Racial/Ethnic and Geographic Variations in Long-Term Survival Among Medicare Beneficiaries After Acute Ischemic Stroke

Xin Tong, MPH; Linda Schieb, MPH; Mary G. George, MD, MSPH; Cathleen Gillespie, MS; Robert K. Merritt, MA; Quanhe Yang, PhD


Prev Chronic Dis. 2021;18(2):e15 

In This Article


Our study's findings suggested that about 2 in 5 Medicare FFS beneficiaries aged 66 or older survived at least 5 years after hospitalization for AIS. Men and women had similar 5-year survival. We found significant racial/ethnic and geographic variations in 5-year survival after AIS. Non-Hispanic Black men had the lowest adjusted 5-year survival. Non-Hispanic White beneficiaries overall had the least variation in adjusted 5-year survival across states; other races/ethnicities had the greatest variation.

Many studies reported racial disparities in stroke risk factors and in stroke hospitalizations, incidence, and mortality,[7–9] but few focused on long-term survival after stroke. An early study using Medicare data suggested that non-Hispanic Black people aged 65 or older, especially men, had significantly lower survival after stroke than non-Hispanic White people, consistent with our findings.[10] Yao et al recently reported that Black Medicare beneficiaries were at higher risk for ischemic stroke than White beneficiaries and more likely to have diabetes or obesity.[7] The Northern Manhattan Stroke study suggested that Black and Caribbean Hispanic people had more stroke risk factors than White people in their community-based multiethnic population study.[8] The Reasons for Geographic and Racial Differences in Stroke (REGARDS) study reported that Black people had a greater age- and sex-adjusted mean 10-year predicted stroke risk than White people, which contributed to disparities in stroke mortality.[9] Reports from the REGARDS study suggested that although management of acute stroke appeared to be more equivalent between Black and White participants, the racial disparity in stroke mortality was largely driven by differences in stroke incidence.[11] Stroke mortality mainly depends on the incidence of stroke associated with the stroke risk profiles in a population,[11,12] and stroke survival depends on prestroke morbidity and frailty, comorbid conditions, severity of stroke, access to stroke treatment, and quality of care.[13,14] Therefore, a population with a higher stroke risk profile, incidence, and mortality could have a better survival rate after stroke than those from a population with lower stroke incidence and mortality. Our findings showed that the crude difference in survival between non-Hispanic White and non-Hispanic Black populations, especially among women, became insignificant after adjusting for demographics, SES, and CCI, suggesting the importance of prestroke comorbidities (Model 1 vs Model 2) in explaining racial differences in stroke survival. Further studies are needed to examine the relative contribution of stroke risk factors, prestroke morbidity and frailty, treatments, and care to racial disparities in stroke survival.

Our study found that Medicare FFS beneficiaries in the southeastern United States region had the lowest 5-year survival following AIS. The findings of recent studies showed significant geographic variations in stroke death rates at the county level, and in the long-established stroke belt in the Southeast.[15,16] In addition, a study based on 2000–2002 Medicare FFS beneficiaries discharged with an incident ischemic stroke reported that the highest recurrent stroke rates occurred in the southern regions.[17]

Our study suggested that the differences in 5-year survival after AIS across the states appeared to be wider for Hispanic people and other races compared with non-Hispanic White and non-Hispanic Black people. The difference between the highest and the lowest survival rates across the states ranged from 9.6 to 28.5 percentage points by race and Hispanic ethnicity. Reasons for these significant differences are not clear. Among Hispanic beneficiaries, the top 5 highest 5-year survival rates were in Massachusetts (49.3%), Washington (50.6%), Maryland (52.0%), Kentucky (52.2%), and Mississippi (55.6%), whereas the 5 lowest survival rates were in Oregon (41.3%), Colorado (40.6%), Alabama (39.7%), Missouri (38.2%), and Delaware (37.9%). With the rapid growth of the Hispanic population in the United States,[18,19] there may be a gap in assessing stroke risk factors, access to health care, and promoting stroke prevention programs across the states among Hispanic residents. Samet et al reported a notably high proportion of Hispanic adults in Texas with obesity and diabetes.[20] The study, which was conducted between 2008 and 2011 and included 15,079 Hispanic participants, reported the pervasive burden of cardiovascular disease risk factors among Hispanic participants and identified the risk factors (hypertension, diabetes, and smoking) associated with stroke.[21] Other studies reported significant disparities in stroke care among racial/ethnic minority groups compared with White participants.[22]

A recent CMS report noted that disparities in clinical care among Hispanic and non-Hispanic White populations varied greatly by geography, especially in rural areas.[23] Although these geographic disparities were not related to stroke care, they may contribute to the wider variations in access to stroke care and survival across the states among Hispanic residents. A few studies also explored the differences in stroke outcomes between non-Hispanic White people and Hispanic, Asian American, and Chinese people.[24–27] A study of participants with AIS over age 65 in the American Heart Association's Get With The Guidelines–Stroke program found that non-Hispanic Black and Hispanic patients had higher adjusted 1-year all-cause rehospitalization than non-Hispanic White patients.[24] A study conducted in Hawaii comparing potentially preventable 30-day readmissions after stroke found that Chinese patients may be at higher risk than non-Hispanic White patients.[25] One Medicare study found that beneficiaries in hospitals with stroke certification had lower stroke mortality, regardless of the size of the hospital, than hospitals without certification.[26] Another Get With The Guidelines–Stroke study with linked Medicare data showed that academic hospitals as compared with nonacademic hospitals and those in the Northeast or West compared with South or Midwest had more favorable stroke outcomes.[27] The higher stroke risk profile, pre-stroke comorbidities, stroke severity, differences in access to health care after stroke, and stroke prevention programs may contribute to the wider variations in 5-year survival after AIS among minority groups across the states. In addition, minority beneficiaries may be underrepresented among Medicare FFS beneficiaries, which may contribute to the wider variation in 5-year stroke survival and limit the generalizability of our findings to minority beneficiaries.[28,29]

Our study had limitations. First, because of the lack of measures of stroke severity, we were unable to examine its impact on overall survival. Second, AIS hospitalizations and deaths were based on administrative records and limited to Medicare FFS beneficiaries aged 66 or older. The first AIS hospitalizations identified in the MEDPAR database might not in fact be the first if the beneficiaries had a stroke before they enrolled in Medicare. Third, the AIS diagnosis was based on ICD-9-CM codes from claims data and was not clinically verified, which could lead to possible misclassification. Fourth, the wider variations in 5-year stroke survival rates observed among Hispanic people and people of other races/ethnicities may be due to the limited sample size for these groups. Lastly, the findings based on FFS beneficiaries in our study may not be generalizable to Medicare patients covered under a health maintenance organization (HMO) plan because of the possible differences in beneficiary characteristics between the 2 types of coverage plans.

Our findings demonstrated significant racial/ethnic and geographic differences in long-term survival after AIS. The variations across states in different racial/ethnic groups call for further study addressing disparities in treatment and access to health care, especially among minority groups. Stroke outcomes could be improved through public health and clinical strategies, such as awareness of risk factors, early diagnosis, and aggressive management of risk factors. Further research may explain the reasons for the significant geographic variations in survival after AIS and help develop prevention strategies to reduce these gaps across the states.