Skip secondary menu

Proceedings of the Stroke Disparities Advisory Panel Meeting

National Institute of Neurological Disorders and Stroke

Proceedings of the Stroke Disparities Advisory Panel Meeting

November 7-8, 2002
Bethesda Marriott
Bethesda, Maryland

Table of Contents

Executive Summary
NINDS Stroke Disparities Advisory Panel Meeting
November 7-8, 2002
Bethesda, Maryland


In 2001, the NINDS charged its Stroke Progress Review Group (Stroke PRG) with the task of setting overall priorities for stroke research. Because racial and ethnic disparities in the stroke burden present a critical research and treatment challenge that must be addressed, the Stroke PRG in its April 2002 report discussed the need to develop a research agenda on stroke disparities. In response, NINDS formed a Stroke Disparities Advisory Panel, which convened a preliminary meeting on June 11, 2002, to plan a full-scale workshop on future research directions in stroke disparities. The workshop took place on November 7-8, 2002, in Bethesda, Maryland. During this meeting, nine panels of experts were formed to discuss and consider recommendations for advancing research on stroke disparities.


Stroke ranks as the third leading cause of death and is the leading cause of serious, long-term disability in the United States . Racial and ethnic minority populations, in some age groups, have a higher relative risk of stroke death when compared with the U.S. non-Hispanic white population. For example, among African-Americans, the relative risk of stroke death is 4 times higher at ages 35-54, 3 times higher at ages 55-64, and almost 2 times higher at ages 65-74. The disparities in stroke mortality rates closely parallel racial and ethnic differences in the prevalence of stroke risk factors. For example, compared to white Americans, African-Americans have higher rates of hypertension, diabetes, smoking, obesity, and physical inactivity. These differences often persist even after statistical adjustments are made for differences in the socioeconomic status.

A number of possible mechanisms involved in stroke disparities have been hypothesized. They include a wide range of cultural and environmental factors, such as racial and ethnic variations in lifestyle, access to healthcare, quality of healthcare received, differences in health beliefs, religiosity, health literacy, adherence to prescribed therapy, stress, and exposures to environmental toxins. In addition, recent progress in stroke genetics has enabled exploration of possible racial and ethnic differences in genetic susceptibility to stroke or stroke risk factors. The panels reviewed the state of scientific evidence on the influence of environmental and genetic factors, and their interactions, on stroke incidence and stroke outcomes.


The discussions regarding current state of research on stroke disparities were conducted by nine panels, each focusing on one of the following topics:

  1. Temporal Trends in Disparities in Stroke Incidence, Mortality, and Quality of Life
  2. Disparities in Stroke Risk Factors and Mechanisms
  3. Access to Health Care/Quality of Care
  4. Genetics and Vascular Biology
  5. Design of Clinical Trials Addressing Stroke Disparities
  6. Bridges to the Community: Assuring Ethical Conduct of Studies and Data Integrity
  7. Primary and Secondary Stroke Prevention
  8. Acute Stroke Treatment
  9. Rehabilitation/Outcomes

Each panel summarized the state of current knowledge within their field of expertise and identified barriers and challenges that hamper our understanding of stroke disparities. The panels specified data needs and scientific standards of evidence necessary to design effective interventions and to develop reliable methods of measuring progress toward the elimination of stroke disparities. Several common themes have emerged across the panel discussions:

  1. Data needs

    The panels have consistently noted the urgent need to collect more data on racial and ethnic stroke disparities, going beyond the well-documented differences in stroke mortality. Racial and ethnic disparities in stroke mortality rates may reflect disparities in incidence rates or case-fatality rates, or both. However, stroke incidence and case fatality data from population-based studies that include substantial numbers of minorities are limited. Data regarding temporal trends in stroke incidence in minority populations are particularly important for measuring progress toward elimination of stroke disparities.

    Evidence to date has suggested possible racial and ethnic differences in the incidence of various stroke subtypes. For example, existing data suggest greater risk of lacunar and intracranial atherosclerotic stroke in African-Americans and Hispanics. Therefore, the panels have emphasized the need to collect further data on racial and ethnic differences in the incidence of different stroke subtypes and the relative contribution of different risk factors to increased rates of various stroke subtypes.

    There is strong and consistent evidence of low levels of implementation of established guidelines for primary and secondary stroke prevention by healthcare providers, and low levels of patient adherence to physician-prescribed medical therapy or changes in lifestyle. These shortcomings have greatly reduced the potential of existing evidence-based knowledge to reduce the stroke burden. However, little information is available on racial and ethnic differences in the quality of preventive care, including possible provider bias, and patient compliance as mechanisms contributing to the disparities in stroke burden.

    Numerous studies have documented racial and ethnic differences in medical care received for various conditions. For example, evidence shows that African-American cardiac patients are less likely than white American patients with the same condition and similar characteristics to receive diagnostic procedures, revascularization procedures, and thrombolytic therapy. In general, disparities in receipt of appropriate care remain after adjusting for factors known to affect care, such as age, sex, insurance status, co-morbidities, and disease severity. Despite the perception that there are racial disparities in acute stroke treatment, currently little is known about possible ethnic and racial differences in the treatment of acute stroke. For example, it is not known whether there are racial and ethnic differences in the use of thrombolytics for acute stroke treatment.

    Similarly, very little information exists on possible racial and ethnic differences in the rate of recovery from stroke, stroke recurrence, and functional outcomes and quality of life following stroke. Previous studies have shown that functional outcome after ischemic stroke is affected by many factors. Those most commonly reported include age, pre-stroke disability, severity of stroke and/or level of consciousness on presentation, presence or absence of urinary incontinence after stroke, previous stroke, diabetes or elevated serum glucose, cardiac disease, and degree of social support or marital status. However, in the majority of these studies, race was not considered as a variable potentially affecting post-stroke outcome.

  2. Challenges encountered in research on stroke disparities

    The panels consistently emphasized the need to improve and standardize race and ethnicity definitions and classifications. Definitions of race-ethnicity have been mandated by the government organizations, but may be inadequate and differ across available studies. These definitions were based on US census methods and do not take into account cultural distinctions, heterogeneity among race groups, ethnicity, heritage, and the effects of inter-marriage.

    There are major confounding factors when examining racial/ethnic disparities in the burden of stroke, including socioeconomic status, education, religion, cultural factors, dietary patterns, geographic region, and gender. These factors complicate study design and limit our ability to generalize results from one study to the entire U.S. population.

    Stroke is a heterogeneous disease and comprises a number of pathological conditions. The phenotypic characterization of stroke in research is a controversial issue that will require a diversity of approaches. While there is consensus that ischemic stroke, intracerebral hemorrhage, and subarachnoid hemorrhage should be addressed separately, an important question is whether there are subphenotypes or intermediate phenotypes that have a particular relevance to racial and ethnic differences in stroke risk.

    There are many operational challenges to conducting research in different racial and ethnic groups. The lack of legal protection against genetic discrimination in the United States , and general mistrust among minorities toward the medical and research establishment, are substantial impediments to the inclusion of racial and ethnic minorities in biomedical, especially genetic research. Participation in clinical trials is declining in numerous disease areas. Unfortunately, recruitment strategies have focused on problem-solving strategies rather than on proactive, specific strategies targeted to reach historically underserved populations.


Scientific Priorities

  • Epidemiological research

    Characterize temporal trends in racial and ethnic disparities for stroke mortality, incidence, and prevalence rates by age, gender, social class, social environment, and geography.

    Identify population attributable risks and interactions among known risk factors and emerging stroke risk factors in multicultural populations, accounting for differences in sex and socioeconomic status.

    Determine whether intermediate markers of stroke risk are valid markers in different racial and ethnic populations and potential targets for prevention trials.

  • Research on genetic and environmental determinants of stroke risk

    Provide characterization of variation in candidate genes and pathways within each racial and ethnic group, based on adequately powered studies within each racial and ethnic group, and with appropriate consideration of the potential for population stratification bias.

    Examine how genetic and environmental factors and their interactions contribute to the incidence of stroke risk factors and subtype-specific strokes.

    Study racial and ethnic differences in response to drugs used to treat acute stroke, as well as those used for primary and secondary prevention.

    Focus translational research on similarities and differences between racial and ethnic groups in important vascular biology mechanisms, including endothelial function, biomarkers of inflammation, coagulation/fibrinolysis, angiogenesis, and oxidative stress. Further characterization of these phenotypes between racial and ethnic groups must accompany genetic studies of these biological mechanisms.

  • Health services and patient management research

    Evaluate the independent contributions of factors relating access to quality care and stroke health disparities, such as health insurance status, acculturation, religious and spiritual factors, and healthcare provider attitudes. Research focused on prejudice and cultural competency on the part of health care professionals and their influence on access to care and stroke disparities should be undertaken. Investigate whether quality-of-care measurements and organizations can effectively ascertain the extent and possible causes of certain of the stroke-care disparities.

    Evaluate barriers to adherence to stroke prevention strategies that are specific to minority groups.

    Develop and test for effectiveness of novel cost-effective and practical stroke prevention programs for multicultural groups that can be adapted for use within communities.

    Collect data on access to acute stroke care among racial and ethnic subgroups. Evaluate the extent to which socioeconomic and cultural factors, level of public awareness, physician biases, and patient co-morbidities may influence acute stroke care.

    Using multidisciplinary research and considering a variety of sociological, environmental, and access variables, conduct research to learn how race affects access to and quality of stroke rehabilitation services. Investigate whether racial differences occur in the recovery trajectory and in optimal recovery after stroke; and determine whether disparities exist in other long-term sequelae of stroke and what factors influence these outcomes.

Research Methodology

Study diverse racial and ethnic groups in different geographic, social, and cultural environments. Cross-cultural studies or migration studies of racial/ethnic groups are useful scientific strategies to disentangle the role of race as a confounder reflecting both environmental and genetic differences. The effect of race/ethnicity on disease outcome should be examined after stratifying on numerous potential confounders, including measures of socioeconomic status, education, and health care access and utilization.

The need to increase minority representation in clinical trials remains critical. In addition to increasing sample size of minority groups, several strategies were suggested to increase the statistical power to characterize race as an effect modifier. For example, statistical power can be increased by the use of surrogate endpoints, such as a composite endpoint of clinical or silent (MRI) strokes. Another approach might be to expand the sample size to test race as an effect modifier, while minimizing the complexity of the study. In addition, the panels urged standardization of data elements on stroke risk factors and stroke outcomes, and data sharing across studies.

Resources Needed

The foundation toward characterizing the public burden of stroke it to develop a national surveillance system to establish and compare incidence and prevalence rates for stroke, both overall and by stroke subtype, for all major racial and ethnic subgroups. This system should include population-based studies in multiple well-characterized communities in different geographic regions throughout the United States . The surveillance system should employ common methodology and definitions.

Develop additional measures of race and ethnicity that can complement current methods of self-reporting and ensure consistency in reporting race across epidemiologic studies.

Strengthen the regulatory environment that protects against genetic discrimination of individuals and named populations.

Develop Stroke Prevention Centers of Excellence, to test interventions aiming to improve screening for risk factors, increase implementation of risk factor management guidelines by healthcare organizations and healthcare providers, and to support patient adherence. Such interventions should be designed by multidisciplinary teams, drawing on experience from diverse fields of research on healthcare delivery, healthcare outcomes, and health behavior.

A more proactive approach to minority recruitment, both investigator and patient, was recommended. Several strategies were suggested, including providing diversity training for investigators to increase their awareness of and sensitivity to cultural norms of target populations; creating a "special populations" advisory group to work with researchers; and making modifications in the NINDS guidelines to establish research requirements for community engagement.


Panel 1
Temporal Trends in Disparities in Stroke Incidence, Mortality, and Quality of Life

Co-Chairs: Joseph P. Broderick, M.D., and George Howard, Dr.P.H.

Ann M. Malarcher, Ph.D.
Lewis B. Morgenstern, M.D.
George Petty, M.D.
Philip Wolf, M.D.

Statement of the Problem

Although stroke is the third leading cause of death in the United States our current understanding of stroke burden among groups and over time is insufficient to provide the foundation needed to implement effective strategies to reduce stroke mortality and morbidity in the foreseeable future. Members of certain racial and ethnic groups such as African Americans bear a substantially greater share of the burden of stroke than do Caucasian Americans as measured by stroke mortality rates in this country. However stroke incidence data from population-based studies that include substantial numbers of minorities are limited as are data regarding temporal trends in stroke incidence in these minority populations. Few studies have performed assessments of quality of life following stroke in minority populations. Stroke prevalence data concerning racial and ethnic disparities are limited to data from a single question on a national survey asking whether a person has ever been told that they have had a stroke. Without the appropriate data strategies for eliminating racial disparities in stroke are unlikely to be effective and progress will be poorly measured.

  • Racial and ethnic designations in epidemiologic studies of strokes are usually obtained through self-reporting; however, such data are often missing in administrative databases. Methods of reporting race and ethnicity may vary between different epidemiologic studies of stroke.
  • Major confounding factors are found when examining racial and ethnic disparities in the burden of stroke, including socioeconomic status, education, cultural factors, geographic region, and gender. These factors complicate study design and limit our ability to generalize results from one study to the entire U.S. population.
  • Stroke is heterogeneous and comprises a number of pathological conditions; these conditions share the common manifestation of brain injury caused by disease of the cerebral vasculature. Information regarding racial and ethnic disparities is particularly lacking regarding incidence rates and mortality of specific stroke subtypes. Race and ethnic disparities in stroke may vary by major type of stroke (e.g., intracerebral hemorrhage versus ischemic stroke) and even by subtype of ischemic stroke (cardioembolic versus large-vessel atherosclerosis). Incidence studies thus must include stroke subtyping through brain imaging and other diagnostic tests.
  • Racial and ethnic disparities in stroke mortality rates may reflect disparities in incidence rates or case fatality rates or both. Incidence and case fatality data from population-based studies of biracial or multiracial populations are needed to address this issue.

Challenges and Questions/ Barriers

Lack of Data

Overall stroke mortality data are available for the various racial groups by age, gender, and geographic region, but they are limited regarding specific stroke subtypes. Incidence data by racial subgroup has been mostly limited to predominantly all-Caucasian, middle-class populations such as the Rochester Epidemiology Project or cohort studies such as the Framingham Heart Study, which may not adequately reflect incidence data from minority populations or other geographic regions.

The Greater Cincinnati/Northern Kentucky Stroke Study and the Northern Manhattan Stroke Study have reported stroke incidence rates in biracial and multi-ethnic populations at a single point in time, but temporal trends in stroke incidence rates in these diverse populations have yet to be reported. Several NIH-funded population-based studies are in progress, such as Multi-Ethnic Study of Atherosclerosis (MESA), Jackson Heart Study, Reasons for Geographic and Racial Differences in Stroke (REGARDS), and a study of non-Hispanic Caucasians and Hispanic Caucasians in Corpus Christi, Texas, but incidence rates from these studies have not yet been reported. Minimal data are available regarding incidence data for other minority populations such as Asian Americans and Native Americans. These population-based studies have published very limited information concerning measures of functional outcome and quality of life by racial subgroups.

Cohort studies such as the Atherosclerosis Risk in Communities study and the Cardiovascular Health Study are another source of incidence data and case fatality rates in different racial and ethnic subgroups, but they are not population-based studies. The advantage of the cohort studies are well-described profiles of risk factors in addition to subsequent case events. A disadvantage is the small number of cerebrovascular events, particularly for certain stroke subtypes.

