Trends in patient and public involvement in health research 2000–2024

Abstract

Objectives

Public involvement in health research has been defined as research carried out ‘with’ or ‘by’ members of the public, rather than ‘to’, ‘about’, or ‘for’ them. We examined trends in patient and public involvement (PPI) prevalence in health research based on reporting in original research articles published in ten general medical journals over 25 years (2000–2024).

Study design and setting

For this bibliometric analysis, OVID Medline from 1 January 2000 to 31 December 2024 was used to identify a random sample of 1,000 individual studies, with 4 articles per journal per year selected from 10 general medical journals (5 adult and 5 pediatric) published over the 25 year-period. All study designs were included. We defined adult studies as those with participants 19 years or older and pediatric studies as those with participants 18 years or younger. Four reviewers independently classified studies as PPI + or PPI- and extracted data on primary study characteristics and on PPI methods for PPI+ studies. Log-linear regression was used to analyze trends over time.

Results

Of 1000 articles, 97/1000 reported PPI (prevalence 10%, 95% CI: 8%-12%). PPI was more prevalent in adult than pediatric journals (13% vs. 7%; P = 0.002). Log-linear regression showed a significant increasing trend in reporting PPI over time (from 0% in 2000 to 38% in 2024) representing an average annual percentage change (APC) of 15.0% (95% CI: 9.6% – 20.6%). Percent PPI+ articles significantly increased over time in adult journals (APC, 22.2%; 95% CI: 16.5% – 28.2%) but not in pediatric journals (APC, 3.0%; 95% CI: -2.1% – 8.4%). By journal, The BMJ (journal with mandatory PPI reporting) had the highest PPI prevalence (24%). By country, PPI was most common in the United States (32%). PPI+ studies were more likely to be an experimental design, conducted in a community setting, and less likely to be multi-country, compared to PPI- studies.

Conclusions

Over 25 years, PPI reporting in original health research articles published in general medical journals increased on average, 15% per year. PPI reporting increased significantly in adult journals (on average, 22% per year) but not in pediatric journals. Level of PPI increased over time.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40900-026-00940-2.

Plain language summary

Patient and public involvement (PPI) in health research refers to researchers working together with patients and other members of the public on any or all aspects of a research study. For our study, we looked at how often patients and members of the public were involved in health research over time. We did this by reviewing original research publications of all study design types published in 10 general medical journals over 25 years, from 2000 to 2024. We searched a large online library of medical research articles to identify a random sample of 1,000 individual studies, with 4 articles per journal per year selected from 10 major general medical journals (5 adult and 5 pediatric) over the 25 year-period. Four reviewers classified trials as PPI + or PPI- and collected data on variables of interest. We found that 10% (97/1000) of published articles (13% in adult journals and 7% in pediatric journals) reported PPI across all types of health research over 25 years. Reporting of PPI increased over time from 0% in 2000 to 38% in 2024, an average increase of 15% per year. An increase in reporting of PPI was shown in adult journals (on average, 22% per year) but not in pediatric journals.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40900-026-00940-2.

Introduction

National strategies, initiatives, and infrastructure to support and promote patient and public involvement (PPI) in health research have been implemented in countries around the world [1]. Examples include the Strategy for Patient-Oriented Research (SPOR) in Canada, launched in 2011 by the Canadian Institutes of Health Research (CIHR), and the Patient Centered Outcomes Research Institute (PCORI) in the USA, launched by an Act of Congress in 2010 (https://cihr-irsc.gc.ca/e/41204.html) [2–4]. In the UK, the Center for Engagement and Dissemination (2020–2025), was built on decades of support for PPI in health and social care research across the National Institute for Health and Care Research (NIHR) enterprise [5, 6].

There are a variety of overlapping definitions for PPI in health research. CIHR defines patient-oriented research as “engaging patients, their caregivers, and families as partners in the research process” (https://cihr-irsc.gc.ca/e/41204.html) [3]. Likewise, the NIHR defines public involvement in health research as “research carried out ‘with’ or ‘by’ members of the public, rather than ‘to’, ‘about’, or ‘for’ them” (https://www.nihr.ac.uk/get-involved/public-involvement) [7]. Involvement of patients and the public as partners in health research has a long history and varied terminology, e.g., integrated knowledge translation, co-production, and community-based participatory research, often with overlapping principles and methods [8].

