Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Stroke. Author manuscript; available in PMC 2011 April 1.
Published in final edited form as:
PMCID: PMC2885129

Racial disparities in awareness and treatment of atrial fibrillation: The REasons for Geographic and Racial Differences in Stroke (REGARDS) study

James F. Meschia, MD, Peter Merrill, BS, Elsayed Z. Soliman, MD, MSc, MS, Virginia J. Howard, PhD, Kevin M. Barrett, MD, MSc, Neil A. Zakai, MD, MSc, Dawn Kleindorfer, MD, Monika Safford, MD, and George Howard, DrPH, For the REGARDS Investigators


Background and Objective

Warfarin reduces stroke risk by about 60% in patients with atrial fibrillation (AF). Differences in awareness and treatment of AF may contribute to racial and geographic disparities in stroke mortality. The objective was to examine predictors of awareness of the diagnosis of AF and treatment with warfarin.


REasons for Geographic and Racial Differences in Stroke (REGARDS) is a national, population-based, longitudinal study of 30,239 blacks and whites ≥ 45 years old with oversampling from African Americans and the southeastern stroke belt states. Participants were enrolled Jan 2003–Oct 2007. Data were collected using telephone interview, in-home evaluation, and self-administered questionnaires. The main variable of awareness of AF was defined by a positive answer to “Has a doctor or other health professional ever told you that you had atrial fibrillation?” and whether there was evidence of treatment on the basis of an in-home medications inventory.


From baseline ECG, 432 individuals (88 black and 344 white) had AF. Of these, 88% (360/409) had at least one additional CHADS2 stroke risk factor and 60% (258/432) were aware of their AF. The odds of blacks being aware of their AF were one-third that of whites (OR=0.32 {95% CI: 0.20-0.52}). Among those aware, the odds of blacks being treated with warfarin were only one-fourth as great as whites (OR=0.28 {0.13-0.60}).


Blacks were less likely than whites to be aware of having atrial fibrillation or to be treated with warfarin. Potential reasons for the racial disparity in warfarin treatment warrant further investigation.

Keywords: Racial disparities, warfarin, atrial fibrillation


About 2.3 million adults in the US have atrial fibrillation (AF), and the prevalence is anticipated to rise to 5.6 million by 2050.1 Cerebral embolism with subsequent stroke is the main complication of AF. Randomized controlled trials show that warfarin at a dose adjusted to an International Normalized Ratio of 2.0 to 3.0 reduces the risk of stroke by about 60%.2 Efficacy demonstrated in the trials has been shown to translate into effectiveness in clinical practice.3, 4 Despite this, there remains a missed opportunity to prevent stroke and other thromboembolic complications from AF because many individuals may fail to have the condition diagnosed and/or fail to receive treatment with warfarin. The Anticoagulation and Risk Factors in Atrial Fibrillation (ATRIA) study found that only 62% of patients with AF who appeared from chart review to be ideal candidates for warfarin received drug.5 A review of health care claims for about four million members of a managed care organization showed that 61% of patients with AF were candidates for warfarin, but only 39% had had initiation of warfarin therapy.6

Awareness of having AF may be poor, though this has been less studied than use of warfarin. The West Birmingham Atrial Fibrillation Project found that only 49% of patients with AF could name their cardiologic condition, and only about half of the patients with AF knew that AF posed a risk of thromboembolism.7 Greater awareness of AF and its clinical implications might lead to greater utilization of warfarin, which might in turn lead to greater stroke reduction.

The Reasons for Geographic and Racial Differences in Stroke (REGARDS) study, a US national, longitudinal study of black and white adults, age at least 45 years, offered an opportunity to assess awareness of AF and compare this awareness of the diagnosis with findings on electrocardiography within a cohort of nearly 30,000 participants.


