FD‡RC¶CUD

Journal of the American College of Cardiology Vol. 51, No. 17, 2008© 2008 by the American College of Cardiology Foundation ISSN 0735-1097/08/$34.00P

Quality of Care of and Outcomes forAfrican Americans Hospitalized With Heart FailureFindings From the OPTIMIZE-HF (Organized Program to InitiateLifesaving Treatment in Hospitalized Patients With Heart Failure) Registry

Clyde W. Yancy, MD, FACC,* William T. Abraham, MD, FACC,† Nancy M. Albert, PHD, RN,‡Robert Clare, MS,§ Wendy Gattis Stough, PHARMD,�¶ Mihai Gheorghiade, MD, FACC,#Barry H. Greenberg, MD, FACC,** Christopher M. O’Connor, MD, FACC,†† Lilin She, PHD,§Jie Lena Sun, MS,§ James B. Young, MD, FACC,‡‡ Gregg C. Fonarow, MD, FACC§§

Dallas, Texas; Columbus and Cleveland, Ohio; Durham and Research Triangle Park, North Carolina;Chicago, Illinois; and San Diego and Los Angeles, California

Objectives We sought to examine the characteristics, quality of care, and clinical outcomes for a large cohort of African-American patients hospitalized with heart failure (HF) in centers participating in a quality improvement initiative.

Background Heart failure in African Americans is characterized by variations in natural history, lesser response to evidence-based therapies, and disparate health care. We hypothesized that a performance improvement program willachieve similar adherence to quality measures in African Americans admitted with HF compared withnon–African Americans.

Methods The OPTIMIZE-HF (Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients With Heart Fail-ure) registry-based performance-improvement program includes a pre-specified 10% subgroup with 60- to 90-day follow-up. Data on quality of care measures and outcomes were analyzed for 8,608 African-American pa-tients compared with 38,501 non–African-American patients.

Results African Americans were significantly younger and more likely to receive evidence-based medications but lesslikely to receive discharge instructions and smoking cessation counseling. In multivariable analyses, African-American race was an independent predictor of lower in-hospital mortality (odds ratio 0.71; 95% confidenceinterval 0.57 to 0.87; p � 0.001) but similar hospital length of stay. After multivariable adjustment, post-discharge outcomes were similar for American-American and non–African-American patients, but African-American race was associated with higher angiotensin-converting enzyme inhibitor prescription and left ventricu-lar function assessment; no other HF quality indicators were influenced by race.

Conclusions In the context of a performance-improvement program, African Americans with HF received similar or bettertreatment with evidence-based medications, less discharge counseling, had better in-hospital survival, and simi-lar adjusted risk of follow-up death/repeat hospital stay. (Organized Program to Initiate Lifesaving Treatment InHospitalized Patients With Heart Failure [OPTIMIZE-HF]; NCT00344513) (J Am Coll Cardiol 2008;51:1675–84)© 2008 by the American College of Cardiology Foundation

ublished by Elsevier Inc. doi:10.1016/j.jacc.2008.01.028

University Medical Center/Duke Clinical Research Institute, Durham, North Caro-lina; ‡‡Department of Cardiovascular Medicine, Cleveland Clinic Foundation,Cleveland, Ohio; and the §§ Department of Medicine, University of California LosAngeles Medical Center, Los Angeles, California. GlaxoSmithKline funded boththe OPTIMIZE-HF registry and this analysis of registry data. For full authordisclosures, please see the end of this paper. John R. Teerlink, MD, served asGuest Editor for this article.

rom the *Baylor Heart and Vascular Institute, Baylor University Medical Center,allas, Texas; †Division of Cardiology, Ohio State University, Columbus, Ohio;

George M. and Linda H. Kaufman Center for Heart Failure and §Duke Clinicalesearch Institute, Durham, North Carolina; �Department of Clinical Research,ampbell University School of Pharmacy, Research Triangle Park, North Carolina;Department of Medicine, Duke University Medical Center, Durham, Northarolina; #Division of Cardiology, Feinberg School of Medicine, Northwestern

niversity, Chicago, Illinois; **Department of Medicine, University of California Saniego Medical Center, San Diego, California; ††Division of Cardiology, Duke

Manuscript received July 30, 2007; revised manuscript received January 24, 2008,accepted January 29, 2008.

cApHA(HhtvprLAwp

aAraccasmqpoAc

Hqgp

orOcatTdAe(iam

M

SO(pigbadsimobcepOassmalAuts9ohifc1odd

1676 Yancy et al. JACC Vol. 51, No. 17, 2008African Americans Hospitalized With HF: The OPTIMIZE-HF Registry April 29, 2008:1675–84

Marked population differencesexist in the prevalence, morbid-ity, and mortality associated withcardiovascular disease. In con-junction with biological risk fac-tors, social and environmentaldeterminants of cardiovasculardisease might contribute to thetreatment and outcome inequi-ties seen in different populations(1). These differences have thepotential to contribute to ob-served gaps in the overall lifeexpectancy and quality of life andare felt to be partly responsiblefor health care disparities. Afri-can Americans are among thosemost seriously affected by dispar-ate health care in the U.S. (2).

Recent statistics confirm thatardiovascular disease is the leading cause of death forfrican Americans (3,4). Heart failure (HF) is especiallyroblematic for African Americans. Both the prevalence ofF and mortality due to chronic HF are increased infrican Americans compared with the general population

4,5). The etiology of left ventricular dysfunction leading toF in African Americans is also different, and the natural

istory is more aggressive than that seen in other popula-ions. Among African-American patients with HF and leftentricular systolic dysfunction (LVSD), only one-fourth ofatients have a known ischemic etiology of LVSD, with theemainder experiencing a putative nonischemic etiology ofVSD (6). In addition, HF occurs at an earlier age infrican-American patients, with more advanced LVSD andorse clinical class at the time of diagnosis than seen in theopulation at large (7).Despite being at high risk for cardiovascular morbidity

nd mortality, prior studies have indicated that Africanmericans are less likely than non-African Americans to

eceive guideline-recommended, evidence-based care suchs diagnostic testing, drug therapy, and interventional pro-edures (8–10). The reasons for these disparities in healthare are multifactorial, however, and might be related toccess to care, patient preference, or bias (11). Recenttudies also document that African-American patients areore likely to be treated at hospitals that provide poor

uality of care with lower use of systematic processes,rotocols, and evidence-based treatments (12–14). More-ver, even when treated at the same hospital, African-merican patients might not receive the same standard of

are delivered to white patients (12,15).The quality of care for African Americans admitted withF has not been well studied in the context of an HF

uality improvement program. The OPTIMIZE-HF (Or-anized Program to Initiate Lifesaving Treatment in Hos-

