The main objective of this spatial epidemiologic research is to gain greater insights into the geographic dimension displayed by the different duration of mentally unhealthy days (MUDs) across U.S. counties. Mentally unhealthy days (MUDs) are studied in entire cross counties for year of 2014. Using Behavioural Risk Factor Surveillance System (BRFSS) data in 2014, we examine main factors of mental health hazard including health behaviour, clinical care, socioeconomic and physical environment, demographic, community resilience, and extreme climatic conditions. In this study, we take complex design factors such as clustering, stratification and sample weight in the BRFSS data into account by using Complex Samples General Linear Model (CSGLM). Then, spatial regression models, spatial lag and error models, are applied to examine spatial dependencies and heteroscedasticity. Results of the geographic analyses indicate that counties with lower air pollution (PM2.5), higher community resilience (social, economic, infrastructure, and institutional resilience), and higher sunlight exposure had significantly lower average number of MUDs reported in the past 30 days. These findings suggest that policy makers should take air pollution, community resilience, and sunlight exposure into account when designing environmental and health policies and allocating resources to more effectively manage mental health problems.
Keywords: Complex Samples General Linear Model (CSGLM); Mental health; U.S counties; air pollution; community resilience; spatial regression models; sunlight.