Primary Submission Category: Randomized Designs and Analyses
Bayesian Estimation of the Survivor Average Causal Effect for Cluster-Randomized Crossover Trials
Authors: Dane Isenberg, Nandita Mitra, Fan Li, Michael Harhay,
Presenting Author: Dane Isenberg*
In cluster-randomized crossover (CRXO) trials, groups of individuals are assigned to one of two sequences of alternating treatments. Since clusters act as their own control, the CRXO design is typically more statistically efficient than the usual parallel-arm cluster-randomized trial. CRXO trials are increasingly popular in critical care studies where the number of available clusters is generally limited. In trials among severely ill patients, researchers often want to assess the effect of treatments on secondary non-terminal outcomes, but there may be several patients who do not survive to have these measurements fully recorded. To this end, we provide a causal inference framework for addressing truncation by death in the setting of CRXO trials. We target the survivor average causal effect (SACE) estimand, a well-defined subgroup treatment effect represented via principal stratification. We propose structural and standard modeling assumptions to enable SACE identification and estimation within a Bayesian paradigm. We evaluate the small-sample performance of our proposed Bayesian approach for estimation of SACE using CRXO trial data through a simulation study. We apply our methods to a two-period cross-sectional CRXO study examining the impact of proton pump inhibitors as compared to histamine-2 receptor blockers on certain non-mortality outcomes among adults requiring invasive mechanical ventilation.