Racial Disparities in Type of Heart Failure and Hospitalization

Abstract

Heart failure (HF) is one of the leading causes of hospitalization and readmissions. Our study aimed to examine racial disparities in heart failure patients including onset, mortality, length of stay (LOS), direct costs, and readmission rates. This is a secondary data analysis. We analyzed the risk-adjusted inpatient data of all patients admitted with HF to one health academic center. We compared five health outcomes among three racial groups (white, black, and Hispanic). There were 1006 adult patients making 1605 visits from 10/01/2011 to 09/30/2015. Most black patients were admitted in younger age than other racial groups which indicates the needs for more public health preventions. With risk adjustments, the racial differences in LOS and readmission rates remain. We stratified health outcomes by race/ethnic and type of HF. The findings suggest that further studies to uncover underlying causes of these disparities are necessary. Using risk-adjusted hospitalization data allows for comparisons of quality of care across three racial groups. The study suggests that more prevention and protection services are needed for African American patients with heart failure.

Full text

Vol.:(0123456789)
1 3
Journal of Immigrant and Minority Health
https://doi.org/10.1007/s10903-018-0727-4
ORIGINAL PAPER
Racial Disparities inType ofHeart Failure andHospitalization
Wei‑ChenLee1 · HaniSerag1· RobertL.Ohsfeldt2· KarlEschbach3· WissamKhalife4· MohamedMorsy4·
KennethD.Smith1· BenG.Raimer1
© Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract
Heart failure (HF) is one of the leading causes of hospitalization and readmissions. Our study aimed to examine racial dis-
parities in heart failure patients including onset, mortality, length of stay (LOS), direct costs, and readmission rates. This is a
secondary data analysis. We analyzed the risk-adjusted inpatient data of all patients admitted with HF to one health academic
center. We compared five health outcomes among three racial groups (white, black, and Hispanic). There were 1006 adult
patients making 1605 visits from 10/01/2011 to 09/30/2015. Most black patients were admitted in younger age than other
racial groups which indicates the needs for more public health preventions. With risk adjustments, the racial differences
in LOS and readmission rates remain. We stratified health outcomes by race/ethnic and type of HF. The findings suggest
that further studies to uncover underlying causes of these disparities are necessary. Using risk-adjusted hospitalization data
allows for comparisons of quality of care across three racial groups. The study suggests that more prevention and protection
services are needed for African American patients with heart failure.
Keywords Heart failure· Health disparities· Length of stay· Readmission
Introduction
Heart failure (HF) is a globally growing epidemic and one
of the leading causes of hospitalization and readmission
in the US [1]. It affects more than 5.7million Americans
with around half a million new cases [1], costs the nation
$32billion [2], and contributes to 1.02million discharges
every year [3]. A recent longitudinal study reports that the
age-adjusted death rate for HF for African American popu-
lation was 91.5 deaths per 100,000 population, higher than
the HF death rates white and Hispanic populations − 87.3
and 53.3 respectively [4]. Likewise, Sharma and colleagues
pointed out that HF is more prevalent in African Americans
than in whites, imposing higher rates of death, morbidity,
and malignant course development [5]. Accordingly, it is
imperative to explore these notable racial/ethnic disparities
and identify solutions for preventable hospitalizations.
Heart failure refers to a chronic and progressive con-
ditions in which the heart characterized by compromised
pumping capacity resulting in inadequate systemic perfu-
sion to meet the metabolic needs of the body [6, 7]. In this
paper we adopted the classification of systolic (left ventri-
cle), diastolic (right ventricle) or combined heart failure.
Systolic heart failure is also associated with reduced ejection
fraction while diastolic is associated with preserved ejection
fraction [7, 8]. The systolic heart failure is associated with
increase in the left ventricular mass resulting in compro-
mised contractibility to pump blood and disproportionate
perfusion in comparison with the requirements [9]. Diastolic
heart failure is associated with increase of resistance of one
or both ventricles to filling resulting in increased hear rate to
compensate the low volume with rapid pumping [9]. About
half of people with HF die within 5years after being diag-
nosed [2, 6]. Heart failure is categorized as either systolic,
diastolic, or combined, based on the signs of congestion with
or without significant left ventricular contraction [10]. Most
* Wei-Chen Lee
weilee@utmb.edu
1 Office ofHealth Policy andLegislative Affairs, University
ofTexas Medical Branch, 301 University Blvd, Galveston,
TX77555-0920, USA
2 Department ofHealth Policy andManagement, Texas A&M
Health Science Center, CollegeStation, TX, USA
3 Department ofPreventive Medicine andCommunity Health,
University ofTexas Medical Branch, Galveston, TX, USA
4 Department ofInternal Medicine-Cardiology, University
ofTexas Medical Branch, Galveston, TX, USA

Journal of Immigrant and Minority Health
1 3
cases with systolic HF are a result of end-stage coronary
artery disease, which means patients either have a history of
myocardial infraction or a chronically underperfused myo-
cardium. Thus, early recognition of symptoms is important
to prevent HF which causes heavy burdens on patients and
on society by high utilization of medical services [11].
