- Email: [email protected]
Racial Disparities in Revascularization Rates Among Patients With Similar Insurance Coverage
o
r
i
g
i
n
a
l
c
o
m
m
u
n
i
c
a
t
i
o
n
Racial Disparities in Revascularization Rates Among Patients With Similar Insurance Coverage Peter Cram, MD, MBA; Levent Bayman, MS; Ioana Popescu, MD, MPH; Mary S. Vaughan-Sarrazin, PhD
Disclosure: Dr Vaughan-Sarrazin is a research scientist in the Center for Research in the Implementation of Innovative Strategies in Practice at the Iowa City Veterans Administration Medical Center, which is funded through the Department of Veterans Affairs, Veterans Health Administration, Health Services Research, and Development Service. J Natl Med Assoc. 2009;101:xxx–xxx Funding/Support: Dr Cram is supported by a K23 career development award (RR01997201) from the National Center for Research Resources at the National Institutes of Health and the Robert Wood Johnson Physician Faculty Scholars Program. This work is also funded by R01 HL085347-01A1 from National Heart, Lung, and Blood Institute at the NIH. The funding sources had no role in the analyses or drafting of this manuscript. Disclaimer: The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. Background: Racial disparities in coronary revascularization—percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG)—have been extensively documented. However, it is unclear whether disparities are consistent among patients with similar health insurance coverage. Our objective was to assess racial disparities in coronary revascularization among white, black, and Hispanic patients with similar insurance coverage hospitalized with acute myocardial infarction (AMI). Methods: We used 2000-2005 state inpatient data for 9 states to identify white, black, and Hispanic patients hospitalized with AMI. Patients were grouped into 3 health insurance cohorts: (1) Medicare, (2) private insurance, and (3) Medicaid/uninsured. We examined use of revascularization (PCI or CABG) among blacks and Hispanics as compared to whites in each of the 3 insurance cohorts. Author Affiliations: Results: The 418 study hospitals admitted 430 509 AMI patients Corresponding Author:238 956 with private insurance, and 74 926 with Medicare,
patients who were uninsured/Medicaid. In unadjusted analyses, black and Hispanic patients were significantly less likely to receive in-hospital revascularization among the Medicare cohort (38.9% vs 44.9% vs 47.3%, P < .001), privately insured cohort (62.9% vs 69.7% vs 74.2%, P < .001), and uninsured/Medicaid cohort (55.2% vs 61.0% vs 68.4%, P < .001). In Cox models
1132 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
adjusting for patient demographics, comorbidity, and clustering of patients within hospitals, blacks were approximately 25% less likely and Hispanics 5% less likely to receive revascularization as compared to whites with similar insurance. Conclusions: Blacks hospitalized with AMI are significantly less likely to receive revascularization when compared to whites and Hispanics with similar health insurance. Our data suggest that patients’ ability to pay for costly procedures is unlikely to explain racial disparities. Keywords: racial disparities n health insurance n heart J Natl Med Assoc. 2009;101:1132-1139 Author Affiliations: Division of General Internal Medicine, Department of Internal Medicine, University of Iowa Carver College of Medicine (Drs Cram, Bayman, Popescu, and Vaughan-Sarrazin); and Center for Research in the Implementation of Innovative Strategies for Practice (CRIISP), Iowa City Veterans Administration Medical Center (Drs Cram and Vaughan-Sarrazin), Iowa City, Iowa. Corresponding Author: Peter Cram, MD, MBA, Division of General Medicine, University of Iowa Carver College of Medicine, Iowa City VA Medical Center, Mail Stop 152, Iowa City, IA 52242 ([email protected]).
Background
R
acial disparities in cardiac care are well studied.1,2 Decades of research have demonstrated that compared to whites, blacks are less likely to receive an array of high-cost cardiac interventions, including implantable cardioverter defibrillators,3 and coronary revascularization procedures such as percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG).4,5 Moreover, racial disparities in cardiac care have persisted despite significant efforts to insure equal access among all patient populations.6,7 While the existence of disparities in coronary revascularization is well established, the relationship between health insurance and racial disparities is less clear. More specifically, it is uncertain whether racial disparities in revascularization are similar in magnitude among patients with similar insurance coverage. Many studies of disparities in coronary revascularization rates have focused on patients covered under a single insurance plan such as Medicare2,4,8 or the Veterans Administration.5 Far fewer studies have examined disparities in revascuVOL. 101, NO. 11, NOVEMBER 2009
Racial disparities in cardiac revascularization
larization rates among patients with private insurance or underinsured populations.9,10 Moreover, to the best of our knowledge, prior studies have not systematically investigated whether racial disparities in revascularization rates are consistent among white, black, and Hispanic patients with similar types of insurance. The objective of the current study was to examine rates of in-hospital revascularization (PCI and CABG) for white, black, and Hispanic patients with similar insurance coverage (or lack thereof) hospitalized with acute myocardial infarction (AMI). Our a priori hypothesis was that racial disparities in revascularization would be smaller among whites, blacks, and Hispanics with private insurance or Medicare and larger among patients who were uninsured.
Methods Data We used state inpatient data for the years 2000-2005 from Arizona, California, Florida, Massachusetts, Maryland, New Jersey, and Wisconsin; and 2000-2004 from New York and Texas to identify all patients age ≥20 years hospitalized with AMI on the basis of International Classification of Diseases, Ninth Clinical Modification (ICD9-CM) code 410.X (N = 1 623 956). The state inpatient databases were developed by the Agency for Healthcare Research and Quality (AHRQ) as part of the Health Care Utilization Project.11 The state inpatient data include many elements available on the UB-92 hospital discharge abstract, including patient demographics (eg, race, sex); admitting hospital; primary and secondary diagnoses and procedures, as captured by ICD-9-CM codes; the diagnosis-related group (DRG); admission source (eg, emergency department, transfer from another hospital); admission and discharge dates; patient’s primary insurance (categorized as Medicare, private insurance, Medicaid, self-pay, other); type of insurance (fee for service or health maintenance organization); and disposition at the time of hospital discharge (eg, transfer to another acute care hospital, deceased). The state inpatient data for many states are processed and sold by the AHRQ central distributor and thus have uniform coding, but data for other states are processed and sold by individual state planning agencies and thus have variable coding. For example, many states code age as a continuous variable, but others provide age as a categorical variable or choose to mask age of individual patients treated in low-volume hospitals to protect patient confidentiality; thus, merging data from multiple states, while entirely feasible, does require significant finesse such as converting age of all patients into categorical variables. In addition, the state inpatient data do not contain unique patient identifiers and do not allow for tracking patients after discharge or transfer from the admitting hospital. As our primary objective was to assess racial disparities in revascularization, we limited our study to patients whose JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
race was listed as white, black, or Hispanic and excluded patients with race listed as Asian, Native American, missing, or “other” because the number of patients in each category was too small to allow for meaningful comparisons to be made (N = 247 311). We also excluded patients whose insurance status was missing (N = 210 735) and patients admitted to hospitals that did not perform revascularization (1290 hospitals that admitted 630 101 AMI patients).
Statistical Analyses After identifying all white, black, and Hispanic patients hospitalized with AMI, patients were stratified into 3 different cohorts based upon their insurance coverage using a scheme we developed previously.12 In particular, AMI patients were categorized as having either: (1) Medicare (fee for service or managed care); (2) private insurance (fee for service or managed care); and 3) uninsured/Medicaid (uninsured, self-pay, or Medicaid). All analyses described below were conducted separately for each of the 3 insurance cohorts. As the primary objective of the study was to assess in-hospital revascularization rates for white, black, and Hispanic patients admitted with AMI, the first step in our analyses was to identify patients who had received revascularization. We defined revascularization as receipt of either PCI (ICD-9 procedure codes 36013607) or CABG (ICD-9 codes 3610-3619) during the AMI admission. We used bivariate methods to compare the demographic characteristics, comorbidity, and clinical risk factors of white, black, and Hispanic patients within each of the 3 insurance cohorts. Comorbid conditions were identified using algorithms developed by Elixhauser and Deyo13,14 and updated by Quan et al.15 Additional clinical risk factors were identified using coding schemes that have been used in prior studies assessing cardiovascular outcomes using administrative data. This included AMI location (coded as anterlateral [ICD-9 codes 410.0, 410.1, and 410.5], inferposterior [410.2, 410.3, 410.4, and 410.6], subendocardial [410.7] and other [410.8 and 410.9], previous CABG surgery (ICD-9 code V4581), previous PCI (ICD-9 code V4582), use of intra-aortic balloon pump during the admission (procedure codes 3761 and 3762), use of mechanical ventilation during the admission, and whether the patient received revascularization (PCI or CABG) during admission.16,17 We used multivariable survival models to estimate the hazard ratio of receiving revascularization for blacks and Hispanics as compared to whites for patients in each of the 3 insurance cohorts; Cox proportional hazards regression models were used to allow for censoring of patients who died or were discharged prior to receiving revascularization. Thus, the dependant variable was receipt of revascularization during the index hospitalization, while the key independent variable was an indicator of patient race. Models were run using 3 different VOL. 101, NO. 11, NOVEMBER 2009 1133
Racial disparities in cardiac revascularization
definitions of revascularization: (1) receipt of PCI, (2) receipt of CABG, (3) receipt of PCI or CABG. Models were run with progressively increasing levels of adjustment: (1) unadjusted; (2) adjusted for patient demographics, comorbidity, and severity; and (3) adjusted for patient characteristics described above and accounting for clustering of patients within hospitals. Patient characteristics that have been identified as being associated with revascularization rates in prior studies and were independently related to utilization of revascularization in bivariate analyses (P < .05) were included in our models. In the risk adjustment model, age was expressed as 4 indicator variables (50-64, 65-69, 70-74, 75-79, 80-84, and ≥85 years), with a referent category of age 20 to 49 years for uninsured patients and those with private insurance and a referent category of age 65 to 69 for Medicare patients. Race was expressed as 2 indicator vari-
ables for patients identified as black and Hispanic, with white race serving as the referent. To insure the robustness of our findings, we repeated the analyses described above after excluding patients who were admitted with AMI after transfer from another acute care hospital. All p values are 2 tailed, with p values less than .05 deemed statistically significant. All statistical analyses were performed using SAS 9.1.3 (SAS Institute Inc, Cary, North Carolina). This project was approved by the University of Iowa Institutional review board.
