To Predict the Length of Hospital Stay After Total Knee Arthroplasty in an Orthopedic Center in China: The Use of Machine Learning Algorithms

Abstract

Background and Objectives: Total knee arthroplasty (TKA) is widely performed to improve mobility and quality of life for symptomatic knee osteoarthritis patients. The accurate prediction of patients length of hospital stay (LOS) can help clinicians for rehabilitation decision-making and bed assignment planning, which thus makes full use of medical resources. Methods: Clinical characteristics were retrospectively collected from 1,298 patients who received TKA. A total of 36 variables were included to develop predictive models for LOS by multiple machine learning (ML) algorithms. The models were evaluated by the receiver operating characteristic (ROC) curve for predictive performance and decision curve analysis (DCA) for clinical values. A feature selection approach was used to identify optimal predictive factors. Results: The areas under the ROC curve (AUCs) of the nine models ranged from 0.710 to 0.766. All the ML-based models performed better than models using conventional statistical methods in both ROC curves and decision curves. The random forest classifier (RFC) model with 10 variables introduced was identified as the best predictive model. The feature selection indicated the top five predictors: tourniquet time, distal femoral osteotomy thickness, osteoporosis, tibia component size, and post-operative values of Hb within 24 h. Conclusions: By analyzing clinical characteristics, it is feasible to develop ML-based models for the preoperative prediction of LOS for patients who received TKA, and the RFC model performed the best.