Systematic, scoping, and umbrella reviews have repeatedly shown that involvement of patients and the public in health research can lead to a greater focus on patient-identified priorities, improved research feasibility and quality, enhanced dissemination of results, and ultimately, improved health and well-being outcomes [9–12]. For example, Borkhoff et al. showed that PPI+ trials (those with patient or public involvement at any stage of the research process) published in the BMJ had increased dissemination metrics (particularly through non-academic channels), compared with PPI- trials [13].

Several systematic reviews have examined the reported prevalence of PPI in randomised controlled trials (RCTs) published in peer-reviewed journals, with prevalence estimates ranging from < 1% to 6% [14–19]. A cross-sectional survey of all health research articles published in 2020 in BMJ Open – a journal with mandatory reporting of PPI in submitted articles – showed a PPI prevalence of 21% [20].

There are relatively few studies, however, on trends in PPI in health research. To our knowledge, only two studies have specifically examined trends in PPI in health research [21, 22]. Boote et al. examined public involvement in health research in the published literature over the period 1995-2009 [21]. This bibliometric analysis examined individual studies included in four reviews (three systematic and one narrative) and two bibliographies, that examined public involvement in research. In total, 683 individual publications were identified. The published results included a graph showing the number of studies with public involvement in health research per year between 1995 and 2009. The frequency of publications with public involvement in research steadily increased over the time period, however, there was no denominator identified and no formal analysis of trends.

Vanneste et al. examined PPI in randomised controlled trials (and their protocols) published between 2015 and 2023 in four major general medical journals - The BMJ, JAMA, New England Journal of Medicine, and Lancet [22]. The first ten RCTs published by each journal, by year, constituted the sampling frame. Over the nine-year period, 360 peer-reviewed publications of RCTs (and 299 corresponding protocols) were identified and reviewed. The prevalence of PPI in published RCTs was 64/360 (18%). Logistic regression showed a significantly increasing trend in reported PPI in published RCTs over time.

The perceived benefits of PPI in health research have been well studied, however, understanding the evolution of PPI reporting over time is a research gap. To our knowledge, no study has examined PPI prevalence over time across a broad representation of study designs. Another key limitation of prior work, e.g., Boote et al., is the absence of a clear denominator for estimating the proportion of studies reporting PPI [21]. Last, no study has compared PPI trends in ‘adult’ and ‘paediatric’ studies. This is considered important, as there are unique challenges with PPI in child health research such as determining who provides consent, whether the limited life experience of children limits their research contributions, and the clear and obvious power differential between researchers and children [23].

The primary objective of this study was to determine the prevalence and trends in PPI in health research (specifically in adult versus pediatric journals) – based on a bibliometric analysis of peer-reviewed, empirical research publications of all study design types – over 25 years, from 2000 to 2024 in 10 general medical journals. A secondary objective was to examine the trend over time in level of PPI involvement among articles reporting PPI over the 25-year period.

Methods

This study followed the Preferred Reporting Items for Bibliometric Analysis (PRIBA) checklist to ensure consistent reporting of key components for a bibliometric analysis [24]. We did not register a protocol prior to conducting the review.

Search strategy and sampling frame

For this bibliometric analysis, adult and pediatric original research articles published over the 25-year period from 1 January 2000 to 31 December 2024 in 10 general medical journals were identified using an electronic search strategy. We developed an OVID Medline search strategy in conjunction with an information science specialist (JC). We selected 10 high-impact general medical journals: 5 “adult” (The BMJ, Lancet, New England Journal of Medicine, JAMA, and CMAJ) and 5 “pediatric” (Pediatrics, JAMA Pediatrics, Archives of Disease in Childhood, Journal of Pediatrics, and BMC Pediatrics). We defined adult original research articles as those with participants 19 years or older published in an “adult” journal and pediatric original research articles as those with participants 18 years or younger published in a “pediatric” journal. We excluded articles from the adult journals if the study population involved only children or both adults and children. The main inclusion criterion was original research articles reporting empirical data. All study designs were included, e.g., experimental studies (RCT, quasi experimental); observational studies (cohort, case control); descriptive studies (cross-sectional surveys); and qualitative research. We excluded all reviews (scoping, systematic, narrative, umbrella), meta-analyses, commentaries, editorials, letters to the editor, opinion pieces, news, and study protocols. The search strategy, as applied to one journal (JAMA Pediatrics) for one year (2024), is shown in Supplementary Table 1.

