abstracts were published in full following presentation at meetings or publication as a summary report. The estimated proportion in a random‐effects model was 37.3% (95% CI, 35.3% to 39.3%) and it was 46.4% at 120 months using survival analyses. This may represent an underestimate for two reasons. Firstly, some reports had a short follow‐up time of between two and three years and full publication may have occurred later. Secondly, most reports found subsequent publications only by searching electronic databases, and may have missed publications that are not included in these databases. The survival analysis takes into account the first limitation and indicates a somewhat higher publication rate. Abstract results that are published after the stated length of the follow‐up cannot be considered when calculating an average proportion, but they are considered when using a time‐to‐event analysis and so the survival analyses most likely presents a more accurate assessment of the true proportion of abstracts that are published in full. In determining time to publication, none of the reports mentioned whether they considered online publication prior to final journal publication in their consideration of this. Although online publication could shorten the interval between presentation at a meeting and the time to full publication, it would not necessarily change the final proportion of abstracts that are published in full. Given the shift from print to online journals and the increase in the number of journals overall, it is somewhat unexpected that we found a proportion of abstracts published in full that is so similar to that in the previous version of this review (Scherer 2007). However, the main reason given by abstract authors for not publishing the results presented as abstracts in full continues to be lack of time (Scherer 2015a), and if the time that authors have available for writing papers has not increased, this finding is not so surprising regardless of the increase in the number of journals over the past decade.

The survival analysis shows that a somewhat larger proportion of studies describing randomized or controlled trials are published in full than other study designs, with an estimated proportion of full publication at 108 months of 68.5% and at 120 months of 68.7%, compared with 44.9% for other types of studies. The proportion of randomized or controlled trials subsequently published in full is slightly higher than we found in the previous version of this review (68.5% versus 63.1% at 108 months) (Scherer 2007) suggesting that trial authors are more aware of the need to publish trial results or are taking more advantage of easier access to publishing (e.g. in electronic journals).

At least when patients are involved in the research, this under‐reporting constitutes scientific misconduct (Antes 2003; Chalmers 1990b). Many trial participants give consent to the potential harms and inconveniences involved in an experimental study under the assumption that they are making a contribution to science. If the study they took part in remains unpublished, their contribution may be for naught. In addition, those who rely on the scientific literature to make healthcare decisions are faced with a possibly biased subset of scientific evidence. However, recent studies question whether the inclusion of 'grey' literature or unpublished study results in a systematic review has the potential to change the treatment estimate (Mayo‐Wilson 2017; Schmucker 2017).

In this updated review, we used data from 252 meetings in 181 reports in the survival analysis to estimate the proportion of fully published abstracts at time intervals ranging from two to 10 years following presentation. Examination of the time course confirms our earlier findings that the highest annualized rates of publication are in the first three to four years following presentation at a meeting (Scherer 2007). We did not update the cumulative proportion of presented abstracts that we had presented in the earlier version of this review and elsewhere (Scherer 1994; Scherer 2007; von Elm 2003a) because of the large variation in the lengths of follow‐up among reports.

Association of factors with full publication

We found evidence for publication bias in the studies reported as abstracts, in that 'positive' results were associated with full publication, no matter how 'positive' was defined. In the context of studies of the effects of interventions, some authors defined 'positive' results as those showing that the experimental treatment was better than the control treatment, while others defined 'positive' results as those showing a statistically significant difference or a preference for either treatment arm. Only two reports defined 'positive' results for non‐inferiority or equivalence trials (Cheng 1998; Klassen 2002), categorizing study results as 'positive' if the study hypothesis had been supported. Some heterogeneity persisted when results were divided into the subgroups categorized by type of definition for 'positive'. There was no clear difference among reports within the subgroups to identify the source of the heterogeneity. However, the significant association between direction of results and subsequent full publication using either definition indicates that this is a robust finding. Bias favoring positive outcomes is also supported by the findings in unpublished studies (Schmucker 2014), and that significant (i.e. 'positive') results are published sooner than non‐significant results (Hopewell 2007; Ioannidis 1998; Stern 1997). Twenty reports that looked only at abstracts describing randomized or controlled trials also found an association of 'positive' results to publication. Because this association is present in this particular subgroup, authors of systematic reviews of the effects of interventions should be especially aware of this potential for bias.

Studies with larger sample sizes had appeared more likely to be published in full in the previous version of this review (Scherer 2007). In this update, rather than classifying sample size in groups by median, we used the report authors’ definition. We still found a significant association of larger sample size with full publication. It is not possible to isolate the association of sample size from the association of significant results with full publication: significant results are published more often; and studies with small sample sizes are less likely to achieve statistical significance. In this situation, lack of publication may be due to lack of significant results and not necessarily small sample size. In all cases, studies with small sample sizes should still be published because, through a meta‐analysis, they can contribute to pooled estimates in proportion to their size.

Results from clinical research studies presented in abstracts were published in full less often than results from basic science. This finding seems inconsistent with the observation that abstracts describing randomized trials were published more often than abstracts of other study designs, but there are several possible reasons for this apparent discrepancy. First, we used the report authors' definitions of 'clinical research' versus 'basic science'. So, the group of 'clinical' abstracts most likely includes, besides randomized trials, studies with many different types of study design that are often not published, a suggestion supported by the apparent heterogeneity of the meta‐analysis. Second, some 'clinical research' abstracts may describe protocols for studies that may or may not be eventually completed and would not necessarily be expected to be subsequently published. Study discontinuation is prevalent in clinical research, mostly due to problems with recruitment of participants (Kasenda 2014). Third, it may be that less time and effort is needed to complete the process to publish basic science in full.

Acceptance of an abstract for presentation at a meeting (compared to those that were submitted but not accepted) was strongly associated with full publication (RR = 1.65; 95% CI 1.48 to 1.85). Abstract authors may believe that rejected abstracts are also more likely to be rejected for subsequent publication due to implied lack of quality or disinterest by peers. This may represent a disincentive to prepare a manuscript for publication. The suggestion we had found earlier (Scherer 2007) towards an association between full publication of abstracts with 'high' compared with 'low' quality scores is significant in the present update (RR = 1.62; 95% CI, 1.27 to 2.05). However, when we looked only at full publication of abstracts describing randomized or controlled trials, we found no clear evidence of an association with publication (RR = 1.09; 95% CI 0.84 to 1.42). It should be noted that report authors used a variety of methods to gauge abstract quality, so these results may not be robust. Similarly, report authors classified abstracts as having a 'high impact' using different criteria and this was also associated with publication.

Among abstracts accepted for presentation, study results presented orally appear more likely to be subsequently published in full than results presented in poster sessions (RR = 1.46; 95% CI, 1.40 to 1.52). Four reports also looked at the proportion of abstracts presented in plenary sessions that were published in full, finding a high proportion of abstracts being published (70.3%; 95% CI, 47.8% to 86.6%). Typically, there is a perception on the part of authors that research selected for oral presentation is more interesting or of higher quality and thus is more likely to be accepted for full publication subsequently. The most common reason for non‐publication elicited by surveys of authors of non‐published abstracts is lack of time or low priority (Dickersin 1992; Easterbrook 1991; Scherer 2015a; Weber 1998). We also found a significant association between oral presentation and full publication for randomized or controlled trials (RR = 1.26; 95% CI, 1.10 to 1.45), in contrast to the lack of an effect reported in the previous version of this review (Scherer 2007).

Some report authors examined the influence of language and country of origin on full publication of abstract results. Possible influences include the relation between language proficiency and publication, since such a high proportion of scientific journals are published in North America and Europe. It is also possible that full publication of abstracts from other countries may be higher in journals that are not indexed in MEDLINE, Embase and other electronic databases used by report authors for ascertainment of subsequent full publication. When we examined the combined results however, we did not find an association between abstract language (English versus other) and full publication. However, the meetings that were examined in the reports that specifically looked at this factor were either international or European, and arguably the non‐English speaking presenters were more fluent in English than persons not presenting at that meeting. Furthermore, when we looked at language of the report of the methodology study or the correspondence address for those reports in a separate mixed multivariable model, we found a strong association between English language and the estimated proportion of full publication. Possible explanations are that the authors of these reports found it easier to identify the subsequent publications (e.g. due to their language proficiency), or that their samples had a higher proportion of English language abstracts. The association of language of the report with full publication of the abstracts is also reflected in our initial finding (in the previous version of this review (Scherer 2007)) of a significant effect of year of report with full abstract publication, although this report‐related characteristic was not significant in the multivariable model. Most likely, this is because we have now included more reports from non‐native English‐speaking countries. Together, these results are evidence of language bias; that is, research conducted in non‐native English‐speaking countries is published less frequently than that conducted in English‐language countries. Others have found language bias in reporting of randomized trials (Song 2010a) and systematic reviews (Juni 2002). Frequently, the authors of the reports from non‐native English‐speaking countries studied abstracts from national meetings, which predominantly presented studies done in that country. With the lower proportion of these studies subsequently being published in full, one could view language bias not only as bias due to the language of the report, but also due to the location where a study is conducted. This suggests that studies conducted in non‐native English‐speaking countries might not be available to systematic reviewers who are seeking to answer questions that would be relevant to settings in these same countries.

Report authors infrequently evaluated funding source. This is possibly due to the fact that, in general, abstracts do not provide information about funding. Reports that examined funding were not consistent in the approach or definitions used. Overall, while funding itself appears to be associated with subsequent full publication, the effect of funding source is less clear and we found no evidence for an association of industry funding with subsequent full publication.

Overall completeness and applicability of evidence

With 425 reports, we believe that further reports will not add substantially to the body of information presented here, which, in part, influenced the decision not to update our February 2016 searches. Although there are differences by medical specialty, the results are generally applicable to all fields of biomedical science.

Quality of the evidence

We have confidence in the findings presented here in that only 7.5% (32/425) of reports were classified as having high risk of bias. Although 83.1% had a high risk of bias for a single criterion, this failing was usually due to the lack of a multivariable model in the analysis rather than the methods used to identify and match abstracts with full publications. All reports directly addressed the question asking what proportion of abstracts presented at meetings are subsequently published in full.

Potential biases in the review process

When pooling the results of reports looking at factors associated with full publication, we found substantial statistical heterogeneity with I² values exceeding 50% in 22 of 24 analyses. Reasons for heterogeneity may be methodological differences between studies with regard to definition of factors, ascertainment of full publication, study design, and type of abstracts evaluated. To our knowledge, there is no commonly agreed standard methodology for this type of investigation. We were not able to identify any consistent factors that contributed to the heterogeneity and so the association of publication with factors based on pooled results needs to be interpreted cautiously. To account for the differences in the included reports, we used a random‐effects model rather than a fixed‐effect model for all analyses.

Access to unpublished studies

The promise of prospective registration of randomized trials as a means to alleviate many of the problems created by failure to publish study results (Dickersin 2003) has not materialized so far. Complete trial registration, including both initial registration and reporting of trial results, would enable authors of systematic reviews and consumers alike to have access to information about completed or ongoing trials, including their published and unpublished results. However, adherence to trial registration is less than optimal and results are still missing from the registry entries of many trials (Dickersin 2012; Zarin 2017). In recent studies, the proportion of published trials in specific journals ranges from 25% (Rongen 2016) to almost 100% (Harriman 2016). In 2012, the median proportion for registered trials supporting the Food and Drug Administration (FDA) approval of 15 drugs was 57% (interquartile range (IQR) (32% to 83%) (Miller 2015a). In addition, reports show that compliance with posting results in ClinicalTrials.gov is neither timely nor complete (Anderson 2015; Zarin 2015), notwithstanding the requirement to post results by the FDA. Nevertheless, trial registration still provides promise for access to more complete information on unpublished trials than many other sources.

Given that there is evidence to suggest that some trial results cannot reliably be included in systematic reviews of the effects of interventions because of missing data or missing information about study conduct (Mayo‐Wilson 2017), it is even more important to stress the importance of full publication of all relevant study information. With regard to industry‐funded research, good publication practice recommendations state that "Companies should endeavour to publish the results from all of their clinical trials of marketed products. These publications should present the results of the research accurately, objectively, and in a balanced fashion". (http://www.ich.org/products/guidelines/efficacy/article/efficacy‐guidelines.html). The same is true of results from research using any study design, whether in industry or academia, or from a high‐, middle‐, or low‐income country. The large proportion of abstract results that are not published in full is another egregious example of research waste, as described by Glasziou and colleagues (Glasziou 2014). Fully‐published results are essential to provide the evidence needed to inform clinical practice and to reduce the prevailing waste of the limited resources available for biomedical research.

Authors' conclusions

Implication for methodological research.

Investigators who wish to add to the vast literature that has already been produced examining full publication of results initially presented in abstracts should follow several minimal requirements to ensure methodological quality, including: follow‐up time of at least 24 months; searches of more than one database as well as author contact to determine subsequent publication; and use of more than a single criterion to match an abstract with retrieved journal articles. Given the abundance of already available data, the added value of new studies of this kind is questionable, except perhaps for investigators wishing to look at full publication to inform a specific medical subspecialty or specific meeting. When studying factors associated with full publication, investigators should explicitly define 'positive' results and describe how the definition applies to non‐inferiority trials; define 'clinical research' and 'basic science' and examine the impact of funding source on subsequent full publication. The low proportion of results from non‐native English‐speaking countries that are published after meeting presentation needs to be examined further, by classifying the location of the meetings as international, national, or regional. Methodological differences between reports and the existence of heterogeneity need to be addressed in further systematic reviews of this kind. The robustness of the results of this systematic review should be examined further, especially when focusing on reports that look at randomized trials.

What's new

Date	Event	Description
19 July 2018	New citation required and conclusions have changed	As well as the six‐fold increase in the number of included studies, the updated analyses have led to several new findings. For instance, a multivariable model taking into account report‐related characteristics showed an association between English language and native English‐speaking country and the full publication. The type of sample also was significantly associated with full publication, as shown by a log likelihood test.
19 July 2018	New search has been performed	Four additional authors have joined the review team: Joerg Meerpohl, Christine Schmucker, Nadine Pfeifer and Guido Schwarzer. A revised search strategy was used to search MEDLINE. New searches were designed and run for Embase and the Cochrane Library. The new search in February 2016 led to the identification and evaluation of 346 unique reports, all of which have been added to this updated review. Fifty‐four reports provided information used to calculate the proportion of abstracts describing randomized or controlled trials that were subsequently published and to examine association of study characteristics to full publication. Other characteristics examined for an association with subsequent full publication include 'prospective' versus 'retrospective' design, level of evidence, academic origin versus non‐academic origin, female versus male abstract author and high versus low impact.

Appendices

Appendix 1. MEDLINE search strategy

#	Searches
1	exp Publishing/sn
2	*publishing/
3	publication bias/
4	selection bias/
5	exp manuscripts as topic/
6	((data or finding? or information or evidence or study or studies or trial? or paper? or article? or report* or literature or work or manuscript? or abstract* or result?) adj6 (unpublish* or un‐publish* or unreport* or un‐report* or nonpublish* or non‐publish* or nonpublicat* or non‐publicat* or (publication? adj3 rate?) or "not publish*")).ti,ab.
7	(underreport* or under‐report* or selective report* or selective publish* or selective publicat* or (final* adj2 (report* or publish* or publicat* or manuscript? or paper? or article?)) or (full? adj2 (report* or publish* or publicat* or manuscript? or paper? or article?)) or (subsequent* adj2 (report* or article? or paper? or publi* or manuscript?)) or (sub‐sequent* adj2 (report? or article? or paper? or publi* or manuscript?)) or (complete* adj2 (report* or article? or paper? or publish* or publicat* or manuscript?))).ti,ab.
8	(bias* adj3 (publish* or publicat*)).ti,ab.
9	or/1‐8
15	exp animals/ not humans/
26	exp congresses/
27	exp Congresses as Topic/
28	"Abstracting and Indexing as Topic"/
29	abstract?.ti.
30	((abstract? or proceeding? or poster? or presentation? or presented or presenting or manuscript?) adj6 (congress* or conference? or meeting? or symposi*)).ti,ab.
31	(fate and (abstract? or proceeding? or poster? or presentation? or manuscript?)).ti,ab.
32	or/26‐31
33	9 and 32
34	33 not 15
35	remove duplicates from 34

Appendix 2. Embase search strategy

Embase (OvidSP)

1	publishing/
2	publication/
3	((data or finding* or information or evidence or study or studies or trial* or paper* or article* or report* or literature or work or manuscript* or abstract* or result*) adj6 (unpublish* or un‐publish* or unreport* or un‐report* or nonpublish* or non‐publish* or nonpublicat* or non‐publicat* or (publication* adj3 rate*) or "not publish*")).ti,ab.
4	(underreport* or under‐report* or selective report* or selective publish* or selective publicat* or (final* adj2 (report* or publish* or publicat* or manuscript* or paper* or article*)) or (full* adj2 (report* or publish* or publicat* or manuscript* or paper* or article*)) or (subsequent* adj2 (report* or article* or paper* or public* or manuscript*)) or (sub‐sequent* adj2 (report* or article* or paper* or public* or manuscript*)) or (complete* adj2 (report* or article* or paper* or publish* or publicat* or manuscript*))).ti,ab.
5	(bias* adj3 (publish* or publicat*)).ti,ab.
6	or/1‐5
7	Abstract*.ti.
8	((abstract* or proceeding* or poster* or presentation* or presented or presenting or manuscript*) adj6 (congress* or conference* or meeting* or symposi*)).ti,ab.
9	(fate and (abstract* or proceeding* or poster* or presentation* or manuscript*)).ti,ab.
10	or/7‐9
11	6 and 10
12	limit 11 to embase
13	remove duplicates from 12

Appendix 3. The Cochrane Library search strategy

ID	Search
#1	MeSH descriptor Publishing, this term only
#2	MeSH descriptor Publication Bias, this term only
#3	MeSH descriptor Editorial Policies, this term only
#4	MeSH descriptor Peer Review, Research, this term only
#5	MeSH descriptor Journalism explode all trees
#6	MeSH descriptor Selection Bias, this term only
#7	MeSH descriptor Manuscripts as Topic explode all trees
#8	(unpublish* or (un next publish*) or unreport* or (un next report*) or nonpublish* or (non next publish*) or nonpublicat* or (non next publicat*)):ti,ab,kw
#9	(underreport* or under next report* or selective next report* or selective next publish* or selective next publicat* or publication* near rate* or (final* near (report* or publish* or publicat* or manuscript* or paper* or article*)) or (full* near (report* or publish* or publicat* or manuscript* or paper* or article*)) or (subsequent* near (report* or article* or paper* or publi* or manuscript*)) or ((sub next sequent*) near (report* or article* or paper* or publi* or manuscript*)) or (complete* near (report* or article* or paper* or publish* or publicat* or manuscript*))):ti,ab,kw
#10	(bias* near (publish* or publicat*)):ti,ab,kw
#11	(#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10)
#22	((abstract* or proceeding* or poster* or presentation* or presented or presenting or manuscript*) near (congress* or conference* or meeting* or fate*)):ti,ab,kw
#23	abstract*:ti
#24	MeSH descriptor Congresses as Topic explode all trees
#25	MeSH descriptor Abstracting and Indexing as Topic, this term only
#26	(#22 OR #23 OR #24 OR #25)
#27	(#11 AND #26)
#37	Additional Search from Keyword List for Method Studies: "Study identification" next "publication bias"
#38	Additional Search from Keyword List for Method Studies: "Data collection" next "unpublished data"
#86	hm next method
#89	(#27 AND #86)

Appendix 4. Web of Science

Web of Science

Databases and date ranges:

‐ Science Citation Index Expanded (SCI‐EXPANDED) 1945‐present

‐ Social Sciences Citation Index (SSCI) 1975‐present

‐ Arts & Humanities Citation Index (A&HCI) 1975‐present

‐ Conference Proceedings Citation Index‐ Science (CPCI‐S) 1990‐present

‐ Conference Proceedings Citation Index‐ Social Science & Humanities (CPCI‐SSH) 1990‐present

Timespan=All Years

Lemmatization=Off

Set
# 1	ts=(unpublish* or "un‐publish*" or unreport* or "un‐report*" or nonpublish* or "non‐publish*" or nonpublicat* or "non‐publicat*" or (publication$ near/3 rate$) or "not publish*")
# 2	ts=(underreport* or "under‐report*" or "selective report*" or "selective publish*" or "selective publicat*" or (final* near/2 (report* or publish* or publicat* or manuscript$ or paper$ or article$)) or (full$ near/2 (report* or publish* or publicat* or manuscript$ or paper$ or article$)) or (subsequent* near/2 (report* or article$ or paper$ or publi* or manuscript$)) or ("sub‐sequent*" near/2 (report$ or article$ or paper$ or publi* or manuscript$)) or (complete* near/2 (report* or article$ or paper$ or publish* or publicat* or manuscript$)))
# 3	ts=(bias* near/3 (publish* or publicat* or selection))
# 4	#1 or #2 or #3
# 5	ts=((abstract$ or proceeding$ or poster$ or presentation$ or presented or presenting or manuscript$) near/6 (congress* or conference$ or meeting$ or symposi* or fate$))
# 6	ti=(abstract$ or proceeding$ or poster$ or congress* or conference$ or meeting$ or symposi*)
# 7	#16 OR #15
# 8	#4 and #17

Data and analyses

Comparison 1. Proportion of abstracts published.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Overall proportion of abstracts published			Other data	No numeric data

Comparison 2. Time to publication.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cumulative percent publication by month			Other data	No numeric data
2 Mean or median time to publication			Other data	No numeric data

2.1. Analysis.

Comparison 2 Time to publication, Outcome 1 Cumulative percent publication by month.

Cumulative percent publication by month
Study	% published at 6 months	% published at 12 months	% published at 18 months	% published at 24 months	% published at 30 months	% published at 36 months	% published at 42 months	% published at 48 months	% published at 54 months	% published at 60 months
Akbari‐Kamrani 2008	12.1	17.2		27.3				40.0
Al‐Qaoud 2013	5.9	17.1	28.9	37.6	43.0	48.1	51.0	54.9	56.0	56.9
Allart 2013	2.2	5.2	7.8	12.6	16	17.7	20.3	21.2
Allart 2015	5.3	9.1	12.2	15.4	17.6	19.5	21.7
Alonso‐Arroyo 2014		21.7		36.7		42.3		46.7		48.0
Amarilyo 2013	13.6	23.5	34.2	40.3	47.8	53.1 (38 months)	54.9	56.7 (50 months)	57.9	59.1
Arap 2014	5.2	14.1	23.0	27.4	34.9	39.3	44.1	44.4	47.4	48.2
Arrive 2001		3.7		6.5		8.5
Arrive 2004		15.7		27.1		31.4		32.7		33.5
Autorino 2006	5.5	11.9	14.5	16.5	19.2	20.5	21.3
Autorino 2007	8.4	18.5	28.2	34.1	37.7	40.6	43.2	45.4
Autorino 2008		13.4		18.6		22.1
Bakkum 2015		7.5		14.9		17.4		18.3		20
Balasubramanian 2006	12.0		30.7		48.5		52.7		55.2
Bhandari 2002	8.4 (8 months)		23.0 (20 months)		29.0 (32 months)		31.8 (44 months)
Bird 1999		13		27		34		39		41
Bolac 2009				46.1		57.6		61.1
Boldt 1999		13.1		26.5		34.3		39.4
Brazzelli 2009				56.9		63.1
Buchan 2011	3.9	7.3	15.6	21.8	25.1	28.5	30.7	32.4
Bydder 2004				23.7		32.4
Callaham 1998		16.8		30.4		39.1		42.2
Camacho 2005		12			41					63
Cartwright 2007a	9.2	23.8	40.8	45.4	56.2	59.2 (32 months)
Castagnetti 2013		30.7		40.2		45.0		46.2		47.2
Castillo 2000	3.7		7.9		12.0		14.9		16.3
Castillo 2002		10.6		25.2		33.5		39.6		41.5
Chalmers 1990a		23.2		30.1				35.8
Chan 2011	9.5		20.3		29.8		33.4
Chan 2013	4.9 (overall) 5.7 (American Society of Clinical Oncology ) 6.9 American Society of Hematology ) 2.1 (International Society for Pharmacoeconomics and Outcomes Research )	12.8 (overall) 14.8 (American Society of Clinical Oncology ) 24.1 (American Society of Hematology ) 6.4 (International Society for Pharmacoeconomics and Outcomes Research )	20.7 (overall) 22.7 (American Society of Clinical Oncology ) 34.5 (American Society of Hematology ) 12.8 (International Society for Pharmacoeconomics and Outcomes Research )	25.0 (overall) 29.5 (American Society of Clinical Oncology ) 34.5 (American Society of Hematology ) 12.8 International Society for Pharmacoeconomics and Outcomes Research )	30.0 (overall) 37.5 (American Society of Clinical Oncology ) 41.4 (American Society of Hematology) 14.9 (International Society for Pharmacoeconomics and Outcomes Research )	34.1 (overall) 40.9 (American Society of Clinical Oncology ) 17.0 (International Society for Pharmacoeconomics and Outcomes Research )	37.8 (overall) 43.2 (American Society of Clinical Oncology ) 25.5 (nternational Society for Pharmacoeconomics and Outcomes Research )	38.4 (overall) 43.2 (American Society of Clinical Oncology )	39.0 (overall) 45.4 (American Society of Clinical Oncology )	40.2 (overall) 46.6 (American Society of Clinical Oncology )
Chand 2008		18.2		25.1
Cheng 1998		7.9		27.0		37.6		43.3		44.4
Chung 2012	3.7	12.2	19.4	25.0	31.2	34.6	37.8	40.4	41.1	41.9
Chung 2012a		21.2		33.1		36.8		38.4
Ciesla 2001		16.0		21.8
Cifuentes 2007		23.4		43.9		53.2		54.4
Clemency 2014	5.8	14.5	22.8	29.8	34.0	37.2	39.4	40.6	41.6	42.4
Cohen 2012		37.8		65.1		75.9		81.7
Collet 1993		5.1		8.0		11.1		11.9		12.4
Collet 1997		2.8		4.5		7.1		9.9		10.7
Collet 2006	2.0	8.3		14.9		20.4		24.3		26.2
Collier 2010	5.5	9.6	13.6	17.8	19.9	22.2	23.1	23.8	24.9	24.9
Corry 1990	5.3 (9 months)			16.7 (21 months)		22.7 (33 months)		24.0 (45 months)
Craig 2001	(5 months) 13.4 (1980‐84 meetings) 10.4 (1990‐94 meetings)	(11 months) 22.5(1980‐84 meetings) 17.2 (1990‐94 meetings)	(17 months) 28.8 (1980‐84 meetings) 24.5 (1990‐94 meetings)	(23 months) 38.1 (1980‐84 meetings) 31.1 (1990‐94 meetings)	(29 months) 41.9 (1980‐84 meetings) 36.1 (1990‐94 meetings)	(35 months) 45.0 (1980‐84 meetings) 39.3 (1990‐94 meetings)	(41 months) 47.5 (1980‐84 meetings) 41.8 (1990‐94 meetings)	(47 months) 50.3 (1980‐84 meetings) 42.8 (1990‐94 meetings)		(59 months) 52.5 (1980‐84 meetings) 44.4 (1990‐94 meetings)
Czorlich 2016		16.2		29.8		40.4
Dahllof 2008		8.8 (10 months)	15.8 (20 months)		19.5		23.3 (40 months)	25.2 (50 months)		25.8
Daluiski 1998		15		31		43		48		54
Daruwalla 2015	4.7	9.7	16	20.3	21.7	23.3	26.2	27.3	27.8	28
Davies 2002	3.5 (5 months)	11.6 (10 months)	25.6 (20 months)	30.2 (25 months)	37.2	41.3 (35 months)	44.8
De Bellefeuille 1992	11		29		41		52		58	58
de Meijer 2015		22.4		49.0		62.2		66.0		68.6
Delgado Gallego 2007		5.6 (Sociedad Española de Medicina de Familia y Comunitaria) 3.2 (Sociedad Española de Medicina Rural y Generalista) 5.9 (Sociedad Española de Medicina General)		12.3 (Sociedad Española de Medicina de Familia y Comunitaria) 7.3 (Sociedad Española de Medicina Rural y Generalista) 5.9 (Sociedad Española de Medicina General)		14.2 (Sociedad Española de Medicina de Familia y Comunitaria) 8.1 (Sociedad Española de Medicina Rural y Generalista) 5.9 (Sociedad Española de Medicina General)		14.9 (Sociedad Española de Medicina de Familia y Comunitaria) 8.1 (Sociedad Española de Medicina Rural y Generalista) 5.9 (Sociedad Española de Medicina General)
DeMola 2009		11		23.5		30
Dhaliwal 2008		5.0		10.0		15.0		16.0		16.5
Diezel 1999		37		42		44		46		46
Donegan 2010		15.2		30.2		40.0		44.0		48.5
Dressler 2015		15.9		23.4		25.4		27.4		27.9
Eck 2005	8.6 (1997 International Society of Arthroscopy, Knee Surgery and Sports Medicine) 8.7 (1999 Biennial International Society of Arthroscopy, Knee Surgery and Sports Medicine)		19.1 (1997 International Society of Arthroscopy, Knee Surgery and Sports Medicine) 18.4 (1999 Biennial International Society of Arthroscopy, Knee Surgery and Sports Medicine)		24.7 (1997 International Society of Arthroscopy, Knee Surgery and Sports Medicine) 28.1 (1999 Biennial International Society of Arthroscopy, Knee Surgery and Sports Medicine)		29.0 (1997 International Society of Arthroscopy, Knee Surgery and Sports Medicine) 37.8 (1999 Biennial International Society of Arthroscopy, Knee Surgery and Sports Medicine)		32.7 (1997 International Society of Arthroscopy, Knee Surgery and Sports Medicine) 38.8 (1999 Biennial International Society of Arthroscopy, Knee Surgery and Sports Medicine)
Ejnisman 2013		6.3		12.3		17.6		22.2		25.4
Elder 1994				29		35		48
Eloubeidi 2001		6.7		16.2		22.8		25.1
Evers 2000		16.6		34.4		47.0				52.3
Forlin 2013		9.3		13.7		16.3		17.0		17.8
Fosbol 2012	7 (2006 to 2008 American Heart Association) 6 (2006 to 2008 American College of Cardiology) 6 (2006 to 2008 European Society of Cardiology)	16 (2006 to 2008 American Heart Association) 14 (2006 to 2008 American College of Cardiology) 14 (2006 to 2008 European Society of Cardiology)	24 (2006 to 2008 American Heart Association) 20 (2006 to 2008 American College of Cardiology) 18 (2006 to 2008 European Society of Cardiology)	32 (2006 to 2008 American Heart Association) 26 (2006 to 2008 American College of Cardiology) 24 (2006 to 2008 European Society of Cardiology)
Galang 2011		11.1		15.1		16.8		18.1		19.2
Gaundong Mbethe 2008		10.6		19.1		24.6				27.1
Gavazza 1996	8.5	13.6	23.7	35.6	42.0	45.7
Glick 2006	6.8	27.0	35.8	41.9	47.3	48.6	50.7	53.0	53.0
Goldman 1980	5	19	30	40	46	49
Goldman 1982	12	24	34	43	49	54
Gourtaud 2009		13.3		27.1
Greenberg 2008		8.4		18.4			25.9 (40 months)	26.8 (50 months)
Guryel 2006	7.0	13.0	24.3	27.7	30.8	31.8	32.8	33.5
Ha 2008		24.7 (European Congress of Radiology) 11.6 (Radiological Society of North America) 13.6 (Korean Radiological Society)		38.8 (European Congress of Radiology) 26.8 (Radiological Society of North America) 19.9 (Korean Radiological Society)		43.5 (European Congress of Radiology) 32.3 (Radiological Society of North America) 21.3 (Korean Radiological Society)		47.1 (European Congress of Radiology) 34.7 (Radiological Society of North America) 23.0 (Korean Radiological Society)
Hackett 2014			20 months for all 25.1 (2004 International Liver Transplantation Society) 29.0 (2005 International Liver Transplantation Society) 16.6 (2007 International Liver Transplantation Society) 26.2 ( 2006 joint meeting of International Liver Transplantation Society, European Liver and Intestinal Transplant Association, Liver Intensive Care of Europe) 13.9 (2008 joint meeting of International Liver Transplantation Society, European Liver and Intestinal Transplant Association, Liver Intensive Care of Europe)				40 months for all 41.1 (2004 International Liver Transplantation Society) 37.3 (2005 International Liver Transplantation Society) 25.5 (2007 International Liver Transplantation Society) 39.4 (2006 joint meeting of International Liver Transplantation Society, European Liver and Intestinal Transplant Association, Liver Intensive Care of Europe) 25.1 (2008 joint meeting of International Liver Transplantation Society, European Liver and Intestinal Transplant Association, Liver Intensive Care of Europe)			46.8 (2004 International Liver Transplantation Society) 42.6 (2005 International Liver Transplantation Society) 30.7 (2007 International Liver Transplantation Society) 45.2 (2006 joint meeting of International Liver Transplantation Society, European Liver and Intestinal Transplant Association, Liver Intensive Care of Europe) 30.3 (2008 joint meeting of International Liver Transplantation Society, European Liver and Intestinal Transplant Association, Liver Intensive Care of Europe)
Hamlet 1997	8	15	22	29	35	40	42	44	44	45
Hanchanale 2014	20.1	25.1				39.0
Harshavardhana 2009				55.4
Hashkes 2003		7.8		20.2		29.8		34.5		36.5
Hoeg 2009		14.1		31.7		44.7				60.8
Hopewell 2003				51.9				58.6
Hopewell 2015	12.5 (overall)	25.9 (overall) 44.4 (2010 American College of Rheumatology) 37.5 (2010 American Diabetes Association) 31.3 (2010 American Heart Association) 14.3 (2010 American Psychiatric Association) 12.5 (2010 American Society of Anesthesiologists) 19.3 (2010 American Society of Clinical Oncology) 14.7 (2010 American Society of Hematology) 40.0 (2010 American Thoracic Society) 25.0 (2010 Infectious Diseases Society of America)	35.0 (overall)	47.2 (overall) 72.2 American College of Rheumatology (ACR)) 50.0 (2010 American Diabetes Association (ADA)) 50.0 (2010 American Heart Association (AHA)) 28.6 (2010 American Psychiatric Association (PSY)) 25.0 (2010 American Society of Anesthesiologists) 42.1 (2010 American Society of Clinical Oncology) 35.3 (2010 American Society of Hematology) 64.0 (2010 American Thoracic Society) 50.0 (2010 Infectious Diseases Society of America (IDSA))	49.7 (overall)	53.3 (overall) 72.2 American College of Rheumatology) 62.5 (2010 American Diabetes Association) 53.0 (2010 American Heart Association) 42.8 (2010 American Psychiatric Association) 25.0 (2010 American Society of Anesthesiologists) 49.1 (2010 American Society of Clinical Oncology) 44.1 (2010 American Society of Hematology) 68.0 (2010 American Thoracic Society) 62.5 (2010 Infectious Diseases Society of America)
Jackson 2000		10		23		33		40		44
Jamjoom 2014		10.6		12.5		12.8
Jamjoom 2015		14.2		22.9		28.9		32.8		34.0
Jasko 2003		15.2		29.5		34.2		39.3
Jones 2008		10.4		27.3		31.7		33.1		33.8
Juzych 1991	10	20	40	49	53	55	56	58	60
Kim 1998	7.6	31.1	41.5	43.7
Klassen 2002	6.9	15.9	22.7	35.5	43.0	49.3	52.0	54.3	56.6	58.1
Kleine‐Konig 2014		18.4		27.6		31.6		33.9		35.9
Kleweno 2008		13.3		35.2		45.5		53.3		59.4
Kottachchi 2010			29.3 (20 months)		47.6	62.2 (40 months)	69.5 (50 months)			76.8
Krzyzanowska 2003	11.4	22.2	30.0	41.8	49.4	57.6	62.2	65.9	69.2	73.3
Kumar 1995		9.5
Kunadian 2015	13.5	22.5	34.0	42.5	49.0	52.5	56.0	58.0	60.0	61.0
Kwong 2007	9.4	18.0	24.1	28.1	31.3	32.0	35.3	36.7	36.7	38.9
Larian 2001	4.8 (1993 American Academy of Otolaryngology ‐ Head & Neck Surgery) 7.2 (1994 American Academy of Otolaryngology ‐ Head & Neck Surgery) 5.1 (1995 American Academy of Otolaryngology ‐ Head & Neck Surgery)	7.6 (1993 American Academy of Otolaryngology ‐ Head & Neck Surgery) 13.0 (1994 American Academy of Otolaryngology ‐ Head & Neck Surgery) 8.8 (1995 American Academy of Otolaryngology ‐ Head & Neck Surgery)	20.9 (1993 American Academy of Otolaryngology ‐ Head & Neck Surgery) 19.8 (1994 American Academy of Otolaryngology ‐ Head & Neck Surgery) 10.4 (1995 American Academy of Otolaryngology ‐ Head & Neck Surgery)	24.5 (1993 American Academy of Otolaryngology ‐ Head & Neck Surgery) 25.3 (1994 American Academy of Otolaryngology ‐ Head & Neck Surgery) 19.5 (1995 American Academy of Otolaryngology ‐ Head & Neck Surgery)	27.7 (1993 American Academy of Otolaryngology ‐ Head & Neck Surgery) 28.3 (1994 American Academy of Otolaryngology ‐ Head & Neck Surgery) 26.3 (1995 American Academy of Otolaryngology ‐ Head & Neck Surgery)	30.1 (1993 American Academy of Otolaryngology ‐ Head & Neck Surgery) 31.4 (1994 American Academy of Otolaryngology ‐ Head & Neck Surgery) 31.6 (1995 American Academy of Otolaryngology ‐ Head & Neck Surgery)	31.7 (1993 American Academy of Otolaryngology ‐ Head & Neck Surgery) 33.7 (1994 American Academy of Otolaryngology ‐ Head & Neck Surgery)	33.3 (1993 American Academy of Otolaryngology ‐ Head & Neck Surgery) 35.2 (1994 American Academy of Otolaryngology ‐ Head & Neck Surgery)
Lau 2016		15.2		33.5		43.3		49.3		51.7
Lee 2012		7.8		16.2		25.7		28.6
Li 2004		20 (1997 Society for Academic Emergency Medicine) 21 (1999 Society for Academic Emergency Medicine) 22 (2000 Society for Academic Emergency Medicine) 24 (2001 Society for Academic Emergency Medicine)		30 (1997 Society for Academic Emergency Medicine) 31 (1999 Society for Academic Emergency Medicine) 33 (2000 Society for Academic Emergency Medicine) 34 (2001 Society for Academic Emergency Medicine)		35 (1997 Society for Academic Emergency Medicine) 36 (1999 Society for Academic Emergency Medicine) 38 (2000 Society for Academic Emergency Medicine)		40 (1997 Society for Academic Emergency Medicine) 39 (1999 Society for Academic Emergency Medicine)
Lim 2013	4.3	8.9	15.0	21.7	25.6	27.7	29.0	30.9
Livas 2014		22.4		35.9		44.6		49.3		51.4
Lloyd 2006	3.8	18.8	28.4	37.7	45.2	50.0	51.0	53.1		55.8
Loevy 1997		11.6		26.5					48
Manuck 2015	5.3 (overall) 16.2 (oral) 2.8 (poster)	16.2 (overall) 39.3 (oral) 10.6 (poster)	25.5 (overall) 47.4 (oral) 20.3 (poster)	33 (overall) 57.7 (oral) 27.2 (poster)	36.1 (overall) 62 (oral) 35.3 (poster)	34.3 (overall) 65.4 (oral) 34.6 (poster)		44.5 (overall) 68.8 (oral) 38.7 (poster)		47 (overall) 72.2 (oral) 41 (poster)
Martinez 2008		1.3 (1997 Confederación Unificada Bioquímica de la República Argentina) 1.4 (1999 Confederación Unificada Bioquímica de la República Argentina) 5.1 (2001 Confederación Unificada Bioquímica de la República Argentina) 1.3 (2003 Confederación Unificada Bioquímica de la República Argentina) 1.9 (2005 Confederación Unificada Bioquímica de la República Argentina)		2.6 (1997 Confederación Unificada Bioquímica de la República Argentina) 8.3 (1999 Confederación Unificada Bioquímica de la República Argentina) 8.5 (2001 Confederación Unificada Bioquímica de la República Argentina) 2.7 (2003 Confederación Unificada Bioquímica de la República Argentina) 4.8 (2005 Confederación Unificada Bioquímica de la República Argentina)		7.8 (1997 Confederación Unificada Bioquímica de la República Argentina) 13.9 (1999 Confederación Unificada Bioquímica de la República Argentina) 8.5 (2001 Confederación Unificada Bioquímica de la República Argentina) 9.3 (2003 Confederación Unificada Bioquímica de la República Argentina)		13.0 (1997 Confederación Unificada Bioquímica de la República Argentina) 18.1 (1999 Confederación Unificada Bioquímica de la República Argentina) 10.2 (2001 meeting) 14.7 (2003 Confederación Unificada Bioquímica de la República Argentina)		14.3 (1997 Confederación Unificada Bioquímica de la República Argentina) 18.1 (1999 Confederación Unificada Bioquímica de la República Argentina) 10.2 (2001 Confederación Unificada Bioquímica de la República Argentina)
Marx 1999	7	13	22	27	32	34	35
McCormick 1985	6	17	34	39	41	48
McKinley 2010	6.1	17.4	27.0	34.8	37.4	37.4	40.0	40.9	41.7	41.7
Meininger 2011	8.4 (2000 European Congress of Anesthesia) 6.7 (2005 European Congress of Anesthesia) 10.1 (2000 German Congress of Anesthesia) 11.4 (2005 German Congress of Anesthesia)	21.3 (2000 European Congress of Anesthesia) 17.4 (2005 European Congress of Anesthesia) 23.9 (2000 German Congress of Anesthesia) 26.7 (2005 German Congress of Anesthesia)		28.6 (2000 European Congress of Anesthesia) 24.9 (2005 European Congress of Anesthesia) 32.0 (2000 German Congress of Anesthesia) 35.7 (2005 German Congress of Anesthesia)
Meissner 2014				38.4 (26 months)
Meranze 1982	5	17	28	32
Micieli 2012	5	11.5	18	23	26	29	30	31	32	32.5
Miguel‐Dasit 2006a		1.3 (1994 Spanish Congress of Radiology) 2.9 (1996 Spanish Congress of Radiology) 3.3 (1998 Spanish Congress of Radiology)		7.4 (1994 Spanish Congress of Radiology) 9.0 (1996 Spanish Congress of Radiology) 8.6 (1998 Spanish Congress of Radiology)		12.1 (1994 Spanish Congress of Radiology) 12.3 (1996 Spanish Congress of Radiology) 11.4 (1998 Spanish Congress of Radiology)		14.9 (1994 Spanish Congress of Radiology) 14.0 (1996 Spanish Congress of Radiology) 13.5 (1998 Spanish Congress of Radiology)		16.7 (1994 Spanish Congress of Radiology) 15.5 (1996 Spanish Congress of Radiology) 14.3 (1998 Spanish Congress of Radiology)
Miguel‐Dasit 2006b		11.7 (9 months)		27.0 (21 months)		37.2 (33 months)		44.4 (45 months)		47.0 (57 months)
Miguel‐Dasit 2007		12.7 (9 months)		28.2 (21 months)		37.5 (33 months)		42.5 (45 months)		45.3 (57 months)
Montane 2007				19.4		23.4
Muffly 2014		26.7		42.0		51.4		54.1		55.3
Murrey 1999		11.0		28.1		39.8		42.9		43.8
Mutlu 2015				12.1		15.9		19.6		23.8
Ng 2004	11.4	19.1	29.0	35.2	37.6	39.5	40.7	41.3
Nguyen 1998	12.4	22.7	35.5	44.7	50.4	54.5	55.9	58.4		58.8
O' Connor 2015	13.4	22.5	29.9	37.9		43.4		50.7		54.0
Ohlsson 1999		19.9		33.3				46.8		51.8 (57 months)
Ospina 2006	3.8	13.3	27.2	39.1	47.3	52.1	54.1	56.2	57.4	58.6
Ozkösem 2013a	0	13	35	40	47	57	60
Papoutsis 2015		8.4		20.0
Parkar 2013		25.5 (overall) 23.3 (2008 European Society of Musculoskeletal Radiology) 28.1 (2009 European Society of Musculoskeletal Radiology)		38.7 (overall) 41.1 (2008 European Society of Musculoskeletal Radiology) 36.0 (2009 European Society of Musculoskeletal Radiology)
Patel 2011		18.8 (2003, 2004 and 2005 Congress of Neurological Surgeons) 20.9 (2005 American Association of Neurological Surgeons)		25.8 (2003, 2004 and 2005 Congress of Neurological Surgeons) 26.4 (2005 American Association of Neurological Surgeons)		29.9 (2003, 2004 and 2005 Congress of Neurological Surgeons) 29.2 (2005 American Association of Neurological Surgeons)		32.0 (2003, 2004 and 2005 Congress of Neurological Surgeons) 30.5 (2005 American Association of Neurological Surgeons)		32.7 (2003, 2004 and 2005 Congress of Neurological Surgeons) 31.4 (2005 American Association of Neurological Surgeons)
Pinheiro 2009	3.2	7.1	9.1	13.0	14.3	14.9	16.2	16.9	20.1	20.1
Prohaska 2013	3.2	8.1	12.1	15.3	16.8	17.7	18.1	18.3	18.6	18.7
Rao 2006	10.0	17.4	28.1	35.5	41.3	43.0
Riessland 2004	10.1	21.9	30.8	36.9	42.9	44.7	46.4	49.0	49.3	51.0
Rollin 2009		14.8		22.6				29.6
Roy 2001		27.1		51.8		61.7		65.2		66.7
Rubin 2014		26.6		37.0		43.2		45.3		46.6
Saeed 2011	4.8	8.8	14.3	17.5	22.3	23.9	25.5	28.3
Saldanha 2016	5.5	14.4	23	29.2	32.4	35.3	37.2	39.4	40.7	41.9
Sanders 2001	12.9	26.7	37.6	45.5	53.3	60.8	63.9	66.7	68.2	69.8
Sanossian 2006	15.0	25.5	37.4	45.6	50.4	55.0	57.8	60.1
Scherer 1994	16	24	40	50	56	57	60	63
Schoenfeld 2012		12.9		24.3		33.8		38.9		41.9
Scholey 2005	8.0	16.7	24.2	30.0	35.5	39.6	42.1	44.3	46.2
Schulte 2012a	12.7	18.6	22.9	26.3	28.8	31.4	33.1	33.9	34.7	35.6
Schulte 2012b	10.6	15.2	22.0	26.3	31.8	33.9	35.2	36.0	37.7	37.8
Sebel 2001	13.1		29.2		40.3		43.7		47.7
Simons 2007				44.2		56.6
Singh 2015		4.9		7.8		13.7		13.7		13.7
Smart 2013		19.7		50.0		68.0		71.3		73.8
Smith 2007	13.2	23.7	31.7	38.4	42.7
Snedeker 2010	15.0 (Conference of Research Workers in Animal Disease)	22.8 (International symposium on Veterinary Economics and Epidemiology) 22.6 (Conference of Research Workers in Animal Disease)	28.3 (Conference of Research Workers in Animal Disease)	33.6 (International symposium on Veterinary Economics and Epidemiology) 32.7 (Conference of Research Workers in Animal Disease)	37.6 (Conference of Research Workers in Animal Disease)	40.9 (International symposium on Veterinary Economics and Epidemiology) 41.2 (Conference of Research Workers in Animal Disease)	43.4 (Conference of Research Workers in Animal Disease)	45.6 (International symposium on Veterinary Economics and Epidemiology) 45.6 (Conference of Research Workers in Animal Disease)
Song 2010		8.9		35.6
Stolk 2002	5.7 (1995 meeting) 5.5 (1996 meeting) 5.4 (1997 meeting) 1.6 (1998 meeting) 4.1 (1999 meeting)	10.9 (1995 meeting) 11.0 (1996 meeting) 9.6 (1997 meeting) 7.9 (1998 meeting) 8.2 (1999 meeting)	18.4 (1995 meeting) 18.3 (1996 meeting) 16.2 (1997 meeting) 14.0 (1998 meeting) 14.5 (1999 meeting)	23.0 (1995 meeting) 22.5 (1996 meeting) 23.8 (1997 meeting) 18.7 (1998 meeting) 21.2 (1999 meeting)	26.4 (1995 meeting) 26.6 (1996 meeting) 25.4 (1997 meeting) 21.0 (1998 meeting)	28.2 (1995 meeting) 28.4 (1996 meeting) 30.0 (1997 meeting) 22.2 (1998 meeting)	30.0 (1995 meeting) 29.8 (1996 meeting) 30.0 (1997 meeting)	30.0 (1995 meeting) 29.8 (1996 meeting) 30.4 (1997 meeting)	30.0 (1995 meeting) 30.7 (1996 meeting)	30.0 (1995 meeting) 30.7 (1996 meeting)
Stöcker 2009		14.1		28.8		39.4		47.8		51.4
Sun 2011		16.0 (American Academy of Ophthalmology) 14.0 (Association for Research in Vision and Ophthalmology)		32.5 (American Academy of Ophthalmology) 26.0 (Association for Research in Vision and Ophthalmology)		40.0 (American Academy of Ophthalmology) 33.5 (Association for Research in Vision and Ophthalmology)		42.9 (American Academy of Ophthalmology) 37.0 (Association for Research in Vision and Ophthalmology)		44.6 (American Academy of Ophthalmology) 39.0 (Association for Research in Vision and Ophthalmology)
Tam 2008	2.7		20.3		39.2		56.8		66.2
Tam 2011		16.0		37.4		55.7		68.7		77.9
Tambuscio 2010		8.5	12.4
Theman 2014	13	22	31	36		43		46		47
Timmer 2001		19.5		40.7				57.4
Timmer 2002	9.2 (basic science) 6.1 (controlled trials) 5.1 (other clinical research)	19.0 (basic science) 19.0 (controlled trials) 11.0 (other clinical research)	27.0 (basic science) 24.8 (controlled trials) 17.0 (other clinical research)	32.2 (basic science) 31.9 (controlled trials) 24.1 (other clinical research)	37.9 (basic science) 35.9 (controlled trials) 30.1 (other clinical research)	40.8 (basic science) 45.1 (controlled trials) 36.0 (other clinical research)	43.1 (basic science) 47.9 (controlled trials) 36.9 (other clinical research)	46.0 (basic science) 50.9 (controlled trials) 38.1 (other clinical research)	46.0 (basic science) 53.1 (controlled trials) 39.0 (other clinical research)	48.3 (basic science) 54.0 (controlled trials) 39.9 (other clinical research)
Toma 2006	18.3	36.6	55.9	66.1	72.0	75.8	79.0	80.6	83.3	83.3
ul Haq 2011				27.4
Uysal 2008	4.8	9.3		11.7		13.2		13.8
Uzun 2013	6.7	13.3	19.0	24.8	27.4	30.7	31.8	33.0	33.5	34.6
Valderrama‐Zurian 2009	6.8		17.3		26.9		32.4		35.4
van der Steen 2004		30.1 (overall) 24.6 (1995 European Association of Plastic Surgeons) 16.7 (1996 European Association of Plastic Surgeons) 38.3 (1997 European Association of Plastic Surgeons) 28.3 (1998 European Association of Plastic Surgeons) 31.9 (1999 European Association of Plastic Surgeons)		47.7 (overall) 42.1 (1995 European Association of Plastic Surgeons) 50.0 (1996 European Association of Plastic Surgeons) 46.7 (1997 European Association of Plastic Surgeons) 45.0 (1998 European Association of Plastic Surgeons) 53.6 (1999 European Association of Plastic Surgeons)		56.9 (overall) (57.9 (1995 European Association of Plastic Surgeons) 58.3 (1996 European Association of Plastic Surgeons) 55.0 (1997 European Association of Plastic Surgeons) 55.0 (1998 European Association of Plastic Surgeons) 58.0 (1999 European Association of Plastic Surgeons)		59.5 (overall) 59.6 (1995 European Association of Plastic Surgeons) 63.3 (1996 European Association of Plastic Surgeons) 55.0 (1997 European Association of Plastic Surgeons) 61.7 (1998 European Association of Plastic Surgeons) 58.0 (1999 European Association of Plastic Surgeons)		61.4(1995 European Association of Plastic Surgeons) 66.7 (1996 European Association of Plastic Surgeons) 55.0 (1997 European Association of Plastic Surgeons) 65.0 (1998 European Association of Plastic Surgeons)
Wai 2006		12.6		36.3		50.4
Walby 2001				30.9				35.3
Walsh 2013	4.2	8.5	10.5	16.3	19.2	29.8	30.3	31.8	32.7	33.6
Wang 1999		27		37		41		43		44
Weale 2006		15.8 (Association of Surgeons of Great Britain & Ireland) 23.2 (Vascular Surgical Society of Great Britain & Ireland) 17.4 (British Transplantation Society) 9.1 (Association of Coloprotocology of Great Britain & Ireland)		33.8 (Association of Surgeons of Great Britain & Ireland) 53.6 (Vascular Surgical Society of Great Britain & Ireland) 32.6 (British Transplantation Society) 23.1 (Association of Coloprotocology of Great Britain & Ireland)
Wieser 2015	8.1	15.7	30.0	38.2	45.7	50.1	53.6	55.7	57.5	59.6
Wong 2009		17.6		32.7		41.1		45.1		47.1
Yalcinkaya 2013		14.5		21.7		26.1
Yamaguchi 1990	28 (Japanese Society for Bacteriology) 18 (Japanese Society for Bacteriology regional meeting) 29 (Japanese Society of Internal Medicine) 9 (Japanese Society of Internal Medicine regional meeting)	53 (Japanese Society for Bacteriology) 37 (Japanese Society for Bacteriology regional meeting) 59 (Japanese Society of Internal Medicine) 23 (Japanese Society of Internal Medicine regional meeting)	64 (Japanese Society for Bacteriology) 51 (Japanese Society for Bacteriology regional meeting) 67 (Japanese Society of Internal Medicine) 32 (Japanese Society of Internal Medicine regional meeting)	75 (Japanese Society for Bacteriology) 67 (Japanese Society for Bacteriology regional meeting) 80 (Japanese Society of Internal Medicine) 35 (Japanese Society of Internal Medicine regional meeting)	79 (Japanese Society for Bacteriology) 71 (Japanese Society for Bacteriology regional meeting) 81 (Japanese Society of Internal Medicine) 35 (Japanese Society of Internal Medicine regional meeting)	83 (Japanese Society for Bacteriology) 73 (Japanese Society for Bacteriology regional meeting) 84 (Japanese Society of Internal Medicine) 46 (Japanese Society of Internal Medicine regional meeting)	84 (Japanese Society for Bacteriology) 73 (Japanese Society for Bacteriology regional meeting) 86 (Japanese Society of Internal Medicine) 46 (Japanese Society of Internal Medicine regional meeting)	85 (Japanese Society for Bacteriology) 73 (Japanese Society for Bacteriology regional meeting) 87 (Japanese Society of Internal Medicine) 46 (Japanese Society of Internal Medicine regional meeting)
Yentis 1993	7		31		41		44		48
Yoo 2002		14.5 (American Orthopaedic Society for Sports Medicine) 24.0 (Arthroscopy Association of North America)		49.4 (American Orthopaedic Society for Sports Medicine) 32.9 (Arthroscopy Association of North America)		57.8 (American Orthopaedic Society for Sports Medicine) 38.3 (Arthroscopy Association of North America)		62.0 (American Orthopaedic Society for Sports Medicine) 41.9 (Arthroscopy Association of North America)	66.9 (American Orthopaedic Society for Sports Medicine) 46.1 (Arthroscopy Association of North America)
Yoon 2012a	15.6		35.7		42.8		49.4
Zamakhshary 2006		60.1		64.5
Özyurt 2012		4.3 (2004 Turkish National Dermatology) 2.4 (2006 Turkish National Dermatology) 4.9 (2008 Turkish National Dermatology)		10.9 (2004 Turkish National Dermatology) 8.0 (2006 Turkish National Dermatology) 8.3 (2008 Turkish National Dermatology)		12.8 (2004 Turkish National Dermatology) 9.8 (2006 Turkish National Dermatology) 9.8 (2008 Turkish National Dermatology)		15.8 (2004 Turkish National Dermatology) 13.7 (2006 mTurkish National Dermatology) 9.8 (2008 Turkish National Dermatology)

2.2. Analysis.

Comparison 2 Time to publication, Outcome 2 Mean or median time to publication.

