A0880
Title: Heterogeneous quantile treatment effect estimation with high-dimensional confounding
Authors: Huichen Zhu - The Chinese University of Hong Kong (Hong Kong) [presenting]
Abstract: Understanding heterogeneous treatment responses is essential for advancing precision medicine. This is because individuals often respond differently to the same treatment due to their unique characteristics and circumstances, including patient demographics, genetic predispositions, and environmental exposures. Moreover, inferring causal relationships or associations from observational data can be compromised by the presence of confounding factors, which can sometimes be high-dimensional. The focus is on estimating heterogeneous quantile treatment effects in the context of high-dimensional confounding. The innovative approach leverages quantile regression and random forests to capture the variability of treatment effects across both covariates and outcome distributions. Additionally, orthogonal estimation equations are employed to robustly adjust for high-dimensional confounding. The theoretical properties of the proposed estimator are rigorously explored, and its finite-sample performance is demonstrated through comprehensive simulations. By addressing these complexities, the aim is to enhance the reliability of treatment effect estimates, ultimately contributing to more personalized and effective medical interventions.
