Objectives: Historically, less than half of peer-reviewed abstracts are published. We set out to determine how many pancreas-related abstracts are published within 5 years of presentation at gastroenterology conferences and to determine a model that predicts successful transition from abstract to journal publication.
Methods: We collected data on study design from all pancreas-related abstracts at the 2010 Digestive Disease Week (DDW), American College of Gastroenterology, and American Pancreatic Association conferences. We then determined whether an abstract was published by October 2015 using a standardized search algorithm.
Results: Of 412 abstracts, 39.8% were published. Studies that were of basic science or translational design (P = 0.02, 0.01, respectively); had more listed authors (P = 0.05); employed randomized, prospective, and multicenter methodology (P = 0.02); and were accepted to DDW (P = 0.02) were more likely to be published. After regression, basic/translational studies (P = 0.002, 0.02, respectively) and DDW-accepted abstracts (P = 0.004) continued to predict successful publication.
Conclusions: It is not clear why only 40% of the pancreas abstracts from 2010 were published 5 years later. Some abstracts may go unpublished because of methodological flaws that escape detection during abstract peer review. Therefore, physicians should use caution when applying abstract data to their clinical decision making.