Machine Learning Predicts First Elective Bowel Resection in Newly Diagnosed Pediatric Crohn Disease Patients

The authors aim to develop a machine learning model to predict the timing of the first elective bowel resection in pediatric patients newly diagnosed with Crohn disease, as existing risk scores do not provide accurate individual predictions. They created Individual Survival Distributions (ISDs) using clinical and laboratory data from a national cohort, achieving a mean prediction error of approximately 755 days for baseline data and 840 days with additional longitudinal data. This innovative approach offers a potential tool for better treatment planning in pediatric Crohn disease patients.