CARE-LLM: Interactive Dashboard for Medical Information Support
- Partner
- Fraunhofer FIT & University Hospital of Cologne
- Period
- Aug 2024 – Dec 2024
- Focus
- Medical Decision Support
Key outcomeInteractive research demonstrator for patient-facing medical explanations, not yet a clinically deployed or clinically evaluated system


Context
Patients living with chronic kidney disease often need to understand complex medical information, while clinicians have limited time to explain it after discharge.
Inputs
- Medical reports and structured clinical data
- Patient-centered information requirements
Approach
- Designed a multi-agent LLM framework to translate complex clinical information into plain language
- Developed an interactive dashboard focused on patient comprehension, with a chat assistant for follow-up questions
- Explored patient-profile-based explanation, adapting depth and structure to the person reading it
Outcomes
- Interactive research demonstrator for patient-facing medical explanations, not yet a clinically deployed or clinically evaluated system
- Simplifying information does not mean removing important information; a useful explanation keeps its meaning while adapting language, structure, and depth to the reader