Predicting postoperative delirium after hip arthroplasty for elderly patients using machine learning

Abstract

Abstract Background: Postoperative delirium (POD) is a common and severe complication after hip arthroplasty for elderly patients. We aim to develop and validate a machine learning method that determines essential features related to postoperative delirium and predicts POD after hip arthroplasty for elderly patients. Methods: We reviewed preoperative and intraoperative clinical data and laboratory tests of hip arthroplasty elderly patients between January 2017 and April 2021 in Orthopedics of First Affiliated Hospital of Chongqing Medical University. The Confusion Assessment Method (CAM) was administered to the patients in their perioperative period. Machine learning algorithms were trained to predict the POD and determine leading features. The predictive performance was evaluated using the area under the curve (AUC), accuracy (ACC), sensitivity, specificity, and F1-score. Results: 476 arthroplasty elderly patients (POD = 86, non-POD = 390) with general anesthesia were included in this study, and the combination of feature selection method mutual information (MI) and linear binary classifier using logistic regression (LR) achieved an encouraging performance (AUC = 0.94, ACC = 0.88, sensitivity = 0.85, specificity = 0.90, F1-score = 0.87) on a balanced test dataset. Conclusion: The machine learning (ML) model could predict POD for arthroplasty elderly patients with satisfying accuracy and revealed the major risk factors of suffering POD such as age, Cystatin C, GFR, CHE, CRP, LDH, monocyte count (MONO), history of mental illness or psychotropic drug use and intraoperative blood loss. Proper preoperative interventions for these factors could assist clinicians in reducing the incidence of POD in arthroplasty elderly patients.