Causal machine learning methods and use of sample splitting in settings with high-dimensional confounding

2024-05-24

AI Summary

To address the challenge of causal effect estimation under high-dimensional confounding, this study systematically compares the Augmented Inverse Probability Weighting (AIPW) and Targeted Maximum Likelihood Estimation (TMLE) estimators under cross-fitting and data-adaptive modeling (including LASSO, random forests, and gradient boosting machines). Our key contributions are threefold: First, we demonstrate—novelly in a real-world, complex epidemiological setting—that cross-fitting is critical for accurate standard error calibration and nominal confidence interval coverage. Second, full-library Super Learner ensemble learning substantially reduces both bias (−37%) and variance (−29%) relative to single-model approaches. Third, while TMLE achieves point estimation accuracy comparable to AIPW, it exhibits superior stability across specifications. Collectively, these results establish the joint use of cross-fitting, full-library Super Learner, and TMLE as a robust and statistically reliable paradigm for causal inference under high-dimensional confounding.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Observational epidemiological studies commonly seek to estimate the causal effect of an exposure on an outcome. Adjustment for potential confounding bias in modern studies is challenging due to the presence of high-dimensional confounding, which occurs when there are many confounders relative to sample size or complex relationships between continuous confounders and exposure and outcome. Despite recent advances, limited evaluation, and guidance are available on the implementation of doubly robust methods, Augmented Inverse Probability Weighting (AIPW) and Targeted Maximum Likelihood Estimation (TMLE), with data-adaptive approaches and cross-fitting in realistic settings where high-dimensional confounding is present. Motivated by an early-life cohort study, we conducted an extensive simulation study to compare the relative performance of AIPW and TMLE using data-adaptive approaches in estimating the average causal effect (ACE). We evaluated the benefits of using cross-fitting with a varying number of folds, as well as the impact of using a reduced versus full (larger, more diverse) library in the Super Learner ensemble learning approach used for implementation. We found that AIPW and TMLE performed similarly in most cases for estimating the ACE, but TMLE was more stable. Cross-fitting improved the performance of both methods, but was more important for estimation of standard error and coverage than for point estimates, with the number of folds a less important consideration. Using a full Super Learner library was important to reduce bias and variance in complex scenarios typical of modern health research studies.

Problem

• Estimating causal effects with high-dimensional confounding in observational studies
• Evaluating doubly robust methods like AIPW and TMLE with cross-fitting
• Assessing Super Learner library impact on bias and variance reduction

Innovation

• Using doubly robust methods like AIPW and TMLE
• Employing cross-fitting to improve variance estimation
• Implementing full Super Learner library for bias reduction