Race/Ethnicity, Quality of Care, and Outcomes in Ischemic Stroke
Background— Prior studies suggest differences in stroke care associated with race/ethnicity. We sought to determine whether such differences existed in a population of black, Hispanic, and white patients hospitalized with stroke among hospitals participating in a quality-improvement program.
Methods and Results— We analyzed in-hospital mortality and 7 stroke performance measures among 397 257 patients admitted with ischemic stroke to 1181 hospitals participating in the Get With The Guidelines-Stroke program 2003 through 2008. Relative to white patients, black and Hispanic patients were younger and more often had diabetes mellitus and hypertension. After adjustment for both patient- and hospital-level variables, black patients had lower odds relative to white patients of receiving intravenous thrombolysis (odds ratio [OR], 0.84; 95% confidence interval [CI], 0.77 to 0.91), deep vein thrombosis prophylaxis (OR, 0.88; 95% CI, 0.83 to 0.92), smoking cessation (OR, 0.85; 95% CI, 0.79 to 0.91), discharge antithrombotics (OR, 0.88; 95% CI, 0.84 to 0.92), anticoagulants for atrial fibrillation (OR, 0.84; 95% CI, 0.75 to 0.94), and lipid therapy (OR, 0.91; 95% CI, 0.88 to 0.96), and of dying in-hospital (OR, 0.90; 95% CI, 0.85 to 0.95). Hispanic patients received similar care as their white counterparts on all 7 measures and had similar in-hospital mortality. Black (OR, 1.31; 95% CI, 1.28 to 1.35) and Hispanic (OR, 1.16; 95% CI, 1.11 to 1.20) patients had higher odds of exceeding the median length of hospital stay relative to whites. During the study, quality of care improved in all 3 race/ethnicity groups.
Conclusions— Black patients with stroke received fewer evidence-based care processes than Hispanic or white patients. These differences could lead to increased risk of recurrent stroke. Quality of care improved substantially in the Get With The Guidelines-Stroke Program over time for all 3 racial/ethnic groups.
Received May 20, 2009; accepted January 15, 2010.
The burden of stroke is higher in black and Hispanic patients as compared with white patients.1,2 Ischemic stroke occurs at an earlier age in black patients, and the prevalence and age-standardized mortality due to ischemic stroke are higher as compared with the general population.1,3–5 Hispanic individuals appear to be at greater risk for stroke as compared with non-Hispanic whites living in the same communities.6 A few reports have suggested that the quality of care and clinical outcomes for hospitalized stroke patients may vary between different racial/ethnic groups.5,7–10 However, these reports have been limited in size and scope, and their findings have been inconsistent. In addition, few data exist on the trends over time in the quality of care for hospitalized stroke patients of different racial or ethnic backgrounds. It is important to understand stroke treatment patterns and clinical outcomes in minority populations who traditionally have not been well represented in large epidemiological studies or randomized clinical trials. We sought to compare the quality of care in black, white, and Hispanic hospitalized stroke patients over 5 years in a large, national quality-improvement stroke registry and determine whether differences in care and in-hospital outcomes were present.
Clinical Perspective on p 1501
Details of the Get With The Guidelines (GWTG)-Stroke program have been described previously.9,11 Data from patients hospitalized with ischemic stroke at a GWTG-Stroke participating hospitals between April 1, 2003, and September 30, 2008, are included in this analysis. Trained personnel at hospitals were instructed to use GWTG-Stroke to collect data on consecutive patients admitted to the hospital with a principal clinical diagnosis of stroke or transient ischemic attack (TIA). Case ascertainment of admissions for ischemic stroke or TIA was conducted by prospective clinical identification and retrospective identification with International Classification of Diseases (9th revision) discharge codes (433.xx, 434.xx, and 436 for ischemic stroke; 435.xx for TIA), followed by chart review to confirm case eligibility, or a combination of both approaches. To ensure that patients assigned a TIA diagnosis truly had deficits caused by ischemia, hospitals were instructed to include TIA patients only if symptoms were still persistent at the time of hospital arrival, symptoms lasted <24 hours, and no alternative diagnosis was identified.
Data were collected with the Internet-based Patient Management Tool (Outcome Inc, Cambridge, Mass). Data were coded, deidentified, and transmitted in a secure manner. Data were collected for each hospitalization, including patient demographics, medical history, initial head computed tomography findings, in-hospital treatments and events, discharge treatments and counseling, and discharge destination. The data collection tool supports both concurrent and retrospective data entry; concurrent collection was encouraged as a process improvement goal in each hospital. The data abstraction tool includes links to standard data definitions and coding instructions and predefined logic checks to prevent invalid entries.9,11 Training in the use of the tool was provided online and via telephone for all users. Data collected by hospitals were not independently audited by external chart review; however, all sites receive individual data quality reports to promote data completeness and accuracy. GWTG-Stroke also provides annual national reports that examine data quality for every collected data field.
Hospital characteristics, including hospital bed size, academic or nonacademic status, annual volume of stroke discharges, and geographical region, were obtained from the American Hospital Association database.12 Each participating institution received human research approval from their Institutional Review Board to enroll cases in GWTG-Stroke without requiring individual patient consent. The Duke Clinical Research Institute serves as the data analysis center and has an agreement to analyze the aggregate deidentified data for research purposes.