Study population

The goals and design of the REGARDS study have been described in detail elsewhere.8 The study was designed to investigate racial and regional disparities in stroke incidence, and as such, oversampled blacks (42% of the sample size) and residents of the Southeastern “Stroke Belt Region” (56% of the sample from North Carolina, South Carolina, Georgia, Alabama, Mississippi, Tennessee, Arkansas, and Louisiana): two groups that are known to have high stroke mortality rates. The study has approximately equal numbers of men and women (45%/55%). Individuals were recruited from a commercially available list of residents by using a combination of mail and telephone contact with a 49% cooperation rate. Enrollment began in January 2003 and was completed in October 2007. The demographic information, income, medical history, and measures of cognitive function and quality of life were obtained by computer-assisted telephone interview. Standardized physical measures that included height, weight, blood pressure and resting electrocardiogram (ECG) were collected at in-home physical examination performed approximately 3-4 weeks after the telephone interview.

The Institutional Review Boards of participating centers reviewed and approved the study methods. Participants provided written informed consent.

Diagnosis of atrial fibrillation

For this report, AF was diagnosed based on ECG that was recorded during the in-home visits by a health care professional. Staff members were trained in standard procedures with use of centrally-trained supervisors, a web-based program, and continuous quality feedback to individual examiners. The ECGs were sent to a central ECG reading center at Wake Forest University where they were read, coded and interpreted by electrocardiographers blinded to clinical data. The first 8,459 REGARDS participants underwent 7-lead ECG recording acquired by applying the standard 4 limb electrodes and a mid-sternal electrode; the remaining participants underwent standard 12-lead-ECG recording. Whether a participant underwent a 7-lead or 12-lead ECG would not affect the ability to detect atrial fibrillation.

Awareness of AF was defined as a positive response to the question: “Has a physician or a health professional ever told you that you had atrial fibrillation?” The telephone interview that established awareness of AF was conducted prior to the in-home visit where the ECG was performed. Current aspirin and warfarin treatment was defined using an inventory of current medications that was conducted as part of the in-home visit, where all prescription and over-the-counter medications taken in the past two weeks were recorded. The medication inventory was conducted prior to interpretation of the ECG and detection of AF.

Estimating risk of stroke

Because both awareness of AF and aggressiveness of treatment could be higher among participants perceived to be at higher risk of stroke, the risk of future stroke was estimated using the CHADS2 score.9 “CHADS2” is derived from congestive heart failure, hypertension, age > 75 years, diabetes, and history of stroke or TIA. Congestive heart failure was defined as self-report of both orthopnea, defined as answering positively to “Do you ever have to sleep on two or more pillows to help you breathe?” and paroxysmal nocturnal dyspnea, defined as answering positively to “Do you ever wake at night because you are having trouble breathing?”. Hypertension was defined as at least one of the following: systolic blood pressure greater than 140 mmHg, diastolic blood pressure greater than 90 mmHg, or a self-report of currently taking medication to control blood pressure. Diabetes mellitus was defined as at least one of the following: fasting blood glucose ≥ 126 mg/dl, non-fasting glucose ≥ 200 mg/dl, or a self-report of currently taking oral hypoglycemic agent or insulin. History of stroke was defined as positively answering the question “Were you ever told by a physician that you had a stroke?”. Transient ischemic attack was defined as answering positively to the question “Were you ever told by a physician that you had a mini-stroke or TIA, also known as a transient ischemic attack?”. These questions regarding stroke and TIA are the same as used in the Questionnaire for Verifying Stroke-free Status, an instrument of proven reliability and validity.10, 11 Some guidelines recommend aspirin for low risk AF patients, such as those with a CHADS2 score of 0, aspirin or warfarin for moderate risk AF patients, such as those with CHADS2 score of 1, and warfarin for AF patients with a CHADS2 score >1.12 Access to care was defined by self-report as having a “usual source of medical care.” Insurance status was defined by self-report as any kind of health care coverage.

Of the 30,239 participants, 377 were excluded due to missing ECG and 1 missing self-reported AF, resulting in an analysis cohort of 29,861.

