EHR-R1: A Reasoning-Enhanced Foundational Language Model for Electronic Health Record Analysis

2025-10-29

AI Summary

Current large language models (LLMs) face two critical bottlenecks in electronic health record (EHR) analysis: narrow task coverage and weak clinical reasoning capabilities. To address these, we propose EHR-R1—the first reasoning-enhanced LLM family specifically designed for EHRs. Our method introduces (i) EHR-Ins, the first large-scale EHR reasoning instruction dataset, constructed via a novel thought-graph-driven data generation framework; (ii) EHR-Bench, a comprehensive benchmark covering 42 clinically relevant tasks; and (iii) a three-stage training paradigm—domain adaptation, reasoning enhancement, and reinforcement learning. Evaluated on MIMIC-Bench, the 72B-parameter EHR-R1 surpasses GPT-4o by over 30 points; on EHRSHOT, it achieves a 10% zero-shot AUROC improvement. It significantly outperforms both leading open-source and commercial models. EHR-R1 establishes a new paradigm for trustworthy, interpretable clinical AI.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Knowledge Representation and Reasoning: Common-Sense Reasoning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Electronic Health Records (EHRs) contain rich yet complex information, and their automated analysis is critical for clinical decision-making. Despite recent advances of large language models (LLMs) in clinical workflows, their ability to analyze EHRs remains limited due to narrow task coverage and lack of EHR-oriented reasoning capabilities. This paper aims to bridge the gap, specifically, we present EHR-Ins, a large-scale, comprehensive EHR reasoning instruction dataset, comprising 300k high-quality reasoning cases and 4M non-reasoning cases across 42 distinct EHR tasks. Its core innovation is a thinking-graph-driven framework that enables to generate high-quality reasoning data at scale. Based on it, we develop EHR-R1, a series of reasoning-enhanced LLMs with up to 72B parameters tailored for EHR analysis. Through a multi-stage training paradigm, including domain adaptation, reasoning enhancement, and reinforcement learning, EHR-R1 systematically acquires domain knowledge and diverse reasoning capabilities, enabling accurate and robust EHR analysis. Lastly, we introduce EHR-Bench, a new benchmark curated from MIMIC-IV, spanning 42 tasks, to comprehensively assess reasoning and prediction across EHR scenarios. In experiments, we show that the resulting EHR-R1 consistently outperforms state-of-the-art commercial and open-source LLMs (including DeepSeek-V3 and GPT-4o), surpassing GPT-4o by over 30 points on MIMIC-Bench and achieving a 10% higher zero-shot AUROC on EHRSHOT. Collectively, EHR-Ins, EHR-R1, and EHR-Bench have significantly advanced the development for more reliable and clinically relevant EHR analysis.

Problem

• Enhancing EHR analysis with reasoning capabilities for clinical decisions
• Addressing limited task coverage in current clinical language models
• Developing specialized models for complex electronic health record interpretation

Innovation

• Thinking-graph framework generates large-scale EHR reasoning data
• Multi-stage training enhances domain knowledge and reasoning capabilities
• EHR-R1 model outperforms state-of-the-art LLMs on clinical benchmarks