ColaCare introduces a novel framework that bridges the gap between structured EHR data and text-based reasoning through collaborative AI agents inspired by clinical multidisciplinary teams.

• Integrates domain-specific expert models with LLMs to improve EHR analysis
• Employs specialized DoctorAgents and a MetaAgent for collaborative patient data analysis
• Mimics the clinical Multidisciplinary Team (MDT) approach used by healthcare professionals
• Enhances prediction of critical healthcare outcomes such as mortality and readmission

This research matters for healthcare by creating more accurate, context-aware clinical decision support systems that can interpret complex patient data while maintaining the human-like reasoning that medical professionals rely on.

ColaCare: Enhancing Electronic Health Record Modeling through Large Language Model-Driven Multi-Agent Collaboration