Differences in Admitting Hospital Characteristics for Black and White Medicare Beneficiaries With Acute Myocardial InfarctionClinical Perspective
Background—Racial disparities in acute myocardial infarction treatment may be due to differences in admitting hospitals. Little is known about factors associated with hospital selection for black and white acute myocardial infarction patients.
Methods and Results—We identified black and white Medicare beneficiaries with acute myocardial infarction in 63 hospital referral regions with at least 50 black admissions during 2005 (n=65 633). We calculated distance from patient home to hospital referral region hospitals using ZIP code centroids. We assessed hospital quality using a composite score made up of hospital risk-adjusted 30-day mortality and acute myocardial infarction performance measures. Hospitals with a score in the top 20% were categorized as high quality, and those in the lowest 20% as low quality. We used conditional multinomial logit models to examine differences in hospital selection for blacks and whites. On average, blacks lived closer to revascularization hospitals (mean, 3.8 versus 6.8 miles; P<0.001) and to high-quality hospitals (mean, 5.6 versus 9.7 miles; P<0.001). After distance was accounted for, blacks were relatively less likely (P<0.001) to be admitted to revascularization hospitals (risk ratio [RR], 0.87; 95% confidence interval [CI], 0.80 to 0.95) and to high-quality hospitals (RR, 0.88; 95% CI, 0.801 to 0.95) but more likely (P<0.001) to be admitted to low-quality hospitals (RR, 1.17; 95% CI, 1.05 to 1.29). In analyses matched by home ZIP code, differences in admissions to revascularization (RR, 0.92; 95% CI, 0.80 to 1.05), high-quality (RR, 0.94; 95% CI, 0.81 to 1.07), and low-quality (RR, 1.15; 95% CI, 0.94 to 1.35) hospitals were not significant.
Conclusions—Differences in admissions to revascularization and high-quality hospitals may contribute to disparities in acute myocardial infarction care. These differences may be due in part to residential ZIP code characteristics.
Over the past 2 decades, numerous studies have demonstrated large disparities in the management of acute myocardial infarction (AMI).1,–,7 Most notably, black patients with AMI are less likely to receive guideline-recommended treatments, even after accounting for patient characteristics1,–,7 and appropriateness of care.5,6
Clinical Perspective on p 2716
A number of recent studies have suggested that differences in characteristics of admitting hospitals may play an important, perhaps primary, role in racial disparities in AMI treatment.8,–,10 These studies have found that the care of black patients is concentrated in a relatively small number of hospitals8 with lower volumes,9 lower adherence to process measures,10 and higher risk-adjusted mortality.11 Yet, little is known about why blacks and whites tend to receive care from hospitals of varying quality.
Prior research has established that where a patient is admitted for hospital care is determined primarily by distance,12,13 and by an array of hospital and patient characteristics.12,–,15 However, no study has specifically addressed the relationship between patient race and admitting hospital and its impact on healthcare disparities. Yet, examining differences in hospital admission for black and white AMI patients in their given geographic context could help us better understand local, regional, and national disparities in coronary care.
The main objective of this study was to examine the association between patient race, distance, and institutional characteristics that might influence admission to particular hospitals for black and white AMI patients. The study framework is adapted from the traditional hospital choice model,12,13 which assumes that admission to a particular hospital given a set of alternatives (ie, hospitals within a healthcare market for cardiac tertiary care) is dependent on the characteristics of available hospitals (ie, location, scope of service, and quality), individual patient characteristics (ie, race, socioeconomic status, and acuity of illness), and distance to available hospitals. Specifically, we hypothesized that characteristics of admitting hospitals will differ significantly for blacks and whites, and that differences will persist after accounting for distance from patient residence to available hospitals. In secondary analyses, we also sought to explore the role of residence in the final admitting hospital choice, matching black and white patients by the ZIP code areas in which they live.
The study used 4 primary data sources: (1) Medicare Provider Analysis and Review (MedPAR) part A data to identify AMI patients and to calculate hospital-level revascularization volumes and mortality rates, (2) the Dartmouth Atlas to define regional healthcare markets, (3) the Centers for Medicare and Medicaid Services Hospital Compare Web site to obtain hospital process performance measures, and (4) the American Hospital Association 2005 survey to obtain additional hospital characteristics.
We identified all Medicare patients admitted to US hospitals with a primary diagnosis of AMI (International Classification of Diseases, ninth revision, clinical modification [ICD-9-CM] code 410) during 2005. Because hospital choices for patients with a prior AMI are likely to be influenced by a different set of factors (eg, experience with illness and treating hospital), we excluded patients with a primary diagnosis of AMI within the previous 3 years and patients <68 years of age to have 3 years of Medicare data.
