Decoding machine learning in nursing research: A scoping review of effective algorithms.

Area of Science:

• Nursing
• Artificial Intelligence
• Machine Learning

Background:

• Artificial intelligence (AI) is rapidly transforming healthcare, impacting nursing roles and necessitating research into AI-integrated systems.
• This scoping review specifically examines machine learning (ML) applications within the nursing field.

Purpose of the Study:

• To investigate machine learning (ML) algorithms used in nursing research.
• To identify common model evaluation methods and areas of focus.
• To determine the most effective ML algorithms in nursing.

Main Methods:

• A scoping review was conducted following PRISMA-ScR guidelines.
• A systematic search of seven major databases was performed.
• Study quality was assessed using the Medical Education Research Study Quality Instrument (MERSQI).

Main Results:

• Twenty-six articles (2019-2023) were reviewed, with 46% from the US; average MERSQI score indicated moderate- to high-quality studies.
• Random Forest was the most utilized ML algorithm, followed by logistic regression, LASSO, decision trees, and SVMs.
• Common evaluation metrics included sensitivity, specificity, accuracy, ROC, AUROC, and precision. Half of the studies focused on nursing staff/students and hospital readmissions/ED visits.

Conclusions:

• Machine learning (ML) shows significant clinical relevance for nurses, nurse practitioners, and administrators, confirming its benefits in healthcare.
• The review highlights the growing importance of ML in nursing research and its implications for patient care and resource management.
• Recommendations include using experimental designs to optimize ML model application across nursing domains.