Stroke Disparities: From Observations to Actions Inaugural Edward J. Kenton Lecture 2020

Abstract

Numerous epidemiological studies have demonstrated stroke disparities across race and ethnic groups. The goal of the Northern Manhattan Study (NOMAS) was to evaluate race and ethnic differences in stroke within a community with three different race-ethnic groups. Starting as a population-based incidence and case-control study, the study evolved into a cohort study. Results from NOMAS have demonstrated differences in stroke incidence, subtypes, risk factors, and outcomes. Disparities in ideal cardiovascular health can help explain many differences in stroke incidence and call for tailored risk factor modification through innovative portals to shift more diverse subjects to ideal cardiovascular health. The results of NOMAS and multiple other studies have provided foundational data to support interventions. Conceptual models to address health disparities have called for moving from detecting disparities in disease incidence, to determining the underlying causes of disparities and developing interventions, and then to testing interventions in human populations. Further actions to address race and ethnic stroke disparities are needed including innovative risk factor interventions, stroke awareness campaigns, quality improvement programs, workforce diversification, and accelerating policy changes.

Introduction

I knew Ed Kenton and admired his steady leadership behind the scenes. He was a humble, quiet, but strong, leader for many of us, and was actively engaged in multiple professional organizations. It means a great deal to me to be the inaugural speaker for this lecture to honor his legacy that we will continue to recognize for many years. As Bruce Ovbiagele already stated, Dr. Kenton made many contributions, some of them particularly important in stroke and minorities and some regarding the birth of the field of vascular neurology.^1,2 I had the pleasure of working with him on the 2006 American Stroke Association’s guidelines for secondary stroke prevention, in which he was in charge of a specific new section regarding applying guidelines to minority populations.³ Unfortunately, I was the one who had to inform the American Academy of Neurology (AAN) at our leadership dinner of his passing. About two weeks before he died, I received an email from him. In our exchange, there was one quote from Ed that I wanted to share that I think is especially relevant for all of us. He wrote: “I have been most fortunate for having been exposed to such talented colleagues and therefore developed and learned so much through the years.” That was what he was about, and I share those sentiments. I am grateful to the many terrific colleagues with whom I have worked who could be giving this lecture. I have learned so much from them throughout the years. Ed also had another piece of advice, and I share this with all of you: “Don’t retire too late.” He said that in an email to me and Cathy Rydell, who is the retiring CEO of the AAN. Ed Kenton had passion, energy and commitment to reduce disparities. He was a role model and has left a great legacy for many of us to follow.

As I thought about the topic of this lecture, I thought it would be most appropriate to address stroke disparities but focus on moving from observations to actions. Besides reviewing the many observations from multiple studies on disparities in mortality, incidence, stroke subtype, risk factors, and ideal cardiovascular health, I wanted to emphasize the need to shift from observations to actions. Our epidemiological observations provide the foundation for initiating interventions to develop, demonstrate, and disseminate to reduce stroke disparities.⁴ These actions include novel risk factor interventions, culturally tailored cause campaigns that Ed Kenton was instrumental in supporting, quality improvement programs with evidence-based recommendations, workforce development and nurturing talent like this Health Equity & Actionable Disparities in Stroke: Understanding & Problem-solving (HEADS-UP) Symposium, and advocating for policy changes that will affect stroke disparities.

Defining Health Disparities

The definition of a health disparity encompasses an inequality or difference in a condition or rank between groups. It implies a lack of equality, opportunity, treatment or status. Often it is expressed as a quantitative measure such as a rate, percent, or mean, and represents a quantity that separates a group from a reference point on a particular measure of health. Many times, we may not readily identify the cause of the disparity. The disparity calls attention to differences in health between groups regardless of the cause and quantifies the measure in absolute or relative terms. Further studies are needed to identify the drivers of these differences and work to reduce them.

