Assessing the robustness of heterogeneous treatment effects in survival analysis under informative censoring

2025-10-15

AI Summary

This paper addresses bias in conditional average treatment effect (CATE) estimation under informative censoring—e.g., early dropout due to adverse events—in survival analysis. Methodologically: (1) it integrates partial identification theory with machine learning to construct *falsifiable bounds* on CATE, relaxing stringent modeling assumptions about the censoring mechanism; (2) it proposes a novel, doubly robust and quasi-oracle efficient meta-learner, compatible with arbitrary base learners, to estimate these bounds efficiently. The key contribution is the first incorporation of partial identification into robust heterogeneous causal inference, substantially mitigating censoring-induced bias. Extensive evaluation—including simulations and real-world clinical trials in oncology—demonstrates improved subgroup identification accuracy and statistical reliability compared to existing methods. The framework enhances the robustness of causal evidence for clinical decision-making under imperfect survival data.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

Dropout is common in clinical studies, with up to half of patients leaving early due to side effects or other reasons. When dropout is informative (i.e., dependent on survival time), it introduces censoring bias, because of which treatment effect estimates are also biased. In this paper, we propose an assumption-lean framework to assess the robustness of conditional average treatment effect (CATE) estimates in survival analysis when facing censoring bias. Unlike existing works that rely on strong assumptions, such as non-informative censoring, to obtain point estimation, we use partial identification to derive informative bounds on the CATE. Thereby, our framework helps to identify patient subgroups where treatment is effective despite informative censoring. We further develop a novel meta-learner that estimates the bounds using arbitrary machine learning models and with favorable theoretical properties, including double robustness and quasi-oracle efficiency. We demonstrate the practical value of our meta-learner through numerical experiments and in an application to a cancer drug trial. Together, our framework offers a practical tool for assessing the robustness of estimated treatment effects in the presence of censoring and thus promotes the reliable use of survival data for evidence generation in medicine and epidemiology.

Problem

• Assessing treatment effect robustness under informative censoring bias
• Developing assumption-lean framework for survival analysis with dropout
• Identifying effective patient subgroups despite biased censoring data

Innovation

• Proposes assumption-lean framework for CATE robustness
• Uses partial identification to derive informative bounds
• Develops meta-learner with double robustness properties