PICOs-RAG: PICO-supported Query Rewriting for Retrieval-Augmented Generation in Evidence-Based Medicine

2025-10-27

AI Summary

Real-world clinical queries often suffer from missing information or ambiguous phrasing, causing existing retrieval-augmented generation (RAG) methods to retrieve irrelevant evidence and generate unreliable answers in evidence-based medicine (EBM). To address this, we propose PICO-RAG—a novel framework that for the first time deeply integrates the PICO (Patient, Intervention, Comparison, Outcome) structured framework into the RAG pipeline. Our PICO-driven query rewriting module automatically identifies and completes incomplete or non-technical clinical queries by extracting and enriching PICO elements, thereby standardizing and semantically enhancing user inputs. This improves both retrieval relevance and clinical accuracy of generated responses. Evaluated on real-world clinical datasets, PICO-RAG achieves an average 8.8% end-to-end performance gain over strong baselines. Our work establishes a new paradigm for enhancing interpretability, robustness, and clinical utility of RAG in EBM applications.

Technology Category

• Data Mining & Knowledge Management: Intelligent Query Processing
• Natural Language Processing: Question Answering
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

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

Abstract

Evidence-based medicine (EBM) research has always been of paramount importance. It is important to find appropriate medical theoretical support for the needs from physicians or patients to reduce the occurrence of medical accidents. This process is often carried out by human querying relevant literature databases, which lacks objectivity and efficiency. Therefore, researchers utilize retrieval-augmented generation (RAG) to search for evidence and generate responses automatically. However, current RAG methods struggle to handle complex queries in real-world clinical scenarios. For example, when queries lack certain information or use imprecise language, the model may retrieve irrelevant evidence and generate unhelpful answers. To address this issue, we present the PICOs-RAG to expand the user queries into a better format. Our method can expand and normalize the queries into professional ones and use the PICO format, a search strategy tool present in EBM, to extract the most important information used for retrieval. This approach significantly enhances retrieval efficiency and relevance, resulting in up to an 8.8% improvement compared to the baseline evaluated by our method. Thereby the PICOs-RAG improves the performance of the large language models into a helpful and reliable medical assistant in EBM.

Problem

• Enhances query handling for complex clinical scenarios
• Improves retrieval efficiency in evidence-based medicine systems
• Transforms user queries into structured PICO format

Innovation

• PICO format expands and normalizes user queries
• Query rewriting enhances retrieval efficiency and relevance
• PICOs-RAG improves large language model medical assistance