Results
Our final sample included 430 509 AMI patients with Medicare, 238 956 with private insurance, and 74 926 patients who were uninsured/Medicaid who were admitted to 418 hospitals in 9 states. A significantly higher proportion of blacks and Hispanics were categorized as
Table 1. Characteristics of White, Black, and Hispanic Patients Admitted With Acute Myocardial Infarction Medicare Patients in State Inpatient Data Characteristics Demographics Age 20-49, (%) Age 50-64, (%) Age 65-69, (%) Age 70-74, (%) Age 75-79, (%) Age 80-84, (%) Age ≥85, (%) Sex, women, (%) Admission year 2000 2001 2002 2003 2004 2005 Admission source Emergency room, (%) Another acute care facility, (%) Other care facilities, (%) Other, (%) High-risk conditions Mechanical ventilation, (%) Intra-aortic balloon pump, (%) Unscheduled admission (%) Comorbidity COPD, (%) Diabetes (uncomplicated), (%) Diabetes (complicated), (%) Renal failure, (%) Obesity, (%) Depression, (%) Previous CABG, (%) Previous PCI, (%) Cardiac catheterization without PCI, (%)
White n = 370 909
Black n = 24 695
Hispanic n = 34 905
NA NA 57 350 (15.5) 73 555 (19.8) 83 839 (22.6) 76 500 (20.6) 79 665 (21.5) 175 331 (47.3)
NA NA 5755 (23.3) 5863 (23.7) 5268 (21.3) 3988 (16.2) 3821 (15.5) 14 504 (58.7)
NA NA 7274 (20.8) 8234 (23.6) 7994 (22.9) 5960 (17.1) 5443 (15.6) 16 938 (48.5)
63 875 65 116 66 560 67 389 66 002 41 969
(17.2) (17.6) (18.0) (18.2) (17.8) (11.3)
3909 4135 4422 4630 4986 2613
(15.8) (16.7) (17.9) (18.6) (20.2) (10.6)
5084 (14.6) 6125 (17.6) 6619 (19.0) 6954 (19.9) 6765 (19.4) 3358 (9.6)
229 205 (61.8) 77 946 (21.0) 12 290 (3.3) 51 468 (13.9)
17 232 (69.8) 3897 (15.8) 757 (3.1) 2809 (11.4)
23 960 (68.6) 5614 (16.1) 979 (2.8) 4352 (12.5)
30 127 (8.1) 19 047 (5.1) 342 368 (92.3)
2584 (10.5) 807 (3.3) 22 896 (92.7)
34 905 (9.2) 1843 (5.3) 30 820 (88.3)
86 804 (23.4) 85 769 (23.1) 15 229 (4.1) 29 403 (7.9) 13 922 (3.8) 13 820 (3.7) 33 224 (9.0) 26 611 (7.2) 103 437 (27.9)
5206 (21.1) 8285 (33.6) 2019 (8.2) 4213 (17.1) 1232 (5.0) 547 (2.2) 1417 (5.7) 1404 (5.7) 7007 (28.4)
7170 (20.5) 12 625 (36.2) 2661 (7.6) 4470 (12.8) 1334 (3.8) 1049 (3.0) 2867 (8.2) 2088 (6.0) 10 732 (30.8)
Abbreviations: CABG, coronary artery bypass grafting; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention.
1134 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 101, NO. 11, NOVEMBER 2009
Racial disparities in cardiac revascularization
uninsured/Medicaid as compared with whites (21.6% of all black patients vs 23.6% of Hispanics vs 7.6% of whites, P < .001). Black AMI patients were younger and significantly more likely to be women as compared to whites and Hispanics for each insurance cohort (Table 1). In addition, blacks and Hispanics were significantly more likely to be admitted from the emergency department, as compared to whites, and less likely to be admitted after a transfer from another acute care hospital in each of the 3 insurance cohorts. Blacks and Hispanics admitted with AMI had more diabetes and more renal failure and were also significantly less likely to have received a prior revascularization (PCI or CABG) procedure. In unadjusted analyses, a significantly lower proportion of black patients received revascularization (PCI or CABG) than whites or Hispanics (48.3% vs 57.8% vs 55.2%, P < .001) (Table 2). The lower rates of revascu-
larization for blacks as compared to whites or Hispanics were consistent across the 3 categories of insurance (Medicare, private insurance, and Medicaid/uninsured) and when PCI and CABG were examined separately. In Cox regression models, looking at patients with all types of insurance in aggregate, black patients hospitalized with AMI were approximately 19% less likely to receive PCI than whites, and Hispanics were approximately 5% less likely even after adjusting for patient characteristics, comorbidity, and clustering of patients within hospitals (Table 3). These results were fairly consistent when each of the 3 insurance cohorts was examined individually. Alternatively, when considering CABG (Table 4), the results were quite different. In particular, the magnitude of the disparity in revascularization for blacks was markedly larger for CABG than for PCI. The disparity was particularly striking for blacks with poor
Uninsured/Medicaid Patients in State Inpatient Data
Privately Insured Patients in State Inpatient Data White n = 204 369
Black n = 16 223
Hispanic n = 18 364
45 089 (22.1) 125 338 (62.3) 11 737 (5.7) 7801 (3.8) 6100 (3.0) 4386 (2.2) 3918 (1.9) 51 478 (25.2)
4845 (29.9) 9150 (56.4) 886 (5.5) 540 (3.3) 379 (2.3) 234 (1.4) 189 (1.2) 6241 (38.5)
4804 (26.2) 10 186 (55.5) 1185 (6.5) 891 (4.9) 610 (3.4) 368 (2.0) 310 (1.7) 4509 (24.6)
37 416 37 487 37 088 36 066 37 151 21 161
(18.3) (18.3) (18.2) (17.7) (17.2) (10.4)
2825 2734 3022 3002 2968 1672
White n = 47 241
Black n = 11 239
Hispanic n = 16 446
14 765 (31.3) 26 485 (56.1) 1527 (3.2) 1388 (2.9) 1286 (2.7) 880 (1.9) 910 (1.9) 15 400 (32.6)
4142 (36.8) 5883 (52.3) 419 (3.7) 280 (2.5) 217 (1.9) 156 (1.4) 142 (1.3) 5338 (47.5)
4239 (25.8) 8757 (53.3) 1011 (6.2) 868 (5.3) 690 (4.2) 447 (2.7) 434 (2.6) 6215 (37.8)
1685 1710 1980 2183 2334 1347
2289 2694 2961 3395 3431 1676
(17.4) (16.9) (18.6) (18.5) (18.3) (10.3)
3051 (16.6) 3335 (18.2) 3472 (18.9) 3511 (19.1) 3326 (18.1) 1669 (9.1)
106 712 (52.2) 58 484 (28.6) 7164 (3.5) 32 009 (15.7)
9536 (58.8) 3885 (24.0) 562 (3.5) 2240 (13.8)
11 103 (60.5) 3967 (21.6) 614 (3.3) 2680 (14.6)
27 377 (58.0) 11 825 (25.0) 1646 (3.5) 6393 (13.5)
7546 (67.1) 2118 (18.9) 335 (3.0) 1240 (11.0)
10 894 (66.2) 3231 (19.7) 480 (2.9) 1841 (11.2)
9574 (4.7) 12 389 (6.1) 184 597 (90.3)
856 (5.3) 663 (4.1) 14 693 (90.6)
802 (4.4) 1081 (5.9) 15 886 (86.5)
3227 (6.8) 3080 (6.5) 43 323 (91.7)
802 (7.1) 462 (4.1) 10 509 (93.5)
1090 (6.6) 1080 (6.6) 14 698 (89.4)
27 153 (13.3) 40 123 (19.6) 5933 (2.9) 5223 (2.6) 19 954 (9.8) 7108 (3.5) 9676 (4.7) 16 023 (7.8) 56 621 (27.7)
1836 (11.3) 4850 (29.9) 805 (5.0) 1075 (6.6) 1730 (10.7) 325 (2.0) 528 (3.3) 1223 (7.5) 5380 (33.2)
1851 (10.1) 5665 (30.9) 781 (4.3) 835 (4.6) 1549 (8.4) 393 (2.1) 809 (4.4) 1237 (6.7) 5914 (32.2)
9188 (19.5) 10 959 (23.2) 1809 (3.8) 1657 (3.5) 4825 (10.2) 2124 (4.5) 2198 (4.6) 3689 (7.8) 13 436 (28.4)
1825 (16.2) 3389 (30.2) 605 (5.4) 968 (8.6) 1135 (10.1) 367 (3.3) 348 (3.1) 713 (6.3) 3741 (33.3)
2096 (12.7) 5759 (35.0) 1084 (6.6) 1195 (7.3) 1138 (6.9) 445 (2.7) 611 (3.7) 867 (5.3) 5643 (34.3)
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
7386 7686 8440 9041 9226 5462
(15.6) (16.3) (17.9) (19.1) (19.5) (11.6)
(15.0) (15.2) (17.6) (19.4) (20.8) (12.0)
(13.9) (16.4) (18.0) (20.6) (20.9) (10.2)
VOL. 101, NO. 11, NOVEMBER 2009 1135
Racial disparities in cardiac revascularization
insurance who were approximately 39% less likely to receive CABG than similarly insured whites, but the disparity was also present for blacks with Medicare and with private insurance. For Hispanic patients the disparity in CABG rates, while present, was relatively small and was not markedly larger for poorly insured Hispanics when compared to those with better insurance (Table 4). In sensitivity analysis, results were similar after excluding patients who were admitted with AMI following transfer from another acute care hospital.
