Efficient Estimation of Causal Effects Under Two-Phase Sampling with Error-Prone Outcome and Treatment Measurements

2025-06-26

AI Summary

Electronic health record (EHR) data often suffer from measurement error in both outcome and treatment variables, compounded by two-stage sampling—where only a subset of records undergoes manual verification—leading to biased causal inference of the average treatment effect (ATE). Method: We propose a doubly robust, efficient, and stability-enhanced ATE estimator that unifies two previously independent frameworks for efficient two-stage sampling estimation. Our approach integrates inverse probability weighting, influence function construction, and measurement error correction to yield a calibrated estimator. Contribution/Results: The proposed method substantially improves finite-sample accuracy and robustness. Simulation studies and analysis of real-world HIV antiretroviral therapy data demonstrate a 30–50% reduction in mean squared error compared with state-of-the-art alternatives. It provides a generalizable, high-precision statistical solution for causal inference in EHR studies involving measurement error and two-stage sampling.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: State Estimation
• Reasoning under Uncertainty: Causality

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Measurement error is a common challenge for causal inference studies using electronic health record (EHR) data, where clinical outcomes and treatments are frequently mismeasured. Researchers often address measurement error by conducting manual chart reviews to validate measurements in a subset of the full EHR data -- a form of two-phase sampling. To improve efficiency, phase-two samples are often collected in a biased manner dependent on the patients' initial, error-prone measurements. In this work, motivated by our aim of performing causal inference with error-prone outcome and treatment measurements under two-phase sampling, we develop solutions applicable to both this specific problem and the broader problem of causal inference with two-phase samples. For our specific measurement error problem, we construct two asymptotically equivalent doubly-robust estimators of the average treatment effect and demonstrate how these estimators arise from two previously disconnected approaches to constructing efficient estimators in general two-phase sampling settings. We document various sources of instability affecting estimators from each approach and propose modifications that can considerably improve finite sample performance in any two-phase sampling context. We demonstrate the utility of our proposed methods through simulation studies and an illustrative example assessing effects of antiretroviral therapy on occurrence of AIDS-defining events in patients with HIV from the Vanderbilt Comprehensive Care Clinic.

Problem

• Estimating causal effects with error-prone EHR data
• Improving efficiency in biased two-phase sampling
• Developing robust estimators for treatment effects

Innovation

• Doubly-robust estimators for treatment effects
• Efficient two-phase sampling with measurement error
• Modified estimators for improved finite sample performance