Prof. Dr. med. Alexander Ghanem (Asklepios Klinik Nord) Study title: AI-augmented perioperative clinical decision support Despite medical advancements, perioperative mortality following surgery has been at 0.4–0.8% for ten years. Experts expect that mortality and permanent harm could be reduced by improved perioperative risk management. Clinical guidelines support decision-making by specifying risk parameters and calculating scores for the risk of harm. Within this context, the KIPeriOP project is addressing the following questions:
- Can the use of guideline-based clinical decision support (CDS) software in day-to-day clinical practice help to enable reliable and structured recording of patient data (explicit risk factors and additional routine parameters) and therefore improve the documentation quality?
- Is the prediction of postoperative harm by a predictive AI model developed on the basis of patient data recorded using CDS tools more reliable than established risk scores? Can such an internally validated model be used to derive new scientific hypotheses for perioperative risk management (e.g. new factors that are predictive for postoperative harm) that are not yet included in the existing guidelines?
How can the AI predictive model be integrated in the CDS tool in an ethically acceptable way? Can the reliability and transparency of the AI predictive model and the validity of the generated hypotheses be confirmed again with regard to the recording region and recording period of different patient populations? The study is being conducted at the following sites: Prof. Dr. med. Alexander Ghanem, Asklepios Klinik Nord Funding: German Federal Ministry of Health (BMG).