For each journal and year (e.g. JAMA Pediatrics, 2024), 4 original research articles were selected at random using the Excel RANDBETWEEN function which returned a random integer between one and the number of eligible studies per year. The study corresponding to this number in the Excel list of eligible studies for that journal and year was selected. This sampling strategy provided a balanced representation of original research across journals. The study selection process is summarized using a flow diagram as shown in Fig. 1. Full texts for selected studies were retrieved (as well as independently published study protocols, when available). Each individual study selected was reviewed to confirm eligibility, and data were collected on variables of interest (see below). In summary, we reviewed in total, a random sample of 1,000 individual studies, with 4 original research articles per journal per year selected from the 10 general medical journals over the 25 year-period, 2000 to 2024.

Fig. 1.

Preferred reporting items for systematic reviews and meta-analyses flow diagram

In summary, a pragmatic sampling strategy – random selection of four original research articles (all types of health research study designs), per journal, per year - was conducted to examine PPI prevalence across a balanced representation of study designs, across journals, and time. This sampling strategy provided a clear denominator for estimating the proportion of studies reporting PPI.

Data abstraction

A primary reviewer (SC or SA or NH or RS) independently collected data on the variables of interest and completed the data abstraction form for each original research article using the full text, study protocol (when available), and study registration data (when available). Studies were classified as PPI + or PPI-. PPI+ studies were defined as those with patient or public involvement at any stage of the research process. All other studies were defined as PPI-. Data collected on descriptive parameters included: first author, manuscript title, journal, and year of publication. Data on study characteristics included: adult or pediatric study, study setting (primary care, hospital, community, population), study design (experimental, observational, qualitative), country in which the study was conducted, multinational (yes or no), research topic (medical, surgical, allied health, mental health, public health), and sample size.

To ensure accurate and consistent data abstraction, the four reviewers met regularly to resolve any issues. We applied a serial duplicate review process on 10% of articles (100 articles in the sampling frame); after a primary reviewer completed the data abstraction form, a second reviewer (SC or SA or NH or RS) checked each item on the form for completeness and accuracy. Any discrepancies between the primary and secondary reviewers were resolved through discussion and adjudication with a third reviewer (CMB).

Iterative team discussions and consensus-building were conducted to determine how to define PPI+ articles. Meetings focused on reviewing existing language, frameworks, and examples to reconcile differing interpretations and agree on inclusion criteria that balanced sensitivity to diverse forms of involvement while ensuring consistency and reproducibility. This process acknowledged that PPI is not a binary construct and that reporting PPI varies widely in depth and clarity. For an article to qualify as PPI+, the main article, protocol, or registration data must have indicated PPI at one or more stages of the research process. Specifically, there had to be an explicit statement that patients or members of the public were involved as partners in the research process.

For PPI+ studies, we abstracted a verbatim text description of PPI from the original research article as well as information on Who Was Involved? (1) patient partners, (2) caregivers or family members, (3) community partners, (4) patient or public organizations, (5) end-users or clinical staff; How Were They Involved? (1) leadership role (Co-PI or Lead PI determined through authorship or summary of contributions), (2) membership role (research team member providing input on study processes), (3) consultation role (focus groups, pilot testing, structured or informal interviews, surveys, or informal consultations); What Aspects Were They Involved In? (1) choosing/developing research question(s) to ensure research relevance, (2) study protocol development (including REB submission), (3) grant application to funding agency, (4) selection of primary or secondary outcomes, (5) intervention design and/or delivery, (6) recruitment and/or retention strategies, (7) patient-facing materials design and/or development (e.g., consent forms, questionnaires), (8) data collection, (9) data analysis, (10) interpretation of findings, (11) review or editing of manuscript, reports or lay summaries, (12) co-authorship of manuscript, reports or lay summaries, (13) dissemination of findings (e.g., presentations, social media promotion, conferences), (14) evaluation of PPI impact in the primary study; When Were They Involved? (1) study conception, (2) study preparation, (3) study execution, (4) data analysis, (5) reporting of findings, (6) > 1 phase, (7) throughout study: phases 1–5, (8) undefined; and, Recognition of patient partner(s) in published manuscript (1) no, (2) main text only, (3) acknowledgements section, (4) co-author (individual), (5) co-author (group).

