Lower education is a risk factor for alcohol-related deaths, but it is unknown if the impact of education varies by race/ethnicity. We evaluated the interaction between education and race/ethnicity on the odds of alcohol use disorder (AUD) and the significance of employment and poverty as potential mediators. The 2016-19 National Survey on Drug Use and Health was used to obtain weighted prevalence of AUD in participants 26+ years. Using AUD as the outcome, multivariable logistic regression analyses were performed, stratified by four levels of education, ranging from less than high school (<HS) to 4-year college+ (4C); odds ratios (ORs) for each race/ethnicity were compared across models for interaction. The second model included employment and poverty as co-variates; changes in ORs were calculated to assess for >|10%| change implying significant mediating or confounding effect. AUD prevalence was highest among American Indian/Native Alaskans (AI/ANs) (8.06%), and similar among non-Hispanic Whites (5.37%) and Blacks (5.09%). When stratified by education, the odds of AUD among Blacks and AI/ANs increased with decreasing education [(OR 1.45, 95% CI 1.07-1.96) for Blacks v. Whites with <HS, compared to 0.55 (0.46-0.67) for Blacks v. Whites with 4C; 2.55 (1.53-4.24) for AI/ANs v. Whites with <HS v. 1.01 (0.45-2.28) for AI/ANs v. Whites with 4C). Including employment and poverty resulted in <|10%| change in ORs. Lower education significantly increased the odds of AUD among Blacks and AI/ANs. Employment and poverty did not significantly change the association, implying there are likely other mechanisms accounting for racial/ethnic disparities in AUD.
Read full abstract