Abbreviationsand Acronyms

ACC/AHA � AmericanCollege of Cardiology/American Heart Association

ACE � angiotensin-converting enzyme

ARB � angiotensinreceptor blocker

BNP � B-type natriureticpeptide

CI � confidence interval

HF � heart failure

LVEF � left ventricularejection fraction

LVSD � left ventricularsystolic dysfunction

OR � odds ratio

italized Patients with Heart Failure) registry provides an e

pportunity to capture quality indicators as a function oface and investigate their association with outcomes. ThePTIMIZE-HF program is designed to improve medical

are and education for hospitalized HF patients and toccelerate use of evidence-based, guideline-recommendedherapies by initiating them before hospital discharge.hese process-of-care intervention strategies might stan-ardize care for all patients with HF and test whetherfrican Americans with HF who are treated similarly will

xperience the same outcomes as other HF populations16). Our analyses investigated the influence of this quality-mprovement initiative in a large African-American subgroupnd the association with quality of care, post-discharge HFanagement, survival, and repeat hospital stay.

ethods

tudy design. The study design and rationale of thePTIMIZE-HF registry has been previously published

17–20). Briefly, OPTIMIZE-HF is a registry anderformance-improvement program for patients hospital-zed with HF. This nationwide patient registry was used toather data on various patient characteristics with a web-ased information system. Participating hospitals had thebility to view national aggregate hospital data as well asaily patient and performance data benchmarked withimilar hospitals. The OPTIMIZE-HF process-of-caremprovement program provided participating hospitals with

aterials for improving treatment and discharge plans forptimal patient management and included evidence-basedest-practice algorithms (detailed algorithms on indications,ontraindications, dosing, and monitoring steps for eachvidence-based HF therapy) along with comprehensiveatient education materials and resources. To be eligible forPTIMIZE-HF, patients had to be adults hospitalized for

n episode of HF as the primary cause of admission or withignificant HF symptoms that developed during hospitaltay with a primary discharge diagnosis of HF. Withethodology similar to national cardiovascular registries

nd many randomized trials, race and ethnicity were col-ected for the purpose of evaluating subgroup differences.dmission and/or medical staff recorded race/ethnicity,sually as the patient was registered. Patients were assignedo race and ethnicity categories with options defined by thetudy protocol. A pre-specified subgroup (10%) had 60- to0-day follow-up data collected. Participating sites had theption of participating in the hospital registry only orospital registry and follow-up data collection. Site partic-

pation in the follow-up portion required obtaining in-ormed consent. This follow-up cohort was demographi-ally similar to patients in the overall registry (Online Table). Process-of-care improvement tool use was defined by usef either preprinted order sets or discharge checklists asocumented in the medical record. Left ventricular systolicysfunction was defined as a documented left ventricular

jection fraction (LVEF) �40% or qualitative assessment of

mcssSnoimsfAaccv

hpdactvdpmovmcO

r dysfunlood p

1677JACC Vol. 51, No. 17, 2008 Yancy et al.April 29, 2008:1675–84 African Americans Hospitalized With HF: The OPTIMIZE-HF Registry

oderate/severe dysfunction. The registry coordinatingenter was Outcome Sciences, Inc. (Cambridge, Massachu-etts), which provided data collection, management, andource data verification.tatistical analyses. Discrete data are reported as theumber and frequency of eligible patients treated at the timef hospital discharge, excluding patients with documentedntolerance/contraindications to specific therapies. The HF

edication use and contraindications/intolerance were as-essed at discharge from the index hospital stay and duringollow-up. Race was analyzed according to African-merican and non–African-American status. Patient char-

cteristics and evidence-based treatments at hospital dis-harge were compared with the Pearson chi-square test forategorical variables and analysis of variance for continuous

Patient Characteristics

Table 1 Patient Characteristics

CharacteristicAf

Age, mean � SD, yrs

Female, %

Hypertensive etiology, %

Ischemic etiology, %

Atrial arrhythmia, %

Patients with LVF assessed, n (%) 7,3

LVEF (in those measured), mean � SD, %

LVSD (LVEF �40% or moderate/severe LVD), %

Left heart catheterization, %

CABG, %

Dialysis, %

CAD/IHD, %

Hyperlipidemia, %

Insulin-treated diabetes, %

Non–insulin-treated diabetes, %

COPD, %

Depression, %

Cigarette smoker within past yr, %

Serum creatinine, mean � SD, mg/dl

Serum creatinine, mean � SD, �mol/l

Serum creatinine �2 mg/dl, %

Hemoglobin, mean � SD, g/dl

Hemoglobin, mean � SD, g/l

Patients with BNP measured, n (%) 4,3

BNP, median (25th, 75th), pg/ml 9

Troponin I, median (25th, 75th), ng/ml

Troponin I, median (25th, 75th), �g/l

Admission weight, mean � SD, kg

Weight change from admission, mean � SD, kg �

Admission SBP, mean � SD, mm Hg 1

Discharge SBP, mean � SD, mm Hg 1

Admission HR, mean � SD, beats/min

Discharge HR, mean � SD, beats/min

Admission jugular venous pressure elevation, %

Admission rales, %

BNP � B-type natriuretic peptide; CABG � coronary artery bypass graft;obstructive pulmonary disease; HR � heart rate; LVD � left ventriculafunction; LVSD � left ventricular systolic dysfunction; SBP � systolic b

ariables. Models of in-hospital mortality, mortality from f

ospital discharge to 90 days, and the combination ofost-discharge mortality or repeat hospital stay have beeneveloped to determine significant factors to use whenpplying adjusted models (Online Table 2) (19,20). Baselinelinical and treatment factors were applied to model selec-ions. To model in-hospital mortality, 45 potential predictorariables were used in a logistic model. To model post-ischarge mortality in the follow-up period, 19 potentialredictor variables were used in a Cox proportional hazardsodel. To model repeat hospital stay and the combination

f post-discharge mortality or repeat hospital stay, 69ariables were used in a logistic model. To model confor-ity with quality of care measures, 69 variables were

onsidered, and those retained for each model are listed innline Table 2. A p value of 0.05 was used as the criterion

merican,608)

