Abstract
OBJECTIVE: Depression not only diminishes the quality of life and self-care abilities of the elderly but also exacerbates the burden of pre-existing physical diseases. This study aims to explore the influencing factors of depressive events in patients with cardiometabolic disease (CMD) within a division of the Xinjiang Production and Construction Corps (XPCC). Additionally, it seeks to establish a depression risk prediction model to provide a foundation for hierarchical management of this comorbidity in the community and for the stratification of the population. METHOD: Utilizing baseline data from 2016 and follow-up data from 2018 to 2024, a total of 1,648 individuals were included after excluding those with baseline depression, cognitive impairment, or loss to follow-up. Data on demographic characteristics, lifestyle habits, chronic diseases, and health service utilization were collected. Generalized estimating equation (GEE) and adaptive LASSO regression were employed to identify longitudinal predictors. Depression prediction models for the total population and subgroups (non-multimorbidity/multimorbidity) were constructed using GEE and generalized linear mixed model (GLMM), and model performance was evaluated. RESULT: The findings revealed that as of 2024, the cumulative incidence of depression was 68.93% (67.02% for non-multimorbidity patients and 74.30% for multimorbidity patients). Model comparisons indicated that the discrimination and accuracy of GLMM were superior in the total population. GEE demonstrated better performance in non-multimorbidity patients, while GLMM consistently outperformed GEE in the multimorbidity population. Self-rated health status and metabolic equivalent emerged as core predictors, with chronic disease treatment and emergency medical treatment identified as significant characteristics for the general population and non-multimorbidity population, respectively. CONCLUSION: The incidence of cardiometabolic diseases accompanied by depression among the elderly in a division community of the Xinjiang Production and Construction Corps (XPCC) is higher than the national average. Generalized Linear Mixed Models (GLMM) are more suitable for screening individuals with cardiometabolic diseases. This study emphasizes the importance of physical activity, social security, and health service indicators in predicting the risk of depression.