Text this: Enhancing clinical decision-making in closed pelvic fractures with machine learning models