Lack of Accurate Subtype-Specific Data

Published data in these population-based studies regarding incidence rates of stroke subtypes by racial and ethnic subgroups are very limited. For less common stroke subtypes such as subarachnoid hemorrhage, the study population needs to be relatively large in order to examine differences in racial and ethnic subgroups.

Inconsistency in Classification of Race and Ethnicity

Race is not a biological construct; it is a social one. There is also inconsistency is how race is classified between studies and in the presence of racial classification in certain databases (e.g., hospital discharge databases). Mortality for certain racial subgroups is substantially underestimated because of misreporting of race and ethnicity on the death certificate.

Need to Understand Cultural Contexts

Race and ethnicity are closely associated with socioeconomic status and cultural factors. Cultural contexts needed to understand differences by race and ethnicity and socioeconomic status include measures of discrimination, stress, acculturation, language barriers, access to healthcare, and others.

Access to Technology

Technology, such as computed tomography imaging and magnetic resonance imaging, can change detection and classification of stroke and stroke subtype and thus can affect measurements of temporal trends in the incidence rates of stroke overall and for subtypes. If access to this technology varies by race/ethnicity/socioeconomic status, measurements of incidence and prevalence rates of stroke also may be affected.

Poor Understanding of the Epidemiology of Outcome and Quality of Life Following Stroke

Epidemiology of outcome as characterized by cognitive and functional disability and quality of life following stroke is lacking. The identification of factors associated with positive functioning outcomes is the first step of providing interventions that may maintain or improve the quality of life for the stroke patient, as well as for caregivers, following stroke events.

Research and Scientific Priorities

Priority 1:

Characterize temporal trends in racial and ethnic disparities for stroke mortality, incidence, and prevalence rates by age, gender, social class, social environment, and geography.

To characterize the public health burden of stroke, develop a national system to establish and compare incidence and prevalence rates for stroke, both overall and by stroke subtype for all major racial and ethnic subgroups. This surveillance system should include population-based studies in multiple well-characterized communities in different geographic regions throughout the United States. This system needs to provide detailed information on incidence rates, with strata defined by age, geographic region, and race and ethnicity. It should employ common methodology and definitions to the extent possible. These strata-specific incidence estimates will serve as the foundation for estimating the proportion of the well-known differences in stroke mortality rates between the strata that are attributable to differences in incidence, rather than case fatality, as well as the case-mix of stroke subtypes between the strata.

Of equal importance, this resource will provide the mechanism for prospectively tracking temporal changes in incidence, allowing for appropriate shifts in resources and research efforts in response to changes in the incidence rates (both overall and by stroke subtype), as well as shifts between the strata defined by demographic factors including racial and ethnic groups.

Priority 2:

Measure differences in stroke outcome and quality of life for different racial and ethnic subgroups.

A well-designed national system should also include cohorts of the various racial and ethic groups of stroke patients that can be prospectively followed to measure outcome and quality of life. Further methodologic research is needed regarding measurement of quality of life for stroke patients in general as well as the impact of stroke among diverse racial and ethnic subgroups. For example, racial and ethnic subgroups may place different values on various aspects of the quality of life following a stroke.

Information regarding functional outcome and quality of life from these cohorts will serve the secondary aim of establishing the magnitude and determinants of the public health burden associated with the post-stroke period for the racial and ethnic subgroups. Specifically, these cohorts can be used to estimate the costs associated with long-term treatment and the mortality, recurrence, and morbidity (including both recovery and cognitive decline) associated with the stroke event. For each of these domains, the cohort also can be used to establish the determinants that place stroke patients at risk for differential outcomes.

Finally, links between the national surveillance cohort and administrative databases (e.g., Medicare, Medicaid) can be investigated to provide determinants of cost-effectiveness and patterns of resource utilization on the national level.

Priority 3:

Find additional measures of race and ethnicity that can complement current methods of self-reporting and ensure consistency in reporting race across epidemiologic studies.

Race and ethnicity are social constructs that include multiple variables such as ancestry, migration, and acculturation. To compare stroke occurrence and outcome between diverse populations in various geographic regions, consistency of definitions and of reporting is critical. Self-reporting of race or ethnicity will likely remain the primary basis for comparison. However, all epidemiologic studies should indicate the number of cases for which racial and ethnic designation is missing.

Even within racial and ethnic groups, substantial variability may occur in ancestry and genetics as well as cultural environment. Additional information, such as ancestry and migration history, as well as genetic markers, will be useful for comparisons between similar racial or ethnic groups (by self-report) from different geographic regions.

Resources Needed

  • Maintenance of current population-based studies of both Caucasian and non-Caucasian populations and the addition of other population studies (cohort, case control, etc.) for measurement of temporal trends in stroke incidence and prevalence rates as well as determinants of outcome in targeted racial and ethnic subgroups. In particular, additional research should focus on currently understudied minority populations.
  • The development of mechanisms for the linkage of information collected in the epidemiological studies, including the development of common methods and infrastructure for sharing stroke data by "outside" investigators. As part of this resource, consistent documentation, ideally using common definitions, needs to be developed for domains assessed in the studies.
  • The development of links with epidemiological databases from other sources (e.g., insurance companies and health maintenance organizations), including improved methods for clinical informatics.
  • Research approaches that ensure access to patient data while protecting patient confidentiality.


Panel 2

Disparities in Stroke Risk Factors and Mechanisms

Co-Chairs: Ralph L. Sacco, M.S., M.D., Herman A. Taylor, M.D., and K.S. Lawrence Wong, M.D.

Robert Adams, M.D.
Marco DiTullio, M.D.
Karen L. Furie, M.P.H., M.D.
James M. Galloway, M.D.
Edgar J. Kenton, M.D.
Stanley Tuhrim, M.D.

Statement of the Problem

Although stroke is the third leading cause of death in the United States , our current understanding of its etiology and variation among various racial and ethnic populations is not sufficient to develop the strategies needed to reduce stroke mortality and morbidity in the foreseeable future. Indeed, the lack of understanding of variations in risk factors and the associated mechanisms producing disparities in stroke incidence, mortality, and morbidity in various racial, cultural, and gender population subgroups represents a major gap in our knowledge. Some of these deficiencies have been noted in NINDS’ Report of the Stroke Progress Review Group, published in April 2002. Stroke is a heterogeneous disease. In addition to hemorrhagic stroke, there are different subtypes of ischemic stroke, and studies over the past decade have clearly identified factors associated with increased risk of both hemorrhagic and ischemic stroke. Some epidemiologic studies also have demonstrated racial and ethnic variation in stroke subtypes as well as these risk factors, which include a greater risk of lacunar and intracranial atherosclerotic stroke and a greater prevalence of hypertension and diabetes in African Americans and Hispanics than in nonminority populations.

In addition, recent work has shown that a significant proportion of stroke-prone individuals fail to receive appropriate therapy. Although evidence-based recommendations that address identified risk factors have been widely disseminated, stroke-prone women and minority populations, in particular, generally do not receive appropriate guidance for modifying even well defined, highly prevalent, and readily treatable factors. Thus, the greatest burden of risk for stroke is borne by underserved segments of the population—those least likely to benefit from advances in stroke prevention.

Currently, we do not have the data needed to address these disparities in stroke risk factors and associated stroke mechanisms or to develop an underlying approach to the treatment and modification within such subgroups of the population. The pursuit of these data and approaches will become increasingly essential as the 21st century progresses, as by 2030, those at risk for stroke will be more urban, poorer, and have lower educational levels than those at risk today and will comprise a major portion of the population, placing a tremendous burden on our public health and medical services. The first step needed to be better prepared to effectively prevent stroke in these population subgroups involves developing a more complete understanding of the inter-relationships between race and ethnicity, stroke risk factors, and stroke mechanisms.

Challenges and Questions/Barriers

Major questions and challenges center around the inadequacy of current definitions of race and ethnicity, the lack of minority-specific data on stroke risk factors and mechanisms, inadequate definitions and delineation of racial and ethnic, socioeconomic, and cultural confounding factors, and the lack of data regarding stroke outcome and recurrence risks.

Methodological Inadequacy of Definitions of Race/Ethnicity

Definitions of race and ethnicity have been mandated by government organizations, but they may be inadequate, and they differ across available studies. These definitions have been based on U.S. Census methods and do not take into account cultural distinctions, heterogeneity among racial groups, ethnicity, heritage, and the effects of intermarriage.

Lack of Minority-Specific Data

Although there are a few exceptions, most large-scale epidemiological studies of stroke risk factors and mechanisms focus primarily on majority populations. Further advances in the understanding of these risk factors and mechanisms as they pertain to minority populations will require additional data and the use of appropriate study designs to determine the incidence and outcomes of events and the prevalence and potency of risk factors. Studies to identify new risk factors and describe the incidence of stroke subtypes may provide new data that could help design stroke interventions. In addition, more specific data with regard to stroke subtype and mechanism are needed in order to better understand the differential effects of risk factors among and within racial and ethnic groups. This will require including individuals in studies with a broad representation of the factors, such as education and socio-economic status, which frequently confound issues of race and ethnicity.

Poorly Defined Racial and Ethnic, Cultural, and Socioeconomic Confounding Factors

Attempts to understand the effects of potential confounding factors, such as socioeconomic status, are themselves confounded by inadequate definitions and measures of discrimination, acculturation, language barriers, access to health care, effects of stress, the meaning of disease, and the possibility of health interventions within different cultures.

Inadequate Data Regarding Stroke Outcome and Recurrence Risk Among Minority Populations

Certain minority groups, most notably African-Americans and Hispanics appear to be at increased risk for stroke recurrence when compared with majority populations. The risk factors for stroke recurrence generally are poorly understood, and there is a paucity of racial- and ethnic-specific data in this area. The risk factors for subsequent stroke may differ from those for initial stroke, and, similarly, the factors affecting the disparity in recurrence risk may differ from those affecting initial stroke risk. A more profound understanding of these factors is crucial to addressing disparities in stroke recurrence. For example, little is known about the interaction of cognitive and functional disability following stroke and race and ethnicity as related to the risk of recurrence or other quality of life issues.

Endemic Barriers to Identifying Stroke Risk Factor Disparities

Current barriers, which need to be overcome in order to identify stroke risk factor and mechanism disparities, include:

  • A lack of uniformity in defining population subgroups and specific risk factors/mechanisms.
  • A lack of adequately powered studies to control for confounding variables in multicultural populations.
  • The need to identify culturally sensitive and ethical recruitment techniques to ensure adequate numbers of diverse subjects for such studies.
  • The need to overcome case ascertainment bias.
  • The presence of geographic and social barriers to accessing isolated populations.
  • The presence of cultural barriers to the acceptance of clinical research.
  • The presence of legislative and political obstacles to conducting research— particularly genetic studies—in disparate populations.

Research and Scientific Priorities

The official United States multicultural classification scheme divides the population according to self-reported race and "Hispanic" versus "non-Hispanic" ethnicity, a scheme rife with misclassification problems and one that fails to account for the fact that genetic and cultural differences within these groups exist based on geographic origins, generations spent in the United States, and the potential for multiple racial and ethnic contributions through intermarriage. Studies examining population stratification in genetics suggest that inter-racial differences based on the current schemes may actually be quite small. It is critical that innovative definitions of subgroups incorporate cultural, gender, and socioeconomic factors in addition to race and ethnicity and classify groups based on susceptibility to, and risk from, factors that affect the incidence of stroke.

Priority 1:

Identify population attributable risks and interactions among known risk factors and emerging stroke risk factors in multicultural populations, accounting for differences in sex and socioeconomic status.

The prevalence of individual biological and behavioral risk factors is expected to vary in different racial and ethnic groups and differs by age, sex, and socioeconomic status. In addition, a specific risk factor may contribute differentially to stroke risk from population to population. The population attributable risk takes account of both the prevalence and relative risk of individual factors and is therefore a useful means of determining the extent to which an individual risk factor impacts a population. It is important to examine subgroup-specific susceptibility and unique interactions between risk factors. Studies will need to determine whether the prevalence of the condition and control of the condition vary significantly by race and ethnicity and account for differences in stroke incidence. This information is critical for designing epidemiological studies and clinical stroke prevention trials, as well as for focusing public health resources appropriately.

Priority 2:

Examine how genetic and environmental factors and their interactions contribute to the incidence of stroke risk factors and subtype-specific strokes.

The extent to which genetic susceptibility, as opposed to differences in lifestyle based on cultural mores and socioeconomic status, affects the incidence of risk factors and strokes due to specific mechanisms remains unclear. These factors are not mutually exclusive and, indeed, may have additive or synergistic effects on stroke risk. The explanations for observed patterns of cerebrovascular disease have yet to be elucidated. For example, the predominantly intracranial atherosclerosis observed in Asian populations, once believed to be genetically mediated, has been shown to evolve into the more common extracranial pattern seen in Caucasian Westerners with the alteration of environmental factors, but the underlying vascular biology that results in such changes is poorly understood. Other factors may predispose to other stroke mechanisms such as cardio-embolism or the development of lipohyalinosis in specific populations. It is unknown to what extent genetic regulation of factors such as endothelial function, immune response, or thrombosis could increase or decrease risk in a population independently or through amplification of conventional risk factors. Understanding these complex interactions will afford opportunities for screening and early intervention in high-risk populations.

Priority 3:

Determine whether intermediate markers of stroke risk are valid markers in different racial and ethnic populations and potential targets for modification trials.

Identification of preclinical markers of stroke risk and intermediate phenotypes is extremely important because it allows for the early identification of a high-risk population and provides a surrogate target for risk factor modification. In recent years, several markers of risk have emerged, but it is unclear whether they convey the same prognostic significance in all groups. Because mechanism of stroke may differ across racial and ethnic populations, markers that are gaining widespread clinical application may not be generalizable to all populations. For example, risk mediated through novel pathways (e.g., eNOS, vWf, fibrinogen, CRP, homocysteine, lipoprotein (a), IMT, LVH, among others) may be entirely missed by screening strategies developed through research on or for use in a predominantly non-Hispanic Caucasian population.

Priority 4:

Determine the risks, determinants, and mechanisms of adverse outcomes after stroke, including mortality, recurrence, dementia, disability, and quality of life in different racial and ethnic populations.

Outcomes after stroke may vary by racial and ethnic group and may be influenced by the prevalence and control of risk factors, stroke subtype, and social and behavioral factors affecting adherence. In addition, it is important to examine the contributions of the primary health care provider and the family caregiver to the determination of outcome. Further studies are needed to specifically address the determinants of outcomes after stroke in order to design better stroke outcome modification programs.

