Racial disparities in stroke awareness: African Americans and Caucasians

RACIAL DISPARITIES IN STROKE AWARENESS: AFRICAN AMERICANS AND CAUCASIANS MOHAMAD G. ALKADRY Old Dominion University RUCHI BHANDARI CHRISTINA S. WILSON West Virginia University BRANDI BLESSETT University of Central Florida ABSTRACT Considerable evidence supports the existence of racial disparities in incidence, mortality, and morbidity related to stroke. Awareness of risk factors could substantially lower the probability of stroke incidence. Awareness of stroke warning signs and treatment options could significantly alter the outcome of a stroke if patients immediately seek emergency help. This article examines the disparities in awareness of stroke risk factors, stroke signs, and action to be taken when stroke occurs. Survey results from 422 Caucasian Americans and 368 African Americans in West Virginia were analyzed. Significant disparities in recognition of cholesterol, smoking, prior stroke, and race as stroke risk factors were observed. The study also found a significant and substantial difference in awareness of stroke signs. There was also a significant difference in the way African Americans and Caucasians would respond to a stroke. The study found no evidence of disparities in recognition of stroke risk factors, such as hypertension, diabetes, heart disease, obesity, alcoholism, and family history.

A stroke occurs when the blood supply to brain cells is suddenly interrupted, causing some brain cells to immediately die, while damaging others in the region of injury. Permanent injury and disability may be curtailed with immediate intervention that restores blood flow to the compromised brain cells (American Heart Association, 2004). Stroke is the leading cause of disability and the third

JHHSA SPRING 2011

463

leading cause of death (Centers for Disease Control and Prevention, 2000). Each year, 700,000 Americans are predicted to experience a stroke, of which over 150,000 are fatal (American Heart Association, 2003). Consequently, the estimated cost of stroke and disability has reached approximately $62.7 billion nationally (DHHS, 2007). Research consistently shows that stroke mortality and morbidity are more severe among African Americans than any other racial or ethnic group (Gillum, 1999; American Heart Association, 2004). For example, African Americans had the highest incidence of high blood pressure, which is associated with the largest death rates related to coronary heart disease (CHD) and stroke (DHHS, 2007). Disparities in impact and intensity of stroke are well documented, and so are racial disparities in stroke risk factors. Effective stroke prevention encompasses awareness of stroke risk-factors, stroke warning signs, and appropriate actions to be taken in case of a stroke (Alkadry, Wilson & Nicholas, 2005; Becker, Fruin, Gooding, Tirschwell, Love & Mankowski, 2001). However, the contribution of disparities in knowledge about stroke risk factors and warning signs to the disparity in stroke incidence remains largely uninvestigated. This study compares stroke awareness of racial minority residents to their white counterparts in West Virginia. First, we will briefly review the literature regarding racial disparities involving strokes. Next, we review and present the persistent differences in cardiovascular risk factors, stroke incidence and type, and survival following stroke that exist between African Americans and whites. Then, we test some stroke awareness hypotheses using data from 456 Caucasian and 400 African American residents of West Virginia.

464

JHHSA SPRING 2011 DISPARITIES IN STROKE INCIDENCE AND MORTALITY

Stroke is one of the diseases that African Americans experience at higher rates than any other racial or ethnic group (Feldman & Fulwood, 1999; Howard & Howard, 2001). Despite substantial improvement in stroke incidence and survival since the 1970s, African Americans continue to have a higher rate of stroke incidence, 30% higher hospitalizations from stroke than Caucasians (Centers for Disease Control, 2004), and a higher rate of stroke mortality than any other racial group in the nation (Gaines, 1997; Gillum, 1999; Howard & Howard, 2001). The national death rate from stroke among African Americans is 166 per 100,000, which is much higher than 117 per 100,000 for Caucasians (Casper, Barnett, Williams, Halverson, Braham, & Greenlund, 2003). There are also significant racial disparities in the age distribution of stroke deaths. African Americans’ critical age for stroke is closer to 35 while that of Caucasians is 45 (Rimmer, Braunschweig, Silverman, Riley, Creviston, & Nicola 2000; Jacobs, Boden-Albala, Lin, & Sacco, 2002; Morgenstern, Spears, Goff, Grotta, & Nichaman, 1997). While 25% of stroke deaths occur among Caucasians younger than 65, the rate is almost doubled (49%) among African Americans in the same age group (Casper et al., 2003). Stroke among African Americans is considered more lethal (Gillum, 1999) because of a higher percentage of hemorrhagic strokes (Ayala, 2002) and cerebral infarction (Bian, Oddone, & Samsa, 2003) compared to Caucasians. Shen, Washington, and Aponte-Soto (2004) corroborated these findings in their study that demonstrated that cerebral artery disease was more prevalent among African-American and Hispanic populations than among