Non-African American(n � 38,581) p Value

� 15.4 75.2 � 12.7 �0.0001

.7 51.4 0.0287

.2 19.3 �0.0001

.5 49.4 �0.0001

.3 33.8 �0.0001

.0) 32,648 (84.6) 0.002

� 17.8 39.7 � 17.5 �0.0001

.9 47.1 �0.0001

.1 8.5 �0.0001

.4 1.0 �0.0001

.0 4.2 �0.0001

.6 52.2 �0.0001

.7 33.8 �0.0001

.4 16.2 �0.0001

.7 24.7 0.0345

.6 28.9 �0.0001

.4 11.4 �0.0001

.3 14.2 �0.0001

� 1.3 1.6 � 1.0 �0.0001

� 113.8 141.8 � 90.4 �0.0001

.7 19.2 �0.0001

� 2.1 12.1 � 2.0 �0.0001

� 20.5 121.2 � 20.4 �0.0001

.5) 26,086 (67.6) �0.0001

0, 2,130) 785 (403, 1,600) �0.0001

.1, 0.5) 0.1 (0.0, 0.3) �0.0001

1, 0.5) 0.1 (0.0, 0.3) �0.0001

� 29.99 80.79 � 25.03 �0.0001

� 5.33 �2.53 � 4.67 0.4332

� 35.98 140.34 � 31.73 �0.0001

� 23.37 123.74 � 22.10 �0.0001

� 20.85 85.76 � 21.48 �0.0001

� 14.15 75.62 � 14.12 �0.0001

.9 31.6 �0.0001

.7 64.0 0.5493

D � coronary artery disease/ischemic heart disease; COPD � chronicction; LVEF � left ventricular ejection fraction; LVF � left ventricular

ressure.

rican A(n � 8

63.6

52

39

29

16

99 (86

35.4

56

10

0

9

37

24

19

25

21

6

27

1.8

156.5

21

12.0

120.1

49 (50

65 (45

0.2 (0

0.2 (0.

90.67

2.59

53.06

29.27

90.27

77.73

36

63

CAD/IH

or variables to remain in the model. The assumption of

lvttrpiaiCwR

R

Pr2p5OdPtfAaiuisAWm0lEfiAeptTApsAQmdtdtph

(Laa0b0AAdttpAAHAss00sspw(0m0t9

w

1678 Yancy et al. JACC Vol. 51, No. 17, 2008African Americans Hospitalized With HF: The OPTIMIZE-HF Registry April 29, 2008:1675–84

inearity was checked in each model for the continuousariables by using restricted cubic splines. When the rela-ionship was found to be nonlinear, appropriate transforma-ions were applied. When the association between race andepeat hospital stay was examined, all variables from theost-discharge death or repeat hospital stay model werencluded. Generalized estimating equations were used toccount for the correlation of data within the same hospitaln the adjusted models. The SAS version 8.2 (SAS Institute,ary, North Carolina) was used for all statistical analyses,hich were performed independently at Duke Clinicalesearch Institute (Durham, North Carolina).

esults

atient characteristics. The OPTIMIZE-HF registry en-olled patients from March 1, 2003 through December 31,004 and included 48,612 patients from 259 hospitals, withre-specified follow-up data available from 91 centers and,791 patients. Race was available for 47,190 (97.1%) patients.f these, 8,608 (17.7%) were African Americans; follow-up

ata were collected from 1,044 African-American patients.atient characteristics are detailed in Table 1. The charac-

eristics of African Americans hospitalized for HF differedrom those of non–African Americans in the registry.frican-American patients tended to be female and younger

nd were more likely to have systolic dysfunction and renalnsufficiency. Systolic blood pressure (SBP), B-type natri-retic peptide (BNP), and admission weight were all highern hospitalized African Americans. Additionally, the pre-umed etiology of HF differed significantly betweenfrican-American and non–African-American patients.hereas HF etiology in African-American patients wasore likely to be hypertensive (39.2% vs. 19.3%; p �

.0001), in non–African-American patients it was moreikely to be ischemic (49.4% vs. 29.5%; p � 0.0001).tiology attributable to other causes (postpartum, valvular,

amilial, alcohol/other drug, chemotherapy, viral, unknown/diopathic, and other) did not differ between African-merican and non–African-American patients (31.3% for

ach group). Importantly, 70.5% of the African-Americanatients with HF had a nonischemic etiology, either relatedo hypertension or causes other than ischemic heart disease.he mean LVEF was significantly lower in African-merican patients (35.4% vs. 39.7%; p � 0.0001), and theercentage of African-American patients with LVSD wasignificantly higher than that observed in non–African-merican patients (56.9% vs. 47.1%; p � 0.0001).uality of care. Comparison of the use of process-of-careeasures between racial populations analyzed in this study

emonstrated significant differences. African-American pa-ients were significantly less likely to receive completeischarge instructions and smoking cessation advice, buthey were more likely to receive LVEF assessment and arescription for angiotensin-converting enzyme (ACE) in-

ibitors in the absence of contraindications and intolerance c

Fig. 1). At discharge, eligible African Americans withVSD were more likely to be prescribed ACE inhibitors orngiotensin receptor blocker drugs (ARBs) (p � 0.0001),ldosterone antagonists (p � 0.0001), and hydralazine (p �.0001) but just as likely as other patients to be prescribedeta-blocker drugs, warfarin, and nitrates (p � 0.225, p �.118, and p � 0.674, respectively) (Fig. 2). Africanmericans were less likely to receive statin therapy. Whereasfrican Americans were more likely to be prescribed hy-ralazine, the percentage of African Americans with LVSDhat received a combination of hydralazine/isosorbide dini-rate was only 4.5%. There was documented use of therocess-of-care improvement tools in 39.5% of African-merican patients compared with 46.9% non–African-merican patients (p � 0.0001).ospital stay and mortality. In unadjusted analyses,frican-American HF patients experienced a modestly

