Full text of DOI:10.1002/clc.4960251008

Clin. Cardiol. 25, 479–483 (2002)
Does Racial Bias Exist in the Medical Management of Heart Failure?
KISHORE J. HARJAI, M.B., B.S., FACC, EDUARDO NUNEZ, M.D., MEHUL SHAH, M.B., B.S., JEFF NEWMAN
Department of Cardiology, Ochsner Clinic, New Orleans, Louisiana, USA
Summary
Background: It is suspected that effective therapies are of-
ten underutilized in black compared with white patients with
coronary artery disease (CAD).
Hypothesis: We hypothesized that an unfavorable bias may
exist against black patients in the medical management of
heart failure.
1.02–2.69) than white patients at the time of discharge from
the hospital. No multivariate association was seen between
patient race and use of digoxin or diuretics. In an analysis of
subsets of patients with ejection fraction < 45% (n = 260), no
association was seen between patient race and utilization of
effective medical therapy.
Conclusion: Our results show no unfavorable bias against
black patients with decompensated heart failure.
Methods: In 566 consecutive adult subjects who were dis-
charged alive from the hospital with a principal discharge di-
agnosis of heart failure, we assessed the effect of patient race
on utilization of classes of medications (angiotensin-convert-
ing enzyme inhibitors [ACEI], digitalis, diuretic agents) and
combinations of medications (effective vasodilators, i.e., ACEI
or combined hydralazine and nitrate; effective combination
therapy, i.e., effective vasodilator with digitalis and diuretic)
known to be beneficial in symptomatic heart failure.
Results: Compared with black patients (n = 182), white pa-
tients were older, had a higher incidence of coronary artery
disease, lower incidence of hypertension, and lower serum
creatinine and left ventricular end-diastolic diameter. In crude
analyses, the utilization of all medications was similar be-
tween white and black patients. After adjustment for clinical
differences, black patients were more likely to receive ACEI
(adjusted odds ratio [OR] = 1.84; 95% confidence interval
[CI] 1.13–3.01), effective vasodilators (OR = 1.97; CI 1.20–
3.23), and effective combination therapy (OR = 1.66; CI
Key words: heart failure, race, angiotensin-converting en-
zyme inhibitors
Introduction
It is suspected that effective therapies are often underuti-
lized in black compared with white patients with coronary
artery disease (CAD). There are surprisingly few data on racial
differences in medical treatment for heart failure, despite evi-
dence for worse outcomes in black patients with heart failure.1
In an analysis of data from the studies of left ventricular dys-
function (LVD) prevention and treatment trials, blacks with
mild to moderate LV systolic dysfunction were at higher risk
for progression of heart failure and death from any cause than
similarly treated whites.2 National statistics show that the rate
of death from heart failure in 1990 was about four times high-
er in black than in white subjects.3
In this study, we attempted to evaluate the impact of pa-
tients’ race on the utilization of effective medical therapy for
heart failure. We hypothesized that in patients discharged from
the hospital with a principal diagnosis of heart failure, effective
medical therapy was less likely to be utilized in black than in
white subjects.
Address for correspondence:
Kishore J. Harjai, M.B., B.S., FACC
Division of Cardiology
Methods
William Beaumont Hospital
3601 W. 13 Mile Road
Royal Oak, MI 48073, USA
e-mail: kharjai@beaumont.edu
Patient Selection and Data Collection
Over a consecutive 12-month period from September 1994
to August 1995, 614 hospitalizations for diagnosis-related
group (DRG) 127 (heart failure and shock) were recorded at
Received: November 20, 2000
Accepted with revision: January 18, 2002

480
Clin. Cardiol. Vol. 25, October 2002
the Ochsner Foundation Hospital. This hospital is located in
suburban New Orleans and receives patients from a large mul-
tispecialty group practice of over 400 physicians. Of the total
cohort, 592 adult patients were discharged alive from the hos-
pital. We excluded 10 patients whose race was classified as
Hispanic or “other,” and 1 patient who was noncompliant with
medical therapy and had 16 hospitalizations during the study
period. Thus, the study population included 566 patients, con-
sisting of 384 white and 182 black subjects. Using administra-
tive and clinical databases maintained at this institution, we
obtained information relating to patient age, gender, utilization
of heart failure medication at the time of discharge from the
hospital (specifically angiotensin-converting enzyme inhib-
itors [ACEI], digoxin, diuretics, nitrates, hydralazine), select-
ed clinical characteristics such as admission vital signs and
serum chemistry values, comorbid conditions (hypertension,
CAD, atrial fibrillation, chronic obstructive pulmonary dis-
ease), and echocardiographic variables (LV ejection fraction,
septal thickness, posterior wall thickness, LV end-diastolic di-
ameter, right-ventricular [RV] systolic function, and left atrial
diameter). Details of the echocardiographic technique used at
our institution have been described previously.4 Clinical re-
cords were reviewed when such information was not available
from the databases.
