Walsh CM, Fung M, Ginsburg S. Publication of Results of Abstracts Presented at Medical Education Conferences. JAMA. 2013;310(21):2307-2309. doi:10.1001/jama.2013.281671
Conferences represent an important forum for presentation of scholarly activity; however, dissemination beyond meeting registrants is limited. Peer-reviewed publication facilitates knowledge translation,1 and failure to publish may lead to unnecessary duplication and publication bias that can compromise future scholarship.2 This study aimed to determine the rate and time course of peer-reviewed journal publication of abstracts presented at the 2 largest North American medical education conferences (Research in Medical Education Conference [RIME] and the Canadian Conference on Medical Education [CCME]) and to identify characteristics associated with publication.
All 455 abstracts from the 2005 and 2006 RIME and CCME conferences were reviewed. Six were excluded (4 withdrawn, 1 missing, 2 published before abstract deadline). Blinded to publication status, 2 investigators (C.M.W. and M.F.) extracted the following information: format (oral or poster), type of scholarship (research, evaluation of a program or innovation, or other), methods (quantitative, qualitative, or mixed), completion status (completed or work in progress), number of centers involved, and degrees and institutions of the authors.
MEDLINE, EMBASE, ERIC, and Google Scholar were searched (from inception through November 2012) using the names of the first, second, and last author, followed by keywords from the title, abstract, or both. Retrieved peer-reviewed publications were compared with the corresponding abstract to ensure it represented the same work. Two investigators recorded journal type (educational or other), funding source, publication type (full-length or abbreviated article), and publication date.
Associations between publication status and all other parameters were analyzed by simple logistic regression using SPSS version 16.0 (SPSS Inc). Multivariable analysis was conducted using all variables from bivariable screening with P < .10. Statistical significance was set at P < .05 (2-sided).
Of the 449 abstracts, 156 (34.7%) were subsequently published (RIME: 86/232; CCME: 70/217); 149 (95.5%) were full-length articles. Funding was acknowledged in 81 (51.9%). Median time to publication was 20 months (interquartile range, 10-30 months), ranging from 9 months preconference (but after the submission deadline) to 76 months postconference. More than 90% were published within 4 years (Figure). Publications appeared in 46 different journals and 73.1% were in educationally focused journals.
In bivariable analysis, publication was more likely for findings from abstracts describing research studies compared with other scholarly activities (38.9% vs 19.4%, respectively; odds ratio [OR], 2.64 [95% CI, 1.38-5.06]; P = .003), and those including a last author with a PhD (44.3% vs 31.6%; OR, 1.73 [95% CI, 1.00-2.97]; P = .049). In multivariable analysis, subsequent publication was more frequent for oral abstracts (48.5% vs 26.8% for poster abstracts; OR, 2.29 [95% CI, 1.45-3.61]; P < .001) and those outlining completed work (41.7% vs 22.0% for work in progress; OR, 1.99 [95% CI, 1.05-3.76]; P = .03) (Table). Quantitative studies were published less frequently than mixed-methods studies (35.8% vs 55.6%, respectively; OR, 0.41 [95% CI, 0.19-0.87]; P = .02) and least frequently overall. Multi-institutional authorship also predicted publication (50.8% vs 29.0% for single center; OR, 2.12 [95% CI, 1.29-3.48]; P = .003).
The subsequent publication rate of abstracts presented at medical education conferences was 34.7%, well below the mean (44.5%) reported for biomedical research.3 One reason may be that educational scholarship is conceptualized more broadly than traditional research4 and presentation at conferences may obviate the need for journal publication because it already constitutes dissemination by making innovations visible and available for others to build on.5 Many of the significant predictive variables, such as oral presentation format and completed works, may be surrogate markers for study quality. We were surprised that quantitative abstracts were least likely to be subsequently published because quantitative methods are most commonly used.
In terms of limitations, our results cannot be generalized beyond the 2 conferences included. Given the word constraints of abstracts, potential misclassification of studies may have occurred and other important predictors (eg, education level) may have been missed. Additionally, author degrees were not available for CCME and therefore could not be included in multivariable analysis. To move the field forward, it is important to develop a robust literature to document emerging innovations and elevate the relevance and value of educational scholarship. Future research is planned to determine barriers to publication to help mitigate them.
Corresponding Author: Shiphra Ginsburg, MD, MEd, FRCPC, Wilson Centre for Research in Education, 600 University Ave, Ste 433, Toronto, ON M5G 1X5, Canada (firstname.lastname@example.org).
Author Contributions: Dr Walsh had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Walsh, Ginsburg.
Acquisition of data: Walsh, Fung, Ginsburg.
Analysis and interpretation of data: Walsh, Ginsburg.
Drafting of the manuscript: Walsh.
Critical revision of the manuscript for important intellectual content: Walsh, Fung, Ginsburg.
Statistical analysis: Walsh.
Administrative, technical, or material support: Fung, Ginsburg.
Study supervision: Ginsburg.
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: Dr Walsh was funded by a Canadian Institute of Health Research Canadian Child Health Clinician Scientist Program doctoral fellowship and a Hospital for Sick Children, Department of Paediatrics fellowship grant.
Role of the Sponsor: The Canadian Institutes of Health Research and the Hospital for Sick Children had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Previous Presentation: The abstract of an earlier version of this article was presented as a poster at the 2012 Association of American Medical Colleges’ Research in Medical Education Meeting.
Additional Contributions: We gratefully acknowledge Peter Church, MD (Hospital for Sick Children, University of Toronto, Toronto, Ontario, Canada), for assistance with data analysis. Dr Church did not receive compensation.