Vital statistics data from the Centers for Disease Control and Prevention (CDC) have demonstrated clear disparities for stroke and cardiovascular mortality. Mortality is one of the easiest rates you can measure, and there have been well recognized greater mortality rates among African Americans compared to whites for heart disease, as well as, stroke. Less has been noted for Hispanic ethnicity which is a heterogeneous group that makes the study of that population even more difficult. Hispanics in the United States are made up of multiple different groups and even the studies that had begun to define them, have now begun to separate them. There are Mexican Americans which make up the majority of Hispanics in the US. Caribbean Hispanics include people from the Dominican Republic, Puerto Rico, Cuba, and Central America. Some Hispanics also originate from South America. The term Hispanic has been used in the US to denote people who are originally from Spanish-speaking countries in Latin America and the Caribbean, regardless of race. Although there are some who do not like the term, it is more of a social-cultural-political construct than a biological classification.⁵ The nosology is used by the US census, government, and researchers to identify an understudied population.

The US population is changing and becoming much more diverse. When we look at projected population growth for the year 2060 and beyond, 44% will be non-Hispanic Whites, dropping from 62%, 14% will be African Americans, and then the Hispanic/Latino community will increase from 17% in 2014 to 29% by 2060.⁶ Caribbean Hispanics (Cubans, Puerto Ricans, Dominicans) are the 2nd largest US group, and there is inadequate data to address their health needs.⁷ These demographic shifts require us to understand and address disparities in this rapidly growing community. Although there have been declines in mortality for cardiovascular diseases and stroke, race and ethnic disparities persist. Both stroke mortality and incidence have been declining for African Americans and Hispanics, but there are still gaps that require further work.

Design of the Northern Manhattan Study

When we set up the Northern Manhattan Study in the early 1990s, it was known that African Americans had twice the mortality from stroke according to US health statistics. Little was known about Hispanics. There were no Latino-focused epidemiological studies, and they were not included as much in cardiovascular health studies. Prior stroke epidemiology studies were performed in predominately white populations.

Northern Manhattan is a tri-ethnic, densely populated, urban community served by one hospital and had a majority Hispanic population. The population was 63% Hispanic, 20% African American, and 15% White. Prior to initiating the Northern Manhattan Study, we had developed some stroke registry data. When I was a stroke fellow, I worked with the NINDS Stroke Data Bank that included four centers prospectively collecting stroke data.^{8, 9} Columbia was one of the centers, and we began to collect standardized data from hospitalized cases. Using that registry, we could begin to evaluate the characteristics of the stroke cases in the neighborhood. After the NINDS Stroke Data bank ended, we started some retrospective data collection on cases from northern Manhattan and began to follow hospitalized stroke patients to characterize disparities in stroke outcomes. One of the first papers we published described one-year outcomes after cerebral infarction in whites, Blacks, and Hispanics hospitalized at Columbia Presbyterian Medical Center.¹⁰ This gave us some of the early pilot data that we used to form the NINDS grant proposals to support the development of the study that was originally called the Northern Manhattan Stroke Study (NOMASS).

The study started when I was at Columbia and continued when I moved to the University of Miami in 2007. The project was modified to a dual-PI model and Mitch Elkind became the other PI with the Columbia research team who continue to follow the prospective cohort. We have continued our productive multi-institutional collaboration. An important message for younger investigators is that collaborations can remain lifelong. Even if you move, those team connections you build will be connections that you can maintain for many years especially in today’s more connected society.

The goal of the NOMASS was to evaluate for race and ethnic differences in stroke within a community with three different race-ethnic groups. Our design was a population-based study that assessed stroke incidence, stroke subtypes, risk factors, and outcomes. The study methods have evolved over time and was initially a population-based stroke incidence and surveillance study and population-based case-control study. NOMASS was really three case-control studies because we matched our stroke cases by age, gender, and race to community derived controls collected through random-digit dialing. After 5 years during the second grant cycle, we expanded the collection of community subjects and evolved to the current prospective cohort study similar to the Framingham Study. Later, we also developed a family cohort to look at genetics and then in the third cycle developed the MRI cohort which began to focus on cognitive decline.

Observations on Stroke Incidence Disparities from NOMASS and Other Studies

Our initial funding for the NOMASS allowed us to collect surveillance data on all of the stroke cases within the region, not just at Columbia where the majority of cases were hospitalized, but also at other hospitals. We had to enlist the help of colleagues at surrounding hospitals where some community subjects from northern Manhattan could have been hospitalized. We also worked to evaluate local imaging centers and the NYC Department of Health to make sure we also included any non-hospitalized strokes. This effort led to our report on the first recorded age-adjusted stroke incidence rates for White, African American, and Hispanics from the same community of northern Manhattan.¹¹ Blacks had the greatest stroke incidence, both in men and women, compared to the white population in northern Manhattan. For the first time, we also published some data on Hispanics, mainly Dominicans that accounted for nearly two-thirds of the Hispanics in northern Manhattan, followed by Puerto Ricans, and some Cubans. Our Hispanic group had intermediate, but greater age-adjusted stroke incidence rates compared to the white population in the northern Manhattan region. Because this was an incidence surveillance study, we had complete risk factor data on the cases, but the denominator was based on census data and we could not adjust for vascular risk factors. Our concurrent population-based case-control study would provide the early data on risk factor differences across the groups.

