Analysis of the influential factors of depression in the elderly based on CHARLS through partial least squares structural equation and random forest modeling

OBJECTIVE: We aimed to explore how influential factors, such as chronic diseases, disability and social support, affect depressive symptoms in older adults, and to analyze the main influential factors on depression in different age groups. METHODS: Partial Least Squares Structural Equation Modeling (PLS-SEM) and Random Forest Modeling were used to explore the relationship between chronic diseases, cognition, activities of daily living (ADL), social support, and depressive symptoms. Moreover, co-morbidity patterns were analyzed in older people with depressive symptoms. The study also used a causal forest approach to evaluate intervention plans for these patients. RESULTS: In total, 16,627 participants aged 60 years or above from the wave four survey of the China Health and Retired Longitudinal Study (CHARLS) were included. ADL and cognition mediated impacts of chronic diseases on depression. The identification of chronic disease co-morbidity patterns revealed the effect of disease groups on depression, thereby providing clues for early diagnosis and comprehensive therapeutic strategy. Age group comparison indicated that chronic diseases and cognition became the key factors influencing older adults, and intervention measures can be developed to relieve depressive symptoms accordingly. For exemplary purpose, the intervention evaluation analysis also provided detailed information about the appropriate exercise intensity. Besides, enhanced social support was crucial for helping chronic disease patients to improve their ADL and instrumental ADL, thereby alleviating depression. CONCLUSION: These findings improve the understanding of influential factors of older patients with depression and chronic diseases, and provide clues for clinical practice, patient self-management, policymaking and future studies.