Resources Needed

Assuming the goal is to better understand and address racial and ethnic disparities in stroke, especially with respect to risk factors and stroke mechanisms, the following are the resource goals that should be pursued and the resources that will be needed:

The overall resource goals are:

  • To adopt common definitions of terms across studies.
  • To stimulate and support new research to address the stated priorities.
  • To use existing data more effectively by creating a means to access and incorporate diverse sources of information, ranging from administrative data and high-level prospectively acquired data from NINDS, as well as other data sponsored by other granting agencies.
  • To develop innovative methods of pooling data.
  • To eliminate misclassification bias due to inadequate access to diagnostic testing/subspecialty expertise.
  • To expand the interactions among existing programs and facilitate communication across governmental and nonprofit organizations.
  1. Enhanced Data Collection

    Enhanced data collection will require the following:

    • Methods to standardize the operational definitions of race, putative risk factors, and mechanisms; clear definition of risks and mechanisms in future studies to enable data pooling.
    • Data collection from ongoing clinical studies, including randomized trials in diverse racial and ethnic populations.
    • New population-based data collection efforts in various U.S. regions targeting certain racial and ethnic populations.
    • Linked data collection from other administrative data sources. Reliance on investigator-driven special effort in select populations might need to be balanced by broad-scale linking of data from other sources. For example, larger numbers of representative patients can be analyzed through a concerted effort to identify and link disparate and diverse sources of data, including health maintenance organizations, insurance companies, registries, Centers for Medicaid and Medicare Services, Centers for Disease Control and Prevention (CDC), National Heart, Lung, and Blood Institute, Department of Veterans Affairs, Indian Health Service, and state-based data, where available. Attempts should be made to use sources from industry such as HEDIS indicators reported from managed care plans to assess possible differences in racial/ethnic application of stroke-pertinent physician/care plan behaviors (e.g., counseling and screening).
    • Use of data sources not yet exploited, including banked serum and DNA from prior clinical studies (e.g., measuring C-reactive protein or homocysteine in frozen samples from earlier studies with sufficient representation of different racial and ethnic groups).

  2. Centralized Data Collection and Analyses

    To move forward in accomplishing some of the research priorities and resource goals, it may be helpful to develop a centralized, national database for the consolidation of data sources and pooled analyses. This resource could accumulate data from various sources across agencies and organizations to form a data clearinghouse that could be accessed by various groups. This effort will require ongoing infrastructure support.

  3. Expanded Access to Diagnostic Facilities

    Widespread availability and access to diagnostic testing and subspecialty care will be needed to ensure the accurate classification of risk factors and stroke mechanisms in lower socioeconomic populations, for which the lack of availability of certain testing may lead to misclassification and inadequate diagnoses:

    1. Use of noninvasive vascular testing, magnetic resonance imaging and magnetic resonance angiography, and echocardiography, including transesophageal echocardiography, which are often needed to reliably classify stroke mechanisms.
    2. Improved access to subspecialty consultation.
    3. Uniform screening of blood pressure, cholesterol, glucose, weight, and lifestyle factors to identify subjects with undiagnosed risk factors.

  4. Stroke Prevention Research Centers on Racial and Ethnic Disparities

    To address these research priorities and help fill the knowledge gaps regarding racial and ethnic disparities in stroke risk factors and mechanisms, we strongly encourage NINDS to establish specific centers for stroke prevention research. These centers would focus on racial and ethnic disparities and stroke prevention and could include various study designs (e.g., epidemiology, mechanism, pilot trials, basic science, genetic studies) and encompass the range of stroke experience, including stroke precursors and persons at risk, as well as stroke survivors. Just as specific research centers have been established for acute stroke translational research (e.g., SPOTRIAS) and other statewide data collection stroke registries have been initiated (e.g., CDC Coverdell centers), the establishment of stroke prevention research centers would help focus efforts on these issues.

    Such centers would include diverse racial and ethnic and cultural groups and could share definitions of terms and systems and approaches to facilitate the recruitment of under-represented groups. They would facilitate interaction among neurologists, cardiologists, primary care providers, behaviorists, epidemiologists, statisticians, geneticists, and those working in other appropriate disciplines to develop the most effective study designs for addressing disparities in risk factors and mechanisms. The centers should include links to local community hospitals in order to evaluate stroke prevention activities beyond the academic center, and partnerships with other groups, including other federal and state entities as well as nonprofit organizations such as the American Stroke Association and the National Stroke Association, would be encouraged.


Panel 3

Access to Care/ Quality of Care

Co-Chairs: Yvonne Harris M.P.A., and Lewis B. Morgenstern M.D.

Carol Allen Ph.D., R.N.
Hal Barron M.D.
Gloria Bonner Ph.D., R.N.
Edward Cooper M.D.
Cheryl Easley Ph.D., R.N.
Moises Gaviria M.D.
Karen Parko M.D.
Beth Quill M.P.H.

Statement of the Problem

Although access to health care can be thought of as an absence of barriers to obtaining needed health care services—that is, having adequate health insurance and a routine health care provider and health care facility—several other components related to access may be related to the problem of health care disparities among different populations and may increase the stroke burden for certain individuals. These components include socioeconomic status; acculturation, faith, and religion; social determinants of health (including fatalism, optimism, depression); and geographic isolation. A large body of literature shows that members of racial and ethnic minority groups frequently have poor access to medical care and receive lower quality of care than do nonminorities. At the same time, members of minority populations are less likely to have health insurance, have more difficulty accessing health care services, and have fewer provider options than nonminorities, even when they have commensurate insurance and when other factors related to health care access are identical.

The reasons for these disparities are complex, not well defined, and difficult to quantify, but they may reflect a variety of health-related risk factors; differences in access to medical care; and socioeconomic, social and cultural, psychological, and individual patient and provider differences. Access to care as a clear causal explanation for health disparities in stroke remains to be demonstrated, as intervention studies have not yet been conducted to determine whether eliminating access to care deficits will reduce stroke disparities in the United States .

Challenges and Questions

  • Precisely define and prioritize barriers to access to quality health care services for minority populations, and increase research in this area.
  • Formulate and test interventions to improve access to quality health care for minority populations.
  • Apply the most successful and cost-effective interventions, and monitor the overall results against an established baseline for access.

It is important to define the aspects of insurance coverage, both public and private, including prescription drug coverage, that lead to the use of the best stroke prevention and acute stroke practices. Increasing the numbers and improving the location of minority health care professionals are also priorities. Additional consideration and study are needed to determine the impact of increasing the number of minority investigators, as well as minority patients in underserved communities, because the needs and effects may vary for different racial and ethnic groups. Stroke training also needs to be provided to all primary care providers and neurologists. The process of separating out socioeconomic factors and geographic location from access to care will help define the role of race and ethnicity in the area of receiving quality care for stroke, and recognizing acculturation and the stress that accompanies it as a modifier of health care access is critical. The role of faith, religiosity, and social determinants of health require further study to determine their interaction with stroke occurrence and outcome in various populations. It is also important to study emerging populations, including Asian Americans, Arab Americans, and Native Americans, while understanding that not all minority populations have the same access to care issues. Recruiting other federal agencies, industry, and other organizations to partner with NINDS to fund research in access to care and quality care for stroke prevention and acute stroke therapy is likely to enhance opportunities for reducing disparities. In addition, improving the delivery of acute stroke therapy to minority populations is likely to help reduce disparities.


  • Economic Barriers: Financial resources and insurance are insufficient.
  • Service Barriers: Stroke services are lacking, not accessible, or not available.
  • Social and Cultural Barriers: These barriers include language barriers, lack of transportation, lack of awareness of services, service fragmentation, and under-representation of minorities among service providers.
  • Individual Barriers: Individual barriers include bias, discrimination, prejudice, and stereotyping practices at the provider, institutional, patient, and health system levels.
  • Psychological Barriers: These include distrust generated by individual and institutional discrimination and by cultural insensitivity.

A common misconception is that having insurance guarantees access to good care; the reality is that having incomplete insurance is a barrier to accessing quality stroke care and that having no insurance is, of course, a distinct barrier. For example, patients without prescription drug coverage are likely to be less compliant with stroke prevention therapies. Not having a routine health care provider is a common barrier to stroke prevention in underserved areas. Poor stroke training among primary care providers impedes access, as does insufficient training for health care professionals regarding culturally sensitive ways to foster trust. Similarly, few community outreach programs focus on encouraging trust in the medical encounter, a particular problem with acute stroke therapy, which includes an urgent consent process that necessitates such trust. Even within minority groups that have achieved generally high socioeconomic status, psychological stress resulting from racial discrimination is likely to be evident, and this is also true on an institutional level.

Acculturation and the related stress negatively impact health behaviors that are needed for stroke prevention and acute stroke therapy. The role of belief systems, including fatalism and its effects on stroke health behaviors, must be addressed, as must religious practices, which may have any number of effects on health-seeking behavior. Lack of information regarding stroke burden in emerging populations is a major obstacle, and problems involving language, prejudice, and a lack of minority physicians in immigrant groups represent notable barriers. In addition, industry's focus on targeting large groups at risk for stroke may lead to the neglect of minority populations that may benefit the most from stroke prevention services and acute stroke therapy.

Research and Scientific Priorities

  • Descriptive epidemiologic studies are needed to evaluate the independent contributions of factors relating access to quality care and stroke health disparities. A specific focus on research using intragroup and intergroup comparisons is suggested.
  • Research focused on prejudice and cultural competency on the part of health care professionals and their influence on access to care and stroke disparities should be undertaken. At the same time improving trust among all populations in medical care and medical research is a critically important research topic.
  • Multidisciplinary research efforts are needed that combine the efforts of health services researchers, stroke neurologists and other stroke specialists, generalists, emergency medicine personnel, behavioral scientists, and representatives from the populations studied. Population-based studies are the method of choice.


The independent contribution of insurance status separated from socioeconomic status in health care disparities requires continued research, and studies regarding insurance and health disparities among subgroups are needed, as are aggregate studies. Studies of stroke across private insurance plans and publicly funded care, including care provided in managed care systems, outside network systems, and private provider groups, are needed. The method employed should utilize community-based stroke research that explores the impact of community and public health resources on health outcomes. Studying the influence of continuity of care on stroke prevention is critical.


Studies should be undertaken to investigate the different acculturation patterns among minorities in the United States regarding stroke prevention and treatment, keeping in mind the heterogeneity of the different subgroups even within one minority population. Studies are warranted regarding the belief systems among the different racial and ethnic groups to determine their role in health behaviors related to stroke. Studies are needed that test the effectiveness of culturally sensitive interventions in closing the gap between research and health outcomes. Rehabilitation may be a prime focus for studying the role of acculturation and health disparities, as compliance with prescribed therapies for stroke prevention may be influenced by acculturation and acculturative stress. More studies are needed comparing stroke risk in recent immigrants with minority populations that have been in the United States for multiple generations.

Faith, Religion and Social Determinants of Health

Studies on religion and spirituality and stroke are needed that pay careful attention to social and behavioral mechanisms through which religion or spirituality influences health behavior. The relationship between religion and spirituality and the incidence of stroke, stroke risk factors, and lifestyle factors associated with stroke needs to be studied, as does the effect of religion and spirituality on the use of traditional or alternative sources of health care. Religious and spiritual factors that negatively impact health status and behaviors warrant further attention, and determining whether the integration of religious and spiritual issues into the delivery of primary care services would be a valuable approach. Exploring the efficacy of religious organizations as instruments of health promotion, health education, and the delivery of community-based health care services is another important focus of research.

Race and Socioeconomic Status

Better definition of race and ethnicity and socioeconomic status as constructs for research is needed, including consideration of the variations within racial and ethnic groups and the growing category of mixed-race groups. Race may be defined as a biological category for genetic studies or as a social and political category when studying the effect of exposure to individual and institutional racism on the risk of stroke and access to care and quality of care. Ethnicity studies should be used to understand the impact of cultural, lifestyle, and belief systems on risky or health-promoting behavior. Discrimination, skin color, preference in body size, diet, attitudes toward health care professionals, and use of alternative health care and their effects on access to care and stroke disparities are important topics for investigation. Additional studies investigating poverty are needed to relate its effects as a determinant of access to quality care. Poverty should also be studied to determine its impact on lifestyle, how it affects access to recreational facilities, and its contributions to stress and level of education and health literacy, as well as its relationship to emergency medical system response. Investigations of the interactions between race and ethnicity and gender also should be undertaken.

Emerging Populations and Native Americans

More information is needed on the current health care utilization patterns of racial and ethnic minorities for acute stroke treatment and stroke prevention, including knowledge, attitudes, and beliefs about acute stroke and stroke prevention among these groups. The risk of the disease must be clearly delineated in these populations, including the risk that is attributable to limitations in acculturation and access to care. Distinctions must be made between issues that are common to more recently immigrated populations and those inherent to stable United States populations. Emerging populations and Native American populations should be targeted for clinical research for prevention strategies for stroke, as they are currently excluded from such trials by their geographic isolation and their misunderstanding of the goals of such investigations. The communities involved should be queried regarding their willingness to participate in genetic studies, and studies should be encouraged that occur in communities outside of academic medical center communities. The differences in perspectives and recommended treatments between traditional medicine and Western-trained health care providers should be evaluated to understand how the two can interface.

Acute Stroke

More minorities should be recruited to help design, conduct, and participate in acute stroke studies in order to learn more about the value of acute stroke therapies in minority populations, and additional acute stroke care projects involving minority patients and minority health professional organizations should be encouraged. Research is needed to clarify whether discrimination occurs for delivery of acute stroke therapy. Further study is needed to determine the value of a stroke toolbox for health care professionals that includes a module regarding minorities and whether it should be included in such research. An important study focus is the expansion of Internet resources for management of acute stroke. Better studies on intravenous recombinant tissue plasminogen activator efficacy and risk in underserved groups are needed. In addition, information dissemination regarding stroke warning signs, the importance of calling for emergency assistance, and the availability of effective therapy for hard-to-reach minorities is a critical study issue.

Industry and Other Funding Agencies

Research is needed to determine whether industry-sponsored activities improve access to care and reduce stroke health disparities. It is important to determine whether the resources that industry uses for community activities are properly targeted. NINDS should collaborate with other federal funding agencies to facilitate mechanisms to perform these studies.

Multidisciplinary Collaborations

Studies investigating access to care and quality of care in minority populations should employ multidisciplinary investigative teams in order to maximize scientific benefit in the design, implementation, and analysis of data.

Resources Needed

  • Funding is much needed to encourage research on issues related to access to quality care and minority populations in the United States . RFA/PA mechanisms to stimulate observational, epidemiologic, intervention, and outcomes research in this area are necessary. It is critical that funding agencies pool efforts to establish appropriate funding mechanisms.
  • Funding research that monitors the quality of care and that helps facilitate improvement in access to quality care for stroke in underserved populations would be optimal. Cost-effectiveness studies and cost-effectiveness analyses in stroke trials should be incorporated into such research.
  • A special emphasis review panel should be considered to review grant applications in stroke health disparities.
  • Additional resources are needed for community outreach efforts that will increase our understanding of the barriers to and facilitators for care within minority populations. Focus groups that include health care workers, members of the public, and local health advocates would be helpful in designing funding targets for improving access to health care.
  • Efforts to increase representation of minorities in the field of stroke research may help further our understanding of how to help prevent and provide the best treatment for stroke in minority patients.
  • Providing information about acute stroke therapy and stroke prevention in native languages is important in working to reduce disease burden in emerging populations. Web-based information, particularly information provided in an interactive format, would help improve health care access by helping to demystify the health care system and foster trust.

Panel 4

Racial and Ethnic Disparities in Stroke: Genetics and Vascular Biology

Co-Chairs: Steven J. Kittner, M.D., M.P.H., Gary H. Gibbons, M.D., and Joseph P. Broderick, M.D.

David Hess, M.D., M.S.
Richard Levy, Ph.D.
James F. Meschia, M.D.
David J. Pinsky, M.D.
Charles N. Rotimi, Ph.D.
Bradford B. Worrall, M.D., M.Sc.

Statement of the Problem

The impact of race/ ethnicity on health status results from a complex interaction of cultural, social, economic, and biological factors associated with racial categories. There are no distinct biological "races," but at the societal level, the concept of race has strong cultural, political and economic meaning. The operational construct of race becomes relevant in our attempts to understand disease distribution within and between populations. Race, although lacking consistent biological definition, has real biological consequences that are manifested at the physiologic level. For these reasons, racial categories are important descriptors in biomedical research. Ethnicity may be a preferable term that incorporates nongenetic factors including economic, social, religious, linguistic, dietary and other variables to identify individuals as belonging to a group, while allowing for potential differences in gene frequencies and for changes over time in both gene frequencies and nongenetic factors.

