Nursing workload: use of artificial intelligence to develop a classifier model

Area of Science:

• Nursing Informatics
• Artificial Intelligence in Healthcare
• Machine Learning Applications

Background:

• Accurate nursing workload assessment is crucial for optimal patient care and resource allocation.
• Traditional methods of nursing workload assessment can be time-consuming and subjective.
• The integration of artificial intelligence offers potential for more objective and efficient workload prediction.

Purpose of the Study:

• To develop and validate a predictive nursing workload classifier model.
• To utilize artificial intelligence (AI) and machine learning (ML) for workload prediction.
• To identify key variables from electronic patient records that predict nursing workload.

Main Methods:

• Retrospective observational study utilizing electronic patient records.
• Machine learning algorithms applied to a dataset of 43,871 nursing assessments and 11,774 patient records.
• Data analysis performed on the Dataiku® data science platform for exploratory, descriptive, and predictive analysis.

Main Results:

• An AI-enabled classifier model for nursing workload was successfully developed.
• The model achieved 72% accuracy in classifying variables contributing to workload prediction.
• The area under the Receiver Operating Characteristic curve was 82%, indicating good predictive performance.

Conclusions:

• It is feasible to train AI algorithms using electronic health record data to predict nursing workload.
• AI tools demonstrate effectiveness in automating the nursing workload assessment process.
• This predictive model has the potential to enhance efficiency and accuracy in healthcare resource management.