Abstract 4357208: Discrimination in Healthcare and Incident Stroke or Myocardial Infarction

People who experience discrimination in healthcare may be less likely to interact with the healthcare system and receive guideline-recommended care, contributing to disparities in multiple health outcomes. The degree to which discrimination is associated with cardiovascular health is not well understood. Is self-reported discrimination in healthcare associated with a composite outcome of incident stroke or MI? We included participants age 50-80 at baseline from the 2008-2020 Health and Retirement Study who were followed for up to 12 years. Discrimination in healthcare was assessed at baseline by asking respondents how often they “receive poorer service or treatment than other people from doctors or hospitals.” Incident stroke or MI was determined through participant-reported diagnosis by a doctor during follow-up (month/year). Cox proportional hazards models were used to estimate incident stroke or MI (n=1911 events). Due to differences in sociodemographic factors and health status, which are associated with the probability of reporting discrimination, we calculated stabilized inverse probability weights from a logistic propensity score and applied them to our models. Our models adjusted for baseline covariates: sociodemographic factors (age, gender, race, ethnicity, education, being uninsured), health status (BMI, smoking status, disease diagnoses [diabetes, hypertension, heart disease]), events before study baseline (prior stroke or mi), and prior healthcare utilization (doctor visits in past 2 years). We employed a sensitivity analysis which used a shorter follow-up period to assess risk of discrimination closer to exposure time. Among study participants (n=17632, mean age 66, 42% male, 69% Non-Hispanic White, 17% Non-Hispanic Black, 10% Hispanic, 3% Non-Hispanic Other), 19% (n=3347) reported discrimination in healthcare settings. Self-reported healthcare discrimination was associated with an increased hazard of stroke or MI (hazard ratio [HR]=1.23, 95% confidence interval [CI]= 1.10-1.39) over the entire follow-up. The hazard is strongest over a shorter 2-year follow-up (HR=1.46, 95% CI=1.17-1.82), and the hazard plateaus after a 5-year follow-up (HR=1.23, 95% CI=1.05=1.44). Even after adjusting for sociodemographic factors and health status, participants who reported experiencing discrimination in healthcare were at increased risk of stroke or MI over short-term and long-term follow-up.

Duke Scholars

Author Matthew E. Dupre Population Health Sciences