As one examines human genetic variation at a population level, there are geographical differences in the frequency of some genetic markers. Individual genetic make up reflects ancestral geographic origin, migration patterns, and population admixture or isolation. Genetic variability among member of a particular race/ ethnic group likely has a greater health impact than differences across these groups. In relationship to stroke, it is likely that differences in recognized and unrecognized environmental factors play a dominant role in observed disparities in risk across different race/ethnic groups. Thus, the medical and research communities must avoid an overly simplified concept of race as biological and the naPve assumption that differences across ethnic groups are genetically based.

Nonetheless, there are known biological determinants of stroke risk that differ across race/ ethnic groups. Higher frequencies of genetic variants of the human beta A globin gene in people of Mediterranean or African origin influence cerebrovascular risk in these populations. The causative genetic mutations are different in different ethnic populations for two hemorrhagic stroke syndromes (familial cerebral cavernous malformations and hereditary amyloid angiopathy). These examples underscore the potential impact of genetic factors on stroke risk in different ethnic groups. As one moves from disease causing genetic variants (monogenetic traits) to risk modifying or susceptibility genes (polygenic or complex traits), the difficulty of identifying the relevant genes increases. Yet, methods are currently available for studying stroke as a complex genetic trait and the potential public health importance of identifying susceptibility genes makes this an important scientific priority.

The risk of early-onset stroke death is increased among African-Americans, American Indians/Alaskan natives, Asian/Pacific Islanders, and some Hispanic populations. Among African-Americans, a major part of this increased risk is due to a higher prevalence of major established vascular risk factors, such as hypertension and diabetes mellitus, at earlier ages and a higher prevalence of multiple vascular risk factors. There may also be a genetic contribution to the disparity in stroke risk since there are differences between racial/ethnic groups in the prevalence of many candidate genes relevant to stroke. Even candidate genes with similar prevalence between racial/ethnic groups may contribute to the disparity in stroke due to gene-gene and gene-environment interactions.

We need improved strategies for stroke prevention, treatment, and recovery, and these strategies need to be established as effective for all segments of our society. In order to accomplish this goal, we must understand the underlying vascular biology, including the contribution of genetic factors, environmental factors, and their interplay. Diversity of genetic background and environmental circumstances provide opportunities for understanding the pathogenesis of stroke that are not present in homogeneous populations. Furthermore, candidate gene association studies that fail to properly account for race/ethnicity and subgroups with different genetic ancestry can have biased results, that obscure potentially important associations or mislead investigators with results due only to the genetic structure of the study population.

Challenges and Questions

The delineation of the role of genetic and environmental factors in observed racial and ethnic disparities in cerebrovascular disease, and in stroke risk generally, is a major challenge for stroke researchers. The genetic and environmental information conveyed by racial and ethnic groupings is correlated and confounded. Thus, it is critical to avoid misinterpretation of racial and ethnic differences as synonymous with genetic differences. Two challenges to understanding the complex web of genetic and environmental determinants of stroke are difficulties inherent in characterizing race and ethnicity and environmental factors in research studies.

Self-reported race is a proxy for social, cultural, and environmental factors that can predispose an individual to stroke, factors that would be otherwise difficult to measure. Current methods for classifying race and ethnicity in genetic studies of stroke are inadequate for this purpose. For example, consider the particular issues of ethno-history and population admixture that may complicate the generalization of results of studies of Mexican-Americans in San Antonio, Puerto-Ricans in New York, and descendents of African slaves from Guatemala to individuals who may fall under the general rubric of "Hispanic." Similar difficulties could occur in the aggregation of study results regarding Japanese and Pacific Islanders. Thus, self-identification criteria as employed by the Census are not adequate for capturing the social construct of race and ethnicity.

In addition, self-reported race inadequately captures differences between subpopulations in the frequencies of genetic markers. When case and control samples are drawn from different subpopulations, allele frequencies will tend to differ among the subpopulations for most randomly chosen loci. This can lead to spurious or biased estimates of allele-disease associations, because a disease that has a higher incidence in a given subpopulation will be associated with any alleles that have a higher frequency in that subpopulation. Several approaches are available for addressing this issue, including using the well-established transmission/disequilibrium test (TDT), which requires family controls, detailed information on familial ancestry, or the relatively new technique of population substructure analyses. To date, these approaches are infrequently used in stroke research.

The limitations of environmental exposure assessment also hamper the study of the genetic component of racial and ethnic differences, and established vascular risk factors are imperfectly measured. For example, the measurement of a person’s blood pressure at a single point in time underestimates the effect of this factor on stroke risk, even in homogeneous populations. This limitation is accentuated in comparisons that are made between racial and ethnic groups, for which there can be important differences in the age of onset of hypertension as well as in the diurnal variation in blood pressure. In addition, important social and cultural correlates of race and ethnicity are often not measured or are poorly measured. These nontraditional epidemiological variables include health beliefs and participation in medical care, access to quality medical care, and differential physician behavior. Given these difficulties, it is important to avoid the simplistic interpretation that group differences that persist after adjusting for other measured social and biological stroke risk factors are innate differences.

The phenotypic characterization of stroke for genetic studies is a controversial issue that likely will require a diversity of approaches. Although there is consensus that ischemic stroke, intracerebral hemorrhage, and subarachnoid hemorrhage should be addressed separately, an important question remains regarding whether there are subphenotypes or intermediate phenotypes that have a particular relevance to racial and ethnic differences in stroke risk. For example, ischemic stroke due to intracranial versus extracranial atherosclerosis and small subcortical strokes may be informative subphenotypes for ischemic stroke, and cortical versus subcortical intracerebral hemorrhage may be useful subphenotypes for intracerebral hemorrhage. Similarly, magnetic resonance imaging-defined infarcts, cerebral white matter disease, and endothelial dysfunction may be useful intermediate phenotypes for investigating racial and ethnic differences in stroke risk.

Racial and ethnic differences in response to commonly used medications to prevent and treat stroke are not sufficiently understood. Many genes encoding drug targets or enzymes that metabolize drugs exhibit polymorphisms that alter sensitivity to particular medications, and racial and ethnic differences in the distribution of these polymorphisms have been observed. Alternatively, there may be differences in compliance or physiologic differences between racial and ethnic groups that reflect the complex interplay of many environmental factors.


Many operational challenges and barriers exist to conducting genetic research in different racial and ethnic groups. The lack of legal protection against genetic discrimination in the United States is a substantial impediment to genetic research and a particular barrier to the inclusion of racial and ethnic minorities. How can researchers best respect the unique cultural concerns of communities and of different racial and ethnic groups? The disclosure of genetic information derived from research studies may result in psychological or economic harm without an immediate or predictable corresponding medical benefit. The issue of disclosure of information about genetic testing during research is particularly sensitive with regard to minority populations in the United States , as studies such as the Tuskegee Syphilis Study have historically engendered distrust of biomedical research among some minority groups.

Research and Scientific Priorities

Priority 1:

Strengthen the regulatory environment that protects against genetic discrimination of individuals and named populations.

When an individual or a named population is shown to have a predisposition to a disease such as stroke, discrimination and stigmatization can occur. Public policy regarding protection from genetic discrimination has an effect on the recruitment and retention of research subjects, including racial and ethnic minorities. If we wish to fully realize the benefits of genetic research, policymakers have an obligation to protect individuals and groups from genetic discrimination.

Priority 2:

Study diverse racial and ethnic groups in different geographic, social, and cultural environments.

Cross-cultural studies or migration studies of racial and ethnic groups are useful scientific strategies for disentangling the role of race as a confounder reflecting both environmental and genetic differences. Similarly, the effect of race and ethnicity on disease outcome should be examined after stratifying on numerous potential confounders, including measures of socioeconomic status, education, and health care access and utilization. While indirect and subject to alternative explanations, these approaches can help target research into racial and ethnic disparities toward genetic or environmental differences.

Priority 3:

Provide characterization of variation in candidate genes and pathways within each racial and ethnic group, adequately powered studies within each racial and ethnic group, and appropriate consideration of the potential for population stratification bias.

Polymorphisms potentially relevant to stroke should be sought in diverse racial and ethnic groups. Family studies and candidate gene studies of stroke, including studies of recovery after stroke, should be conducted in samples with sufficient power to investigate racial and ethnic differences. Gene environment interactions should be tested in different racial and ethnic groups. These research approaches will reduce the risk of inappropriate generalization across racial and ethnic groups.

Candidate gene studies should use additional strategies beyond self-reported ancestry to control for population stratification bias. In studies of early onset stroke, when it is feasible to obtain parental and sibling DNA, the use of the transmission/ disequilibrium test represents an optimal design. In studies of older populations, population substructure analysis using a modeling approach to identify and control for genetic subpopulations should be more widely utilized.

Priority 4:

Study racial and ethnic differences in response to drugs used to treat acute stroke, as well as those used for primary and secondary prevention.

Particularly relevant drug classes include anti-hypertensives, anti-platelet agents, statins, thrombolytics, and anti-coagulants. NINDS may have an especially important role in sponsoring such studies on commonly used but older drugs for which the pharmaceutical industry may not perceive a commercial advantage. Studies are also required on the biological basis of racial and ethnic differences, namely polymorphisms associated with drug metabolism and drug action. These studies should be performed using populations that self-identify as belonging to a racial or ethnic group, and as well as for subgroups identified by genetic markers. Attention should also be given to environmental factors related to race and ethnicity that could lead to differential outcomes from drug use.

Priority 5:

Focus translational research on similarities and differences between racial and ethnic groups in important vascular biology mechanisms, including endothelial function, biomarkers of inflammation, coagulation/fibrinolysis, angiogenesis, and oxidative stress. Further characterization of these phenotypes between racial and ethnic groups must accompany genetic studies of these biological mechanisms.

  • Endothelial dysfunction: Define racial and ethnic differences in endothelial dysfunction between African Americans, Caucasians, and Hispanics, focusing on groups at high risk for stroke. Novel methods of examining cerebral vasoreactivity should be developed, and the following questions should be addressed: Are there racial and ethnic differences in cerebral endothelial NOS, and endothelin? Does the endothelial dysfunction in African Americans respond to statins and angiotensin-converting enzyme inhibitors and angiotensin receptor blockers as well as Caucasians do? If racial and ethnic differences exist, then studies of the genetic versus environmental bases for these differences would be a high priority.
  • Inflammation: Determine if there are racial and ethnic differences in inflammation, specifically biomarkers such as C-reactive protein levels and other inflammatory markers (adhesion molecules such as ICAM, E selectin). Does inflammation play a role equally among racial and ethnic groups?
  • Coagulation and fibrinolysis: Are there differences in the propensity of cerebral micro vessels to develop thromboses? Could there be gene polymorphisms (such as platelet PLA2 allele, PAI-1 Kosczak variation, Factor V Leiden) that underlie a propensity towards cerebrovascular thrombosis or hemorrhage, and could these polymorphisms or other susceptibility polymorphisms track with racial and ethnic groups? Could there be polygenic variations in coagulant or anticoagulant genes that track with race that increase susceptibility to or vulnerability to microvascular failure in stroke?
  • Angiogenesis: Are there differences between racial and ethnic groups in angiogenesis after stroke? Is there a difference in recovery after stroke that is related to different angiogenic responses? Are there differences in numbers of circulating endothelial progenitor cells?

Resources Needed

  • Coordination and collaboration of existing population-based research efforts with large ethnic cohorts (e.g., the Jackson Heart Study; the San Antonio Heart Study; the Atherosclerosis Risk in Communities Study; the Genetic Determinants of High Blood Pressure [HyperGEN] Study; the Reasons for Geographic and Racial Differences in Stroke Study; the Marshfield Epidemiological Study Area; the Northern Manhattan Stroke Study; the Greater Cincinnati/Northern Kentucky Stroke Study; the Stroke Prevention in Young Women Study). This would allow for the rapid testing of stroke-susceptibility loci and candidate stroke genes identified in Iceland and Europe for their implications in American racial and ethnic populations and would provide adequate power to address gene-environment interactions by race and ethnicity.
  • A new consortium to facilitate adequately powered studies of the genetic and environmental contributions to racial and ethnic differences in stroke risk. A major task of this consortium would be to develop common minimum standards for phenotyping and standardizing the key environmental risk factors and demographic variables that need to be collected with the genetic samples, including common approaches for defining racial and ethnic ancestry and for identifying subpopulations. Studies of the prognostic value and genetic determinants of biomarkers could be performed on prospective cohorts. Stroke registries with high racial and ethnic minority population representation could be utilized to search for genetic variation in candidate genes and apply expression profiling and proteomic analysis.
  • Clinical trials with significant representation of racial and ethnic minority groups and that include DNA banking at the outset. This will permit pharmacogenetic studies that will benefit all segments of society, including racial and ethnic minority groups.
  • Targeted cerebrovascular disease funding to study racial and ethnic differences in vascular biology, genetics, and pharmacogenetics, including gene-environment interactions. The creation of Specialized Centers of Clinical Research Excellence (SCCORS) could facilitate the "translating" of preclinical insights of vascular mechanisms of ischemic cerebral injury into understanding racial and ethnic disparities in stroke incidence and outcomes. These might be modeled after existing SCCORS in atherosclerosis, heart failure, and ischemic heart disease and allow the development of projects and participants from multiple institutions. Although preclinical mechanistic studies are encouraged in the new SCCOR framework, there is a critical emphasis on the second "C," that is, the clinical component. This may require the creation of sufficient infrastructure to allow for the recruitment of adequate numbers of under-represented populations. Integrating a group of investigators across traditional scientific boundaries, especially preclinical and clinical boundaries, could serve as a potent tool for understanding the vascular mechanisms that underlie racial and ethnic disparities in stroke incidence and outcome.

Panel 5A

Design of Clinical Trials to Address Stroke Disparities

Chair: George Howard, Dr.P.H.

John M. Flack, M.D., M.P.H.
Wayne H. Giles, Ph.D.
DeJuran Richardson, Ph.D.

Statement of the Problem

Randomized Clinical trials (RCTs) are the well-accepted "gold standard" of studies that assess the impact of new treatment regimens. They also can offer tangential information that can be used to understand the causes of and thus reduce disparities in stroke risk. Other types of studies can also be effective research tools in addressing stroke disparities. For example, epidemiological studies can provide important data regarding to the prevalence and causes of stroke disparity. In order to maximize the effectiveness of clinical trials and other types of studies in addressing disparities in stroke risk, various study design issues must considered.

Race and ethnicity is a proxy for the differential aggregation of potentially confounding variables such as obesity, proteinuria, dietary sodium intake, smoking, socioeconomic status, and cultural factors, all of which influence stroke risk but vary at the level of the individual. Furthermore, these variables are not uniformly distributed either within or between racial and ethnic groups. The range of baseline stroke risk as well as risk reductions that occur in response to treatment within a racial group are almost assuredly equal to, or larger than, differences in the risk observed between racial and ethnic groups. Thus, an important aspect of future clinical trials for stroke, as well as those focusing on other conditions, will be to identify the modifiers of treatment effects within racial groups.