Statistical analysis

The determinants of AF awareness, warfarin use, and aspirin use were established using PROC LOGISTIC in SAS 9.2. The association of variables with AF awareness and warfarin use was assessed in: 1) univariate analyses, 2) a full multivariable model (including all predictors simultaneously), and 3) a “most parsimonious” model, created using backwards stepwise elimination where the least significant variable was removed from the model until all variables had a p-value of 0.2 or less. The likelihoods of being treated with warfarin based on race and CHADS2 score were determined using logistic regression as well, after eliminating the interactive effect of race and CHADS2 score, since it was not significant.


Figure 1 shows the distribution of groups based on reported history of AF and ECG findings. A total of 27,558 people had agreement between self-report of AF and evidence of AF on electrocardiography. A total of 2,303 people had a discrepancy between self-report of AF and evidence of AF on electrocardiography.

Figure 1
Schematic of populations from REGARDS analyzed in the study. A fib denotes atrial fibrillation; Hx, history. The entire cohort in this case is everyone with non-missing values for atrial fibrillation on ECG and self-reports of atrial fibrillation variables. ...

Demographic and medical characteristics of participants with and without AF on electrocardiography are shown in Table 1. Of those with AF on ECG, 88 (20.4%) were black, and 344 were white. Median age was 74 years (Interquartile range [IQR] = 69 to 79 years), and 61% reported at least some college education. Prior stroke was reported by 11.1% (48/432), 8.9% (34/384) reported prior transient ischemic attack, and 15.9% (61/384) reported stroke symptoms with no history of stroke or transient ischemic attack. Based on CHADS2, 88% (360/409) had at least one risk factor for stroke beyond AF.

Table 1
Description of the REGARDS cohort with and without ECG evidence of AF

A total of 60% (258/432) individuals with AF on ECG were aware of the diagnosis. Race, education and income were univariately associated with awareness of AF. (Table 2). Race and income remained independent predictors in the full and most parsimonious multivariate models, but education did not. Blacks were less than a third as likely as whites to have been aware that they had AF (OR=0.32 {95%CI, 0.19 to 0.54; p<0.001}).

Table 2
Results of Logistic Regression Models on Awareness of Atrial Fibrillation (i.e., of those with positive ECG evidence of AF, what are the predictors of self-reported AF; n = 432)

In all models, race was the only significant predictor of use of warfarin treatment among individuals aware that they had AF (Table 3). Race remained an independent predictor of warfarin treatment in the full multivariable model. The odds of blacks being treated with warfarin were only one-fourth as great as whites (OR=0.26; {0.12 to 0.57}).

Table 3
Results of Logistic Regression Models on Warfarin Treatment for Atrial Fibrillation (i.e., of those self-reporting AF, what are the predictors of being on warfarin; n = 258).

Table 4 shows the distribution by CHADS2 score and race of individuals who did or did not receive warfarin. In a logistic regression model using race and risk of stroke from AF (CHADS2), race remained a strong predictor of treatment with warfarin, but risk of stroke was not (Table 5).

Table 4
Number On and Not On Warfarin Stratified by Race and CHADS2 Score in the Group IB population.
Table 5
Results of Logistic Regression Model Using Race and CHADS2 Score in the Group IB population.

Because aspirin is a proven but less effective means of preventing thromboembolic complications in patients with AF, we explored predictors for its use among participants who self-reported AF and had electrocardiographic evidence of the arrhythmia. The same set of covariates used to identify predictors of warfarin treatment was used to identify aspirin predictors. The only significant predictor in univariate analysis was sex. The odds ratio for women being on aspirin therapy compared to men was 0.35 (95%CI, 0.18 to 0.68; p=0.002). This relationship remained significant in the full multivariable model (OR= 0.25; 95%CI, 0.11 to 0.59; p=0.002). In the most parsimonious multivariable model the odds ratio for women being on aspirin therapy was 0.35 (95%CI, 0.18 to 0.68;p=0.002). Race was not a predictor of aspirin therapy. The odds of a black individual being on aspirin relative to a white individual was 0.98 (95% CI, 0.35 to 2.73;p=0.72).