The MedPAR files contain data on all Medicare fee-for-service hospitalizations, including patient ZIP code and demographics, primary and secondary diagnoses, and procedures, as defined by ICD-9-CM codes, admission source (eg, transfer from another hospital), discharge disposition, and a 6-digit unique identifier for the admitting hospital.
We assigned patients and admitting hospitals to distinct healthcare markets represented by 2005 hospital referral regions. Patients and hospitals were allocated through the use of home and hospital ZIP codes following Dartmouth Atlas algorithms (http://www.dartmouthatlas.org). We excluded patients who lived in ZIP codes outside the market area where they were admitted, (n=25 393 [13%]) because these patients might have been traveling at the time of the AMI, thus rendering distance from home to available hospitals inaccurate. We also excluded patients who were transferred from another hospital (n=15 007 [8%]), because these patients' selection of hospitals would likely be influenced by a set of factors different from those determining the initial admission for treatment. We excluded patients for whom race was missing or defined as other than black or white (n=2093 [4.2%]), limiting our comparisons to black and white patients with a first episode of AMI.
Finally, we excluded hospital referral regions with <50 blacks admitted with AMI during 2005 (n=243) to allow meaningful statistical comparisons. Characteristics of study hospital referral regions are provided in the Material section in the online-only Data Supplement.
Characteristics of Hospitals Within Selected Markets
On the basis of prior literature on hospital choice,12,–,15 we focused primarily on 2 hospital characteristics expected to influence hospital choice: availability of revascularization services (scope of service) and hospital quality of care (defined below).
We categorized hospitals as having revascularization programs if they performed at least 2 coronary artery bypass graft surgeries during 2005 on the basis of calculated MedPAR volumes.
We assessed hospital quality using process performance measures and hospital-specific, risk-adjusted mortality rates. For both types of measures, we assumed that hospital selection is influenced by how patients perceive hospital quality of care before admission; hence, we used data collected before 2005 (ie, 2004 data for process measures and 2002 to 2004 data for mortality rates).
We downloaded 2004 hospital AMI performance measures from the Hospital Compare Web site (www.hospitalcompare.hhs.gov). We selected 5 core AMI process measures for our analyses: (1) use of aspirin within 24 hours of arrival, (2) use of aspirin at discharge, (3) use of β-blockers within 24 hours of arrival, (4) use of β-blockers at discharge, and (5) use of an angiotensin-converting enzyme inhibitor or angiotensin receptor blocker for left ventricular systolic dysfunction. We excluded 3 additional AMI measures available from the Hospital Compare database (time to reperfusion with either thrombolytics or percutaneous coronary intervention and smoking cessation counseling), because they were reported by relatively few hospitals. For each measure, a hospital's score is calculated as the number of patients who received a given treatment divided by the number of patients eligible for the treatment. We used the 5 measures to create a single AMI process performance score for each hospital as the total number of treatments in a given hospital divided by the total number of treatments eligible in each hospital (ie, the sum of individual measure numerators divided by the sum of individual measure denominators).16 We excluded 110 hospitals (1044 patients) that either had not reported measures for 2004 or had reported less than a combined 25 eligible patients for these measures, thus rendering the scores less reliable.16
We calculated risk-adjusted mortality rates using data from Medicare beneficiaries admitted with AMI during 2002 to 2004. Hospital-specific risk-adjusted 30-day mortality rates were constructed as the ratio of predicted to expected mortality at each hospital multiplied by the mean 30-day mortality rate for the population with the use of a methodology similar to the one used by Centers for Medicare and Medicaid Services.17 Patient-level variables drawn from MedPAR records included demographics, comorbidities identified from ICD-9-CM diagnostic codes using well-established algorithms,18,19 and AMI location, defined as anterior and lateral, inferior and posterior, subendocardial, and other unspecified site, on the basis of ICD-9-CM codes. Predicted mortality and expected 30-day mortality were estimated with multivariable mixed models that included hospital random effects. Covariates included in the mortality model are displayed in Appendix I in the online-only Data Supplement.
