Learning Disease Progression Models That Capture Health Disparities

2024-12-20

arXiv.org

AI Summary

Existing disease progression models often overlook health disparities, leading to biased real-world data and compromising diagnostic and therapeutic decisions. This paper introduces the first interpretable Bayesian disease progression model that explicitly characterizes three structural health disparities: delayed presentation, accelerated disease progression, and insufficient follow-up. We theoretically demonstrate that omitting any of these disparities induces systematic bias in severity estimation. Grounded in real-world electronic health records, our model integrates a causal inference framework with an identifiable latent-variable design—constituting the first approach to systematically disentangle and distinguish these three inequity mechanisms within disease progression modeling. Evaluated on heart failure data, the model accurately identifies patient subgroups affected by distinct disparities, substantially refines high-risk classification, and improves both predictive fairness and clinical utility.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Philosophy and Ethics of AI: Bias, Fairness & Equity

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Disease progression models are widely used to inform the diagnosis and treatment of many progressive diseases. However, a significant limitation of existing models is that they do not account for health disparities that can bias the observed data. To address this, we develop an interpretable Bayesian disease progression model that captures three key health disparities: certain patient populations may (1) start receiving care only when their disease is more severe, (2) experience faster disease progression even while receiving care, or (3) receive follow-up care less frequently conditional on disease severity. We show theoretically and empirically that failing to account for any of these disparities can result in biased estimates of severity (e.g., underestimating severity for disadvantaged groups). On a dataset of heart failure patients, we show that our model can identify groups that face each type of health disparity, and that accounting for these disparities while inferring disease severity meaningfully shifts which patients are considered high-risk.

Problem

• Existing disease progression models ignore health disparities in data
• Health disparities include delayed care and faster progression rates
• Unaccounted disparities bias severity estimates, especially for disadvantaged groups

Innovation

• Interpretable Bayesian model for disease progression
• Captures three key health disparities
• Identifies high-risk patients more accurately