Level of engagement was based on the International Association for Public Participation (IAP2) Spectrum: 0 = Inform, 1 = Consult, 2 = Involve, 3 = Collaborate, 4 = Empower. IAP2 is a widely used framework to define level of PPI in research; each PPI+ study was scored on a scale of 1 to 4, with 4 or empower being the highest level [25]. Our definition of PPI in research only considered engagements with two way communication between research staff and patient partners. Therefore, the Inform level of participation in the IAP2 Spectrum was not included in our analysis.

To ensure accuracy in data abstraction for the additional variables for PPI+ studies, we applied a serial review process whereby a primary reviewer (SA or NH) completed the data abstraction form and a second reviewer (SA or NH) checked each item on the form for completeness and accuracy. Any discrepancy between the primary reviewer and secondary reviewer was resolved through discussion and adjudication with a third reviewer (CMB). Record verification used the King & Lashley method whereby the second reviewer verified the first 10 studies for which a primary reviewer completed data abstraction [26]. If no inconsistency was found, then the second reviewer continued the verification by checking every tenth record. If an entry error was found, after correcting the inconsistency, every record was checked until 10 consecutive error-free records were reviewed, after which checking every tenth record was resumed.

Statistical analysis

Descriptive statistics were used to describe the prevalence of PPI, and other study characteristics. Medians and interquartile ranges (IQRs) were used to describe continuous variables that were non-normally distributed. PPI + and PPI- studies were compared using the chi-square test (or Fisher’s exact test as appropriate) for categorical variables and the Mann-Whitney test for continuous variables.

To examine temporal trends in counts and proportions of PPI in adult and pediatric original research articles, we used descriptive analyses to first evaluate the proportion of studies reporting PPI between 2000 and 2024, with the year of publication as the independent variable and proportion (or percent) of studies reporting PPI as the dependent variable. An initial visualization of the data using scatter plots and residual plots revealed a non-linear trend. Trends were assessed using log-linear regression, modeling the natural logarithm of the annual proportion (or percent) of studies reporting PPI as a function of the year of publication to estimate the average annual percent change (APC) and corresponding 95% confidence intervals (CI) to determine statistically significant trends over time overall and by journal type (adult, paediatric). Statistical significance was inferred when the CI excluded zero. In describing trends, the term ‘increase’ was used when the slope (or APC) was positive and statistically significant. To test for a significant difference by journal type, an interaction term was included to determine if the slopes of the lines were different. A sensitivity analysis was performed to evaluate the robustness of our findings by restricting the analysis to 9 of 10 general medical journals, i.e., data from The BMJ were excluded as this was the only journal of the ten with mandatory PPI reporting. A secondary analysis, based on the mean IAP2 score for all PPI+ articles for each year, used simple linear regression to examine the trend over time in level of patient and public involvement in the PPI+ articles across the 25-year period. Statistical significance was defined as P < 0.05; statistical tests were 2-sided. All analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Patient and public involvement

A patient partner (FB, co-author), a key member of our research team was involved in every aspect of this bibliometric analysis, including contributing to the development of the research protocol, refining the research question, and providing input on study design, outcome measures, and analysis, interpreting the findings, writing the manuscript and preparing dissemination plans.

Results

The 10 high-impact general medical journals (5 adult and 5 pediatric) published 16,839 original research articles over the 25-year period, 2000 to 2024. Figure 1 describes the screening process. Four original research articles per journal per year were randomly selected, generating a sample of 1,000 individual studies published between 2000 and 2024. Of the randomly selected studies, 26 were excluded as ineligible (study protocols).

Over the 25-year period from 2000 to 2024, 97 of 1000 articles reported PPI (prevalence 10%, 95% CI: 8%-12%). Supplementary Table 2 lists the citations for the 97 PPI+ studies. Adult journals had a higher PPI+ prevalence compared to pediatric journals (63/500, 13% vs. 34/500, 7%; P = 0.002). Of the 10 high-impact general medical journals, The BMJ accounted for the majority of PPI+ articles and Journal of Pediatrics accounted for the least: The BMJ (n = 24), JAMA (n = 13), Pediatrics (n = 12), The Lancet (n = 11), BMC Pediatrics (n = 9), CMAJ (n = 8), JAMA Pediatrics (n = 7), New England Journal of Medicine (n = 7), Archives of Disease in Childhood (n = 6), and Journal of Pediatrics (n = 0).

