Abstract
Latinx people living in the U.S. report a disproportionately high prevalence of diabetes. This project builds on the existing social determinants of diabetes literature by examining factors associated with a greater likelihood of diabetes and investigates factors correlated with the Latinx/non-Latinx disparity. This project studies the adult sample (18 and older) from the 2010-2018 IPUMS Health: National Health Interview Survey (NHIS) data. Logistic regression analyses are used to examine the patterns between reporting Latinx identity and reporting diabetes with additional subgroup analyses of the Latinx and non-Latinx groups. Then, Oaxaca-Blinder decomposition is used to examine the patterns explaining the difference in self-reported diabetes between the Latinx and non-Latinx population for the whole sample and by age group. The logistic regression analyses show that after adjusting for age and other key social determinants of health, Latinx individuals are approximately 64.5% (OR 1.645, [95% CI, 1.536-1.760]) more likely to report being diagnosed with diabetes than non-Latinx individuals. Individual characteristics of age, race, and smoking behaviors are identified as suppressors of the gap, and conversely, characteristics of income, education, and BMI all contribute to the Latinx diabetes disparity gap. The Oaxaca-Blinder decomposition results show that the measured social determinants of health characteristics explain a meaningful amount of the Latinx diabetes gap. Importantly, differences in education and income (which are more immediately actionable policy areas) make larger contributions to the gap than BMI or other health behaviors.