Previous studies reported that social disorganization/fragmentation could predict mental well-being of residents in a community. The aim of this study is to examine how area and individual level of social distance could predict likelihood of mental health among older people in Japan. We empirically derived an index of "social distance" by taking averaged differences in sociodemographic characteristics that are income, education, hometown of origin, the duration of residency, and life stage, between the study participants and their neighbors. We used the study participants (n = 9147) from the Aichi Gerontological Evaluation Study, which targeted residents with aged 65 years or over in a central part in Japan. Depressive symptoms of the study participants were assessed using the short version of the Geriatric Depression Scale (GDS-15). We also tested if area-level social capital would moderate the association between social distance and depressive symptoms. Using multilevel analyses, we found that higher social distance from neighbors was associated with increased depressive symptoms, independently of respondents' own values of income and educational attainment. At the individual level, each standard deviation in income-based and education-based social distance was associated with an odds ratio for depressive symptoms of 1.15 (95% CI: 1.01-1.30) and 1.17 (95% CI: 1.03-1.32), respectively. However, the area-aggregated indices of social distance were not associated with depressive symptoms. Additionally, area-level social capital indicating higher levels of trust between neighbors and social participation, buffered the adverse effect of social distance on depressive risk. In an instance of the "dark side" of social capital, we also found that stronger social cohesion increased depressive symptoms for residents whose hometown of origin differed from the communities where they currently resided.
Copyright © 2013 Elsevier Ltd. All rights reserved.