Using methodology similar to national cardiovascular registries, hospital staff recorded patient race/ethnicity data. Information on race and ethnicity could have been obtained from various sources, including patient self-designation, administrative personnel during the registration process, admitting providers, or nursing intake forms. The data entry tool supports single-race, multi-race, and ethnic classification. From 2003 to September 2005, Hispanic ethnicity was captured as part of a multi-select option that also included racial categories (black, white, Asian, American Indian/Alaska native, and Native Hawaiian/Pacific Islander). Starting in October 2005, a separate data element for Hispanic ethnicity (yes versus no/not documented) was implemented following the recommendations of Census 2000.13 There were 427 887 ischemic stroke patients admitted to 1188 participating hospitals during the study period. For this study, we restricted analysis to 3 patient groups: non-Hispanic black (referred to as black), non-Hispanic white (referred to as white), or Hispanic regardless of race (referred to as Hispanic). Patients who were classified as Asian, Pacific Islander, mixed race, or any other nonblack or nonwhite categories (n=28 490, 6.7%) were therefore excluded. Only 1770 patients (0.41%) were excluded because of missing information on race/ethnicity. An additional 360 patients (0.084%) were excluded because of missing information on sex. The final analysis sample consisted of 397 257 ischemic stroke admissions from 1181 hospitals.
The following 7 predefined, evidence-based performance measures9,14 were used to compare the quality of care (QOC) in black, Hispanic, and white patients. These 7 measures were the result of harmonization of performance measurement activities between the American Heart Association, the US Centers for Disease Control and Prevention, and the Joint Commission. Detailed specifications are available for download and review online.15 Six of the 7 measures were endorsed by the National Quality Forum; the smoking cessation measure was not endorsed solely because of an National Quality Forum decision to release an identical global smoking measure applicable to all inpatient care populations. A summary defect-free care measure16 (also known as an all-or-none measure) was used as a measure of overall QOC. Defect-free care is defined as the proportion of patients who received all of the interventions for which they were eligible (patients who were not eligible for any of the 7 measures were excluded).
Acute-term performance measures included the following: intravenous (IV) tissue-type plasminogen activator (tPA) administered to patients who presented to the emergency department within 2 hours after symptom onset and received the drug within 3 hours of symptom onset (“IV tPA <2 hours”); antithrombotic medication (ie, antiplatelet or anticoagulant agents) administered by end of hospital day 2 (“early antithrombotics”); deep vein thrombosis (DVT) prophylaxis (ie, warfarin, heparin, other anticoagulants, or pneumatic pressure devices) administered by end of hospital day 2 to patients who were nonambulatory at hospital day 2 (“DVT prophylaxis”).
Discharge Performance Measures
Discharge performance measures included the following: antithrombotic medication administered at discharge (“antithrombotics”); anticoagulation (ie, warfarin, full-dose heparins, or other anticoagulants) administered at discharge for any atrial fibrillation (paroxysmal, persistent, or permanent) (“anticoag for atrial fibrillation”); lipid-lowering agent administered at discharge to patients with low-density lipoprotein >100, or those on lipid-lowering agents before admission, or in whom low-density lipoprotein was not measured in the past 30 days (“low-density lipoprotein 100 or not documented”); counseling or medication to help quit smoking provided at or before discharge for patients who are current smokers (any cigarettes in past year; “smoking cessation”).
In-Hospital Clinical Outcomes
In-hospital clinical outcomes included the following: the in-hospital case fatality rate for all ischemic stroke admissions, excluding those discharged to hospice (“in-hospital mortality”); the proportion of patients discharged to home versus any other destination (“discharge home”); the median (interquartile range) and mean duration of hospitalization (“length of stay”).
Contingency tables were generated to explore the relationship between black, Hispanic, and white race/ethnicity status and other demographic and clinical variables, as well as hospital-level characteristics. Because of prior literature suggesting that there are substantial differences between hospitals that do or do not serve large minority populations, even after accounting for hospital size and teaching status,7 we created a hospital-level variable expressing the percentage of minority stroke patients (defined as black and Hispanics) treated at each hospital. Similar contingency tables were generated to explore the relationship between race/ethnicity and compliance with the individual and defect-free QOC measures. Pearson χ2 test for nominal data and Kruskal-Wallis tests for ordinal and continuous data were used as tests for statistical associations. Statistical significance was defined as P≤0.01.
