HIV Working Group Seminar

Keith Barnatchez
Doctoral Student, Department of Biostatistics, Harvard University

Estimating Causal Effects with Error-Prone Exposures Using Control Variates
Abstract: Exposure measurement error poses a common challenge to performing causal inference in observational studies. Despite the growing body of work at the intersection of causal inference and measurement error, there remains a critical need for flexible, assumption-lean estimation methods that are easy to implement with desirable theoretical properties. In this talk, we will discuss an estimation approach based on control variates that can be implemented in various study designs to address biases induced by exposure measurement error for general causal quantities. Time permitting, we will briefly preview an ongoing data application using data from the Vanderbilt Comprehensive Care Clinic.