The relationship between educational level and mortality. The Reykjavík Study

OBJECTIVE: Epidemiological studies have indicated an association between socioeconomic factors and health. It has not been clearly established whether this association is wholly or partly independent of classical risk factors. Our objective was to estimate the relationship between educational level and coronary artery disease (CAD), mortality and all-cause mortality. The Reykjavík Study involving 18 912 participants followed-up 4-30 years provides an ideal opportunity to address this question. DESIGN AND SUBJECTS: The participants were aged 33-81 years and living in the Reykjavík area. They were divided into four groups according to education. The standard risk factors were assessed on entry and mortality, and cause of death registered during follow-up. Multiple Cox regression analysis was applied to assess the relationship between age at examination, year of examination, educational level and mortality. RESULTS: The all-cause mortality and CAD mortality was significantly related to education, even after adjustment for classical risk factors. For men, 14% (95% CI: 2-24) reduction was found in CAD mortality for those having high school education relative to elementary school. The figures for junior college and university education were 17% (95% CI: 1-31) and 38% (95% CI: 21-32), respectively. These figures were only slightly lower when major CAD risk factors were controlled for and still significant. Similar figures were found for all-cause mortality. For women 34% (95% CI: 18-48) reduction was found in CAD mortality for high school education and 55% (95% CI: 22-74) for junior college, but too few had university education for reliable results. The figures were lower for all-cause mortality, but significant. The figures were reduced when major CAD risk factors were controlled for, but still significant. CONCLUSION: Education is a strong protective factor both for all-cause and CAD mortality. Only a small part of this effect can be explained through conventional risk factors.