The relationship between race/ethnicity (black versus white or Hispanic versus white) and compliance with individual QOC measures, the summary defect-free care measure, and stroke outcomes were then examined using multivariable logistic regression models. Given the large data set, traditional model-building approaches that identify independent predictors based on statistical significance were not used. Instead, the final adjusted models included the following prespecified potential patient-level confounders, which were selected on the basis of published data supporting their association with stroke risk or outcome: age, sex, body mass index, atrial fibrillation, previous stroke/TIA, coronary heart disease or prior myocardial infarction, carotid stenosis, diabetes mellitus, peripheral vascular disease, hypertension, dyslipidemia, and current smoking. In addition, hospital-level characteristics associated with stroke risk or outcome were also included (geographic region, number of beds, academic versus not, annual stroke volume). We also included the hospital-level term for percentage of minority patients treated (specified in quartiles). Inclusion of this model term allows us to more accurately distinguish hospital-level determinants from patient-level determinants of differences in care in the models. Because there were improvements in QOC measure adherence over time, we included adjustments for both calendar time and time in program (specified as years of hospital program participation, which varied between 1 and 6 years). To account for the hierarchical nature of the data, we used hierarchical modeling techniques, including hospital effects, as a random intercept term.17 Separate hierarchical multivariable logistic regression models were developed for each QOC measure, the summary defect-free care measure, in-hospital mortality, discharge status (home versus other), and length of stay >4 days versus ≤4 days (median length of stay, 4 days). We developed a hierarchical multivariable logistic regression model to assess how the summary defect-free care measure changed for each race/ethnicity group with increasing time in program for the participating hospitals (years of program participation could vary between 1 and 6 years, and comparisons were made to the baseline period). The model also adjusted for calendar time. In addition, to further evaluate the potential impact of age on our results, we undertook stratified analyses (using 10-year age groups) by racial/ethnic group for the summary defect-free care measure and in-hospital mortality.
All statistical analyses were performed using SAS version 9.1 software (SAS Institute, Cary, NC).The authors had full access to and take full responsibility for the integrity of the data in the study and the accuracy of the data analysis. All authors have read and agree to the manuscript as written.
There were 397 257 ischemic stroke admissions included for analysis. Just over half (52.7%) of the population were women, and the average age was 71.1±14.5 (SD) years. The race/ethnicity distribution was 79.1% white (n=314 102), 16.0% black (n=63 370), and 4.9% Hispanic (n=19 785). There were significant differences in demographic and clinical characteristics by racial/ethnic group (Table 1). Compared with white patients, black and Hispanic patients were substantially younger and less likely to present by ambulance. Black patients were more likely than white patients to have a past medical history of prior stroke/TIA, diabetes mellitus, hypertension, and smoking and were less likely to have a history of atrial fibrillation, heart disease, carotid stenosis, peripheral vascular disease, and dyslipidemia (all P<0.0001; Table 1). Compared with white patients, Hispanic patients were less likely to have a past medical history of atrial fibrillation, heart disease, and carotid stenosis, but were more likely to have diabetes mellitus.
The 1181 hospitals included in the analysis were from every state in the continental United States, Puerto Rico, and the District of Columbia (median 14 hospitals per state; range 1 to 119). The median hospital admission volume for ischemic stroke was 203 (interquartile range 59 to 464). There was significant variability across the 3 race/ethnicity groups in hospital characteristics, including bed size, hospital type, volume of stroke discharges, and geographic location (Table 1). Black patients were more likely to be treated at larger, academic hospitals than white or Hispanic patients.
When hospitals were stratified into quartiles of the percentage of minority patients treated, significant differences in hospital characteristics between racial groups still remained. Hospitals that admitted a greater proportion of minority stroke patients were larger (median bed size was 188, 292, 311, and 348 for quartiles 1 through 4 of minority stroke admissions, respectively; P<0.0001). The percentage of academic hospitals also increased with greater minority stroke admissions (38.6%, 42.5%, 51.1%, and 59.5% for quartiles 1 through 4, respectively; P<0.0001).
There were statistically significant race/ethnicity differences in unadjusted adherence rates for 4 of 7 QOC measures (DVT prophylaxis, early antithrombotics, lipid-lowering medication prescribed at discharge, and smoking cessation) and the summary defect-free care measure (Table 2). There were no differences in unadjusted rates of use of IV tPA, early antithrombotics, and anticoagulation for atrial fibrillation. When hospitals were stratified by quartile of percentage of minority patients treated, these unadjusted QOC findings did not change (data not shown). When multivariate models were constructed adjusting for patient-level characteristics only, significant and important differences in QOC emerged, predominantly in relation to early antithrombotics, antithrombotics at discharge, and smoking cessation, which, relative to whites, were used less in both black and Hispanic patients (Table 3). When comparing differences in QOC between black and white patients, the magnitude of the odds ratios (ORs) and the statistical significance of the IV tPA, DVT prophylaxis, lipid-lowering treatment, and defect-free care measures were amplified when the fully-adjusted model using both hospital-level and patient-level characteristics was generated (Table 3). Black patients had significantly lower adjusted odds compared with white patients of receiving IV thrombolysis (OR, 0.84; 95% confidence interval [CI], 0.77 to 0.91), DVT prophylaxis (OR, 0.88; 95% CI, 0.83 to 0.92), the discharge use of antithrombotics (OR, 0.88; 95% CI, 0.84 to 0.92), anticoagulation for atrial fibrillation (OR, 0.84; 95% CI, 0.75 to 0.94), smoking cessation (OR, 0.85; 95% CI, 0.79 to 0.91), and cholesterol-lowering medication (OR, 0.91; 95% CI, 0.88 to 0.96), as well as the defect-free measure of QOC (OR, 0.90; 95% CI, 0.88 to 0.93). In contrast, after adjustment for patient- and hospital-level variables, differences between Hispanic and white patients were lessened and remained significant for only antithrombotics at discharge and smoking cessation. In age-stratified analyses of the defect-free care measure, we found that findings were similar with the exception of patients <50 years of age (online-only Data Supplement).
