Previous studies have shown that abstracts presented at scientific meetings are published less than 50% of the time.1 Similarly, Egloff and colleagues found that 47% of abstracts accepted to the 2009 SGIM meeting were subsequently published, most (86%) within 3 years of the presentation.2 They also found that articles focusing on medical education were most likely to be published. Not surprisingly, in multivariable modeling, presentations of randomized controlled trials and multi-center studies were the strongest predictors of eventual publication. Systematic reviews and those that focused on cancer research had the highest impact factor. In contrast to previous studies,3 the authors found no evidence that studies with statistically significant results were more likely to be published than studies without significant findings.

Abstracts presented at the SGIM meeting are selected through a peer review process. It would have been interesting to explore how well the quality ratings from that review process correlate with both the likelihood of publication and the impact factor for published articles. A surrogate for this could have looked at whether abstracts selected for oral presentation were more likely to be published than those accepted for posters. It would also have been interesting to assess the publication rate and impact factor for rejected submissions to the SGIM meeting. We previously demonstrated that the peer review process for the Journal of General Internal Medicine resulted in the publication of articles with a higher impact factor than articles rejected by JGIM and published elsewhere,4 suggesting value in the peer review process. SGIM members expend considerable time and effort rating the quality of scientific abstract submissions, and expanding this study to look at the correlation between ratings, publications, and impact factor could help validate the process. A study weakness was using the journal impact factor as a surrogate for the article's impact.

It is also interesting to speculate as to what would be required to increase the publication rate. The incoming editors for JGIM are exploring whether a mentoring process could help in this respect. A good follow-up study could involve contacting authors of meeting submissions to survey predictors of successful publication.