Generalized and particularized trust for health between urban and rural residents in Japan: A cohort study from the JAGES project.

Previous studies on trust and health have not fully considered the nature of trust in relation to types of trust and socio-cultural background. The present study aimed to examine whether generalized trust (trust in general people; GT) and particularized trust (trust in particular people; PT) in urban and rural areas had different associations with health. This prospective cohort study on older adults used panel data obtained in 2010 and 2013. Surveys were conducted in 24 municipalities in Japan. Of 20,209 respondents, 13,657 participants were followed up. The independent variables were GT and PT in neighbors; the dependent variable was self-rated health (SRH) at follow-up. We examined the interaction term between population density and each trust variable. Age, sex, SRH at the baseline, and other potential confounders were adjusted. The median age was 72 years (females: 53.4%). Percentages of high GT and high PT were 21.0% and 72.4%, respectively. Prevalence of poor SRH at the follow-up was 15.5% and 28.5% in high and low GT, respectively, and 16.9% and 32.8% in high and low PT, respectively. After adjusting for covariates in logistic regression models, low GT and PT were significantly associated with higher odds ratios (ORs) for poor SRH compared to high trust (GT: OR = 1.43 [95% confidence interval (95%CI) = 1.17, 1.75] and PT: OR = 1.44 [95%CI = 1.15, 1.81]). Associations of low PT with poor SRH significantly strengthened when population density increased (interaction term of low PT: OR = 1.16 [95%CI = 1.04, 1.27]). On the other hand, associations of GT with SRH were not significantly interacted by population density. The mediation analysis showed that the direct effects of PT influenced SRH in urban areas only. In urban areas with high social uncertainty, trust in particular neighbors was more beneficial to health.