Racial and ethnic disparities in cardiovascular diseases
including HF are well identified and documented [5, 12,
13]. Nevertheless, it is unknown whether a person’s race or
ethnicity put them more at risk for having a certain type of
heart failure and how it affects different aspects of hospi-
talization. The current study aims to identify and describe
the racial/ethnic disparities in terms of the onset of different
HF types and utilization of inpatient care by analyzing a
4-year hospital inpatient data. The number of HF admissions
of this hospital accounted for more than 90% of total HF
admissions in Galveston County [14]. Given it is a complete
patient data rather than a random survey sample, the findings
may give insights of a profile of racial minorities with HF
compared to white patients. The findings may also inform
how to orient resources and actions to eliminate racial/ethnic
disparities in health systems.
Methods
Study Population
The data of this secondary data analysis was drawn from
University HealthSystem Consortium (UHC) database, a
collection of quarterly outpatient and inpatient data from
academic medical centers (AMCs) nationwide [15]. The
study population included patients with HF as their prin-
cipal diagnosis [16], and patients who discharged from the
inpatient department of one AMC between 10/01/2011 and
09/30/2015.
Admissions for hospice, chemotherapy, radiation therapy,
dialysis, prisoner, and medical tourism were excluded. We
classified each patient’s principal diagnosis as congestive HF
(ICD-9: 428.0), left HF (ICD-9: 428.1), systolic HF (ICD-
9: 428.2), diastolic HF (ICD-9: 428.3), combined systolic
and diastolic HF (ICD-9: 428.4), or unspecified HF (ICD-
9: 428.9). We also categorized patient’s race/ethnicity into
non-Hispanic white, non-Hispanic African American, or
Hispanic populations. Patients younger than 18, patients in
other race categories (e.g., Asian or Native Hawaiians), and
patients admitted for left HF or unspecified HF were eventu-
ally excluded due to small sample size (< 10).
Study Variables
The UHC system incorporates risk adjustment mod-
els for cost, length of stay (LOS), potentially avoidable
complications, readmissions, and mortality to reflect patient
complexity of its hospital members [17]. The variables used
in the adjustment model are age, gender, insurance plan,
admission source, admission status, and a common group of
comorbidities defined by UHC. The categories of primary
payer are commercial plans, Medicare, Medicaid and indi-
gent, self-pay, and others (e.g. VA). The patient’s admission
source is either facility or non-facility. Patient’s status at
admission is either emergency or not emergency.
The independent variables are race/ethnicity, and type of
heart failure. The dependent variables include the prevalence
of each type of HF, UHC-adjusted LOS (duration between
date of admission and date of discharge) per visit, UHC-
adjusted direct cost (e.g. healthcare provider time) per day,
UHC-adjusted mortality, and UHC-adjusted HF-related
readmission rates (i.e. percentage of patients being read-
mitted to the same hospital within 90days after the previ-
ous discharge). With the application of adjustment models,
UHC’s members can directly compare each other’s perfor-
mance to gain valuable insights from other members with
leading practices [18].
Statistical Analysis
The univariate analysis contained both encounter-level and
patient-level information on patients’ characteristics. The
bivariate analysis examined the associations of race/ethnicity
with type of HF and other patients’ characteristics present on
admission. Continuous variables were reported as mean and
range and compared using Student’s t test. Categorical vari-
ables were compared between groups using χ2 statistics. The
multivariate analysis addressed the racial/ethnic disparities
in the risk-adjusted outcome variables [17]. Particularly for
the binary outcome (readmitted in 90 days/not readmitted),
we employed generalized estimating equation (GEE) models
with an independent working correlation matrix. All analy-
ses were conducted by Stata 14 [19] and a p-value less than
0.05 was considered as statistical significance.
