ABSTRACT
Background: To examine the outcomes heterogeneity of clinical trial protocols of coronavirus disease 2019 (COVID-19) to prioritize the establishment of a core outcome set. Methods: Databases of the International Committee of Medical Journal Editors - accepted clinical trial registry platforms were searched on February 14, 2020 and May 31, 2020. Randomized controlled trials and non-randomized controlled trials of COVID-19 were considered. Patient condition was classified as common, severe, or critical. Interventions included traditional Chinese medicine and Western medicine. We excluded trials that involved discharged patients, psychological intervention, and complications of COVID-19. The general information and outcomes, outcome measurement instruments, and measurement times were extracted. The results were analyzed by descriptive analysis. Results: In all, 19 registry platforms were searched. A total of 97 protocols were selected from among 160 protocols for the first search. For protocols of traditional Chinese medicine clinical trials, 76 outcomes from 16 outcome domains were reported, and almost half (34/76, 44.74%) of the outcomes were reported only once;the most frequently reported outcome was time taken for severe acute respiratory syndrome coronavirus 2 RNA to become negative. Twenty-seven (27/76, 35.53%) outcomes provided one or more outcome measurement instruments. Ten outcomes provided one or more measurement time frame. For protocols of Western medicine clinical trials, 126 outcomes from 17 outcome domains were reported;almost half (62/126, 49.21%) of the outcomes were reported only once;the most frequently reported outcome was proportion of patients with negative severe acute respiratory syndrome coronavirus 2. Twenty-seven outcomes provided one or more outcome measurement instruments. Forty (40/126, 31.75%) outcomes provided one or more measurement time frame. There were > 40 duplicated outcomes between the clinical trials protocols of traditional Chinese medicine and western medicine protocols. All of them were included in the Delphi survey when developing core outcome set for COVID-19. A total of 1,027 protocols were selected from 2,741 protocols for the second search. Forty-two new outcomes and 47 new outcome measurement instruments were reported. Conclusion: Outcome reporting in clinical trial protocols of COVID-19 is inconsistent. Thus, establishing a core outcome set is necessary for diagnosis and management.