JHHSA SPRING 2011

465

Caucasians. Bian et al. (2003) found racial differences in stroke incidence and mortality among the elderly. DISPARITIES IN STROKE RISK FACTORS Stroke incidence can often be traced back to the presence of one or more of several non-modifiable and modifiable risk factors. While race, age, family history, and prior stroke are considered non-modifiable stroke risk factors, hypertension, smoking, obesity, cholesterol, diabetes, physical inactivity, and alcohol use are all modifiable risk factors that can be managed to effectively lower the likelihood of future stroke (National Stroke Association, 2003). Disparities in stroke incidence, mortality, and morbidity may be linked to disparities in the prevalence of risk factors between African Americans and Caucasians (Becker, Tuggle, & Prentice, 2001). Gaines and Burke (1995) report greater prevalence of risk factors of stroke among African Americans of all ages. African Americans also have a higher prevalence of modifiable stroke risk factors such as diabetes, hypertension, high cholesterol, and high blood pressure (Feldman & Fulwood, 1999; Becker, Tuggle, & Prentice, 2001, Ruland, Raman, Chaturvedi, Leurgans, & Gorelick, 2003). The American Heart Association (2003) reports that one in three African American adults – compared to the national average of one in four American adults - has hypertension. Subsequently, the effects modifiable stroke risk factors are more frequent and severe in the African American population. Therefore, when adjusted for age, stroke deaths are almost 40% higher for African Americans than whites (CDC, 2005). The racial disparities in risk factors are adequately captured by the Behavioral Risk Factor Surveillance System (BRFSS), conducted by the Centers for Disease Control and Prevention (CDC). According to the BRFSS,

466

JHHSA SPRING 2011

African Americans are at a higher risk than White Americans for all the modifiable stroke risk factors, with the exception of cholesterol, smoking, and alcohol consumption (BRFSS, 2003). Higher percentage of African Americans have hypertension (31.4% compared to 25.8% Whites), higher Body Mass Index (68.5% compared to 57.4% of Whites), diabetes (10.2% compared to 6.6% of Whites), and lack physical activity (30% compared to 21% Whites). DISPARITIES IN STROKE AND SOCIO-ECONOMIC STATUS In addition to a higher prevalence of risk factors among African Americans, there are other explanations for the disparities in stroke mortality and morbidity. Several studies document the relationship between disparities in stroke and socio-economic status (Ayanian, Udvarhelyi, Gastonis, Pashos, & Epstein, 1993; Burstin, Lipsitz, & Brennan 1992; Becker & Newsom, 2003; Cooper, 1993; Fiscella, Franks, Gold, & Clancy, 2000; Gornick, Eggers, Reilly, Mentnech, Fitterman, Kucken, & Vladeck, 1996; Nickens, 1995; Oddone, Horner, Diers, Lipscomb, McIntyre, Cauffman et al., 1998; Dries, Exner, Derek, Gersh, Cooper, Carson, et al. 1999; Whittle, Conigliaro, Good, & Lofgren, 1993). These studies associate disparities in healthcare with socio-economic status, to the extent of regarding it as a key determinant of disease outcome (Chung, 2003), particularly for African-American health (Lillie-Blanton & Laveist, 1996; Gornick, 2000; Burstin, et al., 1992; Becker et al. 2003). Wilson (1987) suggests that socio-economic status is a greater determinant of African-American problems than even race. Low socioeconomic status has been associated with lesser access to care and lesser utilization of healthcare, even among those with health