Mean or median time to publication
Study	Number of abstracts (meeting)	Mean (months)	Standard Deviation	Median (months)	Range (months)	95 % Confidence Interval (months)
Abicht 2012	825 (overall; 1999 to 2008 American College of Foot & Ankle Surgeons) 20 (1999 meeting) 35 (2000 meeting) 37 (2001 meeting) 52 (2002 meeting) 59 (2003 meeting) 74 (2004 meeting) 107 (2005 meeting) 142 (2006 meeting) 150 (2007 meeting) 149 (2008 meeting)	17.6 (overall) 4.8 (1999 meeting) 2.9 (2000 meeting) 19.2 (2001 meeting) 17.9 (2002 meeting) 23 (2003 meeting) 24.1 (2004 meeting) 20.8 (2005 meeting) 22.2 (2006 meeting) 12.7 (2007 meeting) 13.4 (2008 meeting)			‐46 to 75 (overall) ‐12 to 40 (1999 meeting) ‐14 to 16 (2000 meeting) 6 to 34 (2001 meeting) 1 to 46 (2002 meeting) 3 to 75 (2003 meeting) 6 to 60 (2004 meeting) ‐46 to 72 (2005 meeting) ‐24 to 54 (2006 meeting) ‐7 to 33 (2007 meeting) ‐22 to 30 (2008 meeting)
Abzug 2014	1127 (overall; 2000 to 2005 American Society for Surgery ) 249 (2000 meeting) 183 (2001 meeting) 177 (2002 meeting) 150 (2003 meeting) 192 (2004 meeting) 176 (2005 meeting)	25.2 (overall) 28.8 (2000 meeting) 27.6 (2001 meeting) 31.2 (2002 meeting) 25.2 (2003 meeting) 20.4 (2004 meeting) 19.2 (2005 meeting)
Acevedo 2014	1972 (2000 to 2010 American Society of Clinical Oncology, American Society of Hematology, European Society of Medical Oncology, European Cancer Organization, and American Society for Radiation Oncology)				0 to 96 12 to 24 (IQR)
Adhikari 2011	1479 (1999 to 2008 Society for Academic Emergency Medicine & American College of Emergency Physicians)	20.4 (emergency ultrasound) 22.8 (emergency medical services) 21.6 (toxicology)				4.8 to 36 (emergency ultrasound) 6 to 39.6 (emergency medical services) 3.1 to 40.1 (toxicology)
Akbari‐Kamrani 2008	198 (2000 to 2003 American Society for Laser Medicine and Surgery)	57				52 to 61
Al‐Qaoud 2013	439 (2000 to 2010 Quebec Urological Association)	19.5
Allart 2013	231 (2008 Société Française de Médecine Physique et de Réadaptation)	18.4	14	21
Allart 2015	779 (2008 European Congress of Physical and Rehabilitation Medicine)	12	15.7	11
Amarilyo 2013	2149 (2006 American College of Rheumatology and Association of Rheumatology Health Professionals)	15.8	18.3	16	‐9 to 51 1.75 to 29.5 (IQR)
Amirhamzeh 2012	762 (2002 to 2006 Pediatric Orthopaedic Society of North America)	29.4 (2002 meeting) 26.8 (2003 meeting) 26.3 (2004 meeting) 22.7 (2005 meeting) 18.8 (2006 meeting)
Arap 2014	195 (2001 to 2007 American Urological Association)			18.8
Arora 2012	191 (International Society of Paediatric Oncology ‐ only abstracts by authors from India)) 13 (American Society of Clinical Oncology) 275 (International Society of Paediatric Oncology (only from UK))	20 (International Society of Paediatric Oncology ‐ only abstracts by authors from India)) 24 (American Society of Clinical Oncology) 31 (International Society of Paediatric Oncology (only from UK))			‐1 to 95 (International Society of Paediatric Oncology (only abstracts by authors from India))
Autorino 2006	1100 (2001 and 2002 World Congress of Endourology)	14.6			40 (maximum)
Autorino 2007	1406 (2000 and 2001 European Association of Urology)	8.7	4.2
Autorino 2008	1877 (2002 to 2004 Société Internationale d'Urologie)	13			1 to 45
Autorino 2010	94 (2004 to 2006 World Congress of Endourology)	16.4	13.2
Bagheri 2005	446 (1997 to 1999 American Association of Oral and Maxillofacial Surgeons)	23.4	16.2		1 to 72
Balasubramanian 2006	241(1997 Association of Surgeons of Great Britain and Ireland)			18	11 to 27 (IQR)
Beker‐Acay 2015	3192 (2010 to 2012 Turkish National Radiology)	15			0 to 42
Bello 2013	75 (2007 to 2010 Nigerian Association of Urological Surgeons)			15	2 to 40
Bergoli 2011	140 (2001 Brazilian Society of Oral Research)	24.1
Bhandari 2002	465 (1996 American Academy of Orthopedic Surgeons)	17.6	12	14	1 to 56
Bhasin 2007	106 (2001 and 2002 Vascular Society of Great Britain and Ireland)	11		12	7 to 18 (IQR)
Blackwell 2009	90 (2000 to 2002 Society of Maternal Fetal Medicine)			19	4 to 66
Bolac 2009	540 (1999 to 2004 Journées Dermatologiques de Paris)	20.3	16.9	17	0 to 90
Bonitz 2011	284 (2006 and 2007 ASRM)			16	10 to 28
Bowrey 1999	496 (1983 to 1995 Welsh Surgical Society)			17	24 to 168 10.0 to 27.5 (IQR)
Brazzelli 2009	160 (1995 to 2004 International Stroke Conference and European Stroke Conference)	20		16
Buchan 2011	179 (2004 Royal College of Ophthalmologists)	20.1				16.9 to 23.4
Burden 2014	112 (2003 to 2012 South Yorkshire Regional Gastroenterology)	12.8 (abstracts ranked within 'top 3' of each meeting) 19.3 (abstracts not ranked within 'top 3' of each meeting)
Bydder 2004	481 (1996 to 1999 Royal Australian and New Zealand College of Radiologists)	16.5		17
Byerly 2000	501 (American Society of Health‐System Pharmacists) 215 (American College of Clinical Pharmacy )			12.8 (1994 American Society of Health‐System Pharmacists) 14.9 (1994 American College of Clinical Pharmacy )
Callaham 1998	481 (1991 Society for Academic Emergency Medicine)	18
Camacho 2005	275 (1997 American Society of Clinical Oncology)	29		40.8		33.6 to 45.6
Canosa 2011	140 (2003 Congreso Argentino de Pediatría)	27	15	26	7 to 37
Carroll 2003	614 (1998 and 1999 Pediatric Academic Society) 160 (oral) 160 (poster) 136 (poster symposium) 158 (rejected)	21.4 (oral) 26.5 (poster) 20.3 (poster symposium) 23.1 (rejected)	11.8 (oral) 12.3 (poster) 11.9 (poster symposium) 12.1 (rejected)
Cartwright 2007a	130 (2003 International Continence Society)			11
Cartwright 2007b	116 (2003 International Urogynecological Meeting)			27
Castaldi 2015	621 (overall; 2005 to 2007 Italian Society of Hygiene and Public Health) 174 (2005 meeting) 187 (2006 meeting) 260 (2007 meeting)	25.1 (overall) 27.4 (2005 meeting) 30.5 (2006 meeting) 20.2 (2007 meeting)				22.1 to 28.1 (overall) 20.8 to 34.0 (2005 meeting) 24.6 to 36.4 (2006 meeting) 16.6 to 23.8 (2007 meeting)
Castillo 2000	491 (1992 Congress of the Spanish Society of Anesthesiology and Resuscitation)	22			‐24 to 72
Castillo 2002	472 (1995 European Society of Anesthesiologists)	16.8	15.6		24 to 60
Cauchy 2014	897 (2005 to 2012 French Congress of Digestive and Hepato‐biliary Surgery)			16.1 (overall) 16.1 (accepted) 15.1 (rejected)	0 to 67.4 (overall)
Chan 2013	164 (1997 to 2007 International Society for Pharmacoeconomics and Outcomes Research, American Society of Hematology, and American Society of Clinical Oncology)			37.2
Chand 2008	2172 (overall; 1999 to 2005 Cardiac Society of Australia and New Zealand) 253 (1999 meeting) 307 (2000 meeting) 307 (2001 meeting) 279 (2002 meeting) 331 (2003 meeting) 315 (2004 meeting) 380 (2005 meeting)	18.0 (overall) 24.0 (1999 meeting) 19.2 (2000 meeting) 20.4 (2001 meeting) 18.0 (2002 meeting) 15.6 (2003 meeting) 14.4 (2004 meeting) 15.6 (2005 meeting)	all reported as standard error 1.8 (overall) 2.2 (1999 meeting) 1.7 (2000 meeting) 1.6 (2001 meeting) 1.7 (2002 meeting) 1.2 (2003 meeting) 1.2 (2004 meeting) 1.2 (2005 meeting)
Chapman 2012	908 (overall; 1997 to 2007 Cochrane Colloquium ) 63 (1997 meeting) 58 (1998 meeting) 59 (1999 meeting) 32 (2000 meeting) 85 (2001 meeting) 89 (2002 meeting) 115 (2003 meeting) 113 (2004 meeting) 119 (2005 meeting) 133 (2006 meeting) 92 (2007 meeting)	22.8 (overall) 37.2 (1997 meeting) 26.4 (1998 meeting) 28.8 (1999 meeting) 33.6 (2000 meeting) 26.4 (2001 meeting) 31.2 (2002 meeting) 24.0 (2003 meeting) 18.0 (2004 meeting) 16.8 (2005 meeting) 14.4 (2006 meeting) 19.2 (2007 meeting)		24 (overall) 36 (1997 meeting) 24 (1998 meeting) 24 (1999 meeting) 36 (2000 meeting) 24 (2001 meeting) 30 (2002 meeting) 24 (2003 meeting) 24 (2004 meeting) 12 (2005 meeting) 12 (2006 meeting) 12 (2007 meeting)
Cheng 1998	178 (Cochrane Cystic Fibrosis Trials Register)			18 (derived from Kaplan Meier survival table)
Chung 2012	1005 (2006 and 2007 Korean Urological Association)	19.0	12.8
Chung 2012a	1176 (2005 to 2007 Spring and Fall meetings of the Congress of the Korean Society of Plastic and Reconstructive Surgeons)	15.0
Ciesla 2001	257 (1998 International Academy of Pathology ‐ United States and Canadian Academy of Pathology, American Society of Cytopathology, and American Society of Clinical Pathologists/College of American Pathologists)	15.8			2 to 26
Cifuentes 2007	665 (overall; 1999 to 2002 Asociación Espanola para el Estudio del Hígado) 176 (1999 meeting) 159 (2000 meeting) 159 (2001 meeting) 171 (2002 meeting)	20.9 (1999 meeting) 21.5 (2000 meeting) 20.2 (2001 meeting) 21.8 (2002 meeting)	1.1 (1999 meeting) 1.6 (2000 meeting) 1.6 (2001 meeting) 4.0 (2002 meeting)
Clemency 2014	635 (2003 to 2005 American College of Emergency Physicians, Society for Academic Emergency Medicine, National Association of EMS Physicians, Association of Air Medical Services, and National Association of EMS Educators)	22.2	16.5		0 to 94
Cohen 2012	378 (2000 to 2005 Society of Gynecologic Oncologists)			14	1 to 85
Cohen 2012a	2463 (2000 to 2004 American Academy of Otolaryngology, Head and Neck Surgery) 1051 (oral) 1413 (poster)	15.5 (oral) 19.1 (poster) 15.0 (oral by female author) 19.4 (posters by female author) 15.6 (oral by male author) 18.6 (posters by male author)
Collier 2010	623 (2002 to 2006 British Association of Oral and Maxillofacial Surgeons)			8	1 to 18
Cornu 2012	226 (2007 Congrès de l’Association Français d’Urologie)			24
Costa 2009	1620 (2002 ASCO, SABCS, EBCC, and ESMO)			11.6		7.6 to 15.7
Craig 2001	320 (1980 to 1984 British Orthopaedic Association) 685 (1990 to 1994 British Orthopaedic Association)	21.4 (1980 to 1984 meetings) 16.8 (1990 to 1994 meetings)			10.3 to 26.3 (overall)
Dahllof 2008	771 (1999 and 2001 International Association of Paediatric Dentistry)	20	20	18
Daluiski 1998	888 (overall; 1991 to 1993 Orthopaedic Research Society) 296 (1991 meeting) 296 (1992 meeting) 296 (1993 meeting)			20 (overall) 18 (1991 meeting) 23 (1992 meeting) 20 (1993 meeting)
Dangouloff‐Ros 2015	744 (2008 to 2010 French Society of Radiology)	21	13.7	18	10 to 29 (IQR)
Daruwalla 2015	443 (2007 to 2010 and 2011 to 2013 Singapore Orthopaedic Association)			13	‐73 to 76
Das 2013	52 (2000 to 2008 American Epilepsy Society)		3	15
Davies 2002	172 (1997 Perinatal Society of Australia and New Zealand)			18	‐36 to 41 9 to 26 (IQR)
De Bellefeuille 1992	81 (1984 American Society of Clinical Oncology)	22.8
de Meijer 2015	2174 (2007 to 2012 Dutch Surgical Society)			25 (accepted) 36 (rejected) 23 (oral) 36 (poster)		23 to 27 (accepted) 30.1 to 41.9 (rejected) 20.9 to 25.1 (oral) 26.6 to 45.4 (poster)
de Oliveira 2012	348 (2010 Brazilian Congress on Pain)	6.7	3	6.9
DeMola 2009	558 (1999 to 2004 American Academy of Orthopaedic Surgery)	31	28
Dhaliwal 2008	200 (2000 All India Ophthalmological Conference)	22.8	16.4	20	2 to 77
Dirk 1996	147 (abstracts from anesthesiology department)	32.4			‐12 to 96
Doğan 2012	218 overall (2007 to 2010 Turkish National Rhinology 67 (oral) 151 (poster)	21.8 (overall) 23.2 (oral) 20.8 (poster)
Drury 2012	282 (2003 to 2007 Society for Cardiothoracic Surgery)			15	‐24 to 63
Duchini 1997	2512 (1991 Digestive Diseases Week)	20.4	12
Durinka 2014	1938 (2009 American Transplant Congress)	16			1 to 49
Durinka 2014b	335 (2009 American College of Surgeons)	16.8
Duthie 2012	20 (2007 to 2009 Senior Trainees’ Annual Research Symposium)	20.2			5 to 41
Dyson 2006	283 (overall; 1990 to 1999 American College of Veterinary Anesthesiologists) 33 (1990 meeting ) 24 (1991 meeting) 28 (1992 meeting) 26 (1993 meeting) 38 (1995 meeting) 39 (1996 meeting) 30 (1997 meeting) 31 (1998 meeting) 34 (1999 meeting)	24.3 (overall) 39.3 (1990 meeting) 24.8 (1991 meeting) 20.9 (1992 meeting) 23.2 (1993 meeting) 22.9 (1995 meeting) 24.8 (1996 meeting) 20.8 (1997 meeting) 19.4 (1998 meeting) 21.4 (1999 meeting)	21.0 (overall) 33.2 (1990 meeting) 32.6 (1991 meeting) 33.7 (1992 meeting) 21.3 (1993 meeting) 14.4 (1995 meeting) 19.5 (1996 meeting) 16.1 (1997 meeting) 14.3 (1998 meeting) 10.0 (1999 meeting)
Elliot 2016	754 (2005 Congress of Neurological Surgeons)			14.8	‐21.2 to 64.2
Eloubeidi 2001	451 (1994 American Society for Gastrointestinal Endoscopy)	20	11.4	17.8	12.7 to 27.9 (IQR)
Ensom 1998	363 (1992 to 1996 Canadian Society of Hospital Pharmacists)	11
Erdağ 2014	1454 (2008 to 2010 Turkish National Otorhinolaryngology and Head & Neck Surgery)	18.6
Ersoy 2015a	243 (2009 Turkish Association of Gynecology and Obstetrics National Gynecology and Obstetrics Congress) 16 (oral) 227 (poster)	13 (oral) 11 (posters)
Ersoy 2015b	161 (2008 to 2010 National Gynecology and Obstetrics) 19 (oral) 142 (poster)	17.5 20.9 (oral) 16.2 (poster)	12.5 11.1 (oral) 12.7 (poster)
Evans 2015	1496 (2004 to 2007 American College of Clinical Pharmacy (ACCP) Spring and Fall national meetings, and regional meetings: Alcalde, Southwest Leadership, Eastern States, Southeastern Pharmacy Residency, Great Lakes Pharmacy Resident, Midwest Pharmacy Resident, and Western States)	24	15	24	1 to 94
Evers 2000	151 (1992 to 1997 European Society of Human Reproduction and Embryology)			32.5	0 to 79
Facione 2007	239 (overall; 2003 to 2006 Société interdisciplinaire francophone d’urodynamique et de pelvipérinéologie	21.6 (2003 meeting) 19.2 (2004 meeting) 13.5 (2005 meeting) 5.6 (2006 meeting)
Feldman 2015	1152 (2005 to 2012 Advancing Transfusion and Cellular Therapies Worldwide and American Society for Apheresis)			17	0 to 108
Fernandes 2008	63 (2001 American College of Sports Medicine, Brazilian authors) 100 (2001 American College of Sports Medicine; random sample) 100 (2001 Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul)	20.2 (2001 American College of Sports Medicine, Brazilian authors) 18.5 (American College of Sports Medicine, random sample) 23.3 (2001 Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul)	20 (2001 American College of Sports Medicine, Brazilian authors) 2.3 (American College of Sports Medicine, random sample) 5.5 (2001 Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul)	12 (2001 American College of Sports Medicine, Brazilian authors) 15 (American College of Sports Medicine, random sample) 12 (2001 Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul)	‐1 to 55 (2001 American College of Sports Medicine, Brazilian authors) ‐1 to 64 (American College of Sports Medicine, random sample) 6 to 48 (2001 Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul)
Fernandez 2011	279 (2003 to 2007 Australian Cardiac Rehabilitation Association)	19	21.6		‐4 to 78
Fosbol 2012	27208 (overall; 2006 to 2008 American Heart Association, American College of Cardiology, and European Society of Cardiology) 11365 ( American Heart Association) 5005 (American College of Cardiology) 10838 (European Society of Cardiology)			12 (overall) 13 (American Heart Association) 13 (American College of Cardiology) 12 (European Society of Cardiology)	7 to 19 (overall) 7 to 19 (IQR; American Heart Association) 6 to 18 (IQR, American College of Cardiology) 6 to 18 (IQR; European Society of Cardiology)
Freeman 2012	90 (overall) 64 (1995 to 2005 American Society of Clinical Oncology) 26 (2004 to 2005 American Association for Cancer Research)	34.8 (American Society of Clinical Oncology) 22.8 (American Association for Cancer Research)
Galang 2011	370 (2002 and 2003 American Dental Education Association Annual Session and Exhibitions)	16		10
Ginzburg 2014	140 (2008 to 2012 Society of Urodynamics, Female Pelvic Medicine, and Urogenital Reconstruction)	17.4	12
Glick 2006	1147 (2000 American Transplant Congress)	15.8	10.6	12
Goldman 1980	276 (1976 American Federation for Clinical Research, American Society for Clinical Investigation, Association for American Physicians, and American College of Cardiology meetings and the 1975 American Heart Association)			14
Goldman 1982	303 (976 American Federation for Clinical Research, American Society for Clinical Investigation, and Association of American Physicians meetings, and the 1975 American Society of Hematology and American Society of Nephrology)			15
Gorman 1990	104 (1984 American Association of Poison Control Centers) 165 (1986 American Association of Poison Control Centers, American Academy of Clinical Toxicologists, American Board of Medical Toxicologists, Canadian Association of Poison Control Centers)	19 (1984 American Association of Poison Control Centers) 12 (1986 American Association of Poison Control Centers, American Academy of Clinical Toxicologists, American Board of Medical Toxicologists, Canadian Association of Poison Control Centers)		12 (1984 American Association of Poison Control Centers) 9 (1986 American Association of Poison Control Centers, American Academy of Clinical Toxicologists, American Board of Medical Toxicologists, Canadian Association of Poison Control Centers)
Gourtaud 2009	443 (2000 and 2001 French Association of Urology)	16.9			65 (maximum)
Greenberg 2008	239 (2003 Society for Medical Decision Making)	20.5		19	0 to 47	17.6 to 23.3
Gregory 2012	888 (overall) 453 (Canadian Society of Plastic Surgeons) 262 (American Society of Plastic Surgeons) 173 (American Association of Plastic Surgeons)	22 (overall) 24 (Canadian Society of Plastic Surgeons) 19 (American Society of Plastic Surgeons) 22 (American Association of Plastic Surgeons)			1 to 85.9 (overall) 2 to 79.0 (Canadian Society of Plastic Surgeons) 1 to 54.3 (American Society of Plastic Surgeons) 1 to 85.9 (American Association of Plastic Surgeons)
Guryel 2006	415 (overall) 300 (British Orthopaedics Association) 115 (British Association for Surgery of the Knee, British Orthopedic Foot Surgical Sociey; British Elbow and Shoulder Society)	15.6 (overall) 16.2 (British Orthopaedics Association) 15.3 (British Association for Surgery of the Knee, British Orthopedic Foot Surgical Sociey; British Elbow and Shoulder Society)	11.1 (overall) 6.1 (British Orthopaedics Association) 9.3 (British Association for Surgery of the Knee, British Orthopedic Foot Surgical Sociey; British Elbow and Shoulder Society)	14 (overall) 14 (British Orthopaedics Association) 14 (British Association for Surgery of the Knee, British Orthopedic Foot Surgical Sociey; British Elbow and Shoulder Society)	‐7 to 56 (overall) ‐7 to 56 (British Orthopaedics Association) 1 to 48 (British Association for Surgery of the Knee, British Orthopedic Foot Surgical Sociey; British Elbow and Shoulder Society)
Ha 2008	1097 (2001 and 2002 Korean Radiological Society, Radiological Society of North America, and European Congress of Radiology)	15.8	13.8
Hajji 2016	270 (2006 to 2012 Société interdisciplinaire Francophone d’Uro‐dynamique et de Pelvi‐périnéologie)	22	15
Halpern 2002	145 (1994 and 1995 American Society of Anesthesiologists)	28	17
Hamlet 1997	1465 (1990 to 1992 meetings of the American Academy of Orthopaedic Surgeons	20
Harel 2011	300(overall) 100 (2006 National Kidney Foundation) 100 (2006 American Society of Nephrology 100 (2006 Annual European Renal Association ‐ European Dialysis and Transplant Association Congress)			12 (overall) 13 (National Kidney Foundation) 10 (American Society of Nephrology 12.5 (Annual European Renal Association ‐ European Dialysis and Transplant Association Congress)	6 to 22 (IQR; overall) 7 to 21 (IQR; National Kidney Foundation) 5 to 20 (IQR; American Society of Nephrology) 8 to 22.5 (IQR; Annual European Renal Association ‐ European Dialysis and Transplant Association Congress)
Hashkes 2003	331 (986, 1991 and 1998 Park City Pediatric Rheumatology)			24
Herbison 2004	63 (1998 and 1999 International Continence Society meetings)	19.2	9.6
Hernández‐Jiménez 2011	1045 (2000 to 2009 Annual Congress of the Sociedad Mexicana de Nutrición y Endocrinología)	25.2	22.8		12 to 108
Hoag 2006	1584 (2000 American Urological Association)	17
Hoeg 2009	559 (overall; 1997 to 1998 American Society of Clinical Oncology) 126 (1997 meeting) 177 (1998 meeting) 256 (2001 meeting)			41 (overall) 33 (1997 meeting) 41 (1998 meeting) 48 (2001 meeting)		37 to 46 (overall)
Hogan 2009	117 (2000, 2003, and 2006 American Contact Dermatitis Society)	18
Hopewell 2015	197 (overall) 8 (2010 Infectious Diseases Society of America) 25 (2010 American Thoracic Society) 34 (2010 American Society of Hematology) 57 (2010 American Society of Clinical Oncology) 7 (2010 American Psychiatric Association) 32 (2010 American Heart Association) 8 (2010 American Diabetes Association) 18 (2010 American College of Rheumatology)			14 (overall) 12.3 (2010 Infectious Diseases Society of America) 11.8 (2010 American Thoracic Society) 15.9 (2010 American Society of Hematology) 15.9 (2010 American Society of Clinical Oncology) 10 (2010 American Society of Anesthesiologists) 17.4 (2010 American Psychiatric Association) 10.9 (2010 American Heart Association) 8.2 (2010 American Diabetes Association) 5.7 (2010 American College of Rheumatology)	6.6 to 20.1 (IQR; overall) 6.5 to 22.9 (IQR; 2010 Infectious Diseases Society of America) 7.6 to 20 (IQR; 2010 American Thoracic Society) 9 to 23.4 (IQR; 2010 American Society of Hematology) 11.9 to 22.9 (IQR; 2010 American Society of Clinical Oncology) 5.5 to 14.5 (IQR; 2010 American Society of Anesthesiologists) 0.3 to 26.4 (IQR; 2010 American Psychiatric Association) 4.2 to 19.6 (IQR; 2010 American Heart Association) 2.7 to 20.2 (IQR; 2010 American Diabetes Association) 4.7 to 20.8 (IQR; 2010 American College of Rheumatology)
Hussein 2014	367 (2006 to 2011 Vascular Society of Great Britain and Ireland)			12	12 to 24 (IQR)
Jackson 2000	777 (1991 to 1994 Pediatric Orthopaedic Society of North America) 349 (accepted) 428 (rejected)			29 (accepted) 27 (rejected)
Jamjoom 2014	273 (2007 to 2012 Saudi Association of Neurological Surgery)			12
Jamjoom 2015	494 (2001 to 2005 Society of British Neurological Surgeons)	22.8			‐35 to 133
Jara‐Tracchia 2010	4137 (2002 to 2003 International Association for Dental Research)	48			36 to 60
Jasko 2003	336 (1991 to 1992, 1995, 1997 to 1999) Musculoskeletal Tumor Society	21.8	13.5
Jones 2008	139 (2000 International AIDS Conference)	19		16	‐2 to 59
Jones 2014	127 (2009 British Association of Urological Surgeons)	17.2
Juzych 1991	175 (overall) 100 (1985 Association for Research in Vision and Ophthalmology) 75 (1984 American Academy of Ophthalmology)	19.4 (1985 Association for Research in Vision and Ophthalmology) 13.0 (1984 American Academy of Ophthalmology)		13 (overall) 18 (1985 Association for Research in Vision and Ophthalmology) 11 (1984 American Academy of Ophthalmology)
Juzych 1993	327 (1985 Association for Research in Vision and Ophthalmology)	19.7		17
Jörgens 2014	493 (2004 European Association for the Study of Diabetes) 317 (accepted) 176 (rejected)	16.5 (accepted) 19.3 (rejected)	12.7 (accepted) 14 (rejected)
Kaifi 2013	76 (overall) 40 (Academic Surgical Congress meetings) 36 (Society of University Surgeons)	10 (overall) 13 (Academic Surgical Congress meetings) 11 (Society of University Surgeons)		12 (overall) 9 (Academic Surgical Congress meetings) 10 (Society of University Surgeons)	0 to 47 (overall) 0 to 47 (Academic Surgical Congress meetings) 0 to 38 (Society of University Surgeons)
Kalyoncu 2011	799 (2005 to 2009 Ulusal Romatoloji Kongrelerinin)	15.9	9.9
Kaya Mutlu 2013	181 (2002, 2004, 2006 and 2008 Advances in Physiotherapy Symposiums 18 (2002 meeting) 59 (2004 meeting) 42 (2006 meeting) 62 (2008 meeting)	20.9 (overall) 20.7 (2002 meeting) 22.7 (2004 meeting) 23.8 (2006 meeting) 14.8 (2008 meeting)	16.7 (overall) 11.5 (2002 meeting) 15.9 (2004 meeting) 22.2 (2006 meeting) 11.3 (2008 meeting)
Kearney 2012	1005 (1996 to 2005 Sir Peter Freyer Surgical Symposia)			19	0 to 132
Kim 1998	357 (1994 Korean Orthopaedic Association)	13
Kim 2011	288 (2001 to 2005 Korean Society of Spine Surgery)	12.2
Klappenbach 2011	200 (2006 Argentine Congress of Surgery)	15
Kleine‐Konig 2014	646 (2006 to 2008 European Paediatric Orthopaedic Society)	13.9	1.3
Kleweno 2008	165 (1999 to 2001 American Orthopaedic Society for Sports Medicine)			21	1 to 60
Kottachchi 2010	82 (1999 to 2003 Digestive Disease Week)			30	17 to 57 (IQR)
Koçak 2014	188 (2004 to 2012 European League Against Rheumatism Congress)	18.8			1 to 64
Krzyzanowska 2003	510 (1989 to 1998 American Society of Clinical Oncology)			32.4
Kumar 1995	275 (1992 American Society for Gastrointestinal Endoscopy) 46 (oral) 150 (poster)	13 (oral) 15 (poster)			6 to 24 (oral) 4 to 24 (poster)
Kunadian 2015	2119 (2002 to 2012 British Cardiovascular Society)		19.2	16.8
Kwong 2007	278 (1999 and 2001 European Federation of National Associations of Orthopaedics and Traumatology)	25.3	18.5
Lau 2016	460 (1996 to 2013 Otorhinolaryngology Research Society)	27.7		23
Lee 2012	346 (2004 and 2005 International Association for Dental Research)	26.4			0 to 67
Lim 2013	1027 (2008 to 2009 Korean Academy of Rehabilitation Medicine)	17.2 (overall) 15.5 (Spring 2008 meeting) 19.8 (Fall 2008 meeting) 16.6 (Spring 2009 meeting) 15.3 (Fall 2009 meeting)	10.5 (overall) 12.2 ( Spring 2008 meeting) 11.3 (Fall 2008 meeting) 10.3 Spring 2009 meeting) 8.8 (Fall 2009 meeting)
Liu 1996	400 (overall) 100 (American Heart Association) 100 (Federation of American Societies for Experimental Biology) 100 (American Gastroenterological Association) 100 (American Academy of Neurology)	22.8 (overall) 22.8 (American Heart Association) 19.2 (Federation of American Societies for Experimental Biology) 26.4 (American Gastroenterological Association) 22.8 (American Academy of Neurology)	10.8 (overall) 9.6 (American Heart Association) 10.8 (Federation of American Societies for Experimental Biology) 10.8 (American Gastroenterological Association) 10.8 (American Academy of Neurology)
Livas 2014	590 (2006 to 2007 European Orthodontic Society Congress) 333 (2006 meeting) 257 (2007 meeting)	15.8 (overall) 17.2 (2006 meeting) 14.6 (2007 meeting)
Lloyd 2006	292 (1998 to 2001 American Association of Hip and Knee Surgeons 40 (1998 meeting) 53 (1999 meeting) 49 (2000 meeting) 61 (2001 meeting)	21.7 (overall) 21 (1998 meeting) 22 (1999 meetings) 22 (2000 meeting) 16 (2001 meeting)	14.7 (overall)
Maguire 2014	577 (2004 to 2010 Reproductive Health Annual meetings)	18.5			‐6 to 78
Malicki 2014	504 (1989, 1993, 1997, 2001, 2005, and 2009 International Congress on Peer Review and Biomedical Publication)			14		12 to 16
Manuck 2015	3281 (2003 to 2010 Society for Maternal‐Fetal Medicine) 629 (oral) 2652 (poster)			19 (overall) 11 (oral) 21 (poster)	9 to 36 (IQR, overall) 9 to 24 (IQR, oral) 11 to 40 (IQR, poster)
Martinez 2008	388 (997, 1999, 2001, 2003, and 2005 Confederación Unificada Bioquímica de la República Argentina)	24	12		48 (maximum)
Marx 1999	527 (1993 American Society of Neuroradiologists and Radiological Society of North America)	15
Maxwell 1981	65 (1977 and 1978 Oncology Nursing Society)			11.5	1 to 30
McCormick 1985	1238 (overall) 371 (1976 to 1978 Ambulatory Pediatric Association) 194 (1979 to 1980 Ambulatory Pediatric Association) 347 (1976 to 1978 Society of Pediatric Research) 326 (1979 to 1980 Society of Pediatric Research)	20 (1976 to 1978 Ambulatory Pediatric Association) 12.9 (1979 to 1980 Ambulatory Pediatric Association) 23.1 (1976 to 1978 Society of Pediatric Research) 16.7 (1979 to 1980 Society of Pediatric Research)
McCue 2005	672 (1997 to 2000 American Society of Colon and Rectal Surgeons)	18.4
McKelvey 2010	272 (1981, 1991, and 2001 Southeastern Residency Conferences) 90 (1981 meeting) 70 (1991 meeting) 112 (2001 meeting)			22.8 (overall) 22 (1981 meeting) 33 (1991 meeting) 24 (2001 meeeting)	60.9 (maximum) 13 to 25 (IQR; 1981 meeting) 22 to 39 (IQR; 1991 meeting) 20 to 28 (IQR; 2001 meeting)
McKinley 2010	115 (2000, 2003, and 2006 American Contact Dermatitis Society)	21.6			3 to 72
Meissner 2014	172 (2007 and 2010 Canadian Association of Radiation Oncology)	18.5		16.5	‐10 to 54
Meral 2016	1368 (2008 to 2011 European Society for Surgical Research) 341 (2008 meeting) 304 (2009 meeting) 476 (2010 meeting) 247 (2011 meeting) 803 (oral) 565 (poster)	17.5 (overall) 23.2 (2008 meeting) 16.1 (2009 meeting) 15.3 (2010 meeting) 14.8 (2011 meeting) 19.1 (oral) 14.0 (poster)	20.7 (oral) 25.2 (poster)		‐166 to 82 (overall) ‐59 to 82 (2008 meeting) ‐60 to 67 (2009 meeting) ‐89 to 55 (2010 meeting) ‐166 to 46 (2011 meeting)
Meranze 1982	80 (1979 and 1980 International Anesthesia Research Society) 324 (1978 American Society of Anesthesiologists)	8.2 (1979 and 1980 International Anesthesia Research Society) 12.5 (1978 American Society of Anesthesiologists)
Micieli 2012	508 (2005 to 2008 Canadian Ophthalmological Society)			15	8 to 24 (IQR)
Miguel‐Dasit 2006a	2992 (1994, 1996, and 1998 Spanish Congress of Radiology ) 987 (1994 meeting) 1022 (1996 meeting) 983 (1998 meeting)	32 (overall) 35 (1994 meeting) 31 (1996 meeting) 30 (1998 meeting)	16 (overall) 16 (1994 meeting) 17 (1996 meeting) 14 (1998 meeting)		12 to 108 (overall) 12 to 108 (1994 meetiing) 12 to 96 (1996 meeting 12 to 72 (1998 meeting)
Mihok 2013	602 (2000 to 2009 British Association for Surgery of the Knee)	23			0 to 129
Mittal 2011	75 (1995 to 2007 Indian Academy of Pediatrics)	23.4	11.8		2 to 77
Moar 2013	318 (2000 to 2009 Craniofacial Society of Great Britain and Ireland) 24 (2000 meeting) 28 (2001 meeting) 32 (2002 meeting) 38 (2003 meeting) 34 (2004 meeting) 36 (2005 meeting) 31 (2006 meeting) 33 (2007 meeting) 40 (2008 meeting) 22 (2009 meeting)	29 (overall)		26 (overall) 27 (2000 meeting) 23 (2001 meeting) 68 (2002 meeting) 24 (2003 meeting) 48 (2004 meeting) 24 (2005 meeting) 14 (2006 meeting) 48 (2007 meeting) 35 (2008 meeting) 13 (2009 meeting)	2 to 110 (overall) ‐2 to 110 (2000 meeting) ‐2 to 48 (2001 meeting) 3 to 110 (2002 meeting) ‐1 to 62 (2003 meeting) 0 to 59 (2004 meeting) 24 to 37 (2005 meeting) 0 to 28 (2006 meeting) 15 to 59 (2007 meeting) 0 to 49 (2008 meeting) 0 to 26 (2009 meeting)
Montane 2007	248 (1994, 1996, and 1998 Spanish Society of Clinical Pharmacology)			18	2 to 60
Moorthi 2013	73 (2005 American Society of Nephrology)			24	13 to 36 (IQR)
Morgan 2005	131 (2001 British Association of Dermatology)	18
Mowla 2006	890 (2000 to 2003 Annual Research Meeting of Iranian Medical Sciences Students) 156 (oral) 636 (poster)	22.3 (oral) 29.5 (poster)
Muffly 2014	438 (2007 to 2008 American Urogynecologic Society) 147 (oral) 284 (poster)	15.7 (overall) 14 (oral) 17 (poster)	12.1 (overall) 11 (oral) 13 (poster)
Mutlu 2015	214 (2005 to 2008 National Congress of Child and Adolescent Psychiatry) 51 (2005 meeting) 64 (2006 meeting) 59 (2007 meeting) 40 (2008 meeting)	30.7 (overall) 35.3 (2005 meeting) 24.4 (2006 meeting) 33.8 (2007 meeting) 33.0 (2008 meeting)	18.9 (overall) 22.6 (2005 meeting) 16.1 (2006 meeting) 18.9 (2007 meeting) 18.2 (2008 meeting)		4 to 74 (overall)
Nasir 2012	128 (2004, and 2006 to 2009 Association of Paediatric Surgeons of Nigeria)	23.5		24	12 to 60
Nasir 2013	164 (2006, 2008, and 2010 Pan‐African Pediatric Surgical Association)			15	5 to 26 (IQR)
Ng 2004	4302 (1998 to 2000 American Urological Association) 1280 (1998 meeting) 1509 (1999 meeting) 1513 (2000 meeting)	36.9 (1998 meeting) 29.4 (1999 meeting) 23.2 (2000 meeting)				35.2 to 37.4 (1998 meeting) 28.6 to 30.1 (1999 meeting) 22.5 to 23.9 (2000 meeting)
Nguyen 1998	490 (1990 to 1995 Orthopedic Trauma Association)	16
O' Connor 2015	298 (1989 to 2014 Sir Peter Freyer Surgical Symposium) 61 (clinical research) 237 (basic science)	24 (overall) 24 (clinical) 24 (basic science)			‐12 to 108 (clinical) ‐12 to 96 (basic science)
O'Kelly 2015	322 (2005 to 2011 Irish Society of Urology) 34 (2005 meeting) 57 (2006 meeting) 37 (2007 meeting) 45 (2008 meeting) 45 (2009 meeting) 56 (2010 meeting) 49 (2011 meeting)	16.7 (overall) 8 (2005) 17 (2006) 28 (2007) 23 (2008) 13 (2009) 26 (2010) 12 (2011)		15	0 to 55
O'Neill 2014	203 (2002 to 2005 Irish Orthopaedic Association)	30.1			4.7 to 76.5
Odunsi 2015	627 (2007 to 2009 Association of British Neurologists)	18.6	14.6
Ogilvie 2014	392 (2006 to 2010 Canadian Society of Otolaryngology ‐ Head & Neck Surgery)	21			‐61 to 76
Olive 2004	249 (1996 Spanish Society for Rheumatology)	18.5
Oliveira 2009	313 (2003 Urological Brazilian meeting)			14	1 to 51
Oliver 2003	370 (1995 to 1997 British Association of Plastic Surgeons) 56 (1995 Winter meeting) 48 (1997 Summer meeting) 54 (1997 Winter meeting) 107 (1998 Winter meeting) 105 (1999 Winter meeting)	19 (1995 Winter meeting) 25 (1997 Summer meeting) 19 (1997 Winter meeting) 13 (1998 Winter meeting) 16 (1999 Winter meeting)			3 to 46 (1995 Winter meeting) 3 to 42 (1997 Summer meeting) 1 to 46 (1997 Winter meeting) 1 to 28 (1998 Winter meeting) 1 to 26 (1999 Winter meeting)
Olson 2012	270 (1995, 2000, and 2005 Western States Conference) 71 (1995 meeting) 77 (2000 meeting) 122 (2005 meeting)	24 (overall) 17 (1995 meeting) 24 (2000 meeting) 24 (2005 meeting)	10 (overall) 13 (1995 meeting) 13 (2000 meeting) 9 (2005 meeting)
Ospina 2006	383 (1995 to 2003 Society for Academic Emergency Medicine)			32		23 to 41
Ozel 2007	116 (2002 to 2003 Society of Gynecologic Surgeons)	16		9	‐4 to 48
Papagikos 2005	802 (1999 to 2001 American Society of Therapeutic Radiology and Oncology)			11.9
Papoutsis 2015	8411 (2006 to 2010 Deutsche Gesellschaft für Kardiologie)	8.4	14.4
Parkar 2013	137 (2008 to 2009 European Society of Musculoskeletal Radiology)			14.5	‐47 to 39
Paulson 2011	501 (2006 Center for International Blood and Marrow Transplantation Research/American Society for Blood and Marrow Transplantation)			19	1 to 41
Peng 2006	473 (1999 American Academy of Otolaryngology Head & Neck Surgery)			16	12 to 29 (IQR)
Pereyra‐Elias 2011	417 (1998, 2000, 2002, 2004, 2006, and 2008 Congreso de la Sociedad de Gastroenterología del Perú)	10.8	10.4	9	0 to 47
Pinheiro 2009	154 (2001 to 2005 American Society of Clinical Oncology)			13.5
Polyzos 2011	155 (2003 and 2004 European Society of Human Reproduction and Embryology)			15	0 to 75
Post 2013	1329 (overall) 1000 (2007 to 2008 North American Primary Care Research Group) 329 (2007 to 2008 Society of Teachers of Family Medicine) 688 (oral) 641 (poster)	15.4 (overall) 16 (North American Primary Care Research Group) 13.1 (Society of Teachers of Family Medicine) 13.7 (oral) 18.6 (poster)	0.6 (overall, SE) 0.6 ( North American Primary Care Research Group, SE) 1.2 (Society of Teachers of Family Medicine, SE) 0.6 (oral, SE) 1.0 (poster, SE)		‐20 to 51
Prendergast 2013	938 (1995 and 2005 British Society of Gastroenterology) 497 (1995 meeting) 441 (2005 meeting)	22.2 (1995 meeting) 18.6 (2005 meeting)	16.9 (1995 meeting) 13.3 (2005 meeting)
Preston 2005	815 (1994 to 1998 Orthopaedic Trauma Association) 153 (1994 meeting) 168 (1995 meeting) 156 (1996 meeting) 157 (1997 meeting) 181 (1998 meeting) 329 (oral) 486 (poster)	24.8 (all oral) 21.9 (oral; 1994 meeting) 26.1 (oral; 1995 meeting) 31.1 (oral; 1996 meeting) 20.9 (oral; 1997 meeting) 25.4 (oral; 1998 meeting) 21.6 (all poster) 28.2 (poster; 1994 meeting) 19.7 (poster; 1995 meeting) 22.1 (poster; 1996 meeting) 18.6 (poster; 1997 meeting) 21.3 (poster; 1998 meeting)	17.9 (all oral) 18.1 (oral; 1994 meeting) 20.3 (oral; 1995 meeting) 19.3 (oral; 1996 meeting) 13.8 (oral; 1997 meeting) 16.8 (oral; 1998 meeting) 19.4 (all poster) 27.8 (poster; 1994 meeting) 18.8 (poster; 1995 meeting) 17.3 (poster; 1996 meeting) 15.9 (poster; 1997 meeting) 17.3 (poster; 1998 meeting)
Prohaska 2013	1941 (2005 American College of Clinical Pharmacy, American Pharmacists Association, and American Society of Health‐System Pharmacists)	16.8	11.9	15
Propst 2015	867 (2004 to 2012 Society of Gynecologic Surgeons)	12.7	11.7		‐5 to 82
Rabenda 2015	619 (2011 European Congress on Osteoporosis, Osteoarthritis, and Musculoskeletal Disorders) 45 (oral) 574 (poster)			14.8 (oral e‐publication) 16.3 (oral print) 9.2 (poster e‐publication) 11.3 (poster print)	8.3 to 23.3 (IQR, oral e‐pub) 8.4 to 23.3 (IQR, oral print) 2.7 to 17.7 (IQR poster e‐pub) 5.3 to 21.7 (IQR, poster print)
Riessland 2004	347 (1995 American Society for Clinical Pharmacology)	31	47	31
Rollin 2009	318 (2000 International Commission on Occupational Health)	17				13 to 21
Roukis 2011	206 (199 to 2008 American College of Foot and Ankle Surgeons) 29 (1999 meeting) 29 (2000 meeting) 20 (2001 meeting) 20 (2002 meeting) 13 (2003 meeting) 14 (2004 meeting) 21 (2005 meeting) 21 (2006 meeting) 21 (2007 meeting) 18 (2008 meeting)	14.5 (overall) 13.3 (1999 meeting) 19.3 (2000 meeting) 18.6 (2001 meeting) 13.7 (2002 meeting) 16.9 (2003 meeting) 8.4 (2004 meeting) 17.9 (2005 meeting) 11.5 (2006 meeting) 14.7 (2007 meeting) 7.6 (2008 meeting)			‐34 to 60 (overall) ‐34 to 60 (1999 meeting) 2 to 50 (2000 meeting) 12 to 34 (2001 meeting) 2 to 42 (2002 meeting) 6 to 36 (2003 meeting) 0 to 24 (2004 meeting) ‐2 to 46 (2005 meeting) ‐5 to 32 (2006 meeting) ‐ 4 to 36 (2007 meeting) ‐1 to 17 (2008 meeting)
Roy 2001	660 (1978 to 1995 Otorhinolaryngological Research Society)	2235			‐60 to 108
Rubin 2014	847 (2004 to 2007 Digestive Disease Week)			12	12 to 24 (IQR)
Saad 2008	154 (2001 to 2005 American Society of Clinical Oncology)			13.5
Saeed 2011	141 (2002, 2004, and 2006 Canadian Blood and Marrow Transplant Group)			20.7	‐5.3 to 91.2
Salami 2013	184 (2005 American Society of Pediatric Hematology/Oncology)			19	11 to 29 (IQR)
Saldanha 2016	513 (Association for Research in Vision and Ophthalmology)			18	1 to 90 (IQR, 11 to 33)
Sanders 2001	255 (1994 British Society of Gastroenterology)			19	0 to 66
Sanossian 2006	353 (2000 International Stroke Conference of the American Stroke Association)			15	‐18 to 69
Sawatsky 2015	144 (2009 Society of General Internal Medicine)	21
Schnatz 2008	661 (1999 to 2003 North American Menopause Society)	24.0 (overall) 20.4 (oral) 24.0 (poster)	18.0 (overall) 15.6 (oral) 18.0 (poster)		103.2 (maximum; overall)
Schoenfeld 2012	770 (1998 to 2006 Society of Military Orthopaedic Surgeons) 96 (1998 meeting) 89 (1999 meeting) 104 (2000 meeting) 75 (2001 meeting) 84 (2002 meeting) 83 (2003 meeting) 86 (2004 meeting) 76 (2005 meeting) 78 (2006 meeting)	32.4 (overall) 43.0 (1998 meeting) 33.0 (1999 meeting) 33.6 (2000 meeting) 30.6 (2001 meeting) 39.6 (2002 meeting) 29.4 (2003 meeting) 27.0 (2004 meeting) 28.5 (2005 meeting) 19.2 (2006 meeting)
Scholey 2005	546 (1993 European Organisation for Caries Research, European Orthodontic Society, and International Association for Dental Research)			18.0 (overall) 13.0 (European Organisation for Caries Research meeting) 23.5 (European Orthodontic Society meeting) 17.0 (International Association for Dental Research meeting)	9 to 30 (overall; IQR) 5.3 to 24.5 (IQR; European Organisation for Caries Research meeting) 12.3 to 36.0 (IQR; European Orthodontic Society meeting) 8.8 to 29.3 (IQR; International Association for Dental Research meeting)
Schulte 2012a	1100 (2003 German Society of Orthopaedics and Trauma Surgery)	14.9	16.9		‐44 to 58
Schulte 2012b	839 (2000 to 2003 Spine Society of Europe) 259 (2000 meeting) 213 (2001 meeting) 168 (2002 meeting) 199 (2003 meeting)	17.7 (overall) 16.2 (2000 meeting) 18.3 (2001 meeting) 14.4 (2002 meeting) 21.2 (2003 meeting)	15.7 (overall) 16.5 (2000 meeting) 17.0 (2001 meeting) 14.4 (2002 meeting) 14.2 (2003 meeting)		‐36 to 60 (overall) ‐36 to 60 (2000 meeting) ‐30 to 59 (2001 meeting) ‐9 to 57 (2002 Smeeting) ‐2 to 56 (2003 meeting)
Seaton 1983	625 (1978 American Speech‐Hearing‐Language Association)			13
Sebel 2001	581 (1995 International Anesthesia Research Society)	15.6
Secil 2005	4413 (Turkish radiology conferences)	24.4			0 to 91
Secil 2007	276 (2000 and 2001 European Society of Gastrointestinal and Abdominal Radiology)			18	1 to 53 (IQR)
Shelmerdine 2015	363 (overall) 35 (British Society for Paediatric Radiology) 163 (European Society of Pediatric Radiology) 165 (Society of Pediatric Radiology)			12 (British Society for Paediatric Radiology) 10 (European Society of Pediatric Radiology) 15 (Society of Pediatric Radiology)	6 to 17 (British Society for Paediatric Radiology) 2 to 29 (European Society of Pediatric Radiology) 3 to 31 (Society of Pediatric Radiology)
Siegfried 2007	65 (African studies) 62 (North American studies)	22.8 (African studies) 26.4 (North American studies)			12 to 84 (African studies) 12 to 72 (North American studies)
Smart 2013	122 (2000 to 2010 American Association of Oral and Maxillofacial Surgeons and American International Association of Dental Research)	21.9	17.3	19	0 to 99
Smith 2007	1683 (2002 and 2003 American Urological Association)			27.8	25.9 to 39.7
Smith 2011a	823 (2000 to 2004 American Physical Therapy Association)	22.7	15.3	19.3
Smollin 2006	237 (2001 North America Conference of Clinical Toxicology)	15			0 to 29
Snedeker 2010	149 (various conferences with abstracts on pre‐harvest or abattoir‐level intervention against pathogenic bacteria)			13.5	0 to 72
Song 2010	4824 (2005 to 2007 United States and Canadian Academy of Pathology)	18
Stranges 2015	655 (2003, 2005, and 2007 Great Lakes Pharmacy Residency Conference)	24.5	14.3
Stöcker 2009	368 (1999 to 2003 Deutsche Gesellschaft für Allgemeinmedizin und Familienmedizin)	18.4			‐24 to 84
Sullivan 2014	227 (2006 to 2010 Irish Society of Urology)	24 (overall) 26 (2006 meeting) 33 (2007 meeting) 22 (2008 meeting) 20 (2009 meeting) 18 (2010 meeting)
Sun 2011	2080 (2003 to 2006 American Academy of Ophthalmology and Association for Research in Vision and Ophthalmology) 617 (2003 meeting) 615 (2004 meeting) 616 (2005 meeting) 616 (2006 meeting)			16 (overall) 20 (2003 meeting) 18 (2004 meeting) 14.5 (2005 Ameeting) 15 (2006 meeting)	8.0 to 27.1 (IQR, overall) 9.1 to 32.0 (IQR, 2003 meeting) 9.1 to 29 (IQR, 2004 meeting) 8.0 to 26.5 (IQR, 2005 meeting) 7.0 to 22.6 (IQR, 2006 meeting)
Susarla 2015	78 (2009 to 2013 Johns Hopkins Hospital/University of Maryland Plastic Survey Research Day)	15.6	13.6		‐2 to 54
Tam 2008	74 (2000 American Society of Clinical Oncology)	31		29	3.4 to 68.4
Tam 2011	709 (1989 to 2003 American Society of Clinical Oncology)			31.2
Tambuscio 2010	623 (2006 American Academy of Forensic Sciences)	10	9
Theman 2014	719 (2000 to 2010 American Society for Surgery of the Hand) 109 (2000 meeting) 69 (2001 meeting) 64 (2002 meeting) 36 (2003 meeting) 80 (2004 meeting) 53 (2005 meeting) 35 (2006 meeting) 88 (2007 meeting) 62 (2008 meeting) 58 (2009 meeting) 65 (2010 meeting)	18 (overall) 22 (2000 meeting) 25(2001 meeting) 22 (2002 meeting) 15 (2003 meeting) 20 (2004 meeting) 19 (2005 meeting) 21 (2006 meeting) 15 (2007 meeting) 12 (2008 meeting) 15 (2009 meeting) 13 (2010 meeting)		14 (overall) 15 (2000 meeting) 14(2001 meeting) 14 (2002 meeting) 15 (2003 meeting) 18 (2004 meeting) 11 (2005 meeting) 20 (2006 meeting) 14 (2007 meeting) 11 (2008 meeting) 13 (2009 meeting) 12 (2010 meeting)	‐14 to 64 (overall) 120 (maximum); 10 to78 IQR; 2000 meeting) 96 (maximum); 12 to 101 IQR; 2001 meeting) 120 (maximum); 21 to 69 IQR; 2002 meeting) 48 (maximum); 7 to 68 IQR; 2003 meeting) 120 (maximum); 11 to 69 IQR; 2004 meeting) 60 (maximum); 13 to 84 IQR; 2005 meeting) 84 (maximum); 30 to 66 IQR; 2006 meeting) 60 (maximum); 18 to 51 IQR; 2007 meeting) 48 (maximum); 15 to 44 IQR; 2008 meeting) 48 (maximum); 17 to 66 IQR; 2009 meeting) 36 (maximum); 17 to 55 IQR; 2010 meeting) 84
Timmer 2001	594 (1994 to 1995 European Pancreatic Club and American Pancreatic Association)			36
Timmer 2002	863 (1992 to 1995 Digestive Diseases Week)			18
Toma 2006	186 (1999 to 2002 American College of Cardiology) 86 ('late‐breaking' abstracts) 100 (not 'late‐breaking' abstracts)			11.5 ('late‐breaking' abstracts) 22 (not 'late‐breaking' abstracts)
Toro‐Polo 2012	532 (2002 to 2005 and 2006 to 2009 Sociedad Científica Médico Estudiantil Peruana)	22.6	24.0	13	0 to 75
Turpen 2010	126 (2002 and 2003 American Urological Association)	51	5
Tyagi 2013	726 (2009 American Society of Anesthesiologists meeting and Indian Society of Anaesthesiologists)			14 (American Society of Anesthesiologists) 12 (Indian Society of Anaesthesiologists’ Conference)	1 to 21 (American Society of Anesthesiologists) 3 to 24 (Indian Society of Anaesthesiologists’ Conference)
ul Haq 2011	179 (2001 British Orthopaedic Association)	18.6	9.4
Uysal 2008	900 (2002 International Federation of Clinical Chemistry and Laboratory Medicine)	11.9		8
Uzun 2013	179 (2005 Undersea and Hyperbaric Medical Society)	18.5	13.6	15	2 to 60
Valderrama‐Zurian 2009	689 (1999 College on Problems of Drug Dependence ) 194 (oral) 495 (poster)	22.4 (overall) 22.4 (oral) 22.4 (poster)	17.3 (overall) 18.5 (oral) 16.4 (poster)	19.9 (overall) 19.9 (oral) 19.9 (poster)
van der Steen 2004	306 (1995 to 1999 European Association of Plastic Surgeons)	14	23.9	13
Varghese 2011	1191 (2003 to 2005 Pediatric Orthopaedic Society of North America) 431 (2003 meeting) 343 (2004 meeting) 417 (2005 meeting)	24.0 (2003 meeting) 24.8 (2004 meeting) 20.2 (2005 meeting) 21.9 (2003 meeting; oral) 28.3 (2004 meeting; oral) 20.7 (2005 meeting; oral) 21.4 (2003 meeting; rejected abstracts) 22.2 (2004 meeting; rejected abstracts) 17.6 (2005 meeting; rejected abstracts)		18 (2003 meeting) 22 (2004 meeting) 17 (2005 meeting) 18 (2003 meeting; oral) 26 (2004 meeting; oral) 18.5 (2005 meeting; oral) 18 (2003 meeting; rejected abstracts) 19.5 (2004 meeting; rejected abstracts) 15 (2005 meeting; rejected abstracts)		20.3 to 27.8 (2003 meeting) 21.4 to 28.3 (2004 meeting) 17.7 to 22.8 (2005 meeting) 17.6 to 26.2 (2003 meeting; oral) 23.2 to 33.5 (2004 meeting; oral) 17.2 to 24.0 (2005 meeting; oral) 18.5 to 24.2 (2003 meeting; rejected abstracts) 19.0 to 25.4 (2004 meeting; rejected abstracts) 15.0 to 20.1 (2005 meeting; rejected abstracts)
Vecchi 2009	581 (1993 to 2002 College on Problems of Drug Dependence) 161 ('positive results') 51 (negative or null results) 325 (abstracts without reported statistical results) 44 (abstracts with no results)	45.6 (overall) 34.8 ('positive results') 62.4 (negative or null results) 50.4 (abstracts without reported statistical results) 62.4 (abstracts with no results)
Vogt 2008	36 (1997 Deutscher Sportärztekongress)	12
von Hardenberg 2013	732 (2002 and 2009 Der Kongress der Deutschen Gesellschaft für Urologie)			15.1
Walby 2001	207 (1995 to 1998 Australasian College of Emergency Medicine and Australasian Society for Emergency Medicine and 1996 and 1998 International Conference on Emergency Medicine)	12.6		11
Walsh 2013	449 (2005 to 2006 Research in Medical Education Conference and Canadian Conference on Medical Education)			20	‐9 to 76 (10 to 30 IQR)
Whitehouse 2009	1138 (overall) 90 (2004 British Hip Society) 73 (2005 British Hip Society meeting) 68 (2004 European Hip Society meeting) 173 (2006 European Hip Society meeting) 23 (2004 British Orthopedic Association meeting) 31 (2005 British Orthopedic Association meeting) 41 (2006 British Orthopedic Association meeting) 293 (2003 European Federation of Orthopaedics and Trauma meeting) 346 (2005 European Federation of Orthopaedics and Trauma meeting)	4.9 (overall) 6.9 (2004 British Hip Society meeting) 6.8 (2005 British Hip Society meeting) ‐4.1 (2004 European Hip Society meeting) 0.7 (2006 European Hip Society meeting) 8.3 (2004 British Orthopedic Association meeting) 6.0 (2005 British Orthopedic Association meeting) 2.0 (2006 British Orthopedic Association meeting) 9.5 (2003 European Federation of Orthopaedics and Trauma meeting) 8.3 (2005 European Federation of Orthopaedics and Trauma meeting)
Wieser 2015	607 (2004 to 2008 American College of Veterinary Anesthesiologists and Association of Veterinary Anaiesthsists; 2003 and 2006 World Congress of Veterinary Anaesthesiology)	25 (overall) 25 (RCTs) 26 (uncontrolled studies) 15 (cross‐sectional studies)	19		‐35 to 104 (overall) 23 to 27 (RCTs) 23 to 30 (uncontrolled studies) 15 (9 to 22) cross‐sectional studies)
Wong 2006	67 (1998 and 1999 New South Wales State Committee, Registrars Paper Day	23.8		21	4 to 60
Wong 2009	790 (2003 Heart Rhythm Society			16.7	10.6 to 27.6 (IRQ)
Wong 2013	3848 (1994, 1997, 2000, 2003, and 2006 Heart Rhythm Society) 521 (1994 meeting) 740 (1997 meeting) 804 (2000 meeting) 790 (2003 meeting) 993 (2006 meeting)			21 (1994 meeting) 18 (1997 meeting) 16 (2000 meeting) 18 (2003 meeting) 13 (2006 meeting)	10 to 34 (IQR, 1994 meeting) 11 to 32 (IQR, 1997 meeting) 6 to 31 (IQR, 2000 meeting) 8 to 27 (IQR, 2003 meeting) 5 to 22 (IQR, 2006 meeting)
Yalcinkaya 2013	770 (2007 Turkish National Orthopedics and Traumatology Congress)	14.9	16.1		‐33 to 55
Yilmaz 2013	1732 (2008 European League Against Rheumatism)			13	0 to 31
Yolcu 2015	1322 (2007 to 2012 Oral and Maxillofacial Surgery Society of Turkey)	17	15
Yoon 2007	154 (1996 to 2001 American Association of Hip and Knee Surgeons) 10 (1996 meeting) 20 (1997 meeting) 18 (1998 meeting) 20 (1999 meeting) 40 (2000 meeting) 46 (2001 meeting)	23.6 (overall) 20.8 (1996 meeting) 36.5 (1997 meeting) 27.3 (1998 meeting) 23.5 (1999 meeting) 9.7 (2000 meeting) 16.6 (2001 meeting)	15.7 (overall) 19.1 (1996 meeting) 19.0 (1997 meeting) 11.6 (1998 meeting) 9.9 (1999 meeting) 14.5 (2000 meeting) 15.8 (2001 meeting)
Yoon 2012a	154 (2005 First World Conference on Sports Injury Prevention) 14 (RCTs) 135 (observational studies)	17 (RCTs) 12 (observational studies)	13 (RCTs) 14 (observational studies)
Yoon 2012b	614 (2005 to 2009 Urological Society of Australia and New Zealand	13.0 (online journal publications) 14.5 (print journals publications)	13.6 (online journal publications) 13.9 (print journal publications)
Yun 1999	2457 (1992 to 1996 Korean Radiological Society)	13.3			0 to 56
Özyurt 2012	1023 (2004, 2006, and 2008 Turkish National Dermatology) 368 (2004 meeting) 328 (2006 meeting) 327 (2008 meeting)	22.5 (overall) 23.2 (2004 meeting) 27.2 (2006 meeting) 14.5 (2008 meeting)			0 to 86

Comparison 3. 'Positive' versus 'not positive' results.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Publication by 'positive' versus 'not positive' results	64		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 defined as 'significant results'	45	15783	Risk Ratio (M‐H, Random, 95% CI)	1.31 [1.23, 1.40]
1.2 defined by direction of results	34	8794	Risk Ratio (M‐H, Random, 95% CI)	1.17 [1.07, 1.28]
2 Publication by 'positive' versus 'not positive results,' randomized or controlled clinical trials	20		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.1 defined as 'significant results'	15	2616	Risk Ratio (M‐H, Random, 95% CI)	1.21 [1.10, 1.32]
2.2 defined by direction of results	13	2307	Risk Ratio (M‐H, Random, 95% CI)	1.17 [1.04, 1.32]

Comparison 4. Sample size.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Publication by larger versus smaller sample size	23	8689	Risk Ratio (M‐H, Random, 95% CI)	1.09 [0.99, 1.19]
2 Publication by larger versus smaller sample size; randomized or controlled clinical trials	6	727	Risk Ratio (M‐H, Random, 95% CI)	1.25 [1.08, 1.46]

Comparison 5. Oral versus poster presentations.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Publication by oral versus poster presentations	143	115910	Risk Ratio (M‐H, Random, 95% CI)	1.46 [1.40, 1.52]
2 Publication by oral versus poster presentations, randomized or controlled clinical trials	9	1841	Risk Ratio (M‐H, Random, 95% CI)	1.32 [1.13, 1.54]

Comparison 6. Accepted abstracts versus rejected abstracts.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Publication by acceptance versus rejection for presentation	22	22319	Risk Ratio (M‐H, Random, 95% CI)	1.65 [1.48, 1.85]

Comparison 7. Clinical versus basic research.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Publication by clinical research versus basic science	92	97372	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.74, 0.82]

Comparison 8. Study design.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Publication by randomized trial versus other study designs	47	28928	Risk Ratio (M‐H, Random, 95% CI)	1.51 [1.36, 1.67]
2 'Prospective' versus 'retrospective'	30	14938	Risk Ratio (M‐H, Random, 95% CI)	1.17 [1.06, 1.30]
3 Level of evidence	7	3344	Risk Ratio (M‐H, Random, 95% CI)	1.50 [1.20, 1.86]

Comparison 9. Multi‐centered versus single center.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Multi‐centered versus single center	44	22780	Risk Ratio (M‐H, Random, 95% CI)	1.32 [1.21, 1.44]
2 Multi‐centered versus single center, randomized or controlled clinical trials	7	1539	Risk Ratio (M‐H, Random, 95% CI)	1.47 [1.09, 1.98]

Comparison 10. Higher versus lower quality.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Higher versus lower quality	12	3364	Risk Ratio (M‐H, Random, 95% CI)	1.46 [1.23, 1.73]
2 Higher versus lower quality, randomized or controlled clinical trials	3	215	Risk Ratio (M‐H, Random, 95% CI)	1.09 [0.84, 1.42]

Comparison 11. Abstract author characteristics.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Academic center versus non‐academic center	34	16913	Risk Ratio (M‐H, Random, 95% CI)	1.60 [1.34, 1.92]
2 Female versus male	6	10935	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.77, 1.03]

Comparison 12. Impact.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 High versus low importance or impact	11	6982	Risk Ratio (M‐H, Random, 95% CI)	1.60 [1.41, 1.82]
1.1 Award winning	7	6405	Risk Ratio (M‐H, Random, 95% CI)	1.59 [1.39, 1.82]
1.2 Clinically important	1	166	Risk Ratio (M‐H, Random, 95% CI)	1.26 [0.99, 1.60]
1.3 Other	3	411	Risk Ratio (M‐H, Random, 95% CI)	1.97 [1.20, 3.23]

Comparison 13. Funding.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Funding reported versus no funding or funding not reported	13	5832	Risk Ratio (M‐H, Random, 95% CI)	1.48 [1.27, 1.73]
2 Industry funding versus other type or no funding	12	3938	Risk Ratio (M‐H, Random, 95% CI)	1.18 [1.00, 1.40]

Comparison 14. North America versus Europe versus other origin.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 North America versus Other	52	19984	Risk Ratio (M‐H, Random, 95% CI)	1.13 [1.04, 1.22]
2 Europe versus Other	44	21281	Risk Ratio (M‐H, Random, 95% CI)	1.24 [1.11, 1.39]
3 North America versus Europe	41	20669	Risk Ratio (M‐H, Random, 95% CI)	1.04 [0.97, 1.13]
4 USA versus Other	24	45931	Odds Ratio (M‐H, Fixed, 95% CI)	1.05 [1.01, 1.09]
5 UK versus Other	3	1070	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.67, 1.21]
6 Australia/New Zealand versus Other	2	1262	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.54, 0.84]
7 Brazil versus Other	2	966	Risk Ratio (M‐H, Random, 95% CI)	1.26 [0.22, 7.02]

Comparison 15. English language vs non‐English language.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 English language versus non‐English language	8	4825	Risk Ratio (M‐H, Random, 95% CI)	1.16 [0.96, 1.41]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Abes 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by first author Matching criteria not reported
Data	Included 776 abstracts presented at the 1996 to 2000 Turkish Association of Pediatric Surgeons meetings Included all abstracts
Comparisons	Proportion of abstracts published Oral (free paper) versus poster presentation
Outcomes	156 of 776 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included  101/394 abstracts presented orally versus 55/382 abstracts presented as posters published
Notes	Surgery ‐ pediatric surgery Funding not reported Unable to obtain full‐length article; information extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation for association with publication using stratified analysis.

Abicht 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar to April 2011 Search completed by investigator Searched by all authors, keywords, and title Matched abstract to full‐length publication by All authors Title
Data	Included 825 abstracts presented at the 1999 to 2008 American College of Foot & Ankle Surgeons meetings Included all abstracts accepted as posters except student posters and commercially sponsored posters
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting
Outcomes	198 of 825 abstracts published 6/20 abstracts presented at the 1999 meeting, 9/35 at the 2000 meeting, 11/37 at the 2001 meeting, 16/52 at the 2002 meeting, 21/59 at the 2003 meeting, 12/74 at the 2004 meeting, 29/107 at the 2005 meeting, 37/142 at the 2006 meeting, 28/150 at the 2007 meeting, and 29/149 at the 2008 meeting published Proportion of abstracts published by time Mean time to publication = 17.6 months (range = ‐46 to 75 months) Mean time to publication = 4.8 months (range = ‐12 to 40 months) for 1999 American College of Foot & Ankle Surgeons meeting Mean time to publication = 2.9 months (range = ‐14 to 16 months) for 2000 American College of Foot & Ankle Surgeons meeting Mean time to publication = 19.2 months (range = 6 to 34 months) for 2001 American College of Foot & Ankle Surgeons meeting Mean time to publication = 17.9 months (range = 1 to 46 months) for 2002 American College of Foot & Ankle Surgeons meeting Mean time to publication = 23 months (range = 3 to 75 months) for 2003 American College of Foot & Ankle Surgeons meeting Mean time to publication = 24.1 months (range = 6 to 60 months) for 2004 American College of Foot & Ankle Surgeons meeting Mean time to publication = 20.8 months (range = ‐46 to 72 months) for 2005 American College of Foot & Ankle Surgeons meeting Mean time to publication = 22.2 months (range = ‐24 to 54 months) for 2006 American College of Foot & Ankle Surgeons meeting Mean time to publication = 12.7 months (range = ‐7 to 33 months) for 2007 American College of Foot & Ankle Surgeons meeting Mean time to publication = 13.4 months (range = ‐22 to 30 months) for 2008 American College of Foot & Ankle Surgeons meeting No factors other than meeting year related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery ‐ foot and ankle Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all posters with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings before 2008 had at least 48 months follow‐up. The 2008 meeting only had 36 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Abuzeid 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database Web of Science for one month prior until two years following each meeting Person completing the search not reported Searched by first and last author, and keywords Matching criteria not reported
Data	Included 3565 abstracts presented at the 2006 to 2010 Canadian Cardiovascular Congress meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Scientific category Multivariable analysis examining oral versus poster, institutional location, year of presentation, and 'scientific category'
Outcomes	858 of 3565 abstracts published 135/569 abstracts presented at the 2006 meeting, 178/758 at the 2007 meeting, 190/862 at the 2008 meeting, 197/741 at the 2009 meeting, and 158/635 at the 2010 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included Multivariable analysis showed type of presentation (adjusted OR for poster versus oral 0.71, 95% CI = 0.60 to 0.83) and scientific category were predictors of publication
Notes	Cardiology ‐ cardiovascular research Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings only had 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of more than 2 factors with publication using multivariable logistic regression analysis.

Abzug 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar in 2010 Person completing the search not reported Searched by all authors and title Matched abstract to full‐length publication by All authors Keywords Title
Data	Included 1127 abstracts presented at the 2000 to 2005 American Society for Surgery of the Hand meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Oral versus poster presentation Subspecialty
Outcomes	518 of 1127 abstracts published 94/249 abstracts presented at the 2000 meeting, 75/183 at the 2001 meeting, 100/177 at the 2002 meeting, 68/150 at the 2003 meeting, 96/192 at the 2004 meeting, and 85/176 at the 2005 meeting published Proportion of abstracts published by time Mean time to publication = 25.2 months, overall Mean time to publication = 28.8 months for the 2000 American Society for Surgery meeting Mean time to publication = 27.6 months for the 2001 American Society for Surgery meeting Mean time to publication = 31.2 months for the 2002 American Society for Surgery meeting Mean time to publication = 25.2 months for the 2003 American Society for Surgery meeting Mean time to publication = 20.4 months for the 2004 American Society for Surgery meeting Mean time to publication = 19.2 months for the 2005 American Society for Surgery meeting Factors related to proportion of abstracts published included 231/425 abstracts presented orally versus 287/702 abstracts presented as poster published
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 60 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year by type presentation with publication using stratified analysis.

Acevedo 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database Google/Google Scholar in July 2011 Person completing the search not reported Searched by first author and title Matching criteria not reported
Data	Included 1972 abstracts presented at the 2000 to 2010 American Society of Clinical Oncology, American Society of Hematology, European Society of Medical Oncology, European Cancer Organization, and American Society for Radiation Oncology meetings Included all abstracts from odd years from Latin American countries
Comparisons	Proportion of abstracts published IQR for time to publication Oral versus poster presentation Latin American countries
Outcomes	353 of 1972 abstracts published Proportion of abstracts published by time IQR = 12 to 24 months (range = 0 to 96 months) Factors related to proportion of abstracts published included 17/37 abstracts presented orally versus 130/718 abstracts presented as poster published
Notes	Oncology ‐ hematology Reporting having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts from odd years.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All abstracts had at least 48 months follow‐up except the meetings in 2008 to 2010.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation and country with publication using stratified analysis.

Adhikari 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and PubMed to December 2009 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Title Contents
Data	Included 1479 abstracts presented at the 1999 to 2008 Society for Academic Emergency Medicine & American College of Emergency Physicians meetings Included all abstracts accepted for presentation that described emergency ultrasound, emergency medical services, toxicology
Comparisons	Proportion of abstracts published Mean time to publication by topic of abstract Emergency ultrasound versus emergency medical services versus toxicology
Outcomes	333 of 1479 total abstracts published Proportion of abstracts published by time Mean time to publication for abstracts on emergency ultrasound = 20.4 months (95% CI = 4.8 to 36 months) Mean time to publication for abstracts on emergency medical services = 22.8 months (95% CI = 6 to 39.6 months) Mean time to publication for abstracts on toxicology = 21.6 months (95% CI = 3.1 to 40.1 months) Factors related to proportion of abstracts published included  103/452 abstracts related to emergency ultrasound versus 152/651 abstracts related to emergency medical services versus 78/376 abstracts related to toxicology published
Notes	Emergency medicine ‐ ultrasound Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific topics, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings up until 2005 had at least 48 months follow‐up. The meetings from 2006 to 2009 had less follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of emergency medicine topic with publication using stratified analysis with Fisher's Exact and Kruskal‐Wallis test.

Aggarwal 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and IndMed; dates of search not reported Person completing the search not reported Searched by first author Matching criteria not reported
Data	Included 271 abstracts presented at the 2007, 2009, and 2010 Indian Rheumatology Association meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation
Outcomes	19 of 271 abstracts published 2/60 abstracts presented at the 2007 meeting, 11/105 at the 2009 meeting, and 6/106 at the 2010 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 15% of abstracts presented orally versus 4% of abstracts presented as poster published
Notes	Rheumatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and type presentation with publication using stratified analysis.

Ahlers‐Schmidt 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar; dates of search not reported Person completing the search not reported Searched by all authors, keywords, and title Matching criteria not reported
Data	Included 67 abstracts presented at the 1999 to 2006 Midwestern Regional Medical School Annual Research Forum Included all abstracts accepted as posters at meeting
Comparisons	Proportion of abstracts published Presented by residents versus students Surgery versus non‐surgical topics
Outcomes	23 of 67 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included  20/52 abstracts presented by residents versus 3/15 abstracts presented by students published 15/37 abstracts related to surgery versus 8/30 abstracts related to non‐surgical topics published
Notes	General medicine/primary care Funding not reported Information extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented as posters.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type author and topic (surgery versus non‐surgery) with publication using stratified analysis and Chi2 tests.

Akbari‐Kamrani 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database The Cochrane Library, Embase, and PubMed to February 2007 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Same intervention One outcome
Data	Included 198 abstracts presented at the 2000 to 2003 American Society for Laser Medicine and Surgery meetings Included all abstracts describing clinical trials
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Survival analysis of proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation RCT design versus CCT design US versus non‐US origin
Outcomes	87 of 198 abstracts published 19/38 abstracts presented at the 2000 meeting, 17/41 at the 2001 meeting, 30/58 at the 2002 meeting, and 21/60 at the 2003 meeting published Proportion of abstracts published by time Mean time to publication = 57 months (95% CI = 52 to 61 months) Survival analysis of proportion published = 40% at 48 months; author did not include 19 abstracts published before the meeting Factors related to proportion of abstracts published included 23/45 abstracts with 'positive' (defined as significant results) versus 22/44 abstracts without 'positive' results published 59/137 abstracts with 'positive' (defined as experimental better than control or clinically relevant) results versus 3/8 abstracts without 'positive' results published 71/143 abstracts presented orally versus 16/55 abstracts presented as poster published 25/54 abstracts with RCT design versus 62/144 abstracts with CCT design published 65/131 abstracts originating in the US versus 22/67 abstracts not originating in the US published
Notes	Surgery ‐ laser medicine and surgery Funding by the Medical Sciences/University of Tehran
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing clinical trials.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, positive results, type of presentation, study design, and US origin with publication using stratified analysis and Chi2 tests.

Al‐Qaoud 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to May 2012 Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by Title Hypothesis Cohort Study design
Data	Included 439 abstracts presented at the 2000 to 2010 Quebec Urological Association meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published Institution from Quebec versus not Basic science research versus clinical research
Outcomes	248 of 439 abstracts published Proportion of abstracts published by time Mean time to publication = 19.5 months Cumulative proportion of abstracts published at 72 months showed proportion published = 58.1% (255/439 abstracts) Factors related to proportion of abstracts published included 212/388 abstracts from a Quebecois institution versus 36/51 abstracts from a non‐Quebecois institution published = OR 2.13 (95% CI = 1.2 to 3.76) Basic science versus clinical research = OR 0.98 (95% CI = 0.72 to 1.33)
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The 2009 and 2010 meetings had less than 48 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of Quebecois origin, and type science for association with publication using multivariable log‐rank tests or Cox regression models.

Allart 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from January 2007 until December 2011 Search completed by investigators Searched by all authors, and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 231 abstracts presented at the 2008 Société Française de Médecine Physique et de Réadaptation meeting Included all abstracts accepted as posters and oral communications excluding communications by guest speakers
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation RCT design versus 'original research' versus case report versus literature review versus presentation of practice Multi‐centered versus single center French versus North African versus rest of the world origin Academic versus non‐academic affiliation Field of work (neurology, movement disorders, other) Therapeutic versus non‐therapeutic study
Outcomes	49 of 231 abstracts published Proportion of abstracts published by time published Mean time to publication = 18.4 months (SD = 14) Median time to publication = 21 months Cumulative proportion of abstracts published at 48 months showed proportion published = 21.2% (49/231 abstracts) Factors related to proportion of abstracts published included 14/46 abstracts with 'positive' (defined as significant results) versus 3/6 abstracts without 'positive' results published 35/144 abstracts presented orally versus 14/87 abstracts presented as poster published 3/4 abstracts with RCT design versus 38/131 abstracts describing 'original research' versus 7/39 with case report design versus 1/6 abstracts describing literature review versus 3/55 abstracts describing presentations of practice published 4/8 abstracts with multiple centers versus 33/131 with a single center published 43/190 abstracts originating from France versus 1/26 abstracts originating from North Africa versus 5/15 abstracts from rest of the world published 45/180 abstracts originating in an academic center versus 4/51 abstracts from authors not originating in an academic center published
Notes	Physical and rehabilitation medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	No	All meetings only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type presentation, study design, multi‐center status, country of origin, academic origin, and sub‐specialty with publication using multivariable logistic regression analysis.

Allart 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from January 2007 until June 2012 Search completed by investigators Searched by all authors, and keywords Matching criteria not reported
Data	Included 779 abstracts presented at the 2008 European Congress of Physical and Rehabilitation Medicine meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation RCT design versus CCT design versus observational study design versus case report versus literature review versus presentation of practice Multi‐centered versus single center North America versus French versus Middle East versus rest of the world origin Academic versus non‐academic affiliation
Outcomes	169 of 779 abstracts published Proportion of abstracts published by time published Mean time to publication = 12 months (SD = 15.7) Median time to publication = 11 months Cumulative proportion of abstracts published at 42 months showed proportion published = 21.7% (169/779 abstracts) Factors related to proportion of abstracts published included 67/231 abstracts with 'positive' (defined as significant results) versus 7/21 abstracts without 'positive' results published 109/288 abstracts presented orally versus 60/491 abstracts presented as poster published 49/108 abstracts with RCT design versus 59/213 abstracts with CCT design versus 49/276 abstracts with observational study design versus 4/116 abstracts with case report design versus 4/24 abstracts describing literature reviews versus 4/42 abstracts describing presentations of practice published 15/30 abstracts with multiple centers versus 139/531 with a single center published 10/48 abstracts originating from North America versus 141/653 abstracts originating from Europe versus 8/39 abstracts originating in the Middle East versus 10/39 abstracts from rest of the world published 157/616 abstracts originating in an academic center versus 12/163 abstracts not originating in an academic center published
Notes	Physical and rehabilitation medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type presentation, study design, multi‐center status, country of origin, academic origin, and sub‐specialty with publication using multivariable logistic regression analysis.

Alonso‐Arroyo 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database Web of Science, SCOPUS, Science Citation Index‐Expanded, Indice Médico Español, Indice Bibliografico Español en Ciencias de la Salud in December 2012 Person completing the search not reported Searched by all authors, keywords, and institutions Matching criteria not reported
Data	Included 300 abstracts presented at the 2002, 2005, and 2008 La Sociedad Española de Cardiología meetings Included a random sample of all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Subspecialty Cumulative proportion of abstracts published
Outcomes	115 of 300 abstracts published 45/100 abstracts presented at the 2002 meeting, 59/100 at the 2005 meeting, and 43/100 at the 2008 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 84 months showed proportion published = 49% (147/300 abstracts) No factors other than meeting year and subspecialty related to proportion of abstracts published reported
Notes	Cardiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random sample of all abstracts accepted for oral presentation.
Search for publications?	Yes	Searched 5 databases.
Follow‐up time?	Yes	All abstracts had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis.

Alpi 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, PubMed, ERIC, and LISTA (Library, Information Science & Technology Abstracts); dates of search not reported Search completed by trained librarian Searched by all authors and keywords Matching criteria not reported
Data	Included 62 abstracts presented at the 2000 to 2008 Medical Librarians Association meetings Included all award‐winning papers and posters published in the April 2009 issue of 'Hypothesis'
Comparisons	Proportion of abstracts published
Outcomes	23 of 62 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Medical education/library science Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

Amarilyo 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from November 2005 until November 2011 Search completed by investigators Searched by first and last author, and keywords Matched abstract to full‐length publication by One author Methodology Hypothesis
Data	Included 2149 abstracts presented at the 2006 American College of Rheumatology and Association of Rheumatology Health Professionals meeting Included all abstracts presenting scientific research studies except individual case reports
Comparisons	Proportion of abstracts published Mean and median time to publication, overall and by type of research Cumulative proportion of abstracts published Oral versus poster presentation Clinical versus basic science research Presentation in plenary session versus presentation in another format Common versus rare diseases
Outcomes	1269 of 2149 abstracts published Proportion of abstracts published by time Mean time to publication = 15.8 months (SD = 18.3; range = ‐9 to 51 months), overall Mean time to publication = 19.5 months for abstracts describing clinical abstracts Mean time to publication = 16.3 months for abstracts describing basic science research Median time to publication = 16 months (IQR = 1.75 to 29.5 months), overall Median time to publication = 17 months (IQR = 9 to 28 months) for abstracts describing clinical research Median time to publication = 13 months (IQR = 5 to 26 months) for abstracts describing basic science research) Cumulative proportion of abstracts published at 60 months showed proportion published = 59.1% (1269/2149 abstracts) Factors related to proportion of abstracts published included 259/378 abstracts presented orally versus 1012/1772 abstracts presented as poster published 796/1354 abstracts describing clinical research versus 473/796 abstracts describing basic science research published 14/14 abstracts presented in a plenary session versus 1255/2135 abstracts presented in another format published
Notes	Rheumatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Unclear	Included all abstracts that described 'scientific research' but not clear how this subgroup was defined.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of type presentation, type science, and type oral session with publication using stratified analysis and Chi2 or Fisher's Exact tests.

Amirhamzeh 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar; dates of search not reported Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by All authors Title Contents
Data	Included 762 abstracts presented at the 2002 to 2006 Pediatric Orthopaedic Society of North America meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication by meeting Oral versus poster Subspecialty
Outcomes	386 of 762 abstracts published Proportion of abstracts published by time Mean time to publication = 29.4 months for 2002 Pediatric Orthopaedic Society of North America meeting Mean time to publication = 26.8 months for 2003 Pediatric Orthopaedic Society of North America meeting Mean time to publication = 26.3 months for 2004 Pediatric Orthopaedic Society of North America meeting Mean time to publication = 22.7 months for 2005 Pediatric Orthopaedic Society of North America meeting Mean time to publication = 18.8 months for 2006 Pediatric Orthopaedic Society of North America meeting Factors related to proportion of abstracts published reported included Reported that oral presentations were 1.47 times (CI, 1.10 to 1.98) more likely to be published compared with posters Reported that there was no differences in proportion published by subspecialty
Notes	Orthopedics/orthopedic surgery ‐ pediatric Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of type of presentation and subspecialty with publication using stratified analysis and Chi2 or ANOVA tests.

Arap 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and LILACS; dates of search not reported Person completing the search not reported Searched by first and last author, and keywords Matching criteria not reported
Data	Included 195 abstracts presented at the 2001 to 2007 American Urological Association meetings Included all abstracts where at least 2/3 of author institutions were from Brazil
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published
Outcomes	100 of 195 abstracts published 7/14 abstracts presented at the 2001 meeting,16/31 at the 2002 meeting, 12/29 at the 2003 meeting, 23/41 at the 2004 meeting, 13/31 at the 2005 meeting, 9/17 at the 2006 meeting, and 15/32 at the 2007 meeting published Proportion of abstracts published by time Median time to publication = 18.8 months Cumulative proportion of abstracts published at 126 months showed proportion published = 54.1% (105/195 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all Brazilian abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year for association with publication using stratified analysis and Chi2 tests.

Arora 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Scopus, and Google/Google Scholar from November 2009 and January 2010 Search completed by investigators Searched by first and last author, and keywords Matched abstract to full‐length publication by One outcome
Data	Included 479 abstracts presented at the 2001 to 2005 International Society of Paediatric Oncology (only from India), American Society of Clinical Oncology (only conducted in India), and International Society of Paediatric Oncology (only from UK) meeting Included only abstracts of studies conducted in India; excluded abstracts for keynote addresses, symposia, and "publication only"
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting Oral versus poster presentation RCT design versus non‐RCT design Conducted in India versus conducted in UK
Outcomes	139 of 479 abstracts published 113/275 abstracts presented at the 2001 to 2005 International Society of Paediatric Oncology (only from UK) meetings, 2/13 at the 2001 to 2005 American Society of Clinical Oncology meetings, and 24/191 at the 2001 to 2005 International Society of Paediatric Oncology (only from India) meetings published Proportion of abstracts published by time Mean time to publication = 20 months (range ‐1 to 95 months) for the International Society of Paediatric Oncology meetings Mean time to publication = 24 months for the American Society of Clinical Oncology meetings Mean time to publication = 31 months for the International Society of Paediatric Oncology meetings Factors related to proportion of abstracts published (included only from International Society of Paediatric Oncology meetings) included 9/37 abstracts presented orally versus 15/154 abstracts presented as posters published 0/7 abstracts with RCT design versus 24/184 abstracts with non‐RCT design published 24/191 abstracts conducted in India versus 113/275 abstracts conducted in the UK published
Notes	Oncology ‐ pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	Yes	Examined association of meeting year, type presentation, institution, subspecialty, and study design with publication using a multivariable linear model analysis.

Arrive 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from 1997 through 1999 Person completing the search not reported Searched by lead author Matched abstract to full‐length publication by Two authors Contents
Data	Included 456 abstracts presented at the 1996 Journées Francaises de Radiologie meeting Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published
Outcomes	39 of 456 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 36 months showed proportion published = 8.5% (39/456 abstracts) No factors related to proportion of abstracts published reported
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted as oral presentations.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meeting only had 24 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.

Arrive 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2000 Search completed by investigator Searched by first, second, and last author and radiologic subspecialty Matched abstract to full‐length publication by Sample size Contents
Data	Included 1897 abstracts presented at the 1995 Radiological Society of North America meeting Included all abstracts of "original research" presented orally
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Clinical research versus animal research versus technical research North American versus European versus Asian versus rest of the world origin Subspecialty
Outcomes	635 of 1897 abstracts published Proportion of abstracts published Cumulative proportion of abstracts published at 60 months showed proportion published = 33.5% (635/1897 abstracts) Factors related to proportion of abstracts published included 503/1389 abstracts describing clinical research versus 36/113 abstracts describing animal research versus 96/395 abstracts describing technical studies published 407/1241 abstracts originating from North America versus 160/462 abstracts from Europe versus 49/133 abstracts from Asia versus 19/61 abstracts from rest of the world published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Unclear	Included all abstracts that described "original research" and presented orally, but subgroup not defined.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Yes	Examined association of type science and country of origin, and subspecialty with publication using multivariable logistic regression analysis.

Autorino 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2005 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by Title Methodology Hypotheses
Data	Included 1100 abstracts presented at the 2001 and 2002 World Congress of Endourology meetings Included all abstracts except videotape sessions
Comparisons	Proportion of abstracts published Mean time to publication Survival analysis of publication rate Clinical research versus preclinical research North America versus Asia versus South America versus Europe versus Africa 'Prospective' versus 'retrospective' Subspecialty
Outcomes	234 of 1100 abstracts published Proportion of abstracts published by time Mean time to publication = 14.6 months (maximum = 40 months) Survival analysis of proportion published at 42 months = 21.3% Factors related to proportion of abstracts published included 186/940 abstracts describing clinical research versus 45/160 abstracts describing preclinical research published The proportion of abstracts published that originated from North America was 29.2%, from Asia 18.5%, from South America 14.3%, from Europe 14.1%, and from Africa 3.1% 20.8% of prospective studies versus 23.4% of retrospective studies published
Notes	Urology ‐ endourology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The 2001 meeting had 48 months follow‐up, but the 2002 meeting only had 36 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of type science, country of origin, prospective status, and subspecialty with publication using multivariable logistic regression analysis.

Autorino 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed in March 2006 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by Title Methodology Hypotheses
Data	Included 1406 abstracts presented at the 2000 and 2001 European Association of Urology meetings Included all abstracts except videotape sessions
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published Oral versus poster presentation Clinical research versus preclinical research North American versus European versus Japanese Funding reported versus funding not reported Prospective versus retrospective Subspecialty
Outcomes	666 of 1406 abstracts published Proportion of abstracts published by time Mean time to publication = 8.66 months (SD = 4.2) Cumulative proportion of abstracts published at 48 months showed proportion published = 45.4% (639/1406 abstracts) Factors related to proportion of abstracts published included 273/548 abstracts presented orally versus 393/858 abstracts presented as poster published 458/1016 abstracts describing clinical research versus 208/390 abstracts describing preclinical research published 48/98 abstracts originating from North America versus 522/1060 abstracts from Europe versus 16/24 abstracts from Japan published 12/22 abstracts with funding reported versus 654/1384 abstracts with funding not reported published
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of type presentation, type science, country of origin, funding status, and subspecialty with publication using multivariable logistic regression analysis.

Autorino 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from October 2002 to July 2007 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by Title Methodology Hypotheses
Data	Included 1877 abstracts presented at the 2002 to 2004 Société Internationale d'Urologie meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published Clinical research versus preclinical research North American versus European versus Asian versus rest of the world (South America, Oceania, and Africa) 'Prospective' versus 'retrospective' design Subspecialty
Outcomes	415 of 1877 abstracts published Proportion of abstracts published by time Mean time to publication = 13 months (range = 1 to 45 months) Cumulative proportion of abstracts published at 36 months showed proportion published = 22.1% (415/1877 abstracts) Factors related to proportion of abstracts published included 331/1543 abstracts describing clinical research versus 86/334 abstracts describing preclinical research published 95/354 abstracts originating from North America versus 135/613 abstracts from Europe versus 155/753 abstracts from Asia versus 31/157 abstracts from the rest of the world (South America, Oceania, and Africa) published 258/1161 abstracts with 'prospective' study design versus 158/716 abstracts with 'retrospective' study design published
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2004 had at least 48 months follow‐up. The 2004 meeting only had 36 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of type science, country of origin, prospective status, and subspecialty with publication using multivariable logistic regression analysis.

Autorino 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 94 abstracts presented at the 2004 to 2006 World Congress of Endourology meetings Included all abstracts describing RCTs
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication
Outcomes	45 of 94 abstracts published 11/21 abstracts presented at the 2004 meeting, 20/36 at the 2005 meeting, and 14/37 at the 2006 meeting published Proportion of abstracts published by time Mean time to publication = 16.4 months (SD = 13.2 months) No factors other than meeting year related to proportion of abstracts published reported
Notes	Urology ‐ endourology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing RCTs.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis and Chi2, Fisher's Exact, Wilcoxon, or t tests.

Bagheri 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE for "a five year interval" Person completing the search not reported Searched by first, last, and senior author Matching criteria not reported
Data	Included 446 abstracts presented at the 1997 to 1999 American Association of Oral and Maxillofacial Surgeons meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Oral versus poster presentation North American versus European versus Asian versus South American origin Subspecialty
Outcomes	155 of 446 abstracts published Proportion of abstracts published by time Mean time to publication = 23.4 months (SD = 16.2; range = 1 to 72 months) Factors related to proportion of abstracts published included 130/347 abstracts presented orally versus 25/99 abstracts presented as poster published 131/372 abstracts originating from North America versus 14/36 abstracts from Europe versus 11/27 abstracts from Asia versus 1/11 abstracts from South America published
Notes	Otolaryngology head and neck surgery ‐ oral and maxillofacial surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The follow‐up for all meetings was Quote: "a five year interval".
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation, country of origin, and subspecialty with publication using stratified analysis and Chi2 tests.

Bakkum 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Index to the Chiropractic Literature until summer 2012 Search completed by investigators Searched by all authors and keywords Matched abstract to full‐length publication by All authors Sample size Title Contents Methodology Hypothesis
Data	Included 776 abstracts presented at the 2002 to 2008 Chiropractic Colleges Educational Conference/ Research Agenda Conference meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation Clinical research versus basic science research versus educational research versus other types of research
Outcomes	249 of 776 abstracts published 34/122 abstracts presented at the 2002 meeting, 16/96 at the 2003 meeting, 34/110 at the 2004 meeting, 41/92 at the 2005 meeting, 40/131 at the 2006 meeting, 38/115 at the 2007 meeting, and 46/110 at the 2008 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 169/400 abstracts presented orally versus 80/376 abstracts presented as poster published 72/233 abstracts describing clinical research versus 120/310 abstracts describing basic science research versus 15/109 abstracts describing educational research versus 24/124 abstracts describing other types of research published
Notes	Physical and rehabilitation medicine Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 6 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, and type science with publication using stratified analysis.

Bakkum 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and VisionCite in 2014 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Contents Methodology Hypothesis
Data	Included 518 abstracts presented at the 2006 American Academy of Optometry meeting Included all abstracts
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Oral versus poster presentation Clinical research versus basic science research Case report design versus other study design Subspecialty
Outcomes	108 of 518 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 84 months showed proportion published = 20.9% (108/518 abstracts) Factors related to proportion of abstracts published included 32/97 abstracts presented orally versus 76/421 abstracts presented as poster published 88/425 abstracts describing clinical research versus 20/93 abstracts describing basic science research published 9/164 abstracts with case report design versus 99/354 abstracts with other study design published
Notes	Ophthalmology/optometry ‐ optometry Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	The meeting had 96 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of type presentation, type science, and study design with publication using stratified analysis and Chi2 tests.

Balasubramanian 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Embase to July 2003 Search completed by investigator Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 241 abstracts presented at the 1997 Association of Surgeons of Great Britain and Ireland meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published Multivariate analysis examining academic affiliation, multi‐center status, clinical design, study design, specialty, presentation type, mention of statistical methods, sample size, results and author number
Outcomes	136 of 241 abstracts published Proportion of abstracts published by time Median time to publication = 18 months (IQR = 11 to 27 months) Cumulative proportion of abstracts published at 78 months showed proportion published = 56.4% (136/241 abstracts) Odds ratio in multivariate analysis comparing multi‐center versus single center studies was 2.1 (95%CI, 0.90 to 4.57); no other factors were significant
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of type of presentation, academic affiliation, multi‐center status, clinical design, study design, mention of statistical methods, sample size, results and author number and subspecialty with publication using multivariable logistic regression analysis.

Beker‐Acay 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed until September 2014 Search completed by investigators Searched by first and second author, and keywords Matching criteria not reported
Data	Included 3192 abstracts presented at the 2010 to 2012 Turkish National Radiology meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Oral versus poster presentation Case report versus scientific research versus educational exhibit Prospective versus retrospective design Subspecialty
Outcomes	300 of 3192 abstracts published 93/844 abstracts presented at the 2010 meeting, 106/1299 at the 2011 meeting, and 101/1049 at the 2012 meeting published Proportion of abstracts published by time Mean time to publication = 15 months (range = 0 to 42 months) Factors related to proportion of abstracts published included 81/481 abstracts presented orally versus 219/2711 abstracts presented as poster published 143/2107 abstracts describing case report design versus 114/654 abstracts describing 'scientific' research versus 43/431 abstracts describing educational exhibit published 83/533 abstracts with prospective design versus 74/552 abstracts with retrospective design published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings except the 2010 meeting had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation, study design, prospective status, and subspecialty with publication using stratified analysis and Chi2 tests

Bello 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and African Journal Online until September 30, 2012 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Methodology Results
Data	Included 75 abstracts presented at the 2007 to 2010 Nigerian Association of Urological Surgeons meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Observational design versus case report versus case series versus other type of study design Basic versus beyond basic versus no use of statistics Retrospective design versus prospective design Subspecialty
Outcomes	18 of 75 abstracts published 3/19 abstracts presented at the 2007 meeting, 8/16 at the 2008 meeting, 6/27 at the 2009 meeting, and 1/13 at the 2010 meeting published Proportion of abstracts published by time Median time to publication = 15 months (range = 2 to 40 months) Factors related to proportion of abstracts published included 13/48 abstracts describing observational design versus 4/18 abstracts describing case report versus 1/7 abstracts describing case series versus 0/2 abstracts describing other types of study designs published 8/41 abstracts with basic use of statistics versus 5/7 abstracts with beyond basic use of statistics versus 5/27 abstracts with no use of statistics published 9/31 abstracts with retrospective design versus 2/14 abstracts with prospective design published
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The 2009 and 2010 meetings had less than 48 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, study design, use of statistics, prospective status, and subspecialty with publication using multivariable logistic regression analysis.

