PHS 7030 - Applied Modern Causal Inference In Comparative Effectiveness Research

This course will provide an overview of modern causal inference using the framework of counterfactual outcomes to provide rigorous definitions of causal effects, confounding and selection biases. Topics include the development of directed acyclic graphs to characterize patterns of confounding and define assumptions required for causal inferences, adjustment for confounding using propensity scores, inverse probability weighting, instrumental variables, and difference-in-difference approaches. Additionally, the course delves into how causal inference methods interface with techniques for network meta-analysis. Methods for implementing methods for causal inference with standard statistical software will be emphasized. A theoretical supplement to this course will provide a more in-depth theoretical background.