Characteristics of PPI + and PPI- studies are described in Table 1. There were differences between PPI + and PPI- studies in journal type (P = 0.002), study setting (P < 0.001), study design (P < 0.001), and multinational status (P = 0.003) but not research topic (P = 0.08) or sample size (P = 0.05). Comparing PPI+ studies with PPI- studies, a higher proportion were in adult journals (63/97, 65%, compared with 437/903, 48%, respectively); more frequently conducted in a community setting (37/97, 38% vs. 133/903, 15%, respectively), and had an experimental study design (69/97, 71%, in particular cluster RCTs, compared with 493/903, 55%, respectively). PPI+ studies were also less likely than PPI- studies to be a multinational study (4/97, 4% vs. 136/903, 15%, respectively).

Table 1.

Primary characteristics of PPI + and PPI- health research studies 2000–2024 (n = 1000)

Characteristics	Total	PPI +	PPI -	P value
Number of studies	1000	97	903
Journal type, n (%)
Adult	500 (50.0)	63 (65.0)	437 (48.4)	0.002
Pediatric	500 (50.0)	34 (35.0)	466 (51.6)
Study setting, n (%)
Community	170 (17.0)	37 (38.1)	133 (14.7)	< 0.001
Hospital (in-patient, clinic, ED)	568 (56.8)	31 (32.0)	537 (59.5)
Primary Care	114 (11.4)	17 (17.5)	97 (10.7)
Population	148 (14.8)	12 (12.4)	136 (15.1)
Study design, n (%)
Experimental	562 (56.2)	69 (71.1)	493 (54.6)	< 0.001
Observational	407 (40.7)	18 (18.6)	389 (43.1)
Qualitative	31 (3.1)	10 (10.3)	21 (2.3)
Multinational, yes, n (%)	140 (14.0)	4 (4.1)	136 (15.1)	0.003
Research topic, n (%)
Medical	711 (71.1)	58 (59.8)	653 (72.3)	0.08
Public health	136 (13.6)	20 (20.6)	116 (12.9)
Allied health	52 (5.2)	8 (8.3)	44 (4.9)
Mental health	54 (5.4)	7 (7.2)	47 (5.2)
Surgical	47 (4.7)	4 (4.1)	43 (4.8)
Sample size, median (IQR)	445 (133, 2189)	632 (219, 2590)	429 (127, 2179)	0.05

Log-linear regression showed a significantly increasing trend over time in PPI+ studies in published health research articles. Figure 2 shows the trend over time in percent PPI+ articles for all 10 journals. Forty articles were sampled each year from 2000 to 2024. Overall, the percent PPI+ articles increased significantly from 0% (0/40) in 2000 to 38% (15/40) in 2024 representing an annual percentage change (APC) of 15.0% (95% CI: 9.6% – 20.6%). The highest levels of PPI reporting were observed in the last three years of the study (2021 to 2024). Figure 3 compares the trend in PPI+ articles in the 5 adult versus 5 pediatric journals over the 25-year period. Using log-linear regression, the difference in slopes between adult and pediatric journals was statistically significant (P < 0.001), suggesting that the overall increase was because of the large increase in PPI+ articles in adult journals over time. In other words, over the 25-year period from 2000 to 2024, the percent PPI+ articles significantly increased in adult journals (from 0% (0/20) in 2000 to 65% (13/20) in 2024; APC, 22.2%; 95% CI: 16.5% – 28.2%) but not in pediatric journals (from 0% (0/20) in 2000 to 10% (2/20) in 2024; APC, 3.0%; 95% CI: -2.1% – 8.4%).

Fig. 2.

Trend over time in percent PPI+ articles, 2000–2024

Fig. 3.