We found that among individuals confirmed to have AF by ECG, blacks were about one third as likely to be aware that they had AF as whites in this US national biracial large sample of adult men and women. Because atrial fibrillation is such a powerful risk factor for incident stroke, these finding suggest that lower awareness of atrial fibrillation and reduced likelihood of treatment among blacks may place blacks at higher risk of a stroke event, which in turn could contribute to the higher stroke mortality among blacks. The reasons for the racial discrepancy are not known. Many of the study participants may be undiagnosed, as often AF itself is not symptomatic. Alternatively, these persons may have been diagnosed with the condition, but simply did not remember or understand the condition. As these results are based on self-report, we cannot distinguish these two possibilities.

Alternatively, individuals who are unaware that they have the diagnosis of AF might not have ever been diagnosed with the condition. Whether the participant had or lacked insurance coverage was not an independent predictor of awareness of AF but since the majority of study participants were over age 65 (and are covered by Medicare) only a small percentage did not have any insurance coverage. Further, whether the participant had or lacked access to health care was also not an independent predictor of awareness of AF. Other possible explanations for racial differences in awareness of AF might be differences in either utilization or delivery of health care.

We also found that among those who were aware that they had AF and who had confirmation of the diagnosis of AF, blacks were about one quarter as likely to be treated with warfarin as whites. In striking contrast, risk of stroke as stratified by the CHADS2 score was not a predictor of warfarin use. The fact that risk of future stroke did not significantly alter the likelihood of warfarin use would seem to reflect an evidence-practice gap. Evidence-based guidelines on the use of warfarin in non-valvular AF typically recommend that estimated risk of stroke be part of the decision process regarding long-term anticoagulation.13, 14 One possible reason for a lack of correlation between CHADS2 score and likelihood of treatment with warfarin is that some patients may have had stroke due to intracerebral hemorrhage. Our study collected data on self reported physician diagnosed stroke at baseline. This has been shown to have good sensitivity and specificity. Less is known about the sensitivity and specificity of self-reported stroke subtype (i.e., ischemic stroke versus hemorrhagic stroke). Both because intracerebral hemorrhage is a relatively small contributor to all strokes (about 15%) and because the case fatality of this type of stroke is high, we presume that only a small proportion of strokes at baseline represented intracerebral hemorrhage.

We found that women with documented atrial fibrillation who were also aware of their condition had about a third of the chance as men of being treated with aspirin, a treatment known have efficacy in stroke prophylaxis. The reasons for this relative under- treatment with aspirin are unknown. Others have observed that women are less likely to receive aspirin for secondary prevention of ischemic heart disease. 15, 16

Certain limitations of our study need to be acknowledged. It is possible that awareness of AF was underestimated in this cohort as paroxysmal atrial fibrillation may not have been detected at the time of baseline ECG. The decision to anticoagulate is often complex, involving predictions of benefit from stroke reduction balanced against predictions of hemorrhagic complications. Physicians may be reluctant to anticoagulate patients who have uncontrolled hypertension. Whether this might have been a factor in the likelihood of being treated with warfarin could not be assessed in our study as there was only one baseline measure of blood pressure. While CHADS2 is an example of a well-validated clinical tool for predicting stroke risk, no comparably well validated tool exists for predicting hemorrhage risk. Study subjects were not interviewed regarding whether they were offered warfarin therapy in the past but they refused, i.e., we are not able to distinguish between people who were never offered therapy from informed refusers. Information on contraindications to warfarin therapy was not available so it is not known what proportion were actually warfarin candidates. The interview did not explore whether subjects who reported no active warfarin use had taken and subsequently discontinued warfarin in the past. Finally, our study was not originally designed to test the specific and singular question of racial differences in awareness and treatment of AF. This report is of an exploratory analysis and as such is subject to the limitation common to observational studies of a possible spurious association generated by type I error.