Bergoli 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, BBO, and LILACS from 2009 to 2010 Search completed by investigators Searched by first and last author, and keywords Matched abstract to full‐length publication by All authors Sample size Title Results Aims Laboratory tests Conclusions
Data	Included 140 abstracts presented at the 2001 Brazilian Society of Oral Research meeting Included a random sample of abstracts
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	37 of 140 abstracts published Proportion of abstracts published by time Mean time to publication = 24.1 months No factors related to proportion of abstracts published reported
Notes	Oral health ‐ oral research Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random sample of abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The meeting had 9 years follow‐up.
Matching?	Yes	Matched by 7 different criteria.

Bergqvist 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by All authors Title
Data	Included 699 abstracts presented at the 1989 to 1994, 1995 to 2000, and 2001 to 2006 European Society for Vascular Surgery meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	553 of 699 abstracts published, 192/218 presented at the 1989 to 2004 meetings, 162/219 at the 1995 to 2000 meetings, and 199/262 at the 2001 to 2006 meetings. Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Surgery ‐ vascular surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Bernstein 1983.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLARS; date of search not reported Person completing the search not reported Searched by first author Matching criteria not reported Contacted abstract authors directly
Data	Included 177 abstracts submitted to the 1978 American Association for the Study of Liver Diseases meeting Included all abstracts
Comparisons	Proportion of abstracts published Accepted versus rejected for conference presentation Clinical research versus basic science research
Outcomes	106 of 177 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 56/82 abstracts accepted for presentation versus 50/95 rejected abstracts published 53/101 abstracts describing clinical research versus 53/76 abstracts describing basic science research published
Notes	Gastroenterology ‐ hepatology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of acceptance for presentation and type of science with publication using stratified analysis and unspecified statistical tests.

Berookhim 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE until October 2009 Person completing the search not reported Searched by all authors, and title Matched abstract to full‐length publication by One author Institution of any author Study design
Data	Included 208 abstracts presented at the 2006 and 2007 American Urological Association meetings Included all abstracts relating to sexual medicine
Comparisons	Proportion of abstracts published Clinical research versus basic science research US versus non‐US origin Academic versus non‐academic affiliation Subspecialty
Outcomes	107 of 208 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 71/137 abstracts describing clinical research versus 33/71 abstracts describing basic science research published 71/129 abstracts originating from the US versus 36/79 abstracts not originating from the US published 92/150 abstracts originating from an academic center versus 13/58 abstracts originating not originating from an academic center published
Notes	Human reproduction and development ‐ sexual and reproductive medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The 2006 meeting had 36 months, and the 2007 meeting 24 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of type science, US origin, academic affiliation, and subspecialty with publication using multivariable logistic regression analysis.

Bhandari 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and PubMed from January 1995 to February 2001 Search completed by investigator Searched by first, second, and last author, and keywords Matched abstract to full‐length publication by All authors Title Contents
Data	Included 465 abstracts presented at the 1996 American Academy of Orthopedic Surgeons meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Clinical research versus basic science research RCT design versus non‐RCT design North American versus European versus rest of the world origin Funding stated versus not stated Subspecialty
Outcomes	159 of 465 abstracts published Proportion of abstracts published by time Mean time to publication = 17.6 months (SD = 12) Median time to publication = 14 months (range = 1 to 56 months) Cumulative proportion of abstracts published at 44 months showed proportion published = 31.8% (148/465 abstracts) Factors related to proportion of abstracts published included 122/357 abstracts describing clinical research versus 37/107 abstracts describing basic science research published 7/23 abstracts with RCT design versus 152/442 abstracts with non‐RCT design published 148/412 abstracts originating from North America versus 5/33 abstracts from Europe versus 6/20 abstracts from rest of the world published No study reported funding
Notes	Orthopedics/orthopedic surgery Reported receiving no funding from grants or outside sources
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 60 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of type science, study design, country of origin and subspecialty with publication using multivariable logistic regression analysis.

Bhasin 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and PubMed to April 2005 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 106 abstracts presented at the 2001 and 2002 Vascular Society of Great Britain and Ireland meetings Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published Mean and median time to publication
Outcomes	63 of 106 abstracts published Proportion of abstracts published by time Mean time to publication = 11 months Median time to publication = 12 months (IQR = 7 to 18 months) No factors related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ vascular research Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted as oral presentations.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The 2001 meeting had 48 months and the 2002 meeting 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Bird 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database Aquatic Sciences and Fisheries Abstracts (1982 to 1998), Biological Abstracts (1989 to 1998), Medline (1966 to 1998), PsycINFO (1982 to 1998), Zoological Record (1978 to September 1998), CINAHL (1982 to 1998), ERIC (1966 to 1999), PAIS (1972 to 1998), Sociological abstracts (1994 to 1998), UnCover (1988 to 1998) Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author   Keywords Contents Methodology Study time period Contacted abstract authors directly
Data	Included 425 abstracts presented at the 1989 and 1991 Society for Marine Mammalogy meetings Included abstracts randomly selected from 849 abstracts presented at the meeting
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting
Outcomes	249 of 425 abstracts published 123/210 abstracts presented at the 1989 meeting, and 126/215 at the 1991 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 103 months for 1989 meeting showed proportion published = 49% (99/203 abstracts; time to publication not available for 7 full publications) Cumulative proportion of abstracts published at 79 months for 1991 meeting showed proportion published = 48% (99/206 abstracts; time to publication not available for 9 full publications) No factors other than meeting year related to proportion of abstracts published reported
Notes	Other non‐clinical specialties ‐ marine mammology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	Yes	Searched 10 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Blackwell 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase, and Google/Google Scholar; dates of search not reported but reported Quote: "five year minimum" Person completing the search not reported Searched by first and last author, keywords, and title Matching criteria not reported
Data	Included 90 abstracts presented at the 2000 to 2002 Society of Maternal Fetal Medicine meetings Included only abstracts of controlled clinical trials
Comparisons	Proportion of abstracts published Median time to publication 'Positive' versus not 'positive' Oral versus poster presentation RCT design versus CCT design US versus non‐US origin Academic versus non‐academic affiliation Subspecialty
Outcomes	50 of 90 abstracts published Proportion of abstracts published by time Median time to publication = 19 months (range = 4 to 66 months) Factors related to proportion of abstracts published included 30/43 'positive' (defined by direction of results or clinically relevant) versus 13/47 not 'positive' abstract results published 15/21 abstracts presented orally published; proportion presented as posters that were published not reported 50/86 abstracts with RCT design versus 0/4 abstracts with CCT design published 34/63 abstracts originating in the US versus 16/27 abstracts not originating in the US published 43/75 abstracts originating in an academic center 7/15 abstracts not abstracts originating in an academic center published
Notes	Gynecology/obstetrics ‐ maternal‐fetal medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing controlled clinical trials.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type presentation, study design, US origin, academic affiliation, and subspecialty with publication using multivariable logistic regression analysis.

Bolac 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from July to August 2008 Search completed by investigator Searched by first and last author, and keywords Matched abstract to full‐length publication by Two authors Contents
Data	Included 540 abstracts presented at the 1999 to 2004 Journées Dermatologiques de Paris meetings Included all abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Clinical research versus basic science research Observational studies versus case reports versus 'essais cliniques' versus epidemiology studies Multi‐centered versus single center Conducted by free dermatologists, institutes or associations versus members of Société francaise de dermatologie Oncology versus infectious diseases Subspecialty
Outcomes	299 of 540 abstracts published Proportion of abstracts published by time Mean time to publication = 20.3 months (SD = 16.9) Median time to publication = 17 months (range = 0 to 90 months) Cumulative proportion of abstracts published at 48 months showed proportion published = 61.1% (330/540 abstracts) Factors related to proportion of abstracts published included 55.2% of abstracts describing clinical research versus 67.3% of abstracts describing basic science research published 149/279 abstracts with observational studies versus 45/80 abstracts of case reports versus 23/28 abstracts of 'essais cliniques' versus 5/16 epidemiology studies published 58.1% of abstracts with multiple centers versus 57.5% of abstracts with a single center published 4/9 abstracts conducted by free dermatologists, institutes or associations versus 16/24 abstracts conducted by members of Société Francaise de Dermatologie published 79/123 abstracts on oncology studies versus 32/46 abstracts on infectious diseases published
Notes	Dermatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of type science, study design, multi‐center status, presenter status, and subspecialty with publication using stratified analysis and Chi2 or Fisher's Exact tests.

Boldt 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1993 to 1999 Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by All authors Contents
Data	Included 566 abstracts presented at the 1994 Deutscher Anesthäsie Kongress and European Society of Anaesthesiologists meetings Included all abstracts of free papers
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting Oral versus poster presentation
Outcomes	233 of 566 abstracts published 111/244 abstracts presented at the 1994 European Society of Anaesthesiologists meeting, and 122/322 at the 1994 Deutscher Anesthäsie Kongress meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 84 months showed proportion published = 37.9% (122/322 abstracts) for Deutscher Anesthäsie Kongress Cumulative proportion of abstracts published at 84 months showed proportion published = 45.5% (111/244 abstracts) for European Society of Anaesthesiologists Factors related to proportion of abstracts published included 137/349 abstracts presented orally versus 96/217 abstracts presented as posters published
Notes	Anesthesiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of free papers.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and type presentation with publication using stratified analysis.

Bonitz 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to April 2011 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 284 abstracts presented at the 2006 and 2007 American Society for Reproductive Medicine meetings Included all abstracts related to male reproduction
Comparisons	Proportion of abstracts published Median time to publication Subspecialty
Outcomes	107 of 284 abstracts published Proportion of abstracts published by time Median time to publication = 16 months (range = 10 to 28 months) No factors other than subspecialty related to proportion of abstracts published reported.
Notes	Human reproduction and development ‐ male reproduction Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of subspecialty with publication using stratified analysis, and time to publication using unspecified multivariable models.

Bowrey 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1981 through 1997 Person completing the search not reported Searched by all authors Matching criteria not reported
Data	Included 496 abstracts presented at the 1983 to 1995 Welsh Surgical Society meetings Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication Subspecialty
Outcomes	233 of 496 abstracts published Proportion of abstracts published by time Median time to publication = 17 months (IQR = 10 to 27.5 months; range = 24 to 168 months) No factors related to proportion of abstracts published reported
Notes	Surgery Funding not reported Author reports on five additional articles 'accepted' for publication, but not included
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 1994 had at least 48 months follow‐up. The 1994 and 1995 meetings had 36 and 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of sub‐specialty with publication using stratified analysis.

Brace 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Agricola, and CAB; dates of search not reported Search completed by investigator Searched by first, second, and last author, and ("organism + "vacci*"") Matched abstract to full‐length publication by Sample size Contents Description of vaccine Age, weight and sex of animals Dose and route of vaccine
Data	Included 154 abstracts presented at the 1988 to 2003 American Association of Swine Veterinarians Annual Conference and American Association of the Bovine Practitioners Annual Conference meetings Included all abstracts related to vaccines
Comparisons	Proportion of abstracts published, overall and by meeting 'Positive' versus not 'positive'
Outcomes	11 of 154 abstracts published 6/65 abstracts presented at the 1988 to 2003 American Association of the Bovine Practitioners Annual Conference meetings, and 5/89 at the 1988 to 2003 American Association of Swine Veterinarians Annual Conference meetings published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 8/91 'positive' (defined as experimental better than control or clinically relevant) versus 3/63 not 'positive' abstract results published
Notes	Other non‐clinical specialties ‐ veterinary vaccine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of meeting and positive results and meeting for association with publication using stratified analysis.

Brazzelli 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE to November 2006 Search completed by investigator Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 160 abstracts presented at the 1995 to 2004 International Stroke Conference and European Stroke Conference meetings Included all abstracts of diagnostic test accuracy studies
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Survival analysis of publication rate 'Positive' versus not 'positive' Sample size equal to or above the median versus sample size below the median Multi‐centered versus single center Origin from English‐speaking country versus Europe (excluding UK) versus Asia and South America "Accurate" versus "possibly useful" versus "non‐informative" estimates of sensitivity and specificity as defined by authors Blinding reported versus blinding not reported Interobserver agreement assessed versus not reported if assessed 'Prospective' versus 'retrospective' design versus design not reported
Outcomes	121 of 160 abstracts published Proportion of abstracts published by time Mean time to publication = 20 months Median time to publication = 16 months Cumulative proportion of abstracts published at 36 months showed proportion published = 63.1% (101/160 abstracts) for International Stroke Conference meetings Survival analysis of proportion published at 120 months = 84% for European Stroke Conference meetings Factors related to proportion of abstracts published included 107/141 'positive' (defined as experimental better than control or clinically relevant) versus 13/19 not 'positive' abstract results published 58/81 abstracts with sample size equal to or above the median versus 62/78 abstracts with sample size below the median published 10/13 abstracts with multiple centers versus 111/147 abstracts with a single center published 54/71 abstracts originating from English speaking country versus 62/79 abstracts from Europe (excluding UK) versus 5/10 abstracts from Asia and South America published 107/141 abstracts with estimates of sensitivity and specificity defined as "accurate" versus 7/10 abstracts with estimates defined as "possibly useful" versus 6/9 abstracts with estimates defined as "non‐informative" published 34/39 abstracts with studies reporting blinding versus 87/121 abstracts not reporting blinding published 15/17 abstracts with interobserver agreement assessed versus 106/143 abstracts not reporting if assessed published 28/42 abstracts with 'prospective' study design versus 10/14 abstracts with 'retrospective' study design versus 83/104 abstracts with design not reported published
Notes	Neurology ‐ stroke Funding by Scottish Executive Health Department Chief Scientist Office and English Department of Health
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of diagnostic test accuracy studies.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings before 2003 had at least 48 months follow‐up. The meetings in 2003 and 2004 had 36 months and 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of study design, number of authors, multi‐center status, sample size, presence of blinding, interobserver agreement, type of diagnostic test, clinical utility of results, Youden's index and country of origin with publication using stratified analysis, and time to publication using Cox regression models

Brost 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Search completed by investigator Searched by all authors and keywords Matching criteria not reported
Data	Included 288 abstracts presented at the 1998 to 2001 Society of Maternal Fetal Medicine meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published Multi‐centered trial versus clinical research versus basic science research versus animal research versus unclassified design
Outcomes	208 of 288 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 22/28 abstracts describing multi‐center trials versus 113/158 abstracts describing clinical research versus 43/52 abstracts describing basic science research versus 28/40 abstracts describing animal research versus 5/10 abstracts with unclassified design published
Notes	Gynecology/obstetrics ‐ maternal‐fetal medicine Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type of presentation and study design with publication using stratified analysis.

Buchan 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported, but "61 months after meeting" Search completed by investigator Searched by all authors Matched abstract to full‐length publication by Methodology One conclusion
Data	Included 179 abstracts presented at the 2004 Royal College of Ophthalmologists meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation Clinical research versus basic science versus service/delivery‐related studies UK origin versus international collaborations 'Prospective' versus 'retrospective' study design Number of authors 1‐2 versus 3‐5 versus 6 or more Academic author versus non‐academic author
Outcomes	64 of 179 abstracts published Proportion of abstracts published by time Mean time to publication = 20.1 months (95% CI = 16.9 to 23.4 months) Cumulative proportion of abstracts published at 48 months showed proportion published = 32.4% (58/179 abstracts) Factors related to proportion of abstracts published included 44.4% of abstracts with 'positive' (defined as statistically significant) versus 43.6% of abstracts with not 'positive' results published 41% of abstracts presented orally versus 35% of abstracts presented as posters published 36.5% of abstracts describing clinical research versus 41.7% of abstracts with basic science research versus 26.3% of abstracts describing service/delivery‐related studies published 35% of abstracts from the UK versus 45% of abstracts with international collaboration published 33% of abstracts with 'prospective' study design versus 33.9% of abstracts with 'retrospective' study design published 26% of abstracts with 1‐2 authors versus 36% of abstracts with 3 to 5 authors versus 50% of abstracts with 6 or more authors published 39.5% of abstracts with academic author versus 28.3% of abstracts with non‐academic author published
Notes	Ophthalmology/optometry ‐ ophthalmology Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting was followed up for Quote: "61 months".
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, type presentation, type science, UK origin, prospective status, number of authors, and academic affiliation with publication using unspecified multivariable analysis.

Burden 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 112 abstracts presented at the 2003 to 2012 South Yorkshire Regional Gastroenterology meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication by presence in list of top 3 abstracts per meeting Ranked in meeting under top 3 abstracts per meeting versus not in top 3 rank per meeting
Outcomes	37 of 112 abstracts published Proportion of abstracts published by time Mean time to publication = 12.8 months, for abstracts ranked in top 3 per meeting Mean time to publication = 19.3 months, for abstracts not ranked in top 3 per meeting Factors related to proportion of abstracts published included 24/32 abstracts ranked under top 3 per meeting versus 13/80 abstracts not ranked in top 3 per meeting published
Notes	Gastroenterology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database and contacted abstract authors directly.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of abstract ranking with publication using stratified analysis and unspecified statistical test.

Bydder 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to August 2002 Person completing the search not reported Searched by presenter and keywords Matching criteria not reported
Data	Included 481 abstracts presented at the 1996 to 1999 Royal Australian and New Zealand College of Radiologists meetings Included all abstracts of "free" papers
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication Cumulative proportion of abstracts published Oral versus poster presentation Diagnostic versus oncology studies
Outcomes	168 of 481 abstracts published 42/117 abstracts presented at the 1996 meeting, 28/97 at the 1997 meetings, 59/163 at the 1998 meeting, and 39/104 at the 1999 meeting published Proportion of abstracts published by time Mean time to publication = 16.5 months Median time to publication = 17 months Cumulative proportion of abstracts published at 36 months showed proportion published = 32.4% (156/481 abstracts) Factors related to proportion of abstracts published included 108/279 abstracts presented orally versus 60/202 abstracts presented as poster published  74/254 abstracts describing diagnostic studies versus 94/227 abstracts of oncology studies published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of "free" papers.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 1999 had at least 48 months follow‐up. The meeting in 1999 had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, and type science with publication using stratified analysis and Chi2 tests.

Byerly 2000.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Current Contents, International Pharmacy Abstracts from 1966 through 1998 Person completing the search not reported Searched by first author Matched abstract to full‐length publication by All authors Title Methodology Sample size Results Research question
Data	Included 716 abstracts presented at the 1994 American Society of Health System Pharmacists and American College of Clinical Pharmacy meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication by meeting
Outcomes	126 of 716 abstracts published 55/501 abstracts presented at the 1994 American Society of Health System Pharmacists meeting, and 71/215 at the 1994 American College of Clinical Pharmacy meeting published Proportion of abstracts published by time Median time to publication from the 1994 American Society of Health System Pharmacists meeting = 12.8 months Median time to publication from the 1994 American College of Clinical Pharmacy meeting = 14.9 months No factors other than meeting year related to proportion of abstracts published reported
Notes	Pharmacology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 6 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Callaham 1998.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase, the Cochrane Library from 1995 to 1996 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by All author Title Contents Contacted abstract authors directly
Data	Included 492 abstracts submitted to the 1991 Society for Academic Emergency Medicine meeting Included all abstracts, but excluded 11 on teaching methods from analysis
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Sample size equal to or above the median versus sample size below the median Accepted versus rejected for conference presentation 'High' versus 'low' quality as defined by author
Outcomes	214 of 492 abstracts published Proportion of abstracts published by time Mean time to publication = 18 months Cumulative proportion of abstracts published at 48 months showed proportion published = 41.3% (203/492 abstracts) Factors related to proportion of abstracts published included 77/153 'positive' (defined as experimental better than control or clinically relevant) versus 36/74 not 'positive' abstract results published 99/212 abstracts with sample size equal to or above the median versus 114/281 abstracts with sample size below the median published 111/179 abstracts accepted for presentation versus 103/313 rejected abstracts published 106/199 abstracts rated by author as 'high' quality versus 107/294 abstracts rated as 'low' quality published
Notes	Emergency medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, sample size, acceptance for presentation, and author rating with publication using multivariable logistic regression analysis.

Camacho 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to December 2004 Search completed by investigator Searched by first, third, and senior author, and keywords Matched abstract to full‐length publication by Title  Contacted abstract authors directly
Data	Included 275 abstracts submitted at the 1997 American Society of Clinical Oncology meeting Included all abstracts describing phase I studies
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Oral versus poster presentation Accepted versus rejected for conference presentation Use of a 'novel' versus 'non‐novel' agent Sponsored by a private agency versus not sponsored by a private agency
Outcomes	185 of 275 abstracts published Proportion of abstracts published by time Mean time to publication = 29 months (95% CI = 33.6 to 45.6 months) Median time to publication = 40.8 months Cumulative proportion of abstracts published at 90 months showed proportion published = 66.9% (184/275 abstracts) Factors related to proportion of abstracts published included 18/25 abstracts presented orally versus 88/123 abstracts presented as poster published 106/148 abstracts accepted for presentation versus 79/127 rejected abstracts published 98/146 abstracts describing use of a "novel" agent versus 86/129 abstracts describing a "non‐novel" agent published 72% of abstracts sponsored by a private agency versus 64% of abstracts not sponsored by a private agency published
Notes	Oncology Funding from Clinical Cancer Research Award from the Cancer and Leukemia Group B
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing phase I studies.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 7 years follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of type presentation, acceptance for presentation, use of a 'novel' agent, and sponsorship by a private agency with publication using stratified analysis and Fisher's Exact test.

Canosa 2011.

Methods	Identification of subsequent full‐length publications Contacted abstract authors directly
Data	Included 140 abstracts, which was a random sample of 840 abstracts submitted to the 2003 Congreso Argentino de Pediatría meeting
Comparisons	Proportion of abstracts published Mean and median time to publication Oral versus poster presentation Pediatrics versus other topics Male versus female authors
Outcomes	16 of 140 abstracts published Proportion of abstracts published by time Mean time to publication = 27 months (SD = 15) Median time to publication = 26 months (range = 7 to 37 months) Factors related to proportion of abstracts published included 9/38 abstracts presented orally versus 7/77 abstracts presented as posters published 15/129 abstracts related to pediatrics versus 1/8 abstracts related to other topics published 9/87 abstracts with female authors versus 7/53 abstracts with male authors published
Notes	Pediatrics Funding from Sociedad Argentina de Pediatría
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random sample of all submitted abstracts.
Search for publications?	Yes	Contacted abstract authors with > 80% response rate.
Follow‐up time?	Unclear	Date of author contact not reported.
Matching?	Yes	Matched by contact with abstract author.
Adjustment for confounding?	No	Examined association of type presentation, topic, and author gender with publication using stratified analysis and Chi2 or t tests.

Carroll 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to March 2003 Search completed by investigator Searched by all authors Matched abstract to full‐length publication by One author Results for one outcome
Data	Included 614 abstracts submitted to the 1998 and 1999 Pediatric Academic Society meetings Included all abstracts submitted to a meeting that were in sessions related to pediatrics, either including all abstracts in a session (if 80 or less) or a random sample of abstracts
Comparisons	Proportion of abstracts published Mean time to publication by presentation status and type Oral versus poster presentation Accepted versus rejected for conference presentation
Outcomes	243 of 614 abstracts published Proportion of abstracts published by time Mean time to publication for rejected abstracts = 23.1 months (SD = 12.1) Mean time to publication for abstracts presented orally = 21.4 months (SD = 11.8) Mean time to publication for abstracts presented as posters = 26.5 months (SD = 12.3) Mean time to publication for abstracts presented in poster symposia = 20.3 months (SD = 11.9) Factors related to proportion of abstracts published included 84/160 abstracts presented orally versus 124/296 abstracts presented as poster or poster symposium published 208/456 abstracts accepted for presentation versus 35/158 rejected abstracts published
Notes	Pediatrics Funding by the Robert Wood Johnson Clinical Scholars and Generalist Physician Faculty Scholars program
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all or a random sample of abstracts in sessions describing a specific topic.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Yes	Examined association of type presentation and acceptance for presentation with publication using multivariable logistic regression analysis.

Cartwright 2007a.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to February 2006 Search completed by investigator; "two observers" Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 130 abstracts presented at the 2003 International Continence Society meeting Included only abstracts from podium and oral poster sessions
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published
Outcomes	77 of 130 abstracts published Proportion of abstracts published by time Median time to publication = 11 months Cumulative proportion of abstracts published at 32 months showed proportion published = 59.2% (77/130 abstracts) Factors related to proportion of abstracts published included Positive results of interventional trials were not associated with publication US origin was significantly associated with publication (P = 0.043, Fisher's Exact Test) Multi‐center status was not associated with publication
Notes	Gynecology/obstetrics ‐ urogynecology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included only abstracts from podium and oral poster sessions.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	No	The meeting had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results, US origin and multi‐center status with publication using stratified analysis and Fisher's Exact or Mann‐Whitney U tests.

Cartwright 2007b.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to January 2007 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 116 abstracts presented at the 2003 International Urogynecological Meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication Oral versus poster presentation French versus other countries Subspecialty
Outcomes	64 of 116 abstracts published Proportion of abstracts published by time Median time to publication = 27 months Factors related to proportion of abstracts published included 31/39 abstracts presented orally versus 33/77 abstracts presented as posters published 88.9% of abstracts with French authors were published
Notes	Gynecology/obstetrics ‐ urogynecology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation and French origin with publication using stratified analysis and Fisher's Exact or Mann‐Whitney U test.

Castagnetti 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and PubMed in October 2012 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 1194 abstracts presented at the 2003 to 2006 and 2007 to 2010 European Society for Pediatric Urology meetings Included all abstracts except for a single abstract that was excluded due to incomplete data
Comparisons	Proportion of abstracts published, overall and by meeting group Cumulative proportion of abstracts published Oral versus poster presentation Clinical research versus basic science research RCT design versus other type of study design Multi‐centered versus single center English speaking country versus non‐English speaking country North American versus European versus rest of the world origin Prospective design versus retrospective design 'Higher' quality as defined as multi‐center/prospective/randomized versus 'lower' quality Subspecialty
Outcomes	564 of 1194 abstracts published 305/611 abstracts presented at the 2003 to 2006 meetings, and 259/583 at the 2007 to 2010 meetings published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 47.2% (564/1194 abstracts) Factors related to proportion of abstracts published included 173/350 abstracts presented orally versus 370/760 abstracts presented as poster published 444/961 abstracts describing clinical research versus 120/233 abstracts describing basic science research published 20/21 abstracts describing a RCT versus 544/1133 abstracts describing a different study design published 33/45 abstracts with multiple centers versus 532/1105 abstracts with a single center published 169/332 abstracts from an English speaking country versus 395/862 abstracts from a non‐English speaking country published 119/215 abstracts originating from North America versus 330/767 abstracts originating from Europe versus 115/222 abstracts with the rest of the world origin published 47/92 abstracts describing a prospective design versus 16/22 abstracts describing a retrospective design published 83/135 abstracts of 'higher' quality versus 366/826 abstracts of 'lower' quality published
Notes	Urology ‐ pediatrics Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with one reasonable exception.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The 2009 and 2010 meetings had less than 48 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type of presentation, type of science, study design, multi‐center status, origin in English language country, country of origin, prospective status, abstract quality, and subspecialty with publication using stratified analysis and Chi2 or Mann‐Whitney U tests

Castaldi 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database Cochrane Central Register of Controlled Trials, MEDLINE, PubMed, and Google/Google Scholar to July 2012 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Main outcome Searched web sites of major Italian journals in public health and hygiene
Data	Included 621 abstracts presented at the 2005 to 2007 Italian Society of Hygiene and Public Health meetings Included all 'good quality abstracts' defined as those with a score of 19 or higher or between 16 and 19 with unscored item for "Inherence"
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting 'Positive' versus not 'positive' Sample size ≤100 versus >100 Oral versus poster presentation Observational design versus experimental design versus 'revision' Academic versus non‐academic affiliation Very high quality versus high quality versus medium quality Male versus female first author
Outcomes	146 of 621 abstracts published 39/174 abstracts presented at the 2005 meeting, 44/187 at the 2006 meeting, and 63/260 at the 2007 meeting published Proportion of abstracts published by time Mean time to publication = 25.1 months (95% CI = 22.1 to 28.1 months), overall Mean time to publication = 27.4 months (95% CI = 20.8 to 34.0 months) for the 2005 meeting Mean time to publication = 30.5 months (95% CI = 24.6 to 36.4 months) for the 2006 meeting Mean time to publication = 20.2 months (95% CI = 16.6 to 23.8 months) for the 2007 meeting Factors related to proportion of abstracts published included 121/450 'positive' (defined as significant results) versus 4/41 not 'positive' abstract results published 47/220 abstracts with a sample size ≤100 versus 99/400 abstracts with a sample size >100 published 54/149 abstracts presented orally versus 92/472 presented as poster published 114/513 abstracts with observational design versus 21/63 abstracts with experimental design versus 11/44 presenting 'revisions' published 120/422 abstracts originating in an academic center versus 26/199 abstracts not abstracts originating in an academic center published 38/109 'very high quality' abstracts versus 78/326 'high quality' abstracts versus 30/186 'medium quality' abstracts published 62/299 abstracts with male first author versus 84/318 with female first author published
Notes	Other non‐clinical specialties ‐public health Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 3 databases and web sites of journals.
Follow‐up time?	Yes	The meetings had at least 60 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, positive results, sample size, type presentation, study design, academic affiliation, abstract quality, and author gender with publication using multivariable logistic regression analysis.

Castillo 2000.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1990 to 1998 Search completed by investigator Search criteria not reported Matched abstract to full‐length publication by All authors Contents
Data	Included 491 abstracts presented at the 1992 Congress of the Spanish Society of Anesthesiology and Resuscitation meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published
Outcomes	84 of 491 abstracts published Proportion of abstracts published by time Mean time to publication = 22 months (range = ‐24 to 72 months) Cumulative proportion of abstracts published at 78 months showed proportion published = 17.1% (84/491 abstracts) No factors related to proportion of abstracts published reported
Notes	Anesthesiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 6 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.

Castillo 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, DataStar from 1993 to December 2000 Search completed by investigator Searched by lead, second, third, and last author, keywords, and title Matching criteria not reported
Data	Included 472 abstracts presented at the 1995 European Society of Anesthesiologists meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Sample size equal to or above the median versus sample size below the median Oral versus poster presentation RCT design versus non‐RCT design North American versus European versus rest of the world origin Subspecialty
Outcomes	199 of 472 abstracts published Proportion of abstracts published by time Mean time to publication = 16.8 months (SD = 15.6; range = 24 to 60 months) Cumulative proportion of abstracts published at 72 months showed proportion published = 42.2% (199/472 abstracts) Factors related to proportion of abstracts published included 160/361 'positive' (defined as significant results) versus 23/56 not 'positive' abstract results published 73/197 abstracts with sample size equal to or above the median versus 106/230 abstracts with sample size below the median published 83/210 abstracts presented orally versus 116/262 abstracts presented as poster published 69/146 abstracts with RCT design versus 128/326 abstracts with non‐RCT design published 8/14 abstracts originating in North America versus 184/408 abstracts from Europe versus 7/49 abstracts from rest of the world published
Notes	Anesthesiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results, sample size, type presentation, study design, country or origin and subspecialty with publication by stratified analysis and Wilcoxon signed rank sum or U‐tests.

Cauchy 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed up to June 30, 2013 Person completing the search not reported Searched by first and last author, and keywords Matched abstract to full‐length publication by Sample size Title Results Type of study Conclusions
Data	Included 897 abstracts submitted to the 2005 to 2012 French Congress of Digestive and Hepato‐biliary Surgery meetings Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication, overall and by presentation status 'Positive' versus not 'positive' Sample size <100 versus >100 Accepted versus rejected abstracts Oral versus poster presentation Clinical research versus 'experimental' research RCT design versus observational design versus case report design versus literature review versus meta‐analysis Multi‐centered versus single center French versus rest of the world origin Plenary versus other oral presentation Subspecialty
Outcomes	334 of 897 abstracts published Proportion of abstracts published by time Median time to publication = 16.1 months (range = 0 to 67.4 months), overall Median time to publication = 16.1 months for accepted abstracts Median time to publication = 15.1 months for rejected abstracts Factors related to proportion of abstracts published included 245/580 'positive' (undefined) versus 89/317 not 'positive' abstract results published 227/624 abstracts with <100 participants versus 107/273 abstracts with <100 participants published 219/453 accepted abstracts versus 115/444 rejected abstracts published 109/176 abstracts presented orally versus 110/277 abstracts presented as poster published 309/857 abstracts describing clinical research versus 25/40 abstracts describing 'experimental' research published 6/8 abstracts with RCT design versus 316/825 abstracts with observational design versus 7/57 abstracts with case report design versus 2/4 abstracts describing literature reviews versus 3/3 abstracts describing meta‐analyses published 35/69 abstracts with multiple centers versus 299/832 abstracts with a single center published 301/716 abstracts from France versus 33/181 abstracts with rest of the world origin published 22/29 abstracts presented in the plenary versus 109/176 other orally presented abstracts published
Notes	Gastroenterology ‐ digestive and hepato‐biliary surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meeting only had 24 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, sample size, acceptance for presentation, type of presentation, type of science, study design, multi‐center status, French origin, type of oral presentation, and subspecialty with publication using multivariable logistic regression analysis.

Chalmers 1990a.

Methods	Identification of subsequent full‐length publications Searched electronic database Oxford Database of Perinatal Trials, date of search not reported Search completed by investigator Searched by all authors, contents, methodology Matching criteria not reported
Data	176 RCT 'summary reports' found in Oxford Database of Perinatal Trials to 1988 More than 2/3 of summary reports were abstracts, remainder were letters and short reports
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Sample size equal to or above the median versus sample size below the median 'High' versus 'low' quality as defined by author Abstract format versus short report or letter Cumulative proportion of abstracts published Subspecialty
Outcomes	64 of 176 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 48 months showed proportion published = 35.8% (63/176 abstracts) Factors related to proportion of abstracts published included 32/98 'positive' (defined as experimental better than control) versus 32/78 not 'positive' abstract results published 38/85 reports with sample size equal to or above the median versus 23/85 reports with sample size below the median published 3/10 reports rated by author as 'high' quality versus 43/114 reports rated as 'medium' quality versus 18/52 reports rated as 'low quality' published 49/125 in abstract format versus 15/51 reports in short report or letter format published
Notes	Pediatrics ‐ perinatology Funded by the National Center for Health Services Research and Department of Health, London, England
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all summary reports and other documents describing controlled trials.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, sample size, abstract quality, report format, and subspecialty with publication using multivariable logistic regression analysis.

Chan 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and OVID from January 1997 to September 2001 Person completing the search not reported Searched by all authors and institutions where study conducted Matching criteria not reported Handsearched three local journals
Data	Included 105 abstracts presented at the 1997 and 1998 Malaysian Paediatric Association Annual Congress, Perinatal Society of Malaysia Annual Congress, and Academy of Medicine of Malaysia Annual Scientific Congress of Medicine meetings Included abstracts of studies conducted in Malaysia and in children 0 to 16 years of age
Comparisons	Proportion of abstracts published Academic versus Ministry of Health affiliation Subspecialty
Outcomes	37 of 105 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published reported included 30/63 abstracts originating in an academic center versus 6/39 abstracts originating in a Ministry of Health institution published
Notes	Pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of trials that were conducted in a specific region and population.
Search for publications?	Yes	Searched 2 databases and performed handsearches.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of academic affiliation and subspecialty with publication using stratified analysis and Chi2 or Fisher's Exact tests.

Chan 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database "electronic literature search" to December 2007 Search completed by trained librarian Search criteria not reported Matching criteria not reported Contacted abstract authors directly
Data	Included 257 abstracts presented at the 2003 and 2004 International HTA conference meetings Included all abstracts presenting assessments of specific technologies
Comparisons	Proportion of abstracts published
Outcomes	122 of 257 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Medical decision making/health policy‐ health technology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Unclear	Included all abstracts that described 'assessments of specific technologies', but unclear how this subgroup was defined.
Search for publications?	Yes	Searched electronically and contacted abstract authors.
Follow‐up time?	Yes	The 2003 meeting had 48 months follow‐up, the 2004 only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Chan 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed, and Scopus to November 2009 Person completing the search not reported Searched by first, second, and last author, and keywords Matched abstract to full‐length publication by Sample size Methodology Results
Data	Included 389 abstracts presented at the 2006 International Conference on Emergency Medicine meeting Included all abstracts except those that were "cancelled"
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published English versus non‐English language North American versus European versus Australian versus Asian versus rest of the world (South American and African) origin Subspecialty
Outcomes	129 of 389 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 42 months showed proportion published = 33.4% (130/383 abstracts) Factors related to proportion of abstracts published included 116/314 abstracts with English language versus 13/69 abstracts with non‐English language published 74/193 abstracts originating in North America versus 25/74 abstracts from Europe versus 9/27 abstracts from Australia versus19/67 abstracts from Asia versus 2/15 abstracts from the Middle East versus 0/6 abstracts from rest of the world (South America and Africa) published
Notes	Emergency medicine Funding not reported Data extracted from short report
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all posters with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	No	The meeting had 36 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of origin from English language country, country of origin, and subspecialty with publication using stratified analyses and Chi2 tests.

Chan 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, HealthStar, CancerLit, and EconLit on July 1, 2010 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 164 abstracts presented at the 1997 to 2007 International Society for Pharmacoeconomics and Outcomes Research, American Society of Hematology, and American Society of Clinical Oncology meetings Included all abstracts on cost‐effectiveness analyses related to malignancies, with primary outcomes such as incremental cost per life‐year or quality‐adjusted life‐year gained
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published for one meeting US versus rest of the world origin UK versus rest of the world origin Canadian versus rest of the world origin EU (excluding UK) versus rest of the world origin
Outcomes	65 of 164 abstracts published 12/47 abstracts presented at the International Society for Pharmacoeconomics and Outcomes Research, 13/29 at the American Society of Hematology, and 41/88 at the American Society of Clinical Oncology meetings published Proportion of abstracts published by time Median time to publication = 37.2 months Cumulative proportion of abstracts published at 60 months = 40.2% (66/164 abstracts), overall Cumulative proportion of abstracts published at 42 months = 25.5% (12/47 abstracts) for International Society for Pharmacoeconomics and Outcomes Research meetings Cumulative proportion of abstracts published at 36 months = 41.4% (12/29 abstracts) for American Society of Hematology meetings Cumulative proportion of abstracts published at 60 months = 46.6% (41/88 abstracts) for American Society of Clinical Oncology meeting Factors related to proportion of abstracts published included US versus rest of the world origin: HR 1.57 UK versus rest of the world origin: HR 1.09 Canadian versus rest of the world origin: HR 0.46 EU (excluding UK) versus rest of the world origin: HR 1.19
Notes	Oncology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing cost‐effectiveness studies related to malignancies.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Yes	All meetings except the 2007 meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting, presence of sensitivity analyses, base year of analyses, time horizon, discounting, or cost‐utility analyses; and conflict of interest with publication using univariable Cox regression analysis.

Chand 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed in November 2007 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Methodology
Data	Included 2172 abstracts presented at the 1999 to 2005 Cardiac Society of Australia and New Zealand meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting Cumulative proportion of abstracts published Clinical research versus basic science research versus surgery
Outcomes	648 of 2172 abstracts published 71/253 abstracts presented at the 1999 meeting, 101/307 at the 2000 meeting, 96/307 at the 2001 meeting, 82/279 at the 2002 meeting, 97/331 at the 2003 meeting, 95/315 at the 2004 meeting, and 106/380 at the 2005 meeting published Proportion of abstracts published by time Mean time to publication = 18.0 months (SE = 1.8 months) Mean time to publication = 24.0 months (SE = 2.2 months) for 1999 Cardiac Society of Australia and New Zealand meeting Mean time to publication = 19.2 months (SE = 1.7 months) for 2000 Cardiac Society of Australia and New Zealand meeting Mean time to publication = 24.0 months (SE = 2.2 months) for 2001 Cardiac Society of Australia and New Zealand meeting Mean time to publication = 18.0 months (SE = 1.7 months) for 2002 Cardiac Society of Australia and New Zealand meeting Mean time to publication = 15.6 months (SE = 1.2 months) for 2003 Cardiac Society of Australia and New Zealand meeting Mean time to publication = 14.4 months (SE = 1.2 months) for 2004 Cardiac Society of Australia and New Zealand meeting Mean time to publication = 15.6 months (SE = 1.2 months) for 2005 Cardiac Society of Australia and New Zealand meeting Cumulative proportion of abstracts published at 24 months showed proportion published = 25.1% (544/2172 abstracts) Factors related to proportion of abstracts published included 396/1534 abstracts describing clinical research versus 52/53 abstracts describing surgery versus 200/585 abstracts describing basic science research published
Notes	Cardiology ‐ cardiothoracic surgery Funding by National Heart Foundation
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2004 had at least 48 months follow‐up. The meetings in 2004 and 2005 had 36 months and 24 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and type science with publication using stratified analysis.

Chapman 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Embase, Cochrane Methodology Register to May 2009, then to September 2010 for 2007 meeting Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Title  Handsearched and searched citations in review articles
Data	Included 908 abstracts presented at the 1997 to 2007 Cochrane Colloquia meetings Included all abstracts that presented "research relevant to the methods of systematic reviews or other health care evaluations"
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication, overall and by meeting Survival analysis of publication rate Oral versus poster presentation Methodologic/topic area
Outcomes	312 of 908 abstracts published 21/63 abstracts presented at the 1997 meeting, 21/58 at the 1998 meeting, 20/59 at the 1999 meeting, 10/32 at the 2000 meeting, 27/85 at the 2001 meeting, 12/89 at the 2002 meeting, 44/115 at the 2003 meeting, 45/113 at the 2004 meeting, 42/119 at the 2005 meeting, 37/133 at the 2006 meeting, and 33/92 at the 2007 meeting published Proportion of abstracts published by time Mean time to publication = 22.8  months, overall Mean time to publication = 37.2 months for 1997 Cochrane Colloquium Mean time to publication = 26.4 months for 1998 Cochrane Colloquium Mean time to publication = 28.8 months for 1999 Cochrane Colloquium Mean time to publication = 33.6 months for 2000 Cochrane Colloquium Mean time to publication = 26.4 months for 2001 Cochrane Colloquium Mean time to publication = 31.2 months for 2002 Cochrane Colloquium Mean time to publication = 24.0 months for 2003 Cochrane Colloquium Mean time to publication = 18.0 months for 2004 Cochrane Colloquium Mean time to publication = 16.8 months for 2005 Cochrane Colloquium Mean time to publication = 14.4 months for 2006 Cochrane Colloquium Mean time to publication = 19.2 months for 2007 Cochrane Colloquium Median time to publication = 24 months, overall Median time to publication = 36 months for 1997 Cochrane Colloquium Median time to publication = 24 months for 1998 Cochrane Colloquium Median time to publication = 24 months for 1999 Cochrane Colloquium Median time to publication = 36 months for 2000 Cochrane Colloquium Median time to publication = 24 months for 2001 Cochrane Colloquium Median time to publication = 30 months for 2002 Cochrane Colloquium Median time to publication = 24 months for 2003 Cochrane Colloquium Median time to publication = 24 months for 2004 Cochrane Colloquium Median time to publication = 12 months for 2005 Cochrane Colloquium Median time to publication = 12 months for 2006 Cochrane Colloquium Median time to publication = 12 months for 2007 Cochrane Colloquium Survival analysis of proportion published at 144 months = 36.4% Factors related to proportion of abstracts published included 123/312 abstracts presented orally versus 189/596 abstracts presented as poster published
Notes	Medical decision making/health policy‐ systematic reviews, health technology Funding from the National Institute for Health Research; National Health Research and Development Program, the Cochrane Collaboration
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	Yes	Searched 3 databases, and performed handsearches and citation searches.
Follow‐up time?	Yes	All meetings before 2006 had at least 48 months follow‐up. The meetings in 2006 and 2007 only had 36 months follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, and methodologic area with publication using stratified analysis.

Cheng 1998.

Methods	Identification of subsequent full‐length publications Searched electronic database Cochrane Cystic Fibrosis Group register, Embase, MEDLINE to 1995 Search completed by investigator Search criteria not reported Matched abstract to full‐length publication by All authors Titles Content Methodology
Data	Included 178 abstracts found in Cochrane Cystic Fibrosis group's register of trials Included all abstracts of RCTs except duplicates and those published before presentation
Comparisons	Proportion of abstracts published Median time to publication Survival analysis of publication rate 'Positive' versus not 'positive' Sample size equal to or above the median versus sample size below the median
Outcomes	57 of 178 abstracts published Proportion of abstracts published by time Median time to publication = 18 months Survival analysis of proportion published at 60 months = 40.1% Factors related to proportion of abstracts published included 43/113 'positive' (defined as experimental better than control) versus 14/42 not 'positive' abstract results published 36/92 abstracts with sample size equal to or above the median versus 25/78 abstracts with sample size below the median published
Notes	Other clinical specialties ‐ cystic fibrosis Funding by National Health Service
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all RCTs with reasonable exceptions.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Follow‐up of abstracts as short as 3 months to multiple years. Unclear how many abstracts were followed up for < 48 months.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of positive results and sample size with publication using stratified analysis and time to publication with univariate Cox regression models.

Chung 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, KoreaMed from August 2006 to August 2011 Search completed by investigator Searched by all authors, keywords, and names of schools or hospitals Matched abstract to full‐length publication by One author Title Methodology Results
Data	Included 1005 abstracts presented at the 2006 and 2007 Korean Urological Association meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Survival analysis of publication rate Clinical research versus basic science research versus 'other' research Subspecialty
Outcomes	421 of 1005 abstracts published Proportion of abstracts published by time Mean time to publication = 19.0 months (SD = 12.8) Survival analysis of proportion published = 41.9% at 60 months Factors related to proportion of abstracts published included 370/879 abstracts describing clinical research versus 39/82 abstracts describing basic science research versus 12/44 abstracts describing 'other' research published
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of type science and subspecialty with publication using stratified analysis and univariate logistic regression.

Chung 2012a.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/GoogleScholar, KoreaMed, and KMbase to 31 August, 2011 Search completed by investigators Searched by all authors, keywords, title, and affiliations Matched abstract to full‐length publication by One or more authors Conclusions
Data	Included 1176 abstracts presented at the 2005 to 2007 Spring and Fall meetings of the Congress of the Korean Society of Plastic and Reconstructive Surgeons Included all free communications and posters excluding symposia and panel sessions
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published Oral versus poster presentation Clinical research versus basic science research versus other type of research Author affiliation with teaching hospital versus private clinic versus both teaching and private clinic Subspecialty
Outcomes	455 of 1176 abstracts published 63/170 abstracts presented at the Spring 2005 meeting, 88/205 at the Fall 2005 meeting, 71/178 at the Spring 2006 meeting, 67/188 at the Fall 2006 meeting, 77/199 at the Spring 2007 meeting, and 89/236 at the Fall 2007 meeting published Proportion of abstracts published by time Mean time to publication = 15.0 months Cumulative proportion of abstracts published at 48 months showed proportion published = 38.4% (451/1176 abstracts) Factors related to proportion of abstracts published included 301/734 abstracts presented orally versus 154/442 abstracts presented as poster published 237/609 abstracts describing clinical research versus 64/125 abstracts describing basic science research versus 154/442 abstracts describing other type of research published 418/1033 abstracts with authors from teaching hospital versus 15/95 abstracts with authors from private clinic versus 22/48 abstracts with authors from both teaching and private clinics
Notes	Surgery ‐ plastic and reconstructive surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions, so low risk of bias.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, type science, author affiliation, and subspecialty with publication using univariate logistic regression.

Ciesla 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to April 2001 Person completing the search not reported Searched by all authors, keywords, and content Matched abstract to full‐length publication by All author   Title Results Introduction Data
Data	Included 257 abstracts presented at the 1998 International Academy of Pathology ‐ United States and Canadian Academy of Pathology, American Society of Cytopathology, and American Society of Clinical Pathologists/College of American Pathologists meetings Included all abstracts related to cytopathology
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published Academic versus non‐academic affiliation Subspecialty
Outcomes	116 of 257 abstracts published 38/66 abstracts presented at the International Academy of Pathology ‐ United States and Canadian Academy of Pathology meeting, 2/12 at the American Society of Clinical Pathologists/College of American Pathologists meeting, and 76/179 at the American Society of Cytopathology meeting published Proportion of abstracts published by time Mean time to publication = 15.8 months (range = 2 to 26 months) Cumulative proportion of abstracts published at 24 months showed proportion published = 21.8% (56/257 abstracts) Factors related to proportion of abstracts published included 83/241 abstracts originating from an academic center versus 2/16 abstracts not originating from an academic center published
Notes	Pathology ‐ cytopathology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meeting only had 36 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of meeting, academic center origin and subspecialty with publication using stratified analysis.

Cifuentes 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from January 1999 to June 2006 Person completing the search not reported Searched by all authors Matching criteria not reported
Data	Included 1533 abstracts, 620 submitted to the 1999 to 2002 Asociación Espanola para el Estudio del Hígado and 913 submitted for presentation at the 1999 European Association for the Study of the Liver meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting for the four meetings of the Asociación Espanola para el Estudio del Hígado Cumulative proportion of abstracts published for one meeting Oral versus poster presentation Accepted versus rejected for conference presentation
Outcomes	572 of 1533 abstracts published 248/913 abstracts submitted to the 1999 European Association for the Study of the Liver meeting, 93/176 at the 1999, 75/159 at the 2000, 63/114 at the 2001, and 93/171 at the 2002 Asociación Espanola para el Estudio del Hígado meetings published Proportion of abstracts published by time Mean time to publication = 20.9 months (SD 1.1) for 1999 Asociación Espanola para el Estudio del Hígado meeting Mean time to publication = 21.5 months (SD 1.6) for 2000 Asociación Espanola para el Estudio del Hígado meeting Mean time to publication = 20.2 months (SD 1.6) for 2001 Asociación Espanola para el Estudio del Hígado meeting Mean time to publication = 21.8 months (SD 4.0) for 2002 Asociación Espanola para el Estudio del Hígado meeting Cumulative proportion of abstracts published at 48 months for 1999 Asociación Espanola para el Estudio del Hígado meeting showed proportion published = 54.4% (93/176 abstracts published) Factors related to proportion of abstracts published included 127/204 abstracts presented orally versus 239/572 abstracts presented as posters published (2001 AEEH meeting not included) 366/776 abstracts accepted for presentation versus 143/643 rejected abstracts published (2001 AEEH meeting not included)
Notes	Gastroenterology ‐ hepatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting and meeting year, type of presentation, and acceptance for presentation with publication using multivariable logistic regression analysis.

Clemency 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from July 2011 through August 2012 Person completing the search not reported Searched by keywords, title and lead author Matching criteria not reported
Data	Included 635 abstracts presented at the 2003 to 2005 American College of Emergency Physicians, Society for Academic Emergency Medicine, National Association of EMS Physicians, Association of Air Medical Services, and National Association of EMS Educators meetings Included all abstracts that were related to emergency medicine except abstracts published prior to meeting
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published Subspecialty
Outcomes	281 of 635 abstracts published 62/128 at the American College of Emergency Physicians meeting, 72/135 at the Society for Academic Emergency Medicine meeting, 121/282 at the National Association of EMS Physicians meeting, 22/66 at the Association of Air Medical Services meeting, and 4/24 abstracts presented at the National Association of EMS Educators meeting published Proportion of abstracts published by time Mean time to publication = 22.2 months (SD = 16.5, range = 0 to 94 months) Cumulative proportion of abstracts published at 96 months showed proportion published = 43.9% (279/635 abstracts) No factors other than subspecialty related to proportion of abstracts published reported.
Notes	Emergency medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific topic, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting and subspecialty with publication using stratified analysis.

Cobos 1998.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Search completed by the investigators Search criteria not reported Matching criteria not reported
Data	Included 120 abstracts presented at the 1994 to 1995 Congreso Nacional de la Sociedad Española de la Cardiologia meetings Included a random sample of abstracts
Comparisons	Proportion of abstracts published
Outcomes	26 of 120 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Cardiology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random sample of abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

Cohen 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to May 2011 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 378 abstracts presented at the 2000 to 2005 Society of Gynecologic Oncologists meetings Included all abstracts for plenary sessions (main, focused and express)
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Clinical research versus basic science research versus 'other' research RCT design versus observational study design versus chart review versus 'translational' design Multi‐centered versus single center 'Fellowship affiliation' versus no 'fellowship affiliation'
Outcomes	309 of 378 abstracts published Proportion of abstracts published by time Median time to publication = 14 months (range = 1 to 85 months) Cumulative proportion of abstracts published at 48 months showed proportion published = 81.7% (309/378 abstracts) Factors related to proportion of abstracts published included 281/343 'positive' (defined as experimental better than control) versus  28/35 not 'positive' abstract results published 199/243 abstracts describing clinical research versus 66/89 abstracts describing basic science research versus 44/46 'other' studies published 24/26 abstracts with RCT design versus 82/106 abstracts with an observational design versus 93/111 chart reviews versus 66/89 abstracts with 'translational' design published 152/173 abstracts with multiple centers versus 157/205 abstracts with a single center published 226/272 abstracts with authors with 'fellowship affiliation' versus 83/106 abstracts without a 'fellowship affiliation' published
Notes	Oncology ‐ gynecologic cancer Funding by Dr. John A. Kerner Research Fund and Dr. Patricia Robertson and Innovations Fund
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts for plenary sessions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type science, study design, multi‐center status, author status with publication using multivariable logistic regression analysis.

Cohen 2012a.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to Summer 2008 Search completed by investigators Searched by all authors, and keywords Matched abstract to full‐length publication by All authors Sample size Title Methodology Results Hypothesis
Data	Included 2463 abstracts presented at the 2000 to 2004 American Academy of Otolaryngology, Head and Neck Surgery meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication by type of presentation and gender of first author Oral versus poster presentation Female versus male first author Oral presentation by female versus male first author Poster presentation by female versus male first author
Outcomes	1152 of 2463 abstracts published Proportion of abstracts published by time Mean time to publication = 19.1 months for poster presentations, overall Mean time to publication = 15.5 months for oral presentations, overall Mean to publication = 19.4 months for poster presentations by female first authors Mean to publication = 18.6 months for poster presentations by male first authors Mean to publication = 15.0 months for oral presentations by female first authors Mean to publication = 15.6 months for oral presentations by male first authors Factors related to proportion of abstracts published included 638/1051 abstracts presented orally versus 514/1413 abstracts presented as poster published 93/275 abstracts presented as poster by female first author versus 421/1138 abstracts presented as poster by male first author published 91/154 abstracts presented orally by female first author versus 547/897 abstracts presented orally by male first author published 184/429 abstracts presented by female first author versus 969/2035 abstracts presented by male first author published
Notes	Otolaryngology head and neck surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 6 different criteria.
Adjustment for confounding?	No	Examined association of type presentation, type presentation by author gender, and first author gender with publication using stratified analysis and Chi2 or Fisher's Exact tests

Collet 1993.

Methods	Identification of subsequent full‐length publications Manual search of databases Index to Dental Literature, Index Medicus; dates of search not reported Search completed by investigator Search criteria not reported Matched abstract to full‐length publication by All authors Title Contents
Data	Included 747 abstracts presented at the 1980 to 1989 Argentine Division of the International Association for Dental Research meetings Included all abstracts included in author index of meetings
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published
Outcomes	94 of 747 abstracts published 0/66 abstracts presented at the 1980 meeting, 1/86 at the 1981 meeting, 2/112 at the 1982 meeting, 15/92 at the 1983 meeting, 19/88 at the 1984 meeting, 20/67 at the 1985 meeting, 12/73 at the 1986 meeting, 10/82 at the 1987 meeting, 10/81 at the 1988 meeting, and 4/73 at the 1989 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 12.4% (93/747 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Oral health ‐ dentistry Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Collet 1997.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by all authors Matched abstract to full‐length publication by All authors Title
Data	Included 506 abstracts presented at the 1990 to 1995 Argentine Division of the International Association for Dental Research meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published Research center Subspecialty
Outcomes	61 of 506 abstracts published 13/65 abstracts presented at the 1990 meeting, 11/101 at the 1991 meeting, 6/75 at the 1992 meeting, 12/84 at the 1993 meeting, 8/104 at the 1994 meeting, and 7/127 at the 1995 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 96 months showed proportion published = 12.1% (61/506 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Oral health ‐ odontology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis.

Collet 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Search completed by investigator Searched by all authors Matched abstract to full‐length publication by All authors Titles Contents
Data	Included 1671 abstracts presented at the 1993 to 2003 Argentine Division of the International Association for Dental Research meetings Included all abstracts
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Research center Subspecialty
Outcomes	489 of 1671 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 120 months showed proportion published = 28.1% (470/1671 abstracts) No factors other than research center and subspecialty related to proportion of abstracts published reported
Notes	Oral health ‐dental research Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of research center and subspecialty with publication using stratified analysis.

Collier 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Search completed by investigator Searched by first, and last author, and keywords Matching criteria not reported
Data	Included 623 abstracts presented at the 2002 to 2006 British Association of Oral and Maxillofacial Surgeons meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Survival analysis of publication rate Clinical research versus 'scientific' research UK centers
Outcomes	147 of 623 abstracts published 26/112 abstracts presented at the 2002 meeting, 29/119 at the 2003 meeting, 42/124 at the 2004 meeting, 20/114 at the 2005 meeting, 30/154 at the 2006 meeting published Proportion of abstracts published by time Median time to publication = 8 months (range = 1 to 18 months) Survival analysis of proportion published at 60 months = 24.9% Factors related to proportion of abstracts published included 118/563 abstracts describing clinical research versus 29/60 abstracts describing 'scientific' research published
Notes	Otolaryngology head and neck surgery ‐ oral and maxillofacial surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type science, and UK center with publication using stratified analysis and Chi2 or Mann‐Whitney U tests.

Cornu 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and MEDLINE until October 2013 Person completing the search not reported Searched by first and last author Matching criteria not reported Searched publication lists of presenters and last authors with keywords
Data	Included 226 abstracts presented at the 2007 Congrès de l’Association Français d’Urologie Included all abstracts presented orally except 18 presentations (videos, invited speakers, sponsors)
Comparisons	Proportion of abstracts published Median time to publication Presenter an urologist in education versus urologist
Outcomes	96 of 226 abstracts published Proportion of abstracts published by time Median time to publication = 24 months Factors related to proportion of abstracts published included 41/77 abstracts from urologists in education versus 55/149 abstracts from other urologists published
Notes	Urology Funding not reported Thesis (memoire)
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally with reasonable exceptions.
Search for publications?	Yes	Searched 1 database and publication lists of authors.
Follow‐up time?	Yes	The meeting had 72 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of author status with publication using stratified analysis and Chi2 test.

Corry 1990.

Methods	Identification of subsequent full‐length publications Searched electronic database Index to Dental Literature, MEDLINE from 1983 through 1988 Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by All authors Sample size Title Contents Results Conclusions
Data	Included 275 abstracts presented at the 1983 and 1984 American Association of Dental Research meetings Included a random sample of all 2789 abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting
Outcomes	63 of 275 abstracts published 27/125 abstracts presented at 1983 meeting, and 63/150 at the 1984 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 21.6% (27/125 abstracts) for 1983 American Association of Dental Research meeting Cumulative proportion of abstracts published at 45 months showed proportion published = 24% (36/150 abstracts) for 1984 American Association of Dental Research meeting No factors other than meeting year related to proportion of abstracts published reported
Notes	Oral health ‐ dental research Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random sample of abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 6 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Costa 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by first, second, and last author Matching criteria not reported
Data	Included 1620 abstracts presented at the 2002 American Society of Clinical Oncology, San Antonio Breast Cancer Symposium, European Breast Cancer Conference, and European Society for Medical Oncology meetings Included all abstracts related to breast cancer
Comparisons	Proportion of abstracts published Median time to publication Survival analysis of publication rate overall and by meeting Presentation at a North American versus European meeting
Outcomes	721 of 1620 abstracts published Proportion of abstracts published by time Median time to publication = 11.6 months (95% CI = 7.6 to 15.7 months) Survival analysis of proportion published at 60 months = 44.5% Survival analysis of proportion published at 60 month for North American meetings = 52.2% and for European meetings = 32.5% Factors related to proportion of abstracts published included 52.7% of abstracts presented at a North American meeting versus 33.3% of abstracts presented at a European meeting published
Notes	Oncology ‐ breast cancer Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of country of origin with publication using stratified analysis and unspecified statistical tests.

Craig 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from 1966 through winter of 1999/2000 Search completed by investigator Searched by all authors and title Matched abstract to full‐length publication by All authors Contents Methodology Results
Data	Included 1005 abstracts presented at the 1980 to 1984 and 1990 to 1994 British Orthopaedic Association meetings Included all abstracts all except Spring meetings in 1982, 1992, 1993
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting decade Cumulative proportion of abstracts published by meeting decade
Outcomes	495 of 1005 abstracts published 184/320 abstracts presented at the 1980 to 1984 meetings, and 311/685 at the 1990 to 1994 meetings published Proportion of abstracts published by time Mean time to publication = 21.4 months for 1980 to 1984 British Orthopaedic Association meetings Mean time to publication = 16.8 months for 1990 to 1994 British Orthopaedic Association meetings Range for all meetings = 10.3 to 26.3 months Cumulative proportion of abstracts published for 1980 to 1984 British Orthopaedic Association meetings at 192 months showed proportion published = 57.5% (184/320 abstracts) Cumulative proportion of abstracts published for 1990 to 1994 British Orthopaedic Association meetings at 72 months showed proportion published = 45.1% (309/685 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Cromer 1998.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE for 5 years following meeting Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by Title Content
Data	Included 128 abstracts presented at the at the 1974, 1976, 1977, 1983, 1986, and 1993 meetings of Society of Adolescent Medicine meetings Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published
Outcomes	58 of 128 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Pediatrics ‐ adolescent medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted as oral presentations.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had 5 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.

Curry 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Search completed by investigator Search criteria not reported Matching criteria not reported
Data	Included 9 abstracts presented at the 1996 to 2000 British Association of Paediatric Surgeons meetings Included all abstracts describing RCTs
Comparisons	Proportion of abstracts published
Outcomes	1 of 9 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Surgery ‐ pediatric surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing RCTs.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

Czorlich 2016.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to November 30, 2014 Person completing the search not reported Searched by first and last author, and keywords Matched abstract to full‐length publication by Sample size Contents Methodology Results Hypothesis
Data	Included 433 abstracts presented at the 2012 German Society of Neurosurgery meetings Included all abstracts except invited manuscripts by Japanese Neurosurgical Society and 'best presentations out of the DGNC sections' (not peer reviewed)
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Oral versus poster presentation Clinical research versus 'experimental' research versus other type of research Retrospective design versus prospective design versus not applicable versus not identified study design Multi‐centered versus single center Academic versus non‐academic affiliation
Outcomes	175 of 433 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 36 months showed proportion published = 40.4% (175/433 abstracts) Factors related to proportion of abstracts published included 136/339 abstracts presented orally versus 39/94 abstracts presented as poster published 132/338 abstracts describing clinical research versus 37/78 abstracts describing 'experimental' research versus 6/17 abstracts describing other types of research published 50/124 abstracts with retrospective design versus 53/82 abstracts with prospective design versus 39/90 abstracts with other study designs versus 33/139 abstracts with unidentified study designs published 54/126 abstracts with multiple centers versus 121/307 abstracts with a single center published 170/404 abstracts originating in an academic center versus 5/20 abstracts not originating in an academic center published
Notes	Surgery ‐ neurosurgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meeting only had 24 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of type of presentation, type of science, prospective status, multi‐center status, and academic affiliation with publication using stratified analysis and Chi2 or t tests.

Dahllof 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed from 1999 through December 2006 Person completing the search not reported Searched by first, second, and last author, and keywords Matched abstract to full‐length publication by All authors Sample size Results Hypothesis
Data	Included 771 abstracts presented at the 1999 and 2001 International Association of Paediatric Dentistry meetings Included all abstracts except pre‐congress course, lectures, and symposia
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Oral versus poster presentation Clinical research versus basic science research versus 'other' research RCT design versus observational versus case report design US versus European versus Asian versus South American origin for countries contributing more than 20 abstracts Subspecialty
Outcomes	204 of 771 abstracts published Proportion of abstracts published by time Mean time to publication = 20 months (SD = 20) Median time to publication = 18 months Cumulative proportion of abstracts published at 80 months showed proportion published = 26.5% (204/771 abstracts) Factors related to proportion of abstracts published included 92/231 abstracts presented orally versus 110/539 abstracts presented as posters published 157/572 abstracts describing clinical research versus 40/146 abstracts describing basic science research versus 7/53 abstracts describing 'other' research published 24/47 abstracts with RCT design versus 113/373 abstracts with observational design versus 20/152 abstracts of case report design published 10/30 abstracts originating from the USA versus 91/259 abstracts from Europe versus 16/88 abstracts from Asia versus 8/81 abstracts from South America published (for countries submitting 20 or more abstracts); 79/343 abstracts published from countries submitting less than 20 abstracts
Notes	Oral health ‐ pediatric dentistry Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all posters with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of type presentation, type science, study design, US origin, and subspecialty with publication using multivariable logistic regression analysis.

Daluiski 1998.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Melvyl MEDLINE PLUS from January 1991 to June 1997 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Sample size Contents Methodology Results
Data	Included 888 abstracts presented at the 1991, 1992, and 1993 Orthopaedic Research Society meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication, overall and by meeting Cumulative proportion of abstracts published by meeting Subspecialty
Outcomes	463 of 888 abstracts published 147/296 abstracts presented at the 1991 meeting, 162/296 at the 1992 meeting, and 153/296 at the 1993 meeting published Proportion of abstracts published by time Median time to publication = 20 months Median time to publication for 1991 Orthopaedic Research Society = 18 months Median time to publication for 1992 Orthopaedic Research Society = 23 months Median time to publication for 1993 Orthopaedic Research Society = 20 months Cumulative proportion of abstracts published at 72 months showed proportion published = 24% (71/296 abstracts) for 1991 Orthopaedic Research Society meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 30.1% (89/296 abstracts) for 1992 Orthopaedic Research Society meeting Cumulative proportion of abstracts published at 54 months showed proportion published = 27% (80/296 abstracts) for 1993 Orthopaedic Research Society meeting No factors other than meeting year and subspecialty related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis and Chi2 tests.

Dangouloff‐Ros 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from October 2008 to October 2013 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by Sample size Contents
Data	Included 744 abstracts presented at the 2008 to 2010 French Society of Radiology meetings Included all abstracts presented orally except those from continuing medical education courses, letters, reviews, and editorials
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication Sample size <60 versus > 60 Clinical research versus experimental research European versus African versus rest of the world origin < 20 patients versus 20 to 40 patients versus 40 to 60 patients versus >60 patients versus unknown number of patients Retrospective versus prospective design Subspecialty
Outcomes	298 of 744 abstracts published 111/255 abstracts presented at the 2008 meeting versus 89/237 at the 2009 meeting versus 98/252 at the 2010 meeting published Proportion of abstracts published by time Mean time to publication = 21 months (SD = 13.7) Median time to publication = 18 months (IQR = 10 to 29 months) Factors related to proportion of abstracts published included 259/462 abstracts with a sample size < 60 versus 156/241 abstracts with a sample size < 60 published 271/698 abstracts describing clinical research versus 27/46 abstracts describing experimental research published 288/671 abstracts originating from Europe versus 3/53 abstracts originating from Africa versus 7/20 abstracts with rest of the world origin published 100/178 abstracts with < 20 patients versus 96/175 abstracts with 20 to 40 patients versus 63/109 abstracts with 40 to 60 patients versus 156/241 abstracts with > 60 patients versus 10/41 with unknown number of patients published 159/460 abstracts with retrospective design versus 139/284 abstracts with prospective design published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The 2010 meeting only had 36 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, sample size, type science, country of origin, prospective status, and subspecialty with publication using multivariable logistic regression analysis.

Daruwalla 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar; dates of search not reported Person completing the search not reported Searched by all authors, keywords, and title Matching criteria not reported
Data	Included 443 abstracts presented at the 2007 to 2010 and 2011 to 2013 Singapore Orthopaedic Association meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published, overall Podium versus poster presentation International versus local presenter Subspecialty
Outcomes	125 of 443 abstracts published 69/193 abstracts presented at the 2007 to 2010 meeting, and 56/250 at the 2011 to 2013 meeting published Proportion of abstracts published by time Median time to publication = 13 months (range = ‐73 to 76 months) Cumulative proportion of abstracts published at 72 months showed proportion published = 28.2% (125/443 abstracts) Factors related to proportion of abstracts published included Oral versus poster presentation: (OR 2.29; CI 95% 1.12 to 4.68; P = 0.02) International versus local presenter: (OR 1.74; CI 1.07 to 2.84; P = 0.027)
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting year, type presentation, author status, and subspecialty with publication using multivariable logistic regression analysis.

Das 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to June 2012 Person completing the search not reported Searched by first and last author, and keywords Matching criteria not reported
Data	Included 52 abstracts presented at the 2000 to 2008 American Epilepsy Society meetings Included only abstracts of randomized clinical trials; excluded 3 systematic reviews of RCTS, and 2 studies with pooled data from more than 2 different RCTs
Comparisons	Proportion of abstracts published Median time to publication
Outcomes	33 of 52 abstracts published Proportion of abstracts published by time Median time to publication = 15 months (SD = 3) No factors related to proportion of abstracts published reported
Notes	Neurology ‐ epilepsy Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing RCTs
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Davies 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, PubMed to October 2000 Search completed by investigator Searched by all authors Matched abstract to full‐length publication by All authors Sample size Content Methodology Results
Data	Included 172 abstracts presented at the 1997 Perinatal Society of Australia and New Zealand meeting Included all 'publishable' abstracts, defined as containing clear methods and results, from a total of 193 abstracts
Comparisons	Proportion of abstracts published Median time to publication Survival analysis of publication rate Oral versus poster presentation Clinical research versus basic science research Subspecialty
Outcomes	78 of 172 abstracts published Proportion of abstracts published by time Median time to publication = 18 months (range = ‐36 to 41 months; IQR = 9 to 26 months) Survival analysis of proportion published at 42 months = 44.8% Factors related to proportion of abstracts published included 67/125 abstracts presented orally versus 11/47 abstracts presented as posters published 48/127 abstracts describing clinical research versus 30/45 abstracts describing basic science research published
Notes	Pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	No	The meeting had a 36 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of type presentation, type science, and subspecialty with publication using stratified analysis and Chi2 tests.

de Andrade 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and SciELO to August 2010 Person completing the search not reported Searched by 'principal' author and title Matching criteria not reported
Data	Included 347 abstracts presented at the 2008 XXI Panamerican Congress of Trauma, VIII Congress of the Brazilian Society of Integrated Assistance to the Traumatized and X Brazilian Congress of Trauma Leagues meetings Included all abstracts
Comparisons	Proportion of abstracts published Nursing or physiotherapy versus medicine
Outcomes	10 of 347 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 6/49 abstracts related to nursing or physiotherapy versus 4/298 abstracts related to medicine published
Notes	Emergency medicine ‐ trauma Funding from FAPESP and CAPES
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	No	The meeting had 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of clinical topic with publication using stratified analysis and unspecified statistical test.

De Bellefeuille 1992.

Methods	Identification of subsequent full‐length publications Searched electronic database Cancerlink; date of search not reported Search completed by investigator Search criteria not reported Matched abstract to full‐length publication by All authors Keywords Contacted abstract authors directly if no publication found
Data	Included 197 abstracts, which was a random sample of 1058 abstracts submitted to the 1984 American Society of Clinical Oncology meeting
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Accepted versus rejected for conference presentation RCT design versus non‐RCT design
Outcomes	115 of 197 abstracts published Proportion of abstracts published by time Mean time to publication = 22.8 months Cumulative proportion of abstracts published at 66 months showed proportion published = 57.8% (114/197 abstracts) Factors related to proportion of abstracts published included 48/65 'positive' (defined as significant results) versus 67/132 not 'positive' abstract results published 63/81 abstracts accepted for presentation versus 52/116 rejected abstracts published 20/31 abstracts with RCT design versus 83/166 abstracts with non‐RCT design published
Notes	Oncology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random sample of all submitted abstracts.
Search for publications?	Yes	Searched 1 database and contacted authors.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, acceptance for presentation, and study design with publication using multivariable logistic regression.

de Meijer 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from January 1, 2002 to August 8, 2014 Search completed by investigators Searched by first author, and keywords Matched abstract to full‐length publication by All authors Contents Methodology
Data	Included 2174 abstracts submitted to the 2007 to 2012 Dutch Surgical Society meetings Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication, by decision on conference presentation and type of presentation Actuarial time to publication, by decision on conference presentation and type of presentation Accepted versus rejected for conference presentation Oral versus poster presentation Academic versus non‐academic affiliation Subspecialty
Outcomes	1358 of 2174 abstracts published Proportion of abstracts published by time Median time to publication = 25 months (95% CI = 23 to 27) for abstracts accepted for presentation Median time to publication = 36 months (95% CI = 30.1 to 41.9) for abstracts rejected for presentation Median time to publication = 23 months (95% CI = 20.9 to 25.1) for abstracts presented orally Median time to publication = 36 months (95% CI = 26.6 to 45.4) for abstracts presented as poster Actuarial time to publication at 60 months showed proportion published = 68.6% (895/1305 abstracts) for abstracts accepted for presentation Actuarial time to publication at 60 months showed proportion published = 57.7% (501/869 abstracts) for abstracts rejected for presentation Actuarial time to publication at 60 months showed proportion published = 72.6% (700/964 abstracts) for abstracts presented orally Actuarial time to publication at 60 months showed proportion published = 57.6% (196/341 abstracts) for abstracts presented as poster Factors related to proportion of abstracts published included 870/1305 abstracts accepted for presentation versus 488/869 rejected abstracts published 680/964 abstracts presented orally versus 190/341 abstracts presented as poster published 853/1154 abstracts originating in an academic center versus 505/1020 abstracts not originating in an academic center published
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2011 had at least 48 months follow‐up. The meetings in 2011 and 2012 only had 36 and 24 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of acceptance for presentation, type presentation, academic affiliation, and subspecialty with publication using stratified analysis and t tests, Chi2, Mann‐Whitney U, or ANOVA tests.

de Oliveira 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed, Google/Google Scholar, Scielo, and LILACS until March 1, 2012 Person completing the search not reported Searched by first and second author, and title Matched abstract to full‐length publication by One author Searched publication lists of 1st and 2nd author on Lattes platform
Data	Included 348 abstracts presented at the 2010 Brazilian Congress on Pain meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean and median time to publication Oral versus poster presentation
Outcomes	31 of 348 abstracts published Proportion of abstracts published by time Mean time to publication = 6.7 months (SD = 3) Median time to publication = 6.9 months Factors related to proportion of abstracts published included 14/121 abstracts presented orally versus 17/227 abstracts presented as poster published
Notes	Neurology ‐ pain Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 4 databases and searched Lattes platform.
Follow‐up time?	No	The meeting only had 24 months follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis and Fisher's Exact tests.

De Sio 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for a 2‐year period Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by Title Methodology Hypotheses
Data	Included 3110 abstracts presented at the 1998 to 1999 and 2008 to 2009 European Association of Urology meetings Included all abstracts of randomized and nonrandomized controlled trials, excluding video abstracts
Comparisons	Proportion of abstracts published, overall and by meeting RCT design versus non‐RCT design US‐origin versus European origin Rest of the world versus European origin Multi‐centre versus single centre Subspecialty
Outcomes	453 of 3110 abstracts published 190/1336 abstracts presented at the 1998 to 1999 meetings, and 263/1774 at the 2008 to 2009 meetings published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 65/375 abstracts describing RCT versus 388/2764 abstracts describing nonrandomized controlled trials published USA versus Europe: OR 1.03 (95% CI = 0.68 to 1.56) Rest of the world versus Europe: OR 0.96 (95% CI = 0.72 to 1.28 Multi‐institutional versus single institution: OR 1.82 (95% CI = 1.44 to 2.3)
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing RCTs and controlled clinical trials.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meetings only had 24 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, study design, multi‐center status, US origin, country of origin, and subspecialty with publication using multivariable logistic regression analysis.

Del Rio Moro 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Enfispo, and Cuiden until October 31, 2009 Search completed by the investigator Searched by all authors, keywords, and title Matching criteria not reported
Data	Included 610 abstracts presented at the 2001 to 2007 Asociación Española de Enfermería en Cardiología meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	76 of 610 abstracts published 4 of 60 abstracts presented at the 2001 meeting, 5/72 at the 2002 meeting, 4/71 at the 2003 meeting, 14/102 at the 2004 meeting, 6/71 at the 2005 meeting, 14/111 at the 2006 meeting, and 16/120 at the 2007 meeting published Proportion of abstracts published by time not reported No factors other than meeting year related to proportion of abstracts published reported
Notes	Nursing ‐ cardiology nursing Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings except the meetings in 2006 and 2007 had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis and Chi2 tests.