The approaches to understanding the causes of stroke or to reducing disparities through randomized clinical trials that are considered in this report fall into three broad categories: 1) studies conducted in "targeted" populations of a single racial and ethnic group; 2) racially heterogeneous studies that include two or more racial/ethnic groups; and 3) efforts to facilitate the sharing of data between two or more clinical trials. The challenges, recommendations, and resources needed for each category are discussed below.

Targeted Studies

One design that has been used to address ethnic and racial disparities in stroke risk is that of the targeted trial, in which the study population consists of a specific ethnic or racial group. Targeted trials are clearly warranted in addressing salient conditions that exclusively or disproportionately affect a specific population subgroup. For example, a study assessing the relationship between the sickle cell trait and stroke risk would be correctly targeted to African Americans, as would a targeted study focusing on African Americans and t-PA treatment, as African Americans seem to be substantially less likely to receive t-PA treatment, even including those who reach the hospital within the treatment window. Care must be taken in conducting or funding such studies, however, because their results generally will not apply to other populations or groups. Nonetheless, such studies can be useful in providing insights into differential efficacy within racial and ethnic groups by providing a large enough sample size to ensure sufficient power for detecting differential efficacy.

Challenges and Questions

Challenges to conducting such trials include being able to justify the need for targeted studies at all (i.e., being able to show that the targeted population is disproportionately impacted by a salient condition). Recruiting and retaining sufficient numbers of patients can also be a daunting task. Quite often, a community-based infrastructure of some kind must be developed by study staff in order to facilitate a cooperative environment for the study and prevent community backlash. This is an even greater challenge when conducting nationwide studies that may involve 50 or more different communities. Choices of infrastructure include advisory groups composed of community representatives and partnerships between study staff and significant community organizations.

Among investigators there can also be a pervasive lack of knowledge of and sensitivity to the culture within various ethnic groups. This lack of knowledge and sensitivity leads to an inability to attract and retain study participants. Given the relatively small numbers of eligible minority patients that may be available for recruitment, all issues that impact the ability to recruit enrollees are important. Another challenge faced by researchers is determining ways to more strategically divide the limited pool of research funding. It is important to have priorities established and disseminated under the umbrella of NINDS, ensuring that only the most important and relevant proposals are submitted for funding consideration.

Research and Scientific Priorities

Many of these challenges can be addressed by increasing among investigators awareness of and sensitivity to cultural norms within the target population. Insensitivity to history and culture are often at the root of an inability to recruit and retain study subjects. NINDS can encourage (and perhaps mandate) investigators to demonstrate cultural knowledge and sensitivity as part of their grant proposals. Alternatively, NINDS could sponsor workshops devoted to disseminating proven strategies for recruiting and retaining minority study patients. Investigator attendance could be strongly urged or even mandated. Cultural sensitivity and awareness training should be incorporated into the educational training of all future health study investigators, and efforts to increase the relatively small number of minority investigators in this area should also be encouraged. In addition, the inclusion of minority-serving institutions would facilitate awareness of social and cultural issues. Finally, NINDS should form a task force or oversight group to make recommendations on research priorities, especially with respect to targeted studies. Pronouncements by such a group could help ensure that a unifying strategy is being incorporated into the funding decision process.

Racially Heterogeneous Studies

Racially heterogeneous RCTs can play an important role in advancing our understanding of the causes of racial disparities in stroke risk by providing the opportunity to assess for differential treatment efficacy by race (i.e., to assess race as an effect modifier). With appropriate adjustments to provide protection from spurious findings arising from multiple testing and subgroup analyses, the potential for identifying a treatment with a larger reduction in stroke risk among minority groups could play a critical role in suggesting treatment of particular importance for the reduction of excess risk in selected groups.

Challenges and Questions

Assessing race and ethnicity as an effect modifier through clinical trials requires substantial representation in these studies of these groups. The challenges and barriers to recruiting a substantial representation of minorities to a racially heterogeneous trial are similar to those required in recruiting for targeted studies:

  • a lack of appropriate societal and cultural training among investigators and staff;
  • the need for greater involvement of minority investigators and staff;
  • the need for careful selection of centers that provide access to minority populations and have demonstrated experience in recruiting these minority populations; and
  • the difficulty in developing commitment structure in multicenter studies.

In addition to challenges associated with the recruitment and retention of minorities to studies, the lack of statistical power to detect race as an effect modifier looms as the major challenge in the conduct of such studies. Specifically, the majority of RCTs are designed to provide "adequate" power to detect the difference between treatment groups in the overall study population. The potential of a differential efficacy is most appropriately addressed by the addition of interaction terms to models predicting study outcomes. The statistical power needed to detect such interaction terms is substantially less than the power needed to detect the main treatment effects. For example, if the overall study has 80 percent power for detecting the main effect the treatment, there could be less than 50 percent power for detecting clinically significant interactions.

The structure of the power formulas to detect racial and ethnic interactions is another difficulty that adds to the recruitment challenge. Specifically, in the case of an "events" trial (endpoints of stroke, death, or other discrete events) the power to detect treatment differences is tied to the number of events (not the sample size). The power is greatest to detect an interaction when the number of outcome events is equal for the races involved in the interaction test. Although there are differences in event rates between racial and ethnic groups, having an equal number of events in the racial and ethnic groups implies a challenge to "over-recruit" the minority populations as a proportion of the entire study group. Thus, although minorities are traditionally under-represented in RCTs, making the detection of race and ethnicity a goal of the study does not suggest the need to increase efforts to ensure adequate representation; rather it suggests the need to increase efforts to ensure over representation.

Increasing the power of a study also can be achieved by "simply" increasing the sample size. Obviously, the overall cost of the study is the greatest challenge to this theoretically straightforward approach. Under a fixed ceiling for the overall cost of the study, some expansion of the study sample size to address the major questions of the trial can be achieved by eliminating complex and expensive aspects of the study. However, the challenge of designing such a large, simple trial is determining a priori what can really be eliminated without damaging the opportunity for substantial gains in knowledge from secondary analyses.

Research and Scientific Priorities

Encouraging the representative recruitment (or over recruitment) of minorities to studies is a priority shared with targeted studies. However, in cases in which biological underpinnings suggest that the treatment response could be differential between racial and ethnic groups, the examination of race as an effect modifier needs to be formally addressed in the study design. Included in these efforts is the inclusion of formal secondary hypotheses of race and ethnicity as an effect modifier and formal discussion of statistical power to detect race and ethnicity effects (or to detect treatment differences in stratified analysis within racial and ethnic groups).

This noble goal, however, needs to be tempered by practical considerations. If studies were held to the "traditional" power levels of 80 percent or 90 percent to detect race as an effect modifier, the costs of the study would be prohibitive (with increases of perhaps millions of dollars). Thus, flexibility is required in evaluating the ability of a study to detect such a treatment effect. This flexibility can be achieved in several areas:

  • The acceptance of lower power for the detection of interaction terms reflecting the presence of race and ethnic effect modification. Specifically, although the 80 percent or 90 percent power is the standard for the "primary" study hypothesis, acceptance of a lower (but reasonable) level of power to detect racial and ethnic interactions is needed.
  • Power also can be increased by using surrogate endpoints. Although "traditional" or clinical endpoints may be required for assessment of the primary endpoint, the acceptance of surrogate endpoints to evaluate racial and ethnic effect modification could be encouraged. For example, although clinical stroke subsequent to treatment may be the primary endpoint for the assessment of an overall treatment effect (i.e., the primary hypothesis), the use of a composite endpoint of clinical or "silent" strokes (i.e., magnetic resonance imaging) for the assessment of the race and ethnic effect modification could substantially increase power.
  • The trade-off between study complexity and study size is a difficult one. However, acceptance is needed of decisions to expand sample size in order to address important issues of potential racial and ethic effect modification while possibly excluding "attractive" aspects of data collection in other arenas. That is, we must be more willing to accept large simple studies when they are needed to address the important possibility of racial and ethnic effect modification.

Resources Needed

The resources needed to facilitate racially heterogeneous studies with the goal of assessing race and ethnicity as an effect modifier fall into four arenas:

  1. encouraging efforts to recruit minorities;
  2. supporting and encouraging the elevation of the detection of race and ethnicity as an effect modifier to (at least) the level of a major secondary study aim;
  3. accepting alternative criteria for the evaluation of race and ethnicity as an effect modifier (i.e., use of surrogate endpoints and a lower power standard); and
  4. to the extent possible, expanding funding to increase the overall study size to meet the important aim of providing an assessment of racial and ethnic effect modification.

Data Sharing

The major shortcoming of racially heterogeneous studies is a lack of power to assess effect modification. The sharing of data across clinical trials has the advantage of providing an increased sample size that may in turn provide an understanding of racial and ethnic disparities in stroke at reduced costs when compared with heterogeneous studies. However, data sharing also poses a number of challenges, including the reluctance of investigators to share data; lack of standardization of data elements across studies; and the cost and time needed to create pooled data sets. Investigators may be reluctant to share data after they have invested resources and time into the design and conduct of clinical studies unless the Institute makes it explicit to investigators that data sharing will be required once the study results have been published. The lack of standardization of data across clinical studies poses yet another challenge. This includes the lack of standard measurements for risk factor and clinical outcome data. In addition, the financing and timing of data sharing poses yet another challenge. Key questions related to financing and timing of data sharing include whether investigators will be reimbursed for their time and effort in pooling data; whether resources be available for the analysis of pooled data; and the appropriate timing for the sharing of data.

In order to address these challenges the following research priorities are recommended:

  • The development of a well-designed National Institutes of Health repository for the sharing of data. The repository should include documentation and other materials to facilitate the pooling of data across studies.
  • The development of core set of common data elements for all studies. These data elements could include a common set of baseline characteristics likely to be associated with clinical outcomes (demographics and major risk factors), as well as a common set of outcomes assessed at regular intervals (incident stroke including subtyping and severity measures, Barthel or other measures at 6-month intervals, or surrogates for incident stroke). To increase comparability across studies, simple and common methodologies should be used to collect data across studies.
  • The development of a panel or group charged by NINDS to set data sharing standards and arbitrate data requests and to review analysis strategies to ensure appropriate use of shared data.

Panel 5B

Design of Clinical Studies: Bridges to the Community - Assuring Ethical Conduct of Studies and Data Integrity

Co-Chair: Virginia J. Howard, MSPH

Selma C. Kunitz, PhD.
George E.Locke, MD
Ronny A. Bell, Ph.D., MS
Kim Dobson Sydnor, Ph.D.

Statement of the Problem

Participation in clinical trials for a number of disease areas is declining, as high percentages of potential participants from targeted populations are not being identified for screening, while many of those who are targeted choose not to participate. Studies involving stroke also currently face recruitment difficulties, despite the large number of people who either have had or are at high risk of having a stroke; it often takes many years for primary and secondary stroke clinical trials to enroll adequate numbers of participants to meet sample size goals. Unfortunately, recruitment strategies have focused on problem-solving approaches rather than on making use of proactive, specific strategies that are designed to reach historically underserved populations.

Recruitment of minorities in NINDS-funded stroke clinical trials, such as the Asymptomatic Carotid Atherosclerosis Study, the North American Symptomatic Carotid Endarterectomy Trial, ACE, the Trial of Org 10172 in Acute Stroke Treatment, Vitamins for Intervention in Stroke Prevention, and the Warfarin versus Aspirin Recurrent Stroke Study, has ranged from 5 percent to 43 percent. Some stroke studies have successfully focused on and recruited African American participants, including the African-American Antiplatelet Stroke Prevention Study, the Greater Cincinnati/Northern Kentucky Stroke Study, the Northern Manhattan Stroke Study, and perhaps individual sites within other studies. However, aside from these studies, it does not appear that "bridges to the community" have been consistently built into the recruitment strategies used for clinical studies involving stroke.

Because clinical trials for stroke are unlikely to be powered in such a way that differences in racial and ethnic groups can be assessed, increased representation of different racial and ethnic groups in these trials is needed to provide some generalizability of results as well as preliminary data for the generation of hypotheses.

Challenges and Questions

With a focus on community involvement, NINDS faces many challenges in addressing the goal of increasing recruitment levels for underserved populations, including the development of approaches for:

  • creating an interface with communities of interest and effectively conveying the value of and necessity for their representation in clinical studies and clinical trials;
  • minimizing or eliminating barriers to participation that will likely persist even when a cogent argument has been accepted regarding the cost/benefit of community members’ presence in clinical trials; and
  • creating a research environment that recognizes the need for community relationships as an essential element of an effective recruitment strategy.


A variety of barriers impede the achievement of an increased presence of under-represented groups in stroke (and other) clinical studies. These barriers include but are not limited to the following:

Lack of Cultural Competence Among Recruitment Staff

Those with cultural competence recognize that characteristics that distinguish one group from another, including language, norms, attitudes, beliefs, and history, shape behavior and decision making and also influence communication and the establishment of trusting relationships. Someone who has cultural competence, for example, would understand that past encounters with the medical community, such as during the Tuskegee Syphilis Study, have caused many African Americans to be suspicious of the motives of those who conduct medical research and that in reaching out to all subsets of the Asian American population, as another example, it is important to realize that many Asian Americans are more oriented to the group rather than the individual. Without such understanding, as well as sensitivity regarding the need for multilingual study documents and recruitment material and methods, efforts at recruiting members of a number of minority groups are likely to be disappointing. A related issue is raised by delaying interaction with a particular community until a research study requires participants.

Socioeconomic Status of Participants

The stratification of health by socioeconomic status has been widely documented. The relationship of health to ethnicity and sex also has been demonstrated; however, these factors are also associated with socioeconomic status. Education, income, and occupation are the most frequently used measures of socioeconomic status, and education and income are particularly salient barriers to recruitment.

Much of the information conveyed about and related to clinical studies comes in the form of written language, which means that persons of low literacy are at a disadvantage when it comes to study inclusion. Even if literacy is not a barrier, persons with low income who may be disproportionately members of minority groups may not be able to afford to participate in a clinical trial. Financial constraints that limit participation include the high cost of taking time off from work and/or time away from family (childcare costs); high transportation expenses; and/or lack of medical insurance that might cover study ancillary costs and the provision of follow-up care. Also, those with low income are less likely to be aware of and have access to the medical facilities where clinical studies are being conducted.

Lack of Diversity in the Workforce

The under-representation of minority populations in clinical studies mirrors the under-representation of minorities in the fields of medicine and public health. Studies that examine the doctor-patient relationship show that patients are more satisfied with doctor visits when there is congruence (racial/ethnic) between themselves and their physicians. Given the voluntary characteristics of clinical study participation, it seems logical to conclude that congruence between the researcher(s) and the study participant would be valued as well. Having a health care provider of the same ethnicity or race is likely to decrease the level of distrust and enhance the probability of clear communication.

Lack of Homogeneity/Familiarity

Referring to under-represented and minority populations often involves aggregating a highly heterogeneous group of people, with heterogeneity reflecting factors such as race and ethnicity, class (socioeconomic status), sex, location (urban/rural), and other distinctions.

Strategies that fail to take into account the need for community-specific recruitment methods have likely contributed to the limited success in the recruitment and enrollment of groups of interest. Another problem that leads to a lack of recognition of group heterogeneity is a lack of familiarity with these groups. A cornerstone of social marketing is knowledge of the consumer, and strategies/products for recruitment and enrollment that are not consumer oriented are not likely to succeed. To be successful, bridges to the community and recruitment strategies must be tailored to the local constituency. Study factors such as treatment schedule, treatment side effects, convenience of clinic hours, study length, and the invasiveness of procedures also may be barriers to recruitment.

