Characterizing fall risk factors in Belgian older adults through machine learning: a data-driven approach

Area of Science:

• Gerontology
• Public Health
• Biomedical Informatics

Background:

• Falls are a significant health concern for the aging population.
• Existing fall-risk screening tools lack predictive accuracy.
• There is a need for effective fall-risk classification models for community-dwelling older adults.

Purpose of the Study:

• To identify risk factors for falls in older adults.
• To develop a fall-risk classification algorithm.
• To incorporate biological, behavioral, environmental, and socioeconomic factors into fall-risk assessment.

Main Methods:

• A quality-of-life questionnaire was administered to 82,580 older adults.
• 139 selected questions were analyzed for their association with fall incidence.
• A random forest classifier was trained to determine feature importance.

Main Results:

• Twenty-four fall risk factors were identified and included in the classification model.
• These factors comprised 2 biological, 8 behavioral, 11 socioeconomic, and 3 environmental variables.
• Each identified factor contributed 4.5% to 6.5% to explaining fall risk.

Conclusions:

• Machine learning identified 24 significant fall risk factors in older adults.
• Maintaining an active lifestyle and satisfaction with living situation are key to reducing fall risk.
• Further research is needed to develop a practical screening tool for daily use.