In this large biracial cohort, blacks were less likely to be aware of AF and less likely to be treated with warfarin than whites. These findings are consistent with prior studies demonstrating that blacks are less likely to achieve quality of care goals for stroke risk factors such as glycemic control in diabetes and blood pressure in hypertension.17 Such differences may underlie racial disparities in stroke morbidity and mortality and should lend urgency to focused efforts to improve patient education and medical literacy. The additional finding that CHADS2 score was not a predictor of warfarin use highlights an evidence-practice gap that should prompt further efforts focused on practitioner awareness and education. Future investigations to further understand the determinants of racial disparities in stroke and the impact of improved diagnosis and treatment of atrial fibrillation on race-related stroke outcomes would be of public health importance.


This research project is supported by a cooperative agreement U01 NS041588 from the National Institute of Neurological Disorders and Stroke, National Institutes of Health, Department of Health and Human Service. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Neurological Disorders and Stroke or the National Institutes of Health. Representatives of the funding agency have been involved in the review of the manuscript but not directly involved in the collection, management, analysis or interpretation of the data. The authors acknowledge the participating investigators and institutions for their valuable contributions: The University of Alabama at Birmingham, Birmingham, Alabama (Study PI, Statistical and Data Coordinating Center, Survey Research Unit): George Howard DrPH, Leslie McClure PhD, Virginia Howard PhD, Libby Wagner MA, Virginia Wadley PhD, Rodney Go PhD, Monika Safford MD, Ella Temple PhD, Margaret Stewart MSPH, David Rhodes RN; University of Vermont (Central Laboratory): Mary Cushman MD; Wake Forest University (ECG Reading Center): Ron Prineas MD, PhD; Alabama Neurological Institute (Stroke Validation Center, Medical Monitoring): Camilo Gomez MD, Susana Bowling MD; University of Arkansas for Medical Sciences (Survey Methodology): LeaVonne Pulley PhD; University of Cincinnati (Clinical Neuroepidemiology): Brett Kissela MD, Dawn Kleindorfer MD; Examination Management Services, Incorporated (In-Person Visits): Andra Graham; Medical University of South Carolina (Migration Analysis Center): Daniel Lackland, DrPH; Indiana University School of Medicine (Neuropsychology Center): Frederick Unverzagt, PhD; National Institute of Neurological Disorders and Stroke, National Institutes of Health (funding agency): Claudia Moy, Ph.D.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Contributor Information

James F. Meschia, Mayo Clinic, Department of Neurology.

Peter Merrill, University of Alabama at Birmingham.

Elsayed Z. Soliman, Wake Forest University School of Medicine.

Virginia J. Howard, University of Alabama at Birmingham.

Kevin M. Barrett, Mayo Clinic, Department of Neurology.

Neil A. Zakai, University of Vermont.

Dawn Kleindorfer, University of Cincinnati, Department of Neurology.

Monika Safford, University of Alabama at Birmingham.

George Howard, University of Alabama at Birmingham.


