Semi-supervised Concordance Learning for Optimal Individual Treatment Regimes

2026-08-14

AI Summary

This study addresses the label scarcity bottleneck in estimating optimal individualized treatment rules under semi-supervised settings by proposing a semiparametric inference framework that integrates nonparametric imputation with concordance-assisted learning. The method leverages unlabeled samples to enhance information extraction and employs single-index kernel smoothing to improve estimation efficiency and robustness. We establish the consistency and asymptotic normality of the proposed estimator theoretically. Empirical evaluations demonstrate that our approach achieves superior estimation accuracy compared to fully supervised methods in label-limited scenarios and exhibits significant effectiveness on real-world medical data. Ultimately, this work provides a reliable statistical learning tool for precision medicine decision-making, effectively mitigating the challenges posed by limited labeled outcomes in clinical applications.

Abstract

Finding the optimal individualized treatment rule that maps individual characteristics or contextual information to treatment assignments has been extensively investigated in existing literature, with widespread practical applications. This paper considers the estimation of optimal treatment regimes within a semi-supervised data framework (exemplified by electronic medical record data). In such settings, only a tiny proportion of observations have observed outcome labels, owing to high labeling costs, time limitations, data privacy concerns, and other constraints, while covariates and treatment assignments are available for all study subjects. We develop a semi-parametric inference method for optimal treatment regimes, which leverages outcome- unlabeled samples with complete covariate and treatment information to enhance estimation efficiency. The proposed estimation framework consists of two key steps: first, flexible nonparametric imputation via single-index kernel smoothing; second, subsequent estimation of the optimal treatment regime based on concordance-assisted learning. We establish the consistency and asymptotic normality of our proposed estimators. Numerical simulation studies demonstrate that our method achieves higher efficiency and stronger robustness relative to fully supervised estimators under finite-sample settings. We further validate the practical value of our proposed framework using the MIMIC-III and ACTG175 datasets.

Problem

• Optimal Individual Treatment Regimes
• Semi-supervised Learning
• Electronic Medical Records
• Outcome Label Scarcity

Innovation

• Semi-supervised Learning
• Optimal Treatment Regimes
• Concordance-assisted Learning
• Single-index Kernel Smoothing
• Semi-parametric Inference