Discussion
In summary, black patients hospitalized with AMI were younger, more often female, and more often admitted through the emergency department as compared to whites and Hispanics. Blacks were less likely than whites and Hispanics to receive revascularization during their hospitalization, and disparities were particularly striking for underinsured blacks with regards to CABG. Our finding that blacks and Hispanics were less likely to receive revascularization than similarly insured whites after adjusting for patient characteristics and admitting
Table 2. Revascularization Rates for White, Black, and Hispanic Patients Hospitalized With Acute Myocardial Infarction Medicare (N = 370 909 Whites, 24 695 Blacks, And 34 905 Hispanics) PCI White Black Hispanic CABG White Black Hispanic PCI or CABG White Black Hispanic
Private Insurance Uninsured/Medicaid All Patients (N = 204 369 Whites, (N = 47 241 Whites, 11,239 (N = 622 519 Whites, 16 223 Blacks, And Blacks, and 52 157 Blacks, and 18 364 Hispanics) 16 446 Hispanics) 69 715 Hispanics)
128 769 (34.7%) 6815 (27.6%) 11 228 (32.2%)
122 341 (59.9%) 8327 (51.3%) 9945 (54.2%)
25 680 (54.4%) 4822 (42.9%) 7450 (45.3%)
276 790 (44.5%) 19 964 (38.3%) 28 623 (41.1%)
49 742(13.4%) 2409 (9.8%) 4756 (13.6%)
32 295 (15.8%) 2013 (12.4%) 3118 (17.0%)
7174 (15.2%) 1123 (10.0%) 2771 (16.8%)
89 211 (14.3%) 5545 (10.6%) 10 645 (15.3%)
175 574 (47.3%) 9106 (36.9%) 15 689 (44.9%)
151 729 (74.2%) 10 202 (62.9%) 12 797 (69.7%)
32 301 (68.4%) 5872 (52.2%) 10 025 (61.0%)
359 604 (57.8%) 25 180 (48.3%) 38 511 (55.2%)
Abbreviations: CABG, coronary artery bypass grafting; PCI, percutaneous coronary intervention.
Table 3. Hazard Rates for Receipt of Percutaneous Coronary Intervention for Blacks and Hispanics With Varying Insurance as Compared to Whites Medicare
Unadjusted
Hazard Ratios (95% CI) Black 0.697 (0.680-0.714) Hispanic 0.896 (0.878-0.913) Black 0.766 (0.747-0.785)
Private Insurance Uninsured/Medicaid
Hazard Hazard P Ratios Ratios (95% CI) Value (95% CI) 0.705 <.0001 0.628 (0.689-0.720) (0.609-0.648) <.0001 0.840 <.0001 0.742 (0.823-0.858) (0.723-0.762) <.0001 0.839 <.0001 0.806 (0.820-0.858) (0.781-0.833) P Value <.0001
All Patients
Hazard Ratios (95% CI) 0.754 (0.743-0.765) <.0001 0.889 (0.878-0.900) <.0001 0.804 (0.792-0.816) P Value <.0001
P Value <.0001 <.0001
Adjusted <.0001 for patient characteristics, comorbidity, Hispanic 0.947 <.0001 0.917 <.0001 0.866 <.0001 0.911 <.0001 and AMI (0.929-0.966) (0.898-0.937) (0.843-0.889) (0.900-0.923) location Adjusted Black 0.771 <.0001 0.859 <.0001 0.827 <.0001 0.814 <.0001 for patient (0.751-0.791) (0.839-0.880) (0.798-0.856) (0.802-0.827) characteristics, comorbidity AMI location Hispanic 0.969 <0.01 0.937 <.0001 0.965 0.028 0.949 <.0001 stratified by (0.947-0.992) (0.915-0.959) (0.936-0.996) (0.935-0.963) hospital IDs Abbreviation: AMI, acute myocardial infarction; CI, confidence interval.
1136 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 101, NO. 11, NOVEMBER 2009
Racial disparities in cardiac revascularization
hospital provides strong evidence that disparities are not related to differences in insurance, comorbidity, or differences in the hospitals where different racial groups tend to receive care.
A number of our findings merit further discussion. Our finding that black AMI patients were less likely than whites or Hispanics to receive revascularization during their hospitalization both reinforces and extends prior
Table 4. Hazard Rates for Receipt of Coronary Artery Bypass Grafting for Blacks and Hispanics With Varying Insurance as Compared to Whites Medicare Hazard Ratios (95% CI) Black 0.589 (0.565-0.613) Hispanic 0.954 (0.926-0.982) Black 0.635 (0.609-0.662)
Private Insurance
Hazard P Ratios Value (95% CI) <.0001 0.562 (0.537-0.588) <.01 0.906 (0.873-0.940) <.0001 0.634 (0.606-0.664)
Uninsured/ Medicaid
All Patients
Hazard Ratios (95% CI) 0.458 (0.430-0.488) 0.853 (0.816-0.891) 0.558 (0.522-0.595)
Hazard P P P Ratios Value Value (95% CI) Value Unadjusted <.0001 <.0001 0.606 (0.589-0.622)<.0001 <.0001 <.0001 0.979 (0.960-0.999) 0.042 Adjusted <.0001 <.0001 0.615 for patient (0.598-0.633) characteristics, <.0001 comorbidity, Hispanic 0.883 <.0001 0.911 <.0001 0.888 <.0001 0.893 and AMI (0.857-0.911 (0.878-0.946) (0.848-0.930) (0.874-0.911) location <.0001 Adjusted Black 0.664 <.0001 0.660 <.0001 0.611 <.0001 0.646 for patient (0.636-0.694) (0.629-0.692) (0.569-0.655) (0.627-0.665) characteristics, <.0001 comorbidity, AMI location Hispanic 0.938 <.001 0.957 0.042 1.054 0.064 0.964 <.01 stratified by (0.905-0.979) (0.917-0.998) (0.997-1.115) (0.941-0.988) hospital IDs Abbreviation: AMI, acute myocardial infarction; CI, confidence interval.
Table 5. Hazard Rates for Receipt of Revascularization (Either Percutaneous Coronary Intervention or Coronary Artery Bypass Grafting) for Blacks and Hispanics as Compared to Whites Medicare
Unadjusted
Hazard Ratios (95% CI) Black 0.688 (0.673-0.702) Hispanic 0.910 (0.895-0.925) Black 0.725 (0.710-0.741)
Private Insurance
Hazard Ratios (95% CI) 0.710 (0.696-0.725) <.0001 0.856 (0.841-0.872) 0.792 <.0001 (0.775-0.808) P Value <.0001
Uninsured/ Medicaid
All Patients
Hazard Hazard P Ratios Ratios (95% CI) Value (95% CI) 0.632 <.0001 0.745 (0.615-0.650) (0.736-0.755) <.0001 0.768 <.0001 0.905 (0.751-0.785) (0.895-0.914) 0.742 0.754 <.0001 (0.720-0.764) <.0001 (0.744-0.764) P Value <.0001
P Value <.0001 <.0001
Adjusted <.0001 for patient characteristics, comorbidity, Hispanic 0.933 0.912 0.864 0.905 <.0001 and AMI (0.917-0.949) <.0001 (0.896-0.929) <.0001 (0.844-0.884) <.0001 (0.895-0.915) location Adjusted Black 0.742 <.0001 0.819 <.0001 0.772 <.0001 0.773 <.0001 for patient (0.726-0.759) (0.802-0.837) (0.748-0.796) (0.763-0.784) characteristics, comorbidity, AMI location Hispanic 0.969 <.01 0.943 <.0001 0.967 0.016 0.950 <.0001 stratified by (0.950-0.988) (0.924-0.962) (0.941-0.994) (0.938-0.962) hospital IDs Abbreviation: AMI, acute myocardial infarction; CI, confidence interval.
