Improving Precision and Power in Randomized Trials for COVID-19 Treatments Using Covariate Adjustment for Binary, Ordinal, or Time-to-Event Outcomes

• Activity Number: 366 - Trial Design and Analysis Methods for COVID-19 Treatment/Prevention
• Type: Invited
• Date/Time: Thursday, August 12, 2021 : 12:00 PM to 1:50 PM
• Sponsor: ENAR
• Author(s): David Benkeser* and Ivan Diaz and Alex Luedtke and Daniel Scharfstein and Jodi Seagal and Michael Rosenblum
• Companies: Emory University and Weill Cornell Medical College and University of Washington and University of Utah and Johns Hopkins University and Johns Hopkins University
• Keywords: covid19; randomized trials; covariate adjustment; ordinal outcomes

Abstract

There are currently thousands of ongoing randomized trials evaluating products for treating COVID-19. For a substantial number of these trials, there is potential to improve reduce the required sample size through covariate adjustment. Though covariate adjustment is recommended by regulatory agencies, it is underutilized, especially for binary, ordinal and time-to-event outcomes, which are all common in COVID-19 trials. To demonstrate the value of covariate adjustment, we simulated randomized controlled trials of a hypothetical COVID-19 treatment with these types of outcomes. The simulations are based on preliminary data from Weill Cornell Medicine New York Presbyterian Hospital and the Centers for Disease Control. Over sample sizes ranging from 100 to 1000 participants, we found substantial precision gains from using covariate adjustment--equivalent to 4-18\% reductions in the required sample size to achieve a desired power. This was the case for a variety of estimands (targets of inference). From these simulations, we conclude that covariate adjustment is a low-risk, high-reward approach to streamlining COVID-19 treatment trials.