Machine Learning Approaches for the Frailty Screening: A Narrative Review

Area of Science:

• Gerontology
• Biomedical Informatics
• Artificial Intelligence in Healthcare

Background:

• Frailty is a condition of increased vulnerability in older adults, leading to adverse health outcomes.
• Early screening and management of frailty are crucial for improving elderly health and societal well-being.
• Translating controlled research findings into dynamic clinical settings presents significant challenges.

Purpose of the Study:

• To conduct a narrative review of frailty screening procedures.
• To focus on innovative tools, indicators, and machine learning (ML) approaches for frailty detection.
• To assess the potential of ML in identifying frailty risk factors.

Main Methods:

• A narrative review methodology was employed.
• Searched for studies on frailty screening using innovative tools and ML.
• Six selected studies were analyzed, focusing on ML techniques and indicators used.

Main Results:

• Support vector machine was the most frequently utilized ML method in the reviewed studies.
• ML approaches demonstrated the ability to identify risk factors for predicting pre-frailty or frailty.
• Identified limitations include data quality issues, which can impact model performance.

Conclusions:

• Machine learning methods hold substantial potential for early frailty detection.
• Further research addressing data quality is needed to optimize ML applications in clinical frailty screening.
• ML tools can significantly aid in the proactive management of frailty in the elderly population.