Encoding and inference on separable effects for sustained treatments

2025-08-14

AI Summary

This study addresses the evaluation of separable causal effects under sustained treatment strategies—focusing on time-varying interventions where clinical decision-making prioritizes full adherence effects over initial treatment assignment or pointwise exposure. Method: We extend separable effects theory to longitudinal settings for the first time, introducing a novel treatment strategy encoding scheme that simplifies identification assumptions and enables transparent graphical representation. Building upon this, we construct a doubly robust, semiparametrically efficient estimator by integrating time-varying treatment coding, identification function modeling, doubly robust estimation, and semiparametric efficiency theory. Results: Applied to the SPRINT trial data, our method successfully quantifies the separable effect of blood pressure intervention on acute kidney injury risk, demonstrating both statistical validity and clinical interpretability. The approach yields theoretically justified, finite-sample robust estimates while preserving causal clarity in dynamic treatment regimes.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Intelligent Robots: State Estimation

Application Category

• User Modeling, Personalization and Recommendation: Studies of user behavior, including longitudinal effects of personalized systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Sustained treatment strategies are common in many domains, particularly in medicine, where many treatment are delivered repeatedly over time. The effects of adherence to a treatment strategy throughout follow-up are often more relevant to decision-makers than effects of treatment assignment or initiation. Here we consider the separable effect of sustained use of a time-varying treatment. Despite the potential usefulness of this estimand, the theory of separable effects has yet to be extended to settings with sustained treatment strategies. To derive our results, we use an unconventional encoding of time-varying treatment strategies. This allows us to obtain concise formulations of identifying assumptions with better practical properties; for example, they admit frugal graphical representations and formulations of identifying functionals. These functionals are used to motivate doubly robust semiparametrically efficient estimators. The results are applied to the Systolic Blood Pressure Intervention Trial (SPRINT), where we estimate a separable effect of modified blood pressure treatments on the risk of acute kidney injury.

Problem

• Extend separable effects theory to sustained treatment strategies
• Develop concise identifying assumptions with practical graphical representations
• Estimate separable effects of modified blood pressure treatments

Innovation

• Unconventional encoding of time-varying treatment strategies
• Concise formulations with frugal graphical representations
• Doubly robust semiparametrically efficient estimators