JHHSA SPRING 2011

467

insurance (Morgenstern et. al., 1997, Newacheck, Hughes, & Stoddard, 1996; Adler, Boyce, Chesney, Folkman, & Syme, 1993). Gillum and Mussolino (2003) found that African Americans with eight or more years of education had significantly lower risk of stroke compared to those with less than eight years of education. Bone, Hill, Stallings, Gelber, Barker, Baylor, et al. (2000) studied 2,196 adult African Americans and found that respondents with hypertension were more likely to have less than 12 years education, be unemployed, or have low income. Some studies suggest that racial disparities would appear smaller if people with the same education and socioeconomic status are compared to each other (Sowers, Ferdinand, Bakris, & Douglas, 2002; Jamerson, 1993). Some investigations associate the racial disparity in stroke incidence with differential treatment by healthcare providers (Smith, 1998; Watson, 2001). Mitchell, Ballard, Matchar, Whisnant, and Samsa (2000) found that African Americans were less likely to receive current standards for stroke treatment as compared to Caucasians even after accounting for differences in the ability to pay, health provider characteristics, patient demographics, and comorbidity. A report by Lillie-Blanton, Rushing, Ruiz, Mayberry, and Boone (2002) evaluated 81 studies published between 1985 and 2001 that included minorities in adequate numbers. The majority of the studies (68 out of 81) revealed that the minorities received sub-standard care. Racial differences have been confirmed in the use of diagnostic procedures, anticoagulant therapy (Christian, Lapane, & Toppa, 2003), thrombolytic therapy, carotid endarterectomy (Goldstein, Matchar, Hoff-Lindquist, Horner, & Samsa, 2003), and revascularization procedures (Ibrahim, Whittle, Bean-Mayberry, Kelley, Good, & Conigliaro, 2003). Alkadry and Tower (2010) assert that geographical location (e.g. rural West Virginia) also restricts the type of

468

JHHSA SPRING 2011

treatment available to residents, particularly access to quality healthcare, availability of qualified health care providers, facilities and insurance coverage. Uneven access to healthcare services and lower quality of care combined with cultural barriers (Reese & Ahern, 1999) and Tuskegee-like abuses of African Americans tend to decrease minority trust in the healthcare system (Eric, 1997; Shavers & Lynch, 1997). This lack of trust makes it even harder for healthcare professionals to access and recruit African Americans in order to provide stroke education and management. The paucity of medical research information from under-served minorities also amplifies stroke research and treatment challenges. DISPARITIES IN AWARENESS The most easily modifiable barrier to acute and preventive stroke therapy is poor knowledge of stroke risks, warning signs, and the necessity of seeking emergency care as soon as a stroke is suspected (Hachinski, 2002). Researchers assessing American adults, largely composed of Caucasians, consistently report poor stroke knowledge in the general population (Travis, Flemming, Brown, Meissner, McClelland, & Weigand, 2003; Reeves, Hogan, & Rafferty, 2002; Rowe, Frankel, & Sanders, 2001; Pratt, Ha, Levine, & Pratt, 2003; and Hux, Rogers, & Mongar, 2000). Stroke knowledge includes awareness of warning signs, stroke factors, and appropriate emergency action. The ‘act F.A.S.T.’ (Face, Arms, Speech, and Time) campaign as promoted by the National Stroke Association may be a great way to educate people on what to do if someone suffers from a stroke. The acronym prompts people to look for signs (such as a droopy face, numbness or weakness on one side of the body, and slurred speech) and act immediately to prevent permanent physical damage (AHA, 2009). Awareness and action can therefore be the

