Optimal Treatment Policy Estimation for Recurrent Events with a Competing Terminal Event: An Instrumented Difference-in-Differences Approach

2026-06-02

AI Summary

This study addresses the challenge of estimating optimal treatment strategies for recurrent events from administrative health data in the presence of unmeasured confounding and competing terminal events, such as death. The authors propose an instrumental variable–based difference-in-differences method (iDID) that innovatively integrates instrumental variable and difference-in-differences frameworks to achieve identification robust to persistent unobserved heterogeneity. The approach explicitly avoids recommending interventions that reduce recurrent events at the expense of survival. Leveraging two inverse probability weighting mechanisms, the method yields a multiply robust estimator that guarantees consistency under correct specification of any one of several candidate models and possesses asymptotic normality. Simulations demonstrate superior finite-sample performance compared to existing methods. Applied to U.S. Medicare data, the approach successfully optimizes first-line treatment for type 2 diabetes, significantly reducing diabetes-related hospitalizations while preserving survival.

Abstract

Learning reproducible and generalizable optimal treatment policies for chronic diseases requires large, representative populations with long-term follow-up. Administrative health data provide a natural starting point, but their use is often limited by unmeasured confounding. We address this by proposing a novel framework based on Instrumented Difference-in-Differences (iDID) to estimate optimal policies for recurrent event outcomes subject to a terminating event. The iDID design is particularly useful in this setting because it leverages policy-induced treatment variation while allowing for persistent unmeasured differences across populations, relying on assumptions that are more plausible for administrative health data than those required by conventional IV or DID approaches. A key feature of our approach is that it explicitly addresses the fundamental challenge of avoiding policies that trivially reduce recurrent adverse events by increasing mortality. We derive two distinct Inverse Probability Weighted identifications and develop a multiply robust estimator that achieves consistency if any one of several subsets of nuisance models is correctly specified. We establish the estimator's consistency and asymptotic normality through large-sample theory and demonstrate its superior finite-sample performance over existing methods via simulation. Finally, we apply this framework to a national Medicare dataset to optimize first-line Type 2 Diabetes strategies, specifically targeting the minimization of disease-related hospitalizations while accounting for survival.

Problem

• optimal treatment policy
• recurrent events
• competing terminal event
• unmeasured confounding
• administrative health data

Innovation

• Instrumented Difference-in-Differences
• optimal treatment policy
• recurrent events
• competing terminal event
• multiply robust estimation