sought between demographic, clinical, and echocardiograph-
ic variables and the utilization of medications at the time of
hospital discharge.
The independent effect of patient race on utilization of med-
ical therapy at the time of discharge was evaluated using mixed
stepwise multiple logistic regression. For each class of medi-
cation or group of medications listed under study outcomes,
we constructed multivariate models relating patient race to
utilization of that particular class of medication or group of
medications. Patients race was forced into the multivariate
equations. Demographic, clinical, and echocardiographic vari-
ables, shown in Table I, were included in the model for each
medication or group of medications. The results were ex-
pressed as odd ratios (OR) and confidence intervals (CI).
Results
Patient Characteristics
Demographic, clinical, and echocardiographic characteris-
tics of the study patients are shown in Table I. Thus, compared
with blacks, whites were older and more likely to be male.
Whites had a lower number of hospitalizations for any reason
within the 6 months preceding the index hospitalization. Dur-
ing the index hospitalization, both groups of patients had a
similar length of stay and similar frequency of need for inten-
sive care. Whites were more likely to be under the primary
care of a cardiologist; they had a higher incidence of CAD and
atrial fibrillation, and a lower incidence of hypertension as
well as lower systolic and diastolic blood pressures at the time
of admission. In addition, they had lower serum creatinine and
sodium levels, but similar blood urea nitrogen level at the time
of hospitalization. There was no difference between patient
groups in the frequency with which echocardiographic exam-
ination had been performed prior to or during the index hospi-
talization, or in the incidence of LV systolic dysfunction (i.e.,
ejection fraction < 45%); however, whites had larger left atrial
and smaller LV end-diastolic diameters than blacks.
Of the 566 patients, 409 had known LV ejection fraction,
assessed by echocardiography performed prior to or during the
index hospitalization. In 260 patients, LV ejection fraction was
known to be < 45%.
Study Definitions and Outcomes
For the purpose of this study, we defined effective vasodila-
tor therapy as utilization of either an ACEI or a combination
of hydralazine and nitrate at the time of discharge. Effective
combination therapy was defined as combination of effective
vasodilator therapy with digitalis and a diuretic agent. The ac-
tual doses employed at the time of discharge were not includ-
ed in the analysis.
The outcomes of the study were the frequencies with which
individual classes of medications (ACEI, diuretics, digoxin),
effective vasodilator therapy, and effective combination thera-
py were utilized at the time of discharge from the hospital. Use
of beta blockers was not monitored for the purpose of this
study, since the study period preceded the publication of recent
clinical trials documenting the benefits of beta-blocker thera-
py in patients with heart failure.5–7
Utilization of Effective Medical Therapy
The crude frequencies of utilization of medications in
white and black patients are shown in Table II. For all classes
of medications shown, there were no significant differences
between white and black patients. A tendency toward greater
use of effective vasodilator therapy in black compared with
white patients was seen in the analysis of the total cohort (77
vs. 70%, p = 0.07).
