Assessing treatment effects in observational data with missing confounders: A comparative study of practical doubly-robust and traditional missing data methods

2024-12-19

AI Summary

In pharmacoepidemiology, electronic health records frequently exhibit high missingness (>50%) in critical confounders and rare outcomes, leading to substantial bias in causal effect estimation. This study is the first to systematically compare doubly robust estimators—generalized raking and inverse-probability-weighted targeted maximum likelihood estimation (IPW-TMLE)—against conventional approaches—multiple imputation and inverse-probability weighting—within a plasmode simulation framework that preserves real-world data structure. Results demonstrate that both doubly robust estimators markedly reduce bias and mean squared error; IPW-TMLE consistently achieves superior performance across most scenarios. Furthermore, the study proposes a practical, bias–variance trade-off–informed guideline for selecting analytical methods. This work establishes an empirical benchmark and provides actionable recommendations for confounding control in observational studies characterized by high missingness and low event rates.

Technology Category

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

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

In pharmacoepidemiology, safety and effectiveness are frequently evaluated using readily available administrative and electronic health records data. In these settings, detailed confounder data are often not available in all data sources and therefore missing on a subset of individuals. Multiple imputation (MI) and inverse-probability weighting (IPW) are go-to analytical methods to handle missing data and are dominant in the biomedical literature. Doubly-robust methods, which are consistent under fewer assumptions, can be more efficient with respect to mean-squared error. We discuss two practical-to-implement doubly-robust estimators, generalized raking and inverse probability-weighted targeted maximum likelihood estimation (TMLE), which are both currently under-utilized in biomedical studies. We compare their performance to IPW and MI in a detailed numerical study for a variety of synthetic data-generating and missingness scenarios, including scenarios with rare outcomes and a high missingness proportion. Further, we consider plasmode simulation studies that emulate the complex data structure of a large electronic health records cohort in order to compare anti-depressant therapies in a rare-outcome setting where a key confounder is prone to more than 50% missingness. We provide guidance on selecting a missing data analysis approach, based on which methods excelled with respect to the bias-variance trade-off across the different scenarios studied.

Problem

• Evaluating treatment effects when confounder data is missing from observational studies
• Comparing doubly-robust methods with traditional missing data approaches
• Providing guidance for method selection across different missingness scenarios

Innovation

• Doubly-robust estimators handle missing confounders efficiently
• Generalized raking and TMLE methods require fewer assumptions
• Methods compared via simulations with rare outcomes scenarios