Delamere 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by content and publication type Matching criteria not reported
Data	Included 30 abstracts included in the Cochrane Skin Group specialized register and presented at various 2000 to 2002 meetings Included a random sample of all abstracts of RCTs
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive'
Outcomes	13 of 30 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 12/22 'positive' (defined as experimental better than control) versus 1/8 not 'positive' abstract results published
Notes	Dermatology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random sample of all abstracts of RCTs.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results with publication using stratified analysis.

Delgado Gallego 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and Indice Medicao Espanol in July 2005 Person completing the search not reported Searched by first author Matching criteria not reported
Data	Included 528 abstracts presented at the 2001 Sociedad Española de Medicina General, Sociedad Española de Medicina Rural y Generalista, and Sociedad Española de Medicina de Familia y Comunitaria meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting Oral versus poster presentation
Outcomes	57 of 528 abstracts published 6/102 abstracts presented at the Sociedad Española de Medicina General, 10/124 abstracts at the Sociedad Española de Medicina Rural y Generalista, and 41/302 at the Sociedad Española de Medicina de Familia y Comunitaria meetings published Proportion of abstracts published by time Cumulative proportion of abstracts published at 48 months = 5.9% (6/102 abstracts) for the Sociedad Española de Medicina General meeting Cumulative proportion of abstracts published at 48 months = 8.1% (10/124 abstracts) for the Sociedad Española de Medicina Rural y Generalista meeting Cumulative proportion of abstracts published at 48 months = 14.9% (45/302 abstracts) for the Sociedad Española de Medicina de Familia y Comunitaria meeting Factors related to proportion of abstracts published included 32/246 abstracts presented orally versus 25/282 abstracts presented as poster published
Notes	General medicine /primary care Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and type presentation with publication using stratified analysis and Chi2 tests.

DeMola 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to April 2007 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by Sample size Content Methodology Results
Data	Included 558 abstracts presented at the 1999 to 2004 American Academy of Orthopaedic Surgery meetings Included all abstracts in shoulder and elbow sections
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published Sample size Oral versus poster presentation North American origin Subspecialty
Outcomes	321 of 558 abstracts published, 48.4% of abstracts presented at the 1999 meeting, 59.6% at the 2000 meeting, 65.5% at the 2001 meeting, 56.5% at the 2002 meeting, 59.4% at the 2003 meeting, and 57.5% at the 2004 meeting published Proportion of abstracts published by time Mean time to publication = 31 months (SD = 28) Cumulative proportion of abstracts published at 36 months showed proportion published = 30% (167/558 abstracts) Factors related to proportion of abstracts published included Published abstracts had greater sample size (P < 0.01) 154/233 abstracts presented orally versus 166/325 abstracts presented as poster published Published abstracts were more often of North American origin (P < 0.05)
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific topics, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2004 had at least 48 months follow‐up. The meeting in 2004 only had 36 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of sample size, type presentation, North American origin and subspecialty with publication using multivariable logistic regression analysis.

Dhaliwal 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to January 2007 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author One outcome
Data	Included 200 abstracts submitted for the 2000 All India Ophthalmological Conference meeting Included only those published in the proceedings book from 278 abstracts submitted to the meeting
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published RCT design versus experimental design versus observational design versus descriptive design India versus non‐India origin 'Completeness of reporting' Subspecialty
Outcomes	33 of 200 abstracts published Proportion of abstracts published by time Mean time to publication = 22.8 months (SD = 16.4) Median time to publication = 20 months (range = 2 to 77 months) Cumulative proportion of abstracts published at 72 months showed proportion published = 17% (34/200 abstracts; includes two publications from one abstract) Factors related to proportion of abstracts published included 3/13 abstracts with RCT design versus 13/99 abstracts with experimental design versus 9/48 abstracts with observational design versus 8/40 abstracts with descriptive design published 31/194 abstracts originating from India versus 2/6 abstracts originating elsewhere published
Notes	Ophthalmology/optometry ‐ ophthalmology Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 7 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of study design, Indian origin, 'completeness of reporting', and subspecialty with publication using combinations of variables with ANOVA with stratified analysis and Tukey and t tests.

Dicembrino 2014.

Methods	Identification of subsequent full‐length publications Contacted abstract authors directly
Data	Included 232 abstracts accepted to the 2005 to 2009 Sociedad Latinoamérica de Investigación Pediátrica meetings Included 232 out of 325 abstracts with author response
Comparisons	Proportion of abstracts published Funding versus no funding Argentinean versus Bolivian versus Brazilian versus Chilean versus Mexican versus Peruvian versus Paraguayan versus rest of the world origin
Outcomes	136 of 232 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 64/92 abstracts with funding versus 72/140 abstracts without funding published 56/88 abstracts originating from Argentina versus 1/2 abstracts originating from Bolivia versus 24/36 abstracts originating from Brazil versus 37/56 abstracts originating in Chile versus 4/8 abstracts originating in Mexico versus 9/32 abstracts originating in Peru versus 3/8 abstracts originating in Paraguay versus 2/2 abstracts originating elsewhere published
Notes	Pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	No	Included abstracts with author response only.
Search for publications?	No	Contacted abstract authors with response rate < 80%.
Follow‐up time?	Unclear	Date of author contact not reported.
Matching?	Yes	Matched through contact with abstract author.
Adjustment for confounding?	No	Examined association of funding status and Latin American country origin with publication using stratified analysis and Chi2 tests.

Diezel 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE (from 1966 through 1996), PsychLIT (from 1974 to 1996), Embase (from 1988 through 1985), Biological abstracts (from 1985 through 1996), ISI (from 1981 through 1996), Cochrane Central Register of Controlled Trials (from 1996 through 1998) Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 95 abstracts presented at the 1971 Vth World Congress of Psychiatry meeting Included RCT abstracts
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published English versus non‐English language
Outcomes	44 of 95 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 84 months = 46.3% (44/95 abstracts) Factors related to proportion of abstracts published included 32/58 abstracts with English language versus 12/37 abstracts with non‐English language published
Notes	Psychiatry Funding by Amt fur Ausbildungsförderung, Nordrhein‐Westfahlen, Germany, the Medical Research Council, and Anglia and Oxford Regional Health Authority, UK
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described RCTs.
Search for publications?	Yes	Searched 6 databases.
Follow‐up time?	Yes	The meeting had at least 14 years follow up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined origin in English language country with publication using stratified analysis and unspecified statistical tests.

Dirk 1996.

Methods	Identification of subsequent full‐length publications Contacted abstract authors directly
Data	Included 147 abstracts by faculty from single anesthesiology department accepted for various conferences
Comparisons	Proportion of abstracts published Median time to publication
Outcomes	80 of 147 abstracts published Proportion of abstracts published by time Mean time to publication = 32.4 months (range = ‐12 to 96 months) No factors related to proportion of abstracts published reported
Notes	Anesthesiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts from a pre‐specified group of authors.
Search for publications?	Unclear	Author contact, but response rate not reported.
Follow‐up time?	Unclear	Date of author contact not reported.
Matching?	Yes	Matched through contact with abstract author.

Donegan 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed to March 2006 Search completed by investigator Searched by all authors, title, and keywords Matched abstract to full‐length publication by Title Contents "Authorship"
Data	Included 756 abstracts presented at the 2001 American Academy of Orthopaedic Surgeons meeting Included all abstracts
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Oral versus poster presentation
Outcomes	367 of 756 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 48.5% (367/756 abstracts) Factors related to proportion of abstracts published included 149/288 abstracts presented orally versus 218/468 abstracts presented as posters published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis and Chi2 tests.

Dooley 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, International Pharmaceutical Abstracts to November 2002 Person completing the search not reported Searched by first and second author Matched abstract to full‐length publication by Title Contents   Handsearched the Journal of Pharmacy Practice and Research
Data	Included 187 abstracts presented at the 1999 24th Federal Conference of the Society of Hospital Pharmacists of Australia meeting Included abstracts except invited presentations
Comparisons	Proportion of abstracts published Oral versus poster presentation Multi‐centered versus single center
Outcomes	21 of 187 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 15/82 abstracts presented orally versus 6/105 abstracts presented as posters published 15/61 abstracts with multiple centers versus 6/126 abstracts with a single center published
Notes	Pharmacology ‐ hospital pharmacology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases, and performed hand searches.
Follow‐up time?	No	The meeting only had 36 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of type of presentation and multi‐center status with publication using stratified analysis and Chi2 tests.

Doğan 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Tűrk Medline and Ulakbim; dates of search not reported Person completing the search not reported Searched by first author and title Matching criteria not reported
Data	Included 218 abstracts presented at the 2007 to 2010 Turkish National Rhinology Congresses meetings Included all abstracts except those published before congress
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by type of presentation Oral versus poster presentation Clinical research versus experimental research Case report design versus experimental design versus other study design University versus teaching hospital versus other type of institution
Outcomes	61 of 218 abstracts published 15/53 abstracts presented at the 2007 meeting, 18/55 at the 2008 meeting, 14/61 at the 2009 meeting, and 14/49 at the 2010 meeting published Proportion of abstracts published by time Mean time to publication = 21.8 months, overall Mean time to publication = 20.8 months for poster presentations Mean time to publication = 23.2 months for oral presentations Factors related to proportion of abstracts published included 24/67 abstracts presented orally versus 37/151 abstracts presented as poster published 58/209 abstracts describing clinical research versus 3/9 abstracts describing experimental research published 19/92 abstracts describing case reports versus 3/9 abstracts describing experimental design versus 39/117 abstracts describing other type of study design published 26/75 abstracts originating from university versus 33/138 abstracts from teaching hospital versus 2/5 abstracts from other type of institution published
Notes	Otolaryngology head and neck surgery Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type of presentation, type of science, study design, university origin with publication using stratified analysis and Chi2 tests.

Dressler 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and university institution websites to July 2012 Person completing the search not reported Searched by all authors, keywords, and title Matching criteria not reported
Data	Included 402 abstracts presented at the 2005 to 2011 Canadian Association of Radiologists meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published 'Radiologists‐in‐training' exhibits versus scientific exhibits versus educational exhibits versus proffered exhibit versus audit exhibits Canadian university
Outcomes	112 of 402 abstracts published 15/47 abstracts presented at the 2005 meeting, 12/57 at the 2006 meeting, 17/49 at the 2007 meeting, 19/55 at the 2008 meeting, 19/59 at the 2009 meeting, 20/64 at the 2010 meeting, and 10/71 at the 2011 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 27.9% (112/402 abstracts) Factors related to proportion of abstracts published included 39/196 abstracts presented in the radiologists‐in‐training exhibits versus 34/100 abstracts presented in the scientific exhibits versus 34/170 abstracts presented in the educational exhibits versus 5/15 abstracts presented in the proffered exhibits versus 0/11 abstracts presented in audit exhibits published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases and university institution web sites.
Follow‐up time?	Yes	All meetings except the 2009 to 2011 meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year by author status, and Canadian university origin with publication using stratified analysis.

Drury 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by all authors Matching criteria not reported Contacted abstract authors directly
Data	Included 909 abstracts presented at the 1993 to 2007 Society for Cardiothoracic Surgery meetings Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication RCT or systematic review design versus other designs UK region Prospective versus retrospective Subspecialty
Outcomes	606 of 909 abstracts published Proportion of abstracts published by time Median time to publication for 2003 to 2007 Society for Cardiothoracic Surgery meetings = 15 months (range = ‐24 to 63 months) Factors related to proportion of abstracts published included 63/74 abstracts with RCT or systematic review design versus 129/208 abstracts with other types of design published (2003 to 2007 Society for Cardiothoracic Surgery meetings)
Notes	Surgery ‐ cardiovascular surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of type presentation, study design, origin within UK, prospective status, award winning status, and subspecialty with publication using multivariable logistic regression analysis.

Duchini 1997.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; from 1991 to 1996 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 2512 abstracts submitted to the 1991 Digestive Diseases Week meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication, overall and by country Clinical versus basic science research North American versus European versus rest of world origin US versus rest of world origin Subspecialty
Outcomes	1229 of 2512 abstracts published Proportion of abstracts published by time Mean time to publication = 20.4 months (SD =12 months) Factors related to proportion of abstracts published included 47.2% of abstracts describing clinical research versus 54.1% of abstracts describing clinical research published 563/1221 abstracts originating from North America versus 406/766 abstracts originating from Europe versus 89/179 abstracts originating from Japan versus 24/55 abstracts originating from Australia versus 78/129 abstracts with multinational origin published 519/1134 abstracts originating from the US versus 563/1087 abstracts originating from the rest of the world published
Notes	Gastroenterology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database
Follow‐up time?	Yes	The meeting had 5 years of follow‐up
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of country of origin, type science, and subspecialty with publication using stratified analysis and Fisher's exact test.

Dudley 1978.

Methods	Identification of subsequent full‐length publications Manual search of database Index Medicus for 3 years following the meeting Person completing the search not reported Searched by all authors and contents Matching criteria not reported
Data	Included 51 abstracts presented at the 1972 Surgical Research Society of Great Britain meeting Included all abstracts
Comparisons	Proportion of abstracts published
Outcomes	29 of 51 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meeting only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Dumville 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, Embase, MEDLINE, and PubMed to November 2005 Search completed by investigator Searched by first, and last author, and keywords Matched abstract to full‐length publication by Contents Methodology Contacted abstract authors directly
Data	Included 467 abstracts presented at the 2001 and 2002 European Wound Management Association Conference meetings Included all abstracts except those from workshops and historical views of wound care
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Cox regression publication rate Subspecialty
Outcomes	57 of 467 abstracts published Proportion of abstracts published by time Cox regression for time to publication showed RCTs published sooner than other study designs (HR = 2.36, 95% CI = 1.02 to 5.45) Factors related to proportion of abstracts published included Cox regression showed 'positive' abstracts (defined as significant results) more likely to have been published (HR = 1.79, 95% CI = 1.26 to 2.55)
Notes	Other clinical specialties ‐wound management Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors.
Follow‐up time?	Yes	The 2001 meeting had 48 months follow‐up, the 2002 meeting only had 36 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, positive results, intervention, type of presentation, study design, and subspecialty with publication using multivariable Cox regression analysis

Durinka 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed 'in 2013' Search completed by investigator Searched by first, second, and last author, and title Matched abstract to full‐length publication by Content Sample Size
Data	Included 1938 abstracts accepted to the 2009 American Transplant Congress meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Oral versus poster presentation Multi‐center versus single center University affiliation Country of first author
Outcomes	244 of 1938 abstracts published Proportion of abstracts published by time Mean time to publication = 16 months (range = 1 to 49 months) Factors related to proportion of abstracts published included 103/641 abstracts presented orally versus 141/1297 abstracts presented as poster published authors reported number of abstracts published by factor, but not the total number of abstracts within each factor
Notes	Surgery ‐ organ transplantation Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Unclear	Included all abstracts.
Search for publications?	Unclear	Searched only 1 database.
Follow‐up time?	Unclear	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 2 criteria.
Adjustment for confounding?	No	Examined association of type presentation, multi‐center status, university affiliation, country of origin, and subspecialty with publication simple stratified analysis. Also examined association of study design by type of presentation, and multi‐center status by type of presentation with publication using stratified analysis and Chi2 tests.

Durinka 2014b.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Search completed by the investigator Search criteria not reported Matched abstract to full‐length publication by First author Second author Last author Title Contents
Data	Included 335 abstracts presented at the 2009 American College of Surgeons meeting Included all abstracts accepted as poster
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	64 of 335 abstracts published Proportion of abstracts published by time Mean time to publication = 16.8 months No factors related to proportion of abstracts published reported
Notes	Surgery Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented as posters.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 3 different criteria.

Duthie 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed, PsychINFO, ScienceDirect in November 2011 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Title Contacted abstract authors directly
Data	Included 20 abstracts presented at the 2007 to 2009 Senior Trainees’ Annual Research Symposium Included all abstracts
Comparisons	Proportion of abstracts published Cohort/cross‐sectional study versus case‐control study versus survey versus audit versus pilot study Mean time to publication
Outcomes	10 of 20 abstracts published Proportion of abstracts published by time Mean time to publication = 20.2 months (range = 5 to 41 months) Factors related to proportion of abstracts published included 3/6 abstracts describing cohort/cross‐sectional studies versus 1/1 abstract describing case‐control studies versus 4/7 abstracts describing a survey versus 1/4 abstracts describing an audit versus 0/2 abstracts describing a pilot study published
Notes	Psychiatry Funding not reported Data abstracted from 'brief report'
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors directly.
Follow‐up time?	No	All meetings except the 2007 meeting had less than 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of study design with publication using stratified analysis

Dyson 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and CAB direct for 'at least 4 years since most recent meeting' Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Sample size Methodology Results Conclusions Contacted abstract authors directly
Data	Included 283 abstracts presented at the 1990 to 1993 and 1995 to 1999 American College of Veterinary Anesthesiologists meetings Included all abstracts except one educational abstract and two case reports
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting
Outcomes	208 of 283 abstracts published 29/33 abstracts presented at the 1990 meeting, 16/24 at the 1991 meeting, 23/28 at the 1992 meeting, 24/26 at the 1993 meeting, 24/38 at the 1995 meeting, 31/39 at the 1996 meeting, 17/30 at the 1997 meeting, 19/31 at the 1998 meeting, and 25/34 at the 1999 meeting published Proportion of abstracts published by time Mean time to publication = 24.3  months (SD = 21) Mean time to publication for 1990 American College of Veterinary Anesthesiologists meeting = 39.3 months (SD = 33.2) Mean time to publication for 1991 American College of Veterinary Anesthesiologists meeting = 24.8 months (SD = 32.6) Mean time to publication for 1992 American College of Veterinary Anesthesiologists meeting = 20.9 months (SD = 33.7) Mean time to publication for 1993 American College of Veterinary Anesthesiologists meeting = 23.2 months (SD = 21.3) Mean time to publication for 1995 American College of Veterinary Anesthesiologists meeting = 22.9 months (SD = 14.4) Mean time to publication for 1996 American College of Veterinary Anesthesiologists meeting = 24.8 months (SD = 19.5) Mean time to publication for 1997 American College of Veterinary Anesthesiologists meeting = 20.8 months (SD = 16.1) Mean time to publication for 1998 American College of Veterinary Anesthesiologists meeting = 19.4 months (SD = 14.3) Mean time to publication for 1999 American College of Veterinary Anesthesiologists meeting = 21.4 months (SD = 10.0) No factors other than meeting year related to proportion of abstracts published reported
Notes	Other non‐clinical specialties ‐ veterinary anesthesiology Funding by Pet Trust, Ontario Veterinary College, University of Guelph
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases and contacted abstract authors.
Follow‐up time?	Yes	Follow‐up was 'at least 4 years since most recent meeting'.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Eck 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to November 2003 Person completing the search not reported Searched by all authors Matched abstract to full‐length publication by One author Title Methodology Results
Data	Included 358 abstracts presented at the 1997 International Society of Arthroscopy, Knee Surgery and Sports Medicine and 1999 Biennial International Society of Arthroscopy, Knee Surgery and Sports Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting
Outcomes	133 of 358 abstracts published 56/162 abstracts presented at the 1997 meeting, and 77/196 at the 1999 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 78 months showed proportion published = 34.6% (56/162 abstracts) for 1997 International Society of Arthroscopy, Knee Surgery and Sports Medicine meeting Cumulative proportion of abstracts published at 54 months showed proportion published = 38.8% (76/196 abstracts) for 1999 Biennial International Society of Arthroscopy, Knee Surgery and Sports Medicine meeting No factors other than meeting year related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ sports medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Ejnisman 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and LILACS; dates of search not reported Person completing the search not reported Searched by first, second, and last author Matched abstract to full‐length publication by All authors Title Summary
Data	Included 653 abstracts accepted to the 2007 Congresso Brasileiro de Ortopedia meeting Included all abstracts
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Oral versus poster presentation Brazilian versus rest of the world origin Subspecialty
Outcomes	174 of 653 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 26.0% (170/653 abstracts) Factors related to proportion of abstracts published included 118/286 abstracts presented orally versus 56/367 abstracts presented as poster published 173/641 abstracts originating from Brazil versus 1/12 abstracts with rest of the world origin published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of type presentation and Brazilian state origin with publication using stratified analysis.

Elder 1994.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Health Planning & Administration abstracts to 1992 Search completed by trained librarian Searched by presenter name Matched abstract to full‐length publication by All authors Title Contents
Data	Included 475 abstracts presented at the 1987 and 1988 North American Primary Care Research Group and Society for Teachers of Family Medicine meetings Included all abstracts related to research and peer presentations
Comparisons	Proportion of abstracts published Survival analysis of publication rate
Outcomes	226 of 475 abstracts published Proportion of abstracts published by time Survival analysis of proportion published at 48 months = 47.6% No factors related to proportion of abstracts published reported
Notes	General medicine/primary care Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific topics, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.

Elliot 2016.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, SCOPUS, and Google/Google Scholar between January 1, 2011 and April 30, 2011 Search completed by investigators Searched by first, second, last, and corresponding author, and keywords Matched abstract to full‐length publication by All authors Title Methodology Results Objectives
Data	Included 754 abstracts presented at the 2005 Congress of Neurological Surgeons meeting Included all abstracts except 1 withdrawn and 1 duplicate abstract
Comparisons	Proportion of abstracts published Median time to publication 'Positive' versus not 'positive' Oral versus poster presentation Clinical research versus basic science research RCT design versus observational design versus case report versus meta‐analysis US origin versus Canadian versus rest of the world origin Principal author with PhD versus not 'New technology/technique' versus not Subspecialty
Outcomes	383 of 754 abstracts published Proportion of abstracts published by time Median time to publication = 14.8 months (range = ‐21.2 to 64.2 months) Factors related to proportion of abstracts published included 84/135 'positive' (defined as significant results) versus 299/619 not 'positive' abstract results published 73/114 abstracts presented orally versus 308/638 abstracts presented as posters published 295/601 abstracts describing clinical research versus 88/153 abstracts describing basic science research published 10/17 abstracts with RCT design versus 219/429 abstracts describing observational design versus 52/125 abstracts describing case reports versus 5/7 abstracts describing meta‐analyses published 274/534 abstracts originating from the US versus 19/31 abstracts originating from Canada versus 90/189 abstracts originating elsewhere published 91/158 abstracts with the principal author having a PhD versus 292/596 abstracts with principal author without PhD published 146/278 abstracts describing a new technology/technique versus 237/476 abstracts not describing a new technology/technique
Notes	Surgery ‐ neurologic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The meeting had 72 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, type presentation, type science, study design, US origin, author status, new technology, and subspecialty with publication using multivariable logistic regression analysis.

Eloubeidi 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL (1982 to June 1998), MEDLINE (1966 to June 1998), CANCERLIT (1983 to June 1998), Health Star (1975 to June 1998), and Current Contents (1997 to June 1998) Search completed by trained librarian Searched by first and senior author Matched abstract to full‐length publication by Sample size Keywords Title Methodology Results Data
Data	Included 451 abstracts submitted to the 1994 American Society for Gastrointestinal Endoscopy meeting Included all abstracts except 2 withdrawn and 8 electronic poster abstracts
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation Accepted versus rejected for conference presentation RCT design versus non‐RCT design Multi‐centered versus single center 'Prospective' versus 'retrospective' design US versus non‐US origin Academic versus non‐academic affiliation
Outcomes	113 of 451 abstracts published Proportion of abstracts published by time Mean time to publication = 20 months (SD = 11.4) Median time to publication = 17.8 months (IQR = 12.7 to 27.9 months) Cumulative proportion of abstracts published at 48 months showed proportion published = 20.8% (94/451 abstracts) Factors related to proportion of abstracts published included 36/98 'positive' (defined as significant results) versus 77/353 not 'positive' abstract results published 16/40 abstracts presented orally versus 64/207 abstracts presented as posters published 80/247 abstracts accepted for presentation versus 33/204 rejected abstracts published 14/41 abstracts with RCT design versus 99/410 abstracts with non‐RCT design published 19/77 abstracts with multiple centers versus 94/374 abstracts with a single center published 64/216 abstracts with 'prospective' study design versus 49/235 abstracts with 'retrospective' study design published 81/350 abstracts originating in the US versus 32/101 abstracts not originating in the US published 101/392 abstracts originating in an academic center versus 47/59 abstracts not originating in an academic center versus published
Notes	Gastroenterology ‐ endoscopy Funding by Veteran's Affairs Health Services Research ‐ Training Grant & Career Development Program
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 5 databases.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 6 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, acceptance for presentation, study design, multi‐center status, prospective status, US origin, and academic affiliation with publication using multivariable logistic regression analysis.

Ensom 1998.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1992 through March 1998 Search completed by investigator Searched by all authors Matching criteria not reported Contacted abstract authors directly Handsearched Canadian Journal of Hospital Pharmacy from 1992 to "present"
Data	Included 363 abstracts presented at the 1992 to 1996 Canadian Society of Hospital Pharmacists meetings Included all 'award winning' abstracts for 1992 and all abstracts published in the Canadian Journal of Hospital Pharmacists for remaining years
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication 'Award winning' versus not 'award winning'
Outcomes	89 of 363 abstracts published 9/11 abstracts presented at the 1992 meeting, 13/56 at the 1993 meeting, 17/84 at the 1994 meetings, 27/105 at the 1995 meeting, and 23/107 at the 1996 meeting published. Proportion of abstracts published by time Mean time to publication = 11 months, but includes abstracts with < 24 months of follow‐up Factors related to proportion of abstracts published included 28/57 'award winning' abstracts versus 61/306 not 'award winning' abstracts published
Notes	Pharmacology ‐ hospital pharmacology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 1 database, contacted abstract authors, and hand searched a journal.
Follow‐up time?	Yes	All meetings before 1995 had at least 48 months follow‐up. The meetings in 1995 and 1996 only had 36 and 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, and award winning status with publication using stratified analysis and Fisher's Exact tests

Erdağ 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Türk Medline, and Ulakbim until July 1, 2013 Person completing the search not reported Searched by first author, keywords, and title Matching criteria not reported Manually searched the Journal of International Advanced Otology and the journal Sprache Stimme Gehör
Data	Included 1454 abstracts presented at the 2008 to 2010 Turkish National Otorhinolaryngology and Head & Neck Surgery meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Oral versus poster presentation Case report versus experimental research versus clinical research (undefined) University hospital versus training hospital versus other institution
Outcomes	319 of 1454 abstract published 111/395 abstracts presented at the 2008 meeting, 107/571 at the 2009 meeting, and 101/488 at the 2010 meeting published Proportion of abstracts published by time Mean time to publication = 18.6 months Factors related to proportion of abstracts published included 133/357 abstracts presented orally versus 186/1097 abstracts presented as poster published 133/774 abstracts describing case reports versus 25/55 abstracts describing 'experimental' research versus 161/625 abstracts describing clinical research (undefined) published 160/750 abstracts from university hospitals versus 148/644 abstracts from training hospitals versus 10/44 abstracts from other institutions published
Notes	Otolaryngology head and neck surgery Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	Yes	Searched 3 databases and manually searched 2 journals.
Follow‐up time?	Yes	The meeting in 2010 only had 36 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, study design, and university affiliation with publication using stratified analysis and Chi2 tests.

Ersoy 2015a.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, Turkish National Database, and Turkish Citation Index until 01 August, 2014 Person completing the search not reported Searched by all authors, keywords, and title Matching criteria not reported
Data	Included 243 abstracts presented at the 2009 Turkish Association of Gynecology and Obstetrics National Gynecology and Obstetrics Congress meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication, by origin and presentation format Oral versus poster presentation University hospital versus education and research hospital versus private institutions versus state hospitals Subspecialty
Outcomes	84 of 243 abstracts published Proportion of abstracts published by time Mean time to publication = 17 months (SD = 2) for international publications Mean time to publication = 11 months (SD = 4) for national publications Mean time to publication = 13 months for oral presentations Mean time to publication = 11 months for poster presentations Factors related to proportion of abstracts published included 9/16 abstracts presented orally versus 75/227 abstracts presented as poster published 47/135 abstracts from university hospitals versus 26/86 abstracts from education and research hospitals versus 4/10 abstracts from private institutions versus 7/9 abstracts from state hospitals published
Notes	Gynecology/obstetrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Yes	The meeting had 60 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation, university affiliation, and subspecialty with publication using stratified analysis and Fisher's Exact, Chi2 , Mann‐Whitney U, or t tests

Ersoy 2015b.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar in August 2014 Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by One author Hypothesis Study design
Data	Included 161 abstracts presented at the 2008 to 2010 National Gynecology and Obstetrics meetings Included all abstracts on reproductive endocrinology and infertility
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by type of presentation Oral versus poster presentation Clinical research versus experimental research Case report design versus experimental design versus other clinical design University versus training and research hospital versus public hospital versus private institution
Outcomes	46 of 161 abstracts published 10/47 abstracts presented at the 2008 meeting, 11/52 at the 2009 meeting, and 25/62 at the 2010 meeting published Proportion of abstracts published by time Mean time to publication = 17.5 months (SD = 12.5), overall Mean time to publication = 20.9 months (SD = 11.1) for oral presentations Mean time to publication = 16.2 months (SD = 12.7) for poster presentations Factors related to proportion of abstracts published included 13/19 abstracts presented orally versus 33/142 abstracts presented as poster published 35/145 abstracts describing clinical research versus 11/16 abstracts describing 'experimental' research published 4/39 abstracts describing case reports versus 11/16 abstracts with 'experimental' design versus 31/106 abstracts with other clinical study design published 32/96 abstracts from a university versus 14/49 abstracts from training and research hospitals versus 0/2 abstracts from public hospitals versus 0/14 abstracts from private institutions published
Notes	Gynecology/obstetrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts on a specific topic, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, type science, study design, university affiliation with publication using stratified analysis and Chi2 or t tests.

Evans 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and International Pharmaceutical Abstracts; dates of search not reported Search completed by investigator Searched by all authors, keywords, and title Matching criteria not reported
Data	Included 1496 abstracts presented at the 2004 to 2007 American College of Clinical Pharmacy (ACCP) Spring and Fall national meetings, and regional meetings: Alcalde, Southwest Leadership, Eastern States, Southeastern Pharmacy Residency, Great Lakes Pharmacy Resident, Midwest Pharmacy Resident, and Western States Included all abstracts of pharmacy resident projects from the ACCP meetings, and a random sample from each of the regional conferences
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication
Outcomes	109 of 1496 abstracts published 2/41 abstracts presented at the 2004 Western, 2/56 at the 2004 Great lakes, 3/70 at the 2004 Southeastern, 3/68 at the 2004 Eastern, 3/20 at the 2004 Alcalde, 2/14 at the 2004 Fall ACCP, 7/92 at the 2005 Western, 3/32 at the 2005 Midwest, 7/54 at the 2005 Great Lakes, 3/72 at the 2005 Southeastern, 5/70 at the 2005 Eastern, 1/22 at the 2005 Alcalde, 5/22 at the 2005 Spring ACCP, 4/44 at the 2005 Fall ACCP, 6/96 at the 2006 Western, 1/36 at the 2006 Midwest, 2/60 at the 2006 Great Lakes, 5/70 at the 2006 Southeastern, 6/70 at the 2006 Eastern, 2/26 at the 2006 Alcalde, 3/22 at the 2006 Spring ACCP, 2/12 at the 2006 Fall ACCP, 4/102 at the 2007 Western, 2/34 at the 2007 Midwest, 4/66 at the 2007 Great Lakes, 6/80 at the 2007 Southeastern, 7/70 at the 2007 Eastern, 4/28 at the 2007 Alcalde, 5/23 at the 2007 Spring ACCP, and 0/24 at the 2007 Fall ACCP meeting published Proportion of abstracts published by time Mean time to publication = 24 months (SD = 15) Median time to publication = 24 months (range = 1 to 94 months) No factors other than meeting and meeting year related to proportion of abstracts published reported
Notes	Pharmacology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts or a random sample of abstracts of a specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting and meeting year with publication using stratified analysis and Chi2 tests.

Evers 2000.

Methods	Identification of subsequent full‐length publications Searched electronic database Cochrane Central Register of Controlled Trials, Embase, and MEDLINE from year of meeting through February 2000 Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by All authors Contents Handsearched two 'major' journals
Data	Included 151 abstracts presented at the 1992 to 1997 European Society of Human Reproduction and Embryology meetings Included all RCT abstracts
Comparisons	Proportion of abstracts published Median time to publication Survival analysis of publication rate 'Positive' versus not 'positive' Sample size equal to or above the median versus sample size below the median Oral versus poster presentation English versus non‐English language Subspecialty
Outcomes	79 of 151 abstracts published Proportion of abstracts published by time Median time to publication = 32.5 months (range = 0 to 79 months) Survival analysis of proportion published at 60 months = 52.3% Factors related to proportion of abstracts published included 41/69 'positive' (defined as significant results) versus 33/82 not 'positive' abstract results published 46/76 abstracts with sample size equal to or above the median versus 40/75 abstracts with sample size below the median published 50/72 abstracts presented orally versus 33/79 abstracts presented as posters published 22/40 abstracts with English language versus 62/111 abstracts with non‐English language published
Notes	Human reproduction and development Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing RCTs.
Search for publications?	Yes	Searched 3 databases and handsearched 2 journals.
Follow‐up time?	Yes	All meetings before 1997 had at least 48 months follow‐up. The meeting in 1997 only had 36 months follow‐up.
Matching?	Yes	Matched by two different characteristics.
Adjustment for confounding?	No	Examined association of positive results, sample size, type presentation, origin in English language country, and subspecialty with publication using Chi2 tests or log‐rank tests

Facione 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to February 2007 Search completed by investigator Searched by all authors Matched abstract to full‐length publication by Title
Data	Included 239 abstracts presented at the 2003 to 2006 Société Interdisciplinaire Francophone d’Urodynamique et de Pelvipérinéologie meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication by meeting
Outcomes	60 of 239 abstracts published Proportion of abstracts published by time Mean time to publication = 21.6 months for 2003 Société interdisciplinaire francophone d’urodynamique et de pelvipérinéologie meeting Mean time to publication = 19.2 months for 2004 Société interdisciplinaire francophone d’urodynamique et de pelvipérinéologie meeting Mean time to publication = 13.5 months for 2005 Société interdisciplinaire francophone d’urodynamique et de pelvipérinéologie meeting Mean time to publication = 5.6 months for 2006 Société interdisciplinaire francophone d’urodynamique et de pelvipérinéologie meeting No factors related to proportion of abstracts published reported
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	Only the meeting in 2003 had 48 months follow‐up. All other meetings had less than 48 months follow‐up.
Matching?	No	Matched by only 1 criterion.

Fede 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, LILACS; dates of search not reported Search completed by investigator Searched by all authors and title Matching criteria not reported Searched online curriculum vitae on Lattes Platform of teachers who co‐authored papers
Data	Included 408 abstracts submitted to the to 2002 to 2007 Undergraduate Medical Congress of ABC (COMUABC) Foundation School of Medicine meetings Included all abstracts submitted as oral presentations
Comparisons	Proportion of abstracts published, overall and by meeting Clinical research versus basic science research versus monographs Medical subspecialty 'Award winning' versus not 'award winning' Observational versus case studies Subspecialty
Outcomes	138 of 408 abstracts published 33/81 abstracts presented at the 2002 meeting, 27/79 at the 2003 meeting, 14/60 at the 2004 meeting, 19/56 at the 2005 meeting, 20/63 at the 2006 meeting, and 25/69 at the 2007 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 118/323 abstracts describing clinical research versus 11/27 abstracts describing basic science research versus 9/58 abstracts of monographs published 42/86 abstracts with observational designs versus 34/124 abstracts with case report design published 34/71 abstracts that were 'award winning' versus 104/337 abstracts that were not 'award winning' published
Notes	General medicine/primary care Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts submitted as oral presentations.
Search for publications?	Yes	Searched 2 databases and screened curriculum vitae of authors.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting year, type science, study design, award winning status, and subspecialty with publication using multivariable logistic regression analysis.

Feldman 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to September 2014 Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 1152 abstracts submitted to the 2005 to 2012 Advancing Transfusion and Cellular Therapies Worldwide and American Society for Apheresis meetings Included all abstracts related to apheresis
Comparisons	Proportion of abstracts published Median time to publication Academic versus non‐academic affiliation Non‐US location of first author versus US location Aphersis type
Outcomes	196 of 1152 abstracts published Proportion of abstracts published by time Median time to publication = 17 months (range = 0 to 108 months) Factors related to proportion of abstracts published included 166/749 abstracts originating from an academic center versus 33/403 abstracts not originating from an academic center published 125/801 abstracts originating in the US versus 74/351 abstracts not originating in the US published
Notes	Hematology ‐ apheresis Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific topic.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings except the 2011 to 2012 meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting, academic affiliation, US origin and subspecialty with publication using Cox proportional hazard analyses.

Fernandes 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to June 2003 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 1133 abstracts presented at the XXIV Congresso Brasileiro de Cirurgia meeting; year of meeting not reported Included all abstracts
Comparisons	Proportion of abstracts published Oral versus poster presentation
Outcomes	29 of 1133 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 9/442 abstracts presented orally versus 17/614 abstracts presented as posters published
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Year of meeting not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis

Fernandes 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database EMBASE, MEDLINE, Google/Google Scholar, SciELO, Sport Discus, and LILACS for 6 years (until June 2007) Person completing the search not reported Searched by first and last author, and title Matched abstract to full‐length publication by All authors Sample size Searched Lattes platform and the website of the Conselho Nacional de Desenvolvimento Científico e Tecnológico
Data	Included 263 abstracts presented at 2001 American College of Sports Medicine, Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul, and American College of Sports Medicine (Brazilian authors only) meetings Included a random sample of 100 abstracts from the American College of Sports Medicine plus all abstracts by Brazilian authors and a random sample of 100 abstracts from the Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication by meeting Oral versus poster presentation
Outcomes	58 of 263 abstracts published 32/100 abstracts from the American College of Sports Medicine meeting, 19/63 presented by Brazilian authors at the American College of Sports Medicine meeting, and 7/100 at the Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul meeting published Proportion of abstracts published by time Mean time to publication = 18.5 months (SD = 2.3) for the American College of Sports Medicine meeting Mean time to publication = 20.2 months (SD = 20) for the abstracts from Brazilian authors presented at the American College of Sports Medicine meeting Mean time to publication = 23.3 months (SD = 5.5) for the Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul meeting Median time to publication = 15 months (range = ‐1 to 55 months) for the American College of Sports Medicine meeting Median time to publication = 12 months (range = ‐1 to 64 months) for the abstracts from Brazilian authors presented at the American College of Sports Medicine meeting Median time to publication = 22 months (range = 6 to 48 months) for the Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul meeting Factors related to proportion of abstracts published included 26/109 abstracts presented orally versus 32/154 abstracts presented as poster published
Notes	Other clinical specialties ‐ sports medicine Funded by the Conselho Nacional de Desenvolvimento Científico e Tecnológico
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all or a random sample of abstracts.
Search for publications?	Yes	Searched 6 databases, and searched Lattes platform and the website of the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq).
Follow‐up time?	Yes	All meetings had at least 72 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting by type presentation with publication using stratified analysis and Chi2 or ANOVA tests.

Fernandez 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, Embase, and MEDLINE to April 2010 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Title
Data	Included 279 abstracts presented at 2003 to 2007 Australian Cardiac Rehabilitation Association meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication
Outcomes	17 of 279 abstracts published 2/49 abstracts presented at the 2003 meeting, 2/56 at the 2004 meeting, 1/60 at the 2005 meeting, 7/72 at the 2006 meeting, 5/42 at the 2007 meeting published Proportion of abstracts published by time Mean time to publication = 19 months (SD = 21.6; range = ‐4 to 78 months) No factors other than meeting year related to proportion of abstracts published reported
Notes	Physical and rehabilitation medicine ‐ cardiac rehabilitation Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings before 2007 had at least 48 months follow‐up. The meeting in 2007 had 36 months follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis and Chi2 tests

Fesperman 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from July 6 to October 24, 2006 Search completed by investigator Searched by first, second, and senior author, and keywords Matched abstract to full‐length publication by All authors Title Results Hypothesis Study design
Data	Included 1195 abstracts presented at the 1996 to 2005 Southeastern Section of the American Urological Association meetings Included all abstracts
Comparisons	Proportion of abstracts published Female versus male first author Female versus male senior author Study origin Study design Oral versus poster presentation Sample size Reporting of statistical analyses
Outcomes	400 of 1195 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 21/74 abstracts by female first authors versus 379/1121 abstracts by male first authors published 23/64 abstracts by female senior authors versus 377/1131 abstracts by male senior authors published "predictors of publication identified in this study...include multi‐institutional origin, the reporting of statistical hypothesis, and ...private institutions"
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2003 had at least 48 months follow‐up. The meetings in 2003, 2004, and 2005 only had 36, 24, and 12 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	Yes	Examined association of sample size, study design, type of presentation, multi‐center status, institutional status, gender or first or senior author, report of statistical hypothesis, and topic using a multivariable proportional hazards model.

Forlin 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed, Google/Google Scholar, BIREME, LILACS, SciELO, and online RBO until 2012 Search completed by the investigators Searched by all authors and title Matching criteria not reported
Data	Included 267 abstracts presented at the 2004 Brazilian Congress of Orthopedics and Traumatology meeting Included all abstracts of free papers except 3 duplicates
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Level of evidence I versus II versus III versus IV versus V 'Anatomical' versus 'experimental' study Subspecialty
Outcomes	58 of 267 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 96 months showed proportion published = 21.5% (58/270 abstracts) Factors related to proportion of abstracts published included 0/0 abstracts with level I evidence versus 16/27 abstracts with level II evidence versus 4/7 abstracts with level III evidence versus 30/207 abstracts with level IV evidence versus 0/8 abstracts with level V evidence published 8/18 abstracts describing an 'anatomical' study versus 20/34 abstracts describing an 'experimental' study versus 30/215 undefined study published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific subgroup.
Search for publications?	Yes	Searched 6 databases.
Follow‐up time?	Yes	The meeting had more than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of level of evidence and subspecialty with publication using stratified analysis.

Fosbol 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database Web of Science for 1 month prior and 2 years following each conference, up to 5 years for 2006 meeting Search completed by computerized algorithm that was tested using manual search by author Searched by first, second, third, and last author, and keywords Matching criteria not reported
Data	Included 27,208 abstracts presented at the 2006 to 2008 American Heart Association, American College of Cardiology, and European Society of Cardiology meetings Included all abstracts except duplicate entries
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication, overall and by meeting Cumulative proportion of abstracts published by meeting Clinical research versus basic science research versus population science US origin versus German versus Japanese versus Italian versus UK origin versus rest of the world origin
Outcomes	8,335 of 27,208 abstracts published 2931/10838 abstracts presented at the 2006 to 2008 European Society of Cardiology meetings, 1481/5005 at the 2006 to 2008 American College of Cardiology meetings, and 3923/11365 at the American Heart Association meetings published Proportion of abstracts published by time Median time to publication = 12 months (IQR = 7 to 19 months), overall Median time to publication = 13 months (IQR = 7 to 19 months) for 2006 to 2008 American Heart Association meetings Median time to publication = 13 months (IQR = 6 to 18 months) for 2006 to 2008 American College of Cardiology meetings Median time to publication = 12 months (IQR = 6 to 18 months) for 2006 to 2008 European Society of Cardiology meetings Cumulative proportion of abstracts published at 24 months showed proportion published = 32% (3637/11365 abstracts) for 2006 to 2008 American Heart Association meetings Cumulative proportion of abstracts published at 24 months showed proportion published = 26% (1301/5005 abstracts) for 2006 to 2008 American College of Cardiology meetings Cumulative proportion of abstracts published at 24 months showed proportion published = 24% (2601/10838 abstracts) for 2006 to 2008 European Society of Cardiology meetings Factors related to proportion of abstracts published included 5,099/17,168 abstracts describing clinical research versus 2,140/6,203 abstracts describing basic science research versus 1,105/3,836 abstracts describing population science published 3,400/10,366 abstracts originating from USA versus 986/3,102 abstracts originating from Germany versus 848/2,884 abstracts originating from Japan versus 592/2,013 abstracts originating from Italy versus 581/1,850 abstracts originating from UK versus 1,928/6,993 abstracts with rest of the world origin published
Notes	Cardiology Funded through a foundation and through an award from David and Stevie Spina
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The majority of meetings only had 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting by year, type science, country of origin, prospective status, abstract quality, and subspecialty with publication using multivariable logistic regression analysis.

Freeman 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to 2010 Person completing the search not reported Searched by all authors, cancer types, investigational agent, dose range, and schedule Matched abstract to full‐length publication by All Authors Cancer types Investigational agent Dose range Schedule Contacted abstract authors directly
Data	Included 90 abstracts presented at the 2004 to 2005 American Association for Cancer Research and 1995 to 2005 American Society of Clinical Oncology meetings Included all abstracts describing acquisition of biopsies for research purposes
Comparisons	Proportion of abstracts published Mean time to publication by meeting
Outcomes	70 of 90 abstracts published 22/26 abstracts presented at the 2004 to 2005 American Association for Cancer Research and 48/64 at the 1995 to 2005 American Society of Clinical Oncology meeting published Proportion of abstracts published by time Mean time to publication = 22.8 months for the American Association for Cancer Research meetings Mean time to publication = 34.8 months for the American Society of Clinical Oncology meetings No factors other than meeting year related to proportion of abstracts published reported
Notes	Oncology Government funding received and costs of publication of article defrayed in part by the payment of page charges
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific topic.
Search for publications?	Yes	Searched 1 database and contacted abstract authors directly.
Follow‐up time?	Yes	All meetings had at least 60 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of meeting with publication using stratified analysis.

Frost 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to November 2014 Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 764 abstracts presented at the 2009 to 2011 Scoliosis Research Society (SRS) and International Meeting of Advanced Spinal Techniques meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Award‐winning versus not
Outcomes	339 of 764 abstracts published 59/150 abstracts presented at the 2009 International Meeting of Advanced Spinal Techniques, 57/114 at the 2009 SRS meeting, 50/134 at the 2010 International Meeting of Advanced Spinal Techniques meeting, 64/115 at the SRS meeting, 65/135 at the 2011 International Meeting of Advanced Spinal Techniques, and 44/116 at the 2011 SRS meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 49/77 award‐winning abstracts versus 290/684 non award‐winning abstracts published
Notes	Orthopedics/orthopedic surgery ‐ spine Reported receiving no direct funding for the study, but fellowship support from AO Spine
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meetings in 2011 only had 36 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting and award winning status with publication using stratified analysis and t tests.

Galang 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for Quote: "maximum of five years" Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by First or last author Contents
Data	Included 370 abstracts presented at the 2002 and 2003 American Dental Education Association Annual Session and Exhibitions meetings Included all abstracts presented as posters except one abstract not having to do with dental education
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' 'Experimental' versus 'observational' study design Used statistics versus no use of statistics 'Analytic' versus 'descriptive' report Multi‐centered versus single center North American versus elsewhere Funding reported versus without funding reported
Outcomes	71 of 370 abstracts published Proportion of abstracts published by time Mean time to publication = 16 months Median time to publication = 10 months Cumulative proportion of abstracts published at 60 months showed proportion published = 19.2% (71/370 abstracts) Factors related to proportion of abstracts published included 40/234 'positive' (defined as significant results) versus 31/136 not 'positive' abstract results published 19/84 abstracts with 'experimental' study design versus 52/286 abstracts with 'observational' study design published 60/285 abstracts that 'used statistics' versus 11/85 abstracts that did not 'use statistics' published 29/83 abstracts with 'analytical' report versus 31/202 abstracts with 'descriptive' report published 11/29 abstracts with multiple centers versus 60/341 abstracts with a single center published 71/363 abstracts originating from North America versus 1/6 abstracts from elsewhere published 8/28 abstracts with funding reported versus 63/342 abstracts without funding reported published
Notes	Oral health ‐ dental education Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all posters with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meetings had a maximum of 5 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, type science, study design, multi‐center status, North American origin, use of statistics, and funding status with publication using multivariable logistic regression analysis.

Gandhi 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors and content Matching criteria not reported
Data	Included 265 abstracts presented at the 1990 to 1999 Society for Gynecologic Investigation meetings Included all plenary abstracts and a random sample of concurrent session and poster abstracts
Comparisons	Proportion of abstracts published Oral versus poster presentation Basic science versus clinical research Interventional versus observational design 'Prospective' versus 'retrospective' design
Outcomes	190 of 265 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 120/159 abstracts presented orally versus 70/106 abstracts presented as posters published Odds ratio predicting publication of abstracts describing basic science versus clinical research = 1.14 (95% CI = 0.64, 2.06) Odds ratio predicting publication of abstracts with interventional design versus observational design = 0.75 (95% CI = 0.41, 1.37) Odds ratio predicting publication of abstracts with 'prospective' study design versus abstracts with 'retrospective' study design = 0.33 (95% CI = 0.04, 2.91) Odds ratio predicting publication of abstracts without a university affiliation versus abstracts with a university affiliation = 0.2 (95% CI = 0.02, 2.39)
Notes	Gynecology/obstetrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all plenary abstracts and random samples of other presentation formats, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of type of presentation, type science, study design, prospective status, abstract quality, and university affiliation with publication using stratified analysis, Chi2 tests and unspecified multivariable analysis.

Garcia‐Covarrubias 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE between December 2001 and January 2002 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 503 abstracts presented at 1996 to 1998 Undersea and Hyperbaric Medical Society meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	86 of 503 abstracts published 26/145 abstracts presented at the 1996 meeting, 27/182 at the 1997 meeting, and 33/176 at the 1998 meeting published Proportion of abstracts published by time not reported No factors other than meeting year related to proportion of abstracts published reported
Notes	Other clinical specialties‐ undersea and hyperbaric medicine Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Garcia‐Muret 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Dermabase, and Indice Medico Espanol to first quarter 2006 Person completing the search not reported Searched by first, and last author, and title Matching criteria not reported
Data	Included 1471 abstracts presented at 2000 to 2003 Spanish National Dermatology and Venerology Congresses Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation Peer‐reviewed versus not peer‐reviewed versus case report versus poster presentation
Outcomes	200 of 1471 abstracts published 48/372 abstracts presented at the 2000 meeting, 55/379 at the 2001 meeting, 56/356 at the 2002 meeting, and 41/364 at the 2003 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 64/381 abstracts presented orally versus 133/1090 abstracts presented as posters published 9/28 peer‐reviewed abstracts versus 15/93 not peer‐reviewed abstracts versus 44/260 abstracts of case report presentations versus 133/1090 abstracts of poster presentations published
Notes	Dermatology and venereology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings before 2003 had at least 48 months follow‐up. The meeting in 2003 only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, and study design with publication using stratified analysis.

Gaundong Mbethe 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by first, second, and last author, and keywords Matching criteria not reported
Data	Included 303 abstracts presented at the 2000 French National Society of Internal Medicine meeting Included all abstracts
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Oral versus poster presentation Clinical research versus basic science research versus epidemiological research Systemic disease versus infectious disease versus hematology versus other disciplines
Outcomes	82 of 303 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 27.1% (82/303 abstracts) Factors related to proportion of abstracts published included 39/108 abstracts presented orally versus 43/195 abstracts presented as posters published 68/272 abstracts describing clinical research versus 5/6 abstracts describing basic science research versus 9/25 abstracts describing epidemiological research published 33/100 abstracts describing systemic disease versus 23/81 abstracts describing infectious disease versus 6/25 abstracts describing hematology versus 18/97 abstracts describing other disciplines published
Notes	General medicine/primary care Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type of presentation, type science, study design, multi‐center status, and subspecialty with publication using stratified analysis.

Gavazza 1996.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE Plus from January 1990 through January 1995 Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by All authors Title Contents Methodology Results
Data	Included 376 abstracts presented at the 1990 to 1992 American Society in Surgery of the Hand meetings Included all 397 abstracts initially, but deleted 21 abstracts from the study population in which the authors only found partial agreement between the abstract and a subsequent full publication
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Clinical research versus basic science research Presentation by residents or fellows versus others
Outcomes	165 of 376 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 36 months showed proportion published = 52.7% (198/376 abstracts) Factors related to proportion of abstracts published included 105/254 abstracts describing clinical research versus 18/32 abstracts describing basic science research published 42/90 abstracts presented by residents or fellows versus 123/286 abstracts presented by others published
Notes	Surgery ‐ hand Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 1992 had at least 48 months follow‐up. The meeting in 1992 only had 36 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of type science, and author status with publication using stratified and log rank analysis.

Gilbert 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Search completed by investigator Searched by all authors and keywords Matching criteria not reported
Data	Included 176 abstracts presented at the 1991 to 2000 Society for Maternal‐Fetal Medicine meetings Included oral presentations from plenary sessions and an equivalent number of randomly selected abstracts from concurrent sessions for each year
Comparisons	Proportion of abstracts published Plenary versus concurrent session 'Prospective' versus 'retrospective' design 'Original' versus 'confirmatory' 'Interventional' versus 'observational' US versus non‐US origin Basic science versus clinical research Academic versus non‐academic affiliation
Outcomes	120 of 176 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 79/88 abstracts presented at plenary sessions versus 49/88 abstracts presented at concurrent sessions published Multivariate analysis showed the following variables associated with publication: 'Prospective' versus 'retrospective' (OR = 3.09; 95% CI = 1.54 to 6.19) 'Plenary' versus 'concurrent' (OR = 2.57; 95% CI = 1.31to 5.05) 'Original' versus 'confirmatory' (OR = 1.97; 95% CI = 1.05 to 3.69) 'Interventional' versus 'observational' (OR = 1.96; 95% CI = 1.03 to 3.72) US versus elsewhere (OR = 0.83; 95% CI = 0.35 to 1.95) Basic science versus clinical (OR 0.63; 95% CI = 0.29 to 1.38) Non‐university versus university affiliation (OR 0.5, 95% CI 0.16 to 1.60)
Notes	Gynecology/obstetrics ‐ maternal‐fetal medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included oral presentations from plenary sessions and a random sample of concurrent sessions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of prospective status, type oral presentation, type science, original status, study design, US origin, and US region with publication using multivariable logistic regression analysis.

Ginzburg 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to October 2013 Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by One author One conclusion
Data	Included 140 abstracts presented at the 2008 to 2012 Society of Urodynamics, Female Pelvic Medicine, and Urogenital Reconstruction meetings Included all abstracts accepted as oral presentation
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	80 of 140 abstracts published Proportion of abstracts published by time Mean time to publication = 17.4 months (SD = 12) No factors related to proportion of abstracts published reported
Notes	Gynecology/obstetrics ‐ urogynecology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings except the 2008 and 2009 meetings had less than 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.

Glick 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE, and PubMed to November 2004 Person completing the search not reported Searched by first, second, and last author and keywords Matched abstract to full‐length publication by Objectives Methodology Results
Data	Included 1147 abstracts presented at the 2000 American Transplant Congress Included all abstracts except those withdrawn and those published prior to the meeting
Comparisons	Proportion of abstracts published Mean/median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation Clinical research versus basic science research RCT design versus observational versus case series versus other study design Multi‐centered versus single center US versus non‐US origin versus origin not specified Industry sponsorship versus none
Outcomes	607 of 1147 abstracts published Proportion of abstracts published by time Mean time to publication = 15.8 months (SD =10.6) Median time to publication = 12 months Cumulative proportion of abstracts published at 54 months showed proportion published = 53.0% (608/1147 abstracts) Factors related to proportion of abstracts published included 208/397 'positive' (defined as significant results) versus 95/234 not 'positive' abstract results published 308/547 abstracts presented orally versus 299/600 abstracts presented as posters published 306/653 abstracts describing clinical research versus 301/494 abstracts describing basic science research published 39/64 abstracts with RCT design versus 333/647 abstracts with observational design versus 140/284 abstracts with case series design versus 95/152 abstracts with other study design published 220/371 abstracts with multiple centers versus 387/776 abstracts with a single center published 398/804 abstracts originating in the US versus 197/314 abstracts not originating in the US versus 12/29 abstracts with origin not specified published Adjusted multivariate analysis showed the following variables associated with publication: Industry sponsorship versus none (OR = 1.78; 95% CI = 1.04 to 3.06) Basic science versus clinical (OR = 1.68; 95% CI = 1.32 to 2.14) 'Non‐American center' versus 'American center' (OR = 1.67; 95% CI = 1.28 to 2.20) Oral versus poster presentation (OR = 1.36; 95% CI = 1.07 to 1.73)
Notes	Surgery ‐ organ transplantation Funding by the Alberta Heritage Foundation for Medical Research
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, type science, study design, multi‐center status, US origin, organ type, and industry sponsorship with publication using multivariable logistic regression analysis.

Goldman 1980.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLARS from November 1975 through June 1979 Search completed by investigator Searched by first author Matched abstract to full‐length publication by All authors Sample size Title Contents Results
Data	Included 276 abstracts submitted to the 1976 American Federation for Clinical Research, American Society for Clinical Investigation, Association for American Physicians, and American College of Cardiology meetings and the 1975 American Heart Association Included 69 abstracts representing all cardiology presentations and posters and 69 randomly selected cardiology abstracts from those not presented at the 1976 meetings of American Federation for Clinical Research, American Society for Clinical Investigation, and Association of American Physicians; 69 randomly selected abstracts from the 1976 American College of Cardiology scientific sessions; and 69 randomly selected abstracts from the 1975 American Heart Association :
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published Accepted versus rejected for conference presentation Affiliated with US medical school versus other US. and non‐US institutions Oral versus poster presentation
Outcomes	137 of 276 abstracts published 38/69 abstracts presented at the 1975 American Heart Association meeting, 34/69 at the 1976 American College of Cardiology meeting, and 65/138 at the American Federation for Clinical Research, American Society for Clinical Investigation, Association for American Physicians meetings published Proportion of abstracts published by time Median time to publication = 14 months Cumulative proportion of abstracts published at 36 months showed proportion published = 48.9% (135/276 abstracts) Factors related to proportion of abstracts published included 113/207 abstracts accepted for presentation versus 24/69 rejected abstracts published 112/205 abstracts from US medical schools versus 25/71 abstracts from other US and non‐US institutions published Stated that 'oral presentations and posters were equally likely to result in articles'
Notes	Cardiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, and a random sample of accepted and rejected abstracts on other topics, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings in 1976 only had 36 months follow‐up. The meeting in 1975 had 48 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of meeting, acceptance for presentation, institutional affiliation, and type presentation with publication using stratified analysis and Chi2, Kruskal Wallis ANOVA, or Mann‐Whitney U tests

Goldman 1982.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLARS to June 1979 Search completed by investigator Search criteria not reported Matched abstract to full‐length publication by All authors Sample size Title Contents Results
Data	Included 303 abstracts presented at the 1976 American Federation for Clinical Research, American Society for Clinical Investigation, and Association of American Physicians meetings, and the 1975 American Society of Hematology and American Society of Nephrology meetings Included 101 abstracts (48 hematology and 53 nephrology) representing all presentations and posters at the 1976 meetings of the American Federation for Clinical Research, American Society for Clinical Investigation, and Association of American Physicians meetings; 101 (48 hematology and 53 nephrology) abstracts randomly selected from abstracts not selected for presentation at these meetings; 48 abstracts randomly selected from the 1975 American Society of Hematology meeting and 53 abstracts from the 1975 American Society of Nephrology meeting
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published by specialty Accepted versus rejected for conference presentation Clinical research versus basic science research Affiliated with US medical school versus other institutions
Outcomes	171 of 303 abstracts published Proportion of abstracts published by time Median time to publication = 15 months Cumulative proportion of abstracts published at 36 months showed proportion published = 58.3% (84/144 abstracts) for abstracts related to hematology Cumulative proportion of abstracts published at 36 months showed proportion published = 50.9% (81/159 abstracts) for abstracts related to nephrology Factors related to proportion of abstracts published included 127/202 abstracts accepted for presentation versus 44/101 rejected abstracts published 116/254 abstracts describing clinical research versus 192/325 abstracts describing basic science research published (includes abstracts reported in Goldman 1980) 245/432 abstracts from US medical schools versus 63/146 abstracts from other US and non‐US institutions published (includes abstracts reported in Goldman 1980)
Notes	Hematology, nephrology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all accepted abstracts and a random sample of rejected abstracts, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings in 1976 only had 36 months follow‐up. The meeting in 1975 had 48 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting, acceptance for presentation, type science, and institutional affiliation with publication using multivariable logistic regression analysis.

Gorman 1990.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLARS from October 1984 to April 1989 Search completed by investigator Searched by all authors Matched abstract to full‐length publication by Title Contents
Data	Included 269 abstracts presented at the 1984 American Association of Poison Control Centers and 1986 American Association of Poison Control Centers, American Academy of Clinical Toxicologists, American Board of Medical Toxicologists, Canadian Association of Poison Control Centers meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting year Mean and median time to publication by meeting date (1984 and 1986)
Outcomes	134 of 269 abstracts published 40/104 abstracts presented at the 1984 meeting, and 56/165 at the 1986 meetings published Proportion of abstracts published by time Mean time to publication = 19 months for the 1984 American Association of Poison Control Centers meeting Mean time to publication = 12 months for the 1986 American Association of Poison Control Centers, American Academy of Clinical Toxicologists, American Board of Medical Toxicologists, Canadian Association of Poison Control Centers meetings Median time to publication = 12 months for the 1984 American Association of Poison Control Centers meeting Median time to publication = 9 months for the 1986 American Association of Poison Control Centers, American Academy of Clinical Toxicologists, American Board of Medical Toxicologists, Canadian Association of Poison Control Centers meetings No factors other than meeting year related to proportion of abstracts published reported
Notes	Other non‐clinical specialties ‐ toxicology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The 1984 meeting had 48 months follow‐up. All other meetings only had 36 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis and Chi2 tests.

Gourtaud 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to March 2007 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 443 abstracts presented at the 2000 and 2001 French Association of Urology meetings Included all abstracts except video presentations
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published Oral versus poster presentation 'Prospective' versus 'retrospective' design
Outcomes	153 of 443 abstracts published 86/234 abstracts presented at the 2000 meeting, and 67/209 at the 2001 meeting published Proportion of abstracts published by time Mean time to publication = 16.9 months (maximum = 65 months) Cumulative proportion of abstracts published at 24 months showed proportion published = 27.1% (120/443 abstracts) Factors related to proportion of abstracts published included 66/187 abstracts presented orally versus 87/256 abstracts presented as posters published 44.1% of abstracts with 'prospective' study design versus 24.8% of abstracts with 'retrospective' study design published
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type presentation and prospective status with publication using stratified analysis.

Greenberg 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2007 Search completed by investigator Searched by first, second, third, and fourth authors, and keywords Matched abstract to full‐length publication by All authors Title Contents
Data	Included 239 abstracts presented at the 2003 Society for Medical Decision Making meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Oral versus poster presentation Study origin
Outcomes	64 of 239 abstracts published Proportion of abstracts published by time Mean time to publication = 20.5 months (95% CI = 17.6 to 23.3 months) Median time to publication = 19 months (range = 0 to 47 months) Cumulative proportion of abstracts published at 50 months showed proportion published = 26.8% (64/239 abstracts) Factors related to proportion of abstracts published included 35/91 abstracts presented orally versus 29/148 abstracts presented as posters published 'could not find any association between study origin (North America v. Europe) and successful publication.'
Notes	Medical decision making/health policy Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by three different characteristics.
Adjustment for confounding?	No	Examined association of type presentation and country of origin with publication stratified analysis and Chi2 tests or log rank analyses.

Gregory 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed through 2010 to provide a minimum of 36 months follow‐up Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by One author Contents
Data	Included 888 abstracts presented at the 2003 to 2007 Canadian Society of Plastic Surgeons, American Society of Plastic Surgeons, and American Association of Plastic Surgeons meetings Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and for each meeting
Outcomes	400 of 888 abstracts published 88/173 abstracts presented at the 2003 to 2007 American Association of Plastic Surgeons meetings, 151/262 presented at the 2003 to 2007 American Society of Plastic Surgeons meetings, and 161/453 at the Canadian Society of Plastic Surgeons meetings published Proportion of abstracts published by time Mean time to publication = 22 months (range = 1 to 85.9 months) for all meetings Mean time to publication = 24 months (range = 2 to 79.0 months) for Canadian Society of Plastic Surgeons meeting Mean time to publication = 19 months (range = 1 to 54.3 months) for American Society of Plastic Surgeons meeting Mean time to publication = 22 months (range = 1 to 85.9 months) for American Association of Plastic Surgeons meeting No factors other than meeting year related to proportion of abstracts published reported
Notes	Surgery ‐ plastic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted as oral presentations.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2007 had at least 48 months follow‐up. The meeting in 2007 only had 36 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Grzeskowiak 2014.

Methods	Identification of subsequent full‐length publications: Searched electronic database Embase, PubMed, and Informit in September 2013 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 96 abstracts presented at the 2007 to 2011 Society of Hospital Pharmacists of Australia meetings Included all abstracts from specialty areas of pediatrics and perinatal health, except for invited presentations
Comparisons	Proportion of abstracts published Oral versus poster presentation Case report design versus clinical audit versus practice report versus experimental/observational design versus other Single versus multiple authors Subspecialty
Outcomes	10 of 96 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included: 5/30 abstracts presented orally versus 5/66 abstracts presented as poster published 0/18 abstracts describing case report versus 1/39 abstracts describing clinical audits versus 0/18 abstracts describing practice reports versus 4/8 abstracts describing 'experimental/observational' design versus 5/14 abstracts with other design published 0/17 abstracts with a single author versus 10/79 abstracts with multiple authors published
Notes	Pharmacology ‐ pediatrics and perinatal health Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific topic.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings except the 2010 and 2011 meeting had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation, study design, number of authors, Australian state, and subspecialty with publication using stratified analysis and Fisher's Exact test.

Guryel 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed, and PreMEDLINE; dates of search not reported Search completed by investigator Searched by first, second, and third authors Matching criteria not reported
Data	Included 415 abstracts presented at the 1997 and 1998 British Orthopaedics Association, British Association for Surgery of the Knee, British Orthopedic Foot Surgical Society, and British Elbow and Shoulder Society meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication overall and by meetings Cumulative proportion of abstracts published Clinical research versus basic science research RCT design versus observational versus case report design Subspecialty
Outcomes	137 of 415 abstracts published 31/115 abstracts presented at the 1997 to 1998 British Association for Surgery of the Knee, British Orthopedic Foot Surgical Society, and British Elbow and Shoulder Society meetings, and 196/300 at the British Orthopaedics Association meetings published Proportion of abstracts published by time Mean time to publication = 15.6 months (SD = 11.1) Mean time to publication = 16.2 months (SD = 6.1) for British Orthopaedics Association meetings Mean time to publication = 15.3  months (SD = 9.3) for British Association for Surgery of the Knee, British Orthopedic Foot Surgical Society, and British Elbow and Shoulder Society meetings Median time to publication = 14 months (range = ‐7 to 56 months) Median time to publication = 14 months (range = ‐7 to 56 months) for British Orthopaedics Association meetings Median time to publication = 14 months (range = 1 to 48 months) for British Association for Surgery of the Knee, British Orthopedic Foot Surgical Society, and British Elbow and Shoulder Society meetings Cumulative proportion of abstracts published at 48 months showed proportion published = 33.5% (139/415 abstracts) Factors related to proportion of abstracts published included 115/345 abstracts describing clinical research versus 22/70 abstracts describing basic science research published 15/28 abstracts with RCT design versus 99/302 abstracts with observational design versus 1/15 abstracts with case report design published
Notes	Orthopedics/orthopedic surgery Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting, type science, study design, and subspecialty with publication using stratified analysis and unspecified statistical tests.

Ha 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Korean medical database to June 2007 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by One author Results
Data	Included 1097 abstracts presented at the 2001 and 2002 Korean Radiological Society, Radiological Society of North America, and European Congress of Radiology meetings Included all research abstracts by Korean authors, excluding abstracts of imaging technique, education, or special interest
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published by meeting 'Positive' versus not 'positive' Sample size < 20 versus between 20 and 50 versus > 50 Oral versus poster presentation 'Prospective' versus 'retrospective' design With statistical analyses versus without statistical analyses Subspecialty
Outcomes	301 of 1097 abstracts published 43/85 abstracts presented at the 2001 to 2002 European Congress of Radiology meeting, 58/164 at the 2001 to 2002 Radiological Society of North America meetings, and 200/848 of the 2001 to 2002 Korean Radiological Society meetings published Proportion of abstracts published by time Mean time to publication = 15.8 months (SD = 13.8) Cumulative proportion of abstracts published at 48 months showed proportion published = 23.0% (195/848 abstracts) for Korean Radiological Society meetings Cumulative proportion of abstracts published at 48 months showed proportion published = 34.7% (57/164 abstracts) for Radiological Society of North America meetings Cumulative proportion of abstracts published at 48 months showed proportion published = 47.1% (40/85 abstracts) for European Congress of Radiology meetings Factors related to proportion of abstracts published included 288/982 'positive' (defined as experimental better than control) versus 13/115 not 'positive' abstract results published 102/420 abstracts with sample size < 20 versus 100/325 with sample size between 20 and 50 versus 99/352 abstracts with sample size > 50 published 209/732 abstracts presented orally versus 92/365 abstracts presented as posters published 103/308 abstracts with 'prospective' study design versus 198/789 abstracts with 'retrospective' study design published 155/411 abstracts with statistical analyses versus 146/686 abstracts without statistical analyses published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup with reasonable exclusions, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting, positive results, sample size, type presentation, prospective status, presence of statistical analysis, and subspecialty with publication using stratified analysis and Chi2 tests.

Hackett 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for the year before presentation until December 31, 2012 Person completing the search not reported Searched by first author, keywords, and first author affiliation Matched abstract to full‐length publication by All authors Title Abstract Disclosure of prior presentation
Data	Included 2345 abstracts presented at the 2004, 2005, and 2007 International Liver Transplantation Society and 2006 and 2008 Joint International Congress of the International Liver Transplantation Society, European Liver and Intestinal Transplant Association, Liver Intensive Care of Europe meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published, overall and by meetings Oral versus plenary versus poster presentation Clinical research versus basic science research RCT design versus observational design versus case report design English language versus non‐English language Subspecialty
Outcomes	913 of 2345 abstracts published 150/299 abstracts presented at the 2004 International Liver Transplantation Society meeting, 155/338 at the 2005 International Liver Transplantation Society meeting, 245/515 at the 2006 Joint International Congress meeting, 172/561 at the 2007 International Liver Transplantation Society meeting, and 192/634 at the 2008 Joint International Congress meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 48 months showed proportion published = 38.9% (913/2345 abstracts) overall Cumulative proportion of abstracts published at 60 months showed proportion published = 30.3% (192/634 abstracts) for the 2008 Joint International Congress of the International Liver Transplantation Society, European Liver and Intestinal Transplant Association, Liver Intensive Care of Europe Cumulative proportion of abstracts published at 60 months showed proportion published = 30.7% (172/561 abstracts) for the 2007 International Liver Transplantation Society Cumulative proportion of abstracts published at 80 months showed proportion published = 47.6% (245/515 abstracts) for the 2006 Joint International Congress of the International Liver Transplantation Society, European Liver and Intestinal Transplant Association, Liver Intensive Care of Europe Cumulative proportion of abstracts published at 80 months showed proportion published = 45.9% (155/338 abstracts) for the 2005 International Liver Transplantation Society Cumulative proportion of abstracts published at 80 months showed proportion published = 50.2% (150/299 abstracts) for the 2004 International Liver Transplantation Society Factors related to proportion of abstracts published included 300/619 abstracts presented orally versus 60/120 abstracts presented in plenary sessions versus 559/1606 abstracts presented as poster published 815/2183 abstracts describing clinical research versus 103/164 abstracts describing basic science research published 39/59 abstracts describing RCTs versus 699/1904 abstracts describing observational designs versus 77/220 abstracts describing case reports published 237/706 abstracts in English language versus 674/1639 abstracts in non‐English language published
Notes	Surgery ‐ liver transplantation Funding through university departments
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting, type presentation, type science, study design, origin in English language country, and subspecialty with publication using stratified analysis and Chi2 or Fisher's Exact tests and log rank analysis for time to publication.

Hajji 2016.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and PubMed in April 2015 Person completing search not reported Searched by first and last author, and keywords Matched abstract to full‐length publication by One author Contents
Data	Included 270 abstracts presented at 2006 to 2012 Société interdisciplinaire Francophone d’Uro‐dynamique et de Pelvi‐périnéologie meetings Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication
Outcomes	110 of 270 abstracts published 14/22 abstracts presented at the 2006 meeting, 12/24 at the 2007 meeting, 24/56 at the 2008 meeting, 15/26 at the 2009 meeting, 18/42 at the 2010 meeting, 21/41 at the 2011 meeting, and 16/62 at the 2012 meeting published Proportion of abstracts published by time Mean time to publication = 22 months (SD = 15) No factors other than meeting year related to proportion of abstracts published reported
Notes	Gynecology/obstetrics ‐ urogynecology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings except the 2012 meeting had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Halpern 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE, Cochrane Central Register of Controlled Trials to April 2000 Search completed by investigator Searched by first, and senior authors, and keywords Matching criteria not reported Searched author's personal files
Data	Included 145 abstracts presented at 1994 and 1995 American Society of Anesthesiologists meetings Included all abstracts related to obstetrical anesthesia
Comparisons	Proportion of abstracts published Mean time to publication 'Positive' versus not 'positive' Clinical research versus basic science research RCT design versus non‐RCT design Peer‐reviewed funding versus none
Outcomes	51 of 145 abstracts published 9 full publications found and excluded, including 3 abstracts published at another meeting; 4 articles that did not match abstract content, although titles were similar; 1 which was a review of the abstract as published in another journal; and 1 abstract that was a description of the methodology subsequently used in another published manuscript Proportion of abstracts published by time Mean time to publication = 28 months (SD = 17) Factors related to proportion of abstracts published included 29/83 'positive' (defined as significant results) versus 9/47 not 'positive' abstract results published 40/113 abstracts describing clinical research versus 11/32 abstracts describing basic science research published 21/47 abstracts with RCT design versus 30/98 abstracts with non‐RCT design published 9/51 abstracts with peer‐reviewed funding versus 7/94 abstracts without peer‐reviewed funding published
Notes	Gynecology/obstetrics ‐ obstetrical anesthesia Funding from "departmental sources" Author provided unpublished manuscript
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	Yes	Searched 3 databases and author's personal files.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results, type science, and funding status with publication using stratified analysis and Chi2 tests.

Hamlet 1997.

Methods	Identification of subsequent full‐length publications Searched electronic database Melvyl MEDLINE Plus from January 1990 through August 1996 for abstracts from 1990 and 1991 meeting, and from January 1992 through August 1996 for abstracts from 1992 Person completing search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by Keywords Methodology
Data	Included 1465 abstracts presented at the 1990, 1991, and 1992 meetings of the American Academy of Orthopaedic Surgeons meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published by meeting Clinical research versus basic science research Subspecialty
Outcomes	668 of 1465 abstracts published 239/523 abstracts presented at the 1990 meeting, 228/501 at the 1991 meeting, and 201/441 at the 1992 meeting published Proportion of abstracts published by time Mean time to publication = 20 months Cumulative proportion of abstracts published at 72 months showed proportion published = 45.7% (239/523 abstracts) for the 1990 American Academy of Orthopaedic Surgeons meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 45.5% (228/501 abstracts) for the 1991 American Academy of Orthopaedic Surgeons meeting Cumulative proportion of abstracts published at 48 months showed proportion published = 45.5% (201/441 abstracts) for the 1992 American Academy of Orthopaedic Surgeons meeting Factors related to proportion of abstracts published included 650/1437 abstracts describing clinical research versus 18/28 abstracts describing basic science research published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type science, and subspecialty with publication using stratified analysis and Chi2 tests

Hanchanale 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to March 2011 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 538 abstracts presented at 2005 European Association for Palliative Care Congress meeting Included all abstracts except invited abstracts
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published UK versus Dutch versus German versus Italian versus Norwegian versus rest of the world origin
Outcomes	230 of 538 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 36 months showed proportion published = 39.0% (210/538 abstracts) Factors related to proportion of abstracts published included 70/163 abstracts originating from the UK versus 22/47 abstracts originating from the Netherlands versus 15/38 abstracts originating from Germany versus 6/32 abstracts originating from Italy versus 20/29 abstracts originating from Norway versus 97/229 abstracts with rest of the world origin published
Notes	General medicine/primary care ‐ palliative care Reported no funding received Data extracted from research letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of country of origin with publication using stratified analysis.