Statistical Analysis
Continuous variables are expressed as mean ± one standard
deviation (SD), and discrete variables as percentages. Differ-
ences between groups for continuous variables were assessed
using the Wilcoxon rank sum test. For categorical variables,
differences between patient groups were assessed using con-
tingency tables. Univariate analyses were performed to assess
differences in clinical and echocardiographic variables be-
tween white and black patients. Also, a univariate relation was
After multivariate adjustment for clinical differences,
black patients were more likely to receive ACEIs (adjusted
OR = 1.84; 95% CI 1.13–3.01), effective vasodilator therapy
(adjusted OR = 1.97; 95% CI 1.20–3.23), and effective com-
bination therapy (adjusted OR = 1.66; 95% CI 1.02–2.69)
compared with white patients (Fig. 1A). No multivariate as-
sociation was seen between patient race and use of digoxin or
diuretic agents. Variables that remained significantly associ-

K. J. Harjai et al.: Racial bias in heart failure management
481
TABLE I Clinical characteristics of patients hospitalized with heart failure
Whites
Blacks
(n = 384)
(n = 182)
p Value
Age, years
Male gender (%)
Recent hospitalizations a
Length of stay during index hospitalization
Cardiologist care during index hospitalization (%)
Intensive care stay during index hospitalization (%)
Coronary artery disease (%)
Atrial fibrillation (%)
Hypertension (%)
Chronic obstructive pulmonary disease (%)
Systolic blood pressure, mmHg
Diastolic blood pressure, mmHg
Serum sodium, mEq/l
Blood urea nitrogen, mg/dl
Serum creatinine, mg/dl
Echocardiography performed before or during hospitalization (%)
Left ventricular ejection fraction < 45% b (%)
Left atrial diameter, cm, parasternal view b
Left ventricular end-diastolic diameter, cm b
72 ± 13
225 (59)
0.7 ± 1.0
3.9 ± 2.8
269 (70)
36 (9)
177 (46)
153 (40)
199 (52)
58 (15)
134 ± 29
75 ± 16
137 ± 5
28 ± 17
1.6 ± 1.0
277 (72)
167 (60)
4.9 ± 0.9
5.3 ± 1.1
65 ± 15
80 (44)
1.6 ± 3.3
4.0 ± 3.2
97 (53)
20 (11)
54 (30)
32 (18)
117 (64)
29 (16)
146 ± 36
84 ± 22
138 ± 4
28 ± 19
2.0 ± 1.7
132 (72)
93 (66)
4.6 ± 0.8
5.5 ± 1.0
< 0.0001
0.001
0.06
NS
<0.0001
NS
0.0002
<0.0001
0.005
NS
0.0004
<0.0001
0.001
NS
0.036
NS
NS
0.01
0.047
Numbers in parentheses represent percentages, rounded off to the nearest integer value.
a Number of hospitalizations in the prior 6 months.
b Data based on 409 patients in whom echocardiographic examination had been performed before or during the index hospitalization.
TABLE II Frequency of use of medications at hospital discharge
Medication
Whites
Blacks
All patients
n = 384
246 (64)
268 (70)
243 (63)
351 (91)
186 (48)
n = 182
127 (70)
140 (77)
112 (62)
172 (95)
97 (53)
Angiotensin-converting enzyme inhibitors (%)
Effective vasodilator therapy a (%)
Digoxin (%)
Diuretic agent (%)
Effective combination therapy (%)
Patients with known left ventricular dysfunction b
Angiotensin-converting enzyme inhibitors (%)
Effective vasodilator therapy (%)
Digoxin (%)
Diuretic agent (%)
Effective combination therapy (%)
n = 167
125 (75)
141 (84)
130 (78)
151 (90)
107 (64)
n = 89
67 (75)
77 (87)
70 (79)
82 (92)
61 (69)
Numbers in parentheses represent percentages, rounded off to the nearest integer value.
a p = 0.07.
b Left ventricular ejection fraction <45%.
ated with utilization of each of the medication classes in the
final multivariate models are shown in Table III.
Discussion
Our analysis disproves our hypothesis that effective medi-
cal therapy is underutilized in black compared with white pa-
tients with symptomatic heart failure. In crude analyses, no
differences were found between black and white subjects with
respect to the utilization of any class or group of medications,
known to be beneficial in patients with heart failure. Further-
more, after multivariate adjustment for clinical and echocar-
Subset Analysis
Similar univariate and multivariate analyses were per-
formed among patients with LV ejection fraction < 45% (n =
260). No significant association was seen between race and
utilization of any class or group of medications (Fig. 1B).

