Poor reporting quality of observational clinical studies comparing treatments of COVID-19 – a retrospective cross-sectional study

Abstract

Abstract Background During the COVID-19 pandemic, the scientific world is in urgent need for new evidence on the treatment of COVID patients. The reporting quality is crucial for transparent scientific publication. Concerns of data integrity, methodology and transparency were raised. Here, we assessed the adherence of observational studies comparing treatments of COVID 19 to the STROBE checklist in 2020. Methods Design: We performed a retrospective, cross-sectional study. Setting: We conducted a systematic literature search in the Medline database. This study was performed at the RWTH Aachen University Hospital, Department of Anaesthesiology Participants: We extracted all observational studies on the treatment of COVID-19 patients from the year 2020. Main outcome measures: The adherence of each publication to the STROBE checklist items was analysed. The journals’ impact factor (IF), the country of origin, the kind of investigated treatment and the month of publication were assessed. Results We analysed 147 observational studies and found a mean adherence of 45.6% to the STROBE checklist items. The percentage adherence per publication correlated significantly with the journals’ IF (point estimate for the difference between 1st and 4th quartile 11.07%, 95% CI 5.12 to 17.02, p < 0.001). U.S. American authors gained significantly higher adherence to the checklist than Chinese authors, mean difference 9.10% (SD 2.85%, p = 0.023). Conclusions We conclude a poor reporting quality of observational studies on the treatment of COVID-19 throughout the year 2020. A considerable improvement is mandatory.