Support Vector Machine-Based Prediction Model for Healthcare Workforce Transition Success Under Decentralization

Area of Science:

• Health Services Research
• Health Workforce Management
• Machine Learning in Healthcare

Background:

• Healthcare decentralization presents challenges for workforce transitions.
• Understanding factors influencing successful transitions is vital for effective health system reform.
• Previous research has not extensively utilized machine learning for this specific prediction task.

Purpose of the Study:

• To develop and validate predictive models for healthcare workforce transition success under decentralization.
• To identify key determinants of successful transitions using Support Vector Machine (SVM) analysis.
• To compare the performance of SVM with other machine learning algorithms.

Main Methods:

• A cross-sectional study involving 430 healthcare personnel undergoing decentralization in Thailand.
• Analysis of 37 predictors across demographics, benefits, and welfare domains.
• Application of SVM with four kernel functions, 10-fold cross-validation, and Synthetic Minority Oversampling Technique (SMOTE) for class imbalance.

Main Results:

• The SVM model with a linear kernel achieved the best performance (accuracy: 69%).
• SMOTE improved sensitivity to 54% while maintaining specificity at 79%.
• Key predictors identified include competitive compensation, career development, fair promotion, hazardous work compensation, and educational leave.

Conclusions:

• This study is the first to apply machine learning to predict healthcare workforce transition success in decentralization.
• The SVM model effectively identified critical factors influencing workforce transitions.
• Findings offer evidence-based guidance for healthcare administrators managing workforce transitions during health system reforms.