Harel 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar to April 2011 Person completing the search not reported Searched by first, and last author, and title Matched abstract to full‐length publication by One author (either first or last author) Methodology At least one outcome
Data	Included 300 abstracts presented at the 2006 National Kidney Foundation, American Society of Nephrology, and Annual European Renal Association ‐ European Dialysis and Transplant Association Congress meetings Included random sample of abstracts presented as posters
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication overall and by meeting 'Positive' versus not 'positive' Clinical research versus basic science research RCT design versus observational design versus case report versus other design 'Collaborative' versus without 'collaborative' design Multi‐centered versus single center Industry funding versus no industry funding 'American' versus 'non‐American' origin
Outcomes	127 of 300 abstracts published 44/100 abstracts presented at the 2006 Annual European Renal Association ‐ European Dialysis and Transplant Association Congress meeting, 54/100 at the 2006 American Society of Nephrology meeting, and 29/100 at the 2006 National Kidney Foundation meeting published Proportion of abstracts published by time Median time to publication = 12 months (IQR = 6 to 22 months) overall Median time to publication = 13 months (IQR = 7 to 21 months) for the National Kidney Foundation meeting Median time to publication = 10 months (IQR = 5 to 20 months) for the American Society of Nephrology meeting Median time to publication = 12.5 months (IQR = 8 to 22.5 months) for the Annual European Renal Association ‐ European Dialysis and Transplant Association Congress Factors related to proportion of abstracts published included 104/211 'positive' (defined as experimental better than control) versus 23/89 not 'positive' abstract results published 96/258 abstracts describing clinical research versus 31/42 abstracts describing basic science research published 11/18 abstracts with RCT design versus 74/171 abstracts with observational study design versus 3/32 abstracts with case report design versus 39/79 abstracts with other design published 15/26 abstracts with 'collaborative' study design versus 112/274 abstracts without a 'collaborative' design published 55/118 abstracts with multiple centers versus 72/182 abstracts with a single center published 23/46 abstracts with industry funding versus 104/254 abstracts without industry funding published 50/137 abstracts originating in 'America' versus 77/163 abstracts not originating in 'America' published
Notes	Nephrology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random sample of poster presentations.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had 5 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting, positive results, type science, study design, multi‐center status, industry funding, 'American' origin with publication using multivariable logistic regression analysis.

Harris 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE to March 2004 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 200 abstracts presented at the 1998 Australian Orthopaedic Association meeting Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Clinical research versus basic science research RCT design versus observational versus case series versus other design North American versus European versus Australian versus Asian versus Oceanian origin
Outcomes	62 of 200 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 12/24 'positive' (defined as significant results) versus 50/176 not 'positive' abstract results published 45/132 'positive' (defined as experimental better than control) versus 17/68 not 'positive' abstract results published 40/159 abstracts describing clinical research versus 22/41 abstracts describing basic science research published 1/3 abstracts with RCT design versus 17/41 abstracts with observational study design versus 31/103 abstracts with case series design versus 13/53 with other design published 21/55 abstracts originating from North America versus 5/19 abstracts from Europe versus 32/110 abstracts from Australia versus 4/11 abstracts from Asia versus 0/5 abstracts from Oceania published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted as oral presentations
Search for publications?	Yes	Searched 2 databases and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type science, study design, and country of origin using multivariable logistic regression analysis.

Harris 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE to March 2004 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 318 abstracts presented at the 1999 American Academy of Orthopaedic Surgeons meeting Included all abstracts
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Clinical research versus basic science research RCT design versus observational versus case series versus other design North American versus European versus rest of the world origin Presence of 'sponsor' versus none
Outcomes	175 of 318 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 69/101 'positive' (defined as significant results) versus 107/217 not 'positive' abstract results published 123/203 'positive' (defined as experimental better than control) versus 53/115 not 'positive' abstract results published 133/244 abstracts describing clinical research versus 43/74 abstracts describing basic science research published 11/18 abstracts with RCT design versus 45/77 abstracts with observational study design versus 95/184 abstracts with case series design versus 25/39 abstracts with other design published 153/283 abstracts originating from North America versus 14/20 abstracts from Europe versus 9/15 abstracts from rest of the world published 49/90 abstracts with a 'sponsor' versus 127/228 abstracts without a 'sponsor' published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type science, study design, country of origin, and funding status using multivariable logistic regression analysis.

Harshavardhana 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to May 2009 Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by All authors Sample Size Contents Methodology Results Level of evidence Conclusions Conflict of interest
Data	Included 278 abstracts presented at the 2000 to 2004 Cervical Spine Research Society meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published
Outcomes	203 of 278 abstracts published 43/59 abstracts presented at the 2000 meeting, 30/51 at the 2001 meeting, 43/58 at the 2002 meeting, 48/57 at the 2003 meeting, and 39/53 at the 2004 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 24 months showed proportion published = 55.4% (154/278 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 8 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis

Harshavardhana 2009a.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2008 Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by Study cohort Design Conclusion Authorship
Data	Included 1063 abstracts presented at the 2000 to 2004 Scoliosis Research Society meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation
Outcomes	560 of 1063 abstracts published 98/210 abstracts presented at the 2000 meeting, 97/166 at the 2001 meeting, 106/214 at the 2002 meeting, 137/259 at the 2003 meeting, and 92/214 at the 2004 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 295/452 abstracts presented orally versus 265/611 abstracts presented as posters published
Notes	Orthopedics/orthopedic surgery ‐ scoliosis Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year by type presentation with publication using stratified analysis and unspecified statistical tests.

Harvey 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, MEDLINE to March 2008 Person completing the search not reported Searched by first author and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 442 abstracts presented at the 2002 and 2003 Medical Library Association meetings Included all abstracts except those by invited guest speakers
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	122 of 442 abstracts published 47/189 abstracts presented at the 2002 meeting, and 75/253 at the 2003 meeting published Proportion of abstracts published by time not reported No factors other than meeting year related to proportion of abstracts published reported
Notes	Medical education/library science Funded by Medical Library Association Research, Development, and Demonstration Project Grant
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases and contacted abstract authors.
Follow‐up time?	Yes	Each meeting had at least 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Hashkes 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to January 2002 Search completed by investigator Searched by first, last, and senior author, and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 331 abstracts submitted to the 1986, 1991 and 1998 Park City Pediatric Rheumatology meetings Included all abstracts submitted to 1998 meeting, and random sample of the 1991 and 1986 meetings
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Clinical research versus basic science research North American versus European versus rest of the world origin 'Novel' research versus other research Disease studied Subspecialty
Outcomes	134 of 331 abstracts published 16/28 abstracts presented at the 1986 meeting, 26/46 at the 1991 meeting, and 92/257 at the 1998 meeting published Proportion of abstracts published by time Median time to publication = 24 months Cumulative proportion of abstracts published at 45 months showed proportion published = 36.5% (92/257 abstracts) Factors related to proportion of abstracts published included 54/112 'positive' (defined as experimental better than control) versus 38/145 not 'positive' abstract results published 87/245 abstracts describing clinical research versus 5/12 abstracts describing basic science research published 49/135 abstracts originating from North America versus 43/122 abstracts from Europe versus 17/60 from the rest of the world published 29/59 abstracts presenting 'novel' research versus 40/145 abstracts presenting other research published
Notes	Rheumatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of one meeting and random samples of the other meetings.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, positive results, type science, country of origin, 'novel' status, and subspecialty with publication using stratified analysis and nominal regression or Chi2 tests.

Herbison 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, Cochrane Central Register of Controlled Trials, Embase, MEDLINE, PsychLIT, HealthStar to December 2005 Person completing the search not reported Searched by all authors Matched abstract to full‐length publication by Sample size Contents Contacted abstract authors directly
Data	Included 82 abstracts presented at the 1987, 1998 and 1999 International Continence Society meetings Included all abstracts of randomized controlled trials
Comparisons	Proportion of abstracts published Mean time to publication 'Positive' versus not 'positive' Sample size greater than 100 versus sample size at or below 100 Oral versus poster presentation Presence of allocation concealment versus no or unsure allocation concealment Presence of blinded treatment administrator versus no or unsure treatment administrator Presence of blinded outcome assessor versus no or unsure blinded outcome assessor Proportion of withdrawals Use of placebo Device versus drug versus neither as intervention Use of subjective outcome Multi‐centered versus single center 'Company' funding versus none
Outcomes	41 of 82 abstracts published Proportion of abstracts published by time Mean time to publication = 19.2 months (SD = 9.6) for 1998 and 1999 International Continence Society meetings Factors related to proportion of abstracts published from 1998 and 1999 International Continence Society meetings included 20/49 'positive' (defined as experimental better than control) versus 5/14 not 'positive' abstract results published 14/27 abstracts with sample size greater than 100 versus 10/34 abstracts with sample size of 100 or less published 4/10 abstracts presented orally versus 18/43 abstracts presented as posters published 7/20 abstracts with allocation concealment present versus 18/43 abstracts with no or unsure allocation concealment published 7/21 abstracts with blinded treatment administrator versus 18/42 abstracts with no or unsure treatment administrator published 0/7 abstracts with blinded outcome assessor versus 25/56 abstracts with no or unsure blinded outcome assessor published 11/23 abstracts with multiple centers versus 14/40 abstracts with a single center published 3/16 abstracts with 'company' funding versus 22/47 abstracts without 'company' funding published
Notes	Urology ‐ incontinence Funding by university
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific study design, so low risk of bias.
Search for publications?	Yes	Searched 6 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of positive results, sample size, type of presentation, multi‐center status, funding status, and presence of blinding for treatment administrator, or outcome assessor, proportion of withdrawals, use of placebo, type intervention, and use of subjective outcome with publication using stratified analysis and Chi2 tests.

Hernandez‐Garcia 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Indice Medico Espanol to June 2010 Person completing the search not reported Searched by all authors Matched abstract to full‐length publication by Title Contents Results Contacted abstract authors directly
Data	Included 325 abstracts presented at the 2006 Spanish Society for Quality in Healthcare meeting Included all abstracts presented orally
Comparisons	Proportion of abstracts published Hospital versus not hospital Spanish origin versus non‐Spanish origin
Outcomes	15 of 325 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 10/212 abstracts in hospitals versus 5/113 abstracts not in hospitals published 15/320 abstracts originating in Spain versus 0/2 abstracts not originating in Spain published
Notes	Medical decision making/health policy ‐ quality in healthcare Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	Yes	Searched 2 database and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of institutional affiliation, and Spanish origin with publication using stratified analysis and Chi2 or Fisher's Exact tests

Hernández‐Jiménez 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, Medscape, and Biomedexperts systems; dates of search not reported Person completing the search not reported Searched by all authors, keywords, title, and location of research Matching criteria not reported
Data	Included 1045 abstracts presented at the 2000 to 2009 Annual Congress of the Sociedad Mexicana de Nutrición y Endocrinología Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Clinical research versus basic science research RCT design versus case report design versus case series versus case‐control design versus file review versus survey
Outcomes	186 of 1045 abstracts published Proportion of abstracts published by time Mean time to publication = 25.2 months (SD = 22.8; range = 12 to 108 months) Factors related to proportion of abstracts published included 148/930 abstracts describing clinical research versus 38/115 abstracts describing basic science research published 30/112 abstracts describing RCTs versus 12/228 abstracts describing case reports versus 36/154 abstracts describing case series versus 26/233 abstracts describing case‐control studies versus 21/125 abstracts describing file reviews versus 23/79 abstracts describing surveys published
Notes	Endocrinology/nutrition Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type science and study design with publication using stratified analysis.

Herron 1993.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, CIM from 1987 through June 1992 Person completing the search not reported Searched by first, and second author, and title Matching criteria not reported
Data	Included 160 abstracts presented at the 1990 Association for Parental and External Nutrition and Eastern Association of the Surgery of Trauma meetings Included all abstracts Also reported on 68 abstracts presented at the 1990 meetings of Association of Air Medical Services, National Flight Nurses Association, National Flight Paramedics Association, National EMS Pilots Association, but not included in this analysis, and 311 abstracts presented in 1987, 1988, 1989, and 1990 at the American Transport conference meeting and previously reported on by Schwartz 1992
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	72 of 160 abstracts published 29/40 abstracts presented at the Eastern Association of the Surgery of Trauma meeting, and 43/120 at the Association for Parental and External Nutrition meeting published Proportion of abstracts published by time not reported No factors other than meeting related to proportion of abstracts published reported
Notes	Emergency medicine ‐ air medical transport Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	No	Only the meetings before 1989 had at least 48 months follow‐up. The meetings in 1989 and 1990 only had 36 and 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting with publication using stratified analysis.

Hoag 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to May 2005 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by One author One conclusion
Data	Included 1584 abstracts presented at the 2000 American Urological Association meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Oral versus poster presentation North American versus European versus South American versus African versus Asian versus Middle East versus Oceanian origin Northeast US versus south US versus Midwest US versus west US origin Subspecialty
Outcomes	875 of 1584 abstracts published Proportion of abstracts published by time Mean time to publication = 17 months Factors related to proportion of abstracts published included 192/324 abstracts presented orally versus 654/1191 abstracts presented as posters published 576/1029 abstracts originating from North America versus 204/374 abstracts from Europe versus 9/20 abstracts from South America versus 0/2 abstracts from Africa versus 69/131 abstracts from Asia versus 15/23 abstracts from the Middle East versus 2/5 abstracts from Oceania published
Notes	Urology Funding for data analysis and statistical services provided by Analysis Works, Inc.
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of type presentation, country of origin, and subspecialty with publication using stratified analysis and t tests

Hoeg 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, PubMed, Google/Google Scholar to April 2007 Search completed by trained librarian Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 559 abstracts presented at the 1997, 1999, and 2001 American Society of Clinical Oncology meetings Included all abstracts of phase II trials, excluded those with preliminary results, ongoing studies, or those with a 'phase III component'
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication by meeting and overall Cumulative proportion of abstracts published Sample size > 35 versus sample size < 35 Oral or poster presentation versus print only Industry sponsor versus none North American or European versus rest of the world origin Subspecialty Type of treatment
Outcomes	361 of 559 abstracts published 90/126 abstracts presented at the 1997 meeting, 124/177 at the 1999 meeting, and 147/256 at the 2001 meeting published Proportion of abstracts published by time Median time to publication = 41 months (95% CI = 37 to 46 months) overall Median time to publication = 33 months for 1997 American Society of Clinical Oncology meeting Median time to publication = 41 months for 1999 American Society of Clinical Oncology meeting Median time to publication = 48 months for 2001 American Society of Clinical Oncology meeting Cumulative proportion of abstracts published at 84 months showed proportion published = 66.8% (373/559 abstracts) Factors related to proportion of abstracts published included 76/259 abstracts with sample size > 35 versus 119/296 abstracts with sample size < 35 published 32/39 abstracts presented orally versus 163/233 abstracts presented as posters versus 166/287 abstracts appearing in print only published 117/174 abstracts with an industry sponsor versus 207/251 abstracts with a non‐industry sponsor versus 37/134 abstracts without funding or not reporting funding published 192/294 abstracts originating from North America versus 116/177 abstracts originating from Europe versus 53/88 abstracts with rest of world origin published
Notes	Oncology Funding by the Gundersen Lutheran Center for Cancer and Blood Disorders and the Gundersen Lutheran Medical Foundation
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup with reasonable exceptions, so low risk of bias.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Unclear	Examined association of meeting year, sample size, type presentation, industry sponsor, country of origin, type treatment, and subspecialty with publication using stratified analysis and Chi2 tests and Cox regression models for time to publication.

Hogan 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 117 abstracts presented at the 2000, 2003, and 2006 American Contact Dermatitis Society meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	54 of 117 abstracts published Proportion of abstracts published by time Mean time to publication = 18 months No factors related to proportion of abstracts published reported
Notes	Dermatology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

Hopewell 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Cochrane Central Register of Controlled Trials in January and July 2000 Person completing the search not reported Searched by all authors Matched abstract to full‐length publication by Contents Contacted abstract authors directly
Data	Included 91 abstracts presented at the 1998 First Symposium of Systematic Reviews and the Third Cochrane Colloquium (1995) meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	39 of 91 abstracts published 26/61 abstracts presented at the 1995 meeting, and 13/30 at the 1998 meeting published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Medical decision making/health policy ‐ systematic reviews, health technology Funding from the National Health Service Research & Development Programme
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases and contacted abstract authors.
Follow‐up time?	Yes	The meeting in 1995 had 5 years follow‐up, but the meeting in 1998 only had 24 months follow‐up.
Matching?	No	Matched by only 1 criterion.

Hopewell 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Cochrane Central Register of Controlled Trials to May 2002 Search completed by investigator Searched by first, second, and third author, and keywords Matching criteria not reported
Data	Included 962 abstracts published in the 1980 to 2000 Australian New Zealand Journal of Medicine, originally presented at 17 different society meetings Included all abstracts describing RCTs
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published
Outcomes	589 of 962 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 96 months showed proportion published = 61.0% (587/962 abstracts) No factors related to proportion of abstracts published reported
Notes	General medicine/primary care Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific study design, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings before 1999 had at least 48 months follow‐up. The meetings in 1999 and 2000 had 36 and 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Hopewell 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from 1 January, 2010 until June 2013 Search completed by investigators Searched by first, and last author, keywords, and title Matching criteria not reported
Data	Included 197 abstracts presented at the 2010 Infectious Diseases Society of America, American Thoracic Society, American Society of Hematology, American Society of Clinical Oncology, American Society of Anesthesiologists, American Psychiatric Association, American Heart Association, American Diabetes Association, and American College of Rheumatology meetings Included all abstracts of systematic reviews except those of animal studies
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication, overall and by meeting Cumulative proportion of abstracts published, overall and by meeting Oral versus poster presentation
Outcomes	105 of 197 abstracts published 5/8 abstracts presented at Infectious Diseases Society of America meeting, 17/25 at the American Thoracic Society meeting, 15/34 at the American Society of Hematology meeting, 28/57 at the American Society of Clinical Oncology meeting, 2/8 at the American Society of Anesthesiologists meeting, 3/7 at the American Psychiatric Association meeting, 17/32 at the American Heart Association meeting, 5/8 at the American Diabetes Association meeting, and 13/18 at the American College of Rheumatology meeting published Proportion of abstracts published by time Median time to publication = 14.0 months (IQR = 6.6 to 20.1 months), overall Median time to publication = 12.3 months (IQR = 6.5 to 22.9 months) for 2010 Infectious Diseases Society of America meeting Median time to publication = 15.9 months (IQR = 9 to 23.4 months) for 2010 American Society of Hematology meeting Median time to publication = 15.9 months (IQR = 11.9 to 22.9 months) for 2010 American Society of Clinical Oncology meeting Median time to publication = 10.0 months (IQR = 5.5 to 14.5 months) for 2010 American Society of Anesthesiologists meeting Median time to publication = 17.4 months (IQR = 0.3 to 26.4 months) for 2010 American Psychiatric Association meeting Median time to publication = 10.9 months (IQR = 4.2 to 19.6 months) for 2010 American Heart Association meeting Median time to publication = 11.8 months (IQR = 7.6 to 20 months) for 2010 American Thoracic Society meeting Median time to publication = 8.2 months (IQR = 2.7 to 20.2 months) for 2010 American Diabetes Association meeting Median time to publication = 5.7 months (IQR = 4.7 to 20.8 months) for 2010 American College of Rheumatology meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 53.3% (105/197 abstracts), overall Cumulative proportion of abstracts published at 36 months showed proportion published = 62.5% (5/8 abstracts) for 2010 Infectious Diseases Society of America meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 44.1% (15/34 abstracts) for 2010 American Society of Hematology meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 49.1% (28/57 abstracts) for 2010 American Society of Clinical Oncology meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 25.0% (2/8 abstracts) for 2010 American Society of Anesthesiologists meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 42.9% (3/7 abstracts) for 2010 American Psychiatric Association meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 53.1% (17/32 abstracts) for 2010 American Heart Association meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 68.0% (17/25 abstracts) published for 2010 American Thoracic Society meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 62.5% (5/8 abstracts) for 2010 American Diabetes Association meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 72.2% (13/18 abstracts) for 2010 American College of Rheumatology meeting Factors related to proportion of abstracts published included 19/33 abstracts presented orally versus 86/164 abstracts presented as poster published
Notes	Other clinical specialties ‐ rheumatology, endocrinology ‐ diabetes, cardiology, psychiatry, anesthesiology, oncology, hematology, thoracic medicine, infectious diseases Reported not receiving funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific study design with reasonable exception, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting and type presentation with publication using stratified analysis.

Hopper 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, PubMed, Science Citation Index to December 2006 Person completing the search not reported Searched by first, and senior author, and keywords Matching criteria not reported
Data	Included 4096 abstracts presented at the 1994 to 2002 British Society of Gastroenterology meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	1618 of 4096 abstracts published 325/570 abstracts presented at the 1994 meeting, 226/480 at the 1995 meeting, 209/465 at the 1996 meeting, 115/330 at the 1997 meeting, 119/395 at the 1998 meeting, 178/575 at the 1999 meeting, 156/380 at the 2000 meeting, 146/405 at the 2001 meeting, and 144/465 at the 2002 meeting published Proportion of abstracts published by time not reported No factors other than meeting year related to proportion of abstracts published reported
Notes	Gastroenterology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Housri 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to August 2007 Person completing the search not reported Searched by principal investigator and keywords Matched abstract to full‐length publication by Hypothesis Methodology Results
Data	Included 1200 abstracts presented at the 2002 to 2004 Society of University Surgeons and Association for Academic Surgery meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation, by meeting Clinical research versus basic science research
Outcomes	722 of 1200 abstracts published 348/657 abstracts presented at the 2002 to 2004 Association for Academic Surgery meetings, and 374/543 at the 2002 to 2004 Society of University Surgeons meetings published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 204/322 abstracts presented orally versus 144/335 abstracts presented as posters published (Association for Academic Surgery meetings) 247/302 abstracts presented orally versus 96/204 abstracts presented at resident conferences versus 19/25 abstracts presented as posters published (Society of University Surgeons meetings) 710/1188 abstracts describing clinical research versus 12/12 abstracts describing basic science research published
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2004 had at least 48 months follow‐up. The meeting in 2004 only had 36 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting, type presentation by meeting, and type science with publication using stratified analysis and Fisher's Exact or t tests.

Hussein 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 367 abstracts accepted at the 2006 to 2011 Vascular Society of Great Britain and Ireland meetings Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication
Outcomes	239 of 367 abstracts published Proportion of abstracts published by time Median time to publication = 12 months (IQR = 12 to 24 months) No factors related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ vascular medicine Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

Izadpanah 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase and MEDLINE for '10 years from date of presentation' Person completing the search not reported Searched by first and senior author, and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 128 abstracts accepted at the 2003 and 2004 Annual Meeting of European Association of Plastic Surgeons Included all abstracts except guest speakers’ lectures, award papers, or presentations from previous meetings
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Clinical research versus non‐clinical research US origin versus non‐US origin Level of evidence I and II versus III and IV Prospective study design versus not Control groups versus no control groups Blinding versus no blinding
Outcomes	73 of 128 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 34/57 abstracts with 'positive' (defined as significant results) versus 41/71 with not 'positive' abstract results published 57/98 abstracts with 'positive' (defined as experimental better than control) versus 19/30 with not 'positive' abstract results published 49/88 abstracts describing clinical research versus 24/40 abstracts describing non‐clinical research published 6/13 abstracts with US origin versus 72/115 abstracts with non‐US origin published 12/21 abstracts with level I or II evidence versus 69/107 abstracts with level III or IV evidence published 27/46 abstracts with prospective design versus 46/82 abstracts without prospective design published 27/41 abstracts with control groups versus 46/87 abstracts without control groups published 6/9 abstracts with blinding versus 67/119 abstracts without blinding published
Notes	Surgery ‐ plastic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases and contacted abstract authors directly.
Follow‐up time?	Yes	All meetings had 10 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type science, study design, US origin, prospective status, level of evidence, presence of a control group, and presence of blinding with publication using multivariable logistic regression analysis.

Jackson 2000.

Methods	Identification of subsequent full‐length publications Searched electronic database Melvyl MEDLINE Plus from January 1988 through October 1998 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Sample size Title Methodology Results
Data	Included 777 abstracts submitted to the 1991 to 1994 Pediatric Orthopaedic Society of North America meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication (for accepted and rejected abstracts) Cumulative proportion of abstracts published by meeting Accepted versus rejected for conference presentation Subspecialty
Outcomes	348 of 777 abstracts published 91/184 abstracts presented at the 1991 meeting, 94/196 at the 1992 meeting, 87/192 at the 1993 meeting, and 76/195 at the 1994 meeting published Proportion of abstracts published by time Median time to publication = 29 months for accepted papers Median time to publication = 27 months for rejected papers Cumulative proportion of abstracts published at 84 months showed proportion published = 55.2% (48/87 abstracts) for 1991 Pediatric Orthopaedic Society of North America meeting Cumulative proportion of abstracts published at 72 months showed proportion published = 54.5% (48/88 abstracts) for 1992 Pediatric Orthopaedic Society of North America meeting Cumulative proportion of abstracts published at 64 months showed proportion published = 47.6% (39/82 abstracts) for 1993 Pediatric Orthopaedic Society of North America meeting Cumulative proportion of abstracts published at 56 months showed proportion published = 46.7% (43/92 abstracts) for 1994 Pediatric Orthopaedic Society of North America meeting Factors related to proportion of abstracts published included 184/349 abstracts accepted for presentation versus 164/428 rejected abstracts published
Notes	Orthopedics/orthopedic surgery ‐ pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, acceptance for presentation, and subspecialty with publication using stratified analysis and Chi2 tests.

Jamjoom 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar in February 2013 Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 273 abstracts presented at the 2007 to 2012 Saudi Association of Neurological Surgery meetings Included all abstracts presented orally except presentations from invited international guest speakers
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published Consultant versus resident versus other staff Academic versus non‐academic affiliation Saudi Arabian versus rest of the world origin Subspecialty
Outcomes	35 of 273 abstracts published 1/14 abstracts presented at the 2007 meeting, 6/35 at the 2008 meeting, 9/61 at the 2009 meeting, 6/51 at the 2010 meeting, 10/64 at the 2011 meeting, and 3/48 at the 2012 meeting published Proportion of abstracts published by time Median time to publication = 12 months Cumulative proportion of abstracts published at 36 months showed proportion published = 12.8% (35/273 abstracts) Factors related to proportion of abstracts published included 29/214 abstracts presented by consultant versus 6/43 abstract presented by resident versus 0/16 abstracts presented by other staff published 17/84 abstracts originating in an academic center versus 15/153 abstracts not originating in an academic center versus published 32/237 abstracts originating from Saudi Arabia versus 3/36 abstracts with rest of the world origin published
Notes	Surgery ‐ neurologic surgery Reported having received no funding Data extracted from brief communication
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	The meetings in 2010 to 2012 had less than 48 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, author status, academic center origin, Saudi Arabian origin, Saudi Arabian region, and subspecialty with publication using stratified analysis and unspecified statistical tests.

Jamjoom 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Ovid with 13 years of follow‐up and Google/Google Scholar to May 2013 Person completing the search not reported Searched by lead author, keywords, and title Matching criteria not reported
Data	Included 494 abstracts presented at the 2001 to 2005 Society of British Neurological Surgeons Spring and Fall meetings Included all abstracts except those from Spring 2002 meeting
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Clinical research versus animal research RCT design versus non‐RCT design UK versus rest of the world origin Subspecialty
Outcomes	181 of 494 abstracts published 19/52 abstracts presented at the Spring 2001 meeting, 20/46 at the Fall 2001 meeting, 13/50 at the Fall 2002 meeting, 20/58 at the Spring 2003 meeting, 22/52 at the Fall 2003 meeting, 17/50 at the Spring 2004 meeting, 16/51 at the Fall 2004 meeting, 25/50 at the Spring 2005 meeting, and 28/85 at the Fall 2005 meeting published Proportion of abstracts published by time Mean time to publication = 22.8 months (range = ‐35 to 133 months) Cumulative proportion of abstracts published at 60 months showed proportion published = 34.0% (168/494 abstracts) Factors related to proportion of abstracts published included 63/113 abstracts with 'positive' (defined as significant results) versus 118/381 with not 'positive' abstract results published 171/478 abstracts describing clinical research versus 10/16 abstracts describing animal research published 16/17 abstracts describing RCTs versus 165/477 abstracts describing other research designs published 174/469 abstracts originating from the UK versus 7/25 abstracts originating elsewhere published
Notes	Surgery ‐ neurologic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had at least 8 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, positive results, type science, study design, UK origin and multi‐center status, and subspecialty with publication using stratified analysis and Chi2 tests.

Jara‐Tracchia 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2009 Person completing the search not reported Searched by all authors Matched abstract to full‐length publication by Title Contents
Data	Included 4137 abstracts presented at the 2002 to 2003 International Association for Dental Research meetings (sections for Argentina, Brazil, Chile, and Peru) and 2002 and 2005 International Association for Dental Research meetings (section for Venezuela) Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Clinical research versus basic science research Presented at Argentinean versus Brazilian versus Chilean versus Peruvian versus Venezuelan section
Outcomes	1156 of 4137 abstracts published 1/67 abstracts presented at the 2002 to 2003 Peruvian section meetings, 17/104 at the 2002 to 2003 Chilean section meetings, 997/3288 at the Brazilian section meetings, 119/416 at the Argentine section meetings, and 22/262 at the Venezuelan section meetings published Proportion of abstracts published by time Mean time to publication = 48 months (range = 36 to 60 months) Factors related to proportion of abstracts published included 866/3395 abstracts describing clinical research versus 290/742 abstracts describing basic science research published 119/416 abstracts from Argentinean section versus 997/3288 abstracts from Brazilian section versus 17/104 abstracts from Chilean section versus 1/67 abstracts from Peruvian section versus 22/262 abstracts from Venezuelan section published
Notes	Oral health ‐ dental research Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of type science, Latin American country origin, and subspecialty by Latin American country origin with publication using stratified analysis.

Jasko 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by all authors, keywords, title, contents Matched abstract to full‐length publication by 'Focus' Contents
Data	Included 336 abstracts presented at the 1991, 1992, 1995, 1997, 1998, and 1999 Musculoskeletal Tumor Society meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published
Outcomes	137 of 336 abstracts published 10/18 abstracts presented at the 1991 meeting, 16/30 at the 1992 meeting, 38/95 at the 1995 meeting, 11/30 at the 1997 meeting, 42/124 at the 1998 meeting, and 20/39 at the 1999 meeting published Proportion of abstracts published by time Mean time to publication = 21.8 months (SD = 13.5) Cumulative proportion of abstracts published at 48 months showed proportion published = 39.3% (132/336 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Oncology ‐ musculoskeletal tumor Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Joe 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 162 abstracts presented at the 2009 to 2014 American Burn Association meetings Included all survey‐based abstracts
Comparisons	Proportion of abstracts published
Outcomes	27 of 162 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Emergency medicine ‐ burns Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific methodological design.
Search for publications?	Unclear	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

Jones 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from January 1978 to December 2005 Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by All authors Sample size Title Methodology Results Research question Conclusions Contacted abstract authors directly
Data	Included 139 abstracts presented at the 2000 International AIDS Conference meeting Included all abstracts from authors with UK address and reporting on primary research with established scientific methods to systematically collect and analyze data
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published
Outcomes	47 of 139 abstracts published Proportion of abstracts published by time Mean time to publication = 19 months Median time to publication = 16 months (range = ‐2 to 59 months) Cumulative proportion of abstracts published at 60 months showed proportion published = 33.8% (47/139 abstracts) No factors related to proportion of abstracts published reported
Notes	Infectious disease/immunology ‐ HIV, AIDS Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific origin and subgroup, so low risk of bias.
Search for publications?	Yes	Searched 1 database and contacted abstract authors directly.
Follow‐up time?	Yes	The meeting had 60 months follow‐up.
Matching?	Yes	Matched by 7 different criteria.

Jones 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed 'within 4 years of meeting' Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 127 abstracts presented at the 2009 British Association of Urological Surgeons meeting Included all abstracts of RCTs and observational studies
Comparisons	Proportion of abstracts published Mean time to publication UK versus non‐UK origin RCT versus observational design High versus low score on abstracts reporting quality Subspecialty
Outcomes	43 of 127 abstracts published Proportion of abstracts published by time Mean time to publication = 17.2 months Factors related to proportion of abstracts published included 3/10 abstracts with RCT design versus 40/117 with observational design published 37/114 abstracts originating in the UK versus 6/13 originating elsewhere published 36/104 abstracts with 'high' score on abstract reporting quality versus 7/23 with low score published 5/12 abstracts with multiple centers versus 38/118 abstracts with a single center published
Notes	Urology Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of methodological designs, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of study design, UK origin, multi‐center status, abstract quality, and subspecialty with publication using multivariable logistic regression analysis.

Juzych 1991.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from January 1984 to July 1989 Search completed by investigator Searched by first author Matched abstract to full‐length publication by Sample size Keywords Title Contents
Data	Included 175 abstracts presented at the 1984 American Academy of Ophthalmology and 1985 Association for Research in Vision and Ophthalmology meetings Included randomly selected abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting Median time to publication by meeting Cumulative proportion of abstracts published by meeting Oral versus poster presentation
Outcomes	105 of 175 abstracts published 48/75 abstracts presented at the 1984 meeting, and 57/100 at the 1985 meeting published Proportion of abstracts published by time Mean time to publication = 19.4 months for 1985 Association of Research in Vision and Ophthalmology meeting Mean time to publication = 13 months for 1984 American Academy of Ophthalmology meeting Median time to publication = 18 months for 1985 Association of Research in Vision and Ophthalmology meeting Median time to publication = 11 months for 1984 American Academy of Ophthalmology meeting Median time to publication = 13 months, overall Cumulative proportion of abstracts published at 63 months showed proportion published = 57.0% (57/100 abstracts) for the 1985 Association for Research in Vision and Ophthalmology meeting Cumulative proportion of abstracts published at 56 months showed proportion published = 64.0% (48/75 abstracts) for the 1984 American Academy of Ophthalmology meeting Factors related to proportion of abstracts published included 63/88 abstracts presented orally versus 42/87 abstracts presented as posters published
Notes	Ophthalmology/optometry ‐ ophthalmology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and type presentation with publication using stratified analysis and Chi2 or Mann‐Whitney U tests

Juzych 1993.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1984 to August 1992 Search completed by investigator Searched by first author Matched abstract to full‐length publication by Keywords Title Contents
Data	Included 327 abstracts presented at the 1985 Association for Research in Vision and Ophthalmology meeting Included randomly selected abstracts stratified by type of session
Comparisons	Proportion of abstracts published Mean and median time to publication Oral versus poster presentation Clinical research versus basic science research Subspecialty
Outcomes	206 of 327 abstracts published Proportion of abstracts published by time Mean time to publication = 19.7 months Median time to publication = 17 months Factors related to proportion of abstracts published included 127/186 abstracts presented orally versus 79/141 abstracts presented as posters published 71/126 abstracts describing clinical research versus 135/201 abstracts describing basic science research published
Notes	Ophthalmology/optometry ‐ ophthalmology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 7 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of type presentation, type science, and subspecialty with publication using multivariable logistic regression analysis.

Jörgens 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE until August 2008 Person completing the search not reported Search criteria not reported Matching criteria not reported Contacted abstract authors directly
Data	Included 493 abstracts submitted to the 2004 European Association for the Study of Diabetes meeting Included a random sample of abstracts
Comparisons	Proportion of abstracts published Mean time to publication by abstract acceptance Accepted versus rejected for conference presentation Quality reviewer score (1 = high, 5 = low) being 1 to 2 versus 2.1 to 3 versus 3.1 to 4 versus 4.1 to 5
Outcomes	209 of 493 abstracts published Proportion of abstracts published by time Mean time to publication = 16.5 months (SD = 12.7) for abstracts accepted for presentation Mean time to publication = 19.3 months (SD = 14) for abstracts rejected for presentation Factors related to proportion of abstracts published included 162/317 abstracts accepted for presentation versus 47/176 abstracts rejected for presentation published 16/19 abstracts with 'quality' score of 1 to 2 versus 91/201 abstracts with score of 2.1 to 3 versus 80/243 abstracts with score of 3.1 to 4 versus 4/31 abstracts with score of 4.1 to 5
Notes	Endocrinology/nutrition ‐ diabetes Funded by the European Association for the Study of Diabetes Data abstracted from research letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random sample of abstracts.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of acceptance for presentation by abstract quality with publication using stratified analysis and Fisher's Exact, Chi2, Kruskal‐Wallis, or Mann‐Whitney U tests.

Kabay 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Web of Science, date of search not reported Search completed by investigator Searched by first, and second author Matching criteria not reported
Data	Included 2118 abstracts presented at the 1996 to 2004 Turkish National Surgical Congress meetings Included all abstracts
Comparisons	Proportion of abstracts published 'Breast and endocrine surgery' 'Hepato‐pancreato‐biliary surgery'
Outcomes	120 of 2118 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 15/309 abstracts related to breast and endocrine surgery 13/82 abstracts related to hepato‐pancreato‐biliary surgery published
Notes	Surgery Funding not reported Unable to obtain full text; data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of subspecialty with publication using stratified analysis

Kaifi 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to January 2012 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Hypotheses Methodology Results
Data	Included 76 abstracts presented at the 2006 to 2010 Academic Surgical Congress and Society of University Surgeons meetings Included all abstracts accepted for plenary sessions
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication, overall and by meeting
Outcomes	60 of 76 abstracts published 24/36 abstracts presented at the 2006 to 2010 Society of University Surgeons meetings, and 36/40 at the 2006 to 2010 Academic Surgical Congress meetings published Proportion of abstracts published by time Mean time to publication = 10 months, overall Mean time to publication = 13 months for Academic Surgical Congress meetings Mean time to publication = 11 months for Society of University Surgeons meetings Median time to publication = 12 months (range = 0 to 47 months), overall Median time to publication = 9 months (range = 0 to 47 months) for Academic Surgical Congress meetings Median time to publication = 10 months (range = 0 to 38 months) for Society of University Surgeons meetings No factors other than meeting related to proportion of abstracts published reported
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for plenary sessions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2009 had at least 48 months follow‐up. The meetings in 2009 and 2010 only had 36 and 24 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis and Fisher's Exact test.

Kalkan 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, SCOPUS, Google/Google Scholar, Clinical Key/Elsevier, MD Consult, ULAKBIM, EBSCO Discovery Service, Science Direct, and Medscape; dates of search not reported Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by All authors Title Summary
Data	Included 1721 abstracts presented at the 2011 to 2012 European Society for Emergency Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation North American versus European versus Asian origin Australian versus Middle Eastern versus rest of the world origin
Outcomes	437 of 1721 abstracts published 172/626 abstracts presented at the 2001 meeting, and 265/1095 at the 2012 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 160/461 abstracts presented orally versus 277/1260 abstracts presented as posters published 56/161 abstracts originating from North America versus 87/496 abstracts originating from Europe versus 14/28 abstracts originating from Asia versus 8/8 abstracts originating from Australia versus 215/745 abstracts originating from the Middle East versus 57/283 abstracts with rest of the world origin published
Notes	Emergency medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 9 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year by type presentation and country of origin by type presentation, prospective status, abstract quality, and subspecialty with publication using stratified analysis and Chi2 or t tests

Kalyoncu 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to July 2014 Person completing the search not reported Searched by all authors Matching criteria not reported
Data	Included 799 abstracts presented at the 2005 to 2009 Ulusal Romatoloji Kongrelerinin meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Oral versus poster presentation Rheumatology award versus not Subspecialty
Outcomes	173 of 799 abstracts published Proportion of abstracts published by time Mean time to publication = 15.9 months (SD = 9.9) Factors related to proportion of abstracts published included 28/87 abstracts presented orally versus 115/507 abstracts presented as posters published 13/36 abstracts with rheumatology award versus 160/763 abstracts without rheumatology award published
Notes	Rheumatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 60 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation and award winning status with publication using stratified analysis and t test, Chi2 or Mann‐Whitney U tests

Kaya 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed to October 2009 Person completing the search not reported Searched by first and last author Matching criteria not reported
Data	Included 208 abstracts presented at the 2006 and 2007 American Urological Association meetings Included all abstracts related to sexual medicine
Comparisons	Proportion of abstracts published Clinical research versus basic science research Academic center versus non‐academic center Number of authors Country of origin Subspecialty
Outcomes	106 of 208 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 71/137 abstracts describing clinical research versus 33/71 abstracts describing basic science research published 92/150 abstracts originating in an academic center versus 13/58 abstracts not originating in an academic center versus published
Notes	Human reproduction and development ‐ sexual medicine Reported received no funding Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of type science, academic center origin, number of authors, country of origin, and subspecialty with publication using an unspecified multivariable analysis.

Kaya Mutlu 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar on May 15, 2012 Person completing the search not reported Searched by all authors and title Matched abstract to full‐length publication by Title Methodology Materials
Data	Included 181 abstracts presented at the 2002, 2004, 2006, and 2008 Advances in Physiotherapy Symposiums meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting Subspecialty
Outcomes	37 of 181 abstracts published 3/18 abstracts presented at the 2002 meeting, 14/59 at the 2004 meeting, 11/42 at the 2006 meeting, and 9/62 at the 2008 meeting published Proportion of abstracts published by time Mean time to publication = 20.9 months (SD = 16.7), overall Mean time to publication = 20.7 months (SD = 11.5) for the 2002 meeting Mean time to publication = 22.7 months (SD = 15.9) for the 2004 meeting Mean time to publication = 23.8 months (SD = 22.2) for the 2006 meeting Mean time to publication = 14.8 months (SD = 11.3) for the 2008 meeting No factors other than meeting year related to proportion of abstracts published reported
Notes	Physical and rehabilitation medicine ‐ physical therapy Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis.

Kearney 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 1005 abstracts presented at the 1996 to 2005 Sir Peter Freyer Surgical Symposia Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication Plenary versus other oral presentation
Outcomes	371 of 1005 abstracts published Proportion of abstracts published by time Median time to publication = 19 months (range = 0 to 132 months) Factors related to proportion of abstracts published included 58/104 abstracts presented in plenary sessions versus 313/901 abstracts presented orally in other session published
Notes	Surgery Funding not reported Data extracted from abstract 104 abstracts presented in plenary session also reported in O'Connor 2015
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type oral presentation with publication using stratified analysis.

Kim 1998.

Methods	Identification of subsequent full‐length publications Searched electronic database Electronic database (not specified) from October 1994 to April 4, 1997 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 357 abstracts presented at the 1994 Korean Orthopaedic Association meeting Included all abstracts presented orally
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published One resident versus more than one resident Subspecialty
Outcomes	156 of 357 abstracts published Proportion of abstracts published by time Mean time to publication = 13 months Cumulative proportion of abstracts published at 24 months showed proportion published = 43.7% (156/357 abstracts) Factors related to proportion of abstracts published included 4/10 abstracts presented by a single resident versus 152/347 abstracts presented by more than one resident published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	Unclear	Number and name of electronic database or databases not reported.
Follow‐up time?	No	The meeting had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of author status and subspecialty with publication using stratified analysis.

Kim 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed in February 2010 Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 288 abstracts presented at the Spring and Fall 2001 to 2005 Korean Society of Spine Surgery meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Subspecialty
Outcomes	167 of 288 abstracts published 16/25 abstracts presented at the Spring 2001 meeting, 21/29 at the Fall 2001 meeting, 17/27 at the Spring 2002 meeting, 17/27 at the Fall 2002 meeting, 18/27 at the Spring 2003 meeting, 13/23 at the Fall 2003 meeting, 11/28 at the Spring 2004 meeting, 18/28 at the Fall 2004 meeting, 23/41 at the Spring 2005 meeting, and 12/33 at the Fall 2005 meeting published Proportion of abstracts published by time Mean time to publication = 12.2 months No factors other than meeting year related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery ‐ spine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis.

Kinsella 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, and Google/Google Scholar for '3 years between meeting and publication' Search completed by the investigator Searched by all authors, keywords, and title Matched abstract to full‐length publication by All authors Title Abstract
Data	Included 444 abstracts presented at the 2006 to 2010 American Orthopaedic Society for Sports Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation
Outcomes	298 of 444 abstracts published 49 of 84 abstracts presented at the 2006 meeting, 62/89 at the 2007 meeting, 46/73 at the 2008 meeting, 72/108 at the 2009 meeting, and 69/90 at the 2010 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 203/277 abstracts presented orally versus 95/167 abstracts presented as poster published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	No	All meetings only had 36 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Unclear	Examined association of meeting year and type presentation with publication using stratified analysis and Chi2 tests. Logistic regression was performed, but unclear if the regression models included a single variable or was multivariable.

Kiroff 2001.

Methods	Identification of subsequent full‐length publications Contacted abstract authors directly
Data	Included 302 abstracts presented at the 1994 to 1996 Royal Australasian College of Surgeons meetings Included all abstracts except those for foundation lectures, plenary sessions, and invited speakers
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' 'Positive' RCT versus not 'positive' RCT Clinical research versus basic science research RCT design versus non‐RCT design Level of evidence Subspecialty
Outcomes	165 of 302 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 98/139 'positive' (defined as significant results) versus 76/159 not 'positive' abstract results published 11/12 'positive' (defined as significant results) RCT abstracts versus 4/8 not 'positive' RCT abstract results published 121/249 abstracts describing clinical research versus 44/53 abstracts describing basic science research published 15/20 abstracts with RCT design versus 150/288 abstracts with non‐RCT design published 57/95 abstracts with level I or II evidence versus 16/28 abstracts with level III evidence versus 92/179 abstract with level IV evidence published
Notes	Surgery Funding not reported 271 abstracts excluded because abstract author did not respond to survey
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Contacted abstract authors directly but with a response rate < 80%.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched through contact with abstract author.
Adjustment for confounding?	No	Examined association of positive results, type science, study design, level of evidence, and subspecialty with publication using stratified analysis and Chi2 tests.

Klappenbach 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and LILACS; dates of search not reported Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title The order of the first and last author
Data	Included 200 abstracts presented at the 2006 Argentine Congress of Surgery Included all abstracts except film and video presentations
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	22 of 200 abstracts published Proportion of abstracts published by time Mean time to publication = 15 months No factors related to proportion of abstracts published reported
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Yes	Matched by 3 different criteria.

Klassen 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, Cochrane Central Register of Controlled Trials, PubMed, Embase, Web of Science, Current Contents, HealthStar between February and July 2000 Search completed by trained librarian Searched by 'primary' author and keywords Matched abstract to full‐length publication by One common outcome Contacted abstract authors directly (only for subset of authors as reported by Hartling 2004)
Data	Included 447 abstracts presented at the 1992, 1993, 1994, and 1995 Society for Pediatric Research meetings Included all RCTs identified by handsearching abstract books
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published 'Positive' versus not 'positive' 'Positive' versus not 'positive' for equivalence studies Clinically very important or important versus somewhat important or not important Government versus industry versus foundation versus internal versus no funding Subspecialty
Outcomes	264 of 447 abstracts published 55/95 abstracts presented at the 1992 meeting, 70/109 at the 1993 meeting, 78/128 at the 1994 meeting, and 61/115 at the 1995 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 96 months showed proportion published = 58.1% (255/447 abstracts) Factors related to proportion of abstracts published included 162/235 'positive' (defined as experimental better than control ) versus 93/187 not 'positive' abstract results published 64/77 'positive' (defined as significant results) versus 22/42 not 'positive' abstract results published for subset studied by Hartling 2004 88/115 abstracts considered clinically very important or important versus 31/51 abstracts considered somewhat important or not important published for subset studied by Hartling 2004 32/38 abstracts reporting government funding versus 37/43 abstracts reporting industry funding versus 6/8 abstracts reporting foundation funding versus 4/5 abstracts reporting internal funding versus 32/57 abstracts reporting no funding for subset studied by Hartling 2004
Notes	Pediatrics Funding for study by Hartling by Alberta Heritage Foundation Secondary publication by Hartling 2004 surveyed a subset of authors
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all RCTs.
Search for publications?	Yes	Searched 7 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	Yes	Examined association of meeting year, positive results, clinical importance, type funding, and subspecialty with publication using multivariable logistic regression analysis.

Kleine‐Konig 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for period of 5 years before and after each conference Person completing the search not reported Searched by all authors, and keywords Matched abstract to full‐length publication by One author Contents Methodology
Data	Included 646 abstracts presented at the 2006 to 2008 European Paediatric Orthopaedic Society meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' versus significance not specified Oral versus poster presentation RCT design versus observational design versus case report versus epidemiological design versus meta‐analysis versus review versus other study design Multi‐centered versus single center Therapeutic versus diagnostic versus prognostic versus economic clinical study Prospective versus retrospective study Level of evidence being level I versus level II versus level III versus level IV versus level V
Outcomes	237 of 646 abstracts published Proportion of abstracts published by time Mean time to publication = 13.9 months (SD = 1.3) Cumulative proportion of abstracts published at 72 months showed proportion published = 36.7% (237/646 abstracts) Factors related to proportion of abstracts published included 107/186 'positive' (defined as significant results) versus 24/62 not 'positive' abstract results versus 106/398 abstracts with significance not reported published 122/268 abstracts presented orally versus 115/378 abstracts presented as poster published 19/23 abstracts with RCT design versus 188/522 abstracts with observational design versus 5/23 abstracts with case report design versus 13/33 abstracts with epidemiological design versus 2/3 abstracts describing a meta‐analysis versus 0/5 abstracts describing a review versus 2/7 abstracts with other study designs published 33/60 abstracts with multiple centers versus 204/586 abstracts with a single center published 128/400 abstracts describing a therapeutic clinical study versus 32/74 abstracts describing a diagnostic clinical study versus 17/34 abstracts describing a prognostic clinical study versus 1/2 abstracts describing an economic clinical study published 65/136 abstracts describing a prospective study versus 171/510 abstracts describing a retrospective study published 15/22 abstracts with evidence level I versus 25/47 abstracts with evidence level II versus 46/98 abstracts with evidence level III versus 102/344 abstracts with evidence level IV versus 6/24 abstracts with evidence level V published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had 5 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of positive results, type of presentation, study design, multi‐center status, prospective status, level of evidence abstract quality, and subspecialty with publication using stratified analysis and Chi2 tests.

Kleweno 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by first, second, and last author, title, and keywords Matched abstract to full‐length publication by One author Contents
Data	Included 165 abstracts presented at the 1999, 2000, and 2001 American Orthopaedic Society for Sports Medicine meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published
Outcomes	98 of 165 abstracts published 39/64 abstracts presented at the 1999 meeting, 33/56 at the 2000 meeting, and 26/45 at the 2001 meeting published Proportion of abstracts published by time Median time to publication = 21 months (range = 1 to 60 months) Cumulative proportion of abstracts published at 60 months showed proportion published = 59.4% (98/165 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ sports medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis and Chi2 tests.

Koene 1994.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE for 3 years before and 5 years after abstract was published in Nederlands Tijdschrift voor Geneeskunde Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by Contents
Data	Included 803 abstracts published in Nederlands Tijdschrift voor Geneeskunde from January 1988 to January 1989 and presented at 29 various scientific meetings Included all abstracts
Comparisons	Proportion of abstracts published
Outcomes	385 of 803 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	General medicine/primary care Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All abstracts had 5 years follow‐up.
Matching?	No	Matched by only 1 criterion.

Korn 2000.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from January 1995 to November 1999 Person completing the search not reported Searched by all authors Matched abstract to full‐length publication by Contents
Data	Included 1637 abstracts presented at the 1997 American College of Emergency Physicians Research Forum and 1995 to 1997 Society for Academic Emergency Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Clinical research versus basic science research versus surveys 'Prospective' versus 'retrospective' design Subspecialty
Outcomes	629 of 1637 abstracts published 186/396 abstracts presented at the 1995 meeting, 210/467 at the 1996 meeting, 181/548 at the 1997 Society for Academic Emergency Medicine meeting, and 52/226 at the 1997 American College of Emergency Physicians Research Forum meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 388/1019 abstracts describing clinical research versus 96/212 abstracts describing basic science research versus 80/188 abstracts describing surveys published 263/607 abstracts with 'prospective' study design versus 125/412 abstracts with 'retrospective' study design published
Notes	Emergency medicine Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	Only the meeting in 1995 had 48 months follow‐up. All other meetings had less than 48 months follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of meeting and meeting year, type science, prospective status, and subspecialty with publication using stratified analysis.

Kottachchi 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE, PubMed, Google/Google Scholar to July 2008 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 82 abstracts presented at the 1999 to 2003 Digestive Disease Week meetings Included all abstracts describing randomized clinical trials of inflammatory bowel disease
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Survival analysis of publication rate of abstracts with 'positive' results 'Positive' versus not 'positive' Oral versus poster presentation North American versus European versus rest of the world origin Industry funding versus academic/government funding versus mixed funding versus unclear funding
Outcomes	64 of 82 abstracts published 10/12 abstracts presented at the 1999 meeting, 16/22 at the 2000 meeting, 16/21 at the 2001 meeting, 10/12 at the 2002 meeting, and 12/15 at the 2003 meeting published Proportion of abstracts published by time Median time to publication = 30 months (IQR = 17 to 57 months) Survival analysis of proportion of abstracts with 'positive' results published at 60 months = 76.8% Factors related to proportion of abstracts published included 51/56 'positive' (defined as significant results) versus 13/26 not 'positive' abstract results published 29/33 abstracts presented orally versus 34/48 abstracts presented as posters published 18/20 abstracts originating from North America versus 30/42 abstracts from Europe versus 16/20 abstracts from rest of the world published 7/8 abstracts with industry funding versus 1/2 abstracts with academic or government funding versus 2/2 abstracts with mixed funding versus 54/70 abstracts with unclear funding published
Notes	Gastroenterology ‐ inflammatory bowel disease Funding by American Gastroenterology Association Scholar Award
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting year, positive results, type presentation, country of origin, number of authors, sample size, and funding source with publication using and for time to publication using multivariable Cox proportional hazard analysis.

Koçak 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Web of Science, and Google/Google Scholar; dates of search not reported Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 188 abstracts presented at the 2004 to 2012 European League Against Rheumatism Congress meetings Included all abstracts in the physiotherapy section
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	47 of 188 abstracts published Proportion of abstracts published by time Mean time to publication = 18.8 months (range = 1 to 64 months) No factors related to proportion of abstracts published reported
Notes	Physical and rehabilitation medicine ‐ physical therapy Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts from a specific subgroup
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.

Krzyzanowska 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE, PubMed and Cochrane Central Register of Controlled Trials to November 2002 Search completed by research assistant with library science background Searched by first, second, and presenting author, and keywords Matching criteria not reported
Data	Included 510 abstracts presented at the 1989 to 1998 American Society of Clinical Oncology meetings Included all abstracts of randomized clinical trials with sample sizes larger than 200 (510 of 539 abstracts)
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published Survival analysis of publication rate 'Positive' versus not 'positive' Oral versus poster presentation Multi‐centered versus single center Subspecialty
Outcomes	415 of 510 abstracts published Proportion of abstracts published by time Median time to publication = 32.4 months Cumulative proportion of abstracts published at 120 months showed proportion published = 80.8% (412/510 abstract) Survival analysis of proportion published at 60 months = 74% Factors related to proportion of abstracts published included 195/223 'positive' (defined as significant results) versus 220/287 not 'positive' abstract results published 160/183 'positive' (defined as experimental better than control) versus 255/327 not 'positive' abstract results published 232/278 abstracts presented orally versus 100/126 abstracts presented as posters published 364/443 abstracts with multiple centers versus 51/67 abstracts with a single center published
Notes	Oncology Funding by Cancer Care Ontario Research Fellowship
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific study design and specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting year, positive results, type of presentation, multi‐center status, sample size, and subspecialty with publication and for time to publication using Cox proportional models.

Kumar 1995.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to September 1994 Person completing the search not reported Searched by first author and one other author ‐ unspecified Matching criteria not reported
Data	Included 275 abstracts presented at the 1992 American Society for Gastrointestinal Endoscopy meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication by presentation format Cumulative proportion of abstracts published Oral versus poster presentation versus neither
Outcomes	76 of 275 abstracts published Proportion of abstracts published by time Mean time to publication for oral presentations = 13 months (range = 6 to 24 months) Mean time to publication for poster presentations = 15 months (range = 4 to 24 months) Mean time to publication for other presentation format = 13 months (range = 5 to 26 months) Cumulative proportion of abstracts published at 12 months showed proportion published = 9.5% (26/275 abstracts) Factors related to proportion of abstracts published included 15/46 abstracts presented orally versus 51/150 abstracts presented as posters versus 10/79 abstracts presented in other format published
Notes	Gastroenterology ‐ endoscopy Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting only had 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis

Kunadian 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE, Web of Science until September 1, 2014 Search completed by investigator Searched by first and senior author, keywords, and title Matching criteria not reported
Data	Included 2119 abstracts presented at the 2002 to 2012 British Cardiovascular Society meetings Included all abstracts except duplicates, and those where author's primary institution was not based in UK
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published Basic science versus epidemiological/ clinical abstracts Subspecialty
Outcomes	1401 of 2199 abstracts published Proportion of abstracts published by time Median time to publication = 16.8 months (SD = 19.2) Cumulative proportion of abstracts published at 90 months showed proportion published = 63% (1335/2119 abstracts) Factors related to proportion of abstracts published included Basic science versus epidemiological/clinical abstracts: OR 2.35 (95% CI 1.73‐3.19)
Notes	Cardiology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	No	Reported that all abstracts were included, but data were inconsistent within the report.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings except the 2011 and 2012 meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type science, and subspecialty with publication using stratified analysis and Chi2 tests.

Kwong 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, PubMed; dates of search not reported Person completing the search not reported Searched by first, second, and last author, and keywords Matched abstract to full‐length publication by Methodology Results
Data	Included 278 abstracts presented at the 1999 and 2001 European Federation of National Associations of Orthopaedics and Traumatology meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published Clinical research versus basic science research versus other RCT design versus observational design versus case series design versus comparative study design versus systematic reviews Grecian versus United Kingdom versus German versus Belgium versus rest of the world origin Subspecialty
Outcomes	112 of 278 abstracts published Proportion of abstracts published by time Mean time to publication = 25.3 months (SD =18.5) Cumulative proportion of abstracts published at 84 months showed proportion published = 40.3% (112/278 abstracts) Factors related to proportion of abstracts published included 92/239 abstracts describing clinical research versus 13/24 abstracts describing basic science research versus 7/15 abstracts with other study design published 20/25 abstracts with RCT design versus 3/5 abstracts with observational design versus 58/175 case series versus 11/34 abstracts with comparative study design versus 2/3 systematic reviews published 11/52 abstracts originating from Greece versus 24/51 abstracts from the UK versus 13/22 abstracts from Germany versus 10/19 abstracts from Belgium versus 54/134 abstracts from elsewhere published
Notes	Emergency medicine ‐ trauma Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of type science, study design, country of origin, and subspecialty with publication using stratified analysis.

Landry 1996.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from January 1990 through November 1994 Person completing the search not reported Searched by 'principal' author and keywords Matched abstract to full‐length publication by 'Principal author' Keywords Contents
Data	Included 168 abstracts presented at the 1990 American Burn Association meeting Included all abstracts
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Clinical research versus basic science research
Outcomes	44 of 168 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 24/58 'positive' (defined as significant results) versus 20/110 not 'positive' abstract results published 16/65 abstracts describing clinical research versus 15/63 abstracts describing basic science research published
Notes	Emergency medicine ‐ trauma, burns Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of positive results, type of science, university affiliation, doctoral degree, use of statistics, and clarity of statistics with publication using stratified analysis and Chi2 tests.

Larian 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Melvyl MEDLINE Plus to 1 July 1999 Person completing the search not reported Searched by first, second, and last author, and keywords Matched abstract to full‐length publication by Hypothesis Study design Protocol Sample size Results
Data	Included 839 abstracts presented at the 1993, 1994, and 1995 American Academy of Otolaryngology ‐ Head & Neck Surgery meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting Clinical research versus basic science research Subspecialty
Outcomes	270 of 839 abstracts published 83/249 abstracts presented at the 1993 meeting, 103/293 at the 1994 meeting, and 84/297 at the 1995 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published for 1995 American Academy of Otolaryngology ‐ Head & Neck Surgery meeting at 36 months showed proportion published = 31.6% (94/297 abstracts) Cumulative proportion of abstracts published for 1994 American Academy of Otolaryngology ‐ Head & Neck Surgery meeting at 48 months showed proportion published = 35.2% (103/293 abstracts) Cumulative proportion of abstracts published for 1993 American Academy of Otolaryngology ‐ Head & Neck Surgery meeting at 48 months showed proportion published = 33.3% (83/249 abstracts) Factors related to proportion of abstracts published included 131/370 abstracts describing clinical research versus 139/440 abstracts describing basic science research published
Notes	Otolaryngology head and neck surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type science, and subspecialty with publication using stratified analysis and Chi2 tests.

Lau 2016.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Keywords Contents
Data	Included 460 abstracts presented at the 1996 to 2013 Otorhinolaryngology Research Society meetings Included all abstracts presented orally except those of six meetings (unspecified) for which meeting abstracts could not be retrieved
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Clinical research versus laboratory research versus surveys/reviews Case report design versus other research design Subspecialty
Outcomes	259 of 460 abstracts published Proportion of abstracts published by time Mean time to publication = 27.7 months Median time to publication = 23 months Cumulative proportion of abstracts published at 60 months showed proportion published = 51.7% (238/460 abstracts) Factors related to proportion of abstracts published included 146/246 abstracts describing clinical research versus 100/191 abstracts describing laboratory research versus 13/23 abstracts describing surveys/reviews published 7/8 abstracts describing case reports versus 252/452 abstracts describing a different study design published
Notes	Otolaryngology head and neck surgery Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of type science, study design, presence of a professor on the team, and subspecialty with publication using stratified analysis and Chi2 , Mann‐Whitney U, or Kruskal Wallis one‐way ANOVA tests.

Lee 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by first, second, and last author, and keywords Matching criteria not reported
Data	Included 346 abstracts presented at the 2004 and 2005 International Association for Dental Research meetings Included all abstracts in the prosthodontics section; excluded 23 withdrawn abstracts and 2 keynote speaker abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation Clinical research versus laboratory science research North American versus European versus rest of the world origin Industry finding versus government funding versus foundation funding versus university funding versus no funding Subspecialty
Outcomes	128 of 346 abstracts published Proportion of abstracts published by time Mean time to publication = 26.4 months (range = 0 to 67 months) Cumulative proportion of abstracts published at 48 months showed proportion published = 28.6% (99/346 abstracts) Factors related to proportion of abstracts published included 71/218 'positive' (defined as experimental better than control) versus 57/128 not 'positive' abstract results published 35/86 abstracts presented orally versus 93/260 abstracts presented as posters published 34/101 abstracts describing clinical research versus 94/245 abstracts describing laboratory science research published 45/140 abstracts originating from North America versus 55/108 abstracts from Europe versus 51/149 abstracts from rest of the world published 17/35 abstracts with funding from industry versus 14/26 abstracts with funding from government versus 4/11 abstracts with funding from foundations versus 1/4 abstracts with funding from universities versus 92/270 abstracts with no funding published
Notes	Oral health ‐ prosthodontics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic with reasonable exceptions, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, type of science, country of origin, funding status, and subspecialty with publication using multivariable logistic regression analyses.

Leles 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2004 Search completed 'independent examiner' Searched by all authors Matching criteria not reported
Data	Included 775 abstracts submitted to the 1999 Sociedade Brasileira de Pesquisa Odontológigica meeting Included all abstracts
Comparisons	Proportion of abstracts published
Outcomes	116 of 775 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Oral health ‐ odontology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 ears follow‐up.
Matching?	Unclear	Matching criteria not reported.

Lensen 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, Embase, MEDLINE, PubMed, and the Cochrane Menstrual Disorders and Subfertility Group specialized register between December 2014 and May 2015 Search completed by the investigators Search criteria not reported Matching criteria not reported
Data	Included 230 abstracts from conferences from 2007 to 2010 and included in the Cochrane Menstrual Disorders and Subfertility Group specialized register Included all abstracts on subfertility RCTs reporting ≥1 reproductive outcome
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation North American versus European versus Asian versus South American versus Middle Eastern versus rest of the world origin Trial registration status versus not Industry funding versus other/no funding
Outcomes	117 of 230 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 38/80 'positive' (defined as 'significant results') versus 79/150 not 'positive' abstract results published (OR 1.16, 95% CI 0.66‐2.02) 54/115 'positive' (defines as "experimental better than control") versus 63/115 abstracts with not 'positive' abstract results published 55/80 abstracts presented orally versus 56/141 abstracts presented as posters published 23/41 abstracts originating from North America versus 49/87 abstracts originating from Europe versus 10/35 abstracts originating from Asia versus 6/22 abstracts originating from South America versus 26/40 abstracts originating from the Middle East versus 3/5 abstracts with rest of the world origin published 6/8 abstracts registered in a trials registry versus 111/227 abstracts not registered published 17/24 abstracts with industry funding versus 100/206 abstracts with other/no funding published
Notes	Human reproduction and development ‐ subfertility Funded by the University of Auckland Data extracted from abstract and poster
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific subgroup.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Yes	All meetings had at least 60 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results, type of presentation, country of origin, registration in a trials register, and funding status with publication stratified analysis and Chi2 tests.

Levett 2000.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for 'three year period' Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 790 abstracts presented at the 1980 to 1990 Caribbean Health Research Council Annual meetings Included all abstracts
Comparisons	Proportion of abstracts published Oral versus poster presentation
Outcomes	263 of 790 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 194/525 abstracts presented orally versus 69/265 abstracts presented as posters published
Notes	Other non‐clinical specialties ‐ health research Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meetings only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis.

Li 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to 'Fall 2003' Search completed by investigator Searched by first author and keywords Matched abstract to full‐length publication by One author Title Institution Time period
Data	Included 2054 abstracts presented at the 1997 and 1999 to 2001 Society for Academic Emergency Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting Oral versus poster presentation Subspecialty
Outcomes	781 of 2054 abstracts published 217/542 abstracts presented at the 1997 meeting, 214/536 at the 1999 meeting, 190/501 at the 2000 meeting, and 166/475 at the 2001 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 48 months showed proportion published = 40.0% (217/542 abstracts) from 1997 Society for Academic Emergency Medicine meeting Cumulative proportion of abstracts published at 48 months showed proportion published = 39.0% (209/536 abstracts) from 1999 Society for Academic Emergency Medicine meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 38.0% (190/501 abstracts) from 2000 Society for Academic Emergency Medicine meeting Cumulative proportion of abstracts published at 24 months showed proportion published = 34.0% (162/475 abstracts) from 2001 Society for Academic Emergency Medicine meeting Factors related to proportion of abstracts published included 305/610 abstracts presented orally versus 491/1444 abstracts presented as posters published
Notes	Emergency medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2000 had at least 48 months follow‐up. The meetings in 2000 and 2001 only had 36 and 24 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type of presentation, and subspecialty with publication using stratified analysis.

Lim 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and KoreaMed from January 2008 to June 2012 Search completed by the investigators Searched by all authors, keywords, title, and institutions Matched abstract to full‐length publication by Results Study design
Data	Included 1027 abstracts presented at the 2008 and 2009 Korean Academy of Rehabilitation Medicine Spring and Fall meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication, overall and by meeting Cumulative proportion of abstracts published Subspecialty
Outcomes	317 of 1027 abstracts published; 35.7% for Spring 2008 meeting, 32.4% for Fall 2008 meeting, 28.0% for Spring 2009 meeting, and 29.4% for Fall 2009 meeting. Proportion of abstracts published by time Mean time to publication = 17.2 months (SD = 10.5), overall Mean time to publication = 15.5 months (SD = 12.2) for Spring 2008 Korean Academy of Rehabilitation Medicine meeting Mean time to publication = 19.8 months (SD = 11.3) for Fall 2008 Korean Academy of Rehabilitation Medicine meeting Mean time to publication = 16.6 months (SD = 10.3) for Spring 2009 Korean Academy of Rehabilitation Medicine meeting Mean time to publication = 15.3 months (SD = 8.8) for Fall 2009 Korean Academy of Rehabilitation Medicine meeting Cumulative proportion of abstracts published at 48 months showed proportion published = 30.9% (317/1027 abstracts) No factors other than subspecialty related to proportion of abstracts published reported
Notes	Physical and rehabilitation medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The meeting in 2008 had 48 months follow‐up, the meeting in 2009 only had 36 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Unclear	Examined association of meeting year and subspecialty with publication using stratified analysis and logistic regression, although unclear whether multivariable logistic regression used.

Lin 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, China National Knowledge Infrastructure, and wanfangdata.com until 31 July, 2010 Person completing the search not reported Searched by first author, keywords, and institution of first author Matching criteria not reported
Data	Included 25 abstracts presented at the 2003 to 2008 American Heart Rhythm Society meetings Included all abstracts from Chinese Mainland ‐ extracted data on subgroup of original sample which was from 2003 to 2010
Comparisons	Proportion of abstracts published
Outcomes	10 of 25 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Cardiology ‐ arrhythmias Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific subgroup.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings except the 2007 and 2008 meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Liu 1996.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Search completed by investigator Search criteria not reported Matching criteria not reported
Data	Included 400 abstracts presented at the 1992 American Heart Association, Federation of American Societies in Experimental Biology, American Gastroenterology Association, and American Academy of Neurology meetings Included 400 abstracts, 100 randomly selected from each meeting
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting
Outcomes	141 of 400 abstracts published 42/100 abstracts presented at the American Academy of Neurology meeting, 42/100 at the American Gastroenterology Association meeting, 28/100 at the Federation of American Societies in Experimental Biology meeting, and 30/100 at the American Heart Association meeting published Proportion of abstracts published by time Mean time to publication = 22.8 months (SD = 10.8) Mean time to publication = 22.8 months (SD = 9.6) for 1992 American Heart Association meeting Mean time to publication = 19.2 months (SD = 10.8) for 1992 Federation of American Societies in Experimental Biology meeting Mean time to publication = 26.4 months (SD = 10.8) for 1992 American Gastroenterology Association meeting Mean time to publication = 22.8 months (SD = 10.8) for 1992 American Academy of Neurology meeting No factors related to proportion of abstracts published reported
Notes	Cardiology ‐ circulation, basic biomedical science, gastroenterology, neurology Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

Livas 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar in April 2013 Search completed by the investigators Searched by first, second, and last author, keywords, and title Matched abstract to full‐length publication by All authors Title Study design Conclusions
Data	Included 590 abstracts presented at the 2006 and 2007 European Orthodontic Society meetings Included all abstracts except keynote lectures
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting Cumulative proportion of abstracts published 'Positive' versus not 'positive' versus no statistical testing Oral versus poster presentation European versus Asian versus rest of the world origin Academic affiliation versus public health service affiliation versus other governmental affiliation versus private practice affiliation versus private company affiliation Subspecialty
Outcomes	308 of 590 abstracts published 168/333 abstracts presented at the 2006 meeting, and 140/257 at the 2007 meeting published Proportion of abstracts published by time Mean time to publication = 15.8 months, overall Mean time to publication = 17.2 months for 2006 European Orthodontic Society meeting Mean time to publication = 14.6 months for 2007 European Orthodontic Society meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 51.4% (303/590 abstracts) Factors related to proportion of abstracts published included 201/340 'positive' (not defined) versus 81/145 not 'positive' abstract results versus 26/105 abstracts without statistical testing published 115/162 abstracts presented orally versus 193/428 abstracts presented as poster published 203/390 abstracts originating from Europe versus 88/172 abstracts originating from Asia versus 17/28 abstracts from the rest of the world published 286/545 abstracts originating in an academic center versus 5/8 abstracts originating from a public health service versus 8/13 abstracts originating from other governmental institution versus 6/21 abstracts originating from private practice versus 3/3 originating from a private company published
Notes	Oral health ‐ orthodontics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, positive results, type of presentation, country of origin, academic affiliation, and subspecialty with publication using multivariable logistic regression analyses.

Lloyd 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported, but senior author "made the final determination to the accuracy and inclusion or exclusion" Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by 'Focus'
Data	Included 292 abstracts presented at the 1998 to 2001 American Association of Hip and Knee Surgeons meetings Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication overall and by meeting Cumulative proportion of abstracts published
Outcomes	168 of 292 abstracts published 25/40 abstracts presented at the 1998 meeting, 29/53 at the 1999 meeting, 30/49 at the 2000 meeting, and 33/61 at the 2001 meeting published Proportion of abstracts published by time Mean time to publication = 21.7 months (SD = 14.7) Mean time to publication = 21 months for 1998 American Association of Hip and Knee Surgeons meeting Mean time to publication = 22 months for 1999 American Association of Hip and Knee Surgeons meeting Mean time to publication = 22 months for 2000 American Association of Hip and Knee Surgeons meeting Mean time to publication = 16 months for 2001 American Association of Hip and Knee Surgeons meeting Cumulative proportion of abstracts published at 72 months showed proportion published = 57.2% (167/292 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery ‐ hip and knee surgery Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted as oral presentations.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Loevy 1997.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLARS to December 1994 Search completed by investigator Search criteria not reported Matched abstract to full‐length publication by Title Results
Data	Included 189 abstracts presented at the 1989 and 1990 American Academy of Pediatric Dentistry meetings Included all abstracts by American Academy of Pediatric Dentistry members
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published Subspecialty
Outcomes	87 of 189 abstracts published 51/106 abstracts presented at the 1989 meeting, and 36/83 at the 1990 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 24 months showed proportion published = 26.5% (50/189 abstracts) Reported proportion of abstracts published by subspecialty
Notes	Oral health ‐ pediatric dentistry Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year by subspecialty with publication using stratified analysis.

Macdonald 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and PubMed; dates of search not reported Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Methodology Conclusions
Data	Included 862 abstracts presented at the 1999 to 2008 British Association of Paediatric Surgeons meetings Included all abstracts presented orally except those from the 2009 joint congress with the European Paediatric Surgeons Association
Comparisons	Proportion of abstracts published
Outcomes	302 of 862 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Surgery ‐ pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally with reasonable exception.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Yes	Matched by 4 different criteria.

Macmillan 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to December 2005 Person completing the search not reported Searched by first, and last author, and keywords Matching criteria not reported
Data	Included 404 abstracts presented at the 2001 and 2002 British Association of Emergency Medicine and Faculty of Accident and Emergency Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published Oral versus poster presentation
Outcomes	124 of 404 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 83/145 abstracts presented orally versus 41/259 abstracts presented as posters published
Notes	Emergency medicine Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The 2001 meeting had 48 months follow‐up, the 2002 meeting only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis.

Maguire 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to March 2013 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 577 abstracts presented at the 2004 to 2010 Reproductive Health Annual meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	335 of 577 abstracts published Proportion of abstracts published by time Mean time to publication = 18.5 months (range = ‐6 to 78 months) No factors related to proportion of abstracts published reported
Notes	Human reproduction and development ‐ reproductive medicine Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings except the 2010 meeting had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Maleck 1998a.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and MEDLINE PLUS from 1993 to 1996 Search completed by investigator Searched by all authors and title Matched abstract to full‐length publication by Title Contents
Data	Included 98 abstracts presented at the 1993 and 1994 Prehospital Care Research Forum meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation
Outcomes	10 of 98 abstracts published 5/64 abstracts presented at the 1993 meeting, and 5/34 at the 1994 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 4/16 abstracts presented orally versus 3/34 abstracts presented as poster published
Notes	Emergency medicine ‐ pre hospital care Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	No	All meetings had less than 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year by type presentation with publication using stratified analysis.

Maleck 1998b.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1994 to June 1998 and PubMed from May 1993 to June 1998 Search completed by investigator Searched by all authors and title Matched abstract to full‐length publication by Title Contents
Data	Included 109 abstracts presented at the 1994 Pan‐European Conference on Emergency Medical Systems meeting Included all abstracts published in conference book
Comparisons	Proportion of abstracts published
Outcomes	11 of 109 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Emergency medicine Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria

Malicki 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Web of Science, and Peer Review and Biomedical Publication website through August 20, 2013 Search completed by the investigators Searched by all authors Matching criteria not reported
Data	Included 504 abstracts presented at the 1989, 1993, 1997, 2001, 2005, and 2009 International Congress on Peer Review and Biomedical Publication meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication RCT design versus observational design versus other intervention design versus opinion piece
Outcomes	305 of 504 abstracts published 34/45 abstracts presented at the 1989 meeting, 41/56 at the 1993 meeting, 50/93 at the 1997 meeting, 58/106 at the 2001 meeting, 63/93 at the 2005 meeting, and 59/111 at the 2009 meeting published Proportion of abstracts published by time Median time to publication = 14 months (95% CI = 12 to 16 months) Factors related to proportion of abstracts published included 22/27 abstracts with RCT design versus 239/383 abstracts with observational design versus 25/52 abstracts with other intervention design versus 19/40 abstracts describing an opinion piece published
Notes	Medical education/library science ‐ peer review and biomedical publication Funded by the Esteve Foundation Data abstracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and study design with publication using stratified analysis and Chi2 tests.