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 101, NO. 11, NOVEMBER 2009 1137
Racial disparities in cardiac revascularization
research. While an array of prior studies have found evidence of disparities in revascularization, most of these studies have been limited to patients with a single type of health insurance (eg, Medicare),4,5,16, patients receiving care in a single medical center,9 or patients enrolled in clinical registries.18 Our finding of significant disparity among black patients relative to whites, regardless of insurance coverage, confirms results of prior studies based on single types of health insurance. In addition, the majority of studies investigating disparities in cardiovascular disease focus only on black populations. There is less evidence documenting disparities for Hispanic populations, although some studies suggest relatively similar rates of revascularization for Hispanic and whites.19-21 Likewise, our study found that the magnitude of disparity for Hispanics was relatively small, although statistically significant. In general, our findings were remarkably consistent across procedure type (PCI vs CABG) and insurance status. Alternatively, a deeper look at the details of our results reveals some important and interesting differences. While blacks were approximately 19% less likely to receive PCI as compared to whites for all insurance cohorts, blacks were approximately 35% less likely to receive CABG. Our finding that disparity in revascularization is larger for CABG than for PCI has been observed in some but not all prior studies.4,9,16,18 Moreover, the magnitude of the disparities that we observed for CABG and PCI were very similar to those observed in a 1997 study by Peterson et al using data collected between 1984 and 1992. The lack of change in the magnitude of disparity raises concerns about the ability of ongoing efforts to improve health services equity in producing measurable reductions in racial disparities.22,23 Finally, we found that the magnitude of disparity for black patients in the use of PCI was relatively consistent across insurance categories, but the disparity in the use of CABG was substantially smaller for black patients with Medicare or private insurance, compared to patients with Medicaid/uninsured. While it appears that insurance type does, at least partially, impact disparities for CABG, significant disparity nevertheless remained, regardless of insurance type. These findings are consistent with other studies suggesting that the removal of financial barriers will not eliminate disparities in utilization of health services. Our finding that a higher proportion of black and Hispanic AMI patients were admitted from the emergency department as compared to whites is also interesting. One potential explanation for this finding is that blacks and Hispanics were less likely to be admitted as transfers from other acute care hospitals and as a consequence, and thus, by default, a higher proportion of blacks and Hispanics were admitted from the emergency department. Such a mechanism is plausible given evidence from prior studies that black AMI patients are less 1138 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
likely to be transferred than whites.16 Alternatively, it is possible that the greater proportion of blacks and Hispanics admitted through the emergency department is related to a lack of access to primary care for these populations, differences in patient preference, or provider decision making.24 There are a number of limitations to our study that merit brief mention. First, our study relied upon administrative data for our risk adjustment models and thus may have been subject to bias if black, white, and Hispanic patients differed in severity in ways that were not captured. Nevertheless, a growing array of studies has now shown that administrative data generally offer similar results to clinical data sources. Second, our analyses were somewhat limited by the idiosyncrasies of the state inpatient data. For example, the state inpatient data do not assign unique patient identifiers, making it impossible to track patients who are transferred between hospitals. However, it seems unlikely that having unique identifiers would have substantially changed our results. In summary, in an analysis of all-payor data from 9 states we found significant disparities in revascularization for black, white, and Hispanic AMI patients with similar health insurance. These findings have important implications. First, conclusions about racial disparities based on specific types of insurance (eg, Medicaid, Medicare) may differ modestly across insurance types, although in general we found little bias in our estimation of the magnitude of disparity depending on insurance, with the exception of CABG. Possibly, the use of invasive, expensive procedures with significant follow-up is more sensitive to insurance coverage. Second, the fact that significant disparity persisted regardless of insurance type suggests that elimination of financial barriers is not likely to eliminate disparities. Finally, our results call into question the efficacy of various efforts to reduce racial disparities.
References
1. Gillum RF. Coronary artery bypass surgery and coronary angiography in the United States, 1979-1983. Am Heart J. 1987;113(5):1255-1260. 2. Goldberg KC, Hartz AJ, Jacobsen SJ, Krakauer H, Rimm AA. Racial and community factors influencing coronary artery bypass graft surgery rates for all 1986 Medicare patients. JAMA. 1992;267(11):1473-1477. 3. Gauri AJ, Davis A, Hong T, Burke MC, Knight BP. Disparities in the use of primary prevention and defibrillator therapy among blacks and women. Am J Med. 2006;119(2):167e17-21. 4. Ayanian JZ, Udvarhelyi IS, Gatsonis CA, Pashos CL, Epstein AM. Racial differences in the use of revascularization procedures after coronary angiography. JAMA. 1993;269(20):2642-2646. 5. Whittle J, Conigliaro J, Good CB, Lofgren RP. Racial differences in the use of invasive cardiovascular procedures in the Department of Veterans Affairs medical system. N Engl J Med. 1993;329(9):621-627. 6. Mukamel DB, Weimer DL, Buchmueller TC, Ladd H, Mushlin AI. Changes in racial disparities in access to coronary artery bypass grafting surgery between the late 1990s and early 2000s. Med Care. 2007;45(7):664-671. 7. Jha AK, Fisher ES, Li Z, Orav EJ, Epstein AM. Racial trends in the use of major procedures among the elderly. N Engl J Med. 2005;353(7):683-691. 8. Schneider EC, Leape LL, Weissman JS, Piana RN, Gatsonis C, Epstein AM. Racial differences in cardiac revascularization rates: does “over-
VOL. 101, NO. 11, NOVEMBER 2009
Racial disparities in cardiac revascularization use” explain higher rates among white patients? Ann Intern Med. 2001; 135(5):328-337. 9. Peterson ED, Shaw LK, DeLong ER, Pryor DB, Califf RM, Mark DB. Racial variation in the use of coronary-revascularization procedures. Are the differences real? Do they matter? N Engl J Med. 1997;336(7):480-486. 10. Carlisle DM, Leake BD, Shapiro MF. Racial and ethnic differences in the use of invasive cardiac procedures among cardiac patients in Los Angeles County, 1986 through 1988. Am J Public Health. 1995;85(3):352-356. 11. HCUP SID overview. http://www.hcup-us.ahrq.gov/sidoverview.jsp Accessed April 21, 2009. 12. Cram P, Pham HH, Bayman L, Vaughan-Sarrazin MS. Insurance status of patients admitted to specialty cardiac and competing general hospitals: Are accusations of cherry picking justified? Med Care. 2008;46(5):467-475. 13. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8-27. 14. Southern DA, Quan H, Ghali WA. Comparison of the Elixhauser and Charlson/Deyo methods of comorbidity measurement in administrative data. Med Care. 2004;42(4):355-360. 15. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130-1139. 16. Popescu I, Vaughan-Sarrazin MS, Rosenthal GE. Differences in mortality and use of revascularization in black and white patients with acute MI admitted to hospitals with and without revascularization services. JAMA. 2007;297(22):2489-2495. 17. Rosenthal GE, Vaughan SM, Hannan EL. In-hospital mortality following coronary artery bypass graft surgery in Veterans Health Administration and private sector hospitals. Med Care. 2003;41(4):522-535.
