Causal Mediation and Interaction

EPI 542 Subject & Catalog Number

Description

The course will approach concepts and methods for mediation and interaction from the perspective of the counterfactual framework. The first part of the course will be concerned with mediation analysis, that is, assessing the extent to which the effect of an exposure on some outcome is mediated through a particular intermediate and the extent to which it is direct or through other pathways. Definitions, theoretical identification results, and statistical techniques related to mediation analysis will be covered. The material in this part of the course will clarify the assumptions required for the estimation of direct and indirect effect and will extend the approach to mediation typically employed in epidemiology and the social sciences to settings with interactions, non-linearities and time-varying exposures. The second part of the course will cover concepts and methods for interaction. Conceptual issues concerning interaction, effect modification and the relation and non-correspondence of statistical and mechanistic notions of interaction will be discussed. Sensitivity analysis approaches and power and sample size calculations for interaction will be introduced. The course will conclude by providing a unifying framework for the assessment of mediation and interaction.

Pre-Requisites

EPI289 or EPI207 or PHS2000B or permission of the instructor. Familiarity with counterfactuals, and linear and logistic regression will be assumed. Some exposure to inverse probability of treatment weighting, marginal structural models, and causal diagrams will be helpful. Familiarity with SAS, Stata or R will be necessary to complete the assignments.