Results
During 4years, there were 1006 individual patients with
1605 visits (Table1). Half of patients were whites and 45.4%
were female. Around 64.6% of encounters were covered by
Medicare. In the first visit, 37.5% of patients were admitted
for diastolic HF. Among all encounters, diastolic HF also
accounted for the plurality (35.0%). Most patients came from
non-facility (i.e. their homes) in emergency (> 74%).
The bivariate analysis (Table2) demonstrates the asso-
ciation of race/ethnicity with other patient characteristics.
There are statistically significant differences in type of HF,
gender, age, and primary payer for different racial/ethnic

Journal of Immigrant and Minority Health
1 3
groups. While white and Hispanic patients were mainly
admitted for diastolic HF (38.8 and 41.1%), more African
American patients were admitted for systolic HF and com-
bined HF (30.7%). Further, African American patients were
younger and more likely to have Medicaid as a primary pay-
ment source (18.0%) compared to whites (7.8%) or Hispanic
(7.7%). Finally, more Hispanic encounters were visited by
female (50.7%) and self-pay (18.8%) compared to their racial
counterparts.
Table3 illustrates the risk-adjusted health outcomes
by patient’s race/ethnicity and type of HF. If all encoun-
ters are taken into account, African American patients
had significantly longer LOS per visit (6.74days) than
whites (5.91days) or Hispanic (6.33days). Additionally,
a higher percentage of African Americans was readmit-
ted within 90days after the previous admission (25.5%)
and this number is even higher for readmissions related to
Table 1 Analysis of
characteristics by individual-
patient level and encounter level
a Patient’s information in the first visit
Patient (N = 1006) Encounter (N = 1605)
Race/ethnicity
Non-Hispanic white 503 (50.0%) 766 (47.7%)
Non-Hispanic black 364 (36.2%) 632 (39.4%)
Hispanic origin 139 (13.8%) 207 (12.9%)
Gender (female) 457 (45.4%) 691 (43.1%)
Age (mean, range) 66.1 (19–98)a65.4 (19–98)
Primary payer
Commercial 83 (8.3%)a111 (6.9%)
Medicaid 106 (10.4%)a190 (11.8%)
Medicare 623 (61.9%)a1037 (64.6%)
Self-pay 147 (14.6%)a200 (12.5%)
Others 47 (4.7%)a67 (4.2%)
Primary diagnosis (ICD-9-CM)
Congestive HF (428.0) 129 (12.8%)a182 (11.3%)
Systolic HF (428.2) 275 (27.3%)a468 (29.2%)
Diastolic HF (428.3) 377 (37.5%)a561 (35.0%)
Combined HF (428.4) 225 (22.4%)a394 (24.6%)
Source of admission (facility) 91 (9.1%)a134 (8.3%)
Status at admission (emergency) 788 (78.3%)a1189 (74.1%)
Table 2 Patient’s characteristics
by race/ethnicity at encounter
level
Non-Hispanic white Non-Hispanic
African American
Hispanic
Total (p < 0.001) 766 (100.0%) 632 (100.0%) 207 (100.0%)
Congestive HF 89 (11.6%) 65 (10.3%) 28 (13.5%)
Systolic HF 228 (29.8%) 194 (30.7%) 46 (22.2%)
Diastolic HF 297 (38.8%) 179 (28.3%) 85 (41.1%)
Combined HF 152 (19.8%) 194 (30.7%) 48 (23.2%)
Gender (female) (p = 0.018) 307 (40.1%) 279 (44.2%) 105 (50.7%)
Age (mean, range) (p < 0.001) 69.6 (20–98) 60.5 (22–98) 65.2 (19–91)
Primary payer (p < 0.001)
Commercial 44 (5.7%) 54 (8.6%) 13 (6.3%)
Medicaid 60 (7.8%) 114 (18.0%) 16 (7.7%)
Medicare 546 (71.3%) 364 (57.6%) 127 (61.4%)
Self-pay 83 (10.8%) 78 (12.3%) 39 (18.8%)
Others 33 (4.3%) 22 (3.5%) 12 (5.8%)
Source of admission (facility) (p = 0.606) 62 (8.1%) 51 (8.1%) 21 (10.1%)
Status at admission (emergency) (p = 0.249) 559 (73.0%) 482 (76.3%) 148 (71.5%)

Journal of Immigrant and Minority Health
1 3
combined HF (32.5%). However, there are no racial/ethnic
differences in cost or mortality.