Trend over time in percent PPI+ articles in adult versus pediatric journals, 2000–2024

Of note, The BMJ introduced mandatory PPI reporting for authors in 2015 and this single journal accounted for 25% (24/97) of all PPI+ articles in our study. The other nine general medical journals included in our review do not have mandatory PPI reporting. Almost all (22/24) of the PPI+ articles published in The BMJ were published in 2015 or later. A sensitivity analysis involving data from the other nine general medical journals, i.e., after removing articles published in The BMJ, showed similar findings for the overall trend in PPI+ prevalence (APC, 16.1%; 95% CI: 5.1% – 28.2%; P < 0.001) (Supplementary Fig. 1a). Likewise, the trend in PPI+ studies for the four adult journals (APC, 40.7%; 95% CI: 25.1% – 58.2%; P < 0.001) was similar to that for the five adult journals (Supplementary Figs. 1b). Last, while overall PPI+ prevalence for adult and pediatric journals was no longer statistically significantly different (39/400, 10% vs. 34/500, 7%, respectively; P = 0.11), the significant difference in PPI+ trend slopes between the (four) adult and (five) pediatric journals persisted (P < 0.001).

In total, 97 of 1000 (10%) studies reported PPI in the research process. Characteristics of patient and public involvement in these 97 studies are summarized in Table 2 and verbatim text excerpts from the published articles are included in Supplementary Table 3. Patient partners were the most frequently involved stakeholder in the PPI+ studies (n = 54; 55.7%). Most individuals or organizations involved in the research process had either a membership or consultation role; in 9 of 97 (9.3%) they had a leadership role. More than one third of studies (n = 35; 36.1%) described a single PPI activity; one study described 10 of 14 PPI activities [27]. The most common study activity in PPI+ health research was input in intervention design and/or delivery (n = 56; 57.7%), followed by feedback on patient-facing materials design and/or development (n = 44; 45.4%), and participation in study protocol development (n = 42; 43.3%). In almost half of the 97 studies (n = 45; 46.4%), PPI was throughout all 5 phases of the study, from study conception to the reporting of findings. Patient partners were often recognized in the acknowledgments section of the publication (n = 48; 49.5%). Though PPI recognition through co-authorship on the published manuscript was rare, individuals were occasionally named or recognized through group authorship for the patient or public organization. Co-authorship (n = 9; 9.3%) corresponded with those studies where individuals or organizations had a leadership role and level of engagement was at the highest level on the IAP2 spectrum (Empower).

Table 2.

Characteristics of patient and public involvement in 97 studies

Information regarding patient and public involvement	No. of Studies
	N = 97
Who Was Involved?†
Patient partners	54 (55.7)
Caregivers or family members	20 (20.6)
Community partners	18 (18.6)
Patient or public organizations	29 (29.9)
End-users, clinical staff	34 (35.1)
How Were They Involved?
Leadership role	9 (9.3)
Membership role	54 (55.7)
Consultation role	34 (35.1)
What Aspects Were They Involved In?†
Choosing/developing research question(s) to ensure relevant research	17 (17.5)
Study protocol development (including REB submission)	42 (43.3)
Grant application to funding agency	11 (11.3)
Selection of primary or secondary outcomes	15 (15.5)
Intervention design and/or delivery	56 (57.7)
Recruitment and/or retention strategies	23 (23.7)
Patient-facing materials design and/or development	44 (45.4)
Data collection	8 (8.2)
Data analysis	2 (2.1)
Interpretation of findings	17 (17.5)
Review or editing of manuscript, reports or lay summaries	16 (16.5)
Co-authorship of manuscript, reports or lay summaries	7 (7.2)
Dissemination of findings (e.g., social media promotion, conferences)	13 (13.4)
Evaluation of PPI impact in the primary study	1 (1.0)
When Were They Involved?
Phase 1 – study conception	9 (9.3)
Phase 2 – study preparation	18 (18.6)
Phase 3 – study execution	13 (13.4)
Phase 4 – data analysis	0 (0.0)
Phase 5 – reporting of findings	4 (4.1)
Throughout study: phases 1–5	45 (46.4)
>1 phase	7 (7.2)
Undefined	2 (2.1)
Recognition of Patient Partner(s) in published manuscript
No	7 (7.2)
in main text only	33 (34.0)
in Acknowledgements section	48 (49.5)
co-author (named)	6 (6.2)
co-author (group)	3 (3.1)
Level of Engagement based on IAP2 Spectrum
Consult	26 (26.8)
Involve	27 (27.8)
Collaborate	35 (36.1)
Empower	9 (9.3)

†more than one option per study possible

Figure 4 shows the trend over time in level of patient and public involvement in PPI+ articles across the 25-year period, by IAP2 category. As shown in the bar graph, PPI+ studies moved slowly over time from consultation to involvement and collaboration. The first indication of empowering patient partners to lead some or all aspects of the research process appeared in 2016. Over the last three years of the study (2021 to 2024), PPI+ articles had the highest levels of engagement by IAP2 category. There was a significant increase in mean IAP2 score over time from 1.3 in 2001 to 2.6 in 2024 (R² = 0.29, F(1,20) = 8.27, P = 0.009).

