Online Program

The assessment of interaction effects via tree-based methods

Martha Daviglus, Northwestern University 
*Joseph Kang, Northwestern University 
Lei Liu, Northwestern University 
Xiaogang Su, University of Alabama 

Keywords: Causal inference, interaction effect, Chicago healthy aging study cohort

Title: The assessment of interaction effects via tree-based methods

Methods for causal inference with potential outcomes have been extensively developed for estimating average causal effect. Given the recent interest of subgroup-level studies and personalized medicine, research with the potential outcome framework has been developed for effect-modifications and interaction effects. In this talk we review methods for estimating interaction effects such as inverse of the propensity weighted (IPW) method and G-formula estimates, and compare them with our new Tree-based standardization method, which we call the Interaction effect tree (IT). The IT procedure uses a likelihood-based decision rule to divide the subgroups into homogeneous groups where the G-formula can be applied. We applied the IT-based method to assessing the effect of overweight/obesity on the coronary artery calcification (CAC) in the Chicago Healthy Aging Study cohort.