horter length of stay (5.58 days vs. 5.71 days; p � 0.001),ignificantly lower in-hospital mortality (2.2% vs. 4.1%; p �.0001), and lower follow-up mortality (6.5% vs. 9.1%; p �.009) (Table 2). However, this patient population hadimilar readmission events (31.7% vs. 29.3%; p � 0.12) andimilar death/repeat hospital stay events (35.3% vs. 36.6%;� 0.44). In multivariable analyses, African-American raceas an independent predictor of lower in-hospital mortality

odds ratio [OR] 0.71, 95% confidence interval [CI] 0.57 to.87, p � 0.001) but was not predictive of follow-uportality (hazard ratio 1.12, 95% CI 0.80 to 1.58, p �

.50), follow-up repeat hospital stay (OR 1.14, 95% CI 0.93o 1.40, p � 0.22), or death/repeat hospital stay (OR 1.02,5% CI 0.86 to 1.22, p � 0.79) (Table 2).A subgroup analysis revealed African-American patients

ith LVSD experienced a shorter length of stay, signifi-

Figure 1 Heart Failure Measures at Hospital Discharge

Percent of patients who received heart failure measures at hospital dischargeby race. ACE � angiotensin-converting enzyme; LVEF � left ventricular ejectionfraction.

antly lower in-hospital mortality, and similar follow-up

mAdhseaRdawAqAhti

D

TpwfsOdorAmd

od

C

*

1679JACC Vol. 51, No. 17, 2008 Yancy et al.April 29, 2008:1675–84 African Americans Hospitalized With HF: The OPTIMIZE-HF Registry

ortality but more readmission events (Figs. 3A and 3B).nalysis of African-American patients without systolicysfunction revealed that this group also had lower in-ospital mortality but similar length of stay, repeat hospitaltay, and post-discharge mortality compared with thatxperienced by non–African-American patients (Figs. 4And 4B).ace as an independent predictor of quality of care. Ad-itional multivariable analyses were performed to assess thessociation between African-American race and conformityith the American College of Cardiology/American Heartssociation (ACC/AHA) performance measures and otheruality indicators for patients hospitalized with HF.frican-American race was independently associated withigher ACE inhibitor prescription and left ventricular func-ion assessment; all other HF quality indicators were notnfluenced by race (Table 3).

Figure 2 Other Medications at Hospital Discharge

Eligible patients receiving other medications at hospital discharge by race. �ACE inexcluding patients with contraindications to ACE inhibitor and/or ARB. †Beta-blockdrugs. ‡Statin use in patients with medical history of coronary artery disease, cereeral vascular disease. §Aldosterone antagonist, hydralazine, and nitrate use in paACE � angiotensin-converting enzyme; ARB � angiotensin receptor blocker; HYDR

linical Outcomes for African-American Compared With Non–Africa

Table 2 Clinical Outcomes for African-American Compared With

Hospital Cohort

Unadju

Odds, Hazard, orLength-of-Stay

Ratio95% C

In

In-hospital mortality 0.49* 0.4

Length of stay 0.97† 0.9

Follow-up cohort

Post-discharge mortality 0.76‡ 0.5

Repeat hospital stay 1.13* 0.9

Post-discharge mortality/repeat hospital stay 0.96* 0.8

Odds ratio. †Length-of-stay ratio. ‡Hazard ratio.

iscussion

he OPTIMIZE-HF is a registry-based performance im-rovement program that enables quality care for patientsith HF through performance improvement tools and

eedback on patient characteristics, treatments, quality mea-ures, and clinical events. Our findings from thisPTIMIZE-HF analysis of African Americans with HF

emonstrate that, among participating hospitals, provisionf most but not all quality measures were similar to otheraces. Furthermore, when treated similarly, African-merican patients have a lower risk-adjusted inpatientortality but a similar adjusted 60- to 90-day post-

ischarge morbidity and mortality risk due to HF.Process-of-care quality-improvement strategies represent

ne potential mechanism for addressing and reversing racialisparities in cardiovascular care and outcomes (21–23). The

r and/or ARB use in patients with left ventricular systolic dysfunction (LVSD),in patients with LVSD, excluding patients with contraindications to beta-blocker

ascular accident/transient ischemic attack, diabetes, hyperlipidemia, or periph-with LVSD. �Warfarin use in patients with chronic or paroxysmal atrial fibrillation.� hydralazine/isosorbide dinitrate.

erican Patients

–African-American Patients

Adjusted

encel p Value

Odds, Hazard, orLength-of-Stay

Ratio95% Confidence

Interval p Value

9 �0.0001 0.71* 0.57–0.87 �0.001

9 0.002 0.98† 0.96–1.01 0.193

9 0.041 1.12‡ 0.80–1.58 0.508

2 0.146 1.14* 0.93–1.40 0.215

2 0.591 1.02* 0.86–1.22 0.789

hibitoer usebral v

tients/ISDN

n-Am

Non

sted

onfidterva

1–0.5

6–0.9

8–0.9

6–1.3

2–1.1

rimowpdfidtodOmh

sh

creitAamccetopHerpd

1680 Yancy et al. JACC Vol. 51, No. 17, 2008African Americans Hospitalized With HF: The OPTIMIZE-HF Registry April 29, 2008:1675–84

esults of the current study suggest that process-of-carentervention programs might benefit African-American as

uch as non–African-American patients, with similar usef certain evidence-based therapies for HF. Importantly,ithin the context of a performance-improvement program,atient race did not influence the quality of HF careelivered. This is an important observation in light of recentndings suggesting that patients with HF or acute myocar-ial infarction treated at hospitals that disproportionatelyreat African-American patients are at high risk for adverseutcomes and lower quality of care (24,25). Although notesigned to improve disparate health care, it is plausible thatPTIMIZE-HF or other similar performance improve-ent programs might represent a unique tool to address

ealth care disparities in HF.Given a higher cardiovascular risk profile as measured by

everal clinical variables, including a lower mean LVEF,

Figure 3 Outcomes in Patients With LeftVentricular Systolic Dysfunction

(A) In-hospital outcomes in heart failure patients with left ventricular systolicdysfunction (LVSD) by race. (B) The 60- to 90-day post-discharge outcomes inheart failure patients with LVSD by race.