Other stroke studies developed and have continued to demonstrate race and ethnic differences in stroke incidence. The Atherosclerosis Risk in Communities (ARIC) cohort showed that Blacks had a 2.41 relative rate ratio for age-adjusted stroke incidence compared to whites which was attenuated to 1.38 after adjustment form vascular risk factors.¹² The Greater Cincinnati/Northern Kentucky Study (GCNKSS) showed that stroke disparities were greater in younger compared to older, with an overall 1.5–2.5 relative differences for Blacks compared to whites, but greater relative differences under the age of 65.¹³ The REGARDS (Reasons for Geographic and Racial Differences in Stroke) study was developed as a nationwide prospective study looking at African Americans versus Whites, and also showed that the incidence rate differences were greater in the younger group than in the older group, with an incidence rate ratio of 4 between the ages of 45–54.¹⁴ Less was known about Hispanics until the Brain Attack Surveillance in Corpus Christi (BASIC) study demonstrated significant stroke incidence disparities for Hispanics of Mexican background.¹⁵

Race and Ethnic Differences in Stroke Subtype and Risk Factors

What about differences in stroke subtype? For intracerebral hemorrhage, the NOMASS found that Blacks versus White had nearly a fourfold difference in our incidence study for differences in intracerebral hemorrhage risk.¹⁶ Since the majority of strokes were evaluated at the same hospital, we could standardize the diagnostic work up and reduce any differential evaluation biases for ischemic stroke subtype. We observed greater relative risks for extracranial atherosclerotic stroke and nearly 5 fold greater rates for intracranial atherosclerotic stroke among Blacks versus Whites, as well as for Hispanics versus Whites.¹⁷ This suggested that maybe there were some vascular risk factors differences that could contribute to why one group would have more intracranial atherosclerosis than another group.¹⁸ For small vessel disease, we found differences with Black versus White having a relative incidence rate of 3.1 and Hispanic versus Whites with 2.3. The relative differences for cardiac embolism were much less with 1.6 for blacks and 1.4 for Hispanics versus whites, less of a difference in terms of the overall relative incidence rate.

We also began to look at risk factors differences as one way to understand the incidence rate disparities. Our risk factor evaluations were mainly based on the results of our population-based case-control study, where we matched by age within five years, by gender and sex, and race-ethnic groups. ^19,20 For hypertension, we observed slightly greater odds ratios in blacks and Hispanics compared to whites, but with largely overlapping confidence intervals.²¹ REGARDS later showed that hypertension had a greater impact among African Americans for every millimeter of systolic blood pressure greater than that for Whites.²² In NOMASS, the real driver for the disparities for hypertension was the difference in the prevalence among the controls. In the White counterparts it was 43%, 62% in the African American non-stroke cases, and 58% in Hispanics. When you translate this to an etiologic fraction or attributable risk, we showed that 37% of strokes in the African Americans could have been attributed to hypertension and 32% in Hispanics compared to 25% for whites.²¹ We found a very different pattern for atrial fibrillation. Just as we had shown that the incidence rate disparities were lower for cardiac embolism, NOMASS found that the impact of atrial fibrillation was much greater in our White group compared to African Americans and Hispanics. Both in terms of prevalence and the odds ratios, AF was greater in whites compared to blacks and Hispanics, translating into a greater etiologic fraction in our population-based case-control study.

