Poor reporting quality of randomized controlled trials comparing treatments of COVID-19-A retrospective cross-sectional study on the first year of publications.

Abstract

Transparent and complete reporting of randomized controlled trials (RCTs) is essential for critical scientific appraisal of the results. It has been argued whether publications during the COVID-19 pandemic have met reporting standards. In this study, we assessed reporting adherence of RCTs on treatment interventions in COVID-19 patients to the CONSORT checklist and discuss which lessons can be learned to improve reporting in the future. This was a retrospective, cross-sectional study performed at the University Hospital RWTH Aachen, Germany. We conducted a pragmatic systematic literature search in the PubMed database to identify RCTs on treatment interventions in COVID-19 patients in the first year of publications on the topic (March 2020-February 2021). We investigated the adherence of each publication to the CONSORT checklist and assessed the association between specific predictors and percentage adherence in an exploratory multivariable regression model. We analyzed 127 RCTs and found that the median percentage adherence to the CONSORT checklist was 54.3% [IQR 38.9 to 65.7]. In the exploratory multivariable regression model, the impact factor (highest tertile of impact factor compared to lowest tertile ß = 21.77, 95% CI 13.89 to 29.66, p<0.001; middle tertile compared lowest tertile ß = 11.79, 95% CI 5.74 to 17.84, p<0.001)) and authors' referral to the CONSORT statement (ß = 9.29, 95% CI 2.98 to 15.60, p = 0.004) were associated with a higher percentage adherence to the CONSORT checklist. The reporting quality of RCTs on treatment interventions in COVID-19 patients during the first year of publications was poor. Measures to improve reporting quality are urgently needed.