Fig. 4.

Trend over time in IAP2 level of patient and public involvement, 2000–2024

Discussion

This bibliometric analysis examined the prevalence, trends, and nature of PPI reporting in published health research articles in ten major general medical journals over 25 years. The prevalence of PPI+ articles increased significantly from 0% in 2000 to 38% in 2024. Analysis of trends over time showed, on average, a 15% increase in PPI+ articles, year over year. PPI reporting increased significantly in adult journals (on average, 22% per year) but not in pediatric journals. PPI+ studies were more likely to have an experimental design, have a community setting, and be published in an adult journal. With regards to the nature of PPI, among PPI+ studies, mean IAP2 scores increased over the 25 years, suggesting deeper levels of engagement of patients and the public in health research over time.

To our knowledge, this is the first long-term bibliometric analysis of trends in PPI across all types of primary research studies in the adult and pediatric literature. Strengths of our study include random sampling of original research articles from a purposive sample of journals (thus ensuring generalizability while maximizing relevance), inclusion of all study designs, a large sample size, a long period (25 years), comparison of trends in the adult and paediatric published literature, and a sampling strategy (four articles per journal per year) to achieve balanced representation across journals and time from ten high-impact general medical journals. There are some study limitations. First, reporting of PPI was neither standardized nor consistent which may have led to the misclassification of PPI+ studies as PPI- studies. Second, data on PPI involvement (type of activity and level) was limited to review of the published information only as we did not contact authors. Third, there is the potential for bias in our sampling strategy, if the likelihood of PPI+ prevalence is related to journal “size,” given the overrepresentation of articles from journals with fewer articles published per year. Last, reviewers were not blinded to the type of study nor the journal in which it was published.

A recent publication by Vanneste et al., who examined PPI in randomised controlled trials published between 2015 and 2023 in four major general medical journals, provides context for our findings [22]. Of 360 RCTs reviewed, Vanneste et al. reported a PPI prevalence of 18%, with logistic regression analysis showing a significantly increasing trend over time. Of note, a cross-sectional survey by Lang et al. found a PPI prevalence of 21% in 3000 research papers published by the journal BMJ Open in 2020 [20]. Vanneste et al. restricted their review to RCTs only (experimental), whereas our study considered all study designs (qualitative to experimental). While Lang et al. included a broader range of studies, their review was based on articles published in 2020, and as we have demonstrated, there has been an increasing trend in reporting PPI over time. This may explain the higher PPI prevalence estimate in the Vanneste et al. and Lang et al. studies, compared to our PPI prevalence estimate. In other words, the lower overall PPI prevalence in our study may be explained by the broader range of study designs reviewed and the longer time frame. In our study, experimental study designs were more likely to be PPI+ (69/562, 12.2%) compared with observational study designs (18/407, 4.4%). One possible explanation is that information on PPI in experimental studies (as opposed to observational studies) may be more likely to be captured as such information may be contained in study registrations, protocols, and/or published articles.

The findings from our study reinforce those of Vanneste et al., [22] showing a steady increase in PPI in published health research articles over the 25-year period, 2000–2024. Several factors may explain this increasing trend in PPI. First, the impact of numerous national initiatives and strategies to support and promote PPI in health research, such as SPOR in Canada, PCORI in the US, and the NIHR in the UK which funded the Center for Engagement and Dissemination may be important drivers [3–5]. In addition, PPI is often considered a moral obligation for researchers and several systematic and umbrella reviews have established the methodological benefits of PPI in health research, such as more meaningful research questions, feasible interventions, increased recruitment and retention, and increased uptake of findings [9–13]. Furthermore, other initiatives such as the Guidance for Reporting Involvement of Patients and Public (GRIPP2) checklists [28] may have contributed to the observed trends, including improvement in the quality of reporting. Of note, however, in previous studies by our group examining the prevalence of PPI in RCTs, we found that none of the RCTs identified had used a reporting tool to describe PPI [13, 19]. Last, mandatory reporting of PPI in health research protocols and submitted papers is increasingly requested by research funding agencies and academic journals. In this context, Price et al. [29] conducted a ‘before-and-after’ study of PPI in BMJ papers published before (June 2013-May 2014) and after (June 2015-May 2016) the introduction of a journal policy requiring authors to report on PPI. Before the policy introduction, 1/89 (0.5%) research papers published in the journal reported PPI, compared with 16/152 (11%) after the policy was introduced. Likewise, Husson et al. [30] conducted a MEDLINE search of chronic illness RCTs in the BMJ group of journals with mandatory PPI reporting and noted that 40/101 (40%) RCTs reported PPI; a prevalence higher than other systematic reviews of PPI in RCTs [14–19].

