Knowledge Discovery With Machine Learning for Hospital-Acquired Catheter-Associated Urinary Tract Infections.

Area of Science:

• Nursing Informatics
• Health Data Science
• Clinical Informatics

Background:

• The proliferation of health data offers opportunities for big data science to advance healthcare insights.
• Nursing research using big data analytics is crucial for evidence-based practice but remains limited.
• Identifying factors for hospital-acquired infections requires sophisticated data analysis techniques.

Purpose of the Study:

• To explain a knowledge discovery and data mining approach for analyzing nursing data.
• To identify factors associated with hospital-acquired catheter-associated urinary tract infections (CAUTI).
• To demonstrate the application of machine learning techniques in nursing research.

Main Methods:

• Utilized a knowledge discovery and data mining approach.
• Integrated multiple data sources, including electronic health records and nurse staffing data.
• Employed three machine learning techniques: decision trees, logistic regression, and support vector machines.

Main Results:

• Decision tree models identified interpretable rules for factors associated with CAUTI.
• Logistic regression models pinpointed specific factors increasing the risk of CAUTI.
• Support vector machines were used for performance comparison with interpretable models.

Conclusions:

• Machine learning approaches can effectively analyze complex nursing data from multiple sources.
• This study advances the secondary use of electronic health records for clinical insights.
• Findings provide essential evidence to guide nursing professionals in preventing CAUTI.