From Bits to Bedside: Machine Learning at the Point-of-Care
, University Hospital Essen
, University Medicine Essen
The implementation of knowledge gained from research to improve clinical care is known as the bench-to-bedside approach. In a similar fashion, we can bring machine learning to the point-of-care. This bits-to-bedside approach is a central goal of the Institute for Artificial Intelligence in Medicine (IKIM) at the University Hospital Essen in Germany. The evaluation of machine learning models in the clinical context requires access to real-world clinical data in real time. This is realized by our IKIM data integration group, originally founded in the radiology department over 10 years ago. In addition to data access, the deployment of trained algorithms, a seamless workflow integration (UX), explainability, and user trust are all mission-critical. To close the loop between research and the point-of-care, we created mechanisms like a clinical chatbot. This not only displays patient data, but also triggers ML workflows. We'll present examples of our AI-powered SMART hospital ecosystem and workflow.