For each hospital, we created a composite quality score comprising the process performance score and the 30-day hospital-level survival rate (ie, 1 minus the standardized 30-day mortality rate). The quality score was calculated with a methodology similar to the one proposed by Centers for Medicare and Medicaid Services' Hospital Quality Incentive Demonstration (HQID) project,20 which weighs each of the process measures and standardized mortality equally. Thus, global scores for each hospital were calculated according to the following formula: (5/6×process performance score)+(1/6× 30-day hospital survival rate). We defined high-quality hospitals as all hospitals ranking in the top 20% on the composite score (n=248); we defined low-quality hospitals as hospitals ranking in the bottom 20% on the composite score (n=249). All other hospitals (n=747) were defined as intermediate-quality hospitals.
We obtained other hospital characteristics from the 2005 American Hospital Association annual survey. We further excluded 8 hospitals (202 patients) that we were unable to link to the 2005 American Hospital Association survey. The survey provided data on hospitals' structural and operational characteristics, including teaching status, ownership (eg, for-profit status), and Medicaid caseload. We categorized hospitals as safety-net providers if the hospital's 2005 Medicaid caseload (based on American Hospital Association data) exceeded the mean for all hospitals within the state by 1 SD.
We estimated travel distance as the straight-line distance from the patient's home residence to all hospitals within the patient's home hospital referral region using home and hospital 5-digit ZIP code–based centroids.
For black and white patients, we compared patient characteristics and distance to and characteristics of admitting hospitals using χ2 statistics for dichotomous measures and t tests or Wilcoxon rank-sum tests for continuous measures. We used conditional logit (McFadden) models21 to examine the likelihood of admission to particular hospitals, given the characteristics of available hospitals within the patient's market. Conditional logit models, which assume that the selection of a particular hospital is conditional on the characteristics of other hospitals in the market, have been widely used in research evaluating determinants of hospital choice.12,14,15,22 In this case, we estimated the impact of hospital distance, presence of revascularization services, quality, teaching status, safety-net status, and for-profit status on the likelihood of admission to a given hospital. For each hospital characteristic, models incorporated separate effects for black and white patients that provided race-specific odds of admission to specific hospitals. Race-specific model estimates associated with each hospital characteristic were exponentiated to obtain the odds of admission to a hospital with a particular characteristic; race-specific probabilities of admission to a hospital with a particular characteristic were calculated from the following formula: odds/(1+odds). Risk ratios for black relative to white patients were calculated as the ratios of the race-specific probabilities of admission to a hospital with a given characteristic, with confidence intervals based on the pooled variance of the race-specific coefficients.
To evaluate the robustness of our findings, we performed several secondary analyses. To determine whether severity of illness influenced hospital choice, we conducted analyses stratified by AMI location. Furthermore, we calculated patient-level expected 30-day mortality rates (ie, the rate of death that would be expected if the patients had been treated at an average hospital, given the average hospital's quality-of-care effect on mortality) on the basis of the mortality risk adjustment model described above. Patients were stratified into low, intermediate, and high risk of death on the basis of the expected mortality being in the lowest, middle 2, and highest quartiles of the distribution. Analyses were conducted for each risk stratum.
To control for the impact of unmeasured differences in the characteristics of the environment in which blacks and whites live on differences in admitting hospital characteristics, we repeated analyses on a subset of blacks and whites pair matched by home ZIP code (n=10 422). Thus, if a ZIP code had 5 black and 3 white AMI patients, the matched sample would contain 3 black and 3 white patients; 2 white patients would be excluded randomly.
P values were 2 sided. Statistical significance was defined using a criterion of P<0.01. All analyses were performed with SAS statistical software version 9.2 (SAS Institute Inc, Cary, NC).
There were 57 342 whites (87.4%) and 8291 blacks (12.6%) admitted during 2005 with a primary diagnosis of AMI to 1244 US hospitals in markets with at least 50 black AMI admissions for the study year. In unadjusted analyses, blacks were younger, more likely women, and more likely to reside in urban areas with lower ZIP code income (Table 1). Further characteristics of the study population are provided in Table 1.
Blacks lived substantially closer to hospitals with revascularization programs and to both high- and low-quality hospitals relative to whites, and traveled on average less for admission (Table 2). The closest hospital was also a revascularization hospital for 49% blacks and 47% whites, a high-quality hospital for 18% blacks and 23% whites, a low-quality hospital for 15% blacks and 12% whites, a teaching hospital for 22% blacks and 10% whites, and a safety-net hospital for 12% blacks and 74 whites (P<0.001 for all comparisons).
In unadjusted analyses, proportions of admissions to hospitals with revascularization programs were only slightly higher for blacks relative to whites with AMI (Table 3). A lower proportion of blacks were admitted to high-quality hospitals, whereas a higher proportion of blacks were admitted to low-quality hospitals. Although statistically significant, these differences were small. Blacks were also more likely to be admitted to safety-net hospitals and to teaching hospitals relative to whites (Table 3).
