Foundation Model of Electronic Medical Records for Adaptive Risk Estimation

2025-02-10

AI Summary

This study addresses the lack of personalization and interpretability in dynamic risk prediction of clinically critical events from electronic health records (EHRs). We propose ETHOS, a foundational model, and ARES, a clinical risk estimation system. ETHOS introduces Patient Health Timeline (PHT) temporal tokenization—a novel paradigm that converts longitudinal PHTs into structured medical tokens. ARES leverages an enhanced Transformer architecture integrated with dynamic probability calibration and a personalized attribution explanation module, enabling real-time, robust, and clinically customizable risk inference. Evaluated on MIMIC-IV (285,622 PHTs, 357 million tokens), ARES achieves significantly higher AUC than state-of-the-art early-warning systems and machine learning models, while demonstrating superior calibration. The implementation is fully open-sourced.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Hardware-aware ML
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness

Application Category

• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

We developed the Enhanced Transformer for Health Outcome Simulation (ETHOS), an AI model that tokenizes patient health timelines (PHTs) from EHRs. ETHOS predicts future PHTs using transformer-based architectures. The Adaptive Risk Estimation System (ARES) employs ETHOS to compute dynamic and personalized risk probabilities for clinician-defined critical events. ARES incorporates a personalized explainability module that identifies key clinical factors influencing risk estimates for individual patients. ARES was evaluated on the MIMIC-IV v2.2 dataset in emergency department (ED) settings, benchmarking its performance against traditional early warning systems and machine learning models. We processed 299,721 unique patients from MIMIC-IV into 285,622 PHTs, with 60% including hospital admissions. The dataset contained over 357 million tokens. ETHOS outperformed benchmark models in predicting hospital admissions, ICU admissions, and prolonged hospital stays, achieving superior AUC scores. ETHOS-based risk estimates demonstrated robustness across demographic subgroups with strong model reliability, confirmed via calibration curves. The personalized explainability module provides insights into patient-specific factors contributing to risk. ARES, powered by ETHOS, advances predictive healthcare AI by providing dynamic, real-time, and personalized risk estimation with patient-specific explainability to enhance clinician trust. Its adaptability and superior accuracy position it as a transformative tool for clinical decision-making, potentially improving patient outcomes and resource allocation in emergency and inpatient settings. We release the full code at github.com/ipolharvard/ethos-ares to facilitate future research.

Problem

• Predicts future patient health timelines accurately.
• Computes dynamic personalized risk probabilities for events.
• Provides personalized explainability for clinical risk factors.

Innovation

• Transformer-based health outcome simulation
• Dynamic personalized risk estimation system
• Personalized explainability for clinical factors