These barriers share the common thread of a lack of focus on building relationships between the scientific research community and the research participant communities. Although the scientific community carries out its work with an understanding of its own norms, values, and language, that same understanding generally has not been applied to the members of the communities with whom researchers want to work.

Research and Scientific Priorities

The single most important research priority is the creation of an environment that reflects a recognition of the inherent value of building partnerships between the scientific community and research participant communities - the two communities that together comprise the research enterprise. The diagram below represents how the research currently operates and a model of how the research could operate more effectively operate.



Incorporating the following priorities could facilitate the creation of such an environment:

Priority 1:

Create a "special populations" advisory group to work with investigative groups from stroke clinical studies.

Every community is different, and every community is heterogeneous. Thus, in speaking of "community representation," we must be careful to specify the characteristics of a particular community. The special populations advisory group would provide information and guidance from the perspective of community involvement in the implementation of clinical trials and studies. It would help to educate NINDS staff as well as grantees about how community relationships are an integral part of a successful stroke study. The advisory group should:

  • include among its ranks members of target communities (advocacy groups, racial and ethnic groups, minority institutions, and organizations);
  • use its overarching role to promote the building of community relationships with targeted communities outside the domain of a particular study; and
  • recommend the inclusion of community relationship building/community preparedness as a funded element of research design (e.g., establish a local community advisory board).

Note: Minority medical organizations include the National Medical Association, the Association of Black Cardiologists, the National Hispanic Medical Association, the International Society of Hypertension in Blacks, the Association of American Indian Physicians, and the National Rural Health Association.

Priority 2:

Establish research requirements for community engagement.

"You cannot serve someone if you do not know who they are."

  • Identify community needs and "care routes" by conducting a needs assessment.
  • Work with the community in the development of social marketing strategies and materials.
  • Identify/foster/create community health workers (community members trained as an arm of the project but who remain tied to the community).
  • Understand/incorporate an ecological framework in designing and implementing research (i.e., the influence of factors outside the study that may impact the study - apply a logic model).
  • Fund/partner with other disease-specific entities to foster prevention strategies in a comprehensive fashion so that the medical/public health arena is speaking with a more unified voice. This helps facilitate the creation of a clearer and more consistent message to the communities of interest and lessens the burden on local institutions to help disseminate the message.
  • Effectively disseminate research findings back to the community in an ongoing manner.

Priority 3:

Make modifications to the research process and to guidelines at NINDS.

  • Include grant reviewers/study section members who are familiar with community-based research methodologies.
  • Accept methodologies (e.g., qualitative research) that are relevant to a community-based research model.
  • Obtain latitude in the compensation and reimbursement of study participants that reflects an appreciation of the costs incurred (time off from work, day care expenses, transportation expenses, incentives).
  • Include a community outreach component that must be demonstrated during the planning period and throughout the entire project.

All of these processes should be informed by the work of the Office of Minority Health (OMH) and Institutes within the National Institutes of Health (NIH), such as the National Cancer Institute, the National Heart, Lung, and Blood Institute, the National Institute of Diabetes and Digestive and Kidney Diseases, and the National Institute of Allergy and Infectious Diseases, which are already using this type of research model.

Resources Needed

A Community Coordinating Officer/Office (CCO) is needed at the national level, perhaps housed in the Office of Minority Health or within NINDS.

A successful strategy will require oversight of all efforts aimed at this goal. NINDS will need to:

  • develop an inventory of existing community-partnership programs within and outside of NIH;
  • identify potential partners (e.g., the Office of Minority Health, the Centers for Disease Control and Prevention (CDC), the Centers for Medicare and Medicaid Services, the American Heart Association, the National Stroke Association, the Brain Attack Coalition);
  • establish links to and communication with other NIH Institutes;
  • conduct funding analysis and recommendations for allocations;
  • oversee the evaluation of these efforts; and
  • develop a resource center/clearinghouse for sharing/networking among stroke researchers.

The CCO could organize the special populations advisory group and thus develop links at both the national and local levels.

Programs Database

The CCO would be a resource center for stroke researchers to locate:

  • existing programs under way within NINDS/NIH regarding community participation;
  • existing programs outside of NIH with organizations (e.g., CDC, the Robert Wood Johnson Foundation, the W.K. Kellogg Foundation);
  • existing pipeline/diversity development programs, e.g., high school programs; and
  • minority organizations and health advocacy groups.

Funding Support

NINDS needs to provide funding support for:

  • the establishment of local or regional advisory groups that include local community members and local stroke researchers;
  • the development of a diverse professional corps with a K-12 pipeline approach, as potential professionals in this area may be lost to early school failure/dropout and lack of support and encouragement for success in the sciences;
  • community relationship building that begins during the planning stage of clinical trials and clinical studies; and
  • seed money for community mapping.

Panel 6A

Stroke Disparities Related to First and Recurrent Prevention

Co-Chairs: Elaine L. Miller, D.N.S., R.N., Ralph L. Sacco, M.D., M.S., and Herman A. Taylor, M.D.

Bernadette Boden-Albala, M.P.H., Dr.P.H.
Carol Barch, M.S.N., R.N.
Patricia C. Clark, Ph.D., R.N.
S. Claiborne Johnston, M.D., Ph.D.
Deborah Summers, M.S.N., C.S., R.N.

Statement of the Problem

Although stroke is a highly preventable disease, it requires a combination of long-term medical risk management and lifestyle modifications. Stroke prevention clinical trials have primarily focused on stroke prevention in high risk situations where individuals already have a history of stroke and TIA. The major limitation of this approach is that the intervention occurs after the disease has already manifested. Results from several clinical studies such as the NHANES III reaffirm the importance of primary prevention research to expand cost-effective evidence based practice and creative alternative interventions. Moreover, testing the applicability of cost-effective prevention strategies that have been successful in other areas (e.g., diabetes and coronary heart disease) is a priority.

In the process of identifying intervention strategies and developing innovative programs, it is useful to conceptualize prevention as potentially encompassing three key ingredients: knowledge, attitude and behavioral change. However, knowledge alone without attitudinal and behavioral actions cannot foster the lifelong changes required across age, race-ethnic, cultural and socioeconomic groups within varied settings.

The interaction of the individual/family, health care providers (HCPs), and the societal systems (e.g., medical systems such as hospitals and outpatient clinics, managed care organizations, American Stroke Association, National Stroke Association, churches, social community organizations) collectively play a pivotal role in the entire stroke prevention process. Along this continuum of stroke prevention care, all three stakeholders must work together, be responsive to the specific needs of one another, and support the achievement of intermediate as well as long-term prevention goals. This interaction between these three stakeholders becomes particularly challenging among racially, ethnically, and culturally diverse groups.

Challenges and Questions/Barriers

There is little question that the most significant reduction in stroke mortality and morbidity will arise from implementing more effective stroke prevention efforts, including strategies aimed at the identification and control of stroke risk factors. However, stroke prevention efforts are frequently ineffective for a number of reasons, including a general distrust of health care providers, the inadequate identification of underlying differences pertaining to race and ethnicity, and the underutilization of social resources that may affect risk factor identification and control. Patterns of racial discrimination and segregation have influenced the use of such social resources and social support structures, which may contribute to the disproportionate burden of stroke risk among African American and Hispanic populations. Disparities in social conditions have been associated with gradients in overall mortality rates, as well as cardiovascular mortality and morbidity, and there is an increasing body of literature suggesting that racial differences in disease-specific mortality and morbidity may be partially explained by underlying social disparities. However, less well documented is the impact of social disparities on stroke. Disparities in social conditions such as class, education, income, and occupation may reflect underlying inequalities in the utilization of social support structures, health care systems, and the disbursement of power, wealth, and knowledge.

Reducing disparities in stroke morbidity and mortality may require the identification of nontraditional, alternative prevention strategies that will contribute to better designed and more effective prevention programs that encompass knowledge, behavioral, and attitudinal interventions for diverse racial-ethnic groups with varying social resources.

Areas of Challenge

Three general areas should be targeted for the reduction of disparities in stroke morbidity and mortality:

  1. Barriers to risk factor identification

    Even with the identification of major stroke risk factors and the development of stroke prevention guidelines, various racial and ethnic groups may not be adequately screened or monitored for these risk factors. Barriers to risk factor identification in diverse populations include the following:

    1. Limited or no formal core curriculum regarding recognition of stroke risk factors/stroke prevention and acute stroke warning signs among physicians, nurses, occupational therapists, physical therapists, and other health care providers.
    2. Inadequate continuing education focusing on the importance of stroke prevention and risk factor education for the primary care and specialty physician, advanced practice nurses, and other health care providers.
    3. Inadequate adherence of physicians and other health care providers to evidence-based prevention guidelines.
    4. Insufficient knowledge of appropriate culturally sensitive strategies for effectively communicating the importance of risk factor identification for individuals, families, and communities.
    5. Limited access to health care settings for the identification of stroke risk factors

  2. Follow-up and modification of risk factors

    Success of risk modification programs depends upon the participation of both the individual and the family. A number of barriers have been identified that may create differences in the success of modification strategies:

    1. Inadequate individual, family, and community resources for facilitating adherence with risk-reduction activities (e.g., financial resources, dietary requirements/foods, medications, transportation).
    2. Inadequate linkages to existing resources.
    3. Social and geographic isolation of individuals, families, and communities.
    4. Lack of culturally appropriate educational materials to facilitate long-term risk factor management.
    5. Inadequate emphasis on the development of effective stroke risk reduction action plans for individuals and families that are sensitive to diversity issues.

  3. Prevention of recurrent stroke

    Although risk factor control following a stroke is just as critical as risk identification and control before the first stroke occurs, a number of new problems arise with the stroke survivor. Possible disability, change in residence, loss of employment, change in dependency status, the onset of depression, and changes in cognitive status are all factors that may further complicate prevention strategies. Barriers to reducing racial and ethnic disparities for recurrent stroke prevention may include the following:

    1. Inadequate access to medical and rehabilitation management for stroke survivors.
    2. Insufficient continuing education of health care providers in the long-term care of stroke survivors, particularly those who belong to diverse racial and ethnic groups.
  4. Lack of continuity of risk modification across the continuum by health care providers.
  5. Inadequate emphasis on the importance of family structure and the role of caregivers.
  6. Insufficient education and support for stroke survivors and their families related to long-term care needs.

Research and Scientific Priorities

Because effective stroke prevention is contingent upon the collaborative effort of the individual and family, health care providers, and social systems, the following research and scientific priorities are organized accordingly.

1. Evaluate barriers to adherence to stroke prevention strategies that are specific to minority groups (knowledge, attitudes, behaviors).

Some of the barriers to consider include:

  • Lack of knowledge regarding stroke, treatment effectiveness, and cultural diversity (individual and family, health care providers, social systems).
  • Fragmentation of care.
  • Inadequate resources, time, and incentives to implement the prevention guidelines or strategies.
  • Potential bias emanating from interactions and differing expectations of the individual and family, health care providers, and, social systems.
  • Health care provider interactions related to lifestyle modifications, outcomes, expectations, and communication (e.g., mismatches between health care providers, the individual and family, and social systems).

1. Develop and test for effectiveness of novel stroke prevention programs for multicultural groups that can be adapted for use within communities, considering the following:

  • Novel self-management strategies for people at risk for stroke and for stroke survivors (e.g., diabetes and heart failure self-management literature).
  • Increase in multidisciplinary and collaborative efforts across the stroke continuum that facilitate communication and monitoring of outcomes.
  • The effectiveness of a stroke care manager.
  • Predictors of health care providers’ adherence to guidelines (e.g., enhanced education, including core curricula, incentive programs).
  • Educational programs that evaluate efficacy and cost-effectiveness short and long term for individuals, families, and communities.
  • Partnerships within communities that facilitate lifestyle changes (e.g., a safe environment and resources such as food and transportation).
  • Methods to improve communication among individuals, families, and health care providers.
  • Specific programs aimed at socially and geographically isolated groups.

Resources Needed

Some of the resources to assist in achieving the research and scientific priorities are as follows:

  • Educational curricula for health care providers and the general public (for children and adults) that include more stroke-related information.
  • The development of partnerships with professional and community organizations (i.e., the American Stroke Association, the National Stroke Association, health clinics, senior centers, managed care organizations).
  • Media efforts that include providing knowledge that reflects sensitivity to cultural diversity.
  • Funding to perpetuate effective stroke prevention interventions for individuals, families, and communities.
  • Expansion of the availability of cardiac rehabilitation resources to the stroke survivor.
  • Expansion of the use of existing data sets (e.g., for diabetes, cardiac), capitalizing on their applicability to stroke.
  • Development of outcome assessments for monitoring and modifying stroke prevention activities.
  • Utilize outcome measurement data to support evidence-based prevention for culturally sensitive practice.
  • Develop Stroke Prevention Centers of Excellence that will be available as resources to accomplish some of the research priorities already identified.

Panel 6B

Disparities in Acute Stroke Care

Chair: Oscar Benavente, MD

Mitchell Elkind, MD, MS
David Tong, MD
Askiel Bruno, MD

Statement of the Problem

As with other areas of stroke treatment, there are likely disparities in acute stroke care: These disparities may be related to race/ethnicity, practice setting, as well as other uncertain factors.


  • Insufficient data

    Despite the perception that there is a racial disparity in acute stroke treatment, few if any data are available to substantiate this concern. However in patients with acute myocardial infarction, the use of thrombolytics has been reported to be significantly lower in minority groups, particularly in African Americans. Nevertheless, further data are necessary to verify the perceived differences in acute stroke treatment.

  • Diversity:

    One of the major challenges to the collection of accurate data about these issues is the wide variety of race-ethnic groups and health care delivery systems in which acute stroke care is delivered. Each setting is likely to have a different combination of factors challenging the delivery of acute care, such as education, insurance, adequacy of resources, suspicion of the health care system, and the use of emergency services for primary care.

  • Methodology:

    Another challenge to understanding stroke disparities is the lack of appropriate methodology. Although we have well developed methodology to study the clinical and pathophysiological aspects of stroke, similar tools to study disparities in acute stroke care are underdeveloped. These methodologies are likely to require multidisciplinary input from a variety of sources such as medical anthropologists, psychologists and community leaders.


  • Biological:

    Some disparities may relate to true biological differences in stroke presentation. For example, some studies suggest stroke subtypes vary between different ethnic groups. The prevalence of co-morbidities which influence acute stroke care such as severe hypertension may acutely influence the use of rtPA.

  • Socioeconomic:

    Socioeconomic factors likely influence the disparity in acute stroke treatment. For example, level of education and health insurance status have been shown to influence awareness of stroke risk factors and outcomes.

  • Cultural:

    Cultural factors may also impact the disparities in the acute management of stroke patients. For example, language barriers may impair the clinician’s ability to treat the patient effectively. These problems are particularly relevant in acute stroke care where decisions must be made within a very abbreviated period of time. In addition, cultural attitudes towards treatment may influence care through differences in trust between patients and physicians.

  • Public awareness:

    Poor public awareness of acute stroke symptoms and treatment is well known. This problem is likely to be more prevalent among minority groups, further worsening the disparity between different ethnic groups.

  • Physician biases:

    Physicians may have misperceptions about the utility of acute stroke therapies among different minority groups. Physicians may be influenced consciously or unconsciously towards specific treatments based upon race/ethnicity

  • Access to acute stroke care:

    The problems with access to acute care for stroke are a particular manifestation of a more general problem with access to health care, but also reflect specific disparities with regard to acute stroke care. There are several components to the problem with access, including those related to health-care systems, physicians, and patients.