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
18. Sonel AF, Good CB, Mulgund J, et al. Racial variations in treatment and outcomes of black and white patients with high-risk non-ST-elevation acute coronary syndromes: insights from CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of the ACC/AHA Guidelines?). Circulation. 2005;111(10):12251232. 19. Barnhart JM, Fang J, Alderman MH. Differential use of coronary revascularization and hospital mortality following acute myocardial infarction. Arch Intern Med. 2003;163(4):461-466. 20. Bertoni AG, Goonan KL, Bonds DE, Whitt MC, Goff DC, Jr., Brancati FL. Racial and ethnic disparities in cardiac catheterization for acute myocardial infarction in the United States, 1995-2001. J Natl Med Assoc. 2005;97(3):317-323. 21. Yarzebski J, Bujor CF, Lessard D, Gore JM, Goldberg RJ. Recent and temporal trends (1975-1999) in the treatment, hospital, and long-term outcomes of Hispanic and non-Hispanic white patients hospitalized with acute myocardial infarction: a population-based perspective. Am Heart J. 2004;147(4):690-697. 22. Brown CP, Ross L, Lopez I, Thornton A, Kiros GE. Disparities in the receipt of cardiac revascularization procedures between blacks and whites: an analysis of secular trends. Ethn Dis. 2008;18(2 suppl 2):S2-112-117. 23. Lucas FL, DeLorenzo MA, Siewers AE, Wennberg DE. Temporal trends in the utilization of diagnostic testing and treatments for cardiovascular disease in the United States, 1993-2001. Circulation. 2006;113(3):374-379. 24. Cunningham PJ. What accounts for differences in the use of hospital emergency departments across US communities? Health Aff. 2006;25(5):w324-w336. n
VOL. 101, NO. 11, NOVEMBER 2009 1139
r
i
g
i
n
a
l
c
o
m
m
u
n
i
c
a
t
i
o
n
Racial Disparities in Revascularization Rates Among Patients With Similar Insurance Coverage Peter Cram, MD, MBA; Levent Bayman, MS; Ioana Popescu, MD, MPH; Mary S. Vaughan-Sarrazin, PhD
Disclosure: Dr Vaughan-Sarrazin is a research scientist in the Center for Research in the Implementation of Innovative Strategies in Practice at the Iowa City Veterans Administration Medical Center, which is funded through the Department of Veterans Affairs, Veterans Health Administration, Health Services Research, and Development Service. J Natl Med Assoc. 2009;101:xxx–xxx Funding/Support: Dr Cram is supported by a K23 career development award (RR01997201) from the National Center for Research Resources at the National Institutes of Health and the Robert Wood Johnson Physician Faculty Scholars Program. This work is also funded by R01 HL085347-01A1 from National Heart, Lung, and Blood Institute at the NIH. The funding sources had no role in the analyses or drafting of this manuscript. Disclaimer: The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. Background: Racial disparities in coronary revascularization—percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG)—have been extensively documented. However, it is unclear whether disparities are consistent among patients with similar health insurance coverage. Our objective was to assess racial disparities in coronary revascularization among white, black, and Hispanic patients with similar insurance coverage hospitalized with acute myocardial infarction (AMI). Methods: We used 2000-2005 state inpatient data for 9 states to identify white, black, and Hispanic patients hospitalized with AMI. Patients were grouped into 3 health insurance cohorts: (1) Medicare, (2) private insurance, and (3) Medicaid/uninsured. We examined use of revascularization (PCI or CABG) among blacks and Hispanics as compared to whites in each of the 3 insurance cohorts. Author Affiliations: Results: The 418 study hospitals admitted 430 509 AMI patients Corresponding Author:238 956 with private insurance, and 74 926 with Medicare,
patients who were uninsured/Medicaid. In unadjusted analyses, black and Hispanic patients were significantly less likely to receive in-hospital revascularization among the Medicare cohort (38.9% vs 44.9% vs 47.3%, P < .001), privately insured cohort (62.9% vs 69.7% vs 74.2%, P < .001), and uninsured/Medicaid cohort (55.2% vs 61.0% vs 68.4%, P < .001). In Cox models
1132 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
adjusting for patient demographics, comorbidity, and clustering of patients within hospitals, blacks were approximately 25% less likely and Hispanics 5% less likely to receive revascularization as compared to whites with similar insurance. Conclusions: Blacks hospitalized with AMI are significantly less likely to receive revascularization when compared to whites and Hispanics with similar health insurance. Our data suggest that patients’ ability to pay for costly procedures is unlikely to explain racial disparities. Keywords: racial disparities n health insurance n heart J Natl Med Assoc. 2009;101:1132-1139 Author Affiliations: Division of General Internal Medicine, Department of Internal Medicine, University of Iowa Carver College of Medicine (Drs Cram, Bayman, Popescu, and Vaughan-Sarrazin); and Center for Research in the Implementation of Innovative Strategies for Practice (CRIISP), Iowa City Veterans Administration Medical Center (Drs Cram and Vaughan-Sarrazin), Iowa City, Iowa. Corresponding Author: Peter Cram, MD, MBA, Division of General Medicine, University of Iowa Carver College of Medicine, Iowa City VA Medical Center, Mail Stop 152, Iowa City, IA 52242 ([email protected]).
Background
R
acial disparities in cardiac care are well studied.1,2 Decades of research have demonstrated that compared to whites, blacks are less likely to receive an array of high-cost cardiac interventions, including implantable cardioverter defibrillators,3 and coronary revascularization procedures such as percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG).4,5 Moreover, racial disparities in cardiac care have persisted despite significant efforts to insure equal access among all patient populations.6,7 While the existence of disparities in coronary revascularization is well established, the relationship between health insurance and racial disparities is less clear. More specifically, it is uncertain whether racial disparities in revascularization are similar in magnitude among patients with similar insurance coverage. Many studies of disparities in coronary revascularization rates have focused on patients covered under a single insurance plan such as Medicare2,4,8 or the Veterans Administration.5 Far fewer studies have examined disparities in revascuVOL. 101, NO. 11, NOVEMBER 2009
Racial disparities in cardiac revascularization
larization rates among patients with private insurance or underinsured populations.9,10 Moreover, to the best of our knowledge, prior studies have not systematically investigated whether racial disparities in revascularization rates are consistent among white, black, and Hispanic patients with similar types of insurance. The objective of the current study was to examine rates of in-hospital revascularization (PCI and CABG) for white, black, and Hispanic patients with similar insurance coverage (or lack thereof) hospitalized with acute myocardial infarction (AMI). Our a priori hypothesis was that racial disparities in revascularization would be smaller among whites, blacks, and Hispanics with private insurance or Medicare and larger among patients who were uninsured.
Methods Data We used state inpatient data for the years 2000-2005 from Arizona, California, Florida, Massachusetts, Maryland, New Jersey, and Wisconsin; and 2000-2004 from New York and Texas to identify all patients age ≥20 years hospitalized with AMI on the basis of International Classification of Diseases, Ninth Clinical Modification (ICD9-CM) code 410.X (N = 1 623 956). The state inpatient databases were developed by the Agency for Healthcare Research and Quality (AHRQ) as part of the Health Care Utilization Project.11 The state inpatient data include many elements available on the UB-92 hospital discharge abstract, including patient demographics (eg, race, sex); admitting hospital; primary and secondary diagnoses and procedures, as captured by ICD-9-CM codes; the diagnosis-related group (DRG); admission source (eg, emergency department, transfer from another hospital); admission and discharge dates; patient’s primary insurance (categorized as Medicare, private insurance, Medicaid, self-pay, other); type of insurance (fee for service or health maintenance organization); and disposition at the time of hospital discharge (eg, transfer to another acute care hospital, deceased). The state inpatient data for many states are processed and sold by the AHRQ central distributor and thus have uniform coding, but data for other states are processed and sold by individual state planning agencies and thus have variable coding. For example, many states code age as a continuous variable, but others provide age as a categorical variable or choose to mask age of individual patients treated in low-volume hospitals to protect patient confidentiality; thus, merging data from multiple states, while entirely feasible, does require significant finesse such as converting age of all patients into categorical variables. In addition, the state inpatient data do not contain unique patient identifiers and do not allow for tracking patients after discharge or transfer from the admitting hospital. As our primary objective was to assess racial disparities in revascularization, we limited our study to patients whose JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
race was listed as white, black, or Hispanic and excluded patients with race listed as Asian, Native American, missing, or “other” because the number of patients in each category was too small to allow for meaningful comparisons to be made (N = 247 311). We also excluded patients whose insurance status was missing (N = 210 735) and patients admitted to hospitals that did not perform revascularization (1290 hospitals that admitted 630 101 AMI patients).
Statistical Analyses After identifying all white, black, and Hispanic patients hospitalized with AMI, patients were stratified into 3 different cohorts based upon their insurance coverage using a scheme we developed previously.12 In particular, AMI patients were categorized as having either: (1) Medicare (fee for service or managed care); (2) private insurance (fee for service or managed care); and 3) uninsured/Medicaid (uninsured, self-pay, or Medicaid). All analyses described below were conducted separately for each of the 3 insurance cohorts. As the primary objective of the study was to assess in-hospital revascularization rates for white, black, and Hispanic patients admitted with AMI, the first step in our analyses was to identify patients who had received revascularization. We defined revascularization as receipt of either PCI (ICD-9 procedure codes 36013607) or CABG (ICD-9 codes 3610-3619) during the AMI admission. We used bivariate methods to compare the demographic characteristics, comorbidity, and clinical risk factors of white, black, and Hispanic patients within each of the 3 insurance cohorts. Comorbid conditions were identified using algorithms developed by Elixhauser and Deyo13,14 and updated by Quan et al.15 Additional clinical risk factors were identified using coding schemes that have been used in prior studies assessing cardiovascular outcomes using administrative data. This included AMI location (coded as anterlateral [ICD-9 codes 410.0, 410.1, and 410.5], inferposterior [410.2, 410.3, 410.4, and 410.6], subendocardial [410.7] and other [410.8 and 410.9], previous CABG surgery (ICD-9 code V4581), previous PCI (ICD-9 code V4582), use of intra-aortic balloon pump during the admission (procedure codes 3761 and 3762), use of mechanical ventilation during the admission, and whether the patient received revascularization (PCI or CABG) during admission.16,17 We used multivariable survival models to estimate the hazard ratio of receiving revascularization for blacks and Hispanics as compared to whites for patients in each of the 3 insurance cohorts; Cox proportional hazards regression models were used to allow for censoring of patients who died or were discharged prior to receiving revascularization. Thus, the dependant variable was receipt of revascularization during the index hospitalization, while the key independent variable was an indicator of patient race. Models were run using 3 different VOL. 101, NO. 11, NOVEMBER 2009 1133
Racial disparities in cardiac revascularization
definitions of revascularization: (1) receipt of PCI, (2) receipt of CABG, (3) receipt of PCI or CABG. Models were run with progressively increasing levels of adjustment: (1) unadjusted; (2) adjusted for patient demographics, comorbidity, and severity; and (3) adjusted for patient characteristics described above and accounting for clustering of patients within hospitals. Patient characteristics that have been identified as being associated with revascularization rates in prior studies and were independently related to utilization of revascularization in bivariate analyses (P < .05) were included in our models. In the risk adjustment model, age was expressed as 4 indicator variables (50-64, 65-69, 70-74, 75-79, 80-84, and ≥85 years), with a referent category of age 20 to 49 years for uninsured patients and those with private insurance and a referent category of age 65 to 69 for Medicare patients. Race was expressed as 2 indicator vari-
ables for patients identified as black and Hispanic, with white race serving as the referent. To insure the robustness of our findings, we repeated the analyses described above after excluding patients who were admitted with AMI after transfer from another acute care hospital. All p values are 2 tailed, with p values less than .05 deemed statistically significant. All statistical analyses were performed using SAS 9.1.3 (SAS Institute Inc, Cary, North Carolina). This project was approved by the University of Iowa Institutional review board.
