Unpublished completed obstetric randomized clinical trials registered at ClinicalTrials.gov: how big is this issue?

Abstract

A portion of obstetrical randomized clinical trials (RCTs) registered at ClinicalTrials.gov are not published in peer-reviewed journals. To compare the characteristics of completed published versus unpublished randomized clinical trials in obstetrics registered on ClinicalTrials.gov, and to identify barriers to publication. This cross-sectional study queried ClinicalTrials.gov for all completed obstetrical randomized clinical trials registered between January 1, 2009, and December 31, 2018. For each completed obstetrical RCT, we abstracted the following registration fields from ClinicalTrials.gov: ClinicalTrials.gov identifier; recruitment status; trial start and completion dates; study results; type of intervention; study phase; enrollment size; funder type; location; and facilities. Calculated variables included time to completion. In May 2021, we used PubMed and Google Scholar to identify the publication status of completed trials, and we compared the characteristics of published vs. unpublished RCTs. The corresponding authors' email addresses for the unpublished studies were collected from ClinicalTrials.gov and departmental websites. Between September 2021- March 2022, authors of these completed but unpublished obstetrical RCTs were contacted and invited to respond to a survey examining perceptions of barriers to publication, responses of which were collected and presented as counts and percentages. Of the 647 obstetrical randomized clinical trials marked as completed on ClinicalTrials.gov, 378 (58%) were published, and 269 (42%) were unpublished. Unpublished trials were more likely to have an enrollment size of <50 participants (14.5% published, 25.3% unpublished; P < 0.001), and less likely to be conducted at multiple sites (25.4% vs. 17.5%; p < 0.02). The main barriers to publication reported in the survey by authors whose trials were not published included: lack of time (30%), change in employment or completion of training (25%), and results that were not of statistical significance (15%). Among the obstetrical randomized clinical trials registered and marked as completed on ClinicalTrials.gov, over 40% were unpublished. Unpublished trials were more likely to be smaller studies, conducted by researchers who reported experiencing a lack of time as the most common barrier to study publication.