Table 4 shows how the likelihood of being admitted to a particular hospital is affected by specific characteristics of available hospital and distance from home to hospitals within patients' markets. The impact of each hospital characteristic on the likelihood of admission to a particular hospital type is shown separately for blacks and whites (columns 1 and 2) and for blacks relative to whites (ie, the ratio of black and white probabilities of admission to hospitals with particular characteristics, column 3).
For both blacks and whites, the 2 most influential hospital characteristics were proximity to the patient's residence and availability of revascularization services (Table 4). Thus, whites were nearly 4 times more likely to select the closest hospital over hospitals located farther from home, whereas blacks were 2.6 times more likely to select the closest hospital, controlling for distance to hospitals and other hospital characteristics. Similarly, whites and blacks were 3.5 times and 2.6 times more likely, respectively, to select a revascularization hospital over other hospitals without revascularization services.
Measures of hospital quality of care were somewhat less influential on the admitting hospital selection, although the relationship was statistically significant (Table 4). Sensitivity analyses using 15% and 25% cutoffs for quality definitions yielded similar results.
Lastly, blacks were more likely and whites were less likely to select teaching hospitals. Both blacks and whites were less likely to select safety-net hospitals and for-profit hospitals compared with hospitals without those characteristics (Table 4). However, black patients with AMI were relatively less likely to select the closest hospital, hospitals providing coronary revascularization, and high-quality hospitals but relatively more likely to select low-quality hospitals, teaching hospitals, and safety-net hospitals compared with white patients (Table 4, relative risk for black versus white).
Secondary analyses stratified by predicted risk of death or AMI location yielded similar results. For example, blacks with low, intermediate, or high risk of death were 13%, 12%, and 13% less likely to be admitted to high-quality hospitals compared with whites with a similar death risk.
To evaluate the effect of where patients live on hospital admission, we matched 63% of blacks (n=5211) and 9% of whites (n=5211) by home ZIP code. These patients were younger (mean age, 80.2 versus 80.7; P<0.001) and more likely to live in urban areas (73.0% versus 68.9%; P<0.001) and in ZIP codes with lower median household income ($36 900 versus $54 100; P<0.001) relative to unmatched patients.
In matched-sample analyses (Table 5), racial differences in the likelihood of admission to closest hospitals, to revascularization hospitals, and to high- and low-quality hospitals were smaller and not statistically significant, but the statistical power was somewhat reduced (eg, the power to detect a 10% black-white difference in likelihood of admission was 0.83 for high-quality hospitals and 0.39 for low-quality hospitals). Blacks were still more likely than whites to be admitted to teaching and safety-net hospitals.
Despite living substantially closer to hospitals with revascularization programs and to high-quality hospitals, after adjustment for distance, black patients with AMI were overall 13% and 12% less likely to be admitted to these hospitals relative to white patients treated for AMI in the same markets. On the other hand, black patients were 17% relatively more likely to be admitted to low-quality hospitals and 30% more likely to be admitted to teaching and safety-net hospitals. In additional analyses restricted to black and white patients living in similar ZIP codes, hospital choices based on revascularization services and quality did not differ significantly, but black patients were still more likely to be admitted to teaching and safety-net hospitals compared with whites. Several of the study findings have potentially important implications and are discussed below.
The finding that black patients were relatively less likely than white patients to be admitted to the closest hospital may be significant in light of strong evidence that, at least in the case of ST-segment elevation AMI, prompt coronary reperfusion improves survival after AMI.23,24 Although we were not able to identify ST-segment elevation AMI within administrative data, black patients were consistently less likely to select the closest hospital in analyses stratified by AMI location (ie, anterior and lateral, inferior and posterior, and subendocardial). However, given that the absolute difference in proportions of blacks and whites admitted to the closest hospital was very low, the finding is unlikely to have a detectable impact on clinical outcomes for blacks and whites in this patient sample.
