Educational disparities in mortality mediated by modifiable behavioral factors: Findings from the Japan Multi-Institutional Collaborative Cohort Study

OBJECTIVES: The extent to which modifiable behavioral factors mediate educational disparities in mortality nationally remains unelucidated. Therefore, we examined whether educational level affects overall and cause-specific mortalities in Japan. STUDY DESIGN: A multi-centered cohort study. METHODS: This national, multicenter study included 58,831 participants (57 % female) from the Japan Multi-Institutional Collaborative Cohort Study (mean age: 54 years). Baseline data were collected between 2004 and 2014, with a maximum 12-year follow-up period. Outcomes included mortalities from cancer, cardiovascular disease, and other causes. Participants were categorized by educational level: high school or lower, junior college, and university or higher. The hazard ratio (HR) for mortality was estimated using Cox regression, with the university or higher cohort as the reference. Causal mediation analysis was conducted to assess the mediated effect of modifiable behaviors on education-related mortality risk. RESULTS: Among male participants, the high school or lower cohort exhibited a higher overall mortality rate (adjusted HR: 1.23; 95 % confidence interval = 1.08-1.41) and elevated risks for cancer-related and other causes of death (HRs: 1.18 and 1.75, respectively). This disparity was most pronounced in men aged 35-59 years. Smoking, excessive alcohol consumption, and inadequate leisure-time physical activity accounted for 11 %, 9 %, and 12 % of the educational mortality discrepancy, respectively. Female participants did not exhibit a significant educational mortality disparity. CONCLUSIONS: Educational disparities in mortality persist among the younger male population in Japan, with modifiable behavioral risks identified as key mediators.