Biostatistics Weekly Seminar

Federated learning of causal estimands may greatly improve estimation efficiency by leveraging data from multiple study sites, but robustness to heterogeneity and model misspecifications is vital for ensuring validity. In this talk, I will discuss a novel Federated Adaptive Causal Estimation (FACE) framework to incorporate heterogeneous data from multiple sites to provide treatment effect estimation and inference for a flexibly specified target population of interest. FACE accounts for site-level heterogeneity in the distribution of covariates through density ratio weighting. To safely incorporate source site information and avoid negative transfer, I will introduce an adaptive weighting procedure via a penalized regression, which achieves both consistency and optimal efficiency. FACE is communication-efficient and privacy-preserving, allowing participating sites to only share summary statistics once with other sites. I will show both theoretical and numerical evaluations of FACE and apply it to conduct a comparative effectiveness study of BNT162b2 (Pfizer) and mRNA-1273 (Moderna) vaccines on COVID-19 outcomes in U.S. veterans using electronic health records from five VA regional sites. I will also discuss a few extensions of FACE, namely 1) identifying subgroups and providing inference for heterogeneous treatment effects using federated trees, and 2) making robust inference for a prevailing model when no target outcome data exists, and 3) incorporating transfer learning to estimate treatment effects for underrepresented populations

Hybrid: Meeting Room and Zoom Link to Follow
13 January 2023
1:30pm