ABSTRACT NUMBER: ESOC2026A2039 THE INTERSECTION OF RACE/ETHNICITY, SEX AND EDUCATION AND STROKE RISK Authors: Lynda Lisabeth, Madeline Kwicklis, Xu Shi, Christopher Becker, Lewis Morgenstern Published: 2026 ABSTRACT AbstractBackground and aimsDisparities in stroke risk by race/ethnicity, sex, and socioeconomic status are well documented; however, most prior research has examined these factors independently rather than jointly. Our objective was to assess how the intersection of race/ethnicity, sex, and educational attainment influences stroke risk.MethodsFirst-ever ischemic strokes (n=3,401) were identified from a population-based study conducted in south Texas, USA (2012–2019, 2021–2023). Demographic characteristics and education (HS) were collected and imputed when missing. Negative binomial models were fit both unadjusted and adjusted for age, incorporating random effects for each intersectional subgroup (defined by race/ethnicity, sex, education) and an offset for log-transformed population counts. Random effects were pooled, and intraclass correlation coefficients (ICCs) calculated to quantify between-group variation.ResultsMedian age (IQR) was 68 years (59–78); 57% were Mexican American, and 40% had >HS education. In unadjusted analyses, between-group variation accounted for 66% of variance in stroke incidence. For example, non-Hispanic White males with a HS education had incidence rates 1.56 times higher than the across-group average (99%CI: 1.30–1.88), whereas Mexican American females with >HS education had rates 0.61 times the across-group average (99%CI: 0.51–0.72). In age-adjusted analyses, between-group variation and relative rates across intersectional subgroups were attenuated but remained evident (see figures).ConclusionsFindings suggest that efforts to reduce stroke disparities targeting a single dimension—such as race/ethnicity alone—may be insufficient, as disparities are shaped by the intersection of multiple social identities.Conflict of interestLisabeth LD: nothing to disclose, author receives support from the National Institutes of Health; Kwicklis M: nothing to disclose; Shi X: nothing to disclose; Becker C: nothing to disclose; Morgenstern: nothing to disclose, author receives support from the National Institutes of HealthFigure 1 - belongs to Results