Event Date
Speaker: Fan Xia (Assistant Professor, Epidemiology and Biostatistics, UC San Francisco)
Title: Efficient Design-based Inference for the Stepped Wedge Design
Abstract: Stepped wedge designs (SWDs) are increasingly used to evaluate interventions delivered at the cluster level, but they present a range of analytical challenges. These include confounding by time due to staggered rollout, heterogeneous time trends across clusters, incidental imbalances from small number of clusters, complex correlation structures, and unreliable standard error estimation. Conventional approaches either require full model specification over time or inefficiently pooling within-period comparisons. We propose a unified semiparametric framework for effect estimation and inference in SWDs that directly addresses these challenges. The proposed estimator is robust to misspecification of nuisance components of cluster-period means, being consistent and asymptotically normal, and achieves the semiparametric efficiency bound under correct covariance and cluster-period mean specification. Our theory accommodates arbitrary treatment effect models and is derived using a nonstandard extension of semiparametric efficiency theory tailored to SWDs with varying cluster-period sizes. To support inference in trials with few clusters, we introduce a permutation-based standard error estimator and a leave-one-out correction to reduce plug-in bias. Through simulation studies and application to a public health trial, we demonstrate the method’s robustness and efficiency relative to standard approaches.
Faculty Web-page (links to UCSF): https://profiles.ucsf.edu/fan.xia