We also began to investigate behavioral risk factors, such as alcohol use¹⁹ and leisure-time physical activity.²³ No physical activity seemed to have the greatest impact in our Hispanic group – threefold greater risk of stroke in the non-physically active; physical inactivity was also most prevalent amongst Hispanics at 37%, therefore translating to a greater etiologic fraction.²¹

Results from the Northern Manhattan Prospective Cohort Study

There are limitations to a population-based case-control study, so as mentioned above we evolved to the current epidemiologic study design of a prospective cohort study. We dropped the word “stroke” and changed the name when we moved to a cohort study and became known as the Northern Manhattan Study (Figure 1). We expanded the collection of community subjects using all of our community controls who were derived through telephone, random digit dialing. They were population-based, free of stroke and representative of the underlying tri-ethnic community. A prospective cohort study design is always a gold standard approach because you are following people and collecting their information pre-disease and measuring their exposure over time. This study design calls for a larger investment, takes a longer time for strokes to develop in the exposure groups, and requires a larger sample size of subjects to follow.

Figure 1:

Northern Manhattan Study Timeline

NOMAS has evolved over many years and we are now in the 27^th year of funding. The prospective cohort consisted of 3,298 subjects with a mean age of 69, 52% were Hispanic, 24% non-Hispanic African American, and 21% non-Hispanic White. About 37% of the NOMAS cohort were men, 46% had a high school education, and 44% had Medicaid or were uninsured. Vascular risk factors were prevalent: 17% were smokers, only 9% had moderate to heavy physical activity, 74% had hypertension, 22% had diabetes, and 62% had hypercholesterolemia. There were differences across race-ethnic groups for many of these factors. Over the years there have been multiple studies that began to work on this core cohort. Our focus in the last few years has really turned to brain health with our neuropsychological assessments and recently initiated dementia adjudication committee.

The goal of the NOMAS cohort study was to follow people over time. We spent time focusing on traditional and novel vascular risk factors, subclinical determinants, mediators, or moderators of outcomes. (Table 1) We extensively measured subclinical vascular disease including carotid plaque, Intima-medial thickness, carotid distensibility.^{24,25,26,27,28} Working with our cardiology colleagues, we utilized echocardiography to measure left ventricular mass, left atrial size, atrial cardiopathy, aortic atherosclerotic plaque, left ventricular strain, ambulatory blood pressure, and cardiac rhythm monitoring.^{29,30,31,32,33,34,35} We also began to look at brain imaging biomarkers and used Magnetic Resonance Imaging to measure subclinical measurements such as white matter hyperintensities, silent and covert infarcts, cerebral microbleeds, and most recently, perivascular spaces and dolichoectasia.^{36,37,38,39,40,41,42, 43,44} We established a collaborative cohort with a wealth of data that grew over time and has allowed many younger colleagues to build academic careers. The goals were to evaluate the relationships among these subclinical measurements and traditional and novel risk factors and risk scores, as well as the risk of clinical vascular events such as stroke, myocardial infarction, and death. As the cohort aged, we enhanced our evaluations to track and define cognitive decline, Mild Cognitive Impairment, and dementia.

Table 1:

Risk factors evaluated in the Northern Manhattan Study

Sociodemographic	Sex, Race, Ethnicity
Socioeconomic	Education, Insurance status
Vascular Risk Factors	Hypertension, Diabetes, Lipids, Cardiac disease
Behavioral Factors	Alcohol, Smoking, Physical Activity, Obesity, Sleep, Migraine
Diet	Mediterranean Pattern, Salt, Fat, Diet Soda
Social Factors	Isolation, Depression
Metabolic	Metabolic Syndrome, HOMA Index, Homocysteine, GFR, FGF23
Carotid Measures	Intima-media thickness, carotid plaque, stenosis, distensibility
Cardiac Measures	Patent foramen ovale, aortic arch atherosclerosis, left atrial size, atrial cardiopathy, left ventricular mass, ambulatory blood pressure
MRI Measures	White matter hyperintensities, silent and covert infarcts, cranial and cerebral volumes, hippocampal volume, perivascular spaces, intracranial dilatation and stenosis
Inflammation and Infection	Infectious burden, inflammatory markers
Combination Scales	Global vascular risk score, Life’s simple 7

Like many cohort studies, it requires a long follow-up period to accumulate enough cases to adequately evaluate stroke incidence. We have recently evaluated sex, race and ethnic differences in stroke incidence in the NOMAS cohort. Our stroke incidence data from the prospective cohort was presented at the International Stroke Conference, as well as published in a simultaneous publication in Stroke.⁴⁵ The crude stroke incidence rates were 9/1000 for Whites, 13/1000 for Blacks, and 10/1000 for Hispanics. Different from some other studies, we documented disparities even among those over age 70. (Figure 2) In particular, for African American women we found significantly increased stroke incidence compared to whites among those less than 70 and greater than 70. Different from our prior incidence study, using the cohort study we were able to adjust for various baseline risk factors. We documented a decrement in some of the risk differences when you began to adjust for education, socioeconomic status, and other vascular risks.