While only focusing on the effects of race/ethnicity,
Table4 shows several statistically significant differences in
terms of LOS per visit (p < 0.01) and probability of 90-day
readmission (p < 0.01) for African American patients. Sur-
prisingly, the differences remain after adding the type of
HF. Both model 2 and model 8 show that African Ameri-
can patients still stayed longer or had a higher likelihood
to be readmitted than non-Hispanic white patients (coef-
ficient > 0, p < 0.05). Being an African American predicts
a 0.63day increase in LOS per visit compared to being
white regardless of type of HF. Likewise, being an African
American is 1.36 (Odds Ratio = e0.31) more likely to be
readmitted than being a non-Hispanic white.
Finally, Fig.1 demonstrates the accumulative percent-
age of encounters stratified by patient’s gender, age, and
race/ethnicity. Most African American, male patients were
admitted to hospital for HF in younger age (younger than
65years old) compared to other groups, followed by Afri-
can American, female patients. Hispanic, male patients
were the third highest percentage in terms of age younger
than 65 but it has a significant increase from age 65 to 70.
Overall, the increase in number of admissions is associ-
ated with three factors in order: African American > His-
panic > White, Male > Female, and Old > Young.
Discussion
The results of this study contribute significantly by identi-
fying the racial/ethnic disparities in length of stay (LOS)
and 90-day, HF-caused readmission rates after risk adjust-
ments. In particular, African American patients experi-
enced loner LOS and more readmissions than Hispanic
patients which generally are regarded as a group of vulner-
able population with limited English proficiency to seek
care [20]. More details were discussed below.
Prior studies have well documented the cardiovascular
disease disparities associated with social determinants of
health including race, ethnic, gender, and so on [21–23].
The study (Table2) has similar findings such as most
white patients were male, African American were admit-
ted at younger age, and Hispanic patients were uninsured.
The study found that African American patients were more
likely to be admitted for systolic or combined HF (more
severe types of HF than congestive or diastolic HF) that
could cause higher utilization of medical care. Although
previous studies have identified the risk factors for HF in
general [24–27], no literature suggests risk factors for a
particular type of HF yet. Therefore, we suggest that future
studies explore and describe social determinants of health
associated with each type of HF for each racial group.
Table 3 Risk-adjusted health outcomes by race/ethnicity and type of HF at encounter level
The numbers of 30-day and 60-day readmissions of Hispanic patients were too low to allow for statistical tests
UHC risk-adjusted mean and standard error Non-Hispanic white
(N = 766)
Non-Hispanic African American
(N = 632)
Hispanic (N = 207)
Average LOS per visit (days) (p = 0.018) 5.91 (0.17) 6.74 (0.25) 6.33 (0.36)
Congestive HF (p = 0.972) 5.01 (0.20) 5.07 (0.39) 4.94 (0.28)
Systolic HF (p = 0.763) 7.14 (0.48) 7.67 (0.55) 7.29 (0.97)
Diastolic HF (p = 0.086) 5.14 (0.11) 5.60 (0.19) 5.26 (0.21)
Combined HF (p = 0.081) 6.11 (0.39) 7.44 (0.53) 8.12 (1.18)
Average cost per day ($) (p = 0.536) 1913.21 (70.02) 1821.26 (53.95) 1928.78 (105.49)
Congestive HF (p = 0.268) 1409.33 (66.25) 1609.48 (139.59) 1388.93 (69.28)
Systolic HF (p = 0.277) 2558.56 (177.68) 2203.67 (133.86) 2554.25 (359.90)
Diastolic HF (p = 0.780) 1467.45 (58.05) 1475.80 (53.71) 1543.40 (82.49)
Combined HF (p = 0.106) 2111.20 (177.87) 1828.55 (83.99) 2326.72 (218.63)
Mortality (p = 0.483) 0.022 (0.002) 0.020 (0.002) 0.025 (0.005)
Congestive HF (p = 0.334) 0.014 (0.002) 0.015 (0.005) 0.025 (0.009)
Systolic HF (p = 0.348) 0.024 (0.004) 0.018 (0.002) 0.016 (0.003)
Diastolic HF (p = 0.491) 0.016 (0.002) 0.021 (0.004) 0.021 (0.006)
Combined HF (p = 0.222) 0.033 (0.008) 0.022 (0.004) 0.044 (0.017)
90-day readmission (yes) (p = 0.007) 143 (18.7%) 158 (25.0%) 37 (17.9%)
Congestive HF (p = 0.591) 15 (16.9%) 11 (16.9%) 7 (25.0%)
Systolic HF (p = 0.148) 44 (19.3%) 53 (27.3%) 11 (23.9%)
Diastolic HF (p = 0.121) 56 (18.9%) 31 (17.3%) 8 (9.41%)
Combined HF (p = 0.012) 28 (18.4%) 63 (32.5%) 11 (22.9%)

Journal of Immigrant and Minority Health
1 3
Literature has characterized the impacts of race/ethnic
on health outcomes [26]. The study (Table3) has similar
findings while showing more details by each type of HF.
