Statistics Seminar - 04/29/25

Abstract:

We consider assessing causal mediation in the presence of a post-treatment event (examples include noncompliance, a clinical event, or death). We identify natural mediation effects for the entire study population and for each principal stratum characterized by the joint potential values of the post-treatment event. We derive the efficient influence function for each mediation estimand, which motivates a set of multiply robust estimators for inference. The multiply robust estimators are consistent under four types of misspecifications and are efficient when all nuisance models are correctly specified. We also develop a nonparametric efficient estimator that leverages data-adaptive machine learners to achieve efficient inference and discuss sensitivity methods to address key identification assumptions. We illustrate our methods via simulations and two real data examples.

About the Speaker:

Dr. Fan Li is an Associate Professor in the Department of Biostatistics at the Yale School of Public Health. He received his PhD in Biostatistics from Duke University in 2019, and joined the Yale Biostatistics faculty in July, 2019. Dr. Li’s research interests include statistical methods for randomized clinical trials, observational studies and a combination of both. He is an expert in the design, monitoring, analysis of parallel-arm, crossover and stepped-wedge cluster randomized trials, which are increasingly seen in pragmatic clinical trials embedded in the health care delivery systems. He has also contributed novel propensity score methods and software to estimate average causal effects with observational data, aimed at improving overlap and internal validity. His recent methods research include generalizability of randomized trials to external target populations, confirmatory or exploratory heterogeneity of treatment effects analyses, complex endpoints in cluster randomized trials, as well as novel study designs to address patient-centered clinical research questions. His methodological research has been supported by multiple NIH and PCORI grants/awards.