Figure 2:

Age-adjusted stroke incidence rates per 1000 person-years for white, Black and Hispanic men and women in the Northern Manhattan Study stratified by age 70.

Risk Factor Disparities and Stroke Incidence

How much of stroke disparities are explained by differences in stroke risk factors? REGARDS has nicely demonstrated that when you control for various risk factors, hypertension being one of the most important, you reduce the disparity between Blacks and whites.²² Hypertension and diabetes from REGARDS are the largest contributors to the disparity. Moreover, the incremental risk associated with rising systolic blood pressure is much greater in African Americans than Whites at various levels of blood pressure, indicating that BP may have a greater impact.⁴⁶

In NOMAS, we collected a variety of risk factors, including traditional factors such as behavioral, diet, social factors, metabolic factors, and cardiac factors, and more novel ones such as residential proximity to major roadways, inflammation markers and infectious burden.^{47,48,49,50–52,53,54,55,56,57,58,59,60} Our goals were to evaluate the impact of various factors on the risk of stroke and evaluate for any differences across race and ethnicity.

Besides individual risk factors, we were interested in evaluating the combination of risk factors. When the concept of the metabolic syndrome was introduced as a combination of multiple different risk factor levels that may be a marker for insulin resistance, we evaluated that in our cohort. Like some of our other risk factors, we found a greater impact of this combination marker among our Hispanic group with a greater hazard ratio, prevalence, and etiologic fraction as it relates to stroke risk in our Hispanic group.⁶¹ We also developed and published our own Global Vascular Risk Score predictive of stroke, MI and vascular death that included some behavioral variables (physical activity and alcohol use) and continuous measures of BP, fasting glucose, cholesterol, and waist circumference.⁶²

In 2010, the AHA launched a new strategic 2020 goal and defined ideal cardiovascular health. The AHA expanded their mission and made it clear that they were focusing on improving cardiovascular health.^63,64 The goal was launched to improve cardiovascular health by 20% for all Americans which was an unequivocal emphasis on achieving health equity that has become an even greater focus with the AH 2030 goal. They defined for the first time ideal cardiovascular health, or what is known as Life’s Simple 7. Ideal cardiovascular health was defined by seven factors: never smoking, body mass index less than 25 Kg/m², physical activity of at least 150 minutes of moderate or 75 minutes of vigorous activity each week, four to five key components of a healthy diet consistent with AHA guidelines, total cholesterol less than 200 mg/dL, blood pressure below 120/80 mm Hg, and a fasting blood glucose less than 100 mg/dL. To be ideal, you had to have these normal levels without medication. Ideal diet factors were: fruits and vegetables 4.5 or more cups per day, fish consumed two or more times per week, fiber-rich whole grains of 3 or more servings of per day, sodium less than 1500mg per day, and sugar-sweetened beverages less than 450 kCal (36 oz) per week.

In our NOMAS cohort, we evaluated the prevalence of Life’s Simple 7, and nobody in our cohort had all 7 factors.⁶⁵ When using four or more ideal health factors as a category, only 15% of women and 25% of men would have met that definition. There were differences in the prevalence of having four or more factors by race-ethnicity: 29% for White, 21% for African American, and 14% for Hispanic. Ideal cardiovascular health as defined by the American Heart Association (AHA) had clear disparities. We evaluated the incidence of stroke, MI, and vascular health death stratified by the levels of ideal cardiovascular health. Overall, those with the greater number of ideal cardiovascular health factors compared to those who had 0–1 ideal factors had a quantitative decline in incidence of stroke, MI, and vascular death. Moreover, this improvement in risk with increasing numbers of ideal factors was observed across each of the three race-ethnic groups. If we could shift more African Americans and Hispanics into ideal cardiovascular health, we could greatly reduce existing disparities.