African American patients had longer LOS and a higher
90-day readmission rate than another two racial groups
except for combined HF. Whether compliance with recom-
mended prescription and self-care differed among three
racial groups among our study population is unknown.
Nevertheless, financial barriers and educational limita-
tion were found in relation with low compliance and high
admissions [28, 29]. Other factors like smoking, drinking,
and substance use are also common among minorities with
heart failure [26, 27]. Accordingly, more risk factors of
poor health outcomes in African American patients need
to be studied.
Along with UHC’s risk adjustments, we further esti-
mated how race/ethnicity and type of HF influence four
health outcomes of interest. Similar to the findings of
Tables3 and 4 shows that African American patients
with HF experienced longer LOS and more readmissions
than their racial counterparts. One of the reasons about
delaying a discharge of patients with HF would be a pro-
vider decision taking into account that African American
patients are less likely to have any family support, home-
based services, or primary care doctor in their living areas
[30–32]. Especially when we focus on HF-related read-
missions instead of all-cause readmissions, vulnerability
and disadvantages of a particular racial group accumulate
over time. Many studies have suggested different types
of interventions and they require both public and private
financial support to make effective, sustainable changes
[29, 30, 33, 34].
Table 4 Linear or logistic regression analysis of risk-adjusted health outcomes by race/ethnicity and type of HF
M model
*p<0.05; **p<0.01; ***p<0.001
UHC-risk adjusted
coefficient (95%CI)
(1) LOS/visit (2) Cost/day (3) Mortality (4) Readmission
M1 M2 M3 M4 M5 M6 M7 M8
Race/ethnicity
Black 0.83** (0.26–1.41) 0.63* (0.06–1.20) − 91.94 (− 268.7–
84.87)
− 160.06 (− 333.30–
13.18)
−0.002 (− 0.01–
0.004)
−0.003 (− 0.009–
0.003)
0.35** (0.07–0.62) 0.31* (0.04–0.59)
Hispanic 0.42 (− 0.42–1.25) 0.52 (− 0.30–1.35) 15.57 (− 242.19–
273.33)
68.79 (− 181.89–
319.47)
0.004 (− 0.005–
0.013)
0.004 (− 0.005–
0.013)
−0.02 (−0.43–
0.39)
−0.02 (−0.43–
0.39)
Type of HF
Systolic – 2.35*** (1.43–3.27) – 946.35 (666.84–
1225.86)
–0.005 (− 0.005–
0.016)
– 0.17 (−0.28–
0.61)
Diastolic – 0.31 (− 0.059–1.21) –− 1.99 (− 274.66–
270.78)
–0.002 (− 0.007–
0.012)
– −0.06 (−0.51–
0.38)
Combined – 1.92*** (0.97–2.87) – 544.45 (257.20–
831.71)
–0.013* (0.003–0.023) – 0.30 (−0.15–
0.75)
Fig. 1 Accumulative percentage of encounters by gender and race/
ethnicity

Journal of Immigrant and Minority Health
1 3
One of the advantages using UHC data is to be able to
examine a set of hospital core measures [35], indicators
related to quality and safety of care. Information we obtained
including discharge instructions given to HF patients, evalu-
ation of left ventricular systolic function, and prescription
of angiotensin converting enzyme inhibitor and angiotensin
receptor blocker if necessary. The results indicated no racial/
ethnic disparities in the core measures, suggesting the equity
in quality and safety of care.
Finally, this study pointed out that African American,
male patients were the most vulnerable since they were read-
mitted to hospital at their younger age as opposed to other
gender and racial groups. In the meantime, there are sys-
tematic variations in relation to admissions. Figure1 clearly
shows that a particular race (African American) and gender
(male) are associated with admissions which is consistent
with a study focusing on young and middle-aged adults with
HF [36]. In other words, the disease burden heavily affects
quality of life in younger age of patients with both character-
istics (male, African American). Identifying the racial group
most vulnerable to poor health outcomes can help policy
makers and healthcare providers target their efforts to reduce
the risk of admissions not only within the hospital but also
after discharge. Using UHC data provides a unique oppor-
tunity to compare hospitalization outcomes among UHC’s
AMC members which employ the same risk adjustment
model; yet, there are some shortcomings of this study. The
secondary data analysis from one academic medical center
does not allow for generalization. Also, literature has indi-
cated that mortality due to heart failure is higher in African
Americans [5, 37]. However, the current study has no suf-
ficient evidence due to the lack of death data out of this hos-
pital. Further studies to uncover underlying causes of these
disparities against African Americans are recommended.
