Conditional Front-door Adjustment for Heterogeneous Treatment Assignment Effect Estimation Under Non-adherence

2025-05-08

AI Summary

Estimating heterogeneous treatment effects under noncompliance suffers from substantial bias and high variance. Method: We propose the Conditional Front-Door (CFD) adjustment framework, theoretically proving that its asymptotic variance strictly dominates that of standard back-door adjustment when true treatment effects are weak. To enable end-to-end estimation, we introduce a shared-parameter modeling paradigm and design LobsterNet—a multi-task causal neural network that jointly learns multiple auxiliary nuisance parameters. Results: Across multiple semi-synthetic and real-world healthcare datasets, LobsterNet reduces average estimation error by 18.7% over state-of-the-art baselines, significantly improving the reliability and stability of personalized treatment decisions in noncompliance settings.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Relational Probabilistic Models
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Estimates of heterogeneous treatment assignment effects can inform treatment decisions. Under the presence of non-adherence (e.g., patients do not adhere to their assigned treatment), both the standard backdoor adjustment (SBD) and the conditional front-door adjustment (CFD) can recover unbiased estimates of the treatment assignment effects. However, the estimation variance of these approaches may vary widely across settings, which remains underexplored in the literature. In this work, we demonstrate theoretically and empirically that CFD yields lower-variance estimates than SBD when the true effect of treatment assignment is small (i.e., assigning an intervention leads to small changes in patients' future outcome). Additionally, since CFD requires estimating multiple nuisance parameters, we introduce LobsterNet, a multi-task neural network that implements CFD with joint modeling of the nuisance parameters. Empirically, LobsterNet reduces estimation error across several semi-synthetic and real-world datasets compared to baselines. Our findings suggest CFD with shared nuisance parameter modeling can improve treatment assignment effect estimation under non-adherence.

Problem

• Estimates heterogeneous treatment effects under non-adherence conditions
• Compares variance of standard backdoor vs. conditional front-door adjustments
• Introduces LobsterNet for multi-task modeling of nuisance parameters

Innovation

• Conditional Front-door Adjustment for unbiased estimates
• LobsterNet multi-task neural network for nuisance parameters
• Shared nuisance parameter modeling improves estimation