Differences in mortality and incidence for major sites of cancer by education level in a Japanese population.

PURPOSE: We aimed to examine the relationships between education and mortality and incidence for major sites of cancer in a Japanese population.
METHODS: Subjects were 32,883 respondents of questionnaire survey in 1978 with ages younger than 75 years. Cancer cases were ascertained through 2001, and causes of deaths were identified through 2003. Hazard ratios of deaths from cancer or developing cancers were compared among those with 9 or less, 10-12, and 13 years or more of education using Cox proportional hazard models.
RESULTS: As for cancer mortality of all sites combined, a statistically significantly decreasing trend was observed in age-adjusted models in both men and women, but no significant differences were observed in multivariate-adjusted (age, body mass index, smoking, radiation dose, and city) models. Among major cancer sites (stomach, colon/rectum, liver, lung, and female breast) examined, a significantly decreasing trend was observed for male liver cancer in a multivariate-adjusted model. As for incidence, a significantly decreasing trend was observed for cancer of all sites combined in men, and for male liver and prostate cancer and female lung cancer in a multivariate-adjusted model.
CONCLUSIONS: Educational differences in cancer incidence and mortality were generally rather small, but were significant for incidence for male all-site, male liver, prostate, and female lung cancers.