Heterogeneous readmission prediction with hierarchical effect decomposition and regularization

2026-03-19

AI Summary

This work addresses the challenge of accurately and interpretably predicting hospital readmission risk from electronic health records, where patient heterogeneity in primary diagnoses complicates model performance. To this end, the authors propose hierNest, a novel hierarchical modeling framework that leverages the nested structure between diagnosis categories and specific diagnoses. By integrating nested reparameterization with structured regularization, hierNest facilitates information sharing across related subpopulations while decomposing effects across hierarchical levels. The method demonstrates consistently superior predictive performance on both simulated data and a large-scale real-world Medicare dataset, with particularly pronounced gains in settings involving small subgroup sample sizes or substantial heterogeneity in hierarchical effects. Importantly, hierNest maintains model interpretability without sacrificing predictive accuracy.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Structured Learning

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Accurately predicting hospital readmission risks using electronic health records (EHRs) is critical for effective patient management and healthcare resource allocation. Patient populations in health systems are highly heterogeneous across different primary diagnoses, necessitating tailored yet interpretable prediction models. We propose a hierarchical modeling framework incorporating hierarchical nested re-parameterization and structured regularization methods, which we call hierNest. Specifically, our approach leverages the inherent hierarchical structure present in primary diagnoses and groupings of these diagnoses into major diagnostic categories. Our methodology facilitates information borrowing across related patient subgroups and preserves interpretability at different hierarchical levels. Simulation studies demonstrate superior predictive accuracy of the proposed method, particularly with small subgroup sample sizes and varying degrees of hierarchical effects. We apply our methods to a large EHR dataset comprising Medicare patients.

Problem

• readmission prediction
• patient heterogeneity
• hierarchical structure
• electronic health records
• interpretability

Innovation

• hierarchical modeling
• nested re-parameterization
• structured regularization
• heterogeneous readmission prediction
• information borrowing