A prospective cohort study of neighborhood stress and ischemic heart disease in Japan: a multilevel analysis using the JACC study data

BACKGROUND: A body of research has shown that neighborhood environment may have an effect on a variety of health outcomes, including cardiovascular disease. One explanation for the mechanism of the effect of neighborhood on cardiovascular disease is psychosocial pathways. Direct evidence for an effect of neighborhood on cardiovascular disease with adjustment for perceived stress at the individual level has not been obtained, however. The Japan Collaborative Cohort Study for the Evaluation of Cancer Risk provides a unique dataset which has aggregated area-based cohorts from 45 areas throughout Japan. The purpose of the present study was to examine the contextual effect of area-level stress on ischemic heart disease using data from a large prospective cohort in Japan. METHODS: A baseline survey of 110,792 residents of 45 areas aged 40-79 years was conducted between 1988 and 1990. Analysis was restricted to subjects from the 33 of 45 areas providing information about self-rated stress (32183 men and 45896 women). Multilevel Poisson regression models were employed in a two-level structure of individuals nested within the 33 areas. Area-level stress was calculated by sex as the number of persons who rated their stress level as high divided by the total number of subjects in that area. Mortality rate ratios (MRRs) per 1 percentage point increase in area-level stress were estimated with adjustment for compositional individual factors. RESULTS: During 15 years of follow-up (1,116,895 person-years), 936 deaths due to ischemic heart disease were recorded. Area-level stress varied from 6% to 22%. In the multivariable models, MRRs of area-level stress were 1.06 (95% confidence interval: 1.00-1.12, p = 0.043) in men and 1.07 (95% confidence interval: 1.00-1.14, p = 0.057) in women. CONCLUSIONS: Area-level stress affects the likelihood of death due to ischemic heart disease of individuals in men. The present findings may suggest that stress should be considered not only within the individual but also within the neighborhood context.