Regression and Machine Learning Algorithm to Study the LOS of Patients Undergoing Hip Surgery

Today, fracture surgery is a key part of a hospital’s orthopaedic department and usually involves significant clinical cost implications. The evaluation of hospitalization time for subjects suffering of hip fracture assumes a key role in the last years because it can affect the postoperative course and recovery of the patient. Length of stay (LOS) is a useful tool for monitoring patients and useful for hospital administrators to assess the efficiency of the hospital. Our aim is to investigate the LOS prediction for all patients with hip fracture hospitalized in two hospitals located in Campania Region, also comparing the obtained results. Different machine learning models and data analysis methodologies have been applied on a cohort of patients hospitalized in two different hospitals, also evaluating them in terms of accuracy and error.