Treatment Non-Adherence Bias in Clinical Machine Learning: A Real-World Study on Hypertension Medication

2025-02-26

AI Summary

This paper identifies an implicit bias in clinical machine learning arising from treatment nonadherence: discrepancies between prescribed therapies documented in electronic health records (EHRs) and patients’ actual behaviors distort causal inference—e.g., reversing estimated treatment effects—and degrade predictive performance (up to 5% AUC reduction), while exacerbating decision-making unfairness for vulnerable populations. Methodologically, we innovatively integrate large language models to extract adherence status from unstructured clinical notes, coupled with structured EHR analysis, causal sensitivity assessment, and multidimensional fairness auditing. In a hypertension cohort, we quantify 21.7% medication nonadherence prevalence. Our work is the first to systematically characterize the dual harm of this bias—causal distortion and predictive degradation—identify its sociodemographic drivers, and elucidate its fairness implications. The study provides critical empirical evidence and a methodological framework for developing responsible, equitable clinical AI systems.

Technology Category

• Machine Learning: Ethics, Bias, and Fairness
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Philosophy and Ethics of AI: Bias, Fairness & Equity

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Fairness and ethical considerations in crowd work and in human-in-the-loop AI systems
• Social Networks and Social Media: Fairness and bias in social network and social media analysis

Abstract

Machine learning systems trained on electronic health records (EHRs) increasingly guide treatment decisions, but their reliability depends on the critical assumption that patients follow the prescribed treatments recorded in EHRs. Using EHR data from 3,623 hypertension patients, we investigate how treatment non-adherence introduces implicit bias that can fundamentally distort both causal inference and predictive modeling. By extracting patient adherence information from clinical notes using a large language model, we identify 786 patients (21.7%) with medication non-adherence. We further uncover key demographic and clinical factors associated with non-adherence, as well as patient-reported reasons including side effects and difficulties obtaining refills. Our findings demonstrate that this implicit bias can not only reverse estimated treatment effects, but also degrade model performance by up to 5% while disproportionately affecting vulnerable populations by exacerbating disparities in decision outcomes and model error rates. This highlights the importance of accounting for treatment non-adherence in developing responsible and equitable clinical machine learning systems.

Problem

• Addresses treatment non-adherence bias in clinical machine learning.
• Investigates bias impact on causal inference and predictive modeling.
• Highlights disparities in decision outcomes and model error rates.

Innovation

• Uses large language models
• Analyzes EHR data
• Identifies treatment non-adherence