MED-COPILOT: A Medical Assistant Powered by GraphRAG and Similar Patient Case Retrieval
arXiv:2603.00460v1 Announce Type: new Abstract: Clinical decision-making requires synthesizing heterogeneous evidence, including patient histories, clinical guidelines, and trajectories of comparable cases. While large language models (LLMs) offer strong reasoning capabilities, they remain prone to hallucinations and struggle to integrate long, structured medical documents. We present MED-COPILOT, an interactive clinical decision-support system designed for clinicians and medical trainees, which combines guideline-grounded GraphRAG retrieval with hybrid semantic-keyword similar-patient retrieval to support transparent and evidence-aware clinical reasoning. The system builds a structured knowledge graph from WHO and NICE guidelines, applies community-level summarization for efficient retrieval, and maintains a 36,000-case similar-patient database derived from SOAP-normalized MIMIC-IV notes and Synthea-generated records. We evaluate our framework on clinical note completion and medica
arXiv:2603.00460v1 Announce Type: new Abstract: Clinical decision-making requires synthesizing heterogeneous evidence, including patient histories, clinical guidelines, and trajectories of comparable cases. While large language models (LLMs) offer strong reasoning capabilities, they remain prone to hallucinations and struggle to integrate long, structured medical documents. We present MED-COPILOT, an interactive clinical decision-support system designed for clinicians and medical trainees, which combines guideline-grounded GraphRAG retrieval with hybrid semantic-keyword similar-patient retrieval to support transparent and evidence-aware clinical reasoning. The system builds a structured knowledge graph from WHO and NICE guidelines, applies community-level summarization for efficient retrieval, and maintains a 36,000-case similar-patient database derived from SOAP-normalized MIMIC-IV notes and Synthea-generated records. We evaluate our framework on clinical note completion and medical question answering, and demonstrate that it consistently outperforms parametric LLM baselines and standard RAG, improving both generation fidelity and clinical reasoning accuracy. The full system is available at https://huggingface.co/spaces/Cryo3978/Med_GraphRAG , enabling users to inspect retrieved evidence, visualize token-level similarity contributions, and conduct guided follow-up analysis. Our results demonstrate a practical and interpretable approach to integrating structured guideline knowledge with patient-level analogical evidence for clinical LLMs.
Executive Summary
This article presents MED-COPILOT, a clinical decision-support system that integrates structured guideline knowledge with patient-level analogical evidence to support transparent and evidence-aware clinical reasoning. The system combines GraphRAG retrieval with hybrid semantic-keyword similar-patient retrieval and outperforms parametric LLM baselines and standard RAG in clinical note completion and medical question answering tasks. The authors demonstrate a practical and interpretable approach to integrating structured guideline knowledge with patient-level analogical evidence for clinical LLMs. This system has the potential to improve clinical decision-making by providing clinicians and medical trainees with accurate and evidence-based recommendations.
Key Points
- ▸ MED-COPILOT combines GraphRAG retrieval with hybrid semantic-keyword similar-patient retrieval
- ▸ The system outperforms parametric LLM baselines and standard RAG in clinical note completion and medical question answering tasks
- ▸ MED-COPILOT integrates structured guideline knowledge with patient-level analogical evidence
Merits
Strength in Clinical Decision-Support
MED-COPILOT provides accurate and evidence-based recommendations to clinicians and medical trainees, improving clinical decision-making.
Improves Generation Fidelity and Clinical Reasoning Accuracy
The system consistently outperforms parametric LLM baselines and standard RAG in clinical note completion and medical question answering tasks.
Interpretable Approach to Clinical LLMs
MED-COPILOT's approach to integrating structured guideline knowledge with patient-level analogical evidence provides a transparent and evidence-aware clinical reasoning framework.
Demerits
Limited Generalizability
The system's performance may not generalize to diverse clinical settings and patient populations, requiring further evaluation and validation.
Dependence on High-Quality Training Data
The effectiveness of MED-COPILOT relies heavily on the quality and availability of training data, which can be a limitation in resource-constrained settings.
Potential for Over-Reliance on Technology
Clinicians and medical trainees may over-rely on MED-COPILOT's recommendations, potentially compromising their critical thinking and clinical judgment skills.
Expert Commentary
The article presents a significant contribution to the field of clinical decision-support systems, highlighting the potential of MED-COPILOT to improve clinical decision-making. While the system's performance is impressive, further evaluation and validation are necessary to ensure its effectiveness in diverse clinical settings and patient populations. Additionally, the potential for over-reliance on technology is a concern that must be addressed in future research and development. Overall, MED-COPILOT is a promising approach to integrating structured guideline knowledge with patient-level analogical evidence, and its development and evaluation have important implications for the field of medical informatics.
Recommendations
- ✓ Recommendation 1: Future research should focus on evaluating MED-COPILOT's performance in diverse clinical settings and patient populations.
- ✓ Recommendation 2: The development and deployment of MED-COPILOT should involve policy-level considerations to ensure its safe and effective integration into clinical practice.