The role of collaboration and mentorship in the publication of surgical resident research

Abstract

Background

Research is an integral part of surgical training and a mandated competency by national accreditation bodies. Most residents engage in research, but the conversion of this research into peer-reviewed publications is unknown. The objectives of this study were to assess the conversion rate of resident research into published manuscripts and determine what variables predict publication.

Methods

Through a retrospective design, 99 resident research abstracts were identified from the Surgery Research Day at the University of Saskatchewan 2008-2018. Publication status was verified using Google Scholar and PubMed. Variables associated with resident-specific, mentor-specific, and project-specific variables were assessed for their role in predicting publication.

Results

Fifty-two (53%) of the 99 abstracts were published in a peer-reviewed journal, and 43 (43%) were presented at a national conference. Logistic regression analysis revealed multidisciplinary research (OR 4.46, CI 1.8-11.4, p = 0.002), projects involving multiple resident researchers (OR 2.56, CI 1.02-6.43, p = 0.045), and faculty supervisor having > 25 publications (OR 2.46, CI 1.03-5.88, p = 0.042) as significant predictors of publication.

Conclusions

Our study identifies three variables related to collaboration and mentorship that can serve as potential starting points to increase research productivity amongst medical trainees.

Introduction

Research is an essential component of surgical education. Engaging in and interpreting novel literature is paramount in providing up-to-date, evidence-based medicine.^1–3 Furthermore, the publication of resident research projects can also have a positive impact on surgical trainees.^1,2 It is often a barometer for academic productivity and competency in research may be a stimulus for establishing a future career as surgical scientists. Based on previous reports, the conversion rate of resident research into peer-reviewed publications is typically between 30-50%.^4–6 Barriers to publication have been described, including time limitations, lack of quality in study design, and lack of prior research experience by surgical residents.^7,8 A 2015 study of plastic surgery residents at Johns Hopkins University found that increased post-graduate year was inversely correlated with higher abstract conversion.⁴ The same study also found that the academic rank of the mentor is also associated with an increased rate of abstract conversion.⁴ Senior academic mentors can play a pivotal role in nurturing future physician-scientists in helping them achieve their research aspirations.^9,10

The primary objective of this study was to assess the conversion rate of resident research abstracts into peer-reviewed publications in a Canadian Surgical Residency Program. A secondary aim of this study was to identify variables associated with the positive conversion of abstracts.

Methods

We performed a retrospective cohort study by downloading Resident Research Day program booklets from 2008-2020, from the publicly accessible University of Saskatchewan Department of Surgery website. This annual symposium allows surgical residents from various disciplines to present their research completed during their clinical training. We identified one hundred and seventeen abstracts with resident presenters for this study and excluded abstracts presented by faculty or medical students. The median time to publication was 12 months, with a range of 11-20 months; therefore, we excluded the abstracts from 2019 and 2020 to allow for at least 20 months of follow-up from the presentation date for each abstract.

Variables that could influence the publication of abstracts were postulated based on the literature and included resident-specific, mentor-specific, and project-specific variables.^4,7 The thirteen variables included the following: post-graduate year of training (PGY), whether the resident won an award at the symposium, the total number of residents listed as co-authors of the abstract, the presenting resident's surgical subspecialty, whether the resident was engaged in the Clinical Investigator Program (CIP), resident's prior number of publications, faculty mentor's number of years in practice, faculty mentor's number of previous publications, whether faculty mentor was a basic scientist, whether the project was financed by a major grant, type of research, type of data collection, and whether the project was multidisciplinary.

The primary outcome variable was successful conversion to a published article in a peer-reviewed journal. Secondary outcomes included time to publication, the journal’s impact factor, and whether the abstract was presented at a national conference.

We used PubMed and Google Scholar, and a specific search strategy to verify whether each abstract was published or not. First, we searched for the primary author of the abstract. Next, we searched for the primary investigator (faculty mentor) and reviewed their list of publications. If we could not identify a publication with the first two searches, then we searched for key words in the title of the abstract. Once we identified a publication, we ensured that the listed authors and methods matched the corresponding abstract. Before concluding that an abstract was not published, we required that at least five searches done in this manner had failed to find a result.

