SELECTION OF ABSTRACTS FOR AN OTOLARYNGOLOGY MEETING: CORRELATION OF BLINDING WITH SUBSEQUENT PUBLICATION

N Wendell Todd
Department of Surgery and Pediatrics, Emory University School of Medicine, 1365 Clifton Road NE, Atlanta, GA 30322, USA

Objective: To determine whether inclusion in a program was associated with subsequent publication in a MEDLINE-extracted journal.

Design: Retrospective cohort study of 118 abstracts submitted for presentation at a regional otolaryngology meeting. Three practicing otolaryngologists scored each abstract. The abstracts' author(s), institution and geographic location were unknown to 2 scorers. Inter-scorer agreements were poor or nil: kappa = 0.24 between the blinded scorers; k = 0.08 and -0.07 for the un-blinded scorer relative the blinded scorers. All 23 high-scoring abstracts were chosen for presentation at the meeting. Of the next 51 abstracts scored as nearly equivalent, inclusion on the program was determined by (1) balance among the various divisions of otolaryngology, (2) whether the abstract came from within the geographic region, and (3) avoiding multiple presentations from an author. Five years later, a search was conducted for publications by the authors.

Results: Of the 53 abstracts on the program, 35 were subsequently published as full articles in peer-reviewed journals. In contrast, only 8 of the non-selected abstracts were published. Of the near-equivalent abstracts that were included on the program, most (16/28) were published. Conversely, the proportion published for those not on the program was only 4/23 (P<.01). Otolaryngology journals accounted for all but 1 of the published articles. The meeting sponsor's journal, the third most-distributed US otolaryngology journal, accounted for 24 articles, of which 3 had not been on the program.

Conclusions: Arbitrary program inclusion was associated with subsequent publication. Un-blinded scores did not agree with blinded scores.

Return to Session Information