There were racial and ethnic differences in unadjusted stroke-related outcomes at discharge (Table 4). Black (4.37%) and Hispanic patients (4.90%) were less likely than white patients (6.06%) to die in hospital, and black (47.9%) and Hispanic patients (52.6%) were more likely to be discharged home as compared with whites (44.0%), who were more likely to be discharged to a skilled nursing facility as compared with black and Hispanic patients (Table 4). Black (6.60 days) and Hispanic patients (6.34 days) had a longer length of hospital stay than whites (5.49 days).
In contrast to QOC measure adherence, findings based on multivariable models of clinical outcome that adjusted for only patient-level factors were similar to the fully-adjusted models that also included hospital-level factors (Table 5). After multivariable adjustment for both patient- and hospital-level characteristics, black patients had lower odds than whites of dying in hospital (OR, 0.90; 95% CI, 0.85 to 0.95) or being discharged home (OR, 0.83; 95% CI, 0.81 to 0.86) and higher odds of exceeding the median length of stay (OR, 1.31; 95% CI, 1.28 to 1.35). By contrast, Hispanic patients had similar in-hospital mortality to whites but greater odds to exceed the median length of stay (OR, 1.16; 95% CI, 1.11 to 1.20) and be discharged home (OR, 1.13; 95% CI, 1.08 to 1.18; Table 5). Age-stratified analyses showed the in-hospital mortality comparison of Hispanic versus white patients was consistent across all age groups (online-only Data Supplement). The age-stratified analyses of black versus white patients revealed greater differences in younger (50 to 59 years) and older (80 to 89 years and >90 years) patients than in other age groups.
Trends in Care and Outcomes
Finally, we explored whether differences in QOC changed in association with the amount of time that hospitals participated in the GWTG-Stroke program (ie, time in program). To explore this, we analyzed changes in the summary defect-free care measure by years of hospital participation in the program within each race/ethnicity group. There were substantial and significant absolute increases in adherence to the defect-free care measure (absolute increase of 30% or more) for all 3 racial/ethnic groups (Figure). Detailed data on trends over time for each individual measure are available in the online-only Data Supplement. When compared with baseline, this substantial improvement persisted after full multivariable adjustment in all 3 racial/ethnic groups (Table 6). However, when the rate of improvement in white patients was compared in the fully adjusted models with the rate of improvement for blacks and for Hispanics, there was no evidence of significant interaction with time in program (black versus white, P=0.85; Hispanic versus white, P=0.16).
Because of its size, national scope, and duration, the GWTG-Stroke program14 provides a unique opportunity to analyze presenting characteristics, quality indicators, and in-hospital stroke outcomes as a function of race/ethnicity. In this analysis of nearly 400 000 black, Hispanic, and white patients admitted to <1100 US hospitals for acute ischemic stroke, we demonstrated that there are important differences in the baseline characteristics, as well as QOC, among different racial/ethnic groups. We found evidence that black patients had lower odds of receiving certain guideline-based stroke interventions relative to white patients—these include short-term treatments (IV thrombolysis and DVT prophylaxis) and discharge interventions (antithrombotics, anticoagulation for atrial fibrillation, smoking cessation, and cholesterol control), which were reflected in the summary measure of care. Importantly, many of these differences were greater after risk adjustment for patient and hospital characteristics. In contrast, there were limited differences in QOC observed between Hispanic and white ischemic stroke patients. The observation that Hispanic patients receive similar QOC compared with white patients has been previously reported.18 Although the absolute differences and ORs in individual QOC measures were small, taken together they amount to a consistent pattern of modestly reduced odds for black patients of receiving guideline-recommended care compared with white patients.
