Resolving the bias-precision paradox with stochastic causal representation learning for personalized medicine

TL;DR

This paper introduces a novel stochastic causal representation learning method, sMMD, to improve personalized treatment effect estimation by balancing bias reduction and heterogeneity preservation in observational data.

Contribution

It proposes sMMD, a sampling-based alignment strategy, to address the bias-precision paradox in causal representation learning for personalized medicine.

Findings

Improves accuracy under distribution shift, reducing error by up to 11.5%.

Increases recall in high-risk clinical tasks.

Enhances clinician decision accuracy by 14.7% and reduces decision time.

Abstract

Estimating individualized treatment effects from longitudinal observational data is central to data-driven medicine, yet existing methods face a fundamental limitation: reducing confounding bias often suppresses clinically informative heterogeneity, degrading patient-specific predictions. Here, we identify this tension as a bias-precision paradox in causal representation learning and introduce sampling-based maximum mean discrepancy (sMMD), a stochastic alignment strategy that replaces global adversarial balancing with subset-level matching. We instantiate this approach in a framework for counterfactual outcome prediction with attribution-grounded interpretability. Across two large-scale ICU cohorts (n = 27,783), our framework improves accuracy under distribution shift, reducing error by up to 11.5% and substantially increasing recall in high-risk tasks. Mechanistic analyses show that sMMD selectively preserves clinically decisive variables. In human-AI evaluation, our method outperforms clinicians-in-training and large language models, and improves clinician accuracy by 14.7% while reducing decision time, enabling interpretable, real-time clinical decision support.