Systematic Review for Risks of Pressure Injury and Prediction Models Using Machine Learning Algorithms.

Area of Science:

• Medical Informatics
• Clinical Nursing
• Artificial Intelligence in Healthcare

Background:

• Pressure injuries (PIs) represent a growing global health concern with significant patient morbidity and healthcare costs.
• Current methods for identifying pressure injury risks lack the precision and timeliness needed for effective prevention in hospitalized adults.
• There is a critical need for advanced methodologies to predict pressure injuries earlier, before visible skin damage occurs.

Purpose of the Study:

• To systematically review and evaluate machine learning (ML) algorithms used in developing prediction models for pressure injuries in adult hospitalized patients.
• To assess the efficacy of ML-based prediction models in identifying pressure injury risks at an earlier stage.
• To analyze the strengths and limitations of existing ML models for pressure injury prediction and guide future research.

Main Methods:

• A systematic review methodology was employed, searching major databases including CINAHIL, PubMed, Science Direct, IEEE, Cochrane, and Google Scholar.
• Studies included were those that constructed a pressure injury prediction model using machine learning algorithms for adult hospitalized patients.
• Twenty-seven relevant articles were selected and critically analyzed based on predefined inclusion criteria.

Main Results:

• Machine learning algorithms show promise in developing predictive models for pressure injuries in hospitalized adults.
• Evidence suggests these ML-based models can identify patients at risk of pressure injuries earlier compared to traditional assessment tools.
• The reviewed studies indicate variability in model performance, necessitating further research into optimal algorithm selection and validation.

Conclusions:

• Machine learning offers a promising avenue for enhancing the early detection and prevention of pressure injuries in hospitalized patients.
• The critical analysis of current models underscores the need for robust validation and standardization of ML approaches in this field.
• Future research should focus on refining ML algorithms and integrating them into clinical workflows to mitigate the incidence of pressure injuries.