EDUCATION AND MORTALITY: IS A COLLEGE DEGREE UNIVERSALLY PROTECTIVE?



Arthur L. Klatsky, M.D., Kaiser Permanente Division of Research, Oakland, CA, USA

Purpose: To study risk of death in relation to educational attainment in 273,843 persons.

Background: Low socio-economic status (SES) has been associated with increased incidence of many medical conditions and connected to racial and ethnic health disparities. Interacting factors include lifestyle, environmental exposures, and health care quality. Using education as a marker of SES, we studied its apparent impact on long-term mortality in a large multiethnic population.

Methods: Logistic regression with 8 covariates; the education referent was persons with no college.

Results: With average follow-up of >30 years, there were 103,218 deaths. Odds ratios (OR) and 95% confidence intervals for death were 0.84 (0.81-0.85) for persons with some college and 0.65 (0.63-0.67) for college graduates (GRAD). For male GRAD the OR was 0.61; for female GRAD it was 0.71. For White, Black, and Asian GRAD the ORs were 0.65, 0.64, and 0.73 respectively. The p value for each of these estimates was <0.001. GRAD had lower risk (p < 0.01) in strata of specific Asian ethnicity (Chinese, Japanese, Filipino, South Asian), smoking habit, interval to death, and BMI, except for persons >60 years old at baseline (OR = 0.90 [0.80-1-01, p = 0.07]). For selected death causes (cardiovascular, non-cardiovascular, coronary, cancer, respiratory, liver disease, accidents/violence) GRAD ORs ranged from 0.46 (liver) to 0.74 (cancer) with all p values < 0.001.

Conclusion: College graduates across gender, race, and smoking strata have substantially lower risk of all-cause mortality and of death from multiple diagnostic groups.