Primary Submission Category: Heterogeneous Treatment Effects
Comparing Machine Learning Methods for Estimating Heterogeneous Treatment Effects by Combining Data from Multiple Randomized Controlled Trials
Authors: Carly Brantner,
Presenting Author: Carly Brantner*
Individualized treatment decisions can improve health outcomes, but using data to make these decisions in a reliable, powerful, and generalizable way is challenging with a single dataset. Leveraging multiple randomized controlled trials allows for the combination of datasets with unconfounded treatment assignment to improve the power to estimate heterogeneous treatment effects. In this study, we discuss several non-parametric approaches for estimating heterogeneous treatment effects using data from multiple trials. We then compare different single-study methods and different ways of aggregating those to the multi-trial setting through a simulation study, with data generation scenarios that have differing levels of cross-trial heterogeneity. The simulations demonstrate that methods that directly allow for heterogeneity of the treatment effect across trials perform better than methods that do not, and that the choice of single-study method matters based on the functional form of the treatment effect. We determine which methods perform well in each setting and then apply the approaches to three randomized controlled trials comparing the effects of treatments for major depressive disorder.