Predicting first-episode homelessness among US Veterans using longitudinal EHR data: time-varying models and social risk factors

2026-02-02

AI Summary

This study addresses the prediction of first-time homelessness among U.S. veterans to enable early intervention. Leveraging electronic health records (EHR) from 4.27 million veterans, we developed models that integrate both static and time-varying risk factors, encompassing clinical, social, and behavioral features. We employed classical machine learning approaches alongside Transformer-based masked language models and fine-tuned large language models to capture longitudinal patterns in the data. The incorporation of time-varying social risk factors substantially improved predictive performance, yielding positive predictive values of 11.65–13.80% within the top 1% highest-risk individuals over a 12-month horizon and enhancing PR-AUC by 15–30%. Notably, model performance was more equitable across racial groups, effectively identifying actionable targets for preventive interventions.

Technology Category

• Humans and AI: Human-in-the-loop Machine Learning
• Machine Learning: Privacy
• Computer Vision: Large Vision Models

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Homelessness among US veterans remains a critical public health challenge, yet risk prediction offers a pathway for proactive intervention. In this retrospective prognostic study, we analyzed electronic health record (EHR) data from 4,276,403 Veterans Affairs patients during a 2016 observation period to predict first-episode homelessness occurring 3-12 months later in 2017 (prevalence: 0.32-1.19%). We constructed static and time-varying EHR representations, utilizing clinician-informed logic to model the persistence of clinical conditions and social risks over time. We then compared the performance of classical machine learning, transformer-based masked language models, and fine-tuned large language models (LLMs). We demonstrate that incorporating social and behavioral factors into longitudinal models improved precision-recall area under the curve (PR-AUC) by 15-30%. In the top 1% risk tier, models yielded positive predictive values ranging from 3.93-4.72% at 3 months, 7.39-8.30% at 6 months, 9.84-11.41% at 9 months, and 11.65-13.80% at 12 months across model architectures. Large language models underperformed encoder-based models on discrimination but showed smaller performance disparities across racial groups. These results demonstrate that longitudinal, socially informed EHR modeling concentrates homelessness risk into actionable strata, enabling targeted and data-informed prevention strategies for at-risk veterans.

Problem

• homelessness
• veterans
• risk prediction
• social risk factors
• electronic health records

Innovation

• longitudinal EHR modeling
• social risk factors
• time-varying prediction
• large language models
• homelessness prevention