Related Experiment Videos
Charting and Predicting Risk: Artificial Intelligence/Machine Learning Pilot Model for Hospital-Acquired Pressure
Shea Polancich1, Chris Hickman, Tracey Dick
1Author Affiliations: Department of Family, Community, and Health Systems (Drs Polancich and Bordelon), Department of Acute, Chronic, and Continuing Care, School of Nursing, University of Alabama at Birmingham (UAB) (Mr Hickman and Dr Dick), and Department of Health Services Administration, UAB School of Health Professions, Birmingham, Alabama (Drs Hall and Hearld).
Journal of Nursing Care Quality
|June 19, 2026
Summary
Artificial intelligence and machine learning (AI/ML) models show increased precision in predicting hospital-acquired pressure injuries (HAPIs). These advanced AI/ML models outperformed traditional methods in identifying patients at risk for HAPIs.
Area of Science:
- Healthcare Informatics
- Clinical Predictive Modeling
- Artificial Intelligence in Medicine
Background:
- Hospital-acquired pressure injuries (HAPIs) affect millions globally each year.
- HAPIs represent a significant patient safety concern and incur substantial healthcare costs.
Purpose of the Study:
- To assess the efficacy of artificial intelligence and machine learning (AI/ML) in predicting HAPIs.
- To compare the predictive accuracy of AI/ML models against traditional analytical approaches for HAPIs.
Main Methods:
- A pilot study utilizing secondary data analysis.
- Development of a training dataset for comparative evaluation of predictive models.
- Exploratory analysis comparing AI/ML models with traditional methods for pressure injury prediction.
Main Results:
- Tree-based AI/ML models demonstrated superior performance in predicting HAPIs compared to logistic regression.
- Logistic regression offered a reasonable fit with interpretable coefficients, but AI/ML models achieved higher predictive accuracy.
Conclusions:
- AI/ML models offer enhanced precision for identifying patients at risk of developing HAPIs.
- The findings suggest a promising role for AI/ML in improving patient outcomes by enabling early risk identification.