Machine Learning Algorithms to Study Features Affecting the Length of Stay in Patients with Lower Limb Fractures: A Bicentric Study

The estimation of the length of stay (LOS) can support the analysis of practitioners for improving the hospital efficiency. For the appropriate management of beds and to reduce costs, it is necessary to analyse and evaluate procedures for reducing hospital stays. For patients with orthopedic trauma, especially with lower limb fractures, LOS becomes an important parameter to estimate. The aim of this study is to study LOS for all patients with lower limb fractures in the University Hospital “San Giovanni di Dio and Ruggi d’Aragona” of Salerno and A.O.R.N. “Antonio Cardarelli” of Naples. The analysis was conducted by implementing different well-known Artificial Intelligence models, whose accuracy for the Hospital “San Giovanni di Dio and Ruggi d’Aragona” was 92,0% (NB algorithm) while for the A.O.R.N. “A. Cardarelli” the best performance was obtained with RF algorithm with an accuracy of 95,92%. These results obtained from the estimation of LOS show promise and therefore can be a valuable help in the management process of the hospital.