Regression and Machine Learning analysis to predict the length of stay in patients undergoing hip replacement surgery

Abstract

Hip fracture is a serious injury associated with adverse outcomes, including mortality. It occurs mainly in older patients and is often associated with other collateral pathologies. Its treatment generally involves a surgical operation. This could lead to various complications due to long-term hospitalization and low motor capacity. All these cause complications that extend far beyond the orthopaedic injury, with negative impacts on the patient’s quality of life and health care economics. In this contest, the main goal of our work was to identify some of the most relevant parameters to take in account for the treatment of patients with hip fracture. Our analysis involved the 456 patients who were operated on fracture to the hip in 2019 and 2020 at the Complex Operative Unit (C.O.U.) of Orthopaedic and Traumatology of the University Hospital "San Giovanni di Dio e Ruggi d’Aragona" of Salerno. Through the implementation of various algorithms, our aim was to formulate a specific model that could best predict the target value of patients.