Manuck 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed during 2014 and 2015 Search completed by the investigators Search criteria not reported Matched abstract to full‐length publication by Title Methodology Results First author of abstract
Data	Included 3281 abstracts presented at the 2003 to 2010 Society for Maternal‐Fetal Medicine meetings Included all oral abstracts for all meetings and poster abstracts only for meetings in 2003, 2005, 2007, and 2009
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication, overall and by type of presentation Cumulative proportion of abstracts published, overall and by type of presentation Oral versus poster presentation Plenary oral versus fellow plenary oral versus other oral presentation Award winning versus not award winning First author affiliation with US versus non‐US
Outcomes	1780 of 3281 abstracts published 365/656 abstracts presented at the 2003 meeting, 55/68 at the 2004 meeting, 407/684 at the 2005 meeting, 64/85 at the 2006 meeting, 389/792 at the 2007 meeting, 59/82 at the 2008 meeting, 379/828 at the 2009 meeting, and 58/86 at the 2010 meeting published Proportion of abstracts published by time Median time to publication = 19 months (IQR = 9 to 36 months), overall Median time to publication = 11 months (IQR = 9 to 24 months) for abstracts presented orally Median time to publication = 21 months (IQR = 11 to 40 months) for abstracts presented as posters Cumulative proportion of abstracts published at 120 months showed proportion published = 54.2% (1779/3281 abstracts), overall Cumulative proportion of abstracts published at 120 months showed proportion published = 76.9% (484/629 abstracts) for abstracts presented orally Cumulative proportion of abstracts published at 120 months showed proportion published = 48.8% (1295/2652 abstracts) for abstracts presented as posters Factors related to proportion of abstracts published included 485/629 abstracts presented orally versus 1295/2652 abstracts presented as poster published 57/63 abstracts presented in oral plenary session versus 51/64 abstracts presented in fellows oral plenary session versus 377/502 abstracts presented in other oral sessions published 86/104 award winning abstracts versus 1694/3177 non‐award winning abstracts published 528/815 abstracts with non‐US first author affiliation versus 1260/2466 abstracts with US first author affiliation published
Notes	Gynecology/obstetrics ‐ maternal‐fetal medicine Funded by the Eunice Kennedy Shriver National Institute of Child Health and Human Development and National Center for Advancing Translational Sciences/ National Institutes of Health, Center for Clinical and Translational Sciences
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for presentation and all poster presentations for pre‐specified years.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, type of presentation, type of oral presentation, award winning status, and US first author with publication using multivariable regression analyses.

Martinez 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and LILACS; dates of search not reported Person completing the search not reported Searched by all authors, keywords, and titles Matching criteria not reported
Data	Included 388 abstracts presented at the 1997, 1999, 2001, 2003, and 2005 Confederación Unificada Bioquímica de la República Argentina meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Survival analysis of publication rate by meeting Multi‐centered versus single center Academic center versus hospitals versus collaborative group versus other institution Subspecialty
Outcomes	42 of 388 abstracts published 11/77 abstracts presented at the 1997 meeting, 11/72 at the 1999 meeting, 6/59 at the 2001 meeting, 10/75 at the 2003 meeting, and 4/105 at the 2005 meeting published Proportion of abstracts published by time Mean time to publication = 24 months (SD = 12; maximum = 48 months) Survival analysis of proportion published at 120 months = 14.3% for 1997 Confederación Unificada Bioquímica de la República Argentina meeting Survival analysis of proportion published at 96 months = 18.1% for 1999 Confederación Unificada Bioquímica de la República Argentina meeting Survival analysis of proportion published at 72 months = 10.2% for 2001 Confederación Unificada Bioquímica de la República Argentina meeting Survival analysis of proportion published at 48 months = 14.7% for 2003 Confederación Unificada Bioquímica de la República Argentina meeting Survival analysis of proportion published at 24 months = 4.8% for 2005 Confederación Unificada Bioquímica de la República Argentina meeting Factors related to proportion of abstracts published included 17/92 abstracts with multiple centers versus 25/196 abstracts with a single center published 13/133 abstracts originating in an academic center versus 4/105 abstracts originating from hospitals versus 17/92 abstracts originating within collaborative groups versus 8/58 abstracts originating from other institutions published
Notes	Other non‐clinical science ‐ biochemistry Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, multi‐center status, academic affiliation and subspecialty with publication using stratified analysis and time to publication using survival analysis.

Marx 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1993 to 1997 Search completed by investigator Searched by 'senior' author Matched abstract to full‐length publication by One author Keywords
Data	Included 527 abstracts presented at the 1993 American Society of Neuroradiologists and Radiological Society of North America meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published Sample size equal to or above the mean versus sample size below the mean Clinical research versus basic science research Subspecialty
Outcomes	194 of 527 abstracts published 8/25 abstracts presented at the Radiological Society of North America meeting, and 186/502 presented at the American Society of Neuroradiologists meeting published Proportion of abstracts published by time Mean time to publication = 15 months Cumulative proportion of abstracts published at 45 months showed proportion published = 35.9% (189/527 abstracts) Factors related to proportion of abstracts published included Proportions of abstracts published with sample size above and below the mean by meeting but did not report actual numbers of abstracts Proportions of abstracts published describing clinical research versus basic science by meeting but did not report actual numbers of abstracts
Notes	Radiology ‐ neuroradiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting, sample size, type science, prospective status, and subspecialty with publication using stratified analysis and Chi2 tests.

Maxwell 1981.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLARS to March 1981 Person completing the search not reported Searched by first and second author Matching criteria not reported Contacted abstract authors directly Contacted editor of major nursing journals Handsearched Index to International Nursing Literature
Data	Included 65 abstracts presented at the 1977 and 1978 Oncology Nursing Society meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication
Outcomes	28 of 65 abstracts published 13/23 abstracts presented at the 1977 meeting, and 15/42 at the 1978 meeting published Proportion of abstracts published by time Median time to publication = 11.5 months (range = 1 to 30 months) No factors other than meeting year related to proportion of abstracts published reported
Notes	Nursing ‐ oncology nursing Funding not reported Authors presented work for an additional 106 abstracts from 1979 and 1981 meetings, but followed for < 2 years and not included here
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database, and contacted abstract authors, editors, and performed handsearches.
Follow‐up time?	Yes	The 1977 meeting had 48 months follow‐up. The 1978 meeting only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

McCormick 1985.

Methods	Identification of subsequent full‐length publications Searched database Index Medicus to June 1983 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 1238 abstracts submitted to the 1976 to 1980 Ambulatory Pediatric Association and Society for Pediatric Research in American Pediatric Society meetings, 355 accepted for presentation and 883 rejected Included random sample of abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting, year, and acceptance for presentation to the meeting Cumulative proportion of abstracts published by meeting and year Accepted versus rejected for conference presentation
Outcomes	330 of 1238 of submitted abstracts published 49/371 abstracts presented at the 1976 to 1978 Ambulatory Pediatric Association meetings, 48/347 at the 1976 to 1978 Society for Pediatric Research in American Pediatric Society meetings, 40/194 at the 1979 to 1980 Ambulatory Pediatric Association meetings, and 35/326 at the 1979 to 1980 Society for Pediatric Research in American Pediatric Society meetings published Proportion of abstracts published by time Mean time to publication = 20 months for abstracts accepted for presentation at the 1976 to 1978 Ambulatory Pediatric Association meeting Mean time to publication = 12.9 months for abstracts accepted for presentation at the 1979 to 1980 Ambulatory Pediatric Association meeting Mean time to publication = 23.1 months for abstracts accepted for presentation at the 1976 to 1978 Society for Pediatric Research Society meeting Mean time to publication = 16.7 months for abstracts accepted for presentation at the 1979 to 1980 Society for Pediatric Research Society meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 47.0% (47/100 abstracts) for abstracts accepted for presentation at the 1976 to 1978 Ambulatory Pediatric Association meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 45.3% (39/86 abstracts) for abstracts accepted for presentation at the 1979 to 1980 Ambulatory Pediatric Association meeting Cumulative proportion of abstracts published at 28 months showed proportion published = 47.2% (42/89 abstracts) for abstracts accepted for presentation at the 1976 to 1978 Society for Pediatric Research Society meeting Cumulative proportion of abstracts published at 30 months showed proportion published = 42.5% (34/80 abstracts) for abstracts accepted for presentation at the 1979 to 1980 Society for Pediatric Research Society meeting Factors related to proportion of abstracts published included 172/355 abstracts accepted for presentation versus 158/883 rejected abstracts published
Notes	Pediatrics Funding in part by the Ambulatory Pediatric Association and Andrew W. Mellon Foundation
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 1983 had at least 48 months follow‐up. The meeting in 1983 only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year by acceptance for presentation with publication using stratified analysis and Chi2, Wilcoxon rank‐sum, or Kruskal‐Wallis tests.

McCue 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors, keywords, and title Matching criteria not reported
Data	Included 672 abstracts presented at the 1997 to 2000 American Society of Colon and Rectal Surgeons meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Oral versus poster presentation US versus UK versus Korea versus elsewhere origin
Outcomes	400 of 672 abstracts published Proportion of abstracts published by time Mean time to publication = 18.4 months Factors related to proportion of abstracts published included 228/321 abstracts presented orally versus 172/351 abstracts presented as posters published 208/365 abstracts originating in the US versus 54/102 abstracts originating in the UK versus 7/21 abstracts originating in Korea versus 131/184 abstracts originating elsewhere published
Notes	Surgery ‐ colon and rectal surgery Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation and US origin with publication using stratified analysis.

McKelvey 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, International Pharmaceutical Abstracts to March 2008 Search completed by investigator Searched by all authors, keywords, and relevant MeSH term, if available Matched abstract to full‐length publication by One author Project objective
Data	Included 272 abstracts presented at the 1981, 1991, and 2001 Southeastern Residency Conference meetings Included only abstracts of pharmacy resident projects
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication overall and by meeting RCT versus CCT versus observational study versus case report versus topic review design Physician co‐author versus not Reporting results versus not
Outcomes	43 of 272 abstracts published 18/90 abstracts presented at the 1981 meeting, 11/70 at the 1991 meeting, and 14/112 at the 2001 meeting published Proportion of abstracts published by time Median time to publication = 22.8 months (maximum = 60.9 months) Median time to publication = 22 months (IQR = 13 to 25 months) for 1981 Southeastern Residency Conference meeting Median time to publication = 33 months (IQR = 22 to 39 months) for 1991 Southeastern Residency Conference meeting Median time to publication = 24 months (IQR = 20 to 28 months) for 2001 Southeastern Residency Conference meeting Factors related to proportion of abstracts published included 12/48 abstracts with RCT design versus 9/71 abstracts with CCT design versus 15/78 abstracts with observational study design versus 6/30 abstracts with case report design versus 1/45 abstracts of topic review published 13/52 abstracts with a physician co‐author listed versus 30/220 abstracts without a physician co‐author or not reported published 18/67 abstracts reporting results versus 25/205 abstracts not reporting results published
Notes	Pharmacology ‐ residents Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 7 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, study design, physician co‐author, and report of results in abstract with publication using stratified analysis. Stated that numbers were too small to perform statistical testing.

McKinley 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to November 2009 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title
Data	Included 115 abstracts presented at the 2000, 2003, and 2006 American Contact Dermatitis Society meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published
Outcomes	64 of 115 abstracts published Proportion of abstracts published by time Mean time to publication = 21.6 months (range = 3 to 72 months) Cumulative proportion of abstracts published at 66 months showed proportion published = 42.6% (49/115 abstracts) No factors related to proportion of abstracts published reported
Notes	Dermatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2006 had at least 48 months follow‐up. The meeting in 2006 only had 36 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.

McLennan 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by all authors, keywords, and title Matching criteria not reported
Data	Included 92 abstracts presented at the 2002 to 2004 American Urogynecological Society meetings Included all abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published, overall and by meeting Industry funding versus no industry funding
Outcomes	68 of 92 abstracts published 18/23 abstracts presented at the 2002 meeting, 25/39 at the 2003 meeting, and 25/30 at the 2004 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 7/11 abstracts with industry funding versus 61/81 abstracts without industry funding published
Notes	Gynecology/obstetrics ‐ urogynecology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and funding status with publication using stratified analysis and Chi2 tests.

Meininger 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported, but reported five year follow‐up Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by Methodology Results
Data	Included 2207 abstracts presented at the 2000 and 2005 German Congress of Anesthesia and European Congress of Anesthesia meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting
Outcomes	813 of 2207 abstracts published 218/644 abstracts presented at the 2000 European Congress of Anesthesia meeting, 183/465 at the 2000 German Congress of Anesthesia meeting, 233/720 at the 2005 European Congress of Anesthesia meeting, and 179/378 at the 2005 German Congress of Anesthesia meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 24 months showed proportion published = 28.6% (184/644 abstracts) for 2000 European Congress of Anesthesia Cumulative proportion of abstracts published at 24 months showed proportion published = 32.0% (149/465 abstracts) for 2000 German Congress of Anesthesia Cumulative proportion of abstracts published at 24 months showed proportion published = 24.9% (179/720 abstracts) for 2005 European Congress of Anesthesia Cumulative proportion of abstracts published at 24 months showed proportion published = 35.7% (135/378 abstracts) for 2005 German Congress of Anesthesia No factors other than meeting year related to proportion of abstracts published reported
Notes	Anesthesiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meetings had Quote: "a five year follow‐up."
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Meissner 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed between June 10 and July 4, 2013 Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by One author Hypothesis Conclusion
Data	Included 172 abstracts presented at the 2007 and 2010 Canadian Association of Radiation Oncology meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication Cumulative proportion of abstracts published by meeting Clinical research versus basic science research versus technical research versus other research Subspecialty
Outcomes	88 of 172 abstracts published 50/102 abstracts presented at the 2007 meeting, and 38/70 at the 2010 meeting published Proportion of abstracts published by time Mean time to publication = 18.5 months Median time to publication = 16.5 months (range = ‐10 to 54 months) Cumulative proportion of abstracts published at 26 months showed proportion published = 38.4% (66/172 abstracts) Factors related to proportion of abstracts published included 55/91 abstracts describing clinical research versus 5/9 abstracts describing basic science research versus 23/57 abstracts describing technical research versus 5/15 abstracts describing other types of research published
Notes	Oncology ‐ radiation oncology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The follow‐up for both meetings was limited to 36 months.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type science, and subspecialty with publication using stratified analysis and Chi2 or t tests.

Menditto 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from January 2008 to December 2012 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 298 abstracts presented at the 2008 Italian Emergency Medicine meeting Included all abstracts
Comparisons	Proportion of abstracts published Oral versus poster presentation Clinical versus basic science research RCT design versus observational design versus case report design versus narrative study Multi‐centered versus single center 'Good' quality versus not 'good' quality Italian language versus not Italian language Sample size >100 versus < 100 ≥ 1 academic author versus no academic author
Outcomes	43 of 298 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 26/112 abstracts presented orally versus 17/186 abstracts presented as poster published 34/280 abstracts describing clinical research versus 9/18 abstracts describing basic science research published 5/8 abstracts describing RCTs versus 26/91 abstracts describing an observational design versus 3/78 abstracts describing a case report design versus 0/103 abstracts describing a narrative study published 4/21 abstracts with multiple centers versus 39/277 abstracts with a single center published 22/46 'good' quality abstracts versus 12/131 not 'good' quality abstracts published 30/270 abstracts in Italian language versus 13/28 abstracts in another language published 27/86 abstracts with a sample size >100 versus 16/212 abstracts with a sample size <100 published 20/88 abstracts with ≥1 academic author versus 23/210 abstracts with no academic author published
Notes	Emergency medicine Funding not reported Data abstracted from letter to the editor
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type of presentation, type science, study design, multi‐center status, abstract quality, Italian language, sample size, and academic affiliation with publication using stratified analysis and univariate logistic regression.

Meral 2016.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed until May 18, 2015 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Sample size Methodology Results Hypothesis
Data	Included 1368 abstracts presented at the 2008 to 2011 European Society for Surgical Research meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall, by meeting and type of presentation Oral versus poster presentation Clinical versus experimental research Case report design versus review versus clinical study versus experimental study European versus Japanese versus rest of the world origin Very high versus high versus medium/low level of country development Prospective versus retrospective design
Outcomes	559 of 1368 abstracts published 146/341 abstracts presented at the 2008 meeting, 156/304 at the 2009 meeting, 155/476 at the 2010 meeting, and 102/247 at the 2011 meeting published Proportion of abstracts published by time Mean time to publication = 17.5 months (range = ‐166 to 82 months), overall Mean time to publication = 23.2 months (range = ‐59 to 82 months), for the 2008 meeting Mean time to publication = 16.1 months (range = ‐60 to 67 months), for the 2009 meeting Mean time to publication = 15.3 months (range = ‐89 to 55 months), for the 2010 meeting Mean time to publication = 14.8 months (range = ‐166 to 46 months), for the 2011 meeting Mean time to publication = 19.1 months (SD = 20.7), for oral presentations Mean time to publication = 14.0 months (SD = 25.2), for poster presentations Factors related to proportion of abstracts published included 391/803 abstracts presented orally versus 168/565 abstracts presented as poster published 252/737 abstracts describing clinical research versus 279/514 abstracts describing 'experimental' research published 17/87 abstracts with case report design versus 11/29 abstracts describing a review versus 252/737 abstracts describing clinical studies versus 279/514 abstracts describing 'experimental' studies published 417/1051 abstracts originating from Europe versus 33/98 abstracts originating from Japan versus 109/219 abstracts with rest of the world origin published 496/1115 abstracts originating from country with very high development level versus 53/230 abstracts originating from country with high development level versus 10/20 abstracts originating from country with medium/low development level published 436/939 abstracts describing a prospective design versus 123/429 abstracts describing a retrospective design published
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, type science, study design, country of origin, economic status of country of origin, and prospective status with publication using stratified analysis and Chi2 , Fisher's Exact test, t tests, or ANOVA.

Meranze 1982.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLARS for 27 months Search completed by investigator Searched by all authors Matched abstract to full‐length publication by Title Results Dates of study
Data	Included 379 abstracts presented at the 1978 American Society of Anesthesiologists and 1979 International Anesthesia Research Society meetings; also reported on 1980 International Anesthesia Research Society meeting, but follow‐up < 24 months so not included Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting Cumulative proportion of abstracts published by meeting
Outcomes	122 of 379 abstracts published 98/324 abstracts presented at the 1978 meeting, 24/55 at the 1979 meeting published Proportion of abstracts published by time Mean time to publication = 8.2 months for 1979 and 1980 International Anesthesia Research Society meeting, but includes abstracts with < 24 months follow‐up Mean time to publication = 12.5 months for 1978 American Society of Anesthesiologists meeting Cumulative proportion of abstracts published at 24 months showed proportion published = 43.6% (24/55 abstracts) for 1979 International Anesthesia Research Society meeting Cumulative proportion of abstracts published at 27 months showed proportion published = 30.9% (100/324 abstracts) for 1978 American Society of Anesthesiologists meeting No factors other than meeting year related to proportion of abstracts published reported
Notes	Anesthesiology Funding not reported Did not include 62 abstracts from 1980 International Anesthesia Research Society meeting since less than two years of follow‐up
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meetings only had 27 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis

Micieli 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed in July 2011 Person completing the search not reported Searched by all authors, keywords, and title Matching criteria not reported
Data	Included 508 abstracts presented at the 2005 to 2008 Canadian Ophthalmological Society meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published 2 or fewer authors versus more than 2 authors Subspecialty
Outcomes	158 of 508 abstracts published 43/116 abstracts presented at 2005 meeting, 43/111 at the 2006 meeting, 33/134 at the 2007 meeting, and 39/147 at the 2008 meeting published Proportion of abstracts published by time Median time to publication = 15 months (IQR = 8 to 24 months) Survival analysis of abstracts published at 72 months = 34.1% Factors related to proportion of abstracts published included 2 or fewer authors = OR 0.35 (95% CI = 0.22 to 0.55)
Notes	Ophthalmology/optometry ‐ ophthalmology Funding not reported Data abstracted from letter to the editor
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings except the meeting in 2008 had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting year, study design, presence of data in abstract, number of author and subspecialty with publication using multivariable logistic regression analyses.

Miguel‐Dasit 2006a.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from January 1994 to December 2004, Index Medicus Espanol in 2004 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 2992 abstracts presented at the 1994, 1996, and 1998 Spanish Congress of Radiology meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication overall and by meeting Cumulative proportion of abstracts published by meeting Oral versus poster presentation Multi‐centered versus single center
Outcomes	464 of 2992 abstracts published 165/987 abstracts presented at the 1994 meeting, 158/1022 at the 1996 meeting, and 141/983 at the 1998 meeting published Proportion of abstracts published by time Mean time to publication = 32 months (SD = 16, range = 12 to 108 months) Mean time to publication = 35 months (SD = 16, range = 12 to 108 months) for 1994 Spanish Congress of Radiology meeting Mean time to publication = 31 months (SD = 17, range = 12 to 96 months) for 1996 Spanish Congress of Radiology meeting Mean time to publication = 30 months (SD = 14, range = 12 to 72 months) for 1998 Spanish Congress of Radiology meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 16.7% (165/987 abstracts) for 1994 Spanish Congress of Radiology meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 15.5% (158/1022 abstracts) for 1996 Spanish Congress of Radiology meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 14.3% (141/983 abstracts) for 1998 Spanish Congress of Radiology meeting Factors related to proportion of abstracts published included 260/1409 abstracts presented orally versus 204/1583 abstracts presented as poster published 58/216 abstracts with multiple centers versus 324/2112 abstracts with a single center published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type of presentation, and multi‐center status with publication using stratified analysis and Chi2 tests.

Miguel‐Dasit 2006b.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2004 Search completed by investigator Searched by all authors and keywords Matching criteria not reported
Data	Included 1020 abstracts presented at the 2000 European Congress of Radiology meeting Included abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published North American versus European versus rest of the world origin Subspecialty
Outcomes	479 of 1020 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 57 months showed proportion published = 47.0% (479/1020 abstracts) Factors related to proportion of abstracts published included 16/21 abstracts originating from North America versus 354/825 abstracts from Europe versus 59/174 abstracts from rest of the world published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included abstracts accepted for oral presentation.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, country of origin, and subspecialty with publication using stratified analysis and Chi2 Fisher's Exactor Kruskal Wallis tests.

Miguel‐Dasit 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2005 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 991 abstracts presented at the 2001 European Congress of Radiology meeting Included all abstracts accepted as oral presentations except 15 withdrawn abstracts and 10 published prior to meeting
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published North American versus European versus rest of the world origin Subspecialty
Outcomes	449 of 991 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 57 months showed proportion published = 45.3% (449/991 abstracts) Factors related to proportion of abstracts published included 13/18 abstracts originating from North America versus 419/879 abstracts from Europe versus 17/94 abstracts from rest of the world published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted as oral presentations with reasonable exceptions, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of country of origin, and subspecialty with publication using stratified analysis and Chi2 or Fisher's Exact tests.

Mihok 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar; dates of search not reported Person completing the search not reported Searched by first and last author, and keywords Matched abstract to full‐length publication by All authors
Data	Included 602 abstracts presented at 2000 to 2009 British Association for Surgery of the Knee meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	200 of 602 abstracts published Proportion of abstracts published by time Mean time to publication = 23 months (range = 0 to 129 months) No factors related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery ‐ knee surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	No	Matched by only one criterion.
Adjustment for confounding?	Unclear	.

Miller 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, MEDLINE, Web of Science, SCOPUS, and Informit; dates of search not reported Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Contents Contacted abstract author directly
Data	Included 648 abstracts presented at 2006 to 2009 Dietetic Association of Australia meetings Included all abstracts except those of plenaries and invited speakers
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation Clinical research versus community/public health research versus food service/supply/industry domain versus workforce domain versus teaching domain Observational design versus experimental design versus review versus qualitative design versus descriptive design versus audit versus other design > 1 author with academic affiliation versus no author with an academic affiliation Australian versus rest of the world origin Practice domain
Outcomes	165 of 648 abstracts published 43/151 abstracts presented at the 2006 meeting, 41/167 at the 2007 meeting, 50/160 at the 2008 meeting, and 31/170 at the 2009 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 94/275 abstracts presented orally versus 71/373 abstracts presented as poster published 75/264 abstracts describing clinical research versus 60/228 abstracts describing the community/public health domain versus 16/59 abstracts describing the food service/supply/industry domain versus 11/66 abstracts describing the workforce domain versus 3/31 abstracts describing the teaching domain published 96/262 abstracts with observational design versus 30/122 abstracts with experimental design versus 5/26 abstracts describing a review versus 15/60 abstracts describing a qualitative design versus 3/27 abstracts describing an audit versus 16/151 abstracts describing another design published 80/164 abstracts with at least one author with an academic affiliation versus 84/164 abstracts with no authors with an academic affiliation published 155/619 abstracts originating from Australia versus 10/29 abstracts originating from rest of the world published
Notes	Endocrinology/nutrition Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 5 databases and contacted abstract authors.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	No	Matched by only one criterion.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, type science, study design, academic affiliation, author status, Australian origin, Australian state origin, and practice domain with publication using logistic regression analysis.

Mily 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from January 2001 to December 2005 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 128 abstracts presented at 2002 and 2003 European Society of Hypertension meetings Included all abstracts
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Sample size more than 500 versus between 50 and 500 versus less than 50 Oral versus poster presentation International versus European Union country versus non‐European Union country versus non‐European country 'Treatment concealed' versus no 'treatment concealed' published RCT design versus non‐RCT design
Outcomes	34 of 128 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 25/93 'positive' (not defined) versus 9/35 not 'positive' abstract results published 5/19 abstracts with sample size greater than 500 versus 16/47 abstracts with sample size between 50 and 500 versus 13/62 abstracts with sample size below 50 published 6/15 abstracts presented orally versus 28/115 abstracts presented as poster published 15/28 abstracts with an international origin versus 13/57 abstracts originating from a single European Union country versus 2/18 abstracts originating from a single non‐European Union country versus 4/25 abstracts originating from outside Europe published 17/43 abstracts with 'treatment concealed' versus 14/71 abstracts without 'treatment concealed' published 24/65 abstracts with RCT design versus 10/52 abstracts with non‐RCT design published
Notes	General medicine/primary care ‐ hypertension Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results, sample size, type presentation, European origin, study design, and 'treatment concealment' with publication using stratified analysis and unspecified. statistical tests.

Mittal 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and IndMed to January 2011 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 75 abstracts presented at the 1995 to 2007 Indian Academy of Pediatrics meetings Included abstracts of 'award winning' presentations
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published Presented by women versus men 'Descriptive' versus 'experimental' design Academic center versus non‐academic center affiliation
Outcomes	28 of 75 abstracts published Proportion of abstracts published by time Mean time to publication = 23.4 months (SD = 11.8; range = 2 to 77 months) Proportion of abstracts published at 24 months = 26.7% (20/75 abstracts) Factors related to proportion of abstracts published included 13/28 abstracts presented by women versus 15/47 abstracts presented by men published 20/62 abstracts with 'descriptive' design versus 8/13 abstracts with 'experimental' design published 26/54 abstracts originating in an academic center versus 2/21 abstracts not originating in an academic center published
Notes	Pediatrics Reported receiving no funding Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of author gender, type science, academic affiliation, Indian region, and award category with publication using stratified analysis and unspecified statistical tests.

Moar 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for a minimum of 2 years Search completed by the investigators Searched by all authors, keywords, and title Matched abstract to full‐length publication by Title Information/data
Data	Included 318 abstracts presented at the 2000 to 2009 Craniofacial Society of Great Britain and Ireland meetings Included all oral presentations except those of invited keynote lectures
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall Median time to publication, overall and by meeting Cumulative proportion of abstracts published Subspecialty
Outcomes	67 of 318 abstracts published 9/24 abstracts presented at the 2000 meeting, 2/28 at the 2001 meeting, 4/32 at the 2002 meeting, 14/38 at the 2003 meeting, 7/34 at the 2004 meeting, 4/36 at the 2005 meeting, 2/31 at the 2006 meeting, 10/33 at the 2007 meeting, 11/40 at the 2008 meeting, and 5/22 at the 2009 meeting published Proportion of abstracts published by time Mean time to publication = 29 months Median time to publication = 26 months (range = 2 to 110 months), overall Median time to publication = 27 months (range = ‐2 to 110 months) for 2000 Craniofacial Society of Great Britain and Ireland meeting Median time to publication = 23 months (range = ‐2 to 48 months) for 2001 Craniofacial Society of Great Britain and Ireland meeting Median time to publication = 68 months (range = 3 to 110 months) for 2002 Craniofacial Society of Great Britain and Ireland meeting Median time to publication = 24 months (range = ‐1 to 62 months) for 2003 Craniofacial Society of Great Britain and Ireland meeting Median time to publication = 48 months (range = 0 to 59 months) for 2004 Craniofacial Society of Great Britain and Ireland meeting Median time to publication = 24 months (range = 24 to 37 months) for 2005 Craniofacial Society of Great Britain and Ireland meeting Median time to publication = 14 months (range = 0 to 28 months) for 2006 Craniofacial Society of Great Britain and Ireland meeting Median time to publication = 48 months (range = 15 to 59 months) for 2007 Craniofacial Society of Great Britain and Ireland meeting Median time to publication = 35 months (range = 0 to 49 months) for 2008 Craniofacial Society of Great Britain and Ireland meeting Median time to publication = 13 months (range = 0 to 26 months) for 2009 Craniofacial Society of Great Britain and Ireland meeting Proportion of abstracts published at 40 months = 15.8% (50/318 abstracts) No factors other than subspecialty related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ craniofacial medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all oral presentations with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2008 had at least 48 months follow‐up. The meetings in 2008 and 2009 had 36 and 24 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis.

Montane 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to May 2004 Search completed by investigator Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 248 abstracts presented at the 1994, 1996, and 1998 Spanish Society of Clinical Pharmacology meetings Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published Oral versus poster presentation Author affiliation Subspecialty
Outcomes	71 of 248 abstracts published Proportion of abstracts published by time Median time to publication = 18 months (range = 2 to 60 months) Cumulative proportion of abstracts published at 36 months showed proportion published = 23.4% (58/248 abstracts) Factors related to proportion of abstracts published included 15/39 abstracts presented orally versus 59/209 abstracts presented as posters published 39.1% of abstracts with at least one author with an academic affiliation versus 32.3% of abstracts with at least one author affiliated with primary health care versus 27.4% of abstracts with at least one author affiliated with a hospital department versus 25.6% of abstracts with at least one author affiliated with a pharmaceutical company published
Notes	Pharmacology Funding by Astra‐Zeneca Foundation, Spain
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Unclear	Examined association of type presentation, academic affiliation, Spanish origin, and subspecialty with publication using stratified analysis and time to publication using Cox regression analysis.

Moorthi 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed until June 2010 Person completing the search not reported Searched by first and last author, and title Matching criteria not reported
Data	Included 73 abstracts submitted to the 2005 American Society of Nephrology meeting Included all abstracts of completed randomized controlled trials
Comparisons	Proportion of abstracts published Median time to publication Industry funding Oral versus poster presentation Multi‐centered Statistically significant results European versus North American origin Asian versus North American origin
Outcomes	39 of 73 abstracts published Proportion of abstracts published by time Median time to publication = 24 months (IQR = 13 to 36 months) Factors related to proportion of abstracts published included Industry funding = OR 2.92 (95% CI = 1.09 to 7.85) Oral versus poster presentation = OR 3.66 (95% CI = 1.29 to 10.39) Multi‐centered versus single center = OR 1.71 (95% CI = 0.59 to 5.02) Positive results (defined as statistically significant) results versus not positive = OR 1.59 (95% CI = 0.57 to 4.49) European versus North American origin = OR 0.61 (95% CI = 0.19 to 1.89) Asian versus North American origin = OR 0.18 (95% CI = 0.05 to 0.72)
Notes	Nephrology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific study design, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results, type of presentation, country of origin, industry funding, and various aspects of study design with publication using univariate logistic regression analysis.

Morgan 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2004 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 131 abstracts presented at the 2001 British Association of Dermatology meeting Included all abstracts except for those from 'specialty' meetings
Comparisons	Proportion of abstracts published Mean time to publication Plenary versus registrar's forum versus clinicopathological cases versus poster presentation
Outcomes	67 of 131 abstracts published Proportion of abstracts published by time Mean time to publication = 18 months Factors related to proportion of abstracts published included 18/29 abstracts presented at plenary sessions versus 9/12 abstracts presented at 'registrar's forum' versus 3/11 abstracts that presented clinicopathological cases versus 37/79 abstracts presented as poster published
Notes	Dermatology Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	No	The meeting only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis.

Morrison 1994.

Methods	Identification of subsequent full‐length publications Contacted abstract authors directly
Data	Included 72 abstracts by residents of the Department of Obstetrics and Gynecology and presented at various national meetings from 1983 to 1992 Included all abstracts
Comparisons	Proportion of abstracts published Clinical research versus basic science research Subspecialty
Outcomes	52 of 72 of all abstracts published, but 36 of 61 abstracts with 24 or more months of follow‐up published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 32/48 abstracts describing clinical research versus 20/24 abstracts describing basic science research published
Notes	Gynecology/obstetrics Funding not reported Authors present work of 11 additional abstracts followed for < 2 years and not included here
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Unclear	Contacted abstract authors, but response rate not reported.
Follow‐up time?	Yes	All meetings before 1991 had at least 48 months follow‐up. All meetings after and including 1991 had less than 48 months follow‐up.
Matching?	Yes	Matched through contact with abstract author.
Adjustment for confounding?	No	Examined association of type science and subspecialty with publication using stratified analysis.

Mowla 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to January 2006 Search completed by investigator Searched by all authors Matched abstract to full‐length publication by One author One outcome
Data	Included 890 abstracts presented at the 2000 to 2003 Annual Research Meeting of Iranian Medical Sciences Students meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication by type of presentation Oral versus poster presentation
Outcomes	98 of 890 abstracts published Proportion of abstracts published by time Mean time to publication = 22.3 months for oral presentations Mean time to publication = 29.5 months for poster presentations Factors related to proportion of abstracts published included 39/156 abstracts presented orally versus 59/636 abstracts presented as poster published
Notes	General medicine/primary care Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2003 had at least 48 months follow‐up. The meeting in 2003 only had 36 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Yes	Examined association of type presentation with publication using stratified analysis and multivariable logistic regression.

Muffly 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed 'for a 2 year period' Person completing the search not reported Searched by last and lead author, and keywords Matching criteria not reported
Data	Included 438 abstracts presented at the 2007 to 2008 American Urogynecologic Society meeting Included all abstracts except videos, those related to tips and tricks, and 7 that could not be identified as either published/unpublished
Comparisons	Proportion of abstracts published Mean time to publication, overall and by type of presentation Cumulative proportion of abstracts published Oral versus poster presentation University affiliation versus not Retrospective design versus prospective design Multi‐centered versus single center RCT design versus other study design
Outcomes	239 of 438 abstracts published Proportion of abstracts published by time Mean time to publication = 15.7 months (SD = 12.1), overall Mean time to publication = 14 months (SD = 11) for oral presentations Mean time to publication= 17 months (SD = 13) for poster presentations Cumulative proportion of abstracts published at 60 months showed proportion published = 55.3% (242/438 abstracts) Factors related to proportion of abstracts published included 121/147 abstracts presented orally versus 118/284 abstracts presented as poster published University affiliation versus not: OR 0.62 (95% CI = 0.38 to 1.03) Retrospective versus prospective design: OR 1.19 (95% CI = 0.76 to 1.85) Multiple centers versus single center: OR 0.7 (95% CI = 0.32 to 1.51) RCT design versus other study design: OR 0.49 (95% CI = 0.17 to 1.42)
Notes	Gynecology/obstetrics ‐ urogynecology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meetings only had 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of type of presentation, university affiliation, study design, multi‐center status, and prospective status with publication using multivariable logistic regression analyses.

Murrey 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to December 1997 Search completed by investigator Searched by all authors Matching criteria not reported, but measured interobserver reliability for 2 searchers (95.1%)
Data	Included 764 abstracts presented at the 1993 American Academy of Orthopedic Surgeons and 1992 and 1993 Society for Surgical Oncology meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published for one meeting Subspecialty
Outcomes	377 of 764 abstracts published 50/78 abstracts presented at the 1992 meeting, 76/113 at the 1993 Society for Surgical Oncology meeting, and 251/573 at the 1993 American Academy of Orthopedic Surgeons meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 43.8% (251/573 abstracts) for 1993 American Academy of Orthopedic Surgeons meeting No factors other than subspecialty related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis.

Mutlu 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar to March 2013 Person completing the search not reported Searched by first and second author, and title Matched abstract to full‐length publication by Title Results Protocol
Data	Included 214 abstracts presented at the 2005 to 2008 National Congress of Child and Adolescent Psychiatry meetings Included all abstracts accepted for poster presentation
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting Cumulative proportion of abstracts published Research type versus case type
Outcomes	54 of 214 abstracts published 13/51 abstracts presented at the 2005 meeting, 19/64 at the 2006 meeting, 12/59 at the 2007 meeting, and 10/40 at the 2008 meeting published Proportion of abstracts published by time Mean time to publication = 30.7 months (SD = 18.9; range = 4 to 74 months), overall Mean time to publication = 35.3 months (SD = 22.6) for 2005 National Congress of Child and Adolescent Psychiatry meeting Mean time to publication = 24.4 months (SD = 16.1) for 2006 National Congress of Child and Adolescent Psychiatry meeting Mean time to publication = 33.8 months (SD = 18.9) for 2007 National Congress of Child and Adolescent Psychiatry meeting Mean time to publication = 33.0 months (SD = 18.2) for 2008 National Congress of Child and Adolescent Psychiatry meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 23.8% (51/214 abstracts) Factors related to proportion of abstracts published included 40/141 abstracts with research type versus 14/73 abstracts with case type published
Notes	Psychiatry ‐ childhood and adolescents Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for poster presentation.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and study design with publication using stratified analysis and Chi2 or Mann‐Whitney U tests.

Nader 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to January 2007 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 125 abstracts presented at the 1999 to 2001 International Geographic Medicine Congress meetings Included abstracts where author responded to inquiry (125 of 250 randomly selected abstracts)
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Sample size less than or equal to 100 versus between 101 and 500 versus > 500 Oral versus poster presentation 'Experimental' design versus 'observational' design Multi‐centered versus single center 'Specified' sponsor versus no 'specified' sponsor
Outcomes	34 of 125 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 34/96 'positive' (defined as experimental better than control) versus 1/26 not 'positive' abstract results published 23/81 abstracts with sample size of less than or equal to 100 versus 8/33 abstracts with sample size between 101 and 500 versus 3/11 abstracts with sample size > 500 published 14/30 abstracts presented orally versus 20/95 abstracts presented as posters published 19/47 abstracts with 'experimental' design versus 16/78 abstracts with 'observational' design published 10/18 abstracts with multiple centers versus 27/107 abstracts with a single center published 25/46 abstracts 'specifying' a financial sponsor versus 10/79 abstracts not 'specifying' a financial sponsor published
Notes	Other clinical specialties ‐ geographic medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	No	Included abstracts only if author responded to inquiry; response rate was 50% (125/250).
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Yes	Matched by author contact.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, study design, multi‐center status, and funding status with publication using multivariable logistic regression analyses.

Nasir 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar, and African Journal Online until the end of December 2010 Person completing the search not reported Searched by all authors and title Matched abstract to full‐length publication by Title Hypothesis Study design Personal contacts with authors (meetings, email, telephone)
Data	Included 153 abstracts presented at the 2004, and 2006 to 2009 Association of Paediatric Surgeons of Nigeria meetings Included all abstracts presented orally except those accepted but not presented during the meetings
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication Observational design versus case report versus survey versus review versus operative technique Multi‐centered versus single center Retrospective versus prospective design Trainees versus consultants Subspecialty
Outcomes	52 of 128 abstracts published 17/30 abstracts presented at the 2004 meeting, 14/26 at the 2006 meeting, 14/40 at the 2007 meeting, and 7/32 at the 2008 meeting Proportion of abstracts published by time Mean time to publication = 23.5 months (range = 12 to 60 months) Median time to publication = 24 months Factors related to proportion of abstracts published included, but all analyses include some abstracts with < 24 months follow‐up 40/112 abstracts describing an observational design versus 11/35 abstracts describing case reports versus 1/1 abstracts describing surveys versus 0/1 abstracts describing reviews versus 0/4 abstracts describing operative technique published 1/2 abstracts with multiple centers versus 51/151 abstracts with a single center published 32/81 abstracts describing a retrospective design versus 8/31 abstracts describing a prospective design published 11/39 abstracts of trainees versus 41/114 abstracts of consultants published
Notes	Surgery ‐ pediatrics Funding not reported Author reported proportion of abstracts published in full from 2009 meeting, but these data were excluded from analyses of proportion published due to < 24 months follow‐up.
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally with reasonable exception.
Search for publications?	Yes	Searched 2 databases, relevant journals and contacted abstract authors directly.
Follow‐up time?	No	The meetings in 2007 and 2008 had less than 48 months follow‐up. The meetings in 2004 and 2006 had at least 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, study design, multi‐center status, prospective status, abstract quality, author status, and subspecialty with publication using stratified analysis and Chi2 or Fisher's Exact tests

Nasir 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, MEDLINE, Google/Google Scholar, and the African Journal OnLine through October 2012 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Methodology Conclusions
Data	Included 164 abstracts presented at the 2006, 2008, and 2010 Pan‐African Pediatric Surgical Association meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Oral versus poster presentation Observational design versus case report versus review versus technical report North American versus European versus African versus rest of the world origin Retrospective versus prospective design Statistical inference stated versus not Control group included versus not Subspecialty
Outcomes	49 of 164 abstracts published 17/45 abstracts presented at the 2006 meeting, 23/62 at the 2008 meeting, and 9/57 at the 2010 meeting published Proportion of abstracts published by time Median time to publication = 15 months (IQR = 5 to 26 months) Factors related to proportion of abstracts published included 38/118 abstracts presented orally versus 11/46 abstracts presented as poster published 42/135 abstracts describing observational design versus 1/11 abstracts describing case reports versus 3/8 abstracts describing reviews versus 1/6 abstracts describing technical reports versus 2/4 abstracts describing surveys published 3/7 abstracts with origin in North American versus 7/18 with origin in Europe versus 38/133 abstracts with origin in Africa versus 0/5 abstracts with rest of the world origin published 32/96 abstracts describing retrospective design versus 10/39 abstracts describing prospective design published Statistical inference stated versus not: 57.9% vs. 26.2% (OR 3.8, 95% CI = 1.44 to10.34) Control group included versus not: 35.3% vs. 24.7% (OR 1.3, 95% CI = 0.46 to 3.79)
Notes	Surgery ‐ pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The 2010 meeting only had 24 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, study design, country of origin, prospective status, presence of statistical inference, presence of a control group, and subspecialty with publication using stratified analysis and Fisher's Exact or Mann‐Whitney U tests.

Ng 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to June 2002 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Title Methodology Results Hypothesis
Data	Included 4302 abstracts presented at the 1998 to 2000 American Urological Association meetings Included all abstracts except those from specialty forums and videotape sessions
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting Survival analysis of publication rate Oral versus poster presentation Clinical research versus basic science research
Outcomes	1627 of 4302 abstracts published 543/1280 abstracts presented at the 1998 meeting, 570/1509 at the 1999 meeting, and 514/1513 at the 2000 meeting published Proportion of abstracts published by time Mean time to publication = 36.9 months (95% CI = 35.2 to 37.4 months) for 1998 American Urological Association meeting Mean time to publication = 29.4 months (95% CI = 28.6 to 30.1 months) for 1999 American Urological Association meeting Mean time to publication = 23.2 months (95% CI = 22.5 to 23.9 months) for 2000 American Urological Association meeting Survival analysis of proportion published = 41.3% at 48 months Factors related to proportion of abstracts published included 384/979 abstracts presented orally versus 1244/3323 abstracts presented as posters published 1205/3124 abstracts describing clinical research versus 422/1178 abstracts describing basic science research published
Notes	Urology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	Only the meeting in 1998 had 48 months follow‐up. All other meetings had less than 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, type presentation, and type science with publication using multivariable Cox regression or log‐rank analyses.

Nguyen 1998.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1986 to June 1997 Person completing the search not reported Searched by keywords and content/subject area Matched abstract to full‐length publication by All authors Contents Contacted abstract authors directly
Data	Included 490 abstracts presented at the 1990 to 1995 Orthopedic Trauma Association meetings Included all abstracts of 'paper' presentations
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published
Outcomes	292 of 490 abstracts published 54/78 abstracts presented at the 1990 meeting, 53/89 at the 1991 meeting, 61/94 at the 1992 meeting, 61/92 at the 1993 meeting, 44/76 at the 1994 meeting, and 19/61 at the 1995 meeting published Proportion of abstracts published by time Mean time to publication = 16 months Cumulative proportion of abstracts published at 72 months showed proportion published = 59.4% (291/490 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Emergency medicine ‐ orthopedic trauma Funding not reported Date of meeting not given, need to assume that meeting was on or before June 1995 and that follow‐up was at least 2 years
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	All meetings before 1994 had at least 48 months follow‐up. The meetings in 1994 and 1995 only had 36 and 24 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Nqwena 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Search criteria not reported Matched abstract to full‐length publication by One author One conclusion
Data	Included 160 abstracts presented at the 1999 to 2002 Sylvester O'Halloran surgical meeting Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published
Outcomes	47 of 160 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Surgery Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted as oral presentations.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.

O' Connor 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and PubMed; dates of search not reported Person completing the search not reported Searched by first, second and last authors, keywords, and title Matched abstract to full‐length publication by Sample size Aims Conclusions
Data	Included 298 abstracts presented at the 1989 to 2014 Sir Peter Freyer Surgical Symposium meetings Included all plenary abstracts
Comparisons	Proportion of abstracts published Mean time to publication, overall and by type of research Cumulative proportion of abstracts published Clinical research versus basic science research Multi‐centered versus single center Irish versus other institutions
Outcomes	168 of 298 abstracts published Proportion of abstracts published by time Mean time to publication = 24 months (SD = 84) Mean time to publication = 24 months (range = ‐12 to 96 months) for abstracts describing basic science research Mean time to publication = 24 months (range = ‐12 to 108 months) for abstracts describing clinical research Cumulative proportion of abstracts published at 108 months showed proportion published = 56.4% (168/298 abstracts) Factors related to proportion of abstracts published included 34/61 abstracts describing clinical research versus 134/237 abstracts describing basic science research published 50/96 abstracts with multiple centers versus 118/202 abstracts with a single center published 116/228 abstracts from an Irish institution versus 52/70 abstracts from another institution published
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all plenary abstracts.
Search for publications?	Yes	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of type science, multi‐center status, and Irish origin with publication using stratified analysis and Chi2 tests

O'Dell 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to July 2010 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 446 abstracts presented at the 2008 28th Western States Conference for Pharmacy Residents, Fellows, and Preceptors meeting Included all abstracts
Comparisons	Proportion of abstracts published
Outcomes	19 of 446 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Pharmacology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meeting only had 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

O'Kelly 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE, PubMed, and SCOPUS from 1 January, 2005 to 31 December, 2013 Person completing the search not reported Searched by all authors, and keywords Matched abstract to full‐length publication by Hypotheses Study design
Data	Included 203 abstracts presented at the 2005 to 2011 Irish Society of Urology meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting Median time to publication Oral versus poster presentations Subspecialty
Outcomes	144 of 322 abstracts published 18/34 abstracts presented at the 2005 meeting, 20/57 at the 2006 meeting, 15/37 at the 2007 meeting, 26/45 at the 2008 meeting, 25/45 at the 2009 meeting, 27/56 at the 2010 meeting, and 13/49 at the 2011 meeting published Proportion of abstracts published by time Mean time to publication = 16.7 months, overall Mean time to publication = 8 months for the 2005 meeting Mean time to publication = 17 months for the 2006 meeting Mean time to publication = 28 months for the 2007 meeting Mean time to publication = 23 months for the 2008 meeting Mean time to publication = 13 months for the 2009 meeting Mean time to publication = 26 months for the 2010 meeting Mean time to publication = 12 months for the 2011 meeting Median time to publication = 15 months (range = 0 to 55 months), overall Factors related to proportion of abstracts published included Oral presentations were more likely to be published than poster presentations (P < 0.0001)
Notes	Urology Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The meetings in 2010 and 2011 had less than 48 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, and subspecialty with publication using univariable logistic regression.

O'Neill 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors, keywords, and institution names Matching criteria not reported
Data	Included 203 abstracts presented at the 2002 to 2005 Irish Orthopaedic Association meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Clinical research versus basic science research Retrospective versus prospective research
Outcomes	66 of 203 abstracts published 13/49 abstracts presented at the 2002 meeting, 17/52 at the 2003 meeting, 20/50 at the 2004 meeting, and 16/52 at the 2005 meeting published Proportion of abstracts published by time Mean time to publication = 30.1 months (range = 4.7 to 76.5 months) Proportion of abstracts published at 60 months = 31.0% (63/203 abstracts) Factors related to proportion of abstracts published included 49/167 abstracts describing clinical research versus 16/36 abstracts describing basic science research published 24/94 abstracts describing retrospective studies versus 24/71 abstracts describing prospective studies published
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type science, and prospective status using stratified analysis and unspecified statistical tests.

Odunsi 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by first and last author, and title Matching criteria not reported
Data	Included 627 abstracts presented at the 2007 to 2009 Association of British Neurologists meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Oral versus poster presentation Case report versus other design Multi‐centered versus single center
Outcomes	244 of 627 abstracts published Proportion of abstracts published by time Mean time to publication = 18.6 months (SD = 14.6) Factors related to proportion of abstracts published included 135/299 abstracts presented orally versus 109/328 abstracts presented as poster published 32/106 abstracts describing case report design versus 212/489 abstracts describing other design published 126/390 abstracts with multiple centers versus 122/239 abstracts with single center published
Notes	Neurology Reporting having received no funding Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type of presentation, study design, and multi‐center status with publication using stratified analysis and Chi2 tests.

Ogilvie 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and Google/Google Scholar; dates of search not reported Person completing the search not reported Searched by all authors, and keywords Matched abstract to full‐length publication by Two authors Methodology Subject matter Contacted abstract authors directly
Data	Included 392 abstracts accepted at the 2006 to 2010 Canadian Society of Otolaryngology ‐ Head & Neck Surgery meetings Included all abstracts presented orally except those from the Poliquin Resident’s competition
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Subspecialty
Outcomes	198 of 392 abstracts published 21/42 abstracts presented at the 2006 meeting, 40/77 at the 2007 meeting, 44/80 at the 2008 meeting, 45/89 at the 2009 meeting, and 44/103 at the 2010 meeting published Proportion of abstracts published by time Mean time to publication = 21 months (range = ‐61 to 76 months) No factors other than subspecialty and meeting year related to proportion of abstracts published reported
Notes	Otolaryngology head and neck surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases, and contacted abstract authors directly.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Yes	Matched by 3 criteria.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis and Chi2 tests.

Ohlsson 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database Cochrane Central Register of Controlled Trials, MEDLINE, and Embase to March 1999 Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 141 abstracts submitted to the 1993 to 1994 American Pediatric Society/Society for Pediatric Research meetings Included all abstracts of randomized clinical trials
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Accepted versus rejected for conference presentation
Outcomes	73 of 141 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 57 months showed proportion published = 51.8% (73/141 abstracts) Factors related to proportion of abstracts published included 62/107 abstracts accepted for presentation versus 11/34 rejected abstracts published
Notes	Pediatrics ‐ neonatology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific study design, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of acceptance for presentation with publication using stratified analysis.

Ohtori 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for 2 years before each meeting until the end of December 2012 Search completed by the investigators Searched by first author and keywords Matched abstract to full‐length publication by Contents
Data	Included 3205 abstracts presented at the 2006 to 2008 Annual Research Meeting of the Japanese Orthopaedic Association and 2006 to 2007 Japanese Orthopaedic Association meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation Subspecialty
Outcomes	1198 of 3205 abstracts published 232/447 abstracts presented at the 2006 Annual Research Meeting of the Japanese Orthopaedic Association meeting, 210/814 at the 2006 Japanese Orthopaedic Association meeting, 252/489 at the 2007 Annual Research Meeting of the Japanese Orthopaedic Association meeting, 212/862 at the 2007 Japanese Orthopaedic Association meeting, and 292/593 at the 2008 Japanese Orthopaedic Association meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 557/1397 abstracts presented orally versus 641/3205 abstracts presented as poster published
Notes	Orthopedics/orthopedic surgery Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	Each meeting had at least 48 months follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of meeting and meeting year, type presentation, and subspecialty using stratified analysis and t test or one‐way ANOVA tests.

Okafor 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed up to December 2014 Person completing the search not reported Searched by all authors, and title Matching criteria not reported
Data	Included 321 abstracts presented at the 2007 to 2011 Cervical Spine Research Society meetings Included all abstracts presented orally
Comparisons	Proportion of abstracts published, overall and by meeting Award‐winning versus not
Outcomes	211 of 321 abstracts published 42/54 abstracts presented at the 2007 meeting, 37/54 at the 2008 meeting, 38/65 at the 2009 meeting, 46/78 at the 2010 meeting, and 48/70 at the 2011 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 35/45 award‐winning abstracts versus 176/276 non award‐winning abstracts published
Notes	Orthopedics/orthopedic surgery ‐ cervical spine Reporting having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The 2011 meeting only had 36 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and award winning status with publication using stratified analysis and t tests.

Olive 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from '1995 to 2001' Person completing the search not reported Searched by all authors Matching criteria not reported
Data	Included 249 abstracts presented at the 1996 Spanish Society for Rheumatology meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	52 of 249 abstracts published Proportion of abstracts published by time Mean time to publication = 18.5 months No factors related to proportion of abstracts published reported
Notes	Rheumatology Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.

Oliveira 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, LILACS, and SciELO to August 2008 Person completing the search not reported Searched by all authors Matching criteria not reported Contacted abstract authors directly
Data	Included 313 abstracts presented at the 2003 Urological Brazilian meeting Included a random sample of 1400 abstracts, included those presented orally
Comparisons	Proportion of abstracts published Median time to publication Clinical research versus basic science research CCT design versus observational design versus retrospective design Brazilian origin versus elsewhere Subspecialty
Outcomes	122 of 313 abstracts published Proportion of abstracts published by time Median time to publication = 14 months (range = 1 to 51 months) Factors related to proportion of abstracts published included 114/287 abstracts describing clinical research versus 9/25 abstracts describing basic science research published 8/16 abstracts with CCT design versus 51/105 abstracts with observational design versus 56/166 abstracts with retrospective design published 116/300 abstracts originating from Brazil versus 7/13 abstracts from elsewhere published
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type science, study design, Brazilian origin, and subspecialty with publication using stratified analysis and Chi2 or Fisher's Exact tests.

Oliver 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Contents Results
Data	Included 370 abstracts presented at the 1995, 1997, 1998, and 1999 British Association of Plastic Surgeons ‐ Winter meetings and the 1997 British Association of Plastic Surgeons ‐ Summer meeting Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting
Outcomes	118 of 370 abstracts published 17/56 abstracts presented at the 1995 British Association of Plastic Surgeons ‐ Winter meeting, 15/48 abstracts at the 1997 British Association of Plastic Surgeons ‐ Summer meeting, 29/54 at the 1997 British Association of Plastic Surgeons ‐ Winter meeting, 33/107 at the 1998 British Association of Plastic Surgeons ‐ Winter meeting, and 24/105 at the 1999 British Association of Plastic Surgeons ‐ Winter meeting published Proportion of abstracts published by time Mean time to publication = 19 months (range = 3 to 46 months) for 1995 British Association of Plastic Surgeons ‐ Winter meeting Mean time to publication = 25 months (range = 3 to 42 months) for 1997 British Association of Plastic Surgeons ‐ Summer meeting Mean time to publication = 19 months (range = 1 to 46 months) for 1997 British Association of Plastic Surgeons ‐ Winter meeting Mean time to publication = 13 months (range = 1 to 28 months) for 1998 British Association of Plastic Surgeons ‐ Winter meeting Mean time to publication = 16 months (range = 1 to 26 months) for 1999 British Association of Plastic Surgeons ‐ Winter meeting No factors other than meeting related to proportion of abstracts published reported
Notes	Surgery ‐ plastic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Olmos‐de‐Aguilera 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and SciELO; dates of search not reported Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 59 abstracts presented at the 2009 to 2011 Congresos Científicos Nacionales de Estudiantes de Medicina, and Asociación Nacional Científica de Estudiantes de Medicina de Chile meetings Included best scientific works of medical students selected by the organizing commission of the two congresses
Comparisons	Proportion of abstracts published
Outcomes	11 of 59 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	General medicine/primary care Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific subgroup.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.

Olson 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase to 5 years after the meeting Search completed by two investigators Searched by all authors and keywords Matching criteria not reported
Data	Included 270 abstracts presented at the 1995, 2000, and 2005 Western States Conference meetings Included every third abstract published in the corresponding abstract book, skipping those not presented
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication overall and by meeting RCT design versus CCT design versus observational study design versus policy abstract versus other
Outcomes	17 of 270 abstracts published 3/71 abstracts presented at the 1995 meeting, 4/77 at the 2000 meeting, and 10/122 at the 2005 meeting published Proportion of abstracts published Mean time to publication = 24 months (SD =10), overall Mean time to publication = 17 months (SD =13) for 1995 Western States Conference meeting Mean time to publication = 24 months (SD =13) for 2000 Western States Conference meeting Mean time to publication = 24 months (SD = 9) for 2005 Western States Conference meeting Factors related to proportion of abstracts published included 0/17 abstracts with RCT design versus 0/4 abstracts with CCT design versus 12/147 abstracts with observational design versus 3/61 abstracts on policy versus 2/41 abstracts of other design or topics published
Notes	Pharmacology ‐ residents Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a systematically selected sample of abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and study design with publication using stratified analysis.

Ospina 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, MEDLINE, Embase to May 2004 Search completed by two investigators Searched by all authors and keywords Matched abstract to full‐length publication by One author Contents Methodology
Data	Included 383 abstracts presented at the 1995 to 2003 Society for Academic Emergency Medicine meetings Included all abstracts of controlled clinical trials and randomized clinical trials
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Survival analysis of publication rate 'Positive' versus not 'positive' Authors endorsing treatment versus not Parallel design versus crossover or factorial design Multi‐centered versus single center US versus non‐US origin University affiliation versus no university affiliation
Outcomes	194 of 383 abstracts published 23/33 abstracts presented at the 1995 meeting, 24/40 at the 1996 meeting, 28/50 at the 1997 meeting, 27/42 at the 1998 meeting, 28/43 at the 1999 meeting, 27/44 at the 2000 meeting, 17/36 at the 2001 meeting, 17/52 at the 2002 meeting, and 3/43 at the 2003 meeting published Proportion of abstracts published by time Median time to publication = 32 months (95% CI = 23 to 41 months) Survival analysis of proportion published at 108 months = 62.1% Factors related to proportion of abstracts published included 84/161 'positive' (defined as significant results) versus 110/222 not 'positive' abstract results published 80/154 'positive' (defined as experimental better than control) versus 4/7 not 'positive' abstract results published 108/210 abstracts with authors endorsing the treatment versus 86/173 abstracts with authors not endorsing the treatment published 164/327 abstracts with parallel design versus 30/56 abstracts with crossover or factorial design published 29/61 abstracts with multiple centers versus 165/322 abstracts with a single center published 169/339 abstracts originating in the US versus 25/44 abstracts not originating in the US published 138/253 abstracts with university affiliation versus 31/86 abstracts without a university affiliation published
Notes	Emergency medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific study designs, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings before 2001 had at least 48 months follow‐up. The other meetings had less than 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, positive results, sample size, author conclusions, study design, multi‐center status, number of authors, university affiliation, and US origin with publication using multivariable logistic regression analyses.

Ozel 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database Electronic database (not specified) for '3 years time interval between meeting and search' Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 116 abstracts presented at the 2002 to 2003 Society of Gynecologic Surgeons meetings Included all abstracts except tip and tricks and video presentations
Comparisons	Proportion of abstracts published Mean and median time to publication Oral versus poster versus oral/poster presentation RCT design versus non‐RCT design Academic institutions versus not Country of origin
Outcomes	63 of 116 abstracts published Proportion of abstracts published by time Mean time to publication = 16 months Median time to publication = 9 months (range = ‐4 to 48 months) Factors related to proportion of abstracts published included 34/37 abstracts presented orally versus 7/18 abstracts presented as oral posters versus 22/61 abstracts presented as non‐oral poster published 5/7 abstracts describing RCTs versus 58/109 abstracts describing other study designs published 58/94 abstracts from academic institutions versus 5/17 abstracts from non‐academic institutions published 58/105 abstracts with origin in North America versus 5/11 abstracts with rest of the world origin published
Notes	Gynecology/obstetrics Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Unclear	The number and type of electronic databases searched was not reported.
Follow‐up time?	No	The meetings had only 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation, study design, academic affiliation, and country of origin with publication using Chi2 tests

Ozkösem 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database Web of Science and Harzing's Publish or Perish algorithm to June 2013 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 943 abstracts presented at the 2005 to 2006 Society for Free Radical Biology and Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published Basic science Number of authors Country of origin Institution
Outcomes	398 of 943 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included Basic science = OR 3.6 (95% CI: 3.1 to 6.4) Number of authors = OR 1.2 (95% CI: 1.0 to 1.3)
Notes	Other non‐clinical specialties ‐free radical biology and medicine Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of type science, number of authors, country of origin, and institution with publication using multivariable logistic regression analyses.

Ozkösem 2013a.

Methods	Identification of subsequent full‐length publications Searched electronic database Web of Science and Harzing’s Publish or Perish software until July 2013 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 1546 abstracts presented at the 2005 to 2006 Society for the Study of Reproduction meetings Included all abstracts
Comparisons	Proportion of abstracts published Survival analysis of proportion of abstracts published 'Positive' versus not 'positive' versus not specified Oral versus poster presentation Clinical research versus basic science research RCT design versus observational design versus case report versus other design Multi‐centered versus single center North American versus European versus rest of the world origin US‐American versus German versus British versus Japanese versus rest of the world origin One author versus two to four authors versus more than four authors Industry funding versus no industry funding
Outcomes	703 of 1546 abstracts published Proportion of abstracts published by time Survival analyses showed proportion of abstracts published at 42 months = 60.0% Factors related to proportion of abstracts published included 383/696 'positive' (defined as significant results) versus 181/464 not 'positive' abstract results versus 148/309 abstracts without specified level of significance published 317/557 abstracts presented orally versus 386/989 abstracts presented as poster published 269/572 abstracts describing clinical research versus 594/974 abstracts describing basic science research published 48/77 abstracts describing RCT design versus 460/942 abstracts describing observational design versus 85/139 abstracts describing case reports versus 264/388 abstracts describing other study designs published 383/649 abstracts with multiple centers versus 320/897 abstracts with a single center published 506/804 abstracts with origin in North America versus 348/588 abstracts with origin in Europe versus 89/151 abstracts with rest of the world origin published 506/804 abstracts from the US versus 198/325 abstracts from Germany versus 150/263 abstracts from the UK versus 75/108 abstracts from Japan versus 14/43 abstracts with rest of the world origin published 13/43 abstracts with one author versus 458/881 abstracts with two to four authors versus 420/618 abstracts with more than 4 authors published 52/77 abstracts with industry funding versus 651/1469 abstracts without industry funding published
Notes	Human reproduction and development ‐ reproductive biology Funded by the Réseau Québécois en reproduction and Centre for the Study of Reproduction Data extracted from poster
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, type science, study design, multi‐center status, country of origin, US origin, number of authors, and funding status with publication using multivariable logistic regression analyses.

Paes 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Scielo from 2012 to January 2015 Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 701 abstracts presented at the 2011 Congresso Brasileiro de Patologia and Congreso de la Sociedad Latinoamericana de Patología Included all abstracts
Comparisons	Proportion of abstracts published North American versus European versus South American origin Brazilian versus other South American versus Mexican versus Spanish versus US versus origin not specified Subspecialty
Outcomes	63 of 701 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 2/8 abstracts originating from North America versus 0/3 abstracts originating from Europe versus 61/684 abstracts originating from Brazil and other Latin American countries published 58/635 abstracts originating from Brazil versus 3/49 abstracts originating from other Latin American countries versus 2/6 abstracts originating from Mexico versus 0/3 abstracts originating from Spain versus 0/2 abstracts originating from the USA versus 0/6 abstracts with origin not specified published
Notes	Pathology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All abstracts had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of country of origin, Brazilian origin, and subspecialty with publication using stratified analysis.

Papagikos 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to November 2003 Person completing the search not reported Searched by first, second, and senior author Matching criteria not reported
Data	Included 802 abstracts presented at the 1999 to 2001 American Society of Therapeutic Radiology and Oncology meetings Included all abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published Median time to publication 'Positive' versus not 'positive' for phase III studies 'Clinical' research versus 'non‐clinical' research Phase III versus phase I or II versus observational versus other US versus non‐US origin Subspecialty
Outcomes	452 of 802 abstracts published Proportion of abstracts published by time Median time to publication = 11.9 months Factors related to proportion of abstracts published included 23/39 'positive' (defined as significant results) versus 21/45 not 'positive' abstract results of phase III studies published 358/604 abstracts describing clinical research versus 94/198 'non‐clinical' abstracts published 44/84 phase III abstracts versus 79/127 phase I or II abstracts versus 257/433 abstracts with observational study design versus 4/6 abstracts with other design published 319/569 abstracts originating in the US versus 94/198 abstracts not originating in the US published
Notes	Oncology ‐ radiation oncology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	Only the meeting in 1999 had 48 months follow‐up. All other meetings had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results, type science, study design, US origin, prospective status, and subspecialty with publication using stratified analysis and unspecified statistical tests.

Papoutsis 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1 January, 2007 to 1 January, 2012 Search completed by the investigators Searched by first author, keywords, and topics Matching criteria not reported
Data	Included 8411 abstracts submitted to the 2006 to 2010 Deutsche Gesellschaft für Kardiologie/Herz‐Kreislaufforschung meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published Accepted versus rejected abstracts Clinical research versus experimental research Clinical research versus experimental research for accepted abstracts Submitted by female versus male first author Accepted abstracts for conference by female versus male first author Rejected abstracts for conference by female versus male first author
Outcomes	1909 of 8411 abstracts published 403/1660 abstracts presented at the 2006 meeting, 443/1750 at the 2007 meeting, 401/1700 at the 2008 meeting, 339/1600 at the 2009 meeting, and 285/1675 at the 2010 meeting published Proportion of abstracts published by time Mean time to publication = 8.4 months (SD = 14.4) Cumulative proportion of abstracts published at 24 months showed proportion published = 20.0% (1680/8411 abstracts) Factors related to proportion of abstracts published included 1425/5535 abstracts accepted for presentation versus 483/2876 rejected for presentation published 1271/5914 abstracts describing clinical research versus 637/2497 abstracts describing experimental research published 942/3765 abstracts describing clinical research versus 483/1770 abstracts describing experimental research among the accepted abstracts published 366/2090 abstracts submitted by female first author versus 1541/6321 abstracts submitted by male first author published 285/1377 abstracts accepted to publication by female first author versus 1140/4158 abstracts accepted to publication by male first author published 81/713 rejected abstracts by female first author versus 401/2163 rejected abstracts by male first author published
Notes	Cardiology Author received support from Deutsche Forschungsgemeinschaft; project received no funding except utilization of resources of the German Cardiac Society
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2009 had at least 48 months follow‐up. The meetings in 2009 and 2010 only had 36 and 24 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, acceptance for presentation, and type science with publication using stratified analysis and Chi2 or Mann‐Whitney U tests

Papp 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Web of Science, ERIC to October 2010 Search completed by trained librarian Searched by all authors and keywords Matching criteria not reported
Data	Included 201 abstracts presented at the 1995 to 2005 Clerkship Directors in Internal Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published
Outcomes	66 of 201 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Medical education/library science Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.

Parkar 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed until September 2011 Search completed by investigator Searched by all authors and title Matching criteria not reported
Data	Included 137 abstracts presented at the 2008 to 2009 European Society of Musculoskeletal Radiology meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published, overall and by meeting North American versus European versus Asian origin
Outcomes	62 of 137 abstracts published 34/73 abstracts presented at the 2008 meeting, and 28/64 at the 2009 meeting published Proportion of abstracts published by time Median time to publication = 14.5 months (range = ‐47 to 30 months) Cumulative proportion of abstracts published at 24 months showed proportion published = 38.7% (53/137 abstracts), overall Cumulative proportion of abstracts published at 24 months showed proportion published = 41.1% (30/73 abstracts) for the 2008 meeting Cumulative proportion of abstracts published at 24 months showed proportion published = 35.9% (23/64 abstracts) for the 2009 meeting Factors related to proportion of abstracts published included 2/3 abstracts originating in North America versus 60/132 abstracts originating in Europe versus 0/2 abstracts originating in Saudi Arabia published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and country of origin with publication using stratified analysis and Chi2 test.

Patel 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to May 2010 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Contents Conclusions
Data	Included 3827 abstracts presented at the 2003 to 2005 Congress of Neurological Surgeons and American Association of Neurological Surgeons meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published Oral versus poster presentation
Outcomes	1243 of 3827 abstracts published 221/651 abstracts presented at the 2003 American Association of Neurological Surgeons meeting, 193/620 at the 2004 American Association of Neurological Surgeons meeting, 196/634 at the 2005 American Association of Neurological Surgeons meeting, 209/605 at the 2003 Congress of Neurological Surgeons meeting, 200/565 at the 2004 Congress of Neurological Surgeons meeting, and 224/752 at the 2005 Congress of Neurological Surgeons meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 32.7% (628/1922 abstracts) for Congress of Neurological Surgeons meetings Cumulative proportion of abstracts published at 60 months showed proportion published = 31.4% (599/1905 abstracts) for American Association of Neurological Surgeons meetings Factors related to proportion of abstracts published included 445/1078 abstracts presented orally versus 798/2749 abstracts presented as posters published
Notes	Surgery ‐ neurologic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting by type presentation with publication using stratified analysis.

Paulson 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to September 2010 Search completed by investigator Searched by first and last author Matching criteria not reported
Data	Included 501 abstracts presented at the 2006 Center for International Blood and Marrow Transplantation Research/American Society for Blood and Marrow Transplantation meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication 'Positive' versus not 'positive' Sample size above 40 versus sample size equal to or less than 40 Oral versus poster presentation Clinical research versus basic science research versus 'descriptive' Multi‐centered versus single center 'Prospective' versus 'retrospective' design
Outcomes	217 of 501 abstracts published Proportion of abstracts published by time Median time to publication = 19 months (range = 1 to 41 months) Factors related to proportion of abstracts published included 164/327 'positive' (defined as experimental better than control) versus 53/173 not 'positive' abstract results published 73/141 abstracts with sample size above 40 versus 65/145 abstracts with sample size equal to or less than 40 published 36/55 abstracts presented orally versus 181/446 abstracts presented as posters published 141/294 abstracts describing clinical research versus 70/124 abstracts describing basic science research versus 6/83 'descriptive' abstracts published 58/101 abstracts with multiple centers versus 153/381 abstracts with a single center published 62/123 abstracts with 'prospective' study design versus 77/169 abstracts with 'retrospective' study design published
Notes	Oncology ‐ blood and marrow transplantation Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results, sample size, type of presentation, type of science, multi‐center status, and prospective status with publication using stratified analysis and Chi2 or t tests

Payne 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, MEDLINE; dates of search not reported Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 186 abstracts presented at the 1992 American Diabetes Association and European Association for the Study of Diabetes and Australian Diabetes Society meetings Included a random sample of abstracts
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	87 of 186 abstracts published 8/31 abstracts presented at the Australian Diabetes Society meeting, 41/84 at the European Association for the Study of Diabetes meeting, and 38/71 at the American Diabetes Association meeting published Proportion of abstracts published by time not reported No factors other than meeting related to proportion of abstracts published reported
Notes	Endocrinology/nutrition ‐ diabetes Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting with publication using stratified analysis.

Peng 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to February 2005 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Methodology Results
Data	Included 473 abstracts presented at the 1999 American Academy of Otolaryngology Head & Neck Surgery meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication 'Positive' versus not 'positive' Oral versus poster presentation RCT design versus observational study design versus case series versus opinion 'Prospective' versus 'retrospective' versus cross‐sectional design Case reports versus other Grant support versus other 'Research award' versus other
Outcomes	237 of 473 abstracts published Proportion of abstracts published by time Median time to publication = 16 months (IQR = 12 to 29 months) Factors related to proportion of abstracts published included 189/337 'positive' (defined as experimental better than control) versus 13/26 not 'positive' abstract results published 135/213 abstracts presented orally versus 102/260 abstracts presented as posters published 55/84 abstracts with RCT design versus 96/173 abstracts with observational study design versus 82/209 abstracts with case series design versus 2/3 opinion abstracts published 90/215 abstracts with 'prospective' study design versus 92/222 abstracts with 'retrospective' study design versus 20/36 with cross‐sectional study design published 18/70 abstracts of case reports versus 219/403 abstracts not of case reports published 25/37 abstracts reporting grant support versus 211/435 abstracts without or not reporting grant support published 9/10 abstracts with 'research award' versus 227/462 abstracts without 'research award' published
Notes	Otolaryngology head and neck surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 6 years follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, study design, prospective status, presence of grant support, and award winning with publication using multivariable logistic regression analyses.

Pereyra‐Elias 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database Google/Google Scholar to January 2011 Search completed by investigator Searched by all authors, content, place of study, and main results Matched abstract to full‐length publication by One author Contents Results Time of study
Data	Included 417 abstracts presented at the 1998, 2000, 2002, 2004, 2006, and 2008 Congreso de la Sociedad de Gastroenterología del Perú meetings Included all abstracts except repeated presentations and presentations published before the meeting
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication Observational design versus experimental design versus case reports versus other design Multi‐centered versus single center 'Prospective' versus 'retrospective' versus other design
Outcomes	34 of 417 abstracts published 9/170 abstracts presented at the 1998 meeting, 6/72 at the 2000 meeting, 5/66 at the 2002 meeting, 3/57 at the 2004 meeting, 7/27 at the 2006 meeting, and 4/26 at the 2008 meeting published Proportion of abstracts published by time mean time to publication = 10.8 months (SD = 10.4) median time to publication = 9 months (range = 0 to 47 months) Factors related to proportion of abstracts published included 21/306 abstracts with observational design versus 2/11 abstracts with experimental design versus 11/96 abstracts of case reports versus 0/4 abstracts with other design published 2/10 abstracts with multiple centers versus 32/407 abstracts with a single center published 11/115 abstracts with 'prospective' study design versus 16/228 abstracts with 'retrospective' study design versus 0/1 abstracts with other design published
Notes	Gastroenterology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2008 had at least 5 years follow‐up. The meeting in 2008 only had 36 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, study design, multi‐center status, and prospective status with publication using stratified analysis and multivariable logistic regression.

Perez 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE until September 2002 Person completing the search not reported Searched by all authors, keywords, and topic Matching criteria not reported Interviewed senior authors if publication not found via electronic database search
Data	Included 79 abstracts presented at the 1998 to 2000 Israel Society of Rheumatology meetings Included all abstracts
Comparisons	Proportion of abstracts published
Outcomes	63 of 79 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Rheumatology Funding not reported Information extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database and contacted abstract authors directly.
Follow‐up time?	No	All meetings except the meeting in 1998 had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Perron 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, PsychINFO, Social work abstracts, and Google/Google Scholar for more than 4 years after the meeting, although dates of search not reported Search completed by investigator Searched by first author, keywords, and title Matched abstract to full‐length publication by Sample size Methodology Results Also searched faculty biography or curriculum vitae
Data	Included 121 abstracts presented at the 2006 Society for Social Work and Research meeting Included a random sample of all abstracts by type of presentation (oral, poster, symposium)
Comparisons	Proportion of abstracts published Oral versus poster versus symposia presentation
Outcomes	52 of 121 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 37/75 abstracts presented orally versus 12/29 abstracts presented as posters published versus 3/17 abstracts presented in symposia published
Notes	Sociology ‐ social work Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	Yes	Searched 4 databases and searched faculty biography or curriculum vitae.
Follow‐up time?	Yes	The follow‐up was for Quote: "more than 4 years after the meeting".
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of type of presentation with publication using stratified analysis and Chi2 , Fisher's Exact test, one way ANOVA tests and logistic regression analysis.

Petticrew 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database Electronic database not specified; dates of search not reported Person completing the search not reported Search criteria not reported Matching criteria not reported Contacted abstract authors directly
Data	Included 77 abstracts presented at the 1996 Society of Social Medicine meeting Included all abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive'
Outcomes	39 of 77 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 18/36 'positive' (defined as experimental better than control) versus 21/41 not 'positive' abstract results published
Notes	Sociology Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation.
Search for publications?	Yes	Searched database, but number and type of database not reported, and contacted abstract authors.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results with publication using stratified analysis and calculated relative risks.

Petticrew 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Web of Science to February 2005 Search completed by information specialist Search criteria not reported Matched abstract to full‐length publication by Title Contents Contacted abstract authors directly
Data	Included 224 abstracts presented at the 1998 and 1999 British Sociological Association Medical Sociology meetings Included all abstracts accepted for oral presentations that described qualitative research
Comparisons	Proportion of abstracts published
Outcomes	99 of 224 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Sociology Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 2 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.