Results
Our final sample included 430 509 AMI patients with Medicare, 238 956 with private insurance, and 74 926 patients who were uninsured/Medicaid who were admitted to 418 hospitals in 9 states. A significantly higher proportion of blacks and Hispanics were categorized as
Table 1. Characteristics of White, Black, and Hispanic Patients Admitted With Acute Myocardial Infarction Medicare Patients in State Inpatient Data Characteristics Demographics Age 20-49, (%) Age 50-64, (%) Age 65-69, (%) Age 70-74, (%) Age 75-79, (%) Age 80-84, (%) Age ≥85, (%) Sex, women, (%) Admission year 2000 2001 2002 2003 2004 2005 Admission source Emergency room, (%) Another acute care facility, (%) Other care facilities, (%) Other, (%) High-risk conditions Mechanical ventilation, (%) Intra-aortic balloon pump, (%) Unscheduled admission (%) Comorbidity COPD, (%) Diabetes (uncomplicated), (%) Diabetes (complicated), (%) Renal failure, (%) Obesity, (%) Depression, (%) Previous CABG, (%) Previous PCI, (%) Cardiac catheterization without PCI, (%)
White n = 370 909
Black n = 24 695
Hispanic n = 34 905
NA NA 57 350 (15.5) 73 555 (19.8) 83 839 (22.6) 76 500 (20.6) 79 665 (21.5) 175 331 (47.3)
NA NA 5755 (23.3) 5863 (23.7) 5268 (21.3) 3988 (16.2) 3821 (15.5) 14 504 (58.7)
NA NA 7274 (20.8) 8234 (23.6) 7994 (22.9) 5960 (17.1) 5443 (15.6) 16 938 (48.5)
63 875 65 116 66 560 67 389 66 002 41 969
(17.2) (17.6) (18.0) (18.2) (17.8) (11.3)
3909 4135 4422 4630 4986 2613
(15.8) (16.7) (17.9) (18.6) (20.2) (10.6)
5084 (14.6) 6125 (17.6) 6619 (19.0) 6954 (19.9) 6765 (19.4) 3358 (9.6)
229 205 (61.8) 77 946 (21.0) 12 290 (3.3) 51 468 (13.9)
17 232 (69.8) 3897 (15.8) 757 (3.1) 2809 (11.4)
23 960 (68.6) 5614 (16.1) 979 (2.8) 4352 (12.5)
30 127 (8.1) 19 047 (5.1) 342 368 (92.3)
2584 (10.5) 807 (3.3) 22 896 (92.7)
34 905 (9.2) 1843 (5.3) 30 820 (88.3)
86 804 (23.4) 85 769 (23.1) 15 229 (4.1) 29 403 (7.9) 13 922 (3.8) 13 820 (3.7) 33 224 (9.0) 26 611 (7.2) 103 437 (27.9)
5206 (21.1) 8285 (33.6) 2019 (8.2) 4213 (17.1) 1232 (5.0) 547 (2.2) 1417 (5.7) 1404 (5.7) 7007 (28.4)
7170 (20.5) 12 625 (36.2) 2661 (7.6) 4470 (12.8) 1334 (3.8) 1049 (3.0) 2867 (8.2) 2088 (6.0) 10 732 (30.8)
Abbreviations: CABG, coronary artery bypass grafting; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention.
1134 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 101, NO. 11, NOVEMBER 2009
Racial disparities in cardiac revascularization
uninsured/Medicaid as compared with whites (21.6% of all black patients vs 23.6% of Hispanics vs 7.6% of whites, P < .001). Black AMI patients were younger and significantly more likely to be women as compared to whites and Hispanics for each insurance cohort (Table 1). In addition, blacks and Hispanics were significantly more likely to be admitted from the emergency department, as compared to whites, and less likely to be admitted after a transfer from another acute care hospital in each of the 3 insurance cohorts. Blacks and Hispanics admitted with AMI had more diabetes and more renal failure and were also significantly less likely to have received a prior revascularization (PCI or CABG) procedure. In unadjusted analyses, a significantly lower proportion of black patients received revascularization (PCI or CABG) than whites or Hispanics (48.3% vs 57.8% vs 55.2%, P < .001) (Table 2). The lower rates of revascu-
larization for blacks as compared to whites or Hispanics were consistent across the 3 categories of insurance (Medicare, private insurance, and Medicaid/uninsured) and when PCI and CABG were examined separately. In Cox regression models, looking at patients with all types of insurance in aggregate, black patients hospitalized with AMI were approximately 19% less likely to receive PCI than whites, and Hispanics were approximately 5% less likely even after adjusting for patient characteristics, comorbidity, and clustering of patients within hospitals (Table 3). These results were fairly consistent when each of the 3 insurance cohorts was examined individually. Alternatively, when considering CABG (Table 4), the results were quite different. In particular, the magnitude of the disparity in revascularization for blacks was markedly larger for CABG than for PCI. The disparity was particularly striking for blacks with poor
Uninsured/Medicaid Patients in State Inpatient Data
Privately Insured Patients in State Inpatient Data White n = 204 369
Black n = 16 223
Hispanic n = 18 364
45 089 (22.1) 125 338 (62.3) 11 737 (5.7) 7801 (3.8) 6100 (3.0) 4386 (2.2) 3918 (1.9) 51 478 (25.2)
4845 (29.9) 9150 (56.4) 886 (5.5) 540 (3.3) 379 (2.3) 234 (1.4) 189 (1.2) 6241 (38.5)
4804 (26.2) 10 186 (55.5) 1185 (6.5) 891 (4.9) 610 (3.4) 368 (2.0) 310 (1.7) 4509 (24.6)
37 416 37 487 37 088 36 066 37 151 21 161
(18.3) (18.3) (18.2) (17.7) (17.2) (10.4)
2825 2734 3022 3002 2968 1672
White n = 47 241
Black n = 11 239
Hispanic n = 16 446
14 765 (31.3) 26 485 (56.1) 1527 (3.2) 1388 (2.9) 1286 (2.7) 880 (1.9) 910 (1.9) 15 400 (32.6)
4142 (36.8) 5883 (52.3) 419 (3.7) 280 (2.5) 217 (1.9) 156 (1.4) 142 (1.3) 5338 (47.5)
4239 (25.8) 8757 (53.3) 1011 (6.2) 868 (5.3) 690 (4.2) 447 (2.7) 434 (2.6) 6215 (37.8)
1685 1710 1980 2183 2334 1347
2289 2694 2961 3395 3431 1676
(17.4) (16.9) (18.6) (18.5) (18.3) (10.3)
3051 (16.6) 3335 (18.2) 3472 (18.9) 3511 (19.1) 3326 (18.1) 1669 (9.1)
106 712 (52.2) 58 484 (28.6) 7164 (3.5) 32 009 (15.7)
9536 (58.8) 3885 (24.0) 562 (3.5) 2240 (13.8)
11 103 (60.5) 3967 (21.6) 614 (3.3) 2680 (14.6)
27 377 (58.0) 11 825 (25.0) 1646 (3.5) 6393 (13.5)
7546 (67.1) 2118 (18.9) 335 (3.0) 1240 (11.0)
10 894 (66.2) 3231 (19.7) 480 (2.9) 1841 (11.2)
9574 (4.7) 12 389 (6.1) 184 597 (90.3)
856 (5.3) 663 (4.1) 14 693 (90.6)
802 (4.4) 1081 (5.9) 15 886 (86.5)
3227 (6.8) 3080 (6.5) 43 323 (91.7)
802 (7.1) 462 (4.1) 10 509 (93.5)
1090 (6.6) 1080 (6.6) 14 698 (89.4)
27 153 (13.3) 40 123 (19.6) 5933 (2.9) 5223 (2.6) 19 954 (9.8) 7108 (3.5) 9676 (4.7) 16 023 (7.8) 56 621 (27.7)
1836 (11.3) 4850 (29.9) 805 (5.0) 1075 (6.6) 1730 (10.7) 325 (2.0) 528 (3.3) 1223 (7.5) 5380 (33.2)
1851 (10.1) 5665 (30.9) 781 (4.3) 835 (4.6) 1549 (8.4) 393 (2.1) 809 (4.4) 1237 (6.7) 5914 (32.2)
9188 (19.5) 10 959 (23.2) 1809 (3.8) 1657 (3.5) 4825 (10.2) 2124 (4.5) 2198 (4.6) 3689 (7.8) 13 436 (28.4)
1825 (16.2) 3389 (30.2) 605 (5.4) 968 (8.6) 1135 (10.1) 367 (3.3) 348 (3.1) 713 (6.3) 3741 (33.3)
2096 (12.7) 5759 (35.0) 1084 (6.6) 1195 (7.3) 1138 (6.9) 445 (2.7) 611 (3.7) 867 (5.3) 5643 (34.3)
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
7386 7686 8440 9041 9226 5462
(15.6) (16.3) (17.9) (19.1) (19.5) (11.6)
(15.0) (15.2) (17.6) (19.4) (20.8) (12.0)
(13.9) (16.4) (18.0) (20.6) (20.9) (10.2)
VOL. 101, NO. 11, NOVEMBER 2009 1135
Racial disparities in cardiac revascularization
insurance who were approximately 39% less likely to receive CABG than similarly insured whites, but the disparity was also present for blacks with Medicare and with private insurance. For Hispanic patients the disparity in CABG rates, while present, was relatively small and was not markedly larger for poorly insured Hispanics when compared to those with better insurance (Table 4). In sensitivity analysis, results were similar after excluding patients who were admitted with AMI following transfer from another acute care hospital.
