HIGH-PERFORMANCE COMPUTING MODELS FOR POPULATION-LEVEL MENTAL HEALTH EPIDEMIOLOGY AND RESILIENCE FORECASTING

Abstract

This study examined population-level mental health dynamics over a ten-year period using a large, multi-country dataset integrating electronic health records, national surveys, and socio-behavioral indicators from 2015 to 2021. The research investigated how demographic, socioeconomic, environmental, and community-level determinants shape depressive and anxiety outcomes, as well as psychological resilience within diverse populations. Descriptive analyses revealed that average prevalence rates of depression (10.2%) and anxiety (11.8%) varied widely across regions, with the highest distress observed among adults aged 25–44 and consistently lower levels among individuals aged 60 and above. Women exhibited higher anxiety prevalence (13.4%) compared with men (10.1%), while resilience levels averaged 61.8 on a 0–100 scale, indicating moderate adaptive capacity with substantial regional variability. Correlation and partial correlation analyses further illuminated the structural foundations of population mental health. Resilience was strongly and negatively associated with both depression (r = –0.64) and anxiety (r = –0.59), while showing pronounced positive correlations with social cohesion (r = 0.71) and healthcare access (r = 0.68). These relationships persisted even after controlling for GDP per capita, demonstrating that resilience is not merely a function of national wealth but is instead grounded in social infrastructure, community capacity, and equitable access to essential services. Psychometric testing confirmed the reliability and validity of the composite indices used, with Cronbach’s alpha values exceeding 0.85 and confirmatory factor analysis demonstrating strong fit (CFI = 0.94; RMSEA = 0.05). The Resilience Index exhibited a stable four-factor structure encompassing economic stability, social capital, information access, and adaptive capacity. Inferential modeling provided stronger evidence of the multidimensional determinants of population resilience. Multiple regression analysis explained 71% of the variance in resilience (Adjusted R² = 0.71), identifying economic stability (β = 0.38), social cohesion (β = 0.34), and healthcare access (β = 0.31) as significant positive predictors. In contrast, unemployment (β = –0.29) and income inequality (β = –0.27) exerted significant negative effects. Hierarchical regression further demonstrated that adding community-level variables increased model explanatory power by 18%, and environmental stressors contributed an additional 7%, underscoring the multi-level nature of resilience as a systems property. Collinearity diagnostics confirmed the statistical independence and robustness of the predictors.