While the evidence suggests that mandatory PPI reporting by journals increases PPI prevalence in health research articles (or increases submission of PPI papers to these journals), PPI prevalence estimates may be underestimated for journals without mandatory reporting, whether because of word limitations or author choices. Nonetheless, the size and scale of the increasing trend in PPI in health research articles in our study - across a range of adult and paediatric journals (of which only one had mandatory reporting) - suggest that the increasing PPI trend is real.

An equally important question, however, is the “quality” of PPI, in other words, whether the increasing PPI trend is associated with less tokenism. The changing levels of IAP2 engagement over time towards collaboration and empowerment of patient partners suggests higher quality PPI. Of note, our study was limited to reviewing published articles and dichotomizing studies as PPI + or PPI-. Lang et al. have noted that this approach does not comprehensively capture the quality and extent of engagement, the attitudes and beliefs of researchers, or the policies of institutions [20]. Evidence derived from patient partner interviews and formal qualitative analyses would help confirm if the shifts in research culture and PPI practice were authentic and meaningful.

The results of our study also indicate that there may be unique barriers to engaging children, youth, and caregivers in health research and further research on enablers and barriers to engagement in this specific population is warranted [31, 32]. Our data suggest that PPI prevalence is increasing more quickly in adult research, compared to paediatric research. Possible explanations include challenges unique to PPI in child health research (competing priorities for families, scheduling challenges, educational commitments) compared to adult research, differential reporting of PPI in adult and paediatric research articles, or whether institutional support for PPI is growing more quickly in adult research compared to child health research.

Last, our findings may be used to inform the development of policies, initiatives, and training/educational programs by funding agencies, institutions, journals, research teams, and individual researchers to promote and support PPI in health research. Potential recommendations to promote PPI and ensure an accurate assessment of PPI prevalence and trends would be for funding agencies to mandate that research grant applications include a description of the nature and type of PPI proposed, that journals follow the lead of the BMJ and other journals to mandate a PPI paragraph in all original research papers (ideally using a formal PPI reporting checklist such as GRIPP2), and last, that institutions provide the relevant infrastructure and support for researchers and patients interested in authentic and meaningful PPI in health research.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

PPI

Patient and public involvement

APC

Annual percentage change

SPOR

Strategy for Patient-Oriented Research

CIHR

Canadian Institutes of Health Research

PCORI

Patient Centered Outcomes Research Institute

NIHR

National Institute for Health and Care Research

RCT

Randomised controlled trial

PRIBA

Preferred Reporting Items for Bibliometric Analysis

IAP2

International Association for Public Participation

IQR

Interquartile ranges

Author contributions

SC conceptualized and designed the study, designed the data collection instruments, coordinated the data extraction, interpreted the data, drafted the manuscript, critically reviewed the manuscript for important intellectual content. CMB conceptualized and designed the study, designed the data collection instruments, supervised the data extraction, performed the statistical analysis, interpreted the data, drafted the manuscript, critically reviewed and revised the manuscript. FB conceptualized and designed the study, designed the data collection instruments, interpreted the data, critically reviewed the manuscript for important intellectual content. SA, NH, and RS performed the data extraction, interpreted the data, and critically reviewed the manuscript for important intellectual content. JC developed the search strategy, performed data curation, interpreted the data, and critically reviewed the manuscript for important intellectual content. CM conceptualized and designed the study, designed the data collection instruments, supervised the data extraction, interpreted the data, drafted the manuscript, critically reviewed and revised the manuscript. All of the authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Data availability

All data relevant to the study are included in the article or uploaded as supplementary information. Extracted data are available on request to the corresponding author.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.