We report 2 separate models, 1 adjusted for patient characteristics alone and the other adjusted for patient and hospital characteristics (Table 3). The models adjusted for patient characteristics alone are presented because the Institute of Medicine19 recommends that these models be used to provide the most appropriate estimates of potential disparities. Although we found no evidence that these differences in care quality had a measurable adverse impact on short-term in-hospital outcomes, this is not unexpected since most of the process measures being assessed are discharge measures. To assess the association between these care differences and clinical outcomes will require analysis of postdischarge (eg, day 90 or 1 year) health status, stroke disability, and mortality. It is important to note that randomized clinical trials of IV tPA have not shown a difference for early outcomes with this therapy or a difference in mortality; therefore, studies that include postdischarge functional status will be required to assess the relationship of this process measure to outcome. Although there were only modest differences in DVT prophylaxis in black patients, these are unlikely to be large enough to result in a difference in overall in-hospital mortality due to pulmonary embolism. The higher odds of being able to ambulate at discharge or be discharged to home after hospitalization for acute ischemic stroke seen in the black and Hispanic patients as compared with white patients in the study may be due to residual confounding by age or other unmeasured factors, rather than differences in hospital QOC. For example, black and Hispanic patients in this study had a higher prevalence of risk factors (especially hypertension and diabetes mellitus) associated with small vessel stroke, which is generally a milder stroke subtype with lower early recurrence rates and better functional recovery.20–22
There has been evidence for racial differences in short- and long-term ischemic cardiovascular care documented for black patients in prior studies.23–25 Fewer data have been reported for ischemic stroke. In a study of 1837 patients with ischemic strokes from a single state-based registry, black patients with ischemic stroke had lower odds of receiving a door-to-computed tomography time of less than 25 minutes (OR, 0.13; 95% CI, 0.05 to 0.32), obtain cardiac monitoring (OR, 0.54; 95% CI, 0.29 to 1.03]), undergo dysphagia screening (OR, 0.69; 95% CI, 0.50 to 0.95), or receive smoking cessation counseling (OR, 0.27; 95% CI, 0.17 to 0.42).7 In a medical record review from 42 academic medical centers, blacks were one fifth as likely to be administered tPA as whites, a finding that persisted after adjustment (OR, 0.21; 95% CI, 0.06 to 0.68) for sex, severity, and insurance status.8 In another study in community hospitals, black patients had lower odds of receiving IV tPA (OR, 0.54; 95% CI, 0.31 to 0.95) even after controlling for factors including severity, physician specialty, and location.26 However, a study in the Michigan Medicare population using retrospective medical record abstraction27 found no differences in the use of antithrombotics at discharge by race or sex after adjustment for confounders.
Previous studies evaluating racial and ethnic differences in ischemic stroke care have suffered from several important limitations. First, some studies have relied on administrative databases27 or have included chart abstraction data from only a few hospitals or limited geographical regions or populations.18 In contrast, our data are from nearly 400 000 ischemic stroke patients from <1100 hospitals of all types (ie, community, academic, Veterans Administration) and is therefore a more nationally representative sample. In contrast to some of these prior studies, we examined a broad range of guideline-recommended QOC measures, from acute care to secondary prevention.
Although it is possible that observed treatment differences could simply reflect differences in patient characteristics, in this study, racial/ethnic differences persisted or even emerged after controlling for patient-level factors. Controlling for hospital site characteristics further amplified this effect for some QOC measures, particularly among black patients, whereas it had a lesser effect on Hispanic patients. Although certain hospital characteristics such as teaching status and size are commonly thought to play a critical role in QOC differences, our study suggests that this may not be the case. We did not find evidence of poorer performance by sites that admit a disproportionately large number of black or Hispanic patients (the quartile of percentage of minority cases treated) to account for the observed differences. It is of course possible that our findings are the result of other important unmeasured confounders, such as stroke severity, patient preference, or access to a neurologist.
Black patients had lower odds of in-hospital mortality in our study, even after adjustment for age and other factors. Other studies have shown higher rates of risk-adjusted in-hospital mortality and ischemic stroke death for black patients,5,2829 although some reports of poststroke adjusted mortality that extend beyond hospital discharge have suggested lower rates among black patients30 and Hispanic patients when compared with whites.29 Recent data suggest a decline in stroke mortality among all minorities.31 Hispanic patients in our study had in-hospital mortality similar to white patients. The predilection for small vessel lacunar strokes in black patients has already been mentioned as an explanation for their better long-term prognosis.20–22 Studies in Northern Manhattan have shown higher relative rates for lacunar and intracranial atherosclerotic strokes compared with cardioembolic strokes among both Hispanic and black patients compared with whites.32,33 Other groups have not found important differences in severity of initial stroke presentation associated with racial or ethnic groups.34,35 Because ischemic stroke subtype and stroke severity by National Institutes of Health Stroke Scale score were recorded infrequently in the patient charts, our risk-adjusted model could not account for them.
GWTG-Stroke is a program designed to improve provider knowledge and education for hospitalized stroke patients and to systematically accelerate the use of evidence-based and guideline-recommended therapies. Given the enormous economic burden of stroke and its disproportionate impact on black and Hispanic compared with white patients,36 it is critical to implement strategies that increase adherence to evidence-based care in this high-risk group. Our study suggests that the implementation of process-of-care intervention programs such as GWTG-Stroke benefits all patients by increasing rates of evidence-based therapies for ischemic stroke. Further study will needed to determine whether equivalent rates of adherence to inpatient measures of guideline-based care among black and Hispanic patients will translate into similar long-term postdischarge outcomes.
