Construction of disability risk prediction model for the elderly based on machine learning

Area of Science:

• Gerontology
• Artificial Intelligence in Healthcare
• Public Health

Background:

• Assessing disability risk in older adults is crucial for timely interventions.
• Existing methods may lack the predictive power to identify individuals at high risk.
• Novel tools are needed to support healthcare professionals in geriatric care.

Purpose of the Study:

• To develop and validate machine learning models for predicting disability risk in older adults.
• To identify key predictors associated with disability onset in this population.
• To provide a novel tool for healthcare professionals to assess and manage disability risk.

Main Methods:

• Utilized data from the China Health and Retirement Longitudinal Study (2018, 2020 waves).
• Employed five machine learning algorithms (Random Forest, Extreme Gradient Boosting, etc.) to build predictive models.
• Applied Shapley Additive Explanations (SHAP) to determine independent predictors of disability risk.

Main Results:

• 21.9% of participants developed disability during follow-up.
• Random Forest (RF) and Extreme Gradient Boosting (XGBoost) models demonstrated superior performance (F1 scores 0.92, 0.86; accuracies 0.92, 0.85).
• Key predictors included self-rated health, education, sleep, alcohol use, depressive symptoms, hypertension, and arthritis.

Conclusions:

• Machine learning models, particularly RF and XGBoost, show strong predictive capability for disability risk in older adults.
• These models can serve as valuable tools in clinical and public health settings for early risk assessment.
• Further validation and exploration of these AI-driven tools are warranted for practical implementation.