1. Go AS, Hylek EM, Phillips KA, Chang Y, Henault LE, Selby JV, Singer DE. Prevalence of diagnosed atrial fibrillation in adults: National implications for rhythm management and stroke prevention: The anticoagulation and risk factors in atrial fibrillation (atria) study. Jama. 2001;285:2370–2375. [PubMed]
2. Hart RG, Pearce LA, Aguilar MI. Meta-analysis: Antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med. 2007;146:857–867. [PubMed]
3. Go AS, Hylek EM, Chang Y, Phillips KA, Henault LE, Capra AM, Jensvold NG, Selby JV, Singer DE. Anticoagulation therapy for stroke prevention in atrial fibrillation: How well do randomized trials translate into clinical practice? Jama. 2003;290:2685–2692. [PubMed]
4. Kalra L, Yu G, Perez I, Lakhani A, Donaldson N. Prospective cohort study to determine if trial efficacy of anticoagulation for stroke prevention in atrial fibrillation translates into clinical effectiveness. Bmj. 2000;320:1236–1239. [PMC free article] [PubMed]
5. Go AS, Hylek EM, Borowsky LH, Phillips KA, Selby JV, Singer DE. Warfarin use among ambulatory patients with nonvalvular atrial fibrillation: The anticoagulation and risk factors in atrial fibrillation (ATRIA) study. Ann Intern Med. 1999;131:927–934. [PubMed]
6. Darkow T, Vanderplas AM, Lew KH, Kim J, Hauch O. Treatment patterns and real-world effectiveness of warfarin in nonvalvular atrial fibrillation within a managed care system. Curr Med Res Opin. 2005;21:1583–1594. [PubMed]
7. Lane DA, Ponsford J, Shelley A, Sirpal A, Lip GY. Patient knowledge and perceptions of atrial fibrillation and anticoagulant therapy: Effects of an educational intervention programme. The west Birmingham atrial fibrillation project. Int J Cardiol. 2006;110:354–358. [PubMed]
8. Howard VJ, Cushman M, Pulley L, Gomez CR, Go RC, Prineas RJ, Graham A, Moy CS, Howard G. The REasons for Geographic And Racial Differences in Stroke study: Objectives and design. Neuroepidemiology. 2005;25:135–143. [PubMed]
9. Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: Results from the national registry of atrial fibrillation. Jama. 2001;285:2864–2870. [PubMed]
10. Meschia JF, Lojacono MA, Miller MJ, Brott TG, Atkinson EJ, O'Brien PC. Reliability of the Questionnaire for Verifying Stroke-Free Status. Cerebrovasc Dis. 2004;17:218–223. [PubMed]
11. Jones WJ, Williams LS, Meschia JF. Validating the Questionnaire for Verifying Stroke-Free Status (QVSFS) by neurological history and examination. Stroke. 2001;32:2232–2236. [PubMed]
12. Kakar P, Boos CJ, Lip GY. Management of atrial fibrillation. Vasc Health Risk Manag. 2007;3:109–116. [PMC free article] [PubMed]
13. Association EHR, Heart Rhythm Society ZD, Camm AJ, Borggrefe M, Buxton AE, Chaitman B, Fromer M, Gregoratos G, Klein G, Moss AJ, Myerburg RJ, Priori SG, Quinones MA, Roden DM, Silka MJ, Tracy C, Smith SC, Jr, Jacobs AK, Adams CD, Antman EM, Anderson JL, Hunt SA, Halperin JL, Nishimura R, Ornato JP, Page RL, Riegel B, Priori SG, Blanc JJ, Budaj A, Camm AJ, Dean V, Deckers JW, Despres C, Dickstein K, Lekakis J, McGregor K, Metra M, Morais J, Osterspey A, Tamargo JL, Zamorano JL. Cardiology ACo; Force AHAT; Guidelines ESoCCfP. Acc/aha/esc 2006 guidelines for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death: A report of the american college of cardiology/american heart association task force and the european society of cardiology committee for practice guidelines (writing committee to develop guidelines for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death) J Am Coll Cardiol. 2006;48:e247–346. [PubMed]
14. Singer D, Albers GW, Dalen JE, Fang MC, Go AS, Halperin JL, Lip G, Manning WJ. Antithrombotic therapy in atrial fibrillation. American college of chest physicians evidence-based clinical practice guidelines. Chest. (8th) 2008;133:546S–592S. [PubMed]
15. Williams D, Bennett K, Feely J. Evidence for an age and gender bias in the secondary prevention of ischaemic heart disease in primary care. Br J Clin Pharmacol. 2003;55:604–608. [PMC free article] [PubMed]
16. Enriquez JR, Pratap P, Zbilut JP, Calvin JE, Volgman AS. Women tolerate drug therapy for coronary artery disease as well as men do, but are treated less frequently with aspirin, beta-blockers, or statins. Gend Med. 2008;5:53–61. [PubMed]
17. Trivedi AN, Zaslavsky AM, Schneider EC, Ayanian JZ. Relationship between quality of care and racial disparities in medicare health plans. Jama. 2006;296:1998–2004. [PubMed]