Assessment of reporting quality of conference abstracts in sports injury prevention according to CONSORT and STROBE criteria and their subsequent publication rate as full papers

Background: The preliminary results of a study are usually presented as an abstract in conference meetings. The reporting quality of those abstracts and the relationship between their study designs and full paper publication rate is unknown. We hypothesized that randomized controlled trials are more likely to be published as full papers than observational studies. Methods: 154 oral abstracts presented at the World Congress of Sports Injury Prevention 2005 Oslo and the corresponding full paper publication were identified and analysed. The main outcome measures were frequency of publication, time to publication, impact factor, CONSORT (for Consolidated Standards of Reporting Trials) score, STROBE (for Strengthening the Reporting of Observational Studies in Epidemiology) score, and minor and major inconsistencies between the abstract and the full paper publication. Results: Overall, 76 of the 154 (49%) presented abstracts were published as full papers in a peer-reviewed journal with an impact factor of 1.946 +/- 0.812. No significant difference existed between the impact factor for randomized controlled trials (2.122 +/- 1.015) and observational studies (1.913 +/- 0.765, p = 0.469). The full papers for the randomized controlled trials were published after an average (SD) of 17 months (+/- 13 months); for observational studies, the average (SD) was 12 months (+/- 14 months) (p = 0.323). A trend was observed in this study that a higher percentage of randomized controlled trial abstracts were published as full papers (71% vs. 47%, p = 0.078) than observational trials. The reporting quality of abstracts, published as full papers, significantly increased compared to conference abstracts both in randomized control studies (CONSORT: 5.7 +/- 0.7 to 7.2 +/- 1.3; p = 0.018, CI -2.7 to -0.32) and in observational studies (STROBE: 8.2 +/- 1.3 to 8.6 +/- 1.4; p = 0.007, CI -0.63 to -0.10). All of the published abstracts had at least one minor inconsistency (title, authors, research center, outcome presentation, conclusion), while 65% had at least major inconsistencies (study objective, hypothesis, study design, primary outcome measures, sample size, statistical analysis, results, SD/CI). Comparing the results of conference and full paper; results changed in 90% vs. 68% (randomized, controlled studies versus observational studies); data were added (full paper reported more result data) in 60% vs. 30%, and deleted (full paper reported fewer result data) in 40% vs. 30%. Conclusions: No significant differences with respect to type of study (randomized controlled versus observational), impact factor, and time to publication existed for the likelihood that a World Congress of Sports Injury conference abstract could be published as a full paper.