Advances in Bayesian Methods for Analyzing Complex Observational Study Data

This session focuses on advances in Bayesian methods for analyzing complex observational study data, with an emphasis on covariate missingness, confounder selection, mediation, and longitudinal outcomes subject to truncation by death and loss to follow-up. These problems arise routinely in comparative effectiveness research, health services research, and population based studies that inform clinical and health policy decision making. The session is timely for methodologists and analysts working with modern observational and quasi-experimental health data, where missing covariates, complex confounding structures, mediation, continuous exposure, and loss to follow-up are common. It will appeal to methodologists as well as applied statisticians interested in the development and application of Bayesian methods to answer critical comparative effectiveness research questions and support evidence-based care and policy. This session features four research talks followed by a 20-minute synthesis and open-floor Q&A led by the discussant, Dr. Liangyuan Hu (Rutgers University), to highlight progress and open problems in Bayesian causal inference.