COMMUNITY-LEVEL SOCIAL CAPITAL AND SOCIAL ISOLATION IN JAPAN: A MULTILEVEL PANEL STUDY

摘要

Social isolation, which can result in solitary death, is one of major problem in Japan. Although individual characteristics among isolated individuals have been revealed, community-level factors remain unclear. Therefore, we examine the association between social isolation and community-level social capital. We used longitudinal panel data from 2010 and 2013 from the Japan Gerontological Evaluation Study, a nationwide survey involving 58,683 functionally independent older people nested within 381 communities. We defined social isolation as contact with friends less than “a few times a year”. Respondents were categorized to “continuous non-isolation”, “change to isolation”, “continuous isolation”, and “change to non-isolation”. Social capital was measured by community-level health related social capital scale including civic participation, social cohesion, and reciprocity (Saito et al. 2017) from the 2010 survey. Multi-level Poisson regression analysis was adopted, controlling for individual-level social capital, age, sex, educational attainment, marital status, IADL, household income, and population density. A marked difference was observed in the proportion of “change to isolation” at the community level (3.1% to 40.4%; mean: 14.8%). In addition, each individual-level social capital dimension showed protective associations with “change to isolation” and “change to non-isolation”. Furthermore, even after controlling for the above variables, protective associations were observed between community-level reciprocity scores and “change to isolation” (prevalence ratio: 0.95, 95% confidence interval: 0.88–0.97). On the other hand, no association was found between community-level social capital and “change to non-isolation”. These results suggest the potential of a population approach for the prevention of social isolation among older people.