Registration Practices of Randomized Clinical Trials in Rhinosinusitis: A Cross-sectional Review.

Randomized clinical trials (RCTs) play an important role in the development of clinical practice guidelines and in clinical decision making. Little is known about the registration practices of RCTs in the diagnosis and treatment of rhinosinusitis. The primary outcome was the frequency of reported RCT registry numbers by authors of rhinosinusitis RCTs. A secondary outcome was the rates of selective reporting bias in RCTs that were prospectively registered. A tertiary outcome end point was the frequency of publication of RCTs registered on ClinicalTrials.gov. Our sample was derived from a PubMed (MEDLINE) search performed on October 31, 2017, using the keywords "sinusitis OR rhinosinusitis OR rhinitis," and filtering by date (January 1, 2015, through October 31, 2017) and study type (RCT). Studies that were considered an RCT were included for the primary outcome analysis. All RCTs that were registered prior to or during patient enrollment were included for secondary outcome analysis. For the tertiary outcome, a search was performed on ClinicalTrials.gov on September 4, 2018, using the keywords "sinusitis OR rhinosinusitis," and filtering by date (January 1, 2013, through October 31, 2015). All analysis took place between October 29, 2018, and October 31, 2018. A total of 179 RCTs were analyzed for our primary outcome: 94 (52.5%) included a registration number in their publication, and 70 (39.1%) were included for secondary outcome analysis of rates of selective reporting bias. Of these 70 RCTs, 22 (31%) were found to have at least 1 major discrepancy between trial registration and publication. For the tertiary outcome, 52 completed clinical trials were identified on ClinicalTrials.gov, of which 21 (40%) had listed publication in the registry. We found that published rhinosinusitis RCTs frequently do not include a trial registration number and that registered RCTs frequently are not published. Furthermore, RCTs that are registered often display selective reporting of their outcomes and frequently favor positive results. We recommend strict adherence to RCT registration policies and the enforcement of accurate reporting to help strengthen the evidence behind clinical decision making.