Predicting Job Burnout Among Female Nurses in China With Machine Learning and Shapley Additive Explanations

Job burnout among nurses is prevalent globally, particularly in China. However, few studies have been conducted on reliable tools for building predictive models. This cross-sectional study was conducted in four cities of Liaoning Province in China during the period from January to April 2022 by utilizing a self-administered smartphone questionnaire protocol, yielding 1400 responses from female nurses. We applied the least absolute shrinkage and selection operator (LASSO) and Boruta to identify the common predictors of job burnout. We then adopted and optimized three highly applicable machine learning (ML) algorithms-K-nearest neighbor (KNN), EXtreme Gradient Boosting (XGBoost), and random forest (RF)-to predict job burnout among female nurses. The values of area under curve (AUC) of KNN, RF, and XGBoost ML models were 0.85-0.95, with XGBoost performing best (AUC = 0.939). In addition, Shapley additive explanations (SHAP) were used to show the contribution of each predictor to the predicted outcomes. The result confirmed the role of consistency, perceived stress, and physical fatigue as key protective factors, and consistency exhibited interactions with perceived stress, organizational support, psychological detachment, and sense of control to nurses' job burnout. This helps identify nurses at risk of job burnout and provide targeted strategy to alleviate nurses' job burnout.