The collected data were entered into SPSS ver 26 (IBM Corporation, Armonk, NY) for statistical analysis, and Chi-squared testing identified 5 variables with a p-value <0.20. These variables, which included resident had prior publications, project involved multiple resident researchers, resident presentation won an award, faculty mentor had >25 publications, and the project was multidisciplinary, were then entered into a ‘backward’ logistic regression analysis.

Our study was exempt from Research Ethics Board (REB) approval as the Research Day programs are readily available to the public on our department website.

Results

We analyzed 99 resident abstracts from 2008-2018 and found publications for 52 (53%) in peer-reviewed journals with a median impact factor of 2.54. We also found evidence of presentation at a national conference for 43% of the abstracts.

Regarding resident-specific variables (Table 1), residents having prior publications were more likely to publish their findings (27/51 vs. 14/48, p = 0.02). Likewise, projects involving multiple resident researchers reached sufficient significance to be included in the logistic regression analysis (24/51 vs 15/48, p = 0.13). However, there was no statistically significant association with CIP engagement or winning an award on publication. The PGY rank of the resident and the surgical subspecialty likewise was statistically nonsignificant in publication.

Table 1.

Resident-, mentor- and project-specific differences in publication for 99 abstracts presented at the Department of Surgery Annual Resident Research Day at the University of Saskatchewan 2008-2018.

Project Characteristic	Published (51)	Not Published (48)	p-value*
Resident Specific Variables
Post-Graduate Year
PGY1 PGY2 PGY3 PGY4 PGY5+	2 16 14 10 9	4 14 12 9 9	0.92
Surgical subspecialties of residents
Neurosurgery General Surgery Orthopedic Surgery Other Surgical Disciplines	16 24 10 1	17 19 10 2	0.67
Resident had prior publications	27	14	0.02
Project with multiple resident researchers	24	15	0.13
Resident presentation received an award	26	17	0.12
Resident enrolled in Clinical Investigator program	10	5	0.22
Mentor Specific Variables
Faculty mentor has been in practice for over 14 years	25	21	0.66
Faculty mentor has more than 25 publications	30	19	0.07
Supervisor is a basic scientist	11	5	0.13
Project Specific Variables
Type of Research
Retrospective Prospective Basic Science Other (Surveys or Observational Studies)	27 4 10 12	25 9 6 6	0.37
Multidisciplinary project	27	12	0.004
College of Medicine $30,000 Research Grant	4	1	0.39

^*Calculated using the Chi-squared test.

In terms of mentor-specific variables, we found that projects supervised by mentors with >25 publications were more likely to be published (30/51 vs 19/48, p=0.04). However, the years in practice or engagement in basic science research were both statistically nonsignificant in our study.

Regarding project-specific variables, multidisciplinary projects were more likely to be published (27/51 vs. 12/48, p = 0.04), while the type of research was not statistically significantly associated.

In constructing the logistic regression model, “multidisciplinary project” and “mentor is a basic scientist” could not both be included because of collinearity. We selected “multidisciplinary project” as it had a much more significant association with publication (p=0.004 vs. p = 0.13). Logistic regression analysis (Table 2) revealed multidisciplinary research (OR 4.46, CI 1.8-11.4, p = 0.002), projects involving multiple resident researchers (OR 2.56, CI 1.02-6.43, p = 0.045), and faculty supervisor having > 25 publications (OR 2.46, CI 1.03-5.88, p = 0.042) as being significantly associated with conversion to publication.

Table 2.

Multivariate logistic regression analysis of variables influencing publication of Resident Research Day projects from the University of Saskatchewan 2008-2018.