This study has several limitations. Because program participation is voluntary, the observed findings may not be representative of hospitals that chose not to participate. However, the GWTG-Stroke database includes hospitals in every state in the United States, and the population in GWTG-Stroke is similar in racial makeup, although with a lower proportion of Hispanic ethnicity when compared with the US population of 75.1% white, 12.3% black, and 12.5% Hispanic as reflected by Census 2000 data.13 GWTG-Stroke has a similar proportion of female patients and prevalence of stroke risk factors compared with other large stroke registries.10,37 Although no formal data reliability studies were performed within this quality-improvement registry, the Michigan Paul Coverdell prototype (which used a very similar data set) found acceptable inter-rater reliability for most of the data elements.38 In many hospitals, race and ethnicity may have been assigned by administrative personnel during the registration process, by admitting providers, nursing intake forms, or patient self-designation. The attribution of race/ethnicity that is not based on direct self-report may be less reliable.39 Further, the change to documenting race and ethnicity as separate data elements is relatively recent, and not all hospitals may have the tools in place to do this. Because of variability between hospitals in the process to establish patient ethnicity, it is possible that Hispanic ethnicity is under-reported in this registry. However, prior studies suggest, based on medical records review, that this may occur in <5% of cases.40 If so, this misclassification would result in a small underestimation of the magnitude of the Hispanic versus white differences in measure conformity and outcomes. Because GWTG collects only deidentified data that are routinely present in the medical record, socioeconomic data are not collected. This study reports ORs, which should not be interpreted as indicating relative risk.
In conclusion, in this very large, diverse racial/ethnic group of ischemic stroke patients treated at hospitals across the United States over a 6-year period, we found substantial differences in baseline characteristics among black, Hispanic, and white patients, particularly in age of presentation and stroke risk factors. Furthermore, there was evidence of a moderate but consistent reduction in the frequency with which guideline-based care was provided to black as compared with white patients when adjusted for patient and hospital factors. Our findings contrast with some previous reports of substantially poorer QOC for black patients with ischemic stroke; however, our findings are consistent with previous reports showing that Hispanic patients received QOC comparable to that of white patients. Substantial improvements in care quality were observed over time in the GWTG-Stroke program for each racial/ethnic group. Although reassuring, further efforts are needed to improve QOC for ischemic stroke for all patients. Expansion of national strategies for the hospital-based implementation of quality improvement in ischemic stroke care should be considered, and the planned introduction of stroke measures into the Joint Commission core measure set15 and Medicare inpatient prospective payment system update41 should help accelerate these efforts.
The authors wish to thank all the hospital personnel and local QI champions involved in Get With the Guidelines-Stroke for their hard work and dedication to improving patient care over the past four years.
Sources of Funding
Get With the Guidelines-Stroke is funded by the American Heart Association and the American Stroke Association. The program is also supported in part by unrestricted educational grants to the American Heart Association by Pfizer Inc, New York, NY, and the Merck-Schering Plough Partnership (North Wales, Pa), who did not participate in the design, analysis, manuscript preparation, or approval.
Dr Schwamm reports in the past 2 years being supported as a consultant on stroke systems development by the Massachusetts Department of Public Health. Dr Schwamm serves as chair of the GWTG Steering Committee of the American Heart Association (AHA). He has provided expert medical opinions in malpractice lawsuits. Dr Reeves reports in the past 2 years receiving no research support, consulting fees, or speaking fees from pharmaceutical companies. He is supported as a consultant to the Michigan Stroke GWTG Registry and serves as a member of the GWTG Improvement Subcommittee of the AHA. Dr Pan reports in the past 2 years receiving no research support, consulting fees, or speaking fees from pharmaceutical companies. She serves as a member of the Duke Clinical Research Institute, which serves as the AHA GWTG data coordinating center. Dr Smith reports in the past 2 years receiving no research support, consulting fees, or speaking fees from pharmaceutical companies. He serves as a member of the AHA GWTG Science Subcommittee. Dr Frankel reports in the past 2 years receiving no research support, consulting fees, or speaking fees from pharmaceutical companies. He has provided expert medical opinions in malpractice lawsuits. He is supported as a consultant to the Georgia Coverdell Stroke Registry, funded by the US Centers for Disease Control and Prevention (CDC) and Georgia Division of Public Health. He is supported by research funding from the National Institute of Neurological Disorders and Stroke and CDC. He was the Principal Investigator at Emory University for the Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) trial sponsored by Pfizer Inc. He serves as a member of the AHA GWTG Steering Committee. Dr Olson reports in the past 2 years receiving no research support, consulting fees, or speaking fees from pharmaceutical companies. He serves as a member of the Duke Clinical Research Institute, which serves as the AHA GWTG data coordinating center. Zhao reports in the past 2 years receiving no research support, consulting fees, or speaking fees from pharmaceutical companies. She serves as a member of the Duke Clinical Research Institute, which serves as the AHA GWTG data coordinating center. Dr Peterson has received research grants from Bristol-Meyers Squibb-Sanofi and Merck-Schering Plough in the past 2 years and has served as a consultant to Astra Zeneca and Schering Plough. Dr Peterson serves as Principle Investigator of the Data Analytic Center for the GWTG of the AHA. Dr Fonarow reports receiving research grants from National Heart, Lung, and Blood Institute; receiving honoraria in the past 2 years from Bristol-Myers Squibb, Merck, and Pfizer; and serving as a consultant for Bristol-Myers Squibb, Merck, and Pfizer. Dr Fonarow served as former chair of the AHA GWTG Steering Committee.