Pinheiro 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and Lilacs in May 2007 Person completing the search not reported Searched by first and last author, and keywords Matching criteria not reported
Data	Included 154 abstracts presented at the 2001 to 2005 American Society of Clinical Oncology meetings Included all abstracts of Brazilian studies
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published
Outcomes	26 of 154 abstracts published 2/12 abstracts presented at the 2001 meeting, 8/26 at the 2002 meeting, 7/46 at the 2003 meeting, 3/32 at the 2004 meeting, and 6/38 at the 2005 meeting published Proportion of abstracts published by time Median time to publication = 13.5 months Cumulative proportion of abstracts published at 78 months showed proportion published = 20.1% (31/154 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Oncology Funding not reported Abstracts may also be included in report by Chan 2009, Costa 2009
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific country, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	The 2004 and 2005 meeting had less than 48 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis and log rank analysis.

Polyzos 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database Cochrane Central Register of Controlled Trials, Embase, PubMed to December 2010 Person completing the search not reported Searched by first, second, last, and corresponding author Matching criteria not reported
Data	Included 155 abstracts presented at the 2003 and 2004 European Society of Human Reproduction and Embryology meetings Included all abstracts of randomized controlled trials related to reproductive medicine
Comparisons	Proportion of abstracts published Median time to publication 'Positive' versus not 'positive' results 'Positive' versus not 'positive' conclusions Sample size greater than 250 versus between 101 and 250 versus equal to or less than 100 Oral versus poster presentation Multi‐centered versus single center North American versus European versus rest of the world origin Reporting on commercial drug versus not reporting on commercial drug Subspecialty
Outcomes	89 of 155 abstracts published Proportion of abstracts published by time Median time to publication = 15 months (range = 0 to 75 months) Factors related to proportion of abstracts published included 49/83 'positive' (defined as significant results) versus 40/72 not 'positive' abstract results published 44/65 'positive' (defined as experimental better than control) versus 45/89 not 'positive' abstract results published 59/95 abstracts with a 'positive' conclusion versus 30/60 abstracts with a negative or neutral conclusion published 12/18 abstracts with sample size greater than 250 versus 24/41 abstracts with sample size between 101 and 250 versus 44/72 abstracts with sample size equal to or less than 100 published 48/71 abstracts presented orally versus 41/84 abstracts presented as posters published 19/26 abstracts with multiple centers versus 7/129 abstracts with a single center published 5/14 abstracts originating from North America versus 47/79 abstracts from Europe versus 37/62 abstracts from rest of the world published 56/94 abstracts reporting on a commercial drug versus 33/61 abstracts not reporting on a commercial drug published
Notes	Human reproduction and development ‐ reproductive medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific study design and specific topic, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, sample size, type of presentation, multi‐center status, country of origin, commercial drug, and subspecialty with publication using multivariable logistic regression analysis.

Post 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed in April 2012 Search completed by the investigators Searched by first author Matched abstract to full‐length publication by Title Contents Abstract
Data	Included 1329 abstracts presented at the 2007 and 2008 North American Primary Care Research Group and Society of Teachers of Family Medicine meetings Included all abstracts except educational sessions, workshops, and breakfast roundtable sessions
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall, by meeting and by type of presentation Oral versus poster presentation
Outcomes	457 of 1329 abstracts published 367/1000 abstracts presented at the North American Primary Care Research Group meetings and 90/329 at the Society of Teachers of Family Medicine meetings published Proportion of abstracts published by time Mean time to publication = 15.4 months (SE = 0.6; range = ‐20 to 51 months), overall Mean time to publication = 16 months (SE = 0.6) for the 2007 to 2008 North American Primary Care Research Group meetings Mean time to publication = 13.1 months (SE = 1.2) for the 2007 to 2008 Society of Teachers of Family Medicine meetings Mean time to publication = 13.7 months (SE = 0.6) for abstracts presented orally Mean time to publication = 18.6 months (SE = 1.0) for abstracts presented as poster Factors related to proportion of abstracts published included 295/688 abstracts presented orally versus 162/641 abstracts presented as poster published
Notes	General medicine/primary care Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting and type of presentation with publication using stratified analysis and unspecified statistical tests.

Prendergast 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database The Cochrane Library, Embase, MEDLINE, Web of Science, Ovid, and Wiley Interscience databases until September 2012 Person completing the search not reported Searched by first and senior author, and keywords Matching criteria not reported
Data	Included 938 abstracts presented at the 1995 and 2005 British Society of Gastroenterology meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting Proportion published by study design and subspecialty between 1995 and 2005 meetings
Outcomes	165 of 938 abstracts published 88/497 abstracts presented at the 1995 meeting and 77/441 at the 2005 meeting published Proportion of abstracts published by time Mean time to publication = 22.2 months (SD = 16.9) for the 1995 meeting Mean time to publication = 18.6 months (SD = 13.3) for the 2005 meeting Examined individual factors related to study design and subspecialty for comparison of proportion of abstracts published between 1995 and 2005 meetings
Notes	Gastroenterology Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 6 databases.
Follow‐up time?	Yes	All meetings had more than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis using Chi2 or t tests.

Preston 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to January 2004 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Contents Study institution (for posters)
Data	Included 815 abstracts presented at the 1994 to 1998 Orthopaedic Trauma Association meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication by meeting and type of presentation Oral versus poster presentation
Outcomes	476 of 815 abstracts published 96/153 abstracts presented at the 1994 meeting, 98/168 at the 1995 meeting, 79/156 at the 1996 meeting, 91/157 at the 1997 meeting, and 114/181 at the 1998 meeting published Proportion of abstracts published by time Mean time to publication = 24.8 months (SD = 17.9) for oral presentations and 21.6 months (SD = 19.4) for poster presentations Mean time to publication = 21.9 months (SD = 18.1) for oral presentations and 28.2 months (SD = 27.8) for poster presentations for 1994 Orthopaedic Trauma Association meeting Mean time to publication = 26.1 months (SD = 20.3) for oral presentations and 19.7 months (SD = 18.8) for poster presentations for 1995 Orthopaedic Trauma Association meeting Mean time to publication = 31.1 months (SD = 19.3) for oral presentations and 22.1 months (SD = 17.3) for poster presentations for 1996 Orthopaedic Trauma Association meeting Mean time to publication = 20.9 months (SD = 13.8) for oral presentations and 18.6 months (SD = 15.9) for poster presentations for 1997 Orthopaedic Trauma Association meeting Mean time to publication = 25.4 months (SD = 16.8) for oral presentations and 21.3 months (SD = 17.3) for poster presentations for 1998 Orthopaedic Trauma Association meeting Factors related to proportion of abstracts published included 222/329 abstracts presented orally versus 254/486 abstracts presented as posters published
Notes	Emergency medicine ‐ orthopedic trauma Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and type of presentation with publication using stratified analysis and Fisher's Exact test.

Prohaska 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar from 2005 to October 2011 Person completing the search not reported Searched by first and second author, and keywords Matched abstract to full‐length publication by Sample size Title Methodology Results Objectives
Data	Included 1941 abstracts presented at the 2005 American College of Clinical Pharmacy Annual and Spring meetings; and American Pharmacists Association meetings; and American Society of Health‐System Pharmacists Summer and Midyear Clinical meetings Included all abstracts except 58 abstracts previously published in meeting organization's published abstract listing
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication Cumulative proportion of abstracts published Subspecialty Practice setting
Outcomes	384 of 1941 abstracts published 85/826 abstracts presented at the American Society of Health‐System Pharmacists Midyear Clinical meeting, 16/114 at the American Society of Health‐System Pharmacists Summer meeting, 55/281 at the American Pharmacists Association Annual meeting, 61/217 at the American College of Clinical Pharmacy Spring meeting, and 167/503 at the American College of Clinical Pharmacy Annual meeting published Proportion of abstracts published by time Mean time to publication = 16.8 months (SD = 11.9) Median time to publication = 15 months Cumulative proportion of abstracts published at 72 months showed proportion published = 18.9% (377/1941 abstracts) No factors other than subspecialty and practice patterns related to proportion of abstracts published reported
Notes	Pharmacology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had 72 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, practice setting, and subspecialty with publication using stratified analysis and Chi2 tests.

Propst 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors and title Matched abstract to full‐length publication by Objective Data
Data	Included 867 abstracts presented at the 2004 to 2012 Society of Gynecologic Surgeons Annual Scientific meetings Included all abstracts except tips and tricks and video presentations
Comparisons	Proportion of abstracts published, overall and % by meeting Mean time to publication Oral versus oral poster versus poster presentation University versus community setting Author location
Outcomes	483 of 867 abstracts published 95/165 presented at the 2004 meeting, 42/89 at the 2005 meeting, 56/99 at the 2006 meeting, 53/85 at the 2007 meeting, 51/81 at the 2008 meeting, 50/83 at the 2009 meeting, 35/76 at the 2010 meeting, 46/91 at the 2011 meeting, and 54/97 at the 2012 meeting published Proportion of abstracts published by time Mean time to publication = 12.7 months (SD=11.7; range = ‐5 to 82 months) Factors related to proportion of abstracts published included 169/205 abstracts presented orally versus 190/459 abstracts presented as poster versus 124/203 abstracts presented as oral poster published 465/804 abstracts with academic affiliation versus 16/61 abstracts without an academic affiliation published 462/826 abstracts with US first author versus 20/40 with author from elsewhere published
Notes	Gynecology/obstetrics ‐ gynecologic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year by type presentation, university affiliation, and author location with publication using stratified analysis and Chi2 or ANOVA tests.

Rabenda 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for '3 years' Search completed by the investigators Searched by first and second author, and keywords Matched abstract to full‐length publication by All authors Title Contents
Data	Included 619 abstracts presented at the 2011 European Congress on Osteoporosis, Osteoarthritis, and Musculoskeletal Disorders meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication, by type of presentation & publication Oral versus poster presentation
Outcomes	191 of 619 abstracts published Proportion of abstracts published by time Time to publication (not stated if mean or median) = 14.8 months (IQR = 8.3 to 23.3 months) for oral e‐publications Time to publication (not stated if mean or median) = 16.3 months (IQR = 8.4 to 23.3 months) for oral print publications Time to publication (not stated if mean or median) = 9.2 months (IQR = 2.7 to 17.7 months) for poster e‐publications Time to publication (not stated if mean or median) = 11.3 months (IQR = 5.3 to 21.4 months) Factors related to proportion of abstracts published included 34/45 abstracts presented orally versus 157/574 abstracts presented as poster published
Notes	Orthopedics/orthopedic surgery ‐ musculoskeletal disorders Funding by ESCEO and IOF
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meeting had less than 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of type of presentation with publication using stratified analysis and Chi2 t test, or Mann Whitney U tests.

Ramsey 1995.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by all authors Matching criteria not reported
Data	Included 249 abstracts presented at the 1991 American Association of Blood Banks meeting Included all abstracts
Comparisons	Proportion of abstracts published Oral versus poster presentation Commercial versus non‐commercial versus mixed vendors Subspecialty Origin
Outcomes	89 of 249 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 51% of abstracts presented orally versus 23% of abstracts presented as posters published 18% of abstracts with 'commercial venders' versus 38% of abstracts with 'non‐commercial sources' versus 44% of abstracts with 'commercial/non‐commercial collaborations' published Reported that 'Publication rate was not influenced by abstract category or foreign origin.'
Notes	Hematology ‐ transfusion medicine, blood bank Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation, funding, subspecialty and country of origin with publication using stratified analysis.

Rao 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to July 2004 Person completing the search not reported Searched by all authors and title Matched abstract to full‐length publication by Contents Methodology
Data	Included 449 abstracts presented at the 2001 and 2002 British Association of Urological Surgeons meetings Included all abstracts
Comparisons	Proportion of abstracts published Survival analysis of publication rate Oral versus poster presentation United Kingdom versus elsewhere origin
Outcomes	142 of 449 abstracts published Proportion of abstracts published by time Survival analysis of proportion published at 36 months = 43.0% Factors related to proportion of abstracts published included 38/124 abstracts presented orally versus 104/325 abstracts presented as posters published 121/391 abstracts originating in the United Kingdom versus 21/58 abstracts originating elsewhere published
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings had less than 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of publication year, type presentation, and UK origin with publication using stratified and survival analyses.

Ravn 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, Embase, PubMed until May 2013 Search completed by the investigators Searched by all authors, keywords, and title Matched abstract to full‐length publication by Sample size Methodology Results Contacted abstract authors directly
Data	Included 118 abstracts presented at the 2009 to 2011 Danish Society for Emergency Medicine meetings Included all abstracts, but only 63 of 118 abstracts have a follow‐up of 24 months
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation
Outcomes	17 of 63 abstracts with 24 months or more follow‐up published; 35 of 118 of all abstracts published 6/19 abstracts presented at the 2009 meeting, 11/44 at the 2010 meeting, and 8/55 at the 2011 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 4/10 abstracts presented orally versus 21/108 abstracts presented as poster published, but abstracts with less than 24 months included
Notes	Emergency medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors.
Follow‐up time?	No	Only the meeting in 2009 had 48 months follow‐up. All other meetings had less than 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and type presentation with publication using Fisher's Exact test.

Richling 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar until March 2014 Search completed by the investigators Searched by first author, keywords, and title Matched abstract to full‐length publication by Results Hypotheses Procedures Participant characteristics Data sets Searched curriculum vitae or resumes of first authors Contacted abstract authors directly
Data	Included 880 abstracts presented at the 2005 Association for Behavior Analysis meeting Included all abstracts related to autism or developmental disabilities
Comparisons	Proportion of abstracts published Symposium versus other oral versus poster presentation Continuing education approval versus no continuing education approval versus ineligible for applying for continuing education credit
Outcomes	77 of 880 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 25/278 abstracts presented at symposia versus 8/55 abstracts presented orally versus 44/547 abstracts presented as poster published 8/61 abstracts with continuing education approval versus 17/217 abstracts without continuing education approval versus 52/602 abstracts ineligible for continuing education approval published
Notes	Human reproduction and development ‐ autism and developmental intellectual disabilities Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts relating to a specific topic, so low risk of bias.
Search for publications?	Yes	Searched 2 databases, curriculum vitae of authors and contacted abstract authors directly.
Follow‐up time?	Yes	The meeting had 108 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	Yes	Examined association of type presentation and continuing education approval with publication using stratified analysis.

Riessland 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to March 2003 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by 'Data'
Data	Included 347 abstracts presented at the 1995 American Society for Clinical Pharmacology meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Oral versus poster presentation 'Type data'
Outcomes	182 of 347 abstracts published Proportion of abstracts published by time Mean time to publication = 31 months (SD = 47) Median time to publication = 31 months Cumulative proportion of abstracts published at 96 months showed proportion published = 53.0% (184/347 abstracts) Factors related to proportion of abstracts published included 12/24 abstracts presented orally versus 170/323 abstracts presented as poster published
Notes	Pharmacology Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 8 years follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of type of presentation and 'type data' with publication using stratified analysis.

Riordan 2000.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to December 1999 Search completed by investigator Searched by first author Matching criteria not reported Contacted abstract authors directly
Data	Included 88 abstracts presented at the 1996 British Paediatric Association and the Pediatric Research Society spring and autumn meetings Included all abstracts accepted for plenary presentation
Comparisons	Proportion of abstracts published, overall and by meeting RCT design versus non‐RCT design Academic versus non‐academic affiliation
Outcomes	55 of 88 abstracts published 24/48 abstracts presented at the Pediatric Research Society spring and autumn meetings, and 31/40 at the British Paediatric Association meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 8/9 abstracts with RCT design versus 47/79 abstracts with non‐RCT design published 15/28 abstracts originating in an academic center versus 13/20 abstracts not originating in an academic center published
Notes	Pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for plenary presentation.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	No	The meeting only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting, study design, academic affiliation with publication using stratified analysis.

Rodriguez 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2010 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 311 abstracts presented at the 2006 to 2009 American Association of Oral and Maxillofacial Surgeons meetings Included all abstracts with authors from the US
Comparisons	Proportion of abstracts published, overall and by meeting Oral versus poster presentation
Outcomes	85 of 311 abstracts published 4/68 abstracts presented at the 2006 meeting, 5/85 at the 2007 meeting, 13/90 at the 2008 meeting, and 9/68 at the 2009 meeting published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 49/163 abstracts presented orally versus 36/148 abstracts presented as poster published
Notes	Ortolaryngology head and neck surgery ‐ oral and maxillofacial surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	Only the meeting in 2006 had 48 months follow‐up. All other meetings had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and type of presentation with publication using stratified analysis and unspecified statistical tests

Rollin 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Embase from January 1998 to December 2006 Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by One author Title Methodology
Data	Included 318 abstracts presented at the 2000 International Commission on Occupational Health meeting Included a random sample, stratified by oral and poster presentations
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published Oral versus poster presentation English‐speaking country versus non‐English speaking country North American versus Western European versus South American versus Asian versus Russian and Eastern European versus African versus Australian origin Authors from countries with GNP < or > $3000 Quantitative data versus no quantitative data Type of occupational risk
Outcomes	105 of 318 abstracts published Proportion of abstracts published by time Mean time to publication = 17 months (95% CI = 13 to 21 months) Cumulative proportion of abstracts published at 72 months showed proportion published = 33% (105/318 abstracts) Factors related to proportion of abstracts published included 49/158 abstracts presented orally versus 56/160 abstracts presented as posters published 18/32 abstracts from English‐speaking country versus 91/286 abstracts from non‐English speaking country published 7/19 abstracts originating from North America versus 66/168 abstracts from Western Europe versus 2/11 abstracts from South America versus 26/89 abstracts from Asia versus 3/12 abstracts from Russia and Eastern Europe versus 0/16 abstracts from Africa versus 1/3 abstracts from Australia published 3/40 abstracts from authors of nations with GNP <$3000 versus 102/278 abstracts from authors of nations with GNP > $3000 published 97/242 abstracts with quantitative data versus 8/76 abstracts without quantitative data published
Notes	Other clinical specialties ‐ occupational health Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	The meeting had 6 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of type of presentation, origin in English language country, country of origin, economic status of country, presence of quantitative data, and type of occupational risk with publication using multivariable logistic regression analyses.

Rosmarakis 2005.

Methods	Identification of subsequent full‐length publications Searched database Index Medicus to March 2004 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 190 abstracts presented at the 1999 and 2000 Interscience Conferences on Antimicrobial Agents and Chemotherapy meetings Included all abstracts presented in the first session of 7 of 15 thematic categories
Comparisons	Proportion of abstracts published
Outcomes	68 of 190 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Infectious disease/immunology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Roukis 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, Infotrieve‐PubMed/MEDLINE, OVID MEDLINE(R), Google to October 2010 Search completed by investigator Searched by all authors and keywords Matching criteria not reported
Data	Included 206 abstracts presented at the 1999 to 2008 American College of Foot and Ankle Surgeons Annual Scientific Conference meetings Included all abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting
Outcomes	139 of 206 abstracts published 20/29 abstracts presented at the 1999 meeting, 19/29 at the 2000 meeting, 10/20 at the 2001 meeting, 14/20 at the 2002 meeting, 11/13 at the 2003 meeting, 10/14 at the 2004 meeting, 15/21 at the 2005 meeting, 15/21 at the 2006 meeting, 15/21 at the 2007 meeting, and 10/18 at the 2008 meeting published Proportion of abstracts published by time Mean time to publication = 14.5 months (range = ‐34 to 60 months), overall Mean time to publication = 13.3 months (range = ‐34 to 60 months) for 1999 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting Mean time to publication = 19.3 months (range = 2 to 50 months) for 2000 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting Mean time to publication = 18.6 months (range = 12 to 34 months) for 2001 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting Mean time to publication = 13.7 months (range = 2 to 42 months) for 2002 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting Mean time to publication = 16.9 months (range = 6 to 36 months) for 2003 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting Mean time to publication = 8.4 months (range = 0 to 24 months) for 2004 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting Mean time to publication = 17.9 months (range = ‐2 to 46 months) for 2005 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting Mean time to publication = 11.5 months (range = ‐5 to 32 months) for 2006 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting Mean time to publication = 14.7 months (range = ‐4 to 36 months) for 2007 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting Mean time to publication = 7.6 months (range = ‐1 to 17 months) for 2008 American College of Foot and Ankle Surgeons Annual Scientific Conference meeting No factors related to proportion of abstracts published reported.
Notes	Surgery ‐ foot and ankle Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings before 2007 had at least 48 months follow‐up. All other meetings had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Roy 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Contents
Data	Included 660 abstracts presented at the 1978 to 1995 Otorhinolaryngological Research Society meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published Subspecialty
Outcomes	456 of 660 abstracts published Proportion of abstracts published by time Mean time to publication = 22.5 months (range = ‐60 to 108 months) Cumulative proportion of abstracts published at 72 months showed proportion published = 69.1% (456/660 abstracts) No factors other than subspecialty related to proportion of abstracts published reported
Notes	Otolaryngology head and neck surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of subspecialty with publication using stratified analysis.

Rubin 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from May 2004 to November 2013 Search completed by the investigators Searched by first author, last author, title and senior author Matched abstract to full‐length publication by All authors Title Methodology Contacted abstract authors directly
Data	Included 847 abstracts presented at the 2004 to 2007 Digestive Disease Week meetings Included all abstracts included in the Best of Digestive Disease Week published after the Digestive Disease Week meetings
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published Oral versus poster presentation US origin versus international origin
Outcomes	408 of 847 abstracts published 147/340 abstracts presented at the 2004 meeting, 81/160 at the 2005 meeting, 79/147 at the 2006 meeting, and 101/200 at the 2007 meeting published Proportion of abstracts published by time Median time to publication = 12 months (IQR = 12 to 24 months) Cumulative proportion of abstracts published at 96 months showed proportion published = 48.2% (408/847 abstracts) Factors related to proportion of abstracts published included 155/255 abstracts presented orally versus 250/592 abstracts presented as a poster published 156/317 abstracts originating in the US versus 249/530 abstracts not originating in the US published
Notes	Gastroenterology ‐ endoscopy Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Unclear	Examined association of meeting year, type of presentation, US origin, subspecialty by type of presentation, and subspecialty by US origin with publication using Fisher's Exact test and logistic regression analyses, although it was unclear if multivariable analyses were completed.

Saad 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and LILACS; dates of search not reported Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 154 abstracts published in the abstract book for the 2001 to 2005 American Society of Clinical Oncology meetings (included some 'publish only' abstracts) Included abstracts from Brazil, defined those in which at least 2/3rds if the involved institutions were from Brazil
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published
Outcomes	26 of 154 abstracts published 2/12 abstracts presented at the 2001 meeting, 8/26 at the 2002 meeting, 7/46 at the 2003 meeting, 3/32 at the 2004 meeting, and 6/38 at the 2005 meeting published Proportion of abstracts published by time Median time to publication = 13.5 months Proportion of abstracts published at 24 months = 13.6% (21/154 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Oncology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified and log rank analysis.

Saeed 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase, CINAHL, Web of Science to November 2009 Search completed by investigator Searched by first, second, and last author Matching criteria not reported
Data	Included 141 abstracts submitted to the 2002, 2004, and 2006 Canadian Blood and Marrow Transplant Group meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation Clinical research versus 'non‐clinical' research Multi‐centered versus single center Funding versus no funding or not reported
Outcomes	43 of 141 abstracts published 21/57 abstracts presented at the 2002 meeting, 12/50 at the 2004 meeting, and 10/34 at the 2006 meeting published Proportion of abstracts published by time Median time to publication = 20.7 months (range = ‐5.3 to 91.2 months) Cumulative proportion of abstracts published at 48 months showed proportion published = 28.3% (40/141 abstracts) Factors related to proportion of abstracts published included 20/53 'positive' (defined as significant results) versus 23/88 not 'positive' abstract results published 13/22 abstracts presented orally versus 30/119 abstracts presented as poster published 33/86 abstracts describing clinical research versus 10/55 abstracts describing 'non‐clinical' research published 10/22 abstracts with multiple centers versus 33/119 abstracts with a single center published 7/10 abstracts with funding versus 36/131 abstracts with no funding or not reporting funding published
Notes	Oncology ‐ blood and marrow transplantation Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Yes	All meetings before 2006 had at least 5 years follow‐up. The meeting in 2006 only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, positive results, type of presentation, type science, multi‐center status, and funding status with publication using Chi2 tests.

Salami 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed in May 2012 Person completing the search not reported Searched by first and last author Matching criteria not reported
Data	Included 184 abstracts presented at the 2005 American Society of Pediatric Hematology/Oncology meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication 'Positive' versus not 'positive' Oral versus poster presentation Clinical research versus basic science research versus other types of research Multi‐centered versus single center North American versus European versus Asian versus rest of the world origin
Outcomes	114 of 184 abstracts published Proportion of abstracts published by time Median time to publication = 19 months (IQR = 11 to 29 months) Factors related to proportion of abstracts published included 46/66 abstracts with 'positive' (defined as statistically significant) versus 5/10 not 'positive' abstract results published 86/123 abstracts with 'positive' (defined as experimental better than control) versus 28/61 not 'positive' abstract results published 17/22 abstracts presented orally versus 91/142 abstracts presented as poster published 53/93 abstracts describing clinical research versus 49/67 abstracts describing basic science research versus 12/24 abstracts describing other types of research published 44/63 abstracts with multiple centers versus 70/121 abstracts with a single center published 110/170 abstracts originating from North America versus 3/5 abstracts originating from Europe versus 0/7 abstracts originating from Asia versus 1/1 with rest of the world origin published
Notes	Oncology ‐ pediatric hematology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Unclear	Included a subgroup of abstracts but definition not reported.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 7 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, positive results, type of presentation, type science, multi‐center status, and country of origin with publication using stratified analysis and Fisher's Exact tests.

Saldanha 2016.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Cochrane Central Register of Controlled Trials, Embase, Web of Science, SCOPUS, and LILACS to June 2013 Search completed by investigator Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 513 abstracts presented at the 2001 to 2004 Association for Research in Vision and Ophthalmology meetings Included all abstracts of RCTs, except 32 with non‐randomized comparisons nested within RCTs
Comparisons	Proportion of abstracts published Median time to publication Cumulative time to publication 'Positive' versus 'not positive' Oral versus poster Multi‐center versus single center One or more reported conflict of interests versus none Funding source reported versus not reported Funding by industry versus other Academic center versus non‐academic center affiliation
Outcomes	230 of 513 abstracts published Proportion of abstracts published by time Median time to publication = 18 months (IQR = 11 to 33; range = 1 to 90 months) Cumulative proportion of abstracts published at 120 months showed proportion published = 44.8% (230/513 abstracts) Factors related to proportion of abstracts published included 64/117 'positive' (defined as statistically significant results) versus 59/111 not 'positive' versus 107/285 abstracts with statistical results not reported published 48/84 'positive' (defined by direction of results) versus 29/57 not 'positive' abstract results published (derived from 130 abstracts from 2001 to 2003 and reported by in Rodriguez Barraquer 2009) 51/95 abstracts presented orally versus 179/418 abstracts presented as posters published 62/106 abstracts with multiple centers versus 15/46 abstracts with single centers versus 153/361 abstracts with number of centers not reported published 39/103 abstracts with authors reporting at least one conflict of interest versus 113/272 abstracts with no author reporting a conflict of interest published 34/56 abstracts reporting an industry sponsor versus 104/241 abstracts not reporting an industry sponsor published 118/258 abstracts originating in an academic center versus 112/255 abstracts not originating in an academic center versus published
Notes	Ophthalmology/optometry ‐ ophthalmology Funding from the National Eye Institute, National Institutes of Health
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 6 databases and contacted abstract authors directly.
Follow‐up time?	Yes	All meetings had at least 9 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, multi‐center status, author conflict of interest, funding, and academic affiliation with publication using log‐binominal models accounting for interaction between variables.

Sanders 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase; dates of search not reported Person completing the search not reported Searched by first, and senior authors, and keywords Matching criteria not reported
Data	Included 255 abstracts presented at the 1994 British Society of Gastroenterology meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published Concordant acceptance at second meeting 'Positive' versus 'not positive 'type, category or sample size' Sample size
Outcomes	178 of 255 abstracts published Proportion of abstracts published by time Median time to publication = 19 months (range = 0 to 66 months) Cumulative proportion of abstracts published at 63 months showed proportion published = 69.8% (178/255 abstracts) Factors related to proportion of abstracts published included 93 abstracts also submitted to the American Gastroenterology meeting, 58/73 accepted for presentation and published Reported that 'There were no statistical differences between publication rates when sub‐dividing abstracts according to type, category or sample size.'
Notes	Gastroenterology Funding not reported Data extracted from letter and abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of acceptance for presentation, type of science, study design, and sample size with publication using stratified analysis and Chi2 tests.

Sanossian 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to August 2005 Person completing the search not reported Searched by first, second, and last author, and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 353 abstracts presented at the 2000 International Stroke Conference of the American Stroke Association meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Sample size greater than 200 versus sample size less than 200 Oral versus poster presentation Multi‐centered versus single center Academic versus non‐academic affiliation US versus Canada versus Europe versus Asia/Australia versus Mexico/South America origin 'Prospective' versus 'retrospective' design Industry sponsor versus no industry sponsor Subspecialty
Outcomes	220 of 353 abstracts published Proportion of abstracts published by time Median time to publication = 15 months (range = ‐18 to 69 months) Cumulative proportion of abstracts published at 48 months showed proportion published = 60.1% (212/353 abstracts) Factors related to proportion of abstracts published included 136/220 'positive' (defined as experimental better than control) versus 84/133 not 'positive' abstract results published 93/148 abstracts with sample size greater than 200 versus 127/205 abstracts with sample size less than 200 published 82/108 abstracts presented orally versus 138/245 abstracts presented as posters published 77/126 abstracts with multiple centers versus 143/227 abstracts with a single center published 174/269 abstracts originating in an academic center versus 46/84 abstracts not originating in an academic center versus published 153/251 abstracts originating from the US versus 21/28 abstracts from Canada versus 39/61 abstracts from Europe versus 25/36 abstracts from Asia/Australia versus 1/1 abstract from Mexico/South America published 94/160 abstracts with 'prospective' study design versus 34/62 abstracts with 'retrospective' study design published 26/37 abstracts reporting an industry sponsor versus 183/316 abstracts not reporting an industry sponsor published
Notes	Neurology ‐ stroke Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, sample size, type of presentation, multi‐center status, academic affiliation, country of origin, prospective status, funding, and subspecialty with publication using multivariable logistic regression analysis.

Sanz Arrufat 2003.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase, IME, ICYT, and IPA to February 2003 Person completing the search not reported Searched by first and last author Matched abstract to full‐length publication by Title Contents Methodology
Data	Included 638 abstracts presented at the 1996 to 1998 Sociedad Española de Farmacia Hospitalaria meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Subspecialty Conference year
Outcomes	92 of 638 abstracts published 30/192 abstracts presented at the 1996 meeting, 27/202 at the 1997 meeting, and 35/244 at the 1998 meeting published Proportion of abstracts published by time not reported No factors other than meeting year related to proportion of abstracts published reported
Notes	Pharmacology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 5 databases.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Sawatsky 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Web of Science, and Google/GoogleScholar until December 2013 Search completed by the investigators Searched by first, second and last author, and keywords Matched abstract to full‐length publication by All authors Title Methodology Results
Data	Included 144 abstracts presented at the 2009 Society of General Internal Medicine meeting Included all abstracts describing medical education research excluding those focusing on patient education or general medical research
Comparisons	Proportion of abstracts published Mean time to publication, overall and by submission category Oral versus poster presentation Quantitative versus qualitative research Scientific versus innovative research Higher MERSQI (Medical Education Research Study Quality Instrument) score versus lower score
Outcomes	64 of 144 abstracts published Proportion of abstracts published by time Mean time to publication = 21 months Mean time to publication = 18.7 months (SD = 9.9) for abstracts in the scientific research category Mean time to publication = 25.4 months (SD = 10.9) for abstracts in the innovative research category Factors related to proportion of abstracts published included 15/21 abstracts presented orally versus 38/99 abstracts presented as poster published 53/120 abstracts describing quantitative research versus 11/24 abstracts describing qualitative research published 42/79 abstracts from the scientific research category versus 22/65 abstracts from the innovative research category published 27/44 with higher (> 10) MERSQI score versus 25/75 abstracts with lower (< 10) MRSQI score
Notes	Medical education/library science Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic with reasonable exceptions, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria
Adjustment for confounding?	No	Examined association of type of presentation, qualitative research, scientific research or abstract quality with publication using stratified analysis and Chi2 or t tests or Kruskal‐Wallis tests.

Scherer 1994.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from year of presentation through 1992 Search completed by investigator Searched by all authors Matched abstract to full‐length publication by All authors Contents Contacted abstract authors directly
Data	Included 93 abstracts presented at the 1988 and 1989 American Academy of Ophthalmology and Association for Research in Vision and Ophthalmology meetings Included all abstracts of randomized controlled trials as verified by abstract author.
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published 'Positive' versus not 'positive' Sample size equal to or above the median versus sample size below the median Multi‐centered versus single center 'High' versus 'low' quality as defined by author
Outcomes	61 of 93 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 48 months showed proportion published = 63.4% (59/93 abstracts) Factors related to proportion of abstracts published included 33/46 'positive' (defined as significant results) versus 28/47 not 'positive' abstract results published 31/43 abstracts with sample size equal to or above the median versus 24/42 abstracts with sample size below the median 14/19 abstracts with multiple centers versus 45/71 abstracts with a single center published 31/43 abstracts rated by author as 'high' quality versus 29/47 abstracts rated as 'low' quality published
Notes	Ophthalmology/optometry ‐ ophthalmology Government funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific study design, so low risk of bias.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	The meeting in 1988 had 48 months and the meeting in 1989 only 36 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of positive results, sample size, trial design, multi‐center status, abstract quality, and subspecialty with publication using stratified analysis and Chi2 tests.

Schibilisky 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database 'Bibliometric study'; names of databases and dates of search not reported Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 777 abstracts submitted to the 2008 International Society for Heart and Lung Transplantation meeting Included all abstracts
Comparisons	Proportion of abstracts published Accepted versus rejected for presentation Oral versus poster Academic affiliation Multi‐centered with 2 or more centers Prospective cohort study design Sample size
Outcomes	259 of 777 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 189/529 abstracts accepted for presentation versus 71/248 abstracts rejected for presentation published 122/319 abstracts presented orally versus 65/210 abstracts presented as posters published 209/650 abstracts describing clinical research versus 49/128 abstracts describing basic science published 238/683 abstracts with a university affiliation versus 20/94 abstracts without a university affiliation published 98/233 abstracts describing multi‐center studies versus 180/564 abstracts describing single center studies published 7/29 abstracts describing case reports versus 65/232 abstracts describing case series versus 11/25 abstracts describing case control studies versus 3/8 abstracts describing cross‐sectional studies versus 79/262 describing retrospective cohort studies versus 28/55 studies describing prospective cohort studies versus 9/18 non‐randomized trials versus 8/18 RCTs published Enrolment of 150 or more patients: OR 3.26 (95% CI:1.04 to 7.91)
Notes	Surgery ‐ heart and lung transplantation Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Unclear	Number and types of databases not reported.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of acceptance for presentation, type of presentation, multi‐center status, academic affiliation, prospective status, and sample size with publication using multivariable logistic regression.

Schnatz 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to June 2007 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Title Contents Methodology
Data	Included 661 abstracts presented at the 1999 to 2003 North American Menopause Society meetings Included all abstracts except those from invited speakers
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication overall and by type of presentation Oral versus poster presentation
Outcomes	253 of 661 abstracts published 55/138 abstracts presented at the 1999 meeting, 33/92 at the 2000 meeting, 39/104 at the 2001 meeting, 52/154 at the 2002 meeting, and 73/173 at the 2003 meeting published Proportion of abstracts published by time Mean time to publication = 24 months (SD = 18; maximum = 103.2 months) Mean time to publication = 20.4 months (SD = 15.6) for oral presentations Mean time to publication = 24 months (SD = 18) for poster presentations Factors related to proportion of abstracts published included 30/52 abstracts presented orally versus 222/609 abstracts presented as posters published
Notes	Gynecology/obstetrics ‐ menopause Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year and type presentation with publication using stratified analysis and Chi2 or t tests.

Schoenfeld 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar in spring of 2010 Search completed by investigator Searched by all authors, keywords, and lead author Matching criteria not reported
Data	Included 770 abstracts presented at the 1998 to 2006 Society of Military Orthopaedic Surgeons meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting Cumulative proportion of abstracts published Subspecialty (secondary publication by Fuller 2011)
Outcomes	352 of 770 abstracts published 41/96 abstracts presented at the 1998 meeting, 36/89 at the 1999 meeting, 52/104 at the 2000 meeting, 35/75 at the 2001 meeting, 42/84 at the 2002 meeting, 43/83 at the 2003 meeting, 38/86 at the 2004 meeting, 33/76 at the 2005 meeting, and 32/78 at the 2006 meeting published Proportion of abstracts published by time Mean time to publication = 32.4 months, overall Mean time to publication = 43.0 months for the 1998 meeting Mean time to publication = 33.0 months for the 1999 meeting Mean time to publication = 33.6 months for the 2000 meeting Mean time to publication = 30.6 months for the 2001 meeting Mean time to publication = 39.6 months for the 2002 meeting Mean time to publication = 29.4 months for the 2003 meeting Mean time to publication = 27.0 months for the 2004 meeting Mean time to publication = 28.5 months for the 2005 meeting Mean time to publication = 19.2 months for the 2006 meeting Cumulative proportion of abstracts published at 120 months showed proportion published = 47.5% (366/770 abstracts) No factors other than subspecialty related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis.

Scholey 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1 year before the meeting to 5 years after the meeting Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 546 abstracts presented at the 1993 European Organisation for Caries Research, European Orthodontic Society, and International Association for Dental Research meetings Included all abstracts from the European Organisation for Caries Research and European Orthodontic Society meetings and a 10% random sample from the International Association for Dental Research meeting
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication overall and by meeting Cumulative proportion of abstracts published Oral versus poster presentation
Outcomes	252 of 546 abstracts published 116/256 abstracts presented at the International Association for Dental Research meeting, 78/175 at the European Orthodontic Society meeting, and 58/115 at the European Organisation for Caries Research meeting published Proportion of abstracts published by time Median time to publication = 18.0 months (IQR = 9 to 30 months) Median time to publication = 13.0 months (IQR = 5.3 to 24.5 months) for European Organisation for Caries Research meeting Median time to publication = 23.5 months (IQR = 12.3 to 36.0 months) for European Orthodontic Society meeting Median time to publication = 17.0 months (IQR = 8.8 to 29.3 months) for International Association for Dental Research meeting Cumulative proportion of abstracts published at 54 months showed proportion published = 46.2% (252/546 abstracts) Factors related to proportion of abstracts published included 99/173 abstracts presented orally versus 153/373 abstracts presented as posters published
Notes	Oral health ‐ dentistry Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included either all abstracts or a random selection of abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and type presentation with publication using stratified analysis and Chi2 tests.

Schulte 2012a.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from November 2003 to November 2008 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Hypothesis Methods Sample size Results
Data	Included 1100 abstracts presented at the 2003 German Society of Orthopaedics and Trauma Surgery meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation RCT design versus non‐RCT design 'Experimental' versus 'clinical' versus epidemiological studies versus meta‐analysis/reviews versus case reports versus other design 'Level I' versus 'level II' versus 'level III' versus 'level IV' evidence
Outcomes	392 of 1100 abstracts published Proportion of abstracts published by time Mean time to publication = 14.9 months (SD = 16.9; range = ‐44 to 58 months) Cumulative proportion of abstracts published at 60 months showed proportion published = 35.6% (392/1100 abstracts) Factors related to proportion of abstracts published included 120/236 'positive' (defined as significant results) versus 15/45 not 'positive' abstract results published 329/900 abstracts presented orally versus 63/200 abstracts presented as posters published 19/44 abstracts with RCT design versus 373/1056 abstracts with non‐RCT design published 154/327 abstracts of 'experimental' studies versus 204/662 abstracts of 'clinical' studies versus 11/22 abstracts of epidemiological studies versus 3/10 abstracts of 'meta‐analysis/reviews' versus 12/36 abstracts of case reports versus 8/43 abstracts of other types of studies published 16/29 abstracts of 'level I' evidence versus 33/83 abstracts of 'level II' evidence versus 19/65 abstracts of 'level III' evidence versus 136/485 abstracts of 'level IV' evidence published
Notes	Orthopedics/orthopedic surgery ‐ orthopedic and trauma surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Yes	Matched by 5 criteria.
Adjustment for confounding?	No	Examined association of positive results, type presentation, study design, type science, and level of evidence with publication using stratified analysis and Chi2 tests.

Schulte 2012b.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for 5 years following each meeting Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Sample size Methodology Results Hypothesis
Data	Included 839 abstracts presented at the 2000 to 2003 Spine Society of Europe meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication overall and by meeting Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation RCT design versus non‐RCT design Multi‐centered versus single center 'Experimental' versus 'clinical' versus epidemiological studies versus meta‐analysis/reviews versus case reports 'Level I' versus 'level II' versus 'level III' versus 'level IV' evidence 'Prospective' versus 'retrospective' design
Outcomes	317 of 839 abstracts published 78/259 abstracts presented at the 2000 meeting, 86/213 at the 2001 meeting, 68/168 at the 2002 meeting, and 85/199 at the 2003 meeting published Proportion of abstracts published by time Mean time to publication = 17.7 months (SD = 15.7; range = ‐36 to 60 months) Mean time to publication = 16.2 months (SD = 16.5; range = ‐36 to 60 months) for 2000 Spine Society of Europe meeting Mean time to publication = 18.3 months (SD = 17.0; range = ‐30 to 59 months) for 2001 Spine Society of Europe meeting Mean time to publication = 14.4 months (SD = 14.4; range = ‐9 to 57 months) for 2002 Spine Society of Europe meeting Mean time to publication = 21.2 months (SD = 14.2; range = ‐2 to 56 months) for 2003 Spine Society of Europe meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 37.8% (317/839 abstracts) Factors related to proportion of abstracts published included 140/270 'positive' (defined as significant results) versus 159/513 not 'positive' abstract results published 154/318 abstracts presented orally versus 163/521 abstracts presented as posters published 31/59 abstracts with RCT design versus 286/780 abstracts with non‐RCT design published 17/30 abstracts with multiple centers versus 165/502 abstracts with a single center published 124/268 abstracts of 'experimental' studies versus 182/532 abstracts of 'clinical' studies versus 7/19 abstracts of epidemiological studies versus 2/6 abstracts of 'meta‐analysis/reviews' versus 2/14 abstracts of case reports published 22/50 abstracts with 'level I' evidence versus 53/129 abstracts of 'level II' evidence versus 33/77 abstracts of 'level III' evidence versus 74/275 abstracts of 'level IV' evidence versus 0/1 of 'level V' published 105/255 abstracts with 'prospective' study design versus 85/296 abstracts with 'retrospective' study design published
Notes	Orthopedics/orthopedic surgery ‐ spine surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had 5 years follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, positive results, type of presentation, study design, multi‐center status, level of evidence, and prospective status with publication using stratified analysis and Chi2 tests.

Schwartz 1992.

Methods	Identification of subsequent full‐length publications Searched database Index Medicus; date of search not reported Person completing the search not reported Searched by first author Matching criteria not reported Searched all issues of the Journal of Air Medical Transport Contacted abstract authors directly
Data	Included 312 abstracts presented at the 1987 to 1990 Association of Air Medical Transport meetings Included all abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published
Outcomes	56 of 312 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Emergency medicine ‐ air medical services Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation.
Search for publications?	Yes	Searched 1 database, contacted abstract authors, and performed handsearches.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

Seaton 1981.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLARS (only for abstracts related to alarygeal speech) and 'dsh' abstracts for 2 years before to 8 years after presentation Person completing the search not reported Searched 'dsh' by all authors; search criteria for MEDLARS not reported Matched abstract to full‐length publication by Contents
Data	Included 583 abstracts presented at the 1967 to 1976 American Speech‐Language Hearing Association meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Subspecialty
Outcomes	174 of 583 abstracts published 13/50 abstracts presented at the 1967 meeting, 20/70 at the 1968 meeting, 15/55 at the 1969 meeting, 19/56 at the 1970 meeting, 20/42 at the 1971 meeting, 14/55 at the 1972 meeting, 21/48 at the 1973 meeting, 20/59 at the 1974 meeting, 15/66 at the 1975 meeting, and 17/82 at the 1976 meeting published Proportion of abstracts published by time not reported No factors other than meeting year related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ speech, language and communication Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had 8 years follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of meeting year and subspecialty with publication using stratified analysis.

Seaton 1983.

Methods	Identification of subsequent full‐length publications Contacted abstract authors directly 14 months after presentation followed by second questionnaire 26 months after presentation
Data	Included 625 abstracts presented at the 1978 American Speech‐Hearing‐Language Association meeting Included only abstracts with 'usable' questionnaires (625/696)
Comparisons	Proportion of abstracts published Median time to publication
Outcomes	202 of 625 abstracts published Proportion of abstracts published by time Median time to publication = 13 months No factors related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ speech, language and communication Funding by Ohio University's Baker Fund Awards Committee
Risk of bias
Item	Authors' judgement	Description
Sampling method?	No	Included only abstracts with 'usable' questionnaires.
Search for publications?	No	Contacted abstract authors but only had a response rate of 37.3%.
Follow‐up time?	No	The meeting only had 26 months follow‐up.
Matching?	Yes	Matched through contact with abstract author.

Sebel 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from January 1995 to December 1999 Search completed by investigator Searched by first, second, and third author Matching criteria not reported
Data	Included 581 abstracts presented at the 1995 International Anesthesia Research Society's Clinical and Scientific Congress meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published
Outcomes	277 of 581 abstracts published Proportion of abstracts published by time Mean time to publication = 15.6 months Cumulative proportion of abstracts published at 54 months showed proportion published = 47.7% (277/581 abstracts) No factors related to proportion of abstracts published reported
Notes	Anesthesiology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Secil 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to June 2004 Person completing the search not reported Searched by first, second, and last author, and keywords Matching criteria not reported
Data	Included 4413 abstracts presented at 1995 to 2002 'Turkish radiology conferences' meetings Included all abstracts except withdrawn abstracts Selected a single abstract when multiple abstracts of same study identified
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Oral versus poster presentation 'Scientific report' versus 'case report versus 'educational exhibit'
Outcomes	521 of 4413 abstracts published 35/336 abstracts presented at the 1995 meetings, 46/444 at the 1996 meetings, 47/368 at the 1997 meetings, 41/436 at the 1998 meetings, 66/539 at the 1999 meetings, 89/619 at the 2000 meetings, 104/821 at the 2001 meetings, and 93/850 at the 2002 meetings published Proportion of abstracts published by time Mean time to publication = 24.4 months (range = 0 to 91 months) Factors related to proportion of abstracts published included 116/754 abstracts presented orally versus 405/3659 abstracts presented as posters published 261/2116 abstracts describing a 'scientific report' versus 249/1995 abstracts describing case reports versus 11/302 abstracts describing 'educational exhibits' published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2001 had at least 48 months follow‐up. All meetings after 2000 had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, and study design with publication using stratified analysis.

Secil 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2004 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 276 abstracts presented at the 2000 and 2001 European Society of Gastrointestinal and Abdominal Radiology meetings Included all abstracts accepted for oral presentation except for 2 withdrawn abstracts and 6 published before the meeting
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Clinical research versus other research North American versus European versus rest of the world origin 'Prospective' versus 'retrospective' versus other design Interventional versus diagnostic Subspecialty Conference year
Outcomes	109 of 276 abstracts published 46/111 abstracts presented at the 2000 meeting, and 63/165 at the 2001 meeting published Proportion of abstracts published by time Median time to publication = 18 months (IQR = 1 to 53 months) Factors related to proportion of abstracts published included 109/264 abstracts describing clinical research versus 0/12 abstracts describing other research published 10/16 abstracts originating from North America versus 83/212 abstracts from Europe versus 16/48 abstracts from rest of the world published 70/153 abstracts with 'prospective' study design versus 39/111 abstracts with 'retrospective' study design versus 0/12 abstracts with other study design published 28/64 abstracts describing therapy versus 81/212 abstracts describing diagnostic tests published
Notes	Radiology ‐ gastrointestinal and abdominal radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The 2000 meeting had 48 months follow‐up, the 2001 meeting only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type science, country of origin, prospective status, and type clinical study with publication using stratified analysis and Chi2 test or logistic regression analysis for country of origin.

Shelmerdine 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and PubMed; dates of search not reported Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 363 abstracts presented at the 2012 British Society for Paediatric Radiology, European Society of Pediatric Radiology, and Society of Pediatric Radiology meetings Included all abstracts accepted as posters
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication by meeting
Outcomes	57 of 363 abstracts published 3/35 abstracts presented at the British Society for Paediatric Radiology meeting, 17/163 at the European Society of Pediatric Radiology, and 37/165 at the Society of Pediatric Radiology published Proportion of abstracts published by time Median time to publication = 12 months (range = 6 to 17 months) for the British Society for Paediatric Radiology meeting Median time to publication = 10 months (range = 2 to 29 months) for the European Society of Pediatric Radiology meeting Median time to publication = 15 months (range = 3 to 31 months) for the Society of Pediatric Radiology meeting No factors other than meeting year related to proportion of abstracts published reported
Notes	Radiology ‐ pediatric radiology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted as posters.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Siegfried 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and specialized register of African trials to June 2006 Search completed by investigator Search criteria not reported Matching criteria not reported
Data	Included 127 conference abstracts included in AIDSearch or NLM Gateway Included all abstracts of randomized controlled trials conducted in Africa or North America from 1985 to 2003 in Africa and in 2000 in North America
Comparisons	Proportion of abstracts published Mean time to publication by country where study conducted North American versus African
Outcomes	86 of 127 abstracts published Proportion of abstracts published by time Mean time to publication = 22.8 months (range = 12 to 84 months) for studies conducted in Africa Mean time to publication = 26.4 months (range = 12 to 72 months) for studies conducted in North America Factors related to proportion of abstracts published included 32/62 abstracts originating from North America versus 54/65 abstracts from Africa published
Notes	Infectious disease/immunology ‐ HIV, AIDS Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing RCTs within specific time and region.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search unclear.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of continent of origin with publication using stratified analysis.

Simons 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase, Google/Google Scholar, and Answer to 2005 Search completed by investigator Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 113 abstracts, presented at the 2000 American Academy of Allergy, Asthma and Immunology meeting Included a random sample of 1150 abstracts
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Oral versus poster presentation 'High‐profile' oral versus 'high‐profile' poster versus general poster presentations
Outcomes	73 of 113 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 36 months showed proportion published = 56.6% (64/113 abstracts) Factors related to proportion of abstracts published included 16/18 abstracts presented orally versus 57/95 abstracts presented as poster published 16/18 abstracts of 'high‐profile' oral presentations versus 9/10 abstracts of 'high‐profile' poster presentations versus 48/85 abstracts of general poster presentations published
Notes	Infectious disease/immunology ‐ allergy Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	Yes	Searched 4 databases and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type of presentation and 'high‐profile' status with publication using stratified analysis and Fisher's Exact test.

Singh 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and IndMed from 1 January, 2007 to 21 December, 2011 Person completing the search not reported Searched by all authors, keywords and title Matched abstract to full‐length publication by All authors Sample size
Data	Included 102 abstracts presented at the 2006 National Neonatology Forum of India meeting Included all abstracts
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Oral versus poster presentation Clinical research versus basic science research versus other types of research Observational design versus case report design versus community‐based design versus intervention design Academic center versus hospital versus individual institution versus other institution Subspecialty
Outcomes	14 of 102 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 13.7% (14/102 abstracts) Factors related to proportion of abstracts published included 6/36 abstracts presented orally versus 3/58 abstracts presented as poster published 11/90 abstracts describing clinical research versus 3/4 abstracts describing basic science research versus 0/8 abstracts describing other types of research 8/56 abstracts with observational design versus 0/15 abstracts with case report design versus 1/3 abstracts with community‐based design versus 2/16 abstracts with intervention design published 14/59 abstracts originating in an academic center versus 0/35 abstracts originating from a hospital versus 0/3 abstracts originating from individual institutions versus 0/5 abstracts originating from other institutions published
Notes	Pediatrics ‐ neonatology Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of type of presentation, type of science, study design, and academic affiliation with publication using stratified analysis and Chi2 or Fisher's Exact tests.

Sinno 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase to September 2009 Search completed by investigator Searched by first, and senior author, and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 138 abstracts presented at the 2003 Annual Meeting of the American Society of Plastic Surgeons Included all abstracts except those from guest speakers, award papers, and case reports
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Clinical research versus other research 'Controlled' design versus 'uncontrolled' design 'Prospective' versus not 'prospective' design Level I/II versus Level III/IV evidence Sponsor reported versus no report US versus non‐US origin Blinding versus no blinding
Outcomes	78 of 138 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 38/55 'positive' (defined as significant results) versus 40/83 not 'positive' abstract results published 57/103 'positive' (defined as experimental better than control) versus 21/35 not 'positive' abstract results published 53/99 abstracts describing clinical research versus 25/39 abstracts describing other research published 34/56 abstracts with 'controlled' study design versus 44/82 abstracts with 'uncontrolled' study design published 20/26 abstracts with 'prospective' study design versus 58/112 abstracts without 'prospective' study design published 53/104 abstracts originating in the US versus 25/34 abstracts not originating in the US published 17/17 abstracts with sponsor reported versus 61/121 abstracts with no report of sponsor published 3/4 abstracts with blinding versus 75/134 abstracts without blinding published Odds ratio between evidence Level I and II versus Level III and IV = 0.10 (95% CI = 0.0 to 5.93)
Notes	Surgery ‐ plastic surgery Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 2 databases and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type science, study design, prospective status, US origin, funding, and presence of blinding with publication using an unspecified multivariable analyses.

Sivan 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed, Google/Google Scholar for 3 years after the meeting Search completed by investigator Searched by all authors, keywords, and authors' institution Matching criteria not reported
Data	Included 251 abstracts presented at the 2000 to 2006 British Society of Rehabilitation Medicine meetings Included all abstracts from English speaking countries
Comparisons	Proportion of abstracts published 'Original' research versus reviews versus audits versus case reports
Outcomes	86 of 251 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 80/152 abstracts describing 'original' research versus 5/10 abstracts describing reviews versus 0/51 abstracts describing an audit versus 1/22 abstracts describing case reports published
Notes	Physical and rehabilitation medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	No	The meetings only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of study design with publication using stratified analysis.

Smart 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed for '5 years' Person completing the search not reported Searched by title, and presenting or senior author Matched abstract to full‐length publication by All authors Methodology Materials section Contacted abstract authors directly
Data	Included 122 abstracts presented at the 2000 to 2010 American Association of Oral and Maxillofacial Surgeons and American International Association of Dental Research meetings Included all abstracts presented by members of the Oral and Maxillofacial Surgery at Massachusetts General Hospital and Harvard School of Dental Medicine
Comparisons	Proportion of abstracts published Mean and median time to publication Survival analysis of proportion of abstracts published
Outcomes	90 of 122 abstracts published Proportion of abstracts published by time Mean time to publication = 21.9 months (SD = 17.3) Median time to publication = 19 months (range = 0 to 99 months) Survival analysis of proportion of abstracts published at 72 month = 73.8% No factors related to proportion of abstracts published reported
Notes	Otolaryngology head and neck surgery ‐ oral and maxillofacial surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 1 database and contacted abstract authors directly.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched abstracts by 3 different criteria.

Smith 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from June to August 2005 Person completing the search not reported Searched by first, second, and senior author, and contents Matched abstract to full‐length publication by All authors Contents Methodology Results
Data	Included 1683 abstracts presented at the 2002 and 2003 American Urological Association meetings Included all abstracts addressing 'clinical research in human subjects' by excluding abstracts involving basic science or animal studies
Comparisons	Proportion of abstracts published Median time to publication Survival analysis of publication rate 'Positive' versus 'not positive' Oral versus poster presentation Observational study design versus meta‐analyses/RCT versus case series versus other design Multi‐centered versus single center 'Prospective' versus 'retrospective' design US versus non‐US origin Therapy versus 'etiology/cause/harm' versus prognosis versus diagnostic tests 'New technology' versus not 'new technology' Subspecialty
Outcomes	740 of 1683 abstracts published Proportion of abstracts published by time Median time to publication = 27.8 months (range = 25.9 to 39.7 months) Survival analysis of proportion published = 42.7% at 30 months Factors related to proportion of abstracts published included 521/711 'positive' (defined as significant results) versus 86/202 not 'positive' abstract results published 238/511 abstracts presented orally versus 487/1086 abstracts presented as posters published 519/1187 abstracts with observational study design versus 78/163 abstracts describing meta‐analyses or RCTs versus 62/138 abstracts with case series design versus 81/195 abstracts with other design published 253/536 abstracts with multiple centers versus 487/1147 abstracts with a single center published 188/459 abstracts with 'prospective' study design versus 331/728 abstracts with 'retrospective' study design published 461/991 abstracts originating in the US versus 279/692 abstracts not originating in the US published 388/868 abstracts describing therapy versus 74/179 abstracts describing 'etiology/cause/harm' versus 115/252 abstracts describing prognosis versus 121/295 abstracts describing diagnostic tests published 129/332 abstracts describing a 'new technology' versus 611/1351 abstracts not describing a new technology published
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	All meetings had less than 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, study design, multi‐center status, prospective status, abstract quality, US origin, type of clinical study, presence of new 'technology' and subspecialty with publication using stratified analysis and time to publication using multivariable proportional hazards models.

Smith 2011a.

Methods	Identification of subsequent full‐length publications Searched electronic database CINAHL, PubMed, Evidence‐based Medicine Reviews 5 years after abstract presentation Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by 'concordance between information'
Data	Included 823 abstracts presented at the 2000 to 2004 American Physical Therapy Associations Combined Sections Meetings Included all abstracts presented in the Orthopaedic and Sports Physical Therapy sections
Comparisons	Proportion of abstracts published Mean and median time to publication 'Positive' versus not 'positive' Oral versus poster presentation RCT design versus observational versus validation versus non‐experimental design Reporting funding source versus not reporting funding source
Outcomes	209 of 823 abstracts published Proportion of abstracts published by time Mean time to publication = 22.7 months (SD = 15.3) Median time to publication = 19.3 months Factors related to proportion of abstracts published included 97/284 'positive' (defined as significant results) versus 34/184 not 'positive' abstract results published 139/401 abstracts presented orally versus 70/422 abstracts presented as poster published 31/103 abstracts with RCT design versus 99/365 abstracts with observational design versus 27/95 abstracts with validation design versus 48/252 abstracts with non‐experimental design published 40/125 abstracts reporting funding source versus 169/698 abstracts not reporting funding source published
Notes	Physical rehabilitation and medicine ‐ physical therapy Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The meetings had 5 years follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, study design, and funding using multivariable logistic regression analyses.

Smith 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Google/Google Scholar, and MedEd Portal; dates of search not reported Search completed by investigator Search criteria not reported Matching criteria not reported Contacted abstract authors directly
Data	Included 138 abstracts presented at the 1998 to 2008 Council on Medical Student Education in Pediatrics meetings Included all abstracts of presentations of 'scholarly' works (undefined) by Council on Medical Student Education in Pediatrics members
Comparisons	Proportion of abstracts published Oral versus poster presentations
Outcomes	47 of 138 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 17/32 abstracts presented orally versus 32/106 abstracts presented as poster published
Notes	Medical education/library science ‐ pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Unclear	Included all abstracts that described 'scholarly works', but unclear how this subgroup was defined.
Search for publications?	Yes	Searched 3 databases and contacted authors directly.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis and Chi2 tests.

Smollin 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Search completed by investigator Searched by first, second, and last author, keywords, and title Matching criteria not reported
Data	Included 237 abstracts presented at the 2001 North America Conference of Clinical Toxicology meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Clinical research versus basic science research versus other Case report design versus other design Subspecialty
Outcomes	57 of 237 abstracts published Proportion of abstracts published by time Mean time to publication = 15 months (range = 0 to 29 months) Factors related to proportion of abstracts published included 48/203 abstracts describing clinical research versus 6/19 abstracts describing basic science research versus 3/15 other abstracts published 25/109 abstracts with case report design versus 32/128 abstracts with other design published
Notes	Other non‐clinical specialties ‐ toxicology Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type science and study design with publication using stratified analysis and Chi2 tests.

Snedeker 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database Agricola, CAB abstracts, Web of Knowledge, Scholar's portal from January 1993 to December 2008 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Contents Methodology Pathogen being studied Species Meat type Contacted abstract authors directly
Data	Included 149 abstracts presented at the following meetings: 1997, 2000, or 2003 International Symposium on Veterinary Economics and Epidemiology 1995 to 2004 Conference of Research Workers in Animal Disease 2003 International Symposium on Shiga Toxin 1997 to 2004 Society for Veterinary Epidemiology and Preventive Medicine 1996, 1997, 1999 and 2003 International Symposium and Control of Foodborne Pathogens in Pork 1998 International Symposium on Food‐borne Salmonella in Poultry 2003 American Society of Microbiologists Conference on Salmonella 1995 to 2004 North East Conference on Avian Diseases meetings 1995 to 2004 International Association of Food Protection meetings 1998 to 2004 Institute of Food Technologists Annual meetings Included all abstracts that were in vivo clinical trials or field experiments involving a pre‐harvest or abattoir‐level intervention against pathogenic bacteria
Comparisons	Proportion of abstracts published Median time to publication Cumulative proportion of abstracts published Oral versus poster 'Positive' versus 'not positive' Animal species Animal housing Pathogen type
Outcomes	70 of 149 abstracts published Proportion of abstracts published by time Median time to publication = 13.5 months (range = 0 to 72 months) Cumulative proportion of abstracts published at 48 months showed proportion published = 45.6% (68/149 abstracts) Factors related to proportion of abstracts published included OR for publication of abstracts presented orally versus poster = 2.2 (95% CI, 1.0 to 4.8) OR for publication of 'not positive' versus 'positive' (defined as statistically significant) results = 0.4 (95% CI, 0.1 to 1.1)
Notes	Other non‐clinical specialties ‐ food safety research Funding by Canadian Institutes of Health Research, Institute of Population and Public Health/Public Health Agency of Canada Applied Public Health Chair
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	Yes	Searched 4 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 6 different criteria.
Adjustment for confounding?	Yes	Examined data on positive results, type of presentation, animal species, animal housing, and pathogen type using stratified analysis and Chi2 tests or logistic regression and time to publication using multivariable Cox proportional hazard analysis.

Song 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Search completed using custom PERL computer program Searched by first and last author Matched abstract to full‐length publication by Results Conclusions
Data	Included 4824 abstracts presented at the 2005 to 2007 United States and Canadian Academy of Pathology meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Cumulative proportion of abstracts published Oral versus poster presentation Using statistical methods versus not One or more versus no authors from US Subspecialty
Outcomes	1725 of 4824 abstracts published 603/1576 abstracts presented at the 2005 meeting, 555/1588 at the 2006 meeting, and 567/1660 at the 2007 meeting published Proportion of abstracts published by time Mean time to publication = 18 months Cumulative proportion of abstracts published at 24 months showed proportion published = 35.6% (1725/4824 abstracts) Factors related to proportion of abstracts published included 385/838 abstracts presented orally versus 1340/3986 abstracts presented as poster published 637/1733 abstracts using statistical methods versus 1088/3091 abstracts not using statistical methods published 1386/3916 abstracts with at least one US author versus 339/908 abstracts with no US authors published
Notes	Pathology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type of presentation, use of statistical methods, US origin and subspecialty with publication using stratified analysis and Fisher's Exact tests.

St John 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to May 2013 Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 327 abstracts presented at the 2009 Australian Gastroenterology Week meeting Included all abstracts
Comparisons	Proportion of abstracts published 'Advanced' trainee versus other Clinical research versus other University affiliation
Outcomes	112 of 327 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 78/250 abstracts describing clinical research versus 34/112 abstracts describing basic science research published 58/137 abstracts from authors with university affiliation versus 54/190 abstracts from authors without a university affiliation published 9/67 abstracts presented by 'advanced' trainee versus 103/260 abstracts presented by other person published
Notes	Gastroenterology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of type of science, university affiliation, and type of presenter with publication using multivariable logistic regression.

Stolk 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, International Pharmaceutical Abstracts to March 2001 Person completing the search not reported Searched by all authors Matched abstract to full‐length publication by All authors Sample size Methodology Research question
Data	Included 1216 abstracts presented at the 1995 to 1999 International Conference on Pharmacoepidemiology meetings Included all abstracts on pharmacoepidemiology and drug safety, except late‐breaking, workshop, and symposium
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting Oral versus poster presentation North American versus European versus rest of the world origin Subspecialty Conference year
Outcomes	319 of 1216 abstracts published; 262 abstracts with more than 24 months follow‐up published 52/174 abstracts presented at the 1995 meeting, 67/218 at the 1996 meeting, 73/240 at the 1997 meeting, 70/315 at the 1998 meeting, and 57/269 at the 1999 meeting published Proportion of abstracts published by time Cumulative proportion of abstracts published at 72 months showed proportion published = 29.9% (52/174 abstracts) for 1995 International Conference on Pharmacoepidemiology meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 30.7% (67/218 abstracts) for 1996 International Conference on Pharmacoepidemiology meeting Cumulative proportion of abstracts published at 48 months showed proportion published = 30.4% (73/240 abstracts) for 1997 International Conference on Pharmacoepidemiology meeting Cumulative proportion of abstracts published at 36 months showed proportion published = 22.2% (70/315 abstracts) for 1998 International Conference on Pharmacoepidemiology meeting Cumulative proportion of abstracts published at 24 months showed proportion published = 21.2% (57/269 abstracts) for 1999 International Conference on Pharmacoepidemiology meeting Factors related to proportion of abstracts published included 97/249 abstracts presented orally versus 222/967 abstracts presented as posters published 102/408 abstracts originating from North America versus 183/596 abstracts from Europe versus 34/212 abstracts from rest of the world published
Notes	Pharmacology ‐ pharmacoepidemiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic with reasonable exceptions, so low risk of bias.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings before 1998 had at least 48 months follow‐up. The meetings in 1998 and 1999 only had 36 and 24 months follow‐up.
Matching?	Yes	Matched by 4 different characteristics.
Adjustment for confounding?	Yes	Examined association of meeting year, type presentation, country of origin, and subspecialty using multivariable Cox regression analyses.

Stranges 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, SCOPUS, and International Pharmaceutical Abstracts for a '5 year interval' Search completed by the investigator Searched by all authors, and keywords Matching criteria not reported
Data	Included 655 abstracts presented at the 2003, 2005, and 2007 Great Lakes Pharmacy Residency Conference meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Clinical research versus basic science research versus survey versus service development RCT design versus observational design versus other interventional design Academic center versus veterans affairs institution versus community/retail institution versus hospital versus other institution affiliation Prospective versus retrospective design Results presented in abstract versus not Intervention on drugs versus humans versus other topics
Outcomes	76 of 655 abstracts published 22/171 abstracts presented at the 2003 meeting, 27/221 at the 2005 meeting, and 27/273 at the 2007 meeting published Proportion of abstracts published by time Mean time to publication = 24.5 months (SD = 14.3) Factors related to proportion of abstracts published included 58/546 abstracts describing clinical research versus 2/8 abstracts describing basic science research versus 13/39 abstracts describing surveys versus 3/72 abstracts describing service development published 1/20 abstracts describing RCT versus 63/573 abstracts describing observational design versus 13/72 abstracts describing other interventional design published 49/287 abstracts originating in an academic center versus 6/101 abstracts originating from a veterans affairs institution versus 3/24 abstracts originating from community/retail institutions versus 18/248 abstracts originating with hospitals versus 0/5 abstracts originating from other institutions published 16/101 abstracts describing prospective design versus 60/564 abstracts describing retrospective design published 18/111 abstracts with results presented in abstract versus 58/554 abstracts with no results presented in abstract published 26/304 abstracts describing drug intervention versus 28/206 abstracts describing 'human' intervention versus 22/155 abstracts describing other intervention published
Notes	Pharmacology Funded by the National Center for Advancing Translational Sciences
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	All meetings had 60 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting year, type of science, study design, academic affiliation, prospective status, presentation of results, and type clinical study with publication using multivariable logistic regression.

Stöcker 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed from January 1997 to December 2007 Person completing the search not reported Searched by first, and last author, and keywords Matching criteria not reported Contacted abstract authors directly Handsearched two journals
Data	Included 368 abstracts presented at the 1999 to 2003 Deutsche Gesellschaft für Allgemeinmedizin und Familienmedizin meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published
Outcomes	200 of 368 abstracts published Proportion of abstracts published by time Mean time to publication = 18.4 months (range = ‐24 to 84 months) Cumulative proportion of abstracts published at 84 months showed proportion published = 54.3% (200/368 abstracts) No factors related to proportion of abstracts published reported
Notes	General medicine/primary care Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database, contacted abstract authors, and handsearched journals.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Sullivan 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to April 2014 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 227 abstracts presented at the 2006 to 2010 Irish Society of Urology Annual meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting Clinical versus laboratory research
Outcomes	95 of 227 abstracts published 43% of abstracts presented at the 2006 meeting, 32% at the 2007 meeting, 36% at the 2008 meeting, 51% at the 2009 meeting, and 47% at the 2010 meeting Proportion of abstracts published by time Mean time to publication = 24 months, overall Mean time to publication = 26 months for the 2006 meeting Mean time to publication = 33 months for the 2007 meeting Mean time to publication = 22 months for the 2008 meeting Mean time to publication = 20 months for the 2009 meeting Mean time to publication = 18 months for the 2010 meeting Factors related to proportion of abstracts published included 39% of abstracts describing clinical research versus 49% of abstracts describing laboratory research published
Notes	Urology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and type science with publication using stratified analysis and t tests.

Sun 2011.

Methods	Identification of subsequent full‐length publication Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by all authors, and keywords Matching criteria not reported
Data	Included 2080 abstracts presented at the 2003 to 2006 American Academy of Ophthalmology and Association for Research in Vision and Ophthalmology meetings Included a random sample of abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication, overall and by meeting year Cumulative proportion of abstracts published by meeting Experimental versus other design RCT versus other design Sample size
Outcomes	881 of 2080 abstracts published 132/617 abstracts presented at the 2003 meeting, 155/615 at the 2004 meeting, 170/616 at the 2005 meeting, and 159/616 at the 2006 meeting published Proportion of abstracts published by time Median time to publication = 16 months (IQR = 8 to 27.1 months), overall Median time to publication = 20 months (IQR = 9.1 to 32 months) for 2003 American Academy of Ophthalmology and Association for Research in Vision and Ophthalmology meetings Median time to publication = 18 months (IQR = 9.1 to 29 months) for 2004 American Academy of Ophthalmology and Association for Research in Vision and Ophthalmology meetings Median time to publication = 14.5 months (IQR = 8 to 26.5 months) for 2005 American Academy of Ophthalmology and Association for Research in Vision and Ophthalmology meetings Median time to publication = 15 months (IQR = 7 to 22.6 months) for 2006 American Academy of Ophthalmology and Association for Research in Vision and Ophthalmology meetings Cumulative proportion of abstracts published at 72 months showed proportion published = 46.0% (221/480 abstracts) for the 2003 to 2006 American Academy of Ophthalmology meetings Cumulative proportion of abstracts published at 72 months showed proportion published = 39.0% (624/1600 abstracts) for the 2003 to 2006 Association for Research in Vision and Ophthalmology meetings Factors related to proportion of abstracts published included Experimental versus other study design: OR 1.28 (95% CI = 1.06‐1.55) RCT versus other study design: OR 1.59 (95% CI = 1.06‐2.4)
Notes	Ophthalmology/optometry ‐ ophthalmology Funding not reported Data extracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting year, type science, study design, sample size, use of a control group with publication using multivariable logistic regression.

Susarla 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and SCOPUS for a 'minimum of 18 months' Person completing the search not reported Searched by title, and presenting and primary author Matched abstract to full‐length publication by All authors Methodology Materials
Data	Included 78 abstracts presented at the 2009 to 2013 Johns Hopkins Hospital/University of Maryland Plastic Survey Research Day meetings Included all abstracts from a presenting author who was a plastic surgery resident in the Johns Hopkins Hospital/University of Maryland program
Comparisons	Proportion of abstracts published Mean time to publication Year of resident status Resident characteristics Faculty investigator with academic rank of (associate) professor Prior publications Subspecialty
Outcomes	50 of 78 abstracts published Proportion of abstracts published by time Mean time to publication = 15.6 months (SD = 13.6; range = ‐2 to 54 months) Factors related to proportion of abstracts published included Year of resident status: OR 0.56 (95% CI 0.36 to 0.85) Faculty investigator with academic rank of either associate professor or professor: OR 3.3 (95% CI 1.1 to 10.5)
Notes	Surgery ‐ plastic and reconstructive surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts from a specific subgroup of authors.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings except the 2009 and 2010 meeting had less than 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of year of resident status, residency track, number of prior publications, type of science, subspecialty, and present of faculty with academic rank of associate or full professor with publication using multivariable logistic regression.

Tam 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed from May 2000 to May 2006 Search completed by investigator Searched by first, second, and last author, and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 74 abstracts presented at the 2000 the American Society of Clinical Oncology meeting Included all abstracts of phase III trials of chemotherapy, chemoradiotherapy, immunotherapy, hormone therapy
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation versus abstract only North American versus European versus Asian versus African versus South American versus Australian versus 'multiple continents' origin Pharmaceutical versus co‐op/government versus unreported funding
Outcomes	55 of 74 abstracts published Proportion of abstracts published by time Mean time to publication = 31 months Median time to publication = 29 months (range = 3.4 to 68.4 months) Cumulative proportion of abstracts published at 78 months showed proportion published = 74.3% (55/74 abstracts) Factors related to proportion of abstracts published included 17/18 'positive' (defined as significant results) versus 38/56 not 'positive' abstract results published 21/27 'positive' (defined as experimental better than control) versus 34/47 not 'positive' abstract results published 32/39 abstracts presented orally versus 19/26 abstracts presented as posters versus 4/9 presented as abstract only published 13/19 abstracts originating from North America versus 33/44 abstracts from Europe versus 3/4 abstracts from Asia versus 1/1 abstracts from Africa versus 0/1 abstracts from South America versus 0/0 abstracts from Australia versus 5/5 abstracts from 'multiple continents' published 3/3 abstracts with pharmaceutical funding versus 0/1 abstracts with co‐op/government funding versus 52/74 abstracts with funding not reported published
Notes	Oncology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described phase III trials on specific topics, so low risk of bias.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Yes	The meeting had 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined data on more than 2 factors, and performed multivariate logistic regression analyses.

Tam 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed, and Embase from January 1989 to December 2009 Person completing the search not reported Searched by first, second, and last author, and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 709 abstracts presented at the 1989 to 2003 American Society of Clinical Oncology meetings Included all abstracts of phase III or randomized controlled trials with sample size >= 200 and outcome of overall survival, disease‐free survival, or progression‐free survival Excluded abstracts that were "repeat presentations" (n = 8) or were published >= 3 months before presentation (n = 2)
Comparisons	Proportion of abstracts published Median time to publication Survival analysis of publication rate
Outcomes	643 of 709 abstracts published Proportion of abstracts published by time Median time to publication = 31.2 months Survival analysis of proportion published at 120 month = 88.5% No factors related to proportion of abstracts published reported
Notes	Oncology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described RCTs and Phase III trials.
Search for publications?	Yes	Searched 2 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.

Tambuscio 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, FORS Forensic Bibliographic Database; dates of search not reported Search completed by an undergraduate student Searched by all authors, keywords, and titles Matching criteria not reported
Data	Included 623 abstracts presented at the 2006 American Academy of Forensic Sciences meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published Oral versus poster presentation North American versus European versus rest of the world origin US versus non‐US versus 'international collaboration' origin
Outcomes	102 of 623 abstracts published Proportion of abstracts published by time Mean time to publication = 10 months (SD = 9) Cumulative proportion of abstracts published at 18 months showed proportion published = 12.4% (77/623 abstracts) Factors related to proportion of abstracts published included 17.2% of abstracts presented orally versus 14.6% of abstracts presented as poster published 69/500 abstracts originating from North America versus 20/59 abstracts from Europe versus 3/37 abstracts from rest of the world published 66/472 abstracts originating in the US versus 26/124 abstracts not originating in the US versus 10/27 abstracts of 'international collaboration' origin published
Notes	Other non‐clinical specialties ‐ forensic sciences Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type of presentation, number of authors, country of origin, and US origin with publication using stratified analysis and unspecified statistical tests.

