Differences in the Relationship Between Educational Attainment and Health Status Across Racial and Ethnic Groups in a Multi-ethnic United States Older Adult Population: A Cross-Sectional Electronic Health Record-Based Study.

Introduction We aimed to describe the relationship of educational attainment with the prevalence of six health outcomes (ever and current smoking, diabetes, hypertension, coronary artery disease, and chronic obstructive pulmonary disease) in an older adult population, including whether education-health relationships differed by health outcome, by racial and ethnic (racial/ethnic) group, and by racial/ethnic group within the same level of education. Methods This cross-sectional study used 2015-2016 electronic health record data for 149,417 non-Hispanic White (White), 15,398 African-American or other Black (Black), 15,319 Hispanic or Latino (Latino), 10,133 Filipino, and 8810 Chinese Northern California health plan members aged 65-79 years whose preferred language was English. For each racial/ethnic group, sex-specific age-standardized prevalence of the six health outcomes was estimated for four levels of education (non-high school graduate, high school graduate, some college, college graduate). Age-adjusted prevalence ratios were used to compare the prevalence between adjacent levels of education and at lower versus college graduate levels within racial/ethnic groups, and the prevalence between White adults and adults in the other racial/ethnic groups, within each level of education and overall. Results The education-health relationship varied across racial/ethnic groups and health outcomes, with gradient relationships more consistently seen for White, Black, and Latino older adults than Filipino and Chinese older adults. Even when a gradient relationship was not observed, the prevalence at the college graduate level was usually significantly lower than the prevalence at the three lower levels of education. The prevalence of current smoking, diabetes, and hypertension was higher among Black than White adults at most levels of education. Controlling for education level minimally affected comparisons of overall prevalence of health outcomes between adults in the White and the other racial/ethnic groups, with the broadest impact seen for Latino-White comparisons. Conclusions The relationship of level of education and health outcomes differs across racial/ethnic groups and by health outcome. This should be taken into consideration when using education as a risk adjustment factor or predictor of health outcomes in multi-ethnic older adult populations.