Second, although black patients lived closer to revascularization hospitals, this geographic proximity did not translate into higher likelihood of admission to such hospitals. Several other studies have evaluated access to hospitals providing high-technology cardiac services as a primary source of racial disparity in AMI treatment. An earlier study by Blustein and Weitzman25 examined the association between race and admission to hospitals offering coronary revascularization services for patients admitted with AMI in the state of New York during 1986. The study found that black and white AMI patients presented equally to hospitals offering coronary revascularization services, but after adjustment for distance to the nearest high-technology hospital, the white-to-black odds ratio for likelihood of admission to such a hospital was 1.68. This apparent discrepancy was attributed to racial differences in travel patterns (eg, blacks being less likely to travel beyond the closest hospital when the closest hospital was not a revascularization hospital). However, these analyses did not examine other hospital characteristics that may significantly influence hospital choice (eg, hospital quality, teaching status), and were limited to the state of New York. Thus, results may not be generalizable to different hospital characteristics and to other US geographic areas. Indeed, the racial difference in the likelihood of admission to revascularization hospitals in our study was considerably smaller.
Compared with this prior analysis, our study has several distinct advantages. The choice of a particular hospital over alternative hospitals in the market is explicitly modeled on the basis of hospitals available within a patient's market. Moreover, analyses are rooted in a theoretical framework for hospital choice and control for distance and an array of hospital characteristics known to influence choice beyond the technological capabilities of the hospitals.
In fact, our study found racial differences in admitting hospital choice for multiple hospital characteristics. Notably, black patients with AMI were less likely to use high-quality hospitals and more likely to use low-quality hospitals relative to white patients, even though black patients lived substantially closer to high-quality hospitals than white patients.
This finding stands in contrast to our recently published study evaluating racial differences in admission to high-quality hospitals for coronary heart disease.26 The study, which identified high-quality hospitals as hospitals consistently ranking on the US News and World Report's list of best hospitals for cardiac care, found that black AMI patients were more likely to be admitted to these high-quality hospitals relative to white patients. The apparently incongruent findings are likely explained by the use of different definitions for high-quality hospitals in the 2 studies. In the first study, high-quality hospitals were represented by a relatively unique group of highly reputable, large teaching institutions. For the present study, we defined high-quality hospitals on the basis of self-reported process measures and risk-adjusted mortality; this definition captured hospitals with varied characteristics located in a larger number of markets that do not necessarily overlap with the US News and World Report's top ranked hospital markets. Importantly, the discrepant results of these 2 studies once again highlight the role of location and the choice of quality indicators in evaluating racial disparities. The role of location is further strengthened by the finding that racial differences in admission to high-and low-quality hospitals in the present study did not persist in analyses of black and white patients in the same residential areas.
Perhaps the most informative finding of the study is the fact that racial differences in hospital admission persisted after accounting for distance to available hospitals but were generally smaller and not statistically significant in analyses restricted to patients matched by home ZIP code. Before matching, racial differences might be explained by a number of unmeasured factors, including differences in mode of transportation to the hospital,27 differences in patient preferences for particular hospitals based on institutional factors28 (eg, patient and staff demographics), physician referral,29 or perceived discrimination.30 Although the lack of statistical significance in matched samples might be related to the diminished statistical power owing to small sample size, it more likely indicates that racial differences in access to high-quality hospitals are primarily driven not by race but by characteristics of neighborhoods in which black and white patients live, including socioeconomic and cultural factors, and local healthcare networks. Indeed, geography is increasingly recognized as a source of racial disparities in health care,31 leading the Institute of Medicine to conclude that “minority access to better quality facilities is often limited by the geographic distribution of care facilities and patterns of residential segregation, which results in higher quality facilities being less accessible.”32
Several study limitations merit further discussion. First, analyses do not reveal true patient preferences, but rather observed hospital admission patterns. Thus, results should be interpreted cautiously in light of potential racial differences in hospital preference based on social factors (eg, segregation) acting within geographic boundaries and generating distrust in medical institutions and providers. Second, the study sample was limited to black and white Medicare beneficiaries. Prior research has shown that hospital choice is influenced by other factors, including age, disease type and severity, and insurance status. Thus, findings may not be applicable to other age groups, private insurance or indigent patients, and other disease settings. Third, the study did not differentiate between ST-segment elevation AMI and non–ST-segment elevation AMI, yet patterns of admission might differ significantly between these 2 conditions owing to disease acuity. Although blacks tend to be diagnosed slightly less often with ST-segment elevation AMI,33 it is unlikely that these differences would fully explain the results of our study. Moreover, analyses stratified by AMI location showed consistent findings. Lastly, we calculated travel distance as straight-line distance between patient home ZIP code and hospital ZIP code. Although this method has been validated in prior studies,34 it is possible that, especially in urban markets, straight-line distance significantly underestimates actual road travel distance and time.