482
Clin. Cardiol. Vol. 25, October 2002
Angiotensin-converting
enzyme inhibitors
Angiotensin-converting
p = 0.015
p = 0.007
p = NS
p = NS
enzyme inhibitors
Effective vasodilator therapy
Effective vasodilator therapy
Digoxin
p = NS
p = NS
Digoxin
p = NS
Diuretic agent
p = NS
Diuretic agent
Effective combination therapy
p = 0.042
p = NS
Effective combination therapy
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
(A) Utilized more in whites
Utilized more in blacks
(B) Utilized more in whites
Utilized more in blacks
FIG. 1 Multivariate odds ratios and confidence intervals for utilization of heart failure medications in black versus white patients at the time of
discharge from the hospital. (A) All patients, (B) patients with known left ventricular (LV) dysfunction (LV ejection fraction ≤45%). NS = not
significant.
TABLE III Significant multivariate correlates of utilization of medications for heart failure
Medication class
ACE inhibitors
Multivariate correlates
Serum creatinine at the time of hospitalization a
Black race
Male gender
Effective vasodilator therapy
Digoxin
Digoxin use
Serum creatinine at the time of hospitalization a
Black race
In-patient care by cardiologist
Systolic blood pressure at the time of hospitalization a
Atrial fibrillation
Heart rate at the time of hospitalization
Serum creatinine at the time of hospitalization a
In-patient care by cardiologist
Age a
Hospitalization (any cause) in the last 6 months
Length of stay a
Diuretics
Serum creatinine at the time of hospitalization a
Diastolic blood pressure at the time of hospitalization
Age a
Hospitalization (any cause) in the last 6 months
Systolic blood pressure a
Effective combination therapy
Serum creatinine at the time of hospitalization a
Hospitalization (any cause) in the last 6 months
Male gender
In-patient care by cardiologist
Black race
Within each medication class, correlates are arranged in descending order of significance.
a Lower values of these variables were associated with higher utilization of the respective medication class.
diographic differences, ACEIs, effective vasodilator therapy,
and combination therapy were all more likely to be utilized
among black than among white patients. Notably, no signifi-
cant racial differences in utilization were seen within the sub-
group of patients who were known to have depressed ejection
fraction. Thus, it is likely that at least part of the higher adjust-
ed utilization of ACEIs, effective vasodilator therapy, and
combination therapy in black compared with white patients in
the total cohort relates to the higher incidence of depressed
ejection fraction in black patients (66 vs. 60%). Data on LV
ejection fraction were not available in all patients, and hence
were not used in the multivariate analysis of the total cohort.
Differences in utilization of medical therapies and interven-
tions across patient populations are often related to potential
differences in access to medical care,8 disease prevalence and
severity, and socioeconomic and cultural factors.9 The design
employed by our study obviates any confounding effect from
lack of access to care or differences in disease severity, since

K. J. Harjai et al.: Racial bias in heart failure management
483
patients in our study were all hospitalized for symptomatic
heart failure, and by definition had symptomatic heart failure
and full access to care.
charge medications do not represent the entire spectrum of
medical therapy, but are only an index of utilization at one
point in time. In a chronic condition such as heart failure, med-
ical therapy is often modified in an outpatient setting, after dis-
charge from the hospital, to achieve optimum long-term out-
comes. Postdischarge utilization of medications was not
measured in our analysis.
Several other modalities of treatment are often used in the
management of subjects with heart failure. These include in-
vasive procedures, such as angiography or angioplasty, cardiac
surgery for CAD or valvular lesions, and, in advanced heart
failure, referral for heart transplantation, chronic outpatient in-
otropic therapy, or device therapy. The utilization of these
modalities was not monitored for the purpose of this study.
Finally, by design, our study does not attempt to estimate the
clinical impact of utilization of effective medical therapy on
outcomes such as readmission rates and mortality.
