A0231
Title: Identification and estimation of causal effects using non-concurrent controls in platform trials
Authors: Michele Santacatterina - New York University (United States) [presenting]
Ivan Diaz - NYU Langone Health (United States)
Federico Macchiavelli Giron - NYU (United States)
Xinyi Zhang - NYU (United States)
Abstract: Platform trials offer a flexible design for simultaneously evaluating multiple treatments. However, their use of non-concurrent controls raises questions for estimating treatment effects. Specifically, which estimands should be targeted? Under what assumptions can these estimands be identified and estimated? Are there any efficiency gains? The purpose is to discuss issues related to the identification and estimation assumptions of common choices of estimand in platform trials that use non-concurrent controls. It is concluded that the most robust strategy to increase efficiency without imposing unwarranted assumptions is to target the concurrent average treatment effect (cATE), the ATE among only concurrent units, using a covariate-adjusted doubly robust estimator. It is suggested that, for the purpose of obtaining efficiency gains, collecting prognostic variables is more important than relying on non-concurrent controls. The perils of targeting ATE are also discussed due to an untestable extrapolation assumption that will often be invalid. Simulations illustrating the points and an application to the ACTT trial are provided, resulting in a 20\% improvement in precision.
