CDE-Mapper: Using Retrieval-Augmented Language Models for Linking Clinical Data Elements to Controlled Vocabularies

2025-05-07

AI Summary

Clinical data elements (CDEs) exhibit high heterogeneity and structural complexity across disparate healthcare systems, impeding standardization and severely limiting data interoperability and integration. To address this, we propose the first modular retrieval-augmented mapping framework for CDEs, which innovatively integrates query decomposition, rule-guided expert prompt engineering, collaborative multi-retriever fusion, and a human-in-the-loop verified knowledge base to enable high-accuracy, automated mapping of CDEs to controlled terminologies. Our approach synergistically combines retrieval-augmented generation (RAG), large language models (LLMs), in-context learning, and multi-source terminology retrieval—balancing precision with interpretability. Evaluated on four heterogeneous clinical datasets, our method achieves an average mapping accuracy 7.2 percentage points higher than state-of-the-art baselines, significantly enhancing data harmonization for clinical decision support and multicenter research.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

The standardization of clinical data elements (CDEs) aims to ensure consistent and comprehensive patient information across various healthcare systems. Existing methods often falter when standardizing CDEs of varying representation and complex structure, impeding data integration and interoperability in clinical research. We introduce CDE-Mapper, an innovative framework that leverages Retrieval-Augmented Generation approach combined with Large Language Models to automate the linking of CDEs to controlled vocabularies. Our modular approach features query decomposition to manage varying levels of CDEs complexity, integrates expert-defined rules within prompt engineering, and employs in-context learning alongside multiple retriever components to resolve terminological ambiguities. In addition, we propose a knowledge reservoir validated by a human-in-loop approach, achieving accurate concept linking for future applications while minimizing computational costs. For four diverse datasets, CDE-Mapper achieved an average of 7.2% higher accuracy improvement compared to baseline methods. This work highlights the potential of advanced language models in improving data harmonization and significantly advancing capabilities in clinical decision support systems and research.

Problem

• Automating linking clinical data elements to controlled vocabularies
• Handling varying CDE complexity and terminological ambiguities
• Improving accuracy in clinical data standardization and interoperability

Innovation

• Retrieval-Augmented Generation with Large Language Models
• Modular query decomposition for CDEs complexity
• Human-in-loop validated knowledge reservoir integration