Proof of Concept of the Aiinane Preoperative Risk Assessment Tool.

BACKGROUND: Each year, over 13,000 patients in Catalonia and more than 4 million worldwide experience last-minute surgery cancellations (LMC) due to preoperative inefficiencies. As anaesthesia services struggle to meet surgical demands, thorough preoperative evaluations become challenging. Current resource-intensive pre-anaesthetic assessments are undermined by high demand, causing inefficiency. However, proper assessment identifies that most patients (\>70%) are low-risk and ensures high-risk patients are adequately prepared by analyzing risk profiles, health status, medical history, treatment, and lab results. RATIONALE: Previous attempts to improve preoperative risk assessment have mainly relied on self-administered questionnaires to detect at-risk patients. The investigators have identified a care model that enhances quality by adding value to preoperative risk assessment. By combining anaesthesiology expertise with AI techniques, the investigators developed an automated digital environment to detect risks, optimize visits, avoid cancellations, and reduce postoperative complications. This system uses parameterized medical knowledge to verify responses and objectively assess patient risk by integrating multiple data sources. The investigators have developed a Class IIa active diagnostic and monitoring product, a Medical Device Software (MDSW, EMDN V92), to support clinical decision-making in an automated digital preoperative environment. It helps assess patients and flags low-risk from medium/high-risk individuals, reporting personalized needs to the medical team. The software was developed exclusively by independent researchers from Bellvitge University Hospital and the Bellvitge Biomedical Research Institute (IDIBELL). HYPOTHESIS: The tool has been tested successfully on fictitious patients in controlled preclinical scenarios. The aim now is a proof-of-concept study to verify its performance with real patients in uncontrolled, real-world settings. This is a