Despite known differences in clinical outcomes between
white and nonwhite patients with heart failure,1–3 existing lit-
erature on the effect of patient race on processes of care in
heart failure is rather sparse. In an analysis of 9,105 adults hos-
pitalized with one of nine illnesses associated with an average
6-month mortality of 50%, the Study to Understand Prognoses
and Preferences for Outcomes and Risks of Treatment (SUP-
PORT) investigators found that seriously ill black patients re-
ceived less resource-intensive care than other patients after ad-
justment for severity of illness.10 Thus, black patients had less
procedures, lower hospitalization cost, and a lower incidence
of care by cardiologists. Differences in resource use were less
marked, but remained significant after adjustment for special-
ty of the caregiver physician. Illnesses included in this study
were acute respiratory failure, chronic obstructive lung dis-
ease, congestive heart failure, cirrhosis, nontraumatic coma,
metastatic colon cancer, advanced non-small-cell lung cancer,
multiple organ system failure with sepsis, or multiple organ
systems failure with malignancy. In the subset of patients with
heart failure, black patients were less likely to be under the care
of the cardiologist. Contrary to the findings of the SUPPORT
investigators, Philbin and DiSalvo found higher hospitaliza-
tion utilization, measured as length of stay and hospital
charges, among black patients hospitalized with heart failure.9
It is not known whether differences in clinical outcomes be-
tween black and white patients with heart failure are related to
inherent biological characteristics or to differences in clinical
management strategies. It is plausible, yet unproven, that
worse outcomes seen in blacks in some studies1, 3 may be re-
lated to inadequate access to treatment or underutilization of
effective strategies. On the contrary, the higher risk of progres-
sion of heart failure and death from any cause among blacks
enrolled in the SOLVD trials are unlikely to be related to ac-
cess issues; the SOLVD data suggest that differences in the
natural history of ischemic and nonischemic LV dysfunction
account for differences in clinical course between black and
white patients.2
Acknowledgment
The authors are indebted to Mr. Mario Vaz for editorial as-
sistance with the preparation of this manuscript.
References
1. Alexander M, Grumbach K, Selby J, Brown AF, Washington E,
Remy L, Rowell R, Massie BM: Congestive heart failure hospital-
izations and survival in California: Patterns according to race/eth-
nicity. Am Heart J 1999;13:919–927
2. Dries DL, Exner DV, Gersh BJ, Cooper HA, Carson PE, Domanski
MJ: Racial differences in the outcome of left ventricular dysfunc-
tion. N Engl J Med 1999;340:609–616
3. Gillum RV: Epidemiology of heart failure in the United States.
Am Heart J 1993;126:1042–1047
4. Harjai KJ, Nunez E, Turgut T, Shah MP, Humphrey JS, Newman J,
Cheirif J, Smart FW, Ventura HO: The independent effects of left
ventricular ejection fraction on short-term outcomes and resource
utilization following hospitalization for heart failure. Clin Cardiol
1999;22(3):184–190
5. Packer M, Bristow MR, Cohn JN, Colucci WS, Fowler MB,
Gilbert EM, Shusterman NH: Effect of carvedilol on morbidity and
mortality in patients with chronic heart failure. N Engl J Med
1996;334:1349–1355
Study Limitations
6. CIBIS-II Investigators and Committees: The cardiac insufficiency
bisoprolol study II: (CIBIS-II): A randomised trial. Lancet 1999;
353: 9–13
7. MERIT-HF Study Group: Effect of metoprolol CR/XL in chronic
heart failure: Metoprolol CR/XL Randomised Intervention Trial
in Congestive Heart Failure (MERIT-HF). Lancet 1999;353:
2001–2007
8. Alexander M, Grumbach K, Selby J, Brown AF, Washington E:
Hospitalization for congestive heart failure. Explaining racial dif-
ferences. J Am Med Assoc 1995;274:1037–1042
9. Philbin EF, DiSalvo TG: Influence of race and gender on care pro-
cess, resource use, and hospital-based outcomes in congestive heart
failure. Am J Cardiol 1998;82:76–81
Our study has several limitations, including, perhaps most
important, the lack of generalizability of our results. Our study
is a single-center study. Quite unlike community settings
where in-patient care for heart failure may be imparted pre-
dominantly by noncardiologists, 366 (65%) patients from our
cohort were under the direct care of cardiologists during the in-
dex hospitalization. Our center has a dedicated team of heart
failure physicians and personnel; in prior experience, this team
has been directly responsible for the in-patient care of 10–11%
of the total number of DRG 127 hospitalizations. Further-
more, our study does not rule out the possibility of self-selec-
tion bias, that is, patients who chose to receive their care at our
center may not represent the universe of patients with heart
failure from the same geographic region. In addition, dis-
10. Phillips RS, Hamel MB, Teno JM, Bellamy P, Broste SK, Califf
RM, Vidaillet H, Davis RB, Muhlbaier LH, Connors AF Jr, for the
SUPPORT Investigators: Race, resource use, and survival in seri-
ously ill hospitalized adults. J Gen Intern Med 1996;11:387–396