In NOMAS, there were some differences in the prevalence of individual ideal cardiovascular factors. Smoking, for example, was worse for African Americans, while all three race-ethnic groups were equally poor for diet. Nobody had ideal diets in our cohort. Blood pressure was worse for African Americans and Hispanics, as well as blood glucose. Physical activity was worse for the Hispanic group. These observations suggesting heterogeneity for stroke risk factors across race-ethnicity call for tailored risk factor modification through innovative portals to shift more diverse subjects to ideal cardiovascular health.

The association of ideal cardiovascular health with vascular outcomes including stroke has been demonstrated in REGARDS and Atherosclerosis in Risk Communities (ARIC).^66,67 The Multi-ethnic Subclinical Atherosclerosis Study (MESA) has shown that the quantitative risk with increasing number of ideal cardiovascular health factors applies across race-ethnicities.⁶⁸ It is clear that management of lifestyle and risk factors are critically important and that if we could get more people into ideal health, we could eliminate disparities.

Despite some reported declines in stroke incidence, stroke disparities have persisted. GCNKSS showed declines in stroke incidence in whites, but not Blacks from 1993 to 2005.⁶⁹ BASIC has reported on a decline in ischemic stroke incidence from 2000–2010 in whites and Mexican-Americans in Texas, but still noted persistent disparities that were more prominent for those under age 75, and possibly attributed to a greater impact of diabetes in mid-life.^70,71 In ARIC they reported on a decline in stroke incidence and mortality from 1987 to 2011 for both whites and Blacks likely attributed to better risk factor control, however disparities between white and Blacks persisted.^72,73 In an accompanying editorial, we noted that “although there has been progress in reducing smoking and lowering blood pressure and cholesterol, formidable challenges to address stroke disparities and successfully control risk factors and lifestyle behaviors across race, ethnicity, and regions persist. Unless health disparities are addressed and innovative strategies to change behavior are developed and adopted, the cerebrovascular health of the population will be unlikely to improve.”⁷⁴

Moving from Observations to Actions

How do we evolve from observations to actions? Conceptual models to address health disparities have called for moving from detecting disparities in disease incidence or prevalence, to determining the underlying causes of disparities and developing interventions, and then to testing interventions in human populations.⁷⁵ The process is to develop, demonstrate, and disseminate these interventions to mitigate disparities. Actions to address race and ethnic stroke disparities have included innovative risk factor interventions, stroke awareness campaigns, quality improvement programs, workforce diversification, and accelerating policy changes.

The Stroke Prevention Intervention Research Program (SPIRP) was an example from the NINDS of investing in four centers with a goal to develop and then implement high-impact culturally appropriate intervention and prevention programs aimed at minority race-ethnic groups. Each of the 4 centers developed a number of pilot research projects to address disparities.

Kaiser Permanente and UCSF developed an innovative program to eliminate disparities, particularly in African Americans. Management of blood pressure, as REGARDS and other studies have demonstrated, would be a key risk factor to target with novel interventions. “Shake, Rattle, and Roll” was a cluster randomized trial among adult primary care physicians and their hypertensive African American patients with an individualized diet and lifestyle coaching intervention with stepwise behavior change, that showed sustained blood pressure management 2 months after intervention.⁷⁶ The team also did some impact modeling to see if this approach could be expanded across larger populations and have an impact.

At UCLA and USC, they have worked on a Chronic Care Model intervention with a strong community-based component that we know is critically important for reducing disparities. They created a program called “Secondary Stroke Prevention by Uniting Community and Chronic Care Model Teams Early to End Disparities (SUCCEED).”⁷⁷ Their intervention included a community-centered component and mobile health information technology. Although they weren’t able to show significant difference in the primary outcome of systolic blood pressure, they did find a greater utilization of home blood pressure monitoring and statins. They also worked on “Worth the Walk” – a culturally tailored 1-month intervention to increase walking and stroke knowledge among Latino, Korean, Chinese, and African American seniors.⁷⁸ Although they did not demonstrate an increase in walking levels among seniors, they did find improvement in stroke preparedness. The research teams also focused on Community Engagement Symposiums that were critical to their success. The conceptual models that many of the SPIRP programs utilized often involved the community health worker, engaging the community, and working through specific mediators (use of blood pressure monitoring, increasing medication adherence, stroke literacy, self-efficacy, and self-management skills) to ultimately improve physiological risk factors and lifestyles.⁷⁷