JHHSA SPRING 2011

469

key to lessening the detrimental physical and financial implications of stroke related disabilities. Studies (such as Rowe et al., 2001) have demonstrated that respondents agree that a person can reduce the risk of having a stroke, yet several of them cannot even recognize the risk factors. In an interview of 2,512 adults, Reeves et al. (2002) reported that 80% of respondents could report at least one stroke risk factor, but only 28% could correctly identify three. Pancioli, Broderick, Kothari, Brott, Tuchfarber, Miller et al. (1998) revealed an even lower percentage (32%) of respondents who were able to identify one of the three risk factors. Early recognition of stroke symptoms and immediate emergency care are the best ways to reduce the impact of a stroke once it is eminent. However, many people fail to identify a stroke and consequently do no seek timely medical intervention. According to a report by the Centers for Disease Control and Prevention (2004), only 17% of the public recognize enough of the major warning signs of stroke to call 911. Yoon, Heller, Levi, Wiggers, and Fitzgerald (2001) found that only 49.8% of respondents correctly identified more than one warning sign of stroke. Another study reported that 39% of 163 patients admitted to the emergency room with possible stroke were unable to identify a single sign of stroke (Kothari, Sauerbeck, Jauch, Broderick, Brott, Khoury et al., 1997). Rowe et al. (2001) found that of the 602 respondents, none could name all five stroke warning signs and only 39% could spontaneously name at least one warning sign. Rapid action is important because the only Food and Drug Administration-approved acute stroke therapy is effective only if given intravenously within 3-6 hours of stroke symptom onset (American Heart Association, 2004; Morgenstern et. al., 1997). An individual’s inability to identify a stroke causes delay in seeking medical attention, and ultimately leads to more severe impacts of stroke

470

JHHSA SPRING 2011

(Yoon & Byles, 2002). Parahoo, Thompson, Cooper, Stringer, Ennis, and McCollam (2003) found that almost half of those surveyed would not contact an ambulance service if they thought they were having a stroke. Menon, Pandey, and Morgenstern (1998) revealed that compared to 34% of White non-Hispanic patients, 28% of African American, 18% of Hispanic American, and 26% of women patients were seen by a neurologist within that three-hour window. What remains largely unclear is the contribution of racial, or other socioeconomic or regional disparities related to knowledge of stroke and its incidence. Ferris, Robertson, Fabunmin, and Mosca (2005), and Pratt et al. (2003) found that knowledge of stroke warning signs is poor among Americans, particularly among racial and ethnic minorities who have a greater non-modifiable risk burden. Pratt et al. (2003) used structured telephone interviews to assess stroke knowledge among 379 older African American adults who had received previous medical treatment in Detroit, Michigan. They concluded that while these adults were aware of the need for urgent attention in the event of stroke, accurate stroke knowledge such as knowledge of stroke warning signs (29% correct), risk factors (58% correct), and modifiable risks (29%) was extremely poor, substantially below levels reported in separate studies of Caucasian adults. Two additional studies assessed stroke awareness only among women, though limited minority sampling also hampered their ability to directly assess racial disparities in stroke knowledge. The first study of 71 Hispanic and 144 non-Hispanic white hospitalized women over 39 years of age demonstrated that Hispanics were significantly less likely to report stroke risk factors such as hypertension (Kattapong et al., 1998). Ferris et. al. (2005) conducted a national survey of about 1000 women including 12% African Americans and 12% Hispanics to assess trends in

JHHSA SPRING 2011

471

stroke awareness and knowledge. Database constraints did not allow for adjustment of potential confounders in evaluating the association between race and stroke awareness. Data from the National Health and Nutrition Examination Survey (NHANES) concluded that a third of the excess mortality in African Americans over Whites was due to risk factors such as hypertension, smoking, high cholesterol, and obesity, another third resulted from socioeconomic factors, but the remaining third could not be explained (Otten, Teustsch, Williamson, & Marks, 1990; Gaines & Burke, 1995). In the next section, we test whether disparities in awareness of stroke risks, signs, and treatment could explain that remaining portion of the unexplained disparity. METHODS Target Population Twelve West Virginia counties with sizeable African American populations were targeted for data collection. A regular mail survey was used to reach the Caucasian population. The original mail survey had a very poor response rate of African Americans, which is consistent with previous stroke surveys in West Virginia (Alkadry et al., 2005). The difficulty of recruitment and retention of African Americans into research studies can be attributed to historical mistrust of biomedical research, lack of cultural relevance and competence, and less access to care (Loftin, Barnett, Bunn, & Sullivan, 2005; ShaversHornaday & Lynch, 1997). These barriers required the researchers to actively seek alternative ways for the inclusion of African Americans in the study. As a result, local and state African American representatives were consulted for advice on overcoming low response rates. These community leaders suggested data collection at local, county, and state African American community events and

