Rethinking Retrieval-Augmented Generation for Medicine: A Large-Scale, Systematic Expert Evaluation and Practical Insights

2025-11-10

AI Summary

Medical large language models (LLMs) face two critical challenges: knowledge obsolescence and unverifiable reasoning. While retrieval-augmented generation (RAG) is widely adopted, its reliability in clinical evidence retrieval, filtering, and response generation remains poorly characterized. This study—through the first large-scale expert annotation effort—systematically exposes RAG’s performance degradation in medical settings: only 22% of retrieved documents are clinically relevant; evidence utilization precision is merely 41–43%; and response factuality declines by 6%. To address this, we propose a stage-aware evaluation framework and design deployable optimizations—including evidence filtering and query reformulation. Evaluated on GPT-4o and Llama-3.1-8B, our approach improves accuracy on MedMCQA and MedXpertQA by 12.0% and 8.2%, respectively. These results significantly enhance RAG’s trustworthiness and practical utility in evidence-based medicine.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Large language models (LLMs) are transforming the landscape of medicine, yet two fundamental challenges persist: keeping up with rapidly evolving medical knowledge and providing verifiable, evidence-grounded reasoning. Retrieval-augmented generation (RAG) has been widely adopted to address these limitations by supplementing model outputs with retrieved evidence. However, whether RAG reliably achieves these goals remains unclear. Here, we present the most comprehensive expert evaluation of RAG in medicine to date. Eighteen medical experts contributed a total of 80,502 annotations, assessing 800 model outputs generated by GPT-4o and Llama-3.1-8B across 200 real-world patient and USMLE-style queries. We systematically decomposed the RAG pipeline into three components: (i) evidence retrieval (relevance of retrieved passages), (ii) evidence selection (accuracy of evidence usage), and (iii) response generation (factuality and completeness of outputs). Contrary to expectation, standard RAG often degraded performance: only 22% of top-16 passages were relevant, evidence selection remained weak (precision 41-43%, recall 27-49%), and factuality and completeness dropped by up to 6% and 5%, respectively, compared with non-RAG variants. Retrieval and evidence selection remain key failure points for the model, contributing to the overall performance drop. We further show that simple yet effective strategies, including evidence filtering and query reformulation, substantially mitigate these issues, improving performance on MedMCQA and MedXpertQA by up to 12% and 8.2%, respectively. These findings call for re-examining RAG's role in medicine and highlight the importance of stage-aware evaluation and deliberate system design for reliable medical LLM applications.

Problem

• Evaluating RAG reliability in medical applications through expert assessment
• Addressing performance degradation caused by poor evidence retrieval and selection
• Developing strategies to improve medical LLM factuality and completeness

Innovation

• Systematically decomposing RAG pipeline into three components
• Using evidence filtering to mitigate retrieval failures
• Applying query reformulation for improved evidence selection