Another research program by this same group was entitled “Systemic Use of Stroke Averting Interventions (SUSTAIN).”⁷⁹ They used a community-based approach including group clinics to get stroke and TIA survivors to improve their risk factor control. At 12 months, systolic pressure was only slightly better in the intervention group compared to usual care, but low-density lipoprotein levels were significantly lower. These pilot studies also have identified important barriers that need to be addressed prior to community implementation such as transportation (29% did not attend group clinics because of transportation issues). In our urban, diverse communities, it is not easy to get to the main academic center for these meetings, so we have to think of innovative ways to circumvent these barriers. Time spent helping patients navigate the healthcare system takes extra efforts in these research programs, but probably does have benefits.

At NYU and Columbia, they worked on a number of programs regarding implementation of evidence-based programs for blood pressure management and stroke education.⁸⁰ They evaluated home blood pressure telemonitoring among minority stroke survivors and whether it can be supplemented with individualized, culturally tailored telephone-based nurse care management to lead to better blood pressure control and reduce stroke recurrence. They also created the “Center for Stroke Disparities Solution Community Transitions Intervention (CSDS)” by using a mixed-method design to evaluate the effectiveness of a culturally-tailored transition care program delivered by nurse practitioners and community health workers in improving blood pressure control, functional outcomes, and quality of life in homebound post-acute stroke patients.⁸¹ They also studied “Tailored Approaches to Stroke Health Education (TASHE)” by using storytelling in the form of films that were shown at African American and Hispanic churches across New York City.⁸² Olajide Williams was involved with this group and has done amazing work with the stroke hip-hop program.⁸³ This culturally-tailored approach to improve stroke awareness among younger people can be an effective, intergenerational model for increasing stroke preparedness among economically disadvantaged minorities. Such programs can be more broadly implemented through school-based public stroke education intervention programs. We need more of these innovative programs to reach people across the lifespan.

The American Stroke Association’s (ASA) Power to End Stroke, a campaign that Ed Kenton played a role, was another excellent example of a culturally tailored awareness campaign. The aim was to get people from the community to be power ambassadors and speak to their own communities about the importance of stroke to increase education and awareness. Another example of culturally sensitive messaging was done with the ASA FAST Awareness Campaign. The Ad Council created a powerful commercial showing a black woman with a stroke and a reverse time lapse film with the theme “I wish I could go back, I would have called 911.” (https://www.youtube.com/watch?v=0B4Xy806Lq8) We need many more of these culturally tailored awareness campaigns to improve stroke education.

In our SPIRP program at the University of Miami and University of Puerto Rico, we focused on a quality improvement program to address stroke disparities.⁸⁴ Our project was designed to unite all the hospitals providing acute stroke care and participating in the AHA Get With the Guidelines Stroke (GWTG-S) program in Florida and Puerto Rico to create the Florida Puerto Rico Collaboration to Reduce Stroke Disparities.⁸⁵ (Figure 3) Our goal was to facilitate the improvement of stroke care delivery and achieve the best patient outcomes across all race-ethnic, sex and regional groups. The aims were to collate stroke data, identify any disparities in acute stroke care, develop interventions to reduce disparities, demonstrate improvement, and disseminate best practices by leveraging multilevel collaborations across institutions, agencies, and the community. The conceptual model was based on collating data across multiple sources, using the data to track how hospitals were performing, benchmarked to regional, state and national metrics, and identifying disparities.(Figure 4⁸⁶) We could then track, analyze, and identify disparities and apply the three D’s: develop, demonstrate, and disseminate interventions. Our focus was on the healthcare system utilizing the quality improvement loop that is central to GWTG-Stroke to improve evidence-based care and eliminate disparities. Although we have not yet focused on the other determinants of health including the individual and the community, we did influence state policy. We have changed policy in the state of Florida based on our data to now require all stroke centers in the state to participate in the Florida Stroke Registry and we receive annual appropriations from the Department of Health.

Figure 3:

Florida Stroke Registry Hospitals 2019. Pin drops represent the locations of the 121 FSR participating hospitals. The various pin drop colors represent the FSR hospital’s stroke certification type. (Comprehensive Stroke Certification, Thrombectomy Stroke Capable, Primary Stroke Certification, Acute Stroke Ready, Not currently certified)

Figure 4:

Conceptual framework of the Florida Stroke Registry based on socio-ecological model of the spheres of influence.