472

JHHSA SPRING 2011

festivals, as well as through community and church organizations. African American adults approached at these events were asked to voluntarily complete the questionnaire. The use of convenience sampling to reach target populations is a common practice in the field of stroke (Okwumabua, Martin, Clayton-Davis, & Pearson, 1997; Travis et al., 2003; Pratt et al., 2003; Hux et al., 2000). Study Population Using both methods, 400 African Americans and 456 Caucasians responded to the 27-question survey between August 2004 and February 2005. Response rates were 51% for Caucasians and approximately 70% for African Americans. The survey included questions about previous stroke and individuals who reported previous stroke were excluded from these analyses. Data Analysis All data analyses were carried out using SAS. Initial analysis described the characteristics of the sample. A multiple logistic regression was used to understand the contribution of factors in predicting high stroke risk factor awareness (defined as >= 95 percentile of risk factor awareness scores). Further analysis was done to evaluate the awareness of modifiable and non-modifiable risk factors and awareness of signs. High awareness for all of these outcome variables was defined as greater than or equal to the 95 percentile of awareness scores. Awareness scores were based on unweighted sum of scores from a set of questions. Awareness of risk factor scores were computed on seven modifiable (hypertension, diabetes, high cholesterol, heart disease, obesity, alcoholism, and smoking) and five non-modifiable (age, history of stroke, family history of stroke, race, and gender) risk factors. Respondents were

JHHSA SPRING 2011

473

asked to report whether they thought each one of these risk factors was in fact a risk factor for stroke, and an accuracy score was computed for each of the respondents. Composite scores on awareness of signs were also computed on seven items related to clinical signs and symptoms. Table 1 Demographic Comparisons between Study Respondents, West Virginia, and U.S. Residents Description Age Under 35 35-49 Years 50-64 Years 65-74 Years 75-84 Years 85 and Older Gender Male Female Household Income < $15,000 $15,000-$34,999 $35,000-$49,999 $50,000-$99,999 > $100,000 Risk Factors Hypertension Smoking Obesity High Cholesterol Diabetes Physical Inactivity

West Virginia*

United States*

20.3% 25.6% 33.3% 12.8% 6.1% 1.9%

28.5% 29.5% 22.3% 10.6% 6.8% 2.3%

49.5% 38%

13% 87%

32.7% 67.3%

49% 51%

49.1% 50.9%

25.9% 30.4% 20.4% 19.6% 3.6%

27% 29.9% 21.2% 18.5% 3.4%

23.8% 31.5% 18.9% 21.6% 4.1%

25.4% 32.0% 16.4% 21.2% 5%

15.8% 25.6% 16.5% 19.7% 12.3%

48.7% 30.2% 35.3% 40% 16.7% 39.6%

49.5% 13% 30.4% 46% 14.2% 39.1%

48.5% 43.9% 40.9% 33.3% 19% 41.3%

32.5% 23% 28% 38% 10.2% 28.4%

25.6% 28.4% 22.2% 30% 6.7% 24.4%

All 897

Whites 442

Blacks 400

15.8% 24.6% 32.6% 14.8% 9.4% 2.7%

12% 23.5% 32.4% 17% 12.2% 2.9%

22.2% 77.8%

6.5% 4.4% 1.5%

*Source: 2000 Census & BRFSS

RESULTS This study is based on responses from 823 respondents (without history of stroke). Self-identified racial status was not provided by 33 respondents, resulting in 422 Whites and 368 African Americans (AA). Table 1

474

JHHSA SPRING 2011

shows a comparison of socio-demographic factors of this sample with West Virginia census and Behavioral Risk Factor Surveillance System (BRFSS), as well as United States 2000 census. The age and gender distribution of this sample was significantly different from West Virginia 2000 census. West Virginia census data clearly shows the difference in age structure between West Virginia and US population census. This sample also differed from West Virginia BRFSS data on risk factor distribution. Initial univariate analysis showed significant differences in the ages of Whites and AA. The average age of respondents was 53.3 years with African Americans significantly younger than whites (Whites: 54.4 ± 16.6 vs. AA: 49.7 ± 16.9 years, P value