Effects of System-Level Factors on Race/Ethnic Differences in In-Hospital Mortality after Acute Ischemic Stroke.

Introduction: Stroke mortality has declined, with differential changes by race; stroke is now the 5th leading cause of death overall, but 2nd leading cause of death in Black individuals. Little is known about recent race/ethnic and sex trends in in-hospital mortality after acute ischemic stroke (AIS) and whether system-level factors contribute to possible differences. Methods: Using the National Inpatient Sample, adults (≥18 years) with a primary diagnosis of AIS from 2006 to 2017 (n=643,912) were identified. We assessed in-hospital mortality by race/ethnicity (White, Black, Hispanic, Asian/Pacific Islander [API], other), sex, and age. Hospitals were categorized by proportion of non-White patients served: <25% ("predominantly White patients"), 25-50% ("mixed race/ethnicity profile"), and ≥50% ("predominantly non-White patients"). Using survey adjusted logistic regression, the association between race/ethnicity and odds of mortality was assessed, adjusting for key sociodemographic, clinical, and hospital characteristics (e.g., age, comorbidities, stroke severity, do not resuscitate orders, and palliative care). Results: Overall, mortality decreased from 5.0% in 2006 to 2.9% in 2017 (p<0.001). Comparing 2012-2017 to 2006-2011, there was a 68% reduction in mortality odds overall after adjusting for covariates, most prominent in White individuals (69%) and smallest in Black individuals (57%). Compared to White patients, Black and Hispanic patients had lower odds of mortality (adjusted odds ratio (aOR) 0.82, 95% CI 0.78-0.87 and aOR 0.93, 95% CI 0.87-1.00), primarily driven by those >65 years (age x ethnicity interaction p < 0.0001). Compared to White men, Black, Hispanic, and API men, and Black women had lower aOR of mortality. The differences in mortality between White and non-White patients were most pronounced in hospitals predominantly serving White patients (aOR 0.80, 0.74-0.87) compared to mixed hospitals (aOR 0.85, 0.79-0.91) and predominantly non-White hospitals (aOR 0.88, 0.81-0.95; interaction effect: p=0.005). Discussion: AIS mortality decreased dramatically in recent years in all race/ethnic subgroups. Overall, non-White AIS patients had lower mortality than their White counterparts, a difference that was most striking in hospitals predominantly serving White patients. Further study is needed to understand these differences and to what extent biological, sociocultural, and system-level factors play a role.