Reporting quality of journal abstracts for surgical randomized controlled trials before and after the implementation of the CONSORT extension for abstracts

Background Adequate reporting is crucial in full-text publications but even more so in abstracts because they are the most frequently read part of a publication. In 2008, an extension for abstracts of the Consolidated Standards of Reporting Trials (CONSORT-A) statement was published, defining which items should be reported in abstracts of randomized controlled trials (RCTs). Therefore, we compared the adherence of RCT abstracts to CONSORT-A before and after the publication of CONSORT-A. Methods RCTs published in the five surgical journals with the highest impact factor were identified through PubMed for 2005–2007 and 2014–2016. Adherence to 15 CONSORT-A items and two additional items for abstracts of non-pharmacological trials was assessed in duplicate. We compared the overall adherence to CONSORT-A between the two time periods using an unpaired t test and explored adherence to specific items. Results A total of 192 and 164 surgical RCT abstracts were assessed (2005–2007 and 2014–2016, respectively). In the pre-CONSORT-A phase, the mean score of adequately reported items was 6.14 (95% confidence interval [CI] 5.90–6.38) and 8.11 in the post-CONSORT-A phase (95% CI 7.83–8.39; mean difference 1.97, 95% CI 1.60–2.34; p