igher BNP levels, and greater incidence of renal insuffi-

iency, African-American patients in the OPTIMIZE-HFegistry might have been expected to fare less well. How-ver, African Americans with HF had lower rates ofn-hospital mortality, regardless of systolic function. Al-hough this might reflect the younger age of African-merican patients, these findings persisted after multivari-

ble adjustment, although the OR for in-hospital mortalityoved from 0.49 to 0.71 after adjustment for observed

ovariates. Residual confounding by measured variables andonfounding by unmeasured variables must still be consid-red in accounting for these observations. It is also notablehat African-American patients hospitalized with HF weren average 11 years younger than non–African-Americanatients. This finding underscores the need for increasedF prevention efforts in this population. Uncertain differ-

nces in the pathophysiology of HF in African Americansemain a potential explanation for these differences inresentation and are perhaps driven by subtle biologicalifferences including certain genomic variances (26).

Figure 4 Outcomes in Patients Without LVSD

(A) In-hospital outcomes in heart failure patients without left ventricular sys-tolic dysfunction (LVSD) by race. (B) The 60- to 90-day post-discharge out-comes in heart failure patients without LVSD by race.

AhVAhief2AHsctysc

eOAeetaApdpaaH

gAcmnboo

abIRiataAeO(bAtbb4wcihewGfASrwbDdarrorcshfdmddmmmAfOwpm

Io

A

1681JACC Vol. 51, No. 17, 2008 Yancy et al.April 29, 2008:1675–84 African Americans Hospitalized With HF: The OPTIMIZE-HF Registry

Prior studies have shown higher rates of hospital stay infrican Americans with HF, but the data on mortality ratesave varied. Analysis of the SOLVD (Studies of Leftentricular Dysfunction) treatment trial suggested thatfrican Americans with chronic systolic dysfunction HFave increased mortality risk (5). Deswal et al. (27), exam-

ning 4,901 African-American and 17,093 Caucasian Vet-rans Affairs patients discharged from the hospital for HF,ound a lower risk-adjusted OR for mortality at 30 days andyears in African Americans with similar readmission rates.study of 29,372 Medicare beneficiaries hospitalized withF in 1998 and 1999 also revealed higher repeat hospital

tay rates but lower mortality in African-American patientsompared with Caucasians (28). Because a substantial propor-ion of African-American patients hospitalized with HF areounger than 65 years of age, Medicare datasets might not beufficient to characterize the demographic status, quality ofare, and outcomes for this patient population.

The adequacy of medical therapy is often impugned as anxplanation for racially disparate outcomes. In thePTIMIZE-HF registry the differences between African-merican and non–African-American patients cannot be

xplained by medication prescription, because appropriatevidence-based therapies were prescribed at similar levels athe time of hospital discharge. Actual medication compli-nce, however, might have impacted repeat hospital stays.frican-American patients were less likely to receive com-lete discharge instructions and smoking cessation advice atischarge. Several studies have shown that pre-dischargerograms that enhance patient education and consultationnd conduct follow-up reminders for medical adherence aressociated with fewer readmissions and with reductions inF symptoms (29–31).The OPTIMIZE-HF performance-improvement pro-

ram was associated with remarkably high ACE inhibitor/RB and beta-blocker prescription rates at hospital dis-

harge, regardless of race. Notably, African Americans wereore likely to be prescribed ACE inhibitors and ARBs than

on–African Americans in this registry, and this might haveeen responsible in part for better inpatient and short-termutcomes in African Americans with HF than previously

ndependent Associationf African-American Race and Quality of Care

Table 3 Independent Associationof African-American Race and Quality of Care

Performance Measures andOther Quality Indicators

OddsRatio

95% ConfidenceInterval p Value

HF-1: delivery of HF dischargeinstructions

1.02 0.94–1.10 0.701

HF-2: left ventricular functionassessment

1.19 1.05–1.34 0.007

HF-3: ACE inhibitor at discharge 1.18 1.01–1.39 0.039

HF-4: smoking cessationcounseling

0.87 0.75–1.02 0.093

ACE inhibitor or ARB at discharge 1.16 0.97–1.39 0.104

Beta-blocker at discharge 0.89 0.71–1.11 0.292

CE � angiotensin-converting enzyme; ARB � angiotensin receptor blocker; HF � heart failure.

bserved. In addition to ACE inhibitor usage, clinical trials p

nd other registries have shown the importance of beta-locker use in the African-American population with HF.n the community-based COHERE (Carvedilol Heart Failureegistry), carvedilol treatment produced similar results, reduc-

ng hospital stays, HF, and mortality in African-Americannd Caucasian patients (32). The most recent update ofhe ACC/AHA clinical practice guidelines for the man-gement of chronic HF recommends that the African-merican population receive the same application of

vidence-based therapies as the general population (33).ur findings would support that statement. The A-HeFT

