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

Area of Science:

• Nursing
• Occupational Health
• Data Science

Background:

• Job burnout is a significant global issue affecting nurses, with limited research on predictive modeling tools, especially in China.
• Understanding predictors of burnout is crucial for developing effective interventions to support nursing staff.

Purpose of the Study:

• To identify common predictors of job burnout among female nurses in China.
• To develop and optimize machine learning (ML) models for predicting nurse job burnout.
• To explore the interactions of key factors influencing job burnout.

Main Methods:

• A cross-sectional study involving 1400 female nurses in China using smartphone questionnaires (January-April 2022).
• Application of LASSO and Boruta for predictor identification.
• Optimization and evaluation of K-nearest neighbor (KNN), Random Forest (RF), and Extreme Gradient Boosting (XGBoost) ML algorithms.
• Utilizing Shapley Additive Explanations (SHAP) for predictor contribution analysis.

Main Results:

• XGBoost demonstrated the highest predictive performance for job burnout with an Area Under the Curve (AUC) of 0.939, within a range of 0.85-0.95 for all tested ML models.
• Consistency, perceived stress, and physical fatigue were identified as significant protective factors against job burnout.
• Consistency showed interactive effects with perceived stress, organizational support, psychological detachment, and sense of control.

Conclusions:

• Machine learning models, particularly XGBoost, can effectively predict job burnout in female nurses.
• Identifying key protective factors and their interactions provides a basis for targeted strategies to mitigate nurse burnout.
• The findings offer valuable insights for healthcare institutions to support nursing well-being and reduce burnout rates.