Bayesian Mediation Analysis for Individualized Treatment Rules

2026-07-30

AI Summary

This study addresses the limited interpretability of existing individualized treatment rule (ITR) evaluation methods, which focus solely on average clinical benefit without elucidating underlying mechanisms. The authors propose a novel framework that integrates causal mediation analysis into ITR assessment by introducing rule-specific nested potential outcomes. This approach decomposes the value difference between a candidate ITR and a reference rule into direct effects and indirect effects operating through specified mediators. Identification is achieved via a rule-level mediation g-formula, and posterior inference is implemented using Bayesian causal mediation forests. Simulations demonstrate favorable finite-sample properties of the proposed estimators. Application to the TRIUMPH trial successfully disentangles the mediating pathways—neurovascular, cardiopulmonary, and behavioral—through which lifestyle interventions improve cognition, substantially enhancing the interpretability of ITR learning.

Abstract

The value of an individualized treatment rule (ITR), defined as the expected outcome under treatment assignment according to the rule, is useful for assessing average clinical benefit but does not explain how the benefit of a rule is generated. We propose a causal mediation framework for decomposing the value contrast between a prespecified candidate ITR and a clinically meaningful reference rule into direct and indirect components. Using rule-specific nested potential outcomes, we define natural direct and indirect rule effects that quantify the extent to which the improvement in value arises through pathways operating directly on the outcome versus through a specified mediator. We give identification conditions under which these components are identified by a rule-level mediation g-formula. For estimation, we adapt Bayesian causal mediation forests to obtain posterior inference for the value contrast and its path-specific components. Our simulations demonstrate that the proposed estimator achieved near-nominal credible interval coverage with decreasing bias and root mean squared error as sample size increased in settings with varying direct and mediated contributions. We further illustrate the method using data from the TRIUMPH trial, decomposing the cognitive benefit of a lifestyle intervention rule through candidate neurovascular, cardiorespiratory, and behavioral mediators. The proposed framework complements optimal ITR learning with explanation using mediation, providing a natural approach for mechanistic evaluation of ITRs.

Problem

• individualized treatment rules
• causal mediation
• treatment effect decomposition
• mechanistic evaluation
• value contrast

Innovation

• Bayesian causal mediation
• individualized treatment rules
• natural direct and indirect effects
• mediation g-formula
• causal mediation forests