Despite these limitations, the present study sheds important light on factors contributing to racial disparities in AMI treatment. The study confirms the role of differential access to hospitals with revascularization services and high quality of care as a plausible source of disparity. However, racial differences in access to high-quality hospitals appear to be primarily driven not by race in itself, but by differences in where the majority of blacks and whites live and seek care. Effective policy recommendations aimed at reducing disparities need to take local socioeconomic and healthcare system factors into consideration.
Sources of Funding
For data acquisition, management, and analysis, this research was supported in part by an award (REA 09-220) from the Department of Veterans Affairs Health Services Research and Development Service. Dr Popescu is supported by a K08 career development award (HL095930-01) from the National Heart, Lung, and Blood Institute. Dr Vaughan-Sarrazin is supported by a Research Enhancement Award Program (REA 09-220) and by IIR-08067 from the Department of Veterans Affairs Health Services Research and Development Service. Dr Cram is employed by the Department of Veterans Affairs and CADRE at the Iowa City VA. His work is funded by R01 HL085347-01A1 from NHLBI at the NIH. Dr. Cram has received payment for advising Vanguard Health, an operator of for-profit hospitals, on quality improvement efforts. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. The funding sources had no role in the analyses or drafting of this manuscript. None of the authors have any conflicts of interest.
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.110.973628/DC1.
- Received June 21, 2010.
- Accepted April 21, 2011.
- © 2011 American Heart Association, Inc.
- Canto JG,
- Allison JJ,
- Kiefe CI,
- Fincher C,
- Farmer R,
- Sekar P,
- Person S,
- Weissman NW
- Manhapra A,
- Canto JG,
- Barron HV,
- Malmgren JA,
- Taylor H,
- Rogers WJ,
- Weaver WD,
- Every NR,
- Borzak S
- Sabatine MS,
- Blake GJ,
- Drazner MH,
- Morrow DA,
- Scirica BM,
- Murphy SA,
- McCabe CH,
- Weintraub WS,
- Gibson CM,
- Cannon CP
- Sonel AF,
- Good CB,
- Mulgund J,
- Roe MT,
- Gibler WB,
- Smith SC Jr.,
- Cohen MG,
- Pollack CV Jr.,
- Ohman EM,
- Peterson ED
- Skinner J,
- Chandra A,
- Staiger D,
- Lee J,
- McClellan M
- Williams SC,
- Koss RG,
- Morton DJ,
- Loeb JM
- Krumholz HM,
- Wang Y,
- Mattera JA,
- Wang Y,
- Han LF,
- Ingber MJ,
- Roman S,
- Normand SL
CMS/Premier Hospital Quality Initiative Demonstration (HQID). http://www.premierinc.com/quality-safety/tools-services/p4p/hqi/specifications-by-focus-area.jsp. Accessed May 21, 2010.
- Zaremba P
- McFadden D
- De Luca G,
- Suryapranata H,
- Ottervanger JP,
- Antman EM
- Canto JG,
- Zalenski RJ,
- Ornato JP,
- Rogers WJ,
- Kiefe CI,
- Magid D,
- Shlipak MG,
- Frederick PD,
- Lambrew CG,
- Littrell KA,
- Barron HV
- Smedley BD,
- Stith AY,
- Nelson AR
- Shaw LJ,
- Shaw RE,
- Merz CN,
- Brindis RG,
- Klein LW,
- Nallamothu B,
- Douglas PS,
- Krone RJ,
- McKay CR,
- Block PC,
- Hewitt K,
- Weintraub WS,
- Peterson ED
Recent research has identified differential access to high-quality hospital care as an important, perhaps primary, factor associated with racial disparities in the treatment of acute myocardial infarction. However, the reasons why minorities are more often treated at hospitals with relatively lower quality of care are not fully understood. We examined the relationship between patient race, distance, and admissions to hospitals with particular characteristics (eg, provision of revascularization services, quality ranks) available to patients within defined hospital markets. We found that, after accounting for distance from home to available hospitals, black patients were overall less likely to be admitted to the closest hospitals, to hospitals with revascularization programs, and to high-quality hospitals. In secondary analyses limited to black and white patients living in similar ZIP codes, these differences were attenuated and not statistically significant. Our findings suggest that differences in admissions to hospitals with particular characteristics are not driven exclusively by race, but also by characteristics of the environment in which patients live. Such local context factors may include economic factors, community culture (eg, perceived discrimination, high levels of distrust in the healthcare system), and physician referral practices. Practicing physicians need to be aware of and sensitive to such potential local barriers in access to high-quality care.