Machine Learning Techniques, Applications, and Potential Future Opportunities in Pressure Injuries (Bedsores)...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Clinical Research Synthesis

Background:

• Pressure injuries (PI) represent a significant healthcare challenge, particularly for patients in acute or long-term care settings.
• Machine learning (ML) has emerged as a valuable tool for managing PI, with previous reviews highlighting its application in numerous studies.
• This research provides a comprehensive, updated summary of ML's role in PI management from January 2007 to July 2022.

Approach:

• A systematic review adhering to PRISMA guidelines was conducted, searching major databases (PubMed, Web of Science, Scopus, Science Direct) and other resources.
• Ninety eligible studies were identified and categorized based on the timing of PI occurrence: before (48%), at (16%), and after (36%) the event.
• Studies were further classified by medical specialty (sixteen identified), analyzing ML methods, data inputs, and outputs within each.

Key Points:

• ML applications in PI management span across diverse medical specialties, offering tailored solutions.
• Deep learning techniques and hybrid models show particular promise for advanced PI prediction and management.
• Integrating existing risk assessment tools with ML, leveraging Electronic Health Records (EHR), enhances clinical decision-making.

Conclusions:

• This review consolidates current ML contributions to PI management, offering a structured overview of methods and applications.
• Identified gaps and opportunities highlight the potential for further research and development in ML-driven PI care.
• The findings support the growing role of ML in improving patient outcomes and optimizing healthcare delivery for pressure injuries.