A0423
Title: Weighting methods for survivor average causal effect estimation in cluster-randomized trials
Authors: Nandita Mitra - University of Pennsylvania (United States) [presenting]
Dane Isenberg - University of Pennsylvania (United States)
Fan Li - Yale University (United States)
Michael Harhay - University of Pennsylvania (United States)
Abstract: Patient-centered outcomes, such as quality of life and length of hospital stay, are often the focus of clinical studies. However, elderly or critically ill clinical trial participants may have truncated or undefined non-mortality outcomes if they do not survive through the measurement time point. To address truncation by death, the survivor average causal effect (SACE) has been proposed as a causally interpretable subgroup treatment effect defined under the principal stratification framework. However, most methods for estimating SACE have been developed in the context of individually-randomized trials. Only limited discussions have centered on cluster-randomized trials (CRTs), where methods typically involve strong distributional assumptions for outcome modeling. Two weighting methods are proposed to estimate SACE in CRTs that obviate the need for potentially complicated outcome distribution modeling. Assumptions that address latent clustering effects to enable point identification of SACE are established, and computationally-efficient asymptotic variance estimators are provided for each weighting estimator. In simulations, the weighting estimators are evaluated, demonstrating their finite-sample operating characteristics and robustness to certain departures from the identification assumptions. The methods are illustrated using data from a CRT to assess the impact of a sedation protocol on mechanical ventilation among children with acute respiratory failure.
