Completeness and changes in data reporting pharmacological interventions to treat COVID-19

TL;DR

This study shows that data reporting for COVID-19 treatment trials is inconsistent across sources, highlighting the need for more reliable and transparent reporting.

Contribution

The study reveals significant changes and missing information in the reporting of pharmacological interventions for COVID-19 across trial registrations and publications.

Findings

19 of 24 WHO TRDS items were changed during the conduct of 122 RCTs and corresponding publications.

Data-sharing statements were missing in 76% of publications compared to 19% at initial registration.

Lowest reliability was found for intervention description changes between trial registrations and publications.

Abstract

Reporting transparency is essential because studies that test pharmacological interventions for COVID-19 should be based on reliably reported data across dissemination sources. We conducted a cross-sectional study of reporting of WHO Trial Registration Data Set (TRDS) items in ClinicalTrials.gov RCTs from January 1, 2020, to May 31, 2021. Completeness and changes among WHO TRDS items were investigated, whereby two authors evaluated RCTs independently to reach κ ≥ 0.80. The frequency of incomplete or uninformative information and the frequency of changes that altered the meaning of WHO TRDS items were assessed. There were changes during the conduct of the trial to 19 of 24 WHO TRDS items for 122 RCTs and 68 corresponding publications in peer-reviewed journals. Among the items, there were greater missing data-sharing statements in publications (52/68 [76%]) than at the initial (23/122 [19%]) or last (17/122 [14%]) registration. The reliability of extractions was high (kappa range: 0.80–1.00), where the lowest (kappa = 0.80, 95% CI 0.59–1.00) was for intervention description changes between the latest registered data and data in publications. Our findings emphasize the need for more reliable reporting of data between COVID-19 dissemination sources.

Figures