Developing AI-Supported Decision Support Systems in Intensive Care Medicine
Intensivmedizin: Mit KI-Unterstützung zuverlässiger entscheiden - Juraforum
Summary
Researchers from Charité and the University Hospital Essen are developing an AI-supported decision support system for intensive care units. The system aims to detect deteriorating conditions early and optimize antibiotic use by analyzing not only structured numerical data but also unstructured text data, such as doctor's notes.
Details
This project, named REACT (Routine Data Enhanced Analytics for Clinical Trajectories), utilizes one of Europe's largest harmonized intensive care data collections, based on approximately 200,000 patient stays. While traditional analysis relied mainly on structured numerical data and fixed thresholds, this system is innovative because it includes unstructured data (such as doctor's notes and progress reports) in its analysis. The data is converted into a unified format using the internationally established interoperability standard, FHIR (Fast Healthcare Interoperability Resources). The initial tools developed include two dashboards: 1) a dashboard supporting the appropriate use of antibiotics based on existing guidelines, and 2) a dashboard providing early warnings of impending circulatory instability. This technical infrastructure is planned to be released as open-source software after a clinical pilot phase, allowing other hospitals to utilize the developed solutions.
Original content copyright by respective publishers