MedCollab: IBIS-Guided Multi-Agent Collaboration with Hierarchical Disease Relation Chains for Clinical Diagnosis

MedCollab is a novel multi-agent framework designed to enhance clinical diagnosis and report generation by integrating specialist and examination agents based on patient records. It employs an Issue-Based Information System (IBIS) protocol for structured deliberation and constructs Hierarchical Disease Relation Chains (HDRC) to connect hypotheses through disease relationships. Experimental results on ClinicalBench and MIMIC-IV demonstrate that MedCollab significantly outperforms existing LLMs and medical multi-agent systems in diagnostic accuracy and reasoning quality, highlighting its potential for improving clinical decision-making processes.