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.
|Original language||English (US)|
|Journal||Journal of the American Statistical Association|
|State||Accepted/In press - 2022|
- Individualized treatment regime
- Nonconvex problem
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty
FingerprintDive into the research topics of 'Fairness-Oriented Learning for Optimal Individualized Treatment Rules'. Together they form a unique fingerprint.
Fairness-Oriented Learning for Optimal Individualized Treatment Rules
Fang, E. X. (Creator), Wang, Z. (Creator) & Wang, L. (Creator), Taylor & Francis, 2021
DOI: 10.6084/m9.figshare.17064129.v1, https://tandf.figshare.com/articles/dataset/Fairness-Oriented_Learning_for_Optimal_Individualized_Treatment_Rules/17064129/1