Lloyd-Jones D, Adams R, Carnethon M, De Simone G, Ferguson TB, Flegal K, Ford E, Furie K, Go A, Greenlund K, Haase N, Hailpern S, Ho M, Howard V, Kissela B, Kittner S, Lackland D, Lisabeth L, Marelli A, McDermott M, Meigs J, Mozaffarian D, Nichol G, O'Donnell C, Roger V, Rosamond W, Sacco R, Sorlie P, Stafford R, Steinberger J, Thom T, Wasserthiel-Smoller S, Wong N, Wylie-Rosett J, Hong Y. Heart disease and stroke statistics–2009 update: a report from the Am Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation. 2009; 119: 480–486.
Sacco RL, Boden-Albala B, Gan R, Chen X, Kargman DE, Shea S, Paik MC, Hauser WA. Stroke incidence among white, black, and Hispanic residents of an urban community: the Northern Manhattan Stroke Study. Am J Epidemiol. 1998; 147: 259–268.
Smith MA, Risser JM, Lisabeth LD, Moye LA, Morgenstern LB. Access to care, acculturation, and risk factors for stroke in Mexican Americans: the Brain Attack Surveillance in Corpus Christi (BASIC) project. Stroke. 2003; 34: 2671–2675.
Sacco RL, Boden-Albala B, Abel G, Lin IF, Elkind M, Hauser WA, Paik MC, Shea S. Race-ethnic disparities in the impact of stroke risk factors: the Northern Manhattan Stroke Study. Stroke. 2001; 32: 1725–1731.
Stansbury JP, Jia H, Williams LS, Vogel WB, Duncan PW. Ethnic disparities in stroke: epidemiology, acute care, and postacute outcomes. Stroke. 2005; 36: 374–386.
Morgenstern LB, Smith MA, Lisabeth LD, Risser JM, Uchino K, Garcia N, Longwell PJ, McFarling DA, Akuwumi O, Al-Wabil A, Al-Senani F, Brown DL, Moye LA. Excess stroke in Mexican Americans compared with non-Hispanic Whites: the Brain Attack Surveillance in Corpus Christi Project. Am J Epidemiol. 2004; 160: 376–383.
Jacobs BS, Birbeck G, Mullard AJ, Hickenbottom S, Kothari R, Roberts S, Reeves MJ. Quality of hospital care in African American and white patients with ischemic stroke and TIA. Neurology. 2006; 66: 809–814.
Johnston SC, Fung LH, Gillum LA, Smith WS, Brass LM, Lichtman JH, Brown AN. Utilization of intravenous tissue-type plasminogen activator for ischemic stroke at academic medical centers: the influence of ethnicity. Stroke. 2001; 32: 1061–1068.
Schwamm LH, Fonarow GC, Reeves MJ, Pan W, Frankel MR, Smith EE, Ellrodt G, Cannon CP, Liang L, Peterson E, Labresh KA. Get With the Guidelines: stroke is associated with sustained improvement in care for patients hospitalized with acute stroke or transient ischemic attack. Circulation. 2009; 119: 107–115.
Reeves MJ, Arora S, Broderick JP, Frankel M, Heinrich JP, Hickenbottom S, Karp H, LaBresh KA, Malarcher A, Mensah G, Moomaw CJ, Schwamm L, Weiss P. Acute stroke care in the US: results from 4 pilot prototypes of the Paul Coverdell National Acute Stroke Registry. Stroke. 2005; 36: 1232–1240.
Reeves MJ, Fonarow GC, Zhao X, Smith EE, Schwamm LH. Quality of care in women with ischemic stroke in the GWTG program. Stroke. 2009; 40: 1127–1133.
American Hospital Association. American Hospital Association Hospital Statistics. 2007 ed. Chicago, IL: Am Hospital Association; 2007.
United States Census Bureau. Population Profile of the United States: 2000. Washington, DC: US Census Bureau; 2000.
American Heart Association. Get with The Guidelines-Stroke. Available at: http://www.strokeassociation.org/presenter.jhtml?identifier= 3002728. Accessed March 5, 2010.
The Joint Commission. The Joint Commission Core Measure Set. Available at: http://www.jointcommission.org/PerformanceMeasurement/ PerformanceMeasurement/STK+Core+Measures.htm. Accessed March 5, 2010.
Molenberghs G, Verbeke G. Models for Discrete Longitudinal Data. New York: Springer; 2006.
Tuhrim S. Editorial comment: ethnic disparities in stroke: epidemiology, acute care, and postacute outcomes. Stroke. 2005; 36: 386–387.
Smedley BD, Stith AY, Nelson AR. Unequal Treatment: Confronting Racial and Ethnic Disparities in Healthcare. Washington, DC: Institute of Medicine & National Academies Press; 2003.
Sprigg N, Gray LJ, Bath PM, Lindenstrom E, Boysen G, De Deyn PP, Friis P, Leys D, Marttila R, Olsson JE, O'Neill D, Ringelstein EB, van der Sande JJ, Turpie AG. Early recovery and functional outcome are related with causal stroke subtype: data from the tinzaparin in acute ischemic stroke trial. J Stroke Cerebrovasc Dis. 2007; 16: 180–184.
Petty GW, Brown RD Jr, Whisnant JP, Sicks JD, O'Fallon WM, Wiebers DO. Ischemic stroke subtypes: a population-based study of functional outcome, survival, and recurrence. Stroke. 2000; 31: 1062–1068.