	Multivariate Logistic Regression
Variable	p-value	Odds Ratio	95% CI
Multidisciplinary project	0.002	4.5	1.8-11.4
Project with multiple resident researchers	0.045	2.6	1.1-6.4
Faculty mentor had > 25 publications	0.042	2.5	1.1-5.9
Resident presentation received an award	-	-	-
Resident had prior publications	-	-	-

Nagelkerke R²=0.237

χ²=19.392,5 df,p=0.002

95%CI: 95% Confidence Inter

Discussion

In this cohort, 53% of resident abstracts from the Annual Research Day were eventually published in a peer-reviewed journal. Although we are a medium-sized residency program without a particularly strong history in surgical research, the percentage of abstracts that went on to publication is nonetheless similar to that seen in other Canadian and American Surgical Residency Programs.^4,11

Through a logistic regression analysis model, we found that resident projects are more likely to be published when projects are multidisciplinary, when projects involve multiple resident researchers, and when projects are supervised by faculty mentors with greater than 25 previous publications. These are associations only and do not elucidate the mechanism whereby these variables increase the chance of publication.

Through a literature review, we did not encounter any studies documenting the role of multidisciplinary collaboration in the publication of resident research projects. In a centre such as ours with limited ongoing surgical research projects, the role of surgical residents and faculty collaborating with established researchers in other fields likely takes on greater importance.¹²

Another novel finding in our study was that multi-resident involvement in a project improves the odds of publication. This is in agreement with previous literature that identified co-authors' lack of time and interest as one of the main barriers to successful publication.^7,13 The involvement of multiple residents can help overcome time constraints and raise the enthusiasm for pursuing publication.^14–16

We included presentation at a national conference as a secondary outcome. This search identified that 43% of the abstracts were submitted at a national conference, all of which were eventually published, implying that high-quality projects worthy of acceptance at a national conference would subsequently go on to publication in a peer-reviewed journal.

The role of surgical mentors with substantial publication experience has been previously validated as a key variable in converting abstracts into peer-reviewed publications.^4,7 The involvement of experienced co-authors in resident projects can guide mentees with research design and facilitate the development of publishable manuscripts. Our early analysis suggested that a resident with prior publications also increased the likelihood of publication, but this significance was lost after adjusting for co-variables in the logistic regression model. The most likely explanation is that residents with prior publications often engaged in research with the same faculty members. The role of experienced faculty in providing invaluable mentorship has been documented before, and our results agree with the literature.^4,9,10

Our study has several limitations owing to the retrospective design of this project. The only resident projects considered in this study were those accepted at the Annual Resident Research Day, which excludes research projects not presented at this symposium. Also, there are likely other variables that are not well-documented nor easy to verify that might influence the publication of these resident projects, such as supervisors' roles and responsibilities, protected time for resident research, number of rejections by peer-reviewed publications, and others.

Another potential limitation of this analysis is that our search strategy was based on PubMed and Google Scholar. Both engines index publications in journals that meet the criteria for indexing, which exclude non-peer-reviewed journals. Given the time constraints of a surgical residency, it is possible that residents chose to publish their findings in the most likely to accept journal. Finally, the variables identified in our study might not be generalizable to other surgical residency programs and may also not apply to non-surgical programs. Over the 10 years of these annual symposiums, there was also a change in the administration of all the surgical programs as well as the department of surgery. These changes led to a greater emphasis on resident research and a greater volume of resident presentations for the annual symposium; however, we did not identify any significant changes in the percentage of resident publications resulting from these administrative changes.

Conclusions

The results of this study suggest that abstracts presented by surgical residents at annual research days at our center have about a 53% publication rate. Based on our study, resident projects are more likely to be published when there is collaboration between multiple disciplines, when projects involve more than one resident, and when projects are supervised by faculty mentors with greater than 25 previous publications. These variables should be considered as ideas to help boost research productivity and research acumen amongst residents. Further research is warranted into this topic to dissect the mechanisms whereby the identified variables lead to an increased chance of publication. Future research might also include a prospective trial to assess what changes are meaningful in increasing resident research productivity.

Funding Statement

Funding: There was no financial or grant support provided for this project.

Conflicts of Interest

The authors have no relevant conflicts of interest to declare.