A comparison of different regression and classification methods for predicting the length of hospital stay after cesarean sections

Cesarean section (CS) is one of the main causes of hospitalization in developed countries. Although no benefits have been shown for the mother and baby, the frequency of CS has increased over the past few decades. The control of the length of stay (LOS) for such a widespread procedure therefore becomes strategic for any healthcare facility. The aim of this study is to investigate causes and factors that determine an increase in the LOS in the case of CS delivery. Multiple linear regression analysis and machine learning algorithms are used to build and compare different models for LOS prediction, with the purpose of offering a potential support tool for the planning and programming of CS procedures in healthcare facilities.