African-American Heart Failure Trial) study tested isosor-ide dinitrate/hydralazine as adjunctive therapy for HF infrican Americans and demonstrated a substantial 43% mor-

ality benefit and a 33% reduction in hospital stays for HF overackground therapy with ACE inhibitor/ARB and beta-locker drugs (6). Despite these substantial benefits, only.5% of African Americans in the OPTIMIZE-HF registryere treated with isosorbide dinitrate/hydralazine. Re-

ent guidelines from the Heart Failure Society of Amer-ca give the adjunctive use of isosorbide dinitrate/ydralazine the highest-tier recommendation, withxpectations that outcomes for African-American patientsill improve with adherence to guideline-based care (34).iven the poor adoption of isosorbide dinitrate/hydralazine,

urther opportunities remain to improve care for African-merican HF patients (6).tudy limitations. Although the OPTIMIZE-HF registryepresents an opportunity to study HF patients in a real-orld setting, there are several limitations to a registry-ased study that call for careful interpretation of results.ata were collected by medical chart review and are depen-

ent upon the accuracy and completeness of documentationnd abstraction. Race was not a self-reported variable butather was determined as that documented in the medicalecord, thus errors in racial determination could haveccurred. Because of the large number of patients in theegistry, some small differences that might be of littlelinical relevance have p values indicating a high degree oftatistical significance. These findings might not apply toospitals that differ in patient characteristics or care patternsrom OPTIMIZE-HF hospitals. The 60- to 90-day follow-upata are only in a subset of patients in the overall registry andight not be representative of the entire aggregate in the

atabase. Medication use was as reported by patients and asocumented in the medical record. Actual adherence ratesight have been lower than reported, and compliance withedical therapy is not a certainty. Decreased compliance withedical therapy due to access of care issues affecting thefrican-American cohort might well have been a contributing

actor to the readmission experiences noted in thePTIMIZE-HF registry. Contraindications and intoleranceere as documented in the medical record, but a proportion ofatients reported to be eligible for treatment but not treatedight have had contraindications or intolerance that were

resent but not documented. There were no direct measures of

schTmcri

C

AcsAgnrcmphltptmf

A

DfIAvbap(GrnCrJhSNPtIMICm

IfKreCECTHfSNlmJMtSsGDrKtScGHGLiEasodtCJJrnOSGCSScMSCor

1682 Yancy et al. JACC Vol. 51, No. 17, 2008African Americans Hospitalized With HF: The OPTIMIZE-HF Registry April 29, 2008:1675–84

ocioeconomic status or bias in OPTIMIZE-HF, thus theontribution of certain elements affecting similar or disparateealth care can only be inferred but not proven with these data.his study was not a prospective randomized trial, and residualeasured and unmeasured confounders might have influenced

linical outcomes. Prospective randomized studies will beequired to further test the influence of performance-mprovement programs on markers of health care disparities.

onclusions

frican-American HF patients, when exposed to a process-of-are improvement initiative, had better-than-previously ob-erved treatment with evidence-based therapies. Furthermore,frican-American HF patients when treated according to

uidelines had similar or better outcomes compared withon–African-American patients. In adjusted analyses, patientace did not influence hospital-based delivery of high-qualityare, as defined by current ACC/AHA HF performanceeasures. Further research is required to test the benefit of a

erformance-improvement program on narrowing evidence ofealth care disparities in African Americans with cardiovascu-

ar disease. In the interim, all appropriate evidence-basedherapies should be employed to improve outcomes in allatients with HF. The OPTIMIZE-HF program suggestshat an in-hospital process-of-care improvement programight help to achieve similar conformity with quality measures

or African Americans with HF.

uthor Disclosures

r. Yancy reported that he has received research grantsrom GlaxoSmithKline, Medtronic, NitroMed, and Scios,nc. He is also a consultant or on the Speakers’ Bureau forstraZeneca, GlaxoSmithKline, Medtronic, NitroMed, No-

artis, and Scios, Inc. He was previously on the advisoryoard for CHF Solutions. He currently serves on the Foodnd Drug Administration (FDA) cardiovascular deviceanel and study section for the National Institutes of HealthNIH). He has received honoraria from AstraZeneca,laxoSmithKline, Medtronic, Novartis, and Scios, Inc. He

eports editorial board involvement for American Heart Jour-al, American Journal of Cardiology (Associate Editor),irculation (Guest Editor), Congestive Heart Failure, Cur-

ent Heart Failure Reports, Journal of Acute Cardiac Care, andournal of Urban Cardiology. Dr. Abraham reported that heas received a research grant from Amgen, Biotronik, CHFolutions, GlaxoSmithKline, HFSA, Medtronic, Myogen,IH, Orqis Medical, Otsuka Maryland Research Institute,aracor, and Scios, Inc. He is/has been a consultant/is on

he Speakers’ Bureau for Amgen, AstraZeneca, Boehringer-ngelheim, CHF Solutions, GlaxoSmithKline, Guidant,

edtronic, Merck, Pfizer, ResMed, Respironics, Scios,nc., and St. Jude Medical. He is on the advisory board ofardioKine, CardioKinetix Inc., CHF Solutions, Depart-

ent of Veterans Affairs Cooperative Studies Program, S

novise, NIH, and Savacor, Inc. He has received honorariarom AstraZeneca, Boehringer-Ingelheim, GlaxoSmith-line, Guidant, Medtronic, Merck, Pfizer, ResMed, Respi-

onics, Scios, Inc., and St. Jude Medical. He reportsditorial board involvement with Congestive Heart Failure,urrent Cardiology Reviews, Current Heart Failure Reports,xpert Review of Cardiovascular Therapy, Journal Watchardiology, PACE—Pacing and Clinical Electrophysiology,he American Heart Hospital Journal, and The Journal ofeart Failure. Dr. Albert reported that she is a consultant

or GlaxoSmithKline and Medtronic. She is also on thepeakers’ Bureau for GlaxoSmithKline, Medtronic,itroMed, and Scios, Inc., and is employed by the Cleve-

and Clinic Foundation. She reports editorial board involve-ent for Progress in Cardiovascular Nursing (Senior Editor),

ournal of Cardiovascular Nursing, and Critical Care Nurse.r. Clare is an employee of DCRI. Dr. Stough reported

hat she has received research grants from Actelion, Glaxo-mithKline, Medtronic, Otsuka, and Pfizer. She is a con-ultant or on the Speakers’ Bureau for Abbott, AstraZeneca,laxoSmithKline, Medtronic, Novacardia, Otsuka, Proteinesign Labs, RenaMed, Sigma Tau, and Scios, Inc. She has

eceived honoraria from Abbott, AstraZeneca, GlaxoSmith-line, Medtronic, and Pfizer. Dr. Gheorghiade reported

