Xuqiao Li | Searcharxiv

arXiv2025

Multicategory Matched Learning for Estimating Optimal Individualized Treatment Rules in Observational Studies with Application to a Hepatocellular Carcinoma Study

One primary goal of precision medicine is to estimate the individualized treatment rules (ITRs) that optimize patients' health outcomes based on individual characteristics. Health studies with multiple treatments are commonly seen in practice. However, most existing ITR estimation methods were developed for the studies with binary treatments. Many require that the outcomes are fully observed. In this paper, we propose a matching-based machine learning method to estimate the optimal ITRs in observational studies with multiple treatments when the outcomes are fully observed or right-censored. We establish theoretical property for the proposed method. It is compared with the existing competitive methods in simulation studies and a hepatocellular carcinoma study.

stat.ME↗

arXiv2025

Matching-Based Policy Learning

The beneficial effects of treatments vary across individuals in most studies. Treatment heterogeneity motivates practitioners to search for the optimal policy based on personal characteristics. A long-standing common practice in policy learning has been estimating and maximizing the value function using weighting techniques. Matching is widely used in many applied disciplines to infer causal effects, which is intuitively appealing because the observed covariates are directly balanced across different treatment groups. Nevertheless, matching is rarely explored in policy learning. In this work, we propose a matching-based policy learning framework. We adapt standard and bias-corrected matching methods to estimate an alternative form of the value function: the advantage function, which can be interpreted as the expected improvement achieved by implementing a given policy compared to the equiprobable random policy. We then learn the optimal policy over a restricted policy class by maximizing the matching estimator of the advantage function. We derive a non-asymptotic high probability bound for the regret of the learned optimal policy. Moreover, we show that the learned policy is almost rate-optimal. The competitive finite sample performance of the proposed method compared to weighting-based and outcome modeling-based learning methods is demonstrated in extensive simulation studies and a real data application.

math.ST↗