Psychosocial Resources, Depression, and Subjective Well-Being among Middle-Aged and Older Adults in China: A Machine Learning and Network Analysis Approach

Subjective well-being (SWB) is a critical component of healthy aging and a core indicator of quality of life. This study aimed to identify key modifiable determinants of SWB among middle-aged and older Chinese adults and to elucidate the complex interrelationships among these factors. A multicenter cross-sectional study was conducted with 7,100 participants aged 55 years and older across four regions in China. Twenty-six variables were analyzed. Feature selection was performed using LASSO regression. Seven machine learning algorithms were trained to predict high SWB, assessed by the WHO-5 Well-Being Index. Model interpretability was evaluated using SHapley Additive exPlanations (SHAP), while network analysis and structural equation modeling characterized variable interrelationships. The Light Gradient Boosting Machine model demonstrated superior performance, achieving an area under the curve of 0.840. SHAP analysis identified social support, attitudes toward aging, self-efficacy, depression, and health-related quality of life as the top predictors. Network analysis revealed that depression served as a central bridge connecting psychosocial and health-related clusters. Structural equation modeling indicated that positive attitudes toward aging and social support influenced well-being directly and indirectly through self-efficacy and depressive symptoms. Psychosocial resources—particularly social support, positive aging attitudes, and self-efficacy—emerged as the strongest determinants of SWB, with depression functioning as a pivotal mediating pathway. Community-based interventions targeting positive aging perceptions, social connectedness, and early depression screening may effectively enhance well-being and quality of life in later life.