Multimorbidity Patterns and Depression: Bridging Epidemiological Associations with Predictive Analytics for Risk...

Area of Science:

• Gerontology
• Psychiatry
• Epidemiology

Background:

• Late-life depression is a major public health issue, especially in older adults with multiple chronic conditions (multimorbidity).
• Current research often overlooks the complex interactions between coexisting diseases, hindering targeted prevention.
• Understanding multimorbidity's impact on depression risk requires advanced analytical approaches.

Purpose of the Study:

• To identify distinct multimorbidity patterns in older adults.
• To analyze the association between these patterns and depression.
• To evaluate predictive models for depression risk stratification using machine learning.

Main Methods:

• Latent class analysis (LCA) was used to cluster multimorbidity patterns in the China Health and Retirement Longitudinal Study (CHARLS).
• Multivariable logistic regression examined associations between patterns and depression.
• XGBoost machine learning assessed predictive performance and interaction effects.

Main Results:

• Four multimorbidity patterns were identified: cardio-metabolic, digestive-joint, respiratory, and cardiovascular-digestive.
• All patterns were significantly associated with depression; the cardiovascular-digestive pattern had the strongest association (OR=4.56).
• The digestive-joint pattern showed the highest predictive effect for depression, amplified by low income, limited education, female gender, and rural residence.

Conclusions:

• This study integrates epidemiological data with predictive analytics for depression risk stratification in older adults with multimorbidity.
• Routine depression screening is recommended for individuals with cardiovascular-digestive diseases and prioritized for women with digestive-joint diseases.
• Low-income and rural older adults with chronic conditions require heightened clinical attention for depression prevention.