hat he has received research grants from the NIH, Otsuka,igma Tau, Merck, and Scios, Inc. He is/has been aonsultant for Debbio Pharm, Errekappa Terapeutici,laxoSmithKline, Protein Design Labs, and Medtronic.e has received honoraria from Abbott, AstraZeneca,laxoSmithKline, Medtronic, Otsuka, Protein Designabs, Scios, Inc., and Sigma Tau. He reports editorial board

nvolvement with Acute Cardiac Care Journal (Associateditor), American Heart Journal, American Journal of Ther-

peutics (Associate Editor), Archives for Chest Disease (As-ociate Editor), Current Cardiology Reviews, Expert Reviewf Cardiovascular Therapy, Heart Disease: A Journal of Car-iovascular Medicine, Heart Failure Reviews, Heart Interna-ional, Journal of Cardiac Failure, Journal of the Americanollege of Cardiology, Italian Heart Journal, The American

ournal of Cardiology, The Journal of Heart Disease, and Theournal of Heart Failure. Dr. Greenberg reported that he haseceived research grant support from Amgen, Cardiody-amics, GlaxoSmithKline, Millennium, Novacardia,tsuka, Pfizer, Sanofi-Aventis, and Titan. He is on thepeakers’ Bureau/is a consultant for Amgen, AstraZeneca,laxoSmithKline, Guidant Corp., Medtronic, Merck &o., NitroMed, Pfizer, Remon Medical Technologies, andcios, Inc. He has served on advisory boards for CHFolutions, GlaxoSmithKline, and NitroMed. He has re-eived honoraria from AstraZeneca, GlaxoSmithKline,

edtronic, Merck & Co., NitroMed, Novartis, Pfizer, andcios, Inc. He reports editorial board involvement forongestive Heart Failure and Journal of the American College

f Cardiology. Dr. O’Connor reported that he has receivedesearch grant support from the NIH. He is on the

peakers’ Bureau and/or a consultant for Amgen, Astra-

ZMaSDrAsNcBtMbMHaoCGtFAGStySMSWMKPWiJC

RBCE

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

1683JACC Vol. 51, No. 17, 2008 Yancy et al.April 29, 2008:1675–84 African Americans Hospitalized With HF: The OPTIMIZE-HF Registry

eneca, Bristol-Myers Squibb, GlaxoSmithKline, Guidant,edtronic, Merck, NitroMed, Novartis, Otsuka, Pfizer,

nd Scios, Inc. He has received honoraria from Glaxo-mithKline, Pfizer, and Otsuka. Dr. She is an employee ofCRI. Ms. Sun is an employee of DCRI. Dr. Young

eported that he has received research grants from Abbott,corn, Amgen, Artesion Therapeutics, AstraZeneca, Bio-

ite, GlaxoSmithKline, Guidant, Medtronic, MicroMed,IH, Scios, Inc., Vasogen, and World Heart. He is a

onsultant for Abbott, Acorn, Amgen, Biomax Canada,iosite, Boehringer-Ingelheim, Bristol-Myers Squibb, Co-

herix, Edwards Lifescience, GlaxoSmithKline, Guidant,edtronic, MicroMed, Novartis, Paracor, Proctor & Gam-

le, Protemix, Scios, Inc., Sunshine, Thoratec, Transworldedical Corporation, Vasogen, Viacor, and World Heart.e reports editorial board involvement for Journal of Heart

nd Lung Transplantation, Evidence-Based Medicine, Journalf the American College of Cardiology, American Heart Journal,leveland Clinic Journal of Medicine, Cardiology Today,raft, TheHeart.org, Transplantation and Immunology Let-

er, and American Society of Transplantation Newsletter. Dr.onarow reported that he has received research grants frommgen, Biosite, Bristol-Myers Squibb, Boston Scientific/uidant, GlaxoSmithKline, Medtronic, Merck, Pfizer,

anofi-Aventis, Scios, Inc., and the NIH. He is/has been onhe Speakers’ Bureau or has received honoraria in the past 5ears from Amgen, AstraZeneca, Biosite, Bristol-Myersquibb, Boston Scientific/Guidant, GlaxoSmithKline, Kos,edtronic, Merck, NitroMed, Pfizer, Sanofi-Aventis,

chering-Plough, Scios, Inc., St. Jude Medical, Takeda, andyeth. He is or has been a consultant for Biosite, Bristol-yers Squibb, Boston Scientific/Guidant, GlaxoSmith-

line, Medtronic, Merck, NitroMed, Orqis Medical,fizer, Sanofi-Aventis, Schering-Plough, Scios, Inc., andyeth. He reports editorial board involvement with Amer-

can Heart Journal, Circulation, Journal of Cardiac Failure,ournal of the American College of Cardiology, and Reviews ofardiovascular Medicine.

eprint requests and correspondence: Dr. Clyde W. Yancy,aylor Heart and Vascular Institute, Baylor University Medicalenter, 3500 Gaston Avenue, Suite H-030, Dallas, Texas 75246.-mail: clydey@baylorhealth.edu.

EFERENCES

1. Ofili E, Igho-Pemu P, Bransford T. The prevention of cardiovasculardisease in blacks. Curr Opin Cardiol 1999;14:169–75.

2. Mensah GA, Mokdad AH, Ford ES, Greenlund KJ, Croft JB. Stateof disparities in cardiovascular health in the United States. Circulation2005;111:1233–41.

3. American Heart Association. African Americans and CardiovascularDiseases—Statistics. Dallas, TX: American Heart Association, 2005.

4. American Heart Association. Heart Disease and Stroke Statistics—

2007 Update. Dallas, TX: American Heart Association, 2007.

5. Dries DL, Exner DV, Gersh BJ, Cooper HA, Carson PE, DomanskiMJ. Racial differences in the outcome of left ventricular dysfunction.N Engl J Med 1999;340:609–16.

6. Taylor AL, Ziesche S, Yancy C, et al. Combination of isosorbidedinitrate and hydralazine in blacks with heart failure. N Engl J Med2004;351:2049–57.

