Prediction of Nursing Workload in Hospital

Area of Science:

• Nursing
• Health Informatics
• Data Science

Background:

• Nursing workload is a critical factor in hospital resource allocation and patient care quality.
• Predicting nursing workload accurately is essential for effective human resource planning and health policy.
• Identifying nursing-sensitive patient characteristics can improve workload prediction models.

Purpose of the Study:

• To investigate which nursing-sensitive patient characteristics can predict nursing workload.
• To compare the effectiveness of different machine learning methods for predictor identification.
• To model nursing workload using identified predictors for prognostic determination.

Main Methods:

• Utilized Support Vector Machine, Random Forest, and Gradient Boosting for predictor identification.
• Employed Feature Importance to compare the results of the initial predictive models.
• Applied Multinomial Logistic Regression to model nursing workload based on identified predictors.
• Data mining techniques were employed for the analysis.

Main Results:

• Identified specific nursing-sensitive patient characteristics as significant predictors of nursing workload.
• Demonstrated the utility of machine learning algorithms in uncovering these predictive characteristics.
• Preliminary results from the data mining process indicate successful predictor identification and modeling.

Conclusions:

• Prognostic determination of nursing workload is feasible through data-driven approaches.
• The identified patient characteristics and modeling methods can inform hospital staffing and health policy.
• This research contributes to optimizing resource management within healthcare settings.