With regard to health-care systems, it is clear that providing acute stroke care requires a effectively. Appropriate resources to provide the necessary infrastructure are not universally available, and this variability contributes to disparities.

With regard to physicians, it remains unclear that there is a general recognition across fields of medical expertise that specific acute therapies for stroke are of benefit. For example, specialists in emergency medicine do not universally endorse the use of intravenous thrombolytic therapy for ischemic stroke, and there is variability in enthusiasm for the use of thrombolytic therapy across hospital settings (community, tertiary care, etc) and locations. Physicians in some settings may not have been trained in acute stroke care. To the extent that patient subpopulations receive acute care in different settings and from different types of health care providers, there are disparities in care provided.

With regard to patients, there are many socioeconomic and cultural differences among different minority groups that contribute to disparities. Level of education, individual resources, family structure, and attitudes toward the local health care system may all contribute to a greater or lesser willingness to access the health care system.

Diminishing the present disparities in access to acute stroke care will require education of physicians, aggressive and widespread efforts to transform systems of acute care with regard to stroke, and massive education campaigns among different, often disenfranchised populations. Further research will also be needed to determine the optimal use of the necessary resources.

Challenges and Questions

There is only limited methodology available for the collection of data on patients’ knowledge about stroke, attitudes toward stroke, perception and understanding of the risk of stroke, and appreciation of the benefits of treatment. Access is only helpful insofar as it is followed by use. The limited studies that have been done suggest that some populations perceive the risk of stroke or its potential disability very differently: a fatalistic approach to disease may undercut efforts to provide acute care. Other populations (e.g., Jehovah’s Witnesses) have access to care but do not use certain available treatments because of their religious beliefs. Apart from identification of factors that may contribute to disparities in treatment, it is also crucial to determine optimal methodologies to study whether interventions can reduce these disparities. For example, early studies have shown that community education programs can lead to improved access in certain populations. Expanding these to other populations will be a significant challenge.

A more general question which was addressed in the report of the Stroke Progress Review Group was whether identification of primary stroke centers would provide a better means of providing access to care. Assuming that the movement toward primary stroke centers continues, a further question is whether the development of these centers can be manipulated to ensure that such centers are available to all communities equally. Otherwise, it is likely that disparities in acute stroke care will only increase.

The major barriers to our knowledge about disparities, the reasons for them, and the optimal ways to limit them are limited resources, diversity of populations of interest, absence of cross-institutional studies, and limited methodologies to study them.

Research and Scientific Priorities

Priority 1:

Which communities are affected?

First, it is crucial to understand on a broad basis where the disparities lie. We have increasing knowledge about national geographic trends in stroke incidence and mortality, but little data on access to stroke care on a national level. Those analyses that have been done locally have serious methodological flaws, often relying on hospitals or physicians to self-evaluate their programs, which leads to biased and inaccurate information. A goal should be to determine the true distribution of resources and availability of care.

Priority 2:

What is it about different communities that lead to different rates of acute stroke care?

The three major components of delivery of acute care—health care systems, physicians, and patients—and all of the subcomponents of each must be broken down and understood more completely. The interactions among them must also be understood before a complete picture can emerge. Regarding health care systems, what are the features that lead to such a disparity of acute stroke care utilization across locations (from 1% nationally to as high as 16% in Houston, TX, for example)? There is some evidence that patients enrolled in HMO’s may have less access to neurologists after a stroke than patients in other delivery care settings. Resources are certainly important, but community involvement, emergency services organization, and governmental support are certainly important factors as well.

Regarding physicians, certain specialists (or generalists) may be best suited to reduce the disparities in use of acute care for stroke. It is unlikely that neurologists will be able to provide all the acute stroke care needed nationally; there simply aren’t enough of them, and most were not trained in the era of thrombolysis. Emergency medicine specialists will need to be involved in the delivery of acute care, but it remains unclear what are the regional variations in physician attitudes toward acute stroke.

Regarding patients, several contributing factors—language, educational level, income, family and social structure, perception of risk, trust of the healthcare system—are all likely to play a role in access to and use of acute stroke care. Valid methodologies to study these issues in regard to acute stroke care are sorely needed. In particular, methods to study some of the less tangible and less directly quantifiable issues, such as perception of risk and trust, are needed. This may require involvement of experts from other areas not directly related to stroke, including social scientists, risk assessment experts, and psychologists and education specialists.

Priority 3:

What interventions can effectively and efficiently improve access to care and limit disparities?

Targeted interventions must be evaluated objectively and carefully to determine whether they can have an impact on access. These studies will need to simultaneously develop methodologies for the evaluation of outcomes after care, using advances derived from the field of outcomes research. These studies must be performed with the same kind of intellectual and scientific rigor as have traditional efficacy studies. For those interventions that demonstrate preliminary efficacy in early phase, pilot studies should be rapidly expanded to broader populations.

Interventions that could be tested would include, but not be limited to:

  1. Education of communities regarding their risk of stroke and how to recognize and respond to the symptoms of acute stroke;
  2. Educational programs targeted at family members of individuals already deemed to be at risk of stroke;
  3. Education of physicians regarding acute stroke care;
  4. Comparisons of systems approaches that focus on emergency medicine physicians vs. neurologists or other health care providers;
  5. Systematic implementation of first responder protocols for notification of hospitals prior to the arrival of potential stroke patients.

Resources Needed

Development of Sociomedical Methodologies Relevant to Stroke

There is a limited, though growing, literature available on the sociological aspects of stroke. There is little information on how communities view stroke, their perception of risk, their perception of the availability of care, and the utilities ascribed to survival after stroke. That information which is available does not generally consider the differences across cultures. Input from those with expertise in the social sciences will be needed for a fuller understanding of issues related to access.

Interdisciplinary Collaboration

Understanding the complexities of why disparities in access to care exist will require input from specialists in emergency medicine and stroke neurology, as well as experts in outcomes research, sociomedical science, psychology, and public policy. Fostering the creation of integrated teams of experts who often use very different methods and professional languages will be important to advancing our understanding of these issues. These teams may need to be formed at individual institutions where this kind of research occurs, as well as in teams created across institutions. It will be crucial to provide the same kind of funding for studies aimed at understanding the use of acute care as have traditionally been devoted to understanding traditional epidemiology and clinical drug efficacy.

Interinstitutional collaborations

Because issues of access to care revolve around differences in communities and populations, it will be necessary to foster the development of inter-institutional protocols as well. The applicability of protocols will need to be tested across a wide range of practice settings if they are truly make an impact on clinical care in the US .

Clinical Infrastructure

Any attempts to study these clinical questions will of necessity rely on practitioners in the relevant communities endorsing the role of acute care for stroke. If the resources to provide acute care are not available in underserved communities, then no amount of research or understanding will improve access. The necessary infrastructure will need to include mechanisms for ongoing physician education, as well as clinical resources.


Panel 6C

Stroke Rehabilitation and Outcomes

Chair: Linda S. Williams, M.D.
Pamela Duncan, Ph.D.
Daniel Hanley, M.D.
Brett Kissela, M.D.
James Stansbury, Ph.D.

Statement of the Problem

Although differentials in the burden of stroke incidence and mortality are unambiguous, differences in the burden of stroke-related disability are largely unknown, and the recovery trajectory is long and may be lengthening for a number of reasons (advances in secondary prevention, stroke occurring at earlier age, enhanced rehabilitation interventions). Therefore, foundational issues in understanding the relationship between disparities and stroke rehabilitation and recovery require us to 1) define the impact of disparities in the burden of disability over time; 2) examine whether disparities in rehabilitation access and quality occur; and 3) investigate possible explanations, including possible disparities in rehabilitation, that account for differing recovery trajectories.

Challenges and Questions

The aim of stroke rehabilitation is to restore the stroke survivor to an optimal level of function and to reestablish them into their community, efforts that are influenced by the following factors:

  1. Intrinsic factors
    • Individual characteristics (e.g., vascular biology, ancestry, genetics)
    • Disease characteristics (e.g., stroke severity, stroke subtype)
  2. Interventions
    • Behavioral interventions (e.g., participation in rehabilitation)
    • Biomedical interventions (e.g., pharmaceutical treatments)
    • System interventions (e.g., organization of stroke care)
  3. Access to rehabilitation and health care services
  4. Sociocultural and environmental factors

The major challenge, therefore, is to understand the ways that these factors interact to affect stroke rehabilitation and recovery outcomes. We believe strongly that understanding disparities in stroke rehabilitation and patient-centered outcomes requires attention to factors from all of these areas.


The major barriers to understanding the effect of racial disparities on the complete trajectory of recovery during the entire duration of stroke survivorship are as follows:

  1. The lack of primary data about racial disparities and rehabilitation processes and patient recovery.

    Very few studies have examined possible racial differences in individual outcomes or in access to and quality of rehabilitation after stroke.

  2. The challenges of longitudinal follow-up beyond the acute care setting.

    The few studies that have examined racial differences in rehabilitation have only measured in-hospital access to rehabilitation. We have little information about racial disparities in outpatient or longer-term rehabilitation settings, and we lack understanding about what recovery means to the patient or the family in the community setting.

  3. Lack of stroke outcome measures that are valid and reliable across different ethnic and linguistic communities.

    Ideally, patients from different ethnic groups but with the same level of function would have similar scores on a given stroke outcome measure. This attribute of stroke outcome tools is rarely evaluated.

  4. Limited range of rehabilitation process measures and patient measures in existing databases.

    Administrative databases generally contain only crude data about the presence or absence of a given rehabilitation service, making it difficult to evaluate quality of rehabilitation and stroke recovery across large and diverse samples.

  5. Absence of information about access/sociocultural/environmental variables in existing databases and in most stroke studies.

    Most studies do not measure variables related to access, sociocultural, or environmental factors that may impact rehabilitation and recovery.

  6. Lack of information about potential ethnic disparities in the efficacy or effectiveness of stroke interventions.

    Most interventional studies in stroke are not adequately powered to assess the impact of the intervention across different ethnic groups.

Research and Scientific Priorities

  1. Determine how race affects access to and quality and timing of rehabilitation across the continuum of care.

    The evidence for both acute and post-acute rehabilitation strongly supports the use of an organized, multidisciplinary, intensive approach to achieve optimal outcomes. In several randomized controlled trials, settings that included stroke unit care or organized inpatient multidisciplinary rehabilitation showed improved outcome compared to "standard" level care. Few studies have examined ethnic disparities in the quality, quantity, and timing of rehabilitation. Furthermore, whether disparities exist in the efficacy and effectiveness of stroke rehabilitation interventions and, if so, what factors underlie these differences are not well known. One existing study shows no difference in the utilization of inpatient rehabilitation but worse functional outcome in African American stroke survivors, and another demonstrated decreased use of inpatient rehabilitation by stroke survivors in the lowest income quintile. Another study showed minorities, less-well-educated individuals, and the oldest old were significantly less likely to receive physical and occupational therapy services in the year after stroke, even when disability and rehabilitating diagnosis were controlled for in multivariate analyses. More detailed studies of racial disparities in the structure, process, and quality of rehabilitation are needed. It will be important to design these studies up front using qualitative and quantitative methodologies to ensure that rehabilitation interventions are culturally appropriate and to consider the breadth of factors that may influence disparities in rehabilitation and recovery.

  2. Investigate whether there are racial differences in the recovery trajectory and in optimal recovery after stroke.

    Previous studies have shown that functional outcome after ischemic stroke is affected by many factors. Those most commonly reported include age, pre-stroke disability, severity of stroke, and/or level of consciousness on presentation, presence or absence of urinary incontinence after stroke, previous stroke, diabetes or elevated serum glucose, cardiac disease, and degree of social support or marital status. However, in the majority of these studies, race was not considered as a variable potentially affecting post-stroke outcome. Some studies have shown that the recovery trajectory in African Americans occurs later and does not catch up to functional recovery of Caucasian stroke survivors. Other recent studies have shown that African American race conferred a two-fold increase in risk of poor functional outcome and that Mexican American stroke survivors had greater psychosocial and functional impairment after stroke. These intriguing findings, and the omission of race from many previous studies of post-stroke outcome, suggest that further research is needed into the effect of race and ethnicity on functional outcome after stroke. If this finding is replicated and race is consistently associated with functional status after stroke, then further research is necessary to determine the underlying ways in which race and ethnicity could affect functional outcome.

    There is also a need to look beyond the activities of daily living and physical status toward potential disparities in social and cognitive function that can have implications for long-term recovery and survival. Patients' perceptions of their health-related quality of life (HRQL) influence their decisions and preferences for care and, conversely, physicians’ perceptions of patients’ HRQL affect recommendations for treatment. Years of research have linked socioeconomic status, race, and gender to health outcomes, regardless of the disease under investigation. Specifically, older age, female gender, non-Caucasian race, and lower socioeconomic status are associated with higher rates of disability and lower patient reports of physical health status. However, relatively little is known about racial variation in quality of life after stroke. While some studies have reported worse HRQL outcomes for female stroke survivors, most studies of HRQL after stroke have not reported results by racial and ethnic groups, so the relationship between race and HRQL after stroke is not known. It will be important to conduct both qualitative and quantitative research to answer the question of whether and in what specific ways race and ethnicity relate to HRQL outcomes after stroke.

  3. Determine whether there are disparities in other long-term sequelae of stroke (e.g. complications, rehospitalization, nursing home placement) and what factors influence these outcomes.

    Few studies address ethnic disparities in these stroke outcomes. One study showed that ethnicity was an independent predictor of rehospitalization after stroke, yet there remain variations between different ethnic groups that are not fully understood. Race and region of the country (specifically South or Midwest residence) have been associated with increased mortality and decreased general medicine and neurology outpatient visits after stroke discharge. Some, but not all, studies in nonstroke populations have shown that patients who are African American or from poor neighborhoods have worse processes of care and greater instability at hospital discharge than other patients. Other studies have identified disparities in the use of elective services expected to improve health and in the use of nonelective services associated with poor management of various chronic diseases. Whether these same patterns of process and utilization variation by race also occur after stroke remains to be determined. It will be important first to measure racial variation in post-stroke process and utilization and then to explore the reasons behind these variations.

  4. Identify the implications for the social and economic burdens of disability given changing demographic patterns in stroke incidence and in the U.S. population as a whole.

    Epidemiologic data suggest that African Americans suffer stroke at younger ages than do Caucasians; however, whether this results in different personal, social, and economic burdens of disability after stroke is not clear. Furthermore, demographic patterns in the United States suggest that minority groups will make up a larger proportion of the population in the coming decades. Data are limited and conflicting regarding whether race affects return to work after ischemic stroke. If disparities in stroke outcomes do exist, these changing demographic patterns may result in an increasing population burden of stroke over time. Longitudinal studies that examine the burden of stroke from a societal perspective and the potential impact of rehabilitation interventions on this burden are needed.

  5. Determine appropriate ways to measure recovery that limit differential functioning across racial and ethnic groups.

    Because stroke recovery is not measured in a standardized way, comparing recovery across studies or ethnic groups is difficult. In addition, most stroke outcome tools have not been evaluated for ethnic or linguistic relevance across different communities. Methodologic work to ensure that stroke outcome is measured without ethnic bias is needed.

