Fairness-Oriented Learning for Optimal Individualized Treatment Rules

• 成果类型: Article
• 署名作者: Fang, Ethan X.; Wang, Zhaoran; Wang, Lan
• 署名单位: Pennsylvania Commonwealth System of Higher Education (PCSHE); Pennsylvania State University; Pennsylvania State University - University Park; Northwestern University; University of Miami; Duke University
• 刊物名称: JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
• ISSN/ISSBN: 0162-1459
• DOI: 10.1080/01621459.2021.2008402
• 发表日期: 2023
• 页码: 1733-1746
• 关键词: statistical treatment choice; Causal Inference; Decision rules; assignment; strategies; program; DESIGN; models

摘要

There has recently been a surge on the methodological development for optimal individualized treatment rule (ITR) estimation. The standard methods in the literature are designed to maximize the potential average performance (assuming larger outcomes are desirable). A notable drawback of the standard approach, due to heterogeneity in treatment response, is that the estimated optimal ITR may be suboptimal or even detrimental to certain disadvantaged subpopulations. Motivated by the importance of incorporating an appropriate fairness constraint in optimal decision making (e.g., assign treatment with protection to those with shorter survival time, or assign a job training program with protection to those with lower wages), we propose a new framework that aims to estimate an optimal ITR to maximize the average value with the guarantee that its tail performance exceeds a prespecified threshold. The optimal fairness-oriented ITR corresponds to a solution of a nonconvex optimization problem. To handle the computational challenge, we develop a new efficient first-order algorithm. We establish theoretical guarantees for the proposed estimator. Furthermore, we extend the proposed method to dynamic optimal ITRs. The advantages of the proposed approach over existing methods are demonstrated via extensive numerical studies and real data analysis.