Safe Individualized Treatment Rules with Controllable Harm Rates

Abstract: Estimating individualized treatment rules (ITRs) is crucial for tailoring interventions in precision medicine. Typical ITR estimation methods rely on conditional average treatment effects (CATEs) to guide treatment assignments. However, such methods overlook individual-level harm within covariate-specific subpopulations, potentially leading many individuals to experience worse outcomes under CATE-based ITRs. In this article, we aim to estimate ITRs that maximize the reward while ensuring that the harm rate induced by the ITR remains below a pre-specified threshold. We derive the explicit form of the optimal ITR, propose a plug-in estimator when both the CATE and harm rate are identifiable, and establish the large-sample properties of the proposed estimator. Given that the harm rate is generally unidentifiable due to its dependence on the joint distribution of potential outcomes, we further propose {three} strategies for estimating ITRs with a controllable harm rate under partial identification. By accounting for both reward and harm, our method offers a reliable solution for developing ITRs in high-stakes domains where harm is crucial. Extensive simulations demonstrate the effectiveness of the proposed methods in controlling harm rates. We apply the method to the Right Heart Catheterization dataset, assessing the potential reduction in harm rate compared to historical interventions.