Resources Needed

To adequately address the above scientific issues will require the following:

  1. Multidisciplinary research

    • Integration of multiple disciplines in the conduct of disparities research: medical anthropology, sociology, psychology, health services researchers, primary care providers, and stroke specialists. This multidisciplinary approach is essential to designing studies that can assess the complex relationships between ethnicity and stroke recovery.
    • Integration of social, environmental, and access variables in stroke rehabilitation and recovery research. Researchers should be encouraged to integrate both qualitative and quantitative approaches in incorporating social, environmental, and access to care data into planned stroke studies. NINDS could foster the collection of these data in its funded stroke research.
  2. Enhanced research infrastructure and focus on disparities in stroke rehabilitation and recovery

    • Support for studies that measure intrinsic, intervention, access, and sociocultural variables that may account for the relationship between disparities and the recovery trajectory. Attention to the measurement of variables from each of these key areas should be encouraged in all types of stroke research. In addition to measurement, we need to move forward with the development and testing of interventions that will decrease racial and ethnic disparities in recovery and access to rehabilitation services.
    • Constructing population-based stroke databases and linked existing databases that address disparities over the long term. Partnering with other government and nongovernmental agencies (e.g., the Centers for Disease Control and Prevention/Coverdale Registry, National Health and Nutrition Examination Survey) to incorporate stroke disparities research into ongoing research efforts may foster the development of this infrastructure. Follow-up must be continued beyond the traditional three to six months in order to address long-term recovery and rehabilitation utilization.
    • The development of community research networks to enhance the involvement of minority groups in stroke research both as researchers and research participants. Such networks are currently being developed in primary care; these could be used in stroke research and/or could serve as a model for similar neurology-based community research networks that would extend our ability to recruit minorities into the communities and expand from the current primarily university-based stroke studies. These efforts may also improve our ability to include minorities to serve as study advisors, colleagues, and researchers, which would further our ability to conduct disparities research in a meaningful way in the community.



NINDS Stroke Disparities Advisory Panel Meeting

November 7-8, 2002 Bethesda Marriott

5151 Pooks Hill Road, Bethesda, MD 20814

Agenda for November 7

8:00 a.m. - 8:30 a.m.
Continental breakfast
8:30 a.m. - 8:40 a.m.
Welcome and introductions
Barbara Radziszewska, PhD, MPH (Clinical Trials, NINDS)
8:40 a.m. - 8:45 a.m.
Charge to the panel
Ronnie D. Horner, PhD (Office of Minority Health and Research, NINDS)
8:45 a.m. - 9:00 a.m.
Review of the agenda
Philip B. Gorelick, MD, MPH (Rush -Presbyterian, St. Lukes Medical Center)
9:00 a.m. - 10:30 a.m.
Breakout Session I
(Panels 3, 4, 5A, 5B, 6A, 6B, 6C)
Anyone can attend any breakout session if schedule permits

Access to health care/quality of care (Panel #3)
Lewis B. Morgenstern, MD (University of Michigan) Yvonne Harris, MPA, CCRA (Rush-Presbyterian- St. Lukes Medical Center)


Edward Cooper, MD (University of Pennsylvania)
Carol M. Allen, PhD, RN (Oakwood College, Huntsville, AL)
Cheryl E. Easley, PhD, RN (Saginaw Valley State University)
Moises Gaviria, MD (University of Illinois)
Hal V. Barron, MD (UCSF/Genentech)
Karen L. Parko, MD (Indian Health Services, Shiprock, AZ)
Beth E. Quill, MPH (University of Texas at Houston)
Gloria J. Bonner, PhD, RN (University of Illinois)

Genetics and vascular biology (Panel #4)

Steven J. Kittner, MD, MPH (University of Maryland)
Gary H. Gibbons, MD (Morehouse School of Medicine)
Joseph P. Broderick, MD (University of Cincinnati)

James F. Meschia, MD (Mayo Clinic at Jacksonville, FL)
David Hess, MD, MS (Medical College of Georgia )
Charles N. Rotimi, PhD (Howard University)
Bradford B. Worrall, MD, MSc, (University of Virginia)
David J. Pinsky, MD (Columbia University)
Richard Levy, Ph.D. (National Pharmaceutical Council)

Design of clinical trials addressing stroke disparities (Panel #5A)


George Howard, DrPH (University of Alabama at Birmingham)

DeJuran Richardson, PhD (Rush-Presbyterian-St. Lukes Medical Center)
Wayne H. Giles, PhD (CDC)
John M. Flack, MD, MPH (Wayne State University)

Bridges to community - assuring ethical conduct of studies and data integrity (Panel #5B)

Virginia J. Howard, MSPH (University of Alabama at Birmingham)

Selma C. Kunitz, PhD (KAI & Associates)
George E. Locke, MD (King/Drew Medical Center)
Ronny A. Bell, PhD, MS (Wake Forest University School of Medicine)
Kim Dobson Sydnor, PhD (Morgan State University)

Stroke primary and secondary prevention (Panel #6A)

Elaine L. Miller, DNS, MN (University of Cincinnati)
Ralph L. Sacco, MD, MS (Columbia University)
Herman A. Taylor, MD (University of Mississippi)

S. Claiborne Johnston, MD, PhD (UCSF)
Bernadette Boden-Albala, MPH, DrPH (Columbia University)
Deborah Summers, RN, CS, MSN (Mid America Brain and Stroke Institute)
Carol Barch, RN, MSN (University of Pittsburgh)
Patricia C. Clark, PhD, RN (Emory University)

Acute stroke treatment (Panel #6B)

Oscar Benavente, MD (University of Texas at San Antonio)

Mitchell S.V. Elkind, MD, MS (Columbia University)
David C. Tong, MD (Stanford University)
Askiel Bruno, MD (Indiana University)

Rehabilitation (Panel #6C)

Linda S. Williams, MD (Indiana University)

Pamela W. Duncan, PhD, PT (University of Florida)
Daniel F. Hanley, MD (John Hopkins University)
Brett M. Kissela, MD (University of Cincinnati)

10:30 a.m. - 12:00 p.m.
Breakout Session II (Panels 1 and 2)

Temporal trends in disparities in stroke incidence, mortality, and quality of life (Panel #1)

Joseph P. Broderick, MD (University of Cincinnati)
George Howard, DrPH (University of Alabama at Birmingham)

Michelle L. Casper, PhD (CDC)
Lewis B. Morgenstern, MD (University of Michigan)
George W. Petty, MD (Mayo Clinic, Rochester)
Philip A. Wolf, MD (Boston University)
Peter MacLeish, PhD (Morehouse School of Medicine)

Disparities in stroke risk factors and mechanisms (Panel #2)

Ralph L. Sacco, MD, MS (Columbia University)
K.S. Lawrence Wong, MD (The Chinese University of Hong Kong)
Herman A. Taylor, MD (University of Mississippi)

Robert Adams, MD (Medical College of Georgia )
Karen L. Furie, MD (Harvard University)
Marco DiTullio, MD (Columbia University)
Edgar J. Kenton, MD (Thomas Jefferson University)
Stan Tuhrim, MD (Mount Sinai School of Medicine)
James M. Galloway, MD (University of Arizona)

12:00 p.m. - 1:30 p.m.
1:30 p.m. - 4:45 p.m.
General Session: Discussion

Panels will take turns and informally present results of their conference calls and breakout sessions, for about 10 minutes. Each informal presentation will be followed by general, 10 minute discussion.

1:30 p.m. - 1:50 p.m.
Temporal trends in disparities in stroke incidence, mortality, and quality of life (Panel #1)
1:50 p.m. - 2:10 p.m.
Disparities in stroke risk factors and mechanisms (Panel #2)
2:10 p.m. - 2:30 p.m.
Access to health care/quality of care (Panel #3)
2:30 p.m. - 2:40 p.m.
2:40 p.m. - 3:00 p.m.
Genetics and vascular biology (Panel #4)
3:00 p.m. - 3:20 p.m.
Design of clinical trials addressing stroke disparities (Panel #5A)
3:20 p.m. - 3:40 p.m.
Bridges to community - assuring ethical conduct of studies and data integrity (Panel #5B)
3:40 p.m. - 3:50 p.m.
3:50 p.m. - 4:10 p.m.
Stroke primary and secondary prevention (Panel #6A)
4:10 p.m. - 4:30 p.m.
Acute stroke treatment (Panel #6B)
4:30 p.m. - 4:50 p.m.
Rehabilitation (Panel #6C)

Agenda for November 8

8:00 a.m. - 8:30 a.m.
Continental breakfast
8:30 a.m. - 12:00 p.m.
Panels go to breakout rooms to finalize their chapters

List of Participants

Philip B. Gorelick, MD, MPH
Professor and Head
University of Illinois College of Medicine

Robert J. Adams, MD, MS
Regents Professor
Presidential Distinguished Chair
Department of Neurology
Medical College of Georgia

Carol M. Allen, PhD, RN (not attending)
Chair and Professor
Department of Nursing
Oakwood College

Carol Ann Barch, MN
Stroke Institute Program Coordinator
Department of Neurology
University of Pittsburgh Medical Center

Hal V. Barron, MD
Vice President
Medical Affairs
Genentech, Inc.

Ronny A. Bell, PhD, MS
Associate Professor
Department of Public Health Sciences
Wake Forest University School of Medicine

Oscar Benavente, MD
Associate Professor
Department of Neurology
The University of Texas
Health Science Center at San Antonio

Bernadette Boden-Albala, MPH, DrPH
Research Director
Department of Neurology
Columbia University/New York Presbyterian Hospital

Gloria J. Bonner, PhD
Assistant Professor
Medical Surgical Nursing
College of Nursing
University of Illinois at Chicago

Joseph P. Broderick, MD
Professor and Chairman,
Department of Neurology
U. of Cincinnati, College of Medicine Med. Center

Askiel Bruno, MD
Associate Professor
Department of Neurology
Indiana University School of Medicine

Michelle L. Casper, PhD (not attending)
Cardiovascular Health Branch
Centers for Disease Control and Prevention

Patricia C. Clark, PhD, RN
Assistant Professor
Nursing/Adult and Elder Health
Emory School of Medicine
Emory University

Edward Cooper, MD
Professor Emeritus, University of Pennsylvania

Marco R. DiTullio, MD
Associate Professor of Clinical Medicine
Department of Cardiology
Columbia University/New York
Presbyterian Hospital

Pamela W. Duncan, PhD
Program Director and Professor
Brooks Center for Rehabilitation Studies
University of Florida
College of Health Professionals
Health Science Center

Cheryl E. Easley, PhD, RN
Crystal M. Lange College of
Nursing & Health Sciences
Saginaw Valley State University

Mitchell S. V. Elkind, MD, MS
Assistant Professor of Neurology
Cerebrovascular Diseases, Neurology
Columbia University/New York Presbyterian Hospital

John M. Flack, MD, MPH
Professor and Associate Chairman for Academic
Affairs and Chief Quality Officer
Department of Internal Medicine
Wayne State University
University Health Center

Karen L. Furie, MD, MPH
Assistant Professor of Neurology
Harvard Medical School
Massachusetts General Hospital

James M. Galloway, MD
Native American Cardiology Program
University of Arizona
Indian Health Service

Moises Gaviria, MD
Director, Neuropsychiatry Professor
Neuropsychiatric Institute
Department of Psychiatry
University of Illinois

Gary H. Gibbons, MD
Professor and Director
Cardiovascular Research Institute
Department of Medicine
Morehouse School of Medicine

Wayne H. Giles, PhD
Associate Director for Science
Division of Adult and Community Health
Centers for Disease Control and Prevention

Daniel F. Hanley, M.D. (not attending)
Division Director and Jeffrey and
Harriet Professor
Department of Neurology
Division of Brain Injury Outcome
The Johns Hopkins University School of Medicine

Yvonne Harris, MPA, CCRA
Project Manager
African-American Antiplatelet Stroke
Prevention Study
Department of Neurology
Rush-Presbyterian-St. Luke’s Medical Center
The Center for Stroke Research

David C. Hess, MD
Chairman and Professor
Department of Neurology
Medical College of Georgia

George Howard, DrPH
Professor and Chair
Department of Biostatistics
The University of Alabama at Birmingham
School of Public Health

Virginia J. Howard, MSPH
Assistant Professor
Department of Epidemiology &
International Health
University of Alabama at Birmingham

S. Claiborne Johnston, MD, PhD
Assistant Professor
Department of Neurology
University of California, San Francisco

Edgar J. Kenton, III, MD
Chief, Cerebrovascular Diseases
Mainline/Jefferson Health System

Brett M. Kissela, MD (not attending)
Assistant Professor
University of Cincinnati

Steven J. Kittner, MD, MPH
Department of Neurology
University of Maryland

Selma C. Kunitz, PhD

Richard Levy, PhD
Vice President
Scientific Affairs
National Pharmaceutical Council

George E. Locke, MD
Professor and Chairman
Department of Neuroscience
King/Drew Medical Center

Ann M. Malarcher, Ph.D.
Senior Scientist, Epidemiologist
Cardiovascular Health Branch
Centers for Disease Control and Prevention

Peter R. MacLeish, PhD
Department of Anatomy and Neurobiology
Morehouse School of Medicine

James F. Meschia, MD
Department of Neurology
Mayo Clinic

Elaine L. Miller, DNS, MN, BSN
Professor of Nursing
College of Nursing
University of Cincinnati

Lewis B. Morgenstern, MD
Director of the Stroke Program
University of Michigan

Karen L. Parko, MD
Director, Neurological Services
United State Public Health Service
Northern Navajo Medical Center

George W. Petty, MD
Department of Neurology
Mayo University

David J. Pinsky, MD
Associate Professor of Medicine
Columbia University College of
Physicians and Surgeons
Division of Medicine
Department of Cardiology

Beth E. Quill, MPH
Associate Professor, MAPS
Director, Center for Excellence
in Public Health Practice
University of Texas-Houston Health
Science Center
School of Public Health
Community Health Practice

DeJuran Richardson, PhD
Professor and Director of Data Management
Center for Stroke Research
Rush-Presbyterian-St. Luke’s Medical
Center /Lake Forest College

Charles N. Rotimi, PhD
Associate Professor and Director
Microbiology/National Human Genome Center
Howard University, College of Medicine

Ralph L. Sacco, MD, MS
Associate Chairman of Neurology
Department of Neurology
Columbia University/New York
Presbyterian Hospital

James P. Stansbury, PhD
Postdoctoral Fellow
Rehabilitation Outcomes Research Center
Veterans Health Administration

Deborah R. Summers, RN, MSN, CS
Advanced Practice Nurse and Stroke
Program Coordinator
Neuroscience Department
Mid America Brain and Stroke Institute

Kim Dobson Sydnor, PhD
W.K. Kellogg Minority Health Disparities Postdoctoral Fellow
Public Health Program
Morgan State University

Herman A. Taylor, Jr., MD
Director and Principal Investigator
Jackson Heart Study
University of Mississippi Medical Center

David C. Tong, MD
Associate Director
Department of Neurology
Stanford Medical Center

Stanley Tuhrim, MD
Professor of Neurology
Department of Neurology
Mount Sinai School of Medicine

Linda S. Williams, MD
Staff Physician, Department of Neurology
Roudebush VAMC HSR&D
Assistant Professor, Department of Neurology
Indiana University School of Medicine

Philip A. Wolf, MD
Professor of Neurology
Department of Neurology
Boston University School of Medicine

Lawrence K.S. Wong, MD
Associate Professor of Neurology
Department of Medicine & Therapeutics
The Chinese University of Hong Kong
Prince of Wales Hospital

Bradford Burke Worrall, MD, MSc
Assistant Professor of Neurology
Department of Neurology
University of Virginia Health System

Last updated April 8, 2011