Discussion
In summary, black patients hospitalized with AMI were younger, more often female, and more often admitted through the emergency department as compared to whites and Hispanics. Blacks were less likely than whites and Hispanics to receive revascularization during their hospitalization, and disparities were particularly striking for underinsured blacks with regards to CABG. Our finding that blacks and Hispanics were less likely to receive revascularization than similarly insured whites after adjusting for patient characteristics and admitting
Table 2. Revascularization Rates for White, Black, and Hispanic Patients Hospitalized With Acute Myocardial Infarction Medicare (N = 370 909 Whites, 24 695 Blacks, And 34 905 Hispanics) PCI White Black Hispanic CABG White Black Hispanic PCI or CABG White Black Hispanic
Private Insurance Uninsured/Medicaid All Patients (N = 204 369 Whites, (N = 47 241 Whites, 11,239 (N = 622 519 Whites, 16 223 Blacks, And Blacks, and 52 157 Blacks, and 18 364 Hispanics) 16 446 Hispanics) 69 715 Hispanics)
128 769 (34.7%) 6815 (27.6%) 11 228 (32.2%)
122 341 (59.9%) 8327 (51.3%) 9945 (54.2%)
25 680 (54.4%) 4822 (42.9%) 7450 (45.3%)
276 790 (44.5%) 19 964 (38.3%) 28 623 (41.1%)
49 742(13.4%) 2409 (9.8%) 4756 (13.6%)
32 295 (15.8%) 2013 (12.4%) 3118 (17.0%)
7174 (15.2%) 1123 (10.0%) 2771 (16.8%)
89 211 (14.3%) 5545 (10.6%) 10 645 (15.3%)
175 574 (47.3%) 9106 (36.9%) 15 689 (44.9%)
151 729 (74.2%) 10 202 (62.9%) 12 797 (69.7%)
32 301 (68.4%) 5872 (52.2%) 10 025 (61.0%)
359 604 (57.8%) 25 180 (48.3%) 38 511 (55.2%)
Abbreviations: CABG, coronary artery bypass grafting; PCI, percutaneous coronary intervention.
Table 3. Hazard Rates for Receipt of Percutaneous Coronary Intervention for Blacks and Hispanics With Varying Insurance as Compared to Whites Medicare
Unadjusted
Hazard Ratios (95% CI) Black 0.697 (0.680-0.714) Hispanic 0.896 (0.878-0.913) Black 0.766 (0.747-0.785)
Private Insurance Uninsured/Medicaid
Hazard Hazard P Ratios Ratios (95% CI) Value (95% CI) 0.705 <.0001 0.628 (0.689-0.720) (0.609-0.648) <.0001 0.840 <.0001 0.742 (0.823-0.858) (0.723-0.762) <.0001 0.839 <.0001 0.806 (0.820-0.858) (0.781-0.833) P Value <.0001
All Patients
Hazard Ratios (95% CI) 0.754 (0.743-0.765) <.0001 0.889 (0.878-0.900) <.0001 0.804 (0.792-0.816) P Value <.0001
P Value <.0001 <.0001
Adjusted <.0001 for patient characteristics, comorbidity, Hispanic 0.947 <.0001 0.917 <.0001 0.866 <.0001 0.911 <.0001 and AMI (0.929-0.966) (0.898-0.937) (0.843-0.889) (0.900-0.923) location Adjusted Black 0.771 <.0001 0.859 <.0001 0.827 <.0001 0.814 <.0001 for patient (0.751-0.791) (0.839-0.880) (0.798-0.856) (0.802-0.827) characteristics, comorbidity AMI location Hispanic 0.969 <0.01 0.937 <.0001 0.965 0.028 0.949 <.0001 stratified by (0.947-0.992) (0.915-0.959) (0.936-0.996) (0.935-0.963) hospital IDs Abbreviation: AMI, acute myocardial infarction; CI, confidence interval.
1136 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 101, NO. 11, NOVEMBER 2009
Racial disparities in cardiac revascularization
hospital provides strong evidence that disparities are not related to differences in insurance, comorbidity, or differences in the hospitals where different racial groups tend to receive care.
A number of our findings merit further discussion. Our finding that black AMI patients were less likely than whites or Hispanics to receive revascularization during their hospitalization both reinforces and extends prior
Table 4. Hazard Rates for Receipt of Coronary Artery Bypass Grafting for Blacks and Hispanics With Varying Insurance as Compared to Whites Medicare Hazard Ratios (95% CI) Black 0.589 (0.565-0.613) Hispanic 0.954 (0.926-0.982) Black 0.635 (0.609-0.662)
Private Insurance
Hazard P Ratios Value (95% CI) <.0001 0.562 (0.537-0.588) <.01 0.906 (0.873-0.940) <.0001 0.634 (0.606-0.664)
Uninsured/ Medicaid
All Patients
Hazard Ratios (95% CI) 0.458 (0.430-0.488) 0.853 (0.816-0.891) 0.558 (0.522-0.595)
Hazard P P P Ratios Value Value (95% CI) Value Unadjusted <.0001 <.0001 0.606 (0.589-0.622)<.0001 <.0001 <.0001 0.979 (0.960-0.999) 0.042 Adjusted <.0001 <.0001 0.615 for patient (0.598-0.633) characteristics, <.0001 comorbidity, Hispanic 0.883 <.0001 0.911 <.0001 0.888 <.0001 0.893 and AMI (0.857-0.911 (0.878-0.946) (0.848-0.930) (0.874-0.911) location <.0001 Adjusted Black 0.664 <.0001 0.660 <.0001 0.611 <.0001 0.646 for patient (0.636-0.694) (0.629-0.692) (0.569-0.655) (0.627-0.665) characteristics, <.0001 comorbidity, AMI location Hispanic 0.938 <.001 0.957 0.042 1.054 0.064 0.964 <.01 stratified by (0.905-0.979) (0.917-0.998) (0.997-1.115) (0.941-0.988) hospital IDs Abbreviation: AMI, acute myocardial infarction; CI, confidence interval.
Table 5. Hazard Rates for Receipt of Revascularization (Either Percutaneous Coronary Intervention or Coronary Artery Bypass Grafting) for Blacks and Hispanics as Compared to Whites Medicare
Unadjusted
Hazard Ratios (95% CI) Black 0.688 (0.673-0.702) Hispanic 0.910 (0.895-0.925) Black 0.725 (0.710-0.741)
Private Insurance
Hazard Ratios (95% CI) 0.710 (0.696-0.725) <.0001 0.856 (0.841-0.872) 0.792 <.0001 (0.775-0.808) P Value <.0001
Uninsured/ Medicaid
All Patients
Hazard Hazard P Ratios Ratios (95% CI) Value (95% CI) 0.632 <.0001 0.745 (0.615-0.650) (0.736-0.755) <.0001 0.768 <.0001 0.905 (0.751-0.785) (0.895-0.914) 0.742 0.754 <.0001 (0.720-0.764) <.0001 (0.744-0.764) P Value <.0001
P Value <.0001 <.0001
Adjusted <.0001 for patient characteristics, comorbidity, Hispanic 0.933 0.912 0.864 0.905 <.0001 and AMI (0.917-0.949) <.0001 (0.896-0.929) <.0001 (0.844-0.884) <.0001 (0.895-0.915) location Adjusted Black 0.742 <.0001 0.819 <.0001 0.772 <.0001 0.773 <.0001 for patient (0.726-0.759) (0.802-0.837) (0.748-0.796) (0.763-0.784) characteristics, comorbidity, AMI location Hispanic 0.969 <.01 0.943 <.0001 0.967 0.016 0.950 <.0001 stratified by (0.950-0.988) (0.924-0.962) (0.941-0.994) (0.938-0.962) hospital IDs Abbreviation: AMI, acute myocardial infarction; CI, confidence interval.
