Lee Simon A, Jain Sujay, Chen Alex, Ono Kyoka, Biswas Arabdha, Rudas Ákos, Fang Jennifer, Chiang Jeffrey N
Department of Computational Medicine, University of California Los Angeles, Los Angeles, CA, USA.
LA Health Services, Los Angeles, CA, USA.
NPJ Digit Med. 2025 Jul 2;8(1):394. doi: 10.1038/s41746-025-01777-x.
Electronic health records (EHR) contain data from disparate sources, spanning various biological and temporal scales. In this work, we introduce the Multiple Embedding Model for EHR (MEME), a deep learning framework for clinical decision support that operates over heterogeneous EHR. MEME first converts tabular EHR into "pseudo-notes", reducing the need for concept harmonization across EHR systems and allowing the use of any state-of-the-art, open source language foundation models. The model separately embeds EHR domains, then uses a self-attention mechanism to learn the contextual importance of these multiple embeddings. In a study of 400,019 emergency department visits, MEME successfully predicted emergency department disposition, discharge location, intensive care requirement, and mortality. It outperformed traditional machine learning models (Logistic Regression, Random Forest, XGBoost, MLP), EHR foundation models (EHR-shot, MC-BEC, MSEM), and GPT-4 prompting strategies. Due to text serialization, MEME also exhibited strong few-shot learning performance in an external, unstandardized EHR database.
电子健康记录（EHR）包含来自不同来源的数据，涵盖各种生物和时间尺度。在这项工作中，我们介绍了用于EHR的多重嵌入模型（MEME），这是一种用于临床决策支持的深度学习框架，可在异构EHR上运行。MEME首先将表格形式的EHR转换为“伪笔记”，减少了跨EHR系统进行概念协调的需求，并允许使用任何最先进的开源语言基础模型。该模型分别对EHR领域进行嵌入，然后使用自注意力机制来学习这些多重嵌入的上下文重要性。在一项对400,019次急诊科就诊的研究中，MEME成功预测了急诊科处置、出院地点、重症监护需求和死亡率。它优于传统机器学习模型（逻辑回归、随机森林、XGBoost、多层感知器）、EHR基础模型（EHR-shot、MC-BEC、MSEM）和GPT-4提示策略。由于文本序列化，MEME在外部未标准化的EHR数据库中也表现出强大的少样本学习性能。
