EHR-RAGp: Retrieval-Augmented Prototype-Guided Foundation Model for Electronic Health Records

TL;DR

EHR-RAGp is a retrieval-augmented foundation model that dynamically integrates relevant patient history to enhance clinical prediction accuracy from electronic health records.

Contribution

It introduces a prototype-guided retrieval module that effectively aligns and retrieves the most relevant historical clinical data for improved predictions.

Findings

EHR-RAGp outperforms state-of-the-art models across multiple clinical tasks.

Integrating EHR-RAGp with existing models yields significant performance improvements.

The framework efficiently leverages long-range clinical context for better downstream results.

Abstract

Electronic Health Records (EHR) contain rich longitudinal patient information and are widely used in predictive modeling applications. However, effectively leveraging historical data remains challenging due to long trajectories, heterogeneous events, temporal irregularity, and the varying relevance of past clinical context. Existing approaches often rely on fixed windows or uniform aggregation, which can obscure clinically important signals. In this work, we introduce EHR-RAGp, a retrieval-augmented foundation model that dynamically integrates the most relevant patient history across diverse clinical event types. We propose a prototype-guided retrieval module that acts as an alignment mechanism and estimates the relevance of retrieved historical chunks with respect to a given prediction task, guiding the model towards the most informative context. Across multiple clinical prediction tasks, EHR-RAGp consistently outperforms state-of-the-art EHR foundation models and transformer-based baselines. Furthermore, integrating EHR-RAGp with existing clinical foundation models yields substantial performance gains. Overall, EHR-RAGp provides a scalable and efficient framework for leveraging long-range clinical context to improve downstream performance.