7. Yancy CW. Heart failure in African Americans: a cardiovascularengima. J Card Fail 2000;6:183–6.

8. Davis SK, Liu Y, Gibbons GH. Disparities in trends of hospitalizationfor potentially preventable chronic conditions among African Ameri-cans during the 1990s: implications and benchmarks. Am J PublicHealth 2003;93:447–55.

9. Manhapra A, Canto JG, Barron HV, et al. Underutilization ofreperfusion therapy in eligible African Americans with acute myocar-dial infarction: role of presentation and evaluation characteristics. AmHeart J 2001;142:604–10.

0. Bertoni AG, Goonan KL, Bonds DE, Whitt MC, Goff DC Jr.,Brancati FL. Racial and ethnic disparities in cardiac catheterization foracute myocardial infarction in the United States, 1995–2001. J NatlMed Assoc 2005;97:317–24.

1. Yancy CW, Sica DA. Cardiovascular disease in African Americans.J Clin Hypertens (Greenwich) 2004;6:54–6.

2. Kressin NR. Separate but not equal: the consequences of segregatedhealth care. Circulation 2005;112:2582–4.

3. Konety SH, Vaughan Sarrazin MS, Rosenthal GE. Patient andhospital differences underlying racial variation in outcomes aftercoronary artery bypass graft surgery. Circulation 2005;111:1210–6.

4. Bradley EH, Herrin J, Wang Y, et al. Racial and ethnic differences intime to acute reperfusion therapy for patients hospitalized withmyocardial infarction. JAMA 2004;292:1563–72.

5. Kressin NR, Petersen LA. Racial differences in the use of invasivecardiovascular procedures: review of the literature and prescription forfuture research. Ann Intern Med 2001;135:352–66.

6. Yancy CW. Heart failure therapy in special populations: the same ordifferent? Rev Cardiovasc Med 2004;5 Suppl 1:S28–35.

7. Fonarow GC, Abraham WT, Albert NM, et al. Organized Programto Initiate Lifesaving Treatment in Hospitalized Patients with HeartFailure (OPTIMIZE-HF): rationale and design. Am Heart J 2004;148:43–51.

8. Fonarow GC, Abraham WT, Albert NM, et al. Influence of aperformance-improvement initiative on quality of care for patientshospitalized with heart failure: results of the Organized Program toInitiate Lifesaving Treatment in Hospitalized Patients With HeartFailure (OPTIMIZE-HF). Arch Intern Med 2007;167:1493–502.

9. Gheorghiade M, Abraham WT, Albert NM, et al. Systolic bloodpressure at admission, clinical characteristics, and outcomes in patientshospitalized with acute heart failure. JAMA 2006;296:2217–26.

0. Fonarow GC, Abraham WT, Albert NM, et al. Association betweenperformance measures and clinical outcomes for patients hospitalizedwith heart failure. JAMA 2007;297:61–70.

1. Fiscella K, Franks P, Gold MR, Clancy CM. Inequality in quality:addressing socioeconomic, racial, and ethnic disparities in health care.JAMA 2000;283:2579–84.

2. Lavizzo-Mourey R, Jung M. Fighting unequal treatment: the RobertWood Johnson Foundation and a quality-improvement approach todisparities. Circulation 2005;111:1208–9.

3. Watson SD. Equity Measures and Systems Reform as Tools forReducing Racial and Ethnic Disparities in Health Care. New York:The Commonwealth Fund, 2005.

4. Mehta SK, Toto KH, Nelson LL, Drazner MH. Therapy of heartfailure in African Americans: lessons from an urban public hospital.Congest Heart Fail 2004;10:40–3.

5. Skinner J, Chandra A, Staiger D, Lee J, McClellan M. Mortality afteracute myocardial infarction in hospitals that disproportionately treatblack patients. Circulation 2005;112:2634–41.

6. McNamara DM, Tam SW, Sabolinski ML, et al. Aldosteronesynthase promoter polymorphism predicts outcome in African Amer-icans with heart failure: results from the A-HeFT Trial. J Am CollCardiol 2006;48:1277–82.

7. Deswal A, Petersen NJ, Souchek J, et al. Impact of race on health careutilization and outcomes in veterans with congestive heart failure.

J Am Coll Cardiol 2004;43:778–84.

2

2

3

3

3

3

3

1684 Yancy et al. JACC Vol. 51, No. 17, 2008African Americans Hospitalized With HF: The OPTIMIZE-HF Registry April 29, 2008:1675–84

8. Rathore SS, Foody JM, Wang Y, et al. Race, quality of care, andoutcomes of elderly patients hospitalized with heart failure. JAMA2003;289:2517–24.

9. Kornowski R, Zeeli D, Averbuch M, et al. Intensive home-caresurveillance prevents hospitalization and improves morbidity ratesamong elderly patients with severe congestive heart failure. Am Heart J1995;129:762–66.

0. Shah NB, Der E, Ruggerio C, Heidenreich PA, Massie BM. Preven-tion of hospitalizations for heart failure with an interactive homemonitoring program. Am Heart J 1998;135:373–8.

1. Rich MW, Beckham V, Wittenberg C, Leven CL, Freedland KE,Carney RM. A multidisciplinary intervention to prevent the readmis-sion of elderly patients with congestive heart failure. N Engl J Med1995;333:1190–5.

2. Abraham WT, Massie BM, Franciosa JA, et al. Tolerability, safety,

and efficacy of beta-blockade in black patients with heart failure in the F

community setting: insights from a large prospective beta-blockerregistry. J Card Fail 2003;9:S94.

3. Hunt SA, Abraham WT, Chin MH, et al. ACC/AHA 2005guideline update for the diagnosis and management of chronicheart failure in the adult: a report of the American College ofCardiology/American Heart Association Task Force on PracticeGuidelines (Writing Committee to Update the 2001 Guidelines forthe Evaluation and Management of Heart Failure). J Am CollCardiol 2005;46:1116 – 43.

4. Adams KF, Lindenfeld J, Arnold JMO, et al. HFSA 2006comprehensive heart failure practice guideline. J Card Fail 2006;12:e1–22.

APPENDIX

or supplementary tables, please see the online version of this article.