FEASIBILITY IN USING DE-IDENTIFIED PATIENT DATA TO ENRICH ARTIFICIAL APPLICATIONS IN MEDICAL EDUCATION

Medical schools are working to provide high-quality medical education (ME) that will support students in attaining better learning outcomes while ensuring an increased quality of care and improved patient safety. The use of Artificial Intelligence (AI) in ME is growing, with a specific focus on curriculum, learning support and assessment of learning. Further development of AI in ME needs more and richer data sets. We hypothesized that clinical data from the healthcare sector might enrich educational data by providing untapped big data sets with contextual relevance concerning the patient population and health system. The aim of this study was therefore to evaluate the value of using real clinical data in AI-driven medical education applications and explore use cases enabled by access to data from the healthcare sector. Over five million de-identified patient records from New York State hospitals were used in this study. Characterization of the clinical data identified new data at the population and patient levels. The identified use cases were an enriched medical education data warehouse, bridging the gap between academic and clinical practices, addressing new learning outcomes in a medical curriculum, more opportunities for AI research in medical education and the exploration of AI in continuing ME.