medIKAL: Integrating Knowledge Graphs as Assistants of LLMs for Enhanced Clinical Diagnosis on EMRs
Abstract: Electronic Medical Records (EMRs), while integral to modern healthcare, present challenges for clinical reasoning and diagnosis due to their complexity and information redundancy. To address this, we proposed medIKAL (\textbf{I}ntegrating \textbf{K}nowledge Graphs as \textbf{A}ssistants of \textbf{L}LMs), a framework that combines Large Language Models (LLMs) with knowledge graphs (KGs) to enhance diagnostic capabilities. medIKAL assigns weighted importance to entities in medical records based on their type, enabling precise localization of candidate diseases within KGs. It innovatively employs a residual network approach, allowing initial diagnosis by the LLM without external knowledge before merging with KG search results. A path-based reranking algorithm and a specialized prompt template further refine the diagnostic process. We validated medIKAL's effectiveness through extensive experiments on a newly introduced open-sourced Chinese EMR dataset, demonstrating its potential to improve clinical diagnosis and decision-making in real-world settings.
Paper Type: Long
Research Area: NLP Applications
Research Area Keywords: healthcare applications, clinical NLP, knowledge graphs
Languages Studied: Chinese
Submission Number: 4293
Loading