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 101, NO. 11, NOVEMBER 2009 1137
Racial disparities in cardiac revascularization
research. While an array of prior studies have found evidence of disparities in revascularization, most of these studies have been limited to patients with a single type of health insurance (eg, Medicare),4,5,16, patients receiving care in a single medical center,9 or patients enrolled in clinical registries.18 Our finding of significant disparity among black patients relative to whites, regardless of insurance coverage, confirms results of prior studies based on single types of health insurance. In addition, the majority of studies investigating disparities in cardiovascular disease focus only on black populations. There is less evidence documenting disparities for Hispanic populations, although some studies suggest relatively similar rates of revascularization for Hispanic and whites.19-21 Likewise, our study found that the magnitude of disparity for Hispanics was relatively small, although statistically significant. In general, our findings were remarkably consistent across procedure type (PCI vs CABG) and insurance status. Alternatively, a deeper look at the details of our results reveals some important and interesting differences. While blacks were approximately 19% less likely to receive PCI as compared to whites for all insurance cohorts, blacks were approximately 35% less likely to receive CABG. Our finding that disparity in revascularization is larger for CABG than for PCI has been observed in some but not all prior studies.4,9,16,18 Moreover, the magnitude of the disparities that we observed for CABG and PCI were very similar to those observed in a 1997 study by Peterson et al using data collected between 1984 and 1992. The lack of change in the magnitude of disparity raises concerns about the ability of ongoing efforts to improve health services equity in producing measurable reductions in racial disparities.22,23 Finally, we found that the magnitude of disparity for black patients in the use of PCI was relatively consistent across insurance categories, but the disparity in the use of CABG was substantially smaller for black patients with Medicare or private insurance, compared to patients with Medicaid/uninsured. While it appears that insurance type does, at least partially, impact disparities for CABG, significant disparity nevertheless remained, regardless of insurance type. These findings are consistent with other studies suggesting that the removal of financial barriers will not eliminate disparities in utilization of health services. Our finding that a higher proportion of black and Hispanic AMI patients were admitted from the emergency department as compared to whites is also interesting. One potential explanation for this finding is that blacks and Hispanics were less likely to be admitted as transfers from other acute care hospitals and as a consequence, and thus, by default, a higher proportion of blacks and Hispanics were admitted from the emergency department. Such a mechanism is plausible given evidence from prior studies that black AMI patients are less 1138 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
likely to be transferred than whites.16 Alternatively, it is possible that the greater proportion of blacks and Hispanics admitted through the emergency department is related to a lack of access to primary care for these populations, differences in patient preference, or provider decision making.24 There are a number of limitations to our study that merit brief mention. First, our study relied upon administrative data for our risk adjustment models and thus may have been subject to bias if black, white, and Hispanic patients differed in severity in ways that were not captured. Nevertheless, a growing array of studies has now shown that administrative data generally offer similar results to clinical data sources. Second, our analyses were somewhat limited by the idiosyncrasies of the state inpatient data. For example, the state inpatient data do not assign unique patient identifiers, making it impossible to track patients who are transferred between hospitals. However, it seems unlikely that having unique identifiers would have substantially changed our results. In summary, in an analysis of all-payor data from 9 states we found significant disparities in revascularization for black, white, and Hispanic AMI patients with similar health insurance. These findings have important implications. First, conclusions about racial disparities based on specific types of insurance (eg, Medicaid, Medicare) may differ modestly across insurance types, although in general we found little bias in our estimation of the magnitude of disparity depending on insurance, with the exception of CABG. Possibly, the use of invasive, expensive procedures with significant follow-up is more sensitive to insurance coverage. Second, the fact that significant disparity persisted regardless of insurance type suggests that elimination of financial barriers is not likely to eliminate disparities. Finally, our results call into question the efficacy of various efforts to reduce racial disparities.
References
1. Gillum RF. Coronary artery bypass surgery and coronary angiography in the United States, 1979-1983. Am Heart J. 1987;113(5):1255-1260. 2. Goldberg KC, Hartz AJ, Jacobsen SJ, Krakauer H, Rimm AA. Racial and community factors influencing coronary artery bypass graft surgery rates for all 1986 Medicare patients. JAMA. 1992;267(11):1473-1477. 3. Gauri AJ, Davis A, Hong T, Burke MC, Knight BP. Disparities in the use of primary prevention and defibrillator therapy among blacks and women. Am J Med. 2006;119(2):167e17-21. 4. Ayanian JZ, Udvarhelyi IS, Gatsonis CA, Pashos CL, Epstein AM. Racial differences in the use of revascularization procedures after coronary angiography. JAMA. 1993;269(20):2642-2646. 5. Whittle J, Conigliaro J, Good CB, Lofgren RP. Racial differences in the use of invasive cardiovascular procedures in the Department of Veterans Affairs medical system. N Engl J Med. 1993;329(9):621-627. 6. Mukamel DB, Weimer DL, Buchmueller TC, Ladd H, Mushlin AI. Changes in racial disparities in access to coronary artery bypass grafting surgery between the late 1990s and early 2000s. Med Care. 2007;45(7):664-671. 7. Jha AK, Fisher ES, Li Z, Orav EJ, Epstein AM. Racial trends in the use of major procedures among the elderly. N Engl J Med. 2005;353(7):683-691. 8. Schneider EC, Leape LL, Weissman JS, Piana RN, Gatsonis C, Epstein AM. Racial differences in cardiac revascularization rates: does “over-
VOL. 101, NO. 11, NOVEMBER 2009
Racial disparities in cardiac revascularization use” explain higher rates among white patients? Ann Intern Med. 2001; 135(5):328-337. 9. Peterson ED, Shaw LK, DeLong ER, Pryor DB, Califf RM, Mark DB. Racial variation in the use of coronary-revascularization procedures. Are the differences real? Do they matter? N Engl J Med. 1997;336(7):480-486. 10. Carlisle DM, Leake BD, Shapiro MF. Racial and ethnic differences in the use of invasive cardiac procedures among cardiac patients in Los Angeles County, 1986 through 1988. Am J Public Health. 1995;85(3):352-356. 11. HCUP SID overview. http://www.hcup-us.ahrq.gov/sidoverview.jsp Accessed April 21, 2009. 12. Cram P, Pham HH, Bayman L, Vaughan-Sarrazin MS. Insurance status of patients admitted to specialty cardiac and competing general hospitals: Are accusations of cherry picking justified? Med Care. 2008;46(5):467-475. 13. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8-27. 14. Southern DA, Quan H, Ghali WA. Comparison of the Elixhauser and Charlson/Deyo methods of comorbidity measurement in administrative data. Med Care. 2004;42(4):355-360. 15. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130-1139. 16. Popescu I, Vaughan-Sarrazin MS, Rosenthal GE. Differences in mortality and use of revascularization in black and white patients with acute MI admitted to hospitals with and without revascularization services. JAMA. 2007;297(22):2489-2495. 17. Rosenthal GE, Vaughan SM, Hannan EL. In-hospital mortality following coronary artery bypass graft surgery in Veterans Health Administration and private sector hospitals. Med Care. 2003;41(4):522-535.
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
18. Sonel AF, Good CB, Mulgund J, et al. Racial variations in treatment and outcomes of black and white patients with high-risk non-ST-elevation acute coronary syndromes: insights from CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of the ACC/AHA Guidelines?). Circulation. 2005;111(10):12251232. 19. Barnhart JM, Fang J, Alderman MH. Differential use of coronary revascularization and hospital mortality following acute myocardial infarction. Arch Intern Med. 2003;163(4):461-466. 20. Bertoni AG, Goonan KL, Bonds DE, Whitt MC, Goff DC, Jr., Brancati FL. Racial and ethnic disparities in cardiac catheterization for acute myocardial infarction in the United States, 1995-2001. J Natl Med Assoc. 2005;97(3):317-323. 21. Yarzebski J, Bujor CF, Lessard D, Gore JM, Goldberg RJ. Recent and temporal trends (1975-1999) in the treatment, hospital, and long-term outcomes of Hispanic and non-Hispanic white patients hospitalized with acute myocardial infarction: a population-based perspective. Am Heart J. 2004;147(4):690-697. 22. Brown CP, Ross L, Lopez I, Thornton A, Kiros GE. Disparities in the receipt of cardiac revascularization procedures between blacks and whites: an analysis of secular trends. Ethn Dis. 2008;18(2 suppl 2):S2-112-117. 23. Lucas FL, DeLorenzo MA, Siewers AE, Wennberg DE. Temporal trends in the utilization of diagnostic testing and treatments for cardiovascular disease in the United States, 1993-2001. Circulation. 2006;113(3):374-379. 24. Cunningham PJ. What accounts for differences in the use of hospital emergency departments across US communities? Health Aff. 2006;25(5):w324-w336. n
VOL. 101, NO. 11, NOVEMBER 2009 1139