Achieving Clinically Meaningful Treatment Efficacy Summaries: A COVID-19 Case Study

Dr. L.J. Wei explores the importance of appropriate endpoints and survival analysis methods in critical care studies.

Critical care studies, especially those that are short-term, are up against a data analysis challenge when recovery or improvement rate is determined to be a primary or secondary endpoint and death is a competing risk. Unfortunately, the most commonly used survival analysis techniques and methods for quantifying treatment effects are often not appropriate or ideal when there are competing risks or a need to interpret for a positive event. The result can be an efficacy summary that is unnecessarily ambiguous and difficult for healthcare professionals and regulatory agencies to comprehend and apply.

In a recent study that included Dr. L.J. Wei, researchers looked at two randomized comparative trials evaluating COVID-19 treatments and outlined the problematic gaps created by the use of the Kaplan-Meier method and hazard ratio. According to the published study, “The Kaplan–Meier curve may be severely biased if the mortality and censoring rates are elevated during follow-up.”

The study authors demonstrate how the use of the less well-known area under the cumulative incidence curve results in a more accurate cumulative recovery rate estimate across the study period without the need to independently censor the recovery times of patients who have died. Additionally, the study proposes an “intuitive and clinically interpretable summary of treatment efficacy based on this curve.”

Finally, in the journey to achieve clinically meaningful treatment summaries based on the most appropriate data analysis, the study authors highlight the importance of carefully considering whether a time-period recovery endpoint or a time-to-recover endpoint is more clinically relevant.