Development of Fall Risk Classification Models for Community-Dwelling Older Adults using Latent Class Analysis and...

Area of Science:

• Gerontology
• Public Health
• Data Science

Background:

• Community-dwelling older adults face significant fall risks.
• Accurate identification of fall-risk groups is crucial for effective prevention.

Purpose of the Study:

• To identify distinct fall-risk groups among community-dwelling older adults in South Korea.
• To develop a classification model for investigating factors associated with fall risk.

Main Methods:

• Utilized data from 9,231 older adults from the 2020 Korea Elderly Survey.
• Employed latent class analysis to define fall-risk groups based on fall indicators.
• Developed classification models (XGBoost) to predict group membership.

Main Results:

• A three-class model (low, moderate, high fall risk) demonstrated optimal interpretability and fit.
• The XGBoost model achieved high performance (accuracy=0.70, F1-score=0.68).
• Significant risk factors included self-rated health, cognitive function, healthcare utilization, and need for assistance with daily living.

Conclusions:

• The study successfully differentiated fall-risk levels, supporting a preventive healthcare approach.
• Identified key risk factors provide a foundation for developing personalized fall prevention programs for older adults.