A prediction model for patient classification according to nursing need: Using data mining techniques

Area of Science:

• Nursing Informatics
• Machine Learning in Healthcare
• Oncology Nursing

Background:

• Accurate patient classification is crucial for effective nursing resource allocation.
• Hospitalized cancer patients present complex and varied nursing care needs.
• Data mining offers potential for developing predictive models in clinical settings.

Purpose of the Study:

• To construct a predictive model for classifying hospitalized cancer patients based on their nursing needs.
• To evaluate the efficacy of different data mining techniques for patient classification.

Main Methods:

• Utilized three data mining techniques: logistic regression, decision tree, and neural network.
• Applied these methods to classify hospitalized cancer patients according to nursing requirements.
• Validated model performance using Receiver Operating Characteristic (ROC) curve analysis.

Main Results:

• Neural network demonstrated superior prediction power compared to logistic regression and decision tree.
• The neural network-based prediction model achieved an accuracy of 84.06% for patient classification.
• ROC curve verification confirmed the enhanced predictive capability of the neural network model.

Conclusions:

• Neural networks are effective for developing accurate patient classification models in nursing.
• The developed model can aid in optimizing nursing care for hospitalized cancer patients.
• Data mining techniques, particularly neural networks, hold significant promise for improving nursing practice and patient outcomes.