The Florida Stroke Registry adapted performance metrics that were part of the GWTG-Stroke program and added others such as lifestyle interventions, treatment performance metrics, and outcome metrics. Our main intervention was to create and disseminate hospital disparities dashboards. Many people have stated “what gets measured, gets done.” If you measure and track performance and provide some benchmarks regarding stroke disparities, it is the first step in making providers aware of where improvement is needed. We created annual hospital dashboards that go across every hospital so they can see their stroke performance for White, African Americans, and Hispanics. (Figure 5) We have annual meetings and webinars, interactive educational modules, and communication portals to share best practices to improve performance and reduce disparities. We measure, track and provide reports on sex disparities and have analyzed differences by region.

Figure 5:

Annual Hospital Disparities Dashboards demonstrates overall stroke performance and by race-ethnic group benchmarked to region and state.

The Florida Stroke Registry confirmed that when you adhere to evidence-based guidelines, performance improved over time and disparities were reduced. ^84,87 In 2010, there were more stroke disparities in defect-free care and other metrics, but these were less apparent the longer the hospitals participated in the program by 2017 and 2018. We did identify differences for Puerto Rico, which unfortunately we have not been able to continue in the study as the Florida Department of Health assumed funding. We were also able to show that there were door-to-needle time differences that improved over time after following evidence-based guidelines.⁸⁸ There are still some significant disparities with worse door-to-CT times and rt-PA treatment utilization in the 4.5 hour window for African Americans than our White groups in Florida.(Figure 6) We created and disseminated an interactive educational module on door-to-needle time that could be deployed in centers to help them learn about the evidence to reduce door-to-needle time using culturally appropriate characters and providing the data that will change practices locally.

Figure 6:

Temporal trends in IV rt-PA utilization within 4.5 hours among eligible patients arriving within 3.5 hours from 2010 to 2018 across race and ethnic subgroups showing that non-Hispanic Blacks were less likely to receive treatment compared to non-Hispanic Whites.

There are other major actions needed to reduce stroke disparities. We need to advance our Diversity Pipeline of healthcare providers and researchers to address and reduce stroke disparities. How do we get more people of diverse backgrounds into our field? We need to support diversity initiatives across the professional lifetime. At past-President of the American Academy of Neurology, we expanded a number of programs including the diversity pipeline with some recommendations to the AAN to engage diverse members, advocate for under-represented members, and provide patient support and promote initiatives to increase diversity in neurologic research. We need to do everything we can to get people in our field, support their career development, and assure that they are successful. Ed Kenton would have been very excited about the goals and plans of this HEADS-UP program.

Lastly, we all need to be public policy advocates for change. We need to go to our governmental leaders and let them know how important it is to create policies that will lead to changes to improve health equity. As you know, the AHA 2030 goal is “together, we will equitably increase healthy life expectancy from 66 to at least 68 years by 2030.”⁸⁹ We need to improve health equity across all race-ethnic groups. The policy and environmental change advocacy strategies that can help achieve these goals are: healthier eating and outdoor physical activity, reduce tobacco and youth vaping, effective diagnosis, treatment, and control of risk factors, expanding care coverage for prevention benefits, and affordable housing, employment opportunities, and opportunities to increase educational attainment. These are fundamental policy changes that can address disparities and all of us need to be advocates for these changes.

The observations and data are clear that African Americans and Hispanics have a greater stroke risk. Differences in risk factors contribute to these disparities and improving ideal cardiovascular health for all groups could reduce stroke disparities. Innovative stroke risk factor intervention and awareness campaigns need to be culturally tailored to more effectively reduce risk. Enhanced participation in quality improvement programs such as GWTG-Stroke can also markedly reduce disparities. We all need to be advocates for actions to enhance the diversity pipeline and advocate for policy and environmental change to improve health equity.

In closing I would also emphasize that progress requires teamwork. I have had the terrific advantage of working with tremendously talented people over the years who have made a huge difference in my career and I am forever thankful to them. I have had the benefit of working with great interdisciplinary teams in New York at Columbia and at the University of Miami. I am also thankful for the support of the NINDS for the Northern Manhattan Study and the initiation of our Florida Stroke Registry. With teams working together, we can effectively make the changes needed to eliminate stroke disparities. Thank you.