Harnessing Machine Learning to Predict Nurse Turnover Intention and Uncover Key Predictors: A Multinational...

Area of Science:

• Nursing Workforce Research
• Health Informatics
• Predictive Analytics in Healthcare

Background:

• Nurse turnover is a significant issue impacting healthcare systems globally.
• Identifying predictors of turnover intention is crucial for developing effective retention strategies.
• Machine learning offers advanced capabilities for analyzing complex healthcare data.

Purpose of the Study:

• To predict nurses' turnover intention using machine learning (ML).
• To identify key psychosocial, organizational, and demographic predictors of turnover intention.
• To compare ML model performance across three countries: the United States, Türkiye, and Malta.

Main Methods:

• Cross-sectional, multinational survey of 1625 nurses.
• Assessed 20 variables including job satisfaction, psychological safety, and work engagement.
• Employed six ML algorithms (logistic regression, random forest, XGBoost, etc.) for prediction and feature importance analysis.

Main Results:

• Logistic regression demonstrated the highest predictive performance (AUC=0.890).
• Job satisfaction was the most influential predictor across all models.
• Other significant predictors included country (USA), work experience, depression, and psychological safety.

Conclusions:

• Machine learning effectively predicts nurse turnover intention using multidimensional data.
• This data-driven framework supports targeted retention strategies and enhances organizational stability.
• The findings provide actionable insights for healthcare leaders to improve nurse retention.