Tan 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to January 2008 Search completed by investigator Searched by first, and last author, and keywords Matching criteria not reported
Data	Included 54 abstracts presented at the 2004 Asian Conference on Emergency Medicine meeting Included all abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published
Outcomes	10 of 54 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Emergency medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Theman 2014.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed in August 2013 Search completed by investigator Searched by all authors, keywords, and title Matched abstract to full‐length publication by All authors Methodology Results Country of origin Hypothesis Study design Study groups Conclusions
Data	Included 719 abstracts presented at the 2000 to 2010 American Society for Surgery of the Hand meetings Included all abstracts from oral presentation on hand, wrist, and brachial plexus topics excluding all that focused primarily on elbow/shoulder joint and those from resident and fellows conferences
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication, overall and by meeting Cumulative proportion of abstracts published
Outcomes	393 of 719 abstracts published 56/109 abstracts presented at the 2000 meeting, 30/69 at the 2001 meeting, 36/64 at the 2002 meeting, 13/36 at the 2003 meeting, 40/80 at the 2004 meeting, 21/53 at the 2005 meeting, 17/35 at the 2006 meeting, 40/88 at the 2007 meeting, 38/62 at the 2008 meeting, 32/58 at the 2009 meeting, and 33/65 at the 2010 meeting published Proportion of abstracts published by time Mean time to publication = 18 months (range = ‐12 to 122 months), overall Mean time to publication = 22 months (maximum = 120 months) for the 2000 American Society for Surgery of the Hand meeting Mean time to publication = 25 months (maximum = 96 months) for the 2001 American Society for Surgery of the Hand meeting Mean time to publication = 22 months (maximum = 120 months) for the 2002 American Society for Surgery of the Hand meeting Mean time to publication = 15 months (maximum = 48 months) for the 2003 American Society for Surgery of the Hand meeting Mean time to publication = 20 months (maximum = 84 months) for the 2004 American Society for Surgery of the Hand meeting Mean time to publication = 19 months (maximum = 60 months) for the 2005 American Society for Surgery of the Hand meeting Mean time to publication = 21 months (maximum = 84 months) for the 2006 American Society for Surgery of the Hand meeting Mean time to publication = 15 months (maximum = 60 months) for the 2007 American Society for Surgery of the Hand meeting Mean time to publication = 12 months (maximum = 48 months) for the 2008 American Society for Surgery of the Hand meeting Mean time to publication = 15 months (maximum = 48 months) for the 2009 American Society for Surgery of the Hand meeting Mean time to publication = 13 months (maximum = 36 months) for the 2010 American Society for Surgery of the Hand meeting Median time to publication = 14 months (IQR = 14 to 64 months), overall Median time to publication = 15 months (IQR = 10 to 78 months) for the 2000 American Society for Surgery of the Hand meeting Median time to publication = 14 months (IQR = 12 to 101 months) for the 2001 American Society for Surgery of the Hand meeting Median time to publication = 14 months (IQR = 21 to 69 months) for the 2002 American Society for Surgery of the Hand meeting Median time to publication = 15 months (IQR = 7 to 68 months) for the 2003 American Society for Surgery of the Hand meeting Median time to publication = 18 months (IQR = 11 to 69 months) for the 2004 American Society for Surgery of the Hand meeting) Median time to publication = 11 months (IQR = 13 to 84 months) for the 2005 American Society for Surgery of the Hand meeting Median time to publication = 20 months (IQR = 30 to 66 months) for the 2006 American Society for Surgery of the Hand meeting Median time to publication = 14 months (IQR = 18 to 51 months) for the 2007 American Society for Surgery of the Hand meeting Median time to publication = 11 months (IQR = 15 to 44 months) for the 2008 American Society for Surgery of the Hand meeting Median time to publication = 13 months (IQR = 17 to 66 months) for the 2009 American Society for Surgery of the Hand meeting Median time to publication = 12 months (IQR = 17 to 55 months) for the 2010 American Society for Surgery of the Hand meeting Cumulative proportion of abstracts published at 120 months showed proportion published = 49% (352/719 abstracts) No factors other than meeting year related to proportion of abstracts published reported
Notes	Surgery ‐ hand Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts from oral presentation on a specific topic.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2010 had at least 48 months follow‐up. The meeting in 2010 only had 36 months follow‐up.
Matching?	Yes	Matched by 8 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Tieman 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase, CINAHL, PsycINFO to 2004 Person completing the search not reported Searched by all authors, keywords, and index terms Matching criteria not reported
Data	Included 1338 abstracts identified for inclusion in "CareSearch", a database of palliative care grey literature Included all abstracts presented from 1980 up to 1999
Comparisons	Proportion of abstracts published Oral versus plenary versus poster versus workshop presentation Subspecialty
Outcomes	213 of 1338 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 17.4% of abstracts presented orally versus 14.8% of abstracts presented in plenary sessions versus 14.3% of abstracts presented as posters versus 6.5% of abstracts presented as workshops published
Notes	General medicine/primary care ‐ palliative care Funding by Palliative Care Section of the Commonwealth Department of Health and Ageing and the National Reference Group
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts within the CareSearch database.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation and subspecialty with publication using stratified analysis.

Timmer 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to July 1999 Person completing the search not reported Searched by first and last authors Matching criteria not reported
Data	Included 594 abstracts presented at the 1994 to 1995 European Pancreatic Club and American Pancreatic Association meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published overall by meeting Clinical research versus basic science research CCT design versus non‐CCT design Multi‐centered versus single center (defined as 1 or 2) North American versus European versus rest of the world origin
Outcomes	341 of 594 abstracts published 138/254 abstracts presented at the American Pancreatic Association meetings, and 203/340 at the European Pancreatic Club meetings published Proportion of abstracts published by time Median time to publication = 36 months Cumulative proportion of abstracts published at 48 months showed proportion published = 57.4% (341/594 abstracts), overall Cumulative proportion of abstracts published at 48 months showed proportion published = 60.1% (203/340 abstracts) for European Pancreatic Club meetings Cumulative proportion of abstracts published at 48 months showed proportion published = 54.9% (139/254 abstracts) for American Pancreatic Association meetings Factors related to proportion of abstracts published included 126/232 abstracts describing clinical research versus 215/362 abstracts describing basic science research published 24/40 abstracts with CCT design versus 317/554 abstracts with non‐CCT design published 43/82 abstracts with multiple centers (defined as 3 or more) versus 298/512 abstracts with a single center (defined as 1 or 2) published 75/141 abstracts originating from North America versus 245/418 abstracts from Europe versus 21/35 abstracts from rest of the world published
Notes	Gastroenterology ‐ pancreatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting, type of science, study design, multi‐center status, and country of origin with publication using multivariable logistic regression.

Timmer 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase, Cochrane Central Register of Controlled Trials, Bios; dates of search not reported Person completing the search not reported Searched by first and last author Matching criteria not reported Contacted abstract authors directly
Data	Included 836 abstracts submitted to the 1992 to 1995 Digestive Diseases Week, combined meetings for American Gastroenterological Association, American Association for the Study of Liver Diseases, American Society for Gastrointestinal Endoscopy, and Surgical Society for the Alimentary Tract Included a random sample of abstracts
Comparisons	Proportion of abstracts published Median time to publication Survival analysis of publication rate by type of research design 'Positive' versus not 'positive' Sample size equal to or above the median versus sample size below the median Accepted versus rejected for conference presentation Clinical research versus basic science research CCT design versus non‐CCT design 'High' versus 'low' quality as defined by author Multi‐centered versus single center North American versus European versus rest of the world origin
Outcomes	392 of 836 abstracts published Proportion of abstracts published by time Median time to publication = 18 months Survival analysis of proportion published at 72 months = 50.0% for basic science research Survival analysis of proportion published at 72 months = 54.9% for CCTs Survival analysis of proportion published at 72 months = 41.1% for other clinical research Factors related to proportion of abstracts published included 177/354 'positive' (defined as significant results) versus 213/482 not 'positive' abstract results published 181/377 abstracts with sample size equal to or above the median versus 142/338 abstracts with sample size below the median published 292/541 abstracts accepted for presentation versus 98/288 rejected abstracts published 310/662 abstracts describing clinical research versus 82/174 abstracts describing basic science research published 170/326 abstracts with CCT design versus 222/510 abstracts with non‐CCT design published 150/300 abstracts rated by author as 'high' quality versus 240/533 abstracts rated as 'low' quality published 36/46 abstracts with multiple centers versus 134/279 abstracts with a single center published 147/311 abstracts originating from North America versus 181/384 abstracts from Europe versus 62/141 abstracts from rest of the world published
Notes	Gastroenterology Funding by Calgary Regional Health Authority R&D and Searle Canada
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	Yes	Searched 4 databases and contacted abstract authors.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, sample size, acceptance for presentation, type of science, study design, abstract quality, multi‐center status, and country of origin with publication using multivariable Cox or logistic regression

Todd 1997.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE 5 years after presentation Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 118 abstracts submitted to a "regional otolaryngology" meeting Included all abstracts
Comparisons	Proportion of abstracts published Accepted versus rejected for conference presentation
Outcomes	43 of 118 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 35/53 abstracts accepted for presentation versus 8/65 rejected abstracts published
Notes	Otolaryngology head and neck surgery Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis and unspecified statistical tests.

Toma 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed, Embase, Cochrane Central Register of Controlled Trials, Google/Google Scholar, SIGLE, CardioSource, Incirculation to December 2005 Search completed by investigator Searched by all authors, keywords, title, and study acronym Matched abstract to full‐length publication by Authors Study name/acronym Title keyword
Data	Included 186 abstracts presented at the 1999 to 2002 American College of Cardiology scientific meetings Included 86 abstracts of late breaking RCTs reporting clinical outcomes, and 100 other randomly chosen RCTs; RCTs with non‐human participants were excluded
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication by whether 'late‐breaking' or not Survival analysis of publication rate 'Positive' versus not 'positive' Multi‐centered versus single center 'Late‐breaking' versus not 'late‐breaking'
Outcomes	148 of 186 abstracts published 34/43 abstracts presented at the 1999 meeting, 36/48 at the 2000 meeting, 41/49 at the 2001 meeting, and 38/46 at the 2002 meeting published Proportion of abstracts published by time Median time to publication = 11.5 months for 'late‐breaking' abstracts Median time to publication = 22 months for abstracts that were not 'late‐breaking' Survival analysis of proportion published at 60 months = 83.3% Factors related to proportion of abstracts published included 70/84 'positive' (defined as significant results) versus 78/102 not 'positive' abstract results published 74/93 'positive' (defined as experimental better than control) versus 74/93 not 'positive' abstract results published 55/63 abstracts with multiple centers versus 93/123 abstracts with a single center published 79/86 'late‐breaking' abstracts versus 69/100 abstracts that were not 'late‐breaking' published
Notes	Cardiology Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts describing RCTs and al late breaking abstracts describing RCTs.
Search for publications?	Yes	Searched 7 databases.
Follow‐up time?	Yes	All meetings before 2002 had at least 48 months follow‐up. The meeting in 2002 only had 36 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, multi‐center status, and 'late‐breaking' status with publication using Chi2 or tests and time to publication using a multivariable proportional hazards analysis.

Toro‐Polo 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, SciELO, LILACS, RedALyC, Portal Revist@s Peruanas, and Sistema de Bibliotecas de la Universidad Nacional Mayor de San Marcos from June to July 2011 Person completing the search not reported Searched by all authors, location of study, population, and main result Matched abstract to full‐length publication by One author Results Time Population Location
Data	Included 532 abstracts submitted to the 2002 to 2005 and 2006 to 2009 Sociedad Científica Médico Estudiantil Peruana meetings Included all abstracts except abstracts from students with foreign affiliation (162) and those that were presented more than once
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication Collaborative authorship versus other Analytical versus descriptive work
Outcomes	52 of 532 abstracts published 21/254 abstracts presented at the 2002 to 2005 meetings and 31/278 at the 2006 to 2009 meetings published Proportion of abstracts published by time Mean time to publication = 22.6 months (SD = 24.0) Median time to publication = 13 months (range = 0 to 75 months) Factors related to proportion of abstracts published included 3/6 abstracts with collaborative authorship versus 49/526 abstracts without collaborative authorship published 34/272 abstracts presenting analytical work versus 18/260 abstracts presenting descriptive work published
Notes	General medicine/primary care Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 6 databases.
Follow‐up time?	Yes	All meetings had at least 72 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, collaborative authorship and type science with publication using unspecified multivariable analyses.

Turpen 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to August 2007 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 126 abstracts presented at the 2002 and 2003 American Urological Association meetings Included all abstracts of RCTs
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	79 of 126 abstracts published Proportion of abstracts published by time Mean time to publication = 51 months (SD = 5) No factors related to proportion of abstracts published reported
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing RCTs.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Tyagi 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar; dates of search not reported Person completing the search not reported Searched by first author, last author, keywords, and title Matched abstract to full‐length publication by Sample size Hypothesis Study design Protocol
Data	Included 726 abstracts presented at the 2009 American Society of Anesthesiologists and Indian Society of Anaesthesiologists’ Conference meetings Included all abstracts from the Indian Society of Anaesthesiologists’ Conference and a random sample from the American Society of Anesthesiologists meeting
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication
Outcomes	99 of 726 abstracts published 80/363 abstracts presented at the American Society of Anesthesiologists, and 19/363 at the Indian Society of Anaesthesiologists’ Conference published Proportion of abstracts published by time Median time to publication = 14 months (range = 1 to 21 months) for the American Society of Anesthesiologists meeting Median time to publication = 12 months (range = 3 to 24 months) for the Indian Society of Anaesthesiologists’ Conference meeting No factors other than meeting related to proportion of abstracts published reported
Notes	Anesthesiology Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts from one conference and a random sample of another conference.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting with publication using stratified analysis and unspecified statistical tests.

ul Haq 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to November 2009 Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by 'Same conclusion/message'
Data	Included 179 abstracts presented at the 2001 British Orthopaedic Association meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published
Outcomes	65 of 179 abstracts published Proportion of abstracts published by time Mean time to publication = 18.6 months (SD = 9.4) Cumulative proportion of abstracts published at 24 months showed proportion published = 27.4% (49/179 abstracts) No factors related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery ‐ trauma and orthopedics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 8 years follow‐up.
Matching?	No	Matched by only 1 criterion.

Uysal 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to October 2007 Search completed by 'two observers' Searched by first, second, and last author Matched abstract to full‐length publication by Sample size
Data	Included 900 abstracts presented at the 2002 International Federation of Clinical Chemistry and Laboratory Medicine meeting Included all abstracts presented as posters
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Clinical research versus basic science research versus animal research Multi‐centered versus single center 'Prospective' versus 'retrospective' versus observational versus case series versus 'experimental' versus 'technical' design
Outcomes	125 of 900 abstracts published Proportion of abstracts published by time Mean time to publication = 11.9 months Median time to publication = 8 months Cumulative proportion of abstracts published at 48 months showed proportion published = 13.8% (124/900 abstracts) Factors related to proportion of abstracts published included 59/385 abstracts describing clinical research versus 52/357 abstracts describing basic science research versus 10/75 abstracts describing animal research published 81/407 abstracts with multiple centers versus 59/493 abstracts with a single center published 10/92 abstracts with 'prospective' cohort study design versus 2/9 abstracts with 'retrospective' cohort study design versus 4/83 abstracts describing observational research versus 47/284 abstracts describing case series design versus 10/75 abstracts with 'experimental' design versus 52/357 abstracts with 'technical' design published
Notes	Other non‐clinical specialties ‐ clinical chemistry Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented as posters.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	Yes	Examined association of type of science, multi‐center status, prospective status, and study design with publication using multivariable logistic regression.

Uzun 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE and PubMed from June 2005 to July 2010 Search completed by the investigators Searched by all authors, keywords, and title Matching criteria not reported. Searched journals South Pacific Underwater Medicine Society and European Underwater and Baromedical Society until 2007
Data	Included 179 abstracts presented at the 2005 Undersea and Hyperbaric Medical Society meeting Included all abstracts except 2 retracted abstracts and 6 that were published prior to meeting
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published Clinical research versus basic science research versus secondary research Observational design versus case report design versus experimental clinical design Multi‐centered versus single center English language versus non‐English language North American versus European versus Asian versus Australian versus South American origin USA versus non‐USA origin Statistical methods mentioned versus not mentioned 1‐5 authors versus > 5 authors Subspecialty
Outcomes	62 of 179 abstracts published Proportion of abstracts published by time Mean time to publication = 18.5 months (SD = 13.6; range = 2 to 60 months) Median time to publication = 15 months Cumulative proportion of abstracts published at 60 months showed proportion published = 34.6% (62/179 months) Factors related to proportion of abstracts published included 32/112 abstracts describing clinical research versus 28/56 abstracts describing basic science research versus 2/11 abstracts describing secondary research published 20/58 abstracts with observational design versus 4/31 abstracts with case report design versus 8/23 abstracts with experimental clinical design published 20/57 abstracts with multiple centers versus 42/122 abstracts with a single center published 47/134 abstracts in English language versus 15/45 abstracts in non‐English language published 37/115 abstracts originating in North America versus 18/32 abstracts originating from Europe versus 3/19 abstracts originating from Asia versus 3/5 abstracts originating from Australia versus 1/8 abstracts originating from South America published 34/107 abstracts originating from the USA versus 28/72 abstracts with non‐US origin published 16/26 abstracts with statistical methods mentioned versus 35/103 abstracts without statistical methods mentioned published 40/138 abstracts with 1 to 5 authors versus 22/41 abstracts with > 5 authors published
Notes	Other clinical specialties ‐ undersea and hyperbaric medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 1 database and 2 journals.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of type of science, study design, multi‐center status, origin in English language country, country of origin, US origin, presence of statistical methods, number of authors and subspecialty with publication using multivariable logistic regression analyses.

Valderrama‐Zurian 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to July 2006 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Aims Sample size Results/conclusions
Data	Included 689 abstracts presented at the 1999 College on Problems of Drug Dependence meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean and median time to publication overall and by type of presentation Cumulative proportion of abstracts published Oral versus poster presentation Clinical research versus basic science research
Outcomes	254 of 689 abstracts published Proportion of abstracts published by time Mean time to publication = 22.4 months (SD = 17.3) Mean time to publication = 22.4 months (SD = 18.5) for oral presentations Mean time to publication = 22.4 months (SD = 16.4) for poster presentations Median time to publication = 19.9 months Median time to publication = 19.9 months for oral presentations Median time to publication = 19.9 months for poster presentations Cumulative proportion of abstracts published at 90 months showed proportion published = 37.0% (255/689 abstracts) Factors related to proportion of abstracts published included 102/194 abstracts presented orally versus 152/495 abstracts presented as poster published 141/480 abstracts describing clinical research versus 113/209 abstracts describing basic science research published
Notes	Psychiatry ‐ drug addiction Funding by Plan Municipal de Drogodependencias Ayuntamiento de Valencia and the Minstério do Sanidad y Consumo, Red de Trastornos Adictivos RD06/001/0020, and V Sefles Program of the University of Valencia, Spain
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 7 years follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	No	Examined association of type presentation, type science, with publication using stratified analysis and Chi2 and t tests, and binary logistic regression.

Van Balken 2005.

Methods	Method used to identify subsequent full‐length publications not reported
Data	Included 61 abstracts presented at the 2002 and 2003 Dutch Urological Association meetings Included all abstracts
Comparisons	Proportion of abstracts published
Outcomes	29 of 61 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Urology Funding not reported Data extracted from English language abstract; full publication not translated
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Unclear	Method used to identify subsequent full‐length publications not reported.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

van der Steen 2004.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2002 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported Handsearched European Journal of Plastic Surgery for 1997 Electronic search of European Journal of Plastic Surgery website from 1998 through 2002
Data	Included 306 abstracts presented at the 1995 to 1999 European Association of Plastic Surgeons meetings Included all abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication Cumulative proportion of abstracts published overall and by meeting Clinical research versus basic science research versus 'aesthetic surgery' studies Subspecialty
Outcomes	195 of 306 abstracts published 38/57 abstracts presented at the 1995 meeting, 43/60 at the 1996 meeting, 35/60 at the 1997 meeting, 39/60 at the 1998 meeting, and 40/60 at the 1999 meeting published Proportion of abstracts published by time Mean time to publication = 14 months (SD = 23.9) Median time to publication = 13 months Cumulative proportion of abstracts published at 48 months showed proportion published = 59.5% (182/306 abstracts) overall Cumulative proportion of abstracts published at 72 months showed proportion published = 61.4% (35/57 abstracts) for 1995 European Association of Plastic Surgeons meeting Cumulative proportion of abstracts published at 72 months showed proportion published = 66.7% (40/60 abstracts) for 1996 European Association of Plastic Surgeons meeting Cumulative proportion of abstracts published at 72 months showed proportion published = 55.0% (33/60 abstracts) for 1997 European Association of Plastic Surgeons meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 65.0% (39/60 abstracts) for 1998 European Association of Plastic Surgeons meeting Cumulative proportion of abstracts published at 48 months showed proportion published = 58.0% (40/69 abstracts) for 1999 European Association of Plastic Surgeons meeting Factors related to proportion of abstracts published included 120/192 abstracts describing clinical research versus 33/56 abstracts of 'aesthetic surgery' studies versus 42/58 abstracts describing basic science research published
Notes	Surgery ‐plastic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation.
Search for publications?	Yes	Searched 1 database, handsearched a journal and electronically searched another journal.
Follow‐up time?	Yes	All meetings before 1999 had at least 48 months follow‐up. The meeting in 1999 only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type science, and subspecialty with publication using stratified analysis and Chi2 tests.

Van Royen 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Search criteria not reported Matching criteria not reported Contacted abstract authors directly
Data	Included 251 abstracts presented at the 1999 to 2006 European General Practice Research Network meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	113 of 251 abstracts published 11/27 abstracts presented at the 1999 meeting, 9/19 at the May 2000 meeting, 8/16 at the October 2000 meeting, 6/17 at the June 2001 meeting, 9/19 at the October 2001 meeting, 12/29 at the May 2002 meeting, 19/33 at the May 2005 meeting, 13/26 at the October 2005 meeting, 16/38 at the May 2006 meeting, and 10/27 at the October 2006 meeting published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	General medicine/primary care Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database and contacted abstract authors.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.

Varghese 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed to August 2009 Search completed by investigator Searched by first author and keywords Matched abstract to full‐length publication by All authors Title Contents
Data	Included 1191 abstracts submitted to the 2003 to 2005 Pediatric Orthopaedic Society of North America meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean and median time to publication by meeting, oral presentation and rejection Oral versus poster presentation Accepted versus rejected for conference presentation
Outcomes	598 of 1191 abstracts published 221/431 abstracts presented at the 2003 meeting, 169/343 at the 2004 meeting, and 208/417 at the 2005 meeting published Proportion of abstracts published by time Mean time to publication = 24 months (95% CI = 20.3 to 27.8 months) for 2003 Pediatric Orthopaedic Society of North America meeting 21.9 months (95% CI = 17.6 to 26.2 months) for oral presentation 21.4 months (95% CI = 18.5 to 24.2 months) for rejected abstracts Mean time to publication = 24.8 months (95% CI = 21.4 to 28.3 months) for 2004 Pediatric Orthopaedic Society of North America meeting 28.3 months (95% CI = 23.2 to 33.5 months) for oral presentation 22.2 months (95% CI = 19.0 to 25.4 months) for rejected abstracts Mean time to publication = 20.2 months (95% CI = 17.7 to 22.8 months) for 2005 Pediatric Orthopaedic Society of North America meeting 20.7 months (95% CI = 17.2 to 24 months) for oral presentation 17.6 months (95% CI = 15 to 20.1 months) for rejected abstracts Median time to publication = 18.0 months or 2003 Pediatric Orthopaedic Society of North America meeting 18 months for oral presentations 18 months for rejected abstracts Median time to publication = 22 months or 2004 Pediatric Orthopaedic Society of North America meeting 26 months for oral presentations 19.5 months for rejected abstracts Median time to publication = 17 months or 2005 Pediatric Orthopaedic Society of North America meeting 18.5 months for oral presentations 15 months for rejected abstracts Factors related to proportion of abstracts published included 159/265 abstracts presented orally versus 100/175 abstracts presented as poster published 259/440 abstracts accepted for presentation versus 339/751 rejected abstracts published
Notes	Orthopedics/orthopedic surgery ‐ pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of meeting year, type presentation, and acceptance for presentation with publication using stratified analysis.

Vecchi 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase from January 1992 to March 2006 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author Primary outcome
Data	Included 581 abstracts presented at the 1993 to 2002 Annual Meeting of College on Problems of Drug Dependence meetings Included all abstracts of randomized and controlled clinical trials, except those reporting analyses on healthy people, pharmacokinetic and toxicity studies, and preliminary findings reporting characteristics of participants without mention of allocation groups
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication overall and by type of results 'Positive' versus negative or null versus no statistical results versus no results RCT design versus CCT design
Outcomes	359 of 581 abstracts published 37/59 abstracts presented at the 1993 meeting, 34/50 at the 1994 meeting, 13/51 at the 1995 meeting, 17/27 at the 1996 meeting, 26/41 at the 1997 meeting, 44/58 at the 1998 meeting, 40/60 at the 1999 meeting, 48/80 at the 2000 meeting, 56/83 at the 2001 meeting, and 44/72 at the 2002 meeting published Proportion of abstracts published Mean time to publication = 45.6 months Mean time to publication = 34.8 months for 'positive' results Mean time to publication = 62.4 months for negative or null results Mean time to publication = 50.4 months for abstracts without reported statistical results Mean time to publication = 62.4 months for abstracts with no results Factors related to proportion of abstracts published included 120/161 abstracts with 'positive' (defined as significant) results versus 24/51 abstracts with negative or null results versus 198/325 abstracts not reporting any statistical results versus 17/44 abstracts with no results reported published 284/455 abstracts with RCT design versus 75/126 abstracts with CCT design published
Notes	Psychiatry ‐ drug and alcohol addiction Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing RCTs and controlled clinical trials.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, positive results, and study design with publication using log rank tests.

Vo 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and MEDLINE; dates of search not reported Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by Title Methodology Results Conclusions
Data	Included 287 abstracts presented at the 2000 to 2003 International Society for Pharmacoeconomics and Outcomes Research ‐ International and European meetings Included all abstracts on mental health
Comparisons	Proportion of abstracts published
Outcomes	54 of 287 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Psychiatry ‐ mental health Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 4 different criteria.

Vogt 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase, SPOLIT, SPOFOR, SPOMEDIA; dates of search not reported Person completing the search not reported Searched by all authors Matched abstract to full‐length publication by Sample size Methodology Results Conclusions Contacted abstract authors directly
Data	Included 36 abstracts presented at the 1997 Deutscher Sportärztekongress meeting Included all abstracts on sports injuries
Comparisons	Proportion of abstracts published Mean time to publication
Outcomes	20 of 36 abstracts published Proportion of abstracts published by time Mean time to publication = 12 months No factors related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ sports medicine Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described a specific topic, so low risk of bias.
Search for publications?	Yes	Searched 5 databases and contacted abstract authors.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 4 different criteria.

von Hardenberg 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by first, last and an additional author, and keywords Matched abstract to full‐length publication by One author Contents
Data	Included 732 abstracts presented at the 2002 and 2009 Der Kongress der Deutschen Gesellschaft für Urologie meetings Included all abstracts except presentations from video, case report and history of medicine sections
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication by meeting
Outcomes	386 of 732 abstracts published 172/352 abstracts presented at the 2002 meeting, and 214/380 at the 2009 meeting published Proportion of abstracts published by time Median time to publication = 15.1 months No factors related to proportion of abstracts published reported
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 2 different criteria.

Vuckovic‐Dekic 2001.

Methods	Identification of subsequent full‐length publications Contacted abstract authors directly
Data	Included 63 abstracts presented at the 1996 and 1998 First and Second Balkan Congress of Oncology meetings Included all abstracts by authors affiliated to Serbian institutions and responding to enquiry
Comparisons	Proportion of abstracts published Oral versus poster presentation
Outcomes	42 of 63 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 16/23 abstracts presented orally versus 26/40 abstracts presented as poster published
Notes	Oncology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup, so low risk of bias.
Search for publications?	No	Contacted abstract authors, but response rate <80%.
Follow‐up time?	Yes	The meeting in 1996 had 5 years follow‐up, but the meeting in 1998 only had 36 months follow‐up.
Matching?	Yes	Matched through contact with abstract author.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis and unspecified statistical tests.

Wai 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by all authors and title Matching criteria not reported
Data	Included 270 abstracts presented at the 1998 to 2000 International Society for the Study of the Lumbar Spine meetings Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published Experimental versus RCT versus cohort versus case control versus descriptive design Statistical significance versus without versus no difference versus negative result Clinical versus biomechanical versus basic science Blinding English language country
Outcomes	146 of 270 abstracts published Proportion of abstracts published by time Cumulative proportion of abstracts published at 36 months showed proportion published = 50.4% (136/270 abstracts) Factors related to proportion of abstracts published included 68.4% of abstracts describing 'experimental' research versus 70.8% of abstracts describing an RCT versus 67.6% of abstracts describing a cohort versus 25% of abstracts describing a case control study versus 30.8% of abstracts presenting a descriptive study published 61.8% of abstracts presenting statistically significant results versus 50% of abstracts 'without statistical significance' versus 58.8% of abstracts presenting results with no difference versus 50% of abstracts presenting negative results published 48.8% of abstracts describing clinical research versus 57.5% of abstracts describing biomechanical research versus 55.7% of abstracts describing basic science research published 84.6% of abstracts with blinding published 54.8% of abstracts from English language country published
Notes	Orthopedics/orthopedic surgery ‐ spine Reported having received no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of positive results, type science, study design, presence of blinding, and origin in English language country with publication multivariable logistic regression.

Walby 2001.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to December 2000 Person completing the search not reported Searched by keyword, and presenting author Matching criteria not reported Handsearched Emergency Medicine through December 2000
Data	Included 207 abstracts presented at the 1995 to 1998 Australasian College of Emergency Medicine and Australasian Society for Emergency Medicine and 1996 and 1998 International Conference on Emergency Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean and median time to publication Cumulative proportion of abstracts published
Outcomes	73 of 207 abstracts published Proportion of abstracts published by time Mean time to publication = 12.6 months Median time to publication = 11 months Cumulative proportion of abstracts published at 48 months showed proportion published = 35.3% (73/207 abstracts) No factors related to proportion of abstracts published reported
Notes	Emergency medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 1 database and handsearched a journal.
Follow‐up time?	No	Only the meetings in 1995 and 1996 had at least 48 months follow‐up. All other meetings had less than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.

Walsh 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE, Google/Google Scholar, and Eric to November 2012 Person completing the search not reported Searched by first, second, and last author, keywords, and title Matching criteria not reported
Data	Included 449 abstracts presented at the 2005 to 2006 Research in Medical Education Conference and Canadian Conference on Medical Education meetings Included all abstracts excluding 7 (4 withdrawn, 1 missing, 2 published before abstract deadline)
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication Cumulative proportion of abstracts published Oral versus poster presentation Multi‐centered versus single center Mixed methods versus qualitative methods versus quantitative methods Multi‐center authorship versus not Research scholarship versus evaluation of program or innovation scholarship versus other scholarly activities Completed works versus work in progress
Outcomes	156 of 449 abstracts published 86/232 abstracts presented at the Research in Medical Education Conference meetings, and 70/217 at the Canadian Conference on Medical Education meetings published Proportion of abstracts published by time Median time to publication = 20 months (IQR = 10 to 30 months, range = ‐9 to 76 months) Cumulative proportion of abstracts published at 60 months showed proportion published = 33.6% (151/449 abstracts) Factors related to proportion of abstracts published included 80/165 abstracts presented orally versus 76/284 abstracts presented as poster published 33/81 abstracts with multiple centers versus 110/301 abstracts with a single center published 20/36 methods with mixed methods design versus 32/69 abstracts with qualitative design versus 91/254 abstracts with quantitative design published 60/118 abstracts with multi‐center authorship versus 96/331 abstracts without multi‐center authorship published 110/283 abstracts on research versus 33/99 abstracts on evaluation of program or innovation versus 16/67 abstracts on other scholarly activity published 125/300 abstracts of completed works versus 18/82 abstracts with works in progress published
Notes	Medical education/library science Government funding and funding through the Hospital of Sick Children, Department of Paediatrics fellowship grant Data abstracted from letter
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting, type of presentation, multi‐center status, study design, number of authors, type of science, completion status, and presence of author with doctorate degree with publication using multivariable logistic regression analyses.

Wang 1999.

Methods	Identification of subsequent full‐length publications Searched electronic database Melvyl MEDLINE Plus from 1990 through 1997 Person completing the search not reported Searched by all authors and title Matched abstract to full‐length publication by All authors Contents Title Methodology
Data	Included 1188 abstracts presented at the 1990 to 1992 North American Spine Society, 1991 to 1993 Scoliosis Research Society, and 1991 to 1993 International Society for the Study of the Lumbar Spine meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting
Outcomes	517 of 1188 abstracts published 220/545 abstracts presented at the North American Spine Society meetings, 145/308 at the Scoliosis Research Society meetings, and 152/335 at the International Society for the Study of the Lumbar Spine meetings published Proportion of abstracts published by time Cumulative proportion of abstracts published at 48 months showed proportion published = 45.4% (152/335 abstracts) for International Society for the Study of the Lumbar Spine meetings Cumulative proportion of abstracts published at 48 months showed proportion published = 47.1% (145/308 abstracts) for the Scoliosis Research Society meetings Cumulative proportion of abstracts published at 48 months showed proportion published = 43.9% (239/545 abstracts) for the North American Spine Society meetings No factors other than meeting year related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery ‐ spine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Waters 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by first and senior author, and keywords Matching criteria not reported
Data	Included 201 abstracts presented at the 2002 to 2007 British HIV Association meetings Included all abstracts accepted as oral presentations
Comparisons	Proportion of abstracts published Site
Outcomes	107 of 201 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published reported included Proportion of abstracts published by site ranged from 25 to 83% (P = 0.173 for trend).
Notes	Infectious disease/immunology ‐ HIV Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Unclear	Included all abstracts presented orally.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of 'site' with publication using stratified analysis and Chi2 test.

Weale 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date not reported but 'median of 28 months after each meeting' Search completed by investigator Searched by all authors Matched abstract to full‐length publication by Title Contents
Data	Included 768 abstracts presented at the 2001 Association of Surgeons of Great Britain and Ireland, Vascular Surgical Society of Great Britain and Ireland, British Transplantation Society, and Association of Coloprotocology of Great Britain & Ireland meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Survival analysis of publication rate by meeting Oral versus poster presentation
Outcomes	246 of 768 abstracts published 75/308 abstracts presented at the Association of Coloprotocology of Great Britain & Ireland meeting, 47/132 at the British Transplantation Society meeting, 30/56 at the Vascular Surgical Society of Great Britain and Ireland meeting, and 94/272 at the Association of Surgeons of Great Britain and Ireland meeting published Proportion of abstracts published by time Survival analysis of proportion published at 24 months = 33.8% for the Association of Surgeons of Great Britain and Ireland meeting Survival analysis of proportion published at 24 months = 53.6% for Vascular Surgical Society of Great Britain and Ireland meeting Survival analysis of proportion published at 24 months = 32.6% for the British Transplantation Society meeting Survival analysis of proportion published at 24 months = 23.1% for the Association of Coloprotocology of Great Britain & Ireland meeting Factors related to proportion of abstracts published included 171/439 abstracts presented orally versus 75/329 abstracts presented as poster published
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meetings only had a median follow‐up of 28 months.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	Unclear	Examined association of meeting and type of presentation with publication using a Cox proportional hazards model, although unclear if multivariable analysis was completed.

Weitz 2005.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Lilacs, and Scielo Chile; dates of search not reported Person completing the search not reported Searched by all authors, keywords, and title Matching criteria not reported Contacted abstract authors directly
Data	Included 392 abstracts presented at the 1998 to 2002 Congresos Chilenos de Gastroenterología meetings Included all abstracts except works conducted entirely in foreign countries, videos, repeated presentations (10) and abstracts for which authors did not respond (16)
Comparisons	Proportion of abstracts published
Outcomes	100 of 392 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Gastroenterology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors directly.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.

Whitehouse 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Embase, Google/Google Scholar; dates of search not reported Search completed by investigator Searched by first, second, and third author, title, and contents Matched abstract to full‐length publication by Contents
Data	Included 1138 abstracts presented at the 2004 and 2005 British Hip Society, 2004 and 2006 European Hip Society, 2004 to 2006 British Orthopedic Association, and 2003 and 2005 European Federation of Orthopaedics and Trauma meetings Included all abstracts accepted for oral presentation related to hip surgery, except Instructional Course Lectures
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting
Outcomes	252 of 1138 abstracts published 17/90 abstracts presented at the 2004 British Hip Society meeting, 15/73 at the 2005 British Hip Society meeting, 27/68 at the 2004 European Hip Society meeting, 30/173 at the 2006 European Hip Society meeting, 8/23 at the 2004 British Orthopedic Association meeting, 5/31 at the 2005 British Orthopedic Association meeting, 8/41 at the 2006 British Orthopedic Association meeting, 57/293 at the 2003 European Federation of Orthopaedics and Trauma meeting, and 85/346 at the 2005 European Federation of Orthopaedics and Trauma meeting published. Proportion of abstracts published by time Mean time to publication = 4.9 months Mean time to publication = 6.9 months for 2004 British Hip Society meeting Mean time to publication = 6.8 months for 2005 British Hip Society meeting Mean time to publication = ‐4.1 months for 2004 European Hip Society meeting Mean time to publication = 0.7 months for 2006 European Hip Society meeting Mean time to publication = 8.3 months for 2004 British Orthopedic Association meeting Mean time to publication = 6.0 months for 2005 British Orthopedic Association meeting Mean time to publication = 2.0 months for 2006 British Orthopedic Association meeting Mean time to publication = 9.5 months for 2003 European Federation of Orthopaedics and Trauma meeting Mean time to publication = 8.3 months for 2005 European Federation of Orthopaedics and Trauma meeting No factors other than meeting and meeting year related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented orally that described a specific topic with reasonable exceptions, so low risk of bias.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	No	Matched by only 1 criterion.
Adjustment for confounding?	No	Examined association of meeting and meeting year with publication using stratified analysis and Fisher's Exact test.

Wieser 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Scopus, and CAB abstracts; dates of search not reported Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by One author One intervention One species Contacted abstract authors directly
Data	Included 607 abstracts presented at the 2004 to 2008 American College of Veterinary Anesthesiologists and Association of Veterinary Anesthesists meetings, and 2003 and 2006 World Congress of Veterinary Anaesthesiology meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published 'Positive' versus not 'positive' RCT or controlled trial design versus other design North America versus Europe versus rest of world Funding reported versus not reported Animal species
Outcomes	384 of 607 abstracts published Proportion of abstracts published by time Mean time to publication = 25 months (SD = 19; range = ‐35 to 104 months) Survival analysis of proportion published at 102 months= 63.3% Factors related to proportion of abstracts published included 258/415 abstracts with 'positive' results (undefined) versus 126/192 abstracts without 'positive' results published 260/383 abstracts with RCT or controlled trial design versus 124/224 abstracts describing other design published 177/256 abstracts originating in North America versus 149/244 abstracts originating in Europe versus 58/107 abstracts originating elsewhere published 74/99 abstracts reporting funding versus 310/508 abstracts without funding or not reporting funding published
Notes	Other non‐clinical specialties ‐ veterinary anesthesiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors directly.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of positive results, study design, multi‐center status, country of origin, funding status and animal species with publication using stratified analysis and Fisher's Exact tests.

Wilson 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, MEDLINE, and ALOIS; dates of search not reported Search completed by the investigator Searched by first and last author Matching criteria not reported Contacted abstract authors directly
Data	Included 250 abstracts presented at the 2009 International Psychogeriatric Association Congress, Clinical Trials on Alzheimer’s Disease and Alzheimer’s Association International Conference meetings Included all abstracts related to dementia diagnosis
Comparisons	Proportion of abstracts published 'Positive' versus not 'positive' Oral versus poster presentation European versus North American versus rest of world origin Sample size Meeting Clinic versus community versus Alzheimer's Disease Neuroimaging Initiative versus unknown setting Imaging versus cerebrospinal fluid versus other biomarkers versus non‐biomarkers Inclusion of participant with versus without mild cognitive impairment
Outcomes	97 of 250 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 87/209 abstracts with 'positive' (defined as 'experimental better than control’) versus 10/41 abstracts with not 'positive' abstract results published 26/53 abstracts presented orally versus 71/197 abstracts presented as poster published 44/105 abstracts with sample size > 100 versus 35/16 abstracts with sample size < 100 published 36/85 abstracts originating in North America versus 40/91 abstracts originating in Europe versus 10/44 abstracts originating elsewhere versus 11/30 abstracts of international origin published
Notes	Neurology ‐ dementia Received 'grant funding'
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts of a specific topic.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors directly.
Follow‐up time?	Unclear	Dates of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting, positive results, type of presentation, sample size, country of origin, setting, type test, and inclusion criteria with publication using multivariable logistic regression.

Winnik 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Cochrane Central Register of Controlled Trials, Google/Google Scholar for 4 years Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by Last author Sample size Primary endpoint Number of study groups Conclusions
Data	Included 1002 abstracts submitted to the 2006 European Society of Cardiology meeting Included a random selection of abstracts
Comparisons	Proportion of abstracts published Sample size greater than 100 versus equal to or less than 100 Oral versus poster presentation Accepted versus rejected for conference presentation Clinical research versus basic science research RCT design versus observational design versus systematic review versus other design Female versus male first author Female versus male last author Subspecialty
Outcomes	293 of 1002 abstracts published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 117/415 abstracts with sample size greater than 100 versus 120/438 abstracts with sample size equal to or below 100 published 27/74 abstracts presented orally versus 117/302 abstracts presented as posters published 143/376 abstracts accepted for presentation versus 150/626 rejected abstracts published 251/899 abstracts describing clinical research versus 42/103 abstracts describing basic science research published 22/65 abstracts with RCT design versus 214/784 abstracts with observational design versus 4/8 abstracts of systematic reviews versus 11/42 abstracts with other design published 250/810 abstracts with university affiliation versus 43/192 abstracts without a university affiliation published 71/261 abstracts with female first author versus 217/723 abstracts with male first author published 25/131 abstracts with female last author versus 259/842 abstracts with male last author published
Notes	Cardiology Funding by the Foundation for Cardiovascular Research, Zurich, Switzerland, the Swiss National Science Foundation, and the University Research Priority Program Integrative Human Physiology at the University of Zurich
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	Yes	Searched 3 databases.
Follow‐up time?	Yes	The meeting had 48 months follow‐up.
Matching?	Yes	Matched by 5 different criteria.
Adjustment for confounding?	Yes	Examined association of sample size, type presentation, acceptance for presentation, type science, study design, author gender, and subspecialty with publication using multivariable regression.

Wong 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to March 2005 Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 67 abstracts submitted to the 1998 and 1999 New South Wales State Committee, Registrars Paper Day meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean and median time to publication Sample size greater than 50 versus less than 50 Oral versus poster presentation Accepted versus rejected for conference presentation Clinical research versus basic science research 'Prospective' versus 'retrospective' design
Outcomes	28 of 67 abstracts published Proportion of abstracts published by time Mean time to publication = 23.8 months Median time to publication = 21 months (range = 4 to 60 months) Factors related to proportion of abstracts published included 11/33 abstracts with sample size greater than 50 versus 17/34 abstracts with sample size less than 50 published 17/34 abstracts presented orally versus 9/23 abstracts presented as posters published 26/57 abstracts accepted for presentation versus 2/10 rejected abstracts published 23/55 abstracts describing clinical research versus 5/12 abstracts describing basic science research published 17/25 abstracts with 'prospective' study design versus 11/42 abstracts with 'retrospective' study design published
Notes	Surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of sample size, type of presentation, acceptance for presentation, type of science, and prospective status with publication using stratified analysis and Chi2 tests.

Wong 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, Embase for 5 years Search completed by investigator Searched by all authors and keywords Matched abstract to full‐length publication by Contents Title Similarities in authors' names
Data	Included 790 abstracts presented at the 2003 Annual Scientific Sessions of Heart Rhythm Society meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication Survival analysis of publication rate 'Positive' versus not 'positive' Clinical research versus basic science research versus animal research versus simulation studies RCT design versus non‐RCT design
Outcomes	377 of 790 abstracts published Proportion of abstracts published by time Median time to publication = 16.7 months (IQR = 10.6 to 27.6 months) Survival analysis of proportion published at 72 months = 47.7% Factors related to proportion of abstracts published included 185/370 'positive' (defined as significant results) versus 192/420 not 'positive' abstract results published 256/586 abstracts describing clinical research versus 28/43 abstracts describing basic science research versus 78/133 abstracts describing animal research versus 15/28 abstracts describing simulation studies published 37/62 abstracts with RCT design versus 340/728 abstracts with non‐RCT design published
Notes	Cardiology ‐electrophysiology Funding by the National Heart Foundation of Australia, National Heart Foundation of New Zealand, the Kidney Health Foundation of Australia, the University of Adelaide, the Royal Adelaide Hospital, the National Health and Medical Research Council of Australia
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	The meeting had 5 years follow‐up.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, type science, and study design with publication using multivariable logistic regression.

Wong 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE; date of search not reported Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 4848 abstracts presented at the 2003 American Heart Association, and the 1994, 1997, 2000, 2003, and 2006 Heart Rhythm Society meetings Included all abstracts from the Heart Rhythm Society and a random sample from the American Heart Association
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication by meeting year of the Heart Rhythm Society Clinical research versus animal research versus cellular research RCT versus other designs 'Positive' versus not 'positive
Outcomes	2406 of 4848 abstracts published 603/1000 abstracts presented at the American Heart Association and 281/521 at the 1994 meeting, 320/740 at the 1997 meeting, 358/804 at the 2000 meeting, 377/790 at the 2003 meeting, and 467/993 at the 2006 meeting of the Heart Rhythm Society published Proportion of abstracts published by time Median time to publication = 21 months (IQR = 10 to 34 months) for the 1994 Heart Rhythm Society meeting Median time to publication = 18 months (IQR = 11 to 32 months) for the 1997 Heart Rhythm Society meeting Median time to publication = 16 months (IQR = 6 to 31 months) for the 2000 Heart Rhythm Society meeting Median time to publication = 18 months (IQR = 8 to 27 months) for the 2003 Heart Rhythm Society meeting Median time to publication = 13 months (IQR = 5 to 22 months) for the 2006 Heart Rhythm Society meeting Factors related to proportion of abstracts published included 1521/3321 abstracts describing clinical research versus 595/992 abstracts describing animal research versus 334/524 abstracts describing cellular research published RCT design (OR = 1.31, 95% CI:1.01 to 1.70) 'Positive' findings (OR = 1.19, 95% CI:1.05 to 1.36)
Notes	Cardiology ‐ electrophysiology Funded through scholarships and fellowship from Rhodes, National Health, Medical Research Council of Australia, National Heart Foundation of Australia, National Health and Medical Research Council of Australia
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts from one series conference and a random sample of a different conference.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting year, positive results, type of science, and study design with publication using multivariable logistic regression.

Yalcinkaya 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Person completing the search not reported Searched by all authors, keywords, and title Matched abstract to full‐length publication by All authors Title Contents
Data	Included 770 abstracts presented at the 2007 Turkish National Orthopedics and Traumatology Congress meeting Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Cumulative proportion of abstracts published Oral versus poster presentation
Outcomes	227 of 770 abstracts published Proportion of abstracts published by time Mean time to publication = 14.9 months (SD = 16.1, range = ‐33 to 55 months) Cumulative proportion of abstracts published at 36 months showed proportion published = 26.1% (201/770 abstracts) Factors related to proportion of abstracts published included 116/264 abstracts presented orally versus 111/506 abstracts presented as poster published
Notes	Emergency medicine ‐ orthopedics and traumatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by 3 different criteria.
Adjustment for confounding?	No	Examined association of type presentation with publication using stratified analysis and t test.

Yamaguchi 1990.

Methods	Identification of subsequent full‐length publications Contacted abstract authors directly Response rate = 61.9% (599/967)
Data	Included 599 abstracts submitted to the 1986 Japanese Society for Bacteriology, Japanese Society for Bacteriology regional, Japanese Society of Internal Medicine, and Japanese Society of Internal Medicine regional meetings Included all abstracts with author response (61.9% of 967 = 599)
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published by meeting
Outcomes	488 of 599 abstracts published 25/54 abstracts presented at the Japanese Society of Internal Medicine regional meeting, 234/269 at the Japanese Society of Internal Medicine, 37/51 at the Japanese Society for Bacteriology regional meeting, and 192/226 at the Japanese Society for Bacteriology published Proportion of abstracts published by time Cumulative proportion of abstracts published at 48 months = 84.9% (192/226 abstracts) for the Japanese Society for Bacteriology meeting Cumulative proportion of abstracts published at 48 months = 72.5% (37/51 abstracts) for the Japanese Society for Bacteriology regional meeting Cumulative proportion of abstracts published at 48 months = 87.0% (234/269 abstracts) for the Japanese Society of Internal Medicine Cumulative proportion of abstracts published at 48 months = 46.3% (25/54 abstracts) for the Japanese Society of Internal Medicine regional meeting No factors other than meeting related to proportion of abstracts published reported
Notes	Infectious disease/immunology ‐ bacteriology, internal medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	No	Included only abstracts with author response, not a representative sample.
Search for publications?	No	Contacted abstract authors directly but response rate was less than 80%.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Contacted abstract authors directly.
Adjustment for confounding?	No	Examined association of meeting with publication using stratified analysis.

Yentis 1993.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE from 1985 to 1990 Person completing the search not reported Searched by first author Matched abstract to full‐length publication by All authors Contents
Data	Included 215 abstracts, which was a systematic selection of abstracts presented at the 1985 American Society of Anesthesiologists, International Anesthesia Research Society, the Anaesthesia Research Society, and Canadian Anaesthetists' Society meetings
Comparisons	Proportion of abstracts published Cumulative proportion of abstracts published by meeting
Outcomes	108 of 215 abstracts published, 54/114 from the American Society of Anesthesiologists meeting, 22/39 from the International Anesthesia Research Society meeting, 19/33 from the Anaesthesia Research Society, and 13/29 from the Canadian Anaesthetists' Society meeting Proportion of abstracts published by time Cumulative proportion of abstracts published at 60 months showed proportion published = 46.5% (53/114 abstracts) for American Society of Anesthesiologists meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 56.4% (22/39 abstracts) for International Anesthesia Research Society meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 57.6% (19/33 abstracts) for the Anaesthesia Research Society meeting Cumulative proportion of abstracts published at 60 months showed proportion published = 44.8% (13/29 abstracts) for Canadian Anaesthetists' Society meeting No factors other than meeting related to proportion of abstracts published reported
Notes	Anesthesiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included a random selection of abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting with publication using stratified analysis and Chi2 test.

Yilmaz 2013.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed on 20 January, 2011 Person completing the search not reported Searched by all authors Matching criteria not reported
Data	Included 1732 abstracts presented at the 2008 European League Against Rheumatism meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication Oral versus poster presentation Clinical research versus preclinical research RCT design versus non‐RCT design Multi‐centered versus single center Sample size Type disease
Outcomes	601 of 1732 abstracts published Proportion of abstracts published by time Median time to publication = 13 months (range = 0 to 31 months) Factors related to proportion of abstracts published included 119/270 abstracts presented orally versus 482/1462 abstracts presented as poster published 393/1168 abstracts describing clinical research versus 208/564 abstracts describing preclinical research published 65/144 abstracts describing RCTs versus 328/1024 abstracts describing other research designs published 265/590 abstracts with multiple centers versus 336/1042 abstracts with a single center published Sample size (higher than median vs. lower than median: 36.2% vs. 30.3%, P = 0.017)
Notes	Rheumatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	No	The meeting only had 36 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of type of presentation, study design, multi‐center status, sample size, and type disease with publication using stratified analysis and Chi2, Mann‐Whitney U, and ANOVA tests.

Yolcu 2015.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed and Google/Google Scholar until November 2013 Person completing the search not reported Searched by first, second, and last author, and keywords Matching criteria not reported
Data	Included 1322 abstracts presented at the 2007 to 2012 Oral and Maxillofacial Surgery Society of Turkey meetings Included all abstracts
Comparisons	Proportion of abstracts published Mean time to publication Oral versus poster presentation Clinical research versus basic science research versus animal research versus technical note Observational design versus case report versus case series Retrospective design versus prospective design Subspecialty
Outcomes	246 of 1322 abstracts published 34/114 abstracts presented at the 2007 meeting, 59/174 at the 2008 meeting, 31/154 at the 2009 meeting, 42/220 at the 2010 meeting, 50/344 at the 2011 meeting, and 30/316 at the 2012 meeting published Proportion of abstracts published by time Mean time to publication = 17 months (SD = 15) Factors related to proportion of abstracts published included 110/390 abstracts presented orally versus 136/932 abstracts presented as poster published 184/1188 abstracts describing clinical research versus 25/55 abstracts describing basic science research versus 32/70 abstracts describing animal research versus 5/9 abstracts presenting a technical note published 87/436 abstracts describing an observational design versus 91/708 abstracts describing a case report versus 6/44 abstracts describing case series published 32/501 abstracts describing retrospective designs versus 32/235 abstracts describing prospective designs published
Notes	Otolaryngology head and neck surgery ‐ oral and maxillofacial surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 2 databases.
Follow‐up time?	Yes	All meetings except the 2010 to 2012 meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type presentation, type science, study design, prospective status, and subspecialty with publication using stratified analysis and Chi2 tests.

Yoo 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database Melvyl MEDLINE Plus from 1990 to 1998 Search completed by investigator Searched by all authors and title Matched abstract to full‐length publication by All authors Title Contents Methodology
Data	Included 333 abstracts presented at the 1991 to 1993 Arthroscopy Association of North America and the American Orthopaedic Society for Sports Medicine meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published
Outcomes	188 of 333 abstracts published 111/166 abstracts presented at the American Orthopaedic Society for Sports Medicine meetings, and 77/167 at the Arthroscopy Association of North America meetings published Proportion of abstracts published by time Cumulative proportion of abstracts published at 54 months showed proportion published = 46.1% (77/167 abstracts) for the Arthroscopy Association of North America meetings Cumulative proportion of abstracts published at 54 months showed proportion published = 66.9% (111/166 abstracts) for the American Orthopaedic Society for Sports Medicine meetings No factors other than meeting year related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ sports medicine Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had at least 5 years follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Yoon 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed; date of search not reported Search completed by investigator Searched by all authors and keywords Matching criteria not reported
Data	Included 154 abstracts presented at the 1996 to 2001 American Association of Hip and Knee Surgeons meetings Included all abstracts accepted for poster presentation
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication overall and by meeting
Outcomes	73 of 154 abstracts published 4/10 abstracts presented at the 1996 meeting, 11/20 at the 1997 meeting, 6/18 at the 1998 meeting, 13/20 at the 1999 meeting, 19/40 at the 2000 meeting, and 20/46 at the 2001 meeting published Proportion of abstracts published by time Mean time to publication = 23.6 months (SD = 15.7 months) Mean time to publication = 20.8 months (SD = 19.1 months) for 1996 American Association of Hip and Knee Surgeons meeting Mean time to publication = 36.5 months (SD = 19 months) for 1997 American Association of Hip and Knee Surgeons meeting Mean time to publication = 27.3 months (SD = 11.6 months) for 1998 American Association of Hip and Knee Surgeons meeting Mean time to publication = 23.5 months (SD = 9.9 months) for 1999 American Association of Hip and Knee Surgeons meeting Mean time to publication = 9.7 months (SD = 14.5 months) for 2000 American Association of Hip and Knee Surgeons meeting Mean time to publication = 16.6 months (SD = 15.8 months) for 2001 American Association of Hip and Knee Surgeons meeting No factors other than meeting year related to proportion of abstracts published reported
Notes	Orthopedics/orthopedic surgery ‐ hip and knee surgery Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for poster presentation.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Yoon 2010.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Web of Science, Korean Medical Database, KMBase, and National Digital Science LInks between October 2003 and April 2009 Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 475 abstracts submitted to the 2003 to 2008 Korean Society of Gastroenterology and Korea Society of Gastrointestinal Endoscopy meetings Included all abstracts
Comparisons	Proportion of abstracts published Clinical research versus experimental research Subspecialty
Outcomes	95 of 475 abstracts published 25/86 abstracts presented at the 2003 meeting, 34/87 at the 2004 meeting, 15/93 at the 2006 meeting, 13/97 at the 2007 meeting, and 8/112 at the 2008 meetings published Proportion of abstracts published by time not reported Factors related to proportion of abstracts published included 82/422 abstracts describing clinical research versus 13/53 abstracts describing experimental research published
Notes	Gastroenterology ‐ pancreato‐biliary Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 6 databases.
Follow‐up time?	Yes	Only the 2003 and 2004 meetings had more than 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year, type science, and subspecialty with publication using stratified analysis.

Yoon 2012a.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE, PubMed; dates of search not reported Person completing the search not reported Searched by first, second, and last author Matched abstract to full‐length publication by All authors Sample size Title Methodology Results Location
Data	Included 154 abstracts presented at the 2005 First World Conference on Sports Injury Prevention in Oslo meeting Included all abstracts accepted for oral presentation
Comparisons	Proportion of abstracts published Mean time to publication by RCT and observational design Cumulative proportion of abstracts published RCT design versus observational design versus 'experimental' design North American versus European versus rest of the world origin
Outcomes	76 of 154 abstracts published Proportion of abstracts published by time Mean time to publication = 17 months (SD = 13 months) for RCTs Mean time to publication = 12 months (SD = 14 months) for observational studies Cumulative proportion of abstracts published at 42 months showed proportion published = 49.4% (76/154 abstracts) Factors related to proportion of abstracts published included 10/14 abstracts with RCT design versus 63/135 abstracts with observational design versus 3/5 abstracts with 'experimental' design published 18/25 abstracts originating from North America versus 24/52 abstracts from Europe versus 34/77 abstracts from rest of the world published
Notes	Other clinical specialties ‐ sports medicine Reported receiving no funding
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts accepted for oral presentation.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Yes	Matched by six different characteristics.
Adjustment for confounding?	No	Examined association of study design, country of origin, prospective status, abstract quality, and subspecialty with publication using stratified analysis and Chi2 tests.

Yoon 2012b.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed to July 2011 Search completed by investigator Searched by first, second, and senior author, and keywords Matched abstract to full‐length publication by One author Methodology Hypothesis Conclusions
Data	Included 614 abstracts presented at the 2005 to 2009 Annual Scientific Meeting of the Urological Society of Australia and New Zealand meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication for online and print journal publications Clinical research versus basic science research RCT design versus observational design versus 'comparative' design versus survey design Multi‐centered versus single center 'Prospective' versus 'retrospective' design Australian or New Zealand versus international origin Subspecialty
Outcomes	183 of 614 abstracts published 43/143 abstracts presented at the 2005 meeting, 37/93 at the 2006 meeting, 18/80 at the 2007 meeting, 43/147 at the 2008 meeting, and 42/151 at the 2009 meeting published Proportion of abstracts published by time Mean time to publication = 13 months (SD = 13.6 months) for online journal publications Mean time to publication = 14.5 months (SD = 13.9 months) for print journal publications Factors related to proportion of abstracts published included 128/450 abstracts describing clinical research versus 44/102 abstracts describing basic science research published 15/34 abstracts with RCT design versus 122/400 abstracts with observational design versus 64/181 abstracts with 'comparative' design versus 20/98 abstracts with survey design published 8/22 abstracts with multiple centers versus 175/592 abstracts with a single center published 59/208 abstracts with 'prospective' study design versus 63/192 abstracts with 'retrospective' study design published 122/461 abstracts originating in Australia or New Zealand versus 61/153 abstracts of international origin published
Notes	Urology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings before 2008 had at least 48 months follow‐up. All other meetings had less than 48 months follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	No	Examined association of meeting year, type of science, study design, multi‐center status, prospective status, Australian or New Zealand origin, and subspecialty with publication using stratified analysis and Chi2 tests.

Yoshida 2008.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, Cochrane Central Register of Controlled Trials, Scirus, Bireme, LILACS, SciELO, Jornal Vascular Brasileiro to December 2007 Person completing the search not reported Searched by title Matched abstract to full‐length publication by All authors Title
Data	Included 1108 abstracts presented at the 2001 and 2003 Congresso Brasileiro de Angiologia e Cirurgia Vascular meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	70 of 1108 abstracts published 21/541 abstracts presented at the 2001 meeting, and 49/567 abstracts at the 2003 meeting published Proportion of abstracts published by time not reported No factors other than year related to proportion of abstracts published reported
Notes	Other clinical specialties ‐ vascular surgery and angiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 7 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.
Adjustment for confounding?	No	Examined association of meeting year with publication using stratified analysis.

Yuan 2011.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed 5 years after each meeting Person completing the search not reported Searched by all authors and keywords Matching criteria not reported
Data	Included 369 abstracts presented at the 2002 and 2003 American Dental Education Association meetings Included all abstracts accepted as posters and related to education
Comparisons	Proportion of abstracts published
Outcomes	71 of 369 abstracts published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Oral health‐ dental education Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts presented as posters that described specific subgroup, so low risk of bias.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	All meetings had 5 years follow‐up.
Matching?	Unclear	Matching criteria not reported.

Yun 1999.

Methods	Identification of subsequent full‐length publications Searched journal of the Korean Radiological Society Searched until June 19, 1998 Person completing the search not reported Searched by all authors, and title Matching criteria not reported
Data	Included 2457 abstracts presented at the 1992 to 1996 Korean Radiological Society meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication Scientific versus exhibition session Season of the meeting in spring versus autumn
Outcomes	614 of 2457 abstracts published 106/422 abstracts presented at the 1992 meeting, 130/498 at the 1993 meeting, 137/480 at the 1994 meeting, 144/430 at the 1995 meeting, and 97/627 at the 1996 meeting published Proportion of abstracts published by time Mean time to publication = 13.3 months (range = 0 to 56 months) Factors related to proportion of abstracts published included 530/1961 abstracts presented at a scientific session versus 64/496 abstracts presented at a exhibition session published 161/575 abstracts presented in spring versus 453/1882 abstracts presented in autumn published
Notes	Radiology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 journal.
Follow‐up time?	Unclear	The meetings in 1995 and 1996 had less than 48 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting year, type session presentation, and season of presentation with publication using stratified analysis.

Zamakhshary 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database PubMed, MEDLINE; dates of search not reported Search completed by investigator Searched by first, and senior author, and keywords Matching criteria not reported
Data	Included 183 abstracts presented at the 2001 and 2002 Canadian Association of Pediatric Surgeons and American Pediatric Surgery Association meetings Included all abstracts except those describing case reports and guest lectures
Comparisons	Proportion of abstracts published, overall and by meeting Cumulative proportion of abstracts published 'Positive' versus not 'positive' Clinical research versus basic science research Canada versus US versus rest of the world origin
Outcomes	118 of 183 abstracts published 48/83 abstracts presented at the Canadian Association of Pediatric Surgeons meetings, and 70/100 abstracts presented at the American Pediatric Surgery Association meetings published Proportion of abstracts published by time Cumulative proportion of abstracts published at 24 months showed proportion published = 64.5% (118/183 abstracts) Factors related to proportion of abstracts published included 105/151 'positive' (defined as significant results) versus 13/32 not 'positive' abstract results published 91/137 abstracts describing clinical research versus 27/46 abstracts describing basic science research published 18/38 abstracts originating from Canada versus 72/105 abstracts from the US versus 28/40 abstracts from rest of the world published
Notes	Surgery ‐ pediatrics Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts with reasonable exceptions.
Search for publications?	Yes	Searched only 1 database.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Yes	Examined association of meeting, positive results, type science, Canadian or US origin with publication using multivariable logistic regression.

Zaretsky 2002.

Methods	Identification of subsequent full‐length publications Searched electronic database 'computerized database search'; date of search not reported Person completing the search not reported Search criteria not reported Matching criteria not reported
Data	Included 85 abstracts presented at the 1990 to 1992 American Society of Hematology meetings Included all abstracts of phase III trials
Comparisons	Proportion of abstracts published 'Positive' results versus not 'positive' results Sample size > 100 versus < 100 Medical sub‐specialty
Outcomes	69 of 85 abstracts published Proportion of abstracts published by time Time to publication ranged from 2 to 90 months Factors related to proportion of abstracts published included 55/67 abstracts describing 'malignant' discipline versus 10/12 abstracts describing 'thrombosis' versus 1/1 abstract describing 'benign' versus 3/5 abstracts describing 'transfusion' published Studies with sample size > 100 were more likely to be published (P = 0.002) Single center studies were less likely to be published (P = 0.01) Proportion of positive and negative trials published was not significantly different (P ‐0.53)
Notes	Hematology Funding not reported Data extracted from abstract
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts describing phase III trials.
Search for publications?	Unclear	Searched database, but number and type not reported.
Follow‐up time?	Unclear	Date of search not reported.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of positive results, sample size, multi‐center status, and subspecialty with publication using Chi2 tests.

Zheng 2007.

Methods	Identification of subsequent full‐length publications Searched electronic database CBM, Chinese journal full‐text database, Chinese science & technology journal database and Wan Fang database from 1983 to 2005 Search completed by investigators Search criteria not reported Matched abstract to full‐length publication by Contents Authors Other methods of searching not clear
Data	Included 523 abstracts presented at the 1988 and 2001 China National Academic Stomatological Conferences meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting
Outcomes	330 of 523 abstracts published 195/310 abstracts presented at the 1988 meeting, and 135/213 at the 2001 meeting published Proportion of abstracts published by time not reported No factors related to proportion of abstracts published reported
Notes	Oral health ‐ stomatology Funding not reported Article in Chinese, details extracted from abstract of full publication
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	Yes	Searched 4 databases.
Follow‐up time?	Yes	All meetings had at least 48 months follow‐up.
Matching?	Yes	Matched by 2 different criteria.

Özyurt 2012.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE until December 31, 2011 Person completing the search not reported Searched by all authors Matching criteria not reported
Data	Included 1023 abstracts accepted at the 2004, 2006, and 2008 Turkish National Dermatology meetings Included all abstracts
Comparisons	Proportion of abstracts published, overall and by meeting Mean time to publication, overall and by meeting Cumulative proportion of abstracts published by meeting Oral versus poster presentation Subspecialty
Outcomes	135 of 1023 abstracts published 58/368 abstracts presented at the 2004 meeting, 45/328 at the 2006 meeting, and 32/327 at the 2008 meeting published Proportion of abstracts published by time Mean time to publication = 22.5 months (range = 0 to 86 months), overall Mean time to publication = 23.2 months for the 2004 meeting Mean time to publication = 27.2 months for the 2006 meeting Mean time to publication = 14.5 months for the 2008 meeting Cumulative proportion of abstracts published at 48 months showed proportion published = 15.8% (58/368 abstracts) for the 2004 meeting Cumulative proportion of abstracts published at 48 months showed proportion published = 13.7% (45/328 abstracts) for the 2006 meeting Cumulative proportion of abstracts published at 48 months showed proportion published = 9.8% (32/327 abstracts) for the 2008 meeting Factors related to proportion of abstracts published included 29/134 abstracts presented orally versus 106/889 abstracts presented as poster published
Notes	Dermatology Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The 2008 meeting only had 36 months follow‐up. All other meetings had at least 48 months follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	No	Examined association of meeting year and type presentation with publication using stratified analysis and unspecified statistical tests.

CI: confidence nterval
 HR: hazard ratio
 IQR: interquartile range
 OR: odds ratio
 RCT: randomized ontrolled trial
 SD: standard deviation

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Agustsdottir 1995	Authors did not look at proportion of abstracts published, looked at surveys to authors of abstracts related to patient‐related oncology submitted to a meeting.
Aleixandre‐Benavent 2009	Authors examined publication of abstracts presented at a scientometrics and informatics conference rather than biomedical science.
Alfaro 2015	Editorial on abstract submission process.
Bax 2014	Authors did not look at proportion of abstracts published, but described studies examining proportion of abstract results published as full‐length articles from the Spanish Society of Cardiology.
Berger 2000	Authors did not look at proportion of abstracts published, but only commented on subsequent publication of results first appearing in abstracts.
Bhandari 2005	Authors did not look at proportion of abstracts published, but examined duplicate presentations at meetings.
Bydder 2006	Authors did not look at proportion of abstracts published, but examined process for selection of abstracts for presentation at a meeting.
Callaham 2001	Authors did not look at proportion of abstracts published, but examined the citations of studies first presented as abstracts and subsequently published.
Castro 2009	Editorial on recommendation related to trial registration.
Cloft 2001	Authors did not look at proportion of abstracts published, but examined proportion of subsequent definitive publications of published preliminary reports defined by having the word 'preliminary' or 'pilot' in title.
Covington 2012	Number of summary reports or abstracts presented and published not presented and not able to be calculated from report.
Cuellar 2005	Follow‐up of less than 24 months.
de Carvalho 2013	Editorial on abstract submissions for the XVII Brazilian Congress of Assisted Reproduction.
de Castro 2011	Editorial on impact on scientific community of Brazilian investigators.
De Giorgi 2014	Follow‐up of less than 24 months for majority of abstracts.
Decullier 2007	Authors did not look at proportion of abstracts published, but examined follow‐up of protocols approved from ethics committees.
Dyrbye 2008	Authors did not look at proportion of abstracts published, but examined research productivity of residents.
Estes 2009	Authors did not look at proportion of abstracts published, but discussed a study that examined publication rates of conference abstracts.
Fennewald 2005	Authors examine publication of abstracts presented at a library science conference rather than biomedical science.
Floch 2002	Introduction to an issue of a journal publishing the abstracts of a specific conference.
Garvey 1970	Follow‐up of less than 24 months.
Garvey 1971	Inexact totals of 'technical reports' or abstracts given and only with approximate proportion of subsequent publications, e.g. Quote: "We studied over 1,000 technical reports that were produced by psychologists in 1962 and found that the main content of one‐third of these had been published in a scientific journal by 1965".
Gidding 1992	Number of summary reports or abstracts presented and published not presented and not able to be calculated from report.
Godkin 1993	Number of summary reports or abstracts presented and published not presented and not able to be calculated from report.
Graber 2013	Authors did not look at proportion of abstracts published, but compared abstracts and body within a full‐length article
Harris 2010	Authors did not look at proportion of abstracts published, but investigated the time to publication from studies in a specific subspecialty, but look from a specific conference(s) but rather from individual papers.
Hayes 2011	Editorial on publication practices of phase III cancer trials.
Holmes 2006	Authors did not look at proportion of abstracts published, but examined research productivity of residents.
Hopewell 2006	Authors did not look at proportion of abstracts published, but compared trial information in abstracts selected known to have a full publication.
Hopewell 2007a	Systematic review of studies investigating the publication rate and time to publication of a sample of studies.
Huber 2001	Follow‐up of less than 24 months for a part of the meetings, and we could not calculate length of follow‐up from report for individual year; author contact unsuccessful.
Humberto 2011	Comment on whether to include abstracts in bibliographies.
Jourbert 2004	Number of summary reports or abstracts presented and published not presented and not able to be calculated from report.
Kain 2010	Number of summary reports or abstracts presented and published not presented and not able to be calculated from report.
Kelly 1998	Authors did not look at proportion of abstracts published, but Investigated proportion of abstracts published in journals from meetings of a selection of societies. The abstract to full‐text publication rate was not investigated.
Khanna 2013	Authors did not look at proportion of abstracts published, but examined the concordance of trial characteristics as presented in a conference abstract with the corresponding characteristics as presented at the conference.
Kho 2009	Authors did not look at proportion of abstracts published, but examined abstracts related to a specific intervention rather than a set of conference abstracts from a specific meeting.
Knobloch 2009	Commentary to a study investigating the publication rates of meeting abstracts.
Koren 1986	Number of summary reports or abstracts presented and published not presented and we could not calculate from report.
Kurien 2014	Number of summary reports or abstracts presented and published not presented and we could not calculate from report.
Leal 2010	Number of summary reports or abstracts presented and published not presented and we could not calculate from report.
Leder 1985	Editorial on types of conference presentations.
Lohr 2006	Authors did not look at proportion of abstracts published, but examined research productivity of residents.
Mallory 1999	Letter by the editor stating the lack of rigor with which the full‐text publication of a meeting abstract is sought and achieved by authors.
Manjarin 2009	Authors did not look at proportion of abstracts published, but examined research productivity of residents.
Martinez Rubio 2002	Authors did not look at proportion of abstracts published, but examined characteristics of abstracts presented at the 32nd Argentinian Congress of Pediatrics
McLennan 2009	Authors did not look at proportion of abstracts published, but investigated the proportion of articles focusing on urogynecology in selected journals over the years.
Murphy 2001	Authors did not look at proportion of abstracts published, but examined research productivity of residents.
Olsson 2006	Editorial comment for a study investigating the publication rate of meeting abstracts.
Ramsey 2008	Authors did not look at proportion of abstracts published, but examined proportion of trials that had been registered.
Raptis 2010	Authors did not look at proportion of abstracts published, but examined the peer review process of the European Surgical Association in which abstracts are simultaneously reviewed for presentation and as manuscripts for Annals of Surgery.
Ribic 2011	Authors did not look at proportion of abstracts published, but compared primary outcomes reported in full‐length publications with those reported in the abstract of the full‐length report.
Rodriguez 2009	Authors did not look at proportion of abstracts published, but examined subsequent publication of protocols submitted to an ethics review board.
Saad 2010	Authors did not look at proportion of abstracts published, but examined country of origin of presented abstracts rather than subsequent publication.
Singer 1999	Number of summary reports or abstracts presented and published not presented and we could not calculate from report.
Smith 2011	Authors did not look at proportion of abstracts published, but examined proportion of submitted posters that were presented at a specific conference.
Song 2009	Systematic review of studies reporting on the odds of publication by study results.
Song 2009a	Systematic review of studies reporting on the odds of publication by study results.
Stewart 2010	Authors did not look at proportion of abstracts published, but examined acceptance of abstracts for presentation at a conference.
Stewart 2013	Editorial co‐published in numerous journals on the value of having residents present at conferences. (We include only a single reference here.)
Stovold 2011	Authors did not look at proportion of abstracts published, but compared abstracts identified by handsearching with those identified by searching Embase.
Takeda 2008	Systematic review of studies investigating publication bias in studies focusing on a specific range of medication.
Tannenbaum 1978	Commentary about Micropaper Editions of conference abstracts.
Taran 2008	Number of summary reports or abstracts presented and published not presented and we could not calculate from report.
Tas 2008	Authors did not look at proportion of abstracts published, but examined the country of origin of abstracts.
Teel 1990	Authors did not look at proportion of abstracts published, but compared differences between oral presentations and written publications.
Timmer 2001b	Number of summary reports or abstracts presented and published not presented and we could not calculate from report.
Tintinalli 2001	Commentary on a study published in the same journal issue dealing with publication rates.
Tricco 2008	Authors did not look at proportion of abstracts published, but examined proportion of protocols of Cochrane Reviews published.
Unalp 2007	Authors did not look at proportion of abstracts published, but examined the correlation between the quality of presentation and subsequent publication.
von Elm 2001	Commentary to a study investigating the publication rates from meeting abstracts.
von Elm 2003	Systematic review of studies investigating the publication rate for meeting abstracts as full texts.
Wein 2005	Editorial comment to an abstract of a study on full publication of meeting abstracts.
Wong 2010	Authors did not look at proportion of abstracts published, but examined reasons for non‐publication.
Yalcinkaya 2014	Authors did not look at proportion of abstracts published, but examined publications within a single journal rather than a follow‐up of conference abstracts.

Differences between protocol and review

The original protocol evaluated the proportion of abstracts published in full using a weighted mean average. However, because of the substantial heterogeneity with length of follow‐up, we chose to estimate the overall proportions in this updated review using a random‐effects meta‐analysis based on logit‐transformed proportions. The updated point estimates are similar, but their confidence intervals are wider, thus better taking into account the heterogeneity.

Contributions of authors

• This is an update of a previous Cochrane Methodology Review (Scherer 2007).

• Roberta W Scherer reviewed the results of the searches for reports eligible for the review, extracted data from the reports, developed the 'Risk of bias' assessment used for this review, assessed reports for risk of bias, obtained information from authors, entered the data into RevMan, conducted statistical analyses, and drafted the review.

• Joerg J Meerpohl reviewed the results of the updated search for reports eligible for the review, extracted data from the reports, developed the 'Risk of bias' assessment used for this review, assessed reports for risk of bias, contributed to statistical analyses and critically read and edited the review.

• Nadine Pfeifer reviewed the results of the updated search for reports eligible for the review, reviewed citations of included reports for eligible reports, extracted data from the reports, assessed reports for risk of bias, entered the data into RevMan, and critically read and edited the review.

• Christine Schmucker conducted the updated literature search, prepared the study flow chart and critically read and edited the review.

• Guido Schwarzer provided data management and statistical expertise, conducted statistical analyses for the previous systematic review and critically read and edited the updated review.

• Erik von Elm developed the risk of bias assessment used for this review, contributed to statistical analyses and critically read and edited the review.

Sources of support

Internal sources

• Johns Hopkins Bloomberg School of Public Health, USA.

• University of Bern, Department of Social and Preventive Medicine, Switzerland.

• University of Freiburg, Evidence in Medicine / Cochrane Germany, Germany.

External sources

• No sources of support supplied

Declarations of interest

None.

New search for studies and content updated (conclusions changed)