Lovett JK, Coull AJ, Rothwell PM. Early risk of recurrence by subtype of ischemic stroke in population-based incidence studies. Neurology. 2004; 62: 569–573.
Reed SD, Cramer SC, Blough DK, Meyer K, Jarvik JG. Treatment with tissue plasminogen activator and inpatient mortality rates for patients with ischemic stroke treated in community hospitals. Stroke. 2001; 32: 1832–1840.
Lisabeth LD, Roychoudhury C, Brown DL, Levine SR. Do gender and race impact the use of antithrombotic therapy in patients with stroke/TIA? Neurology. 2004; 62: 2313–2315.
Gillum R. Stroke Incidence, Comorbid Conditions, Health Care Utilization, and Posstroke Mortality in Black and White Medicare Beneficiaries. Basel, Switzerland: Karger; 1999.
Ayala C, Greenlund KJ, Croft JB, Keenan NL, Donehoo RS, Giles WH, Kittner SJ, Marks JS. Racial/ethnic disparities in mortality by stroke subtype in the United States, 1995–1998. Am J Epidemiol. 2001; 154: 1057–1063.
Wolfe CD, Smeeton NC, Coshall C, Tilling K, Rudd AG. Survival differences after stroke in a multiethnic population: follow-up study with the South London stroke register. BMJ. 2005; 331: 431.
Keppel KG, Pearcy JN, Wagener DK. Trends in racial and ethnic-specific rates for the health status indicators: United States, 1990–98. Healthy People 2000 Stat Notes. 2002: 1–16.
Sacco RL, Shi T, Zamanillo MC, Kargman DE. Predictors of mortality and recurrence after hospitalized cerebral infarction in an urban community: the Northern Manhattan Stroke Study. Neurology. 1994; 44: 626–634.
White H, Boden-Albala B, Wang C, Elkind MS, Rundek T, Wright CB, Sacco RL. Ischemic stroke subtype incidence among whites, blacks, and Hispanics: the Northern Manhattan Study. Circulation. 2005; 111: 1327–1331.
Jones MR, Horner RD, Edwards LJ, Hoff J, Armstrong SB, Smith-Hammond CA, Matchar DB, Oddone EZ. Racial variation in initial stroke severity. Stroke. 2000; 31: 563–567.
Uchino K, Risser JM, Smith MA, Moye LA, Morgenstern LB. Ischemic stroke subtypes among Mexican Americans and non-Hispanic whites: the BASIC Project. Neurology. 2004; 63: 574–576.
Brown DL, Boden-Albala B, Langa KM, Lisabeth LD, Fair M, Smith MA, Sacco RL, Morgenstern LB. Projected costs of ischemic stroke in the United States. Neurology. 2006; 67: 1390–1395.
Reeves MJ, Broderick JP, Frankel M, LaBresh KA, Schwamm L, Moomaw CJ, Weiss P, Katzan I, Arora S, Heinrich JP, Hickenbottom S, Karp H, Malarcher A, Mensah G, Reeves MJ. The Paul Coverdell National Acute Stroke Registry: initial results from four prototypes. Am J Prev Med. 2006; 31: S202–S209.
Department of Health and Human Services, Centers for Medicare & Medicaid Services. Medicare Program; Changes to the Hospital Inpatient Prospective Payment Systems for Acute Care Hospitals and Fiscal Year 2010 Rates; and Changes to the Long Term Care Hospital Prospective Payment System and Rate Years 2010 and 2009 Rates; Final Rule; 42 CFR Parts 412, 413, 415. Federal Register. 2009; 74: 43753–44236.
Prior studies suggest differences in stroke care associated with race/ethnicity. Get With The Guidelines (GWTG)-Stroke has been shown to improve the quality of care of patients hospitalized with stroke or transient ischemic attack. We sought to determine whether such differences existed in a population of black, Hispanic, and white patients hospitalized with ischemic stroke from 2003 to 2008 among hospitals participating in GWTG-Stroke. Relative to white patients, black and Hispanic patients were younger and had different risk factor profiles, including more frequent diabetes mellitus and hypertension. After adjustment for both patient- and hospital-level variables, we found evidence of moderate but persistent differences in delivery of evidence-based care. Compared with white patients, black patients were less likely to receive specific evidence-based interventions (eg, intravenous thrombolysis, deep vein thrombosis prophylaxis, smoking cessation, discharge antithrombotics, anticoagulants for atrial fibrillation, and lipid therapy), but for Hispanic patients, the care was similar. There were substantial improvements in quality of ischemic stroke care observed over time in the GWTG-Stroke program for each racial/ethnic group (absolute increase of 30% or more in the defect-free summary measure). Our findings of only moderate differences in quality for black patients contrast with some previous reports, which suggested substantially poorer quality of care. Our findings are consistent with previous reports suggesting that Hispanic patients received quality of care comparable to that of white patients. Expansion of national strategies for hospital-based implementation of quality improvement in ischemic stroke care should be undertaken, with a special emphasis on reducing disparity and differences in stroke care delivery.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.109.881490/DC1.