Conclusion
Cardiovascular diseases including heart failure are among
the most widespread and costly health problems facing the
nation today; fortunately, they are also among the most pre-
ventable. Centers for Medicare and Medicaid Services initi-
ated the model of accountable health communities (AHC)
after the model of accountable care organizations (ACO)
[38]. Instead of leaving one another alone, there shall be a
collaborative program between two to prioritize community
health needs and allocate community resources to achieve
health equity. Most important of all, a combination of pri-
mary, secondary, and tertiary interventions is recommended
to achieve a meaningful degree of prevention and protection.
Acknowledgements The researchers would like to express the deep-
est gratitude to the following departments that without their help, this
project would not be possible: (1) University of Texas Medical Brach
Office of the President-Waiver Operations, (2) Clinical Data Manage-
ment, and (3) Office of Health Policy and Legislative Affairs.
Funding The study received financial funding from Texas Medicaid
1115 Waiver (#094092602.1.9).
Compliance with Ethical Standards
Conflict of interest No conflict of interest or others exists.
Ethical Approval The study has acquired the approval of Institutional
Review Board (#16-0128) at University of Texas Medical Branch and
has complied with all requirements for a secondary data analysis to
protect privacy of health information.
References
1. Roger V. Epidemiology of heart failure. Circ Res.
2013;113:646–59.
2. Centers for Disease Control and Prevention (CDC). Heart failure
fact sheet. http://www.cdc.gov/dhdsp /data_stati stics /fact_sheet s/
fs_heart _failu re.htm.
3. American Heart Association. Heart disease and stroke statis-
ticis-2016 update. https ://circ.ahajo urnal s.org/conte nt/early
/2015/12/16/CIR.00000 00000 00035 0.
4. Ni H, Xu JQ. Recent trends in heart failure-related mortality:
United States, 2000–2014. NCHS data brief, no 231. Hyattsville,
MD: National Center for Health Statistics. http://www.cdc.gov/
nchs/data/datab riefs /db231 .htm.
5. Sharma A, Colvin-Adams M, Yancy CW. Heart failure in Afri-
can Americans: disparities can be overcome. Clevel Clin J Med.
2014;81(5):301–11.
6. National Heart, Lung, and Blood Institute. What is heart failure?
https ://www.nhlbi .nih.gov/healt h/healt h-topic s/topic s/hf.
7. Center of Continuous Education at Cleveland Clinic. Diseases
management: heart failure. 2017. http://www.cleve landc linic
meded .com/medic alpub s/disea seman ageme nt/cardi ology /heart
-failu re/.
8. American Heart Association. Types of Heart Failure. 2017. http://
www.heart .org/HEART ORG/Condi tions /Heart Failu re/About
Heart Failu re/Types -of-Heart -Failu re_UCM_30632 3_Artic le.jsp#.
WiBcH DdryV A.
9. Medical Clinics. Pathophysiology of systolic and diastolic heart
failure. 2017. http://www.medic al.thecl inics .com/artic le/S0025
-7125(12)00123 -X/abstr act.
10. Ashley EA, Niebauer J. Chapter7: Heart failure. In: Cardiology
explained. London, UK: Remedica. http://www.ncbi.nlm.nih.gov/
books /NBK22 18/.
11. Borlaug BA, Redfield MM. Are systolic and diastolic heart failure
overlapping or distinct phenotypes within the heart failure spec-
trum? Circulation. 2011;123:2006–14.
12. Davis A, etal. Cardiovascular health disparities. Med Care Res
Rev. 2007;64(5 Suppl):29S–100S.
13. Graham G. Population-based approaches to understanding dispari-
ties in cardiovascular disease risk in the United States. Int J Gen
Med. 2014;7:393–400.
14. Center to Eliminate Health Disparities (CEHD). Disparity docu-
mentation report and improvement plans. Galveston, TX: Uni-
versity of Texas Medical Branch. http://www.utmb.edu/hpla/healt
h-dispa ritie s/publi catio ns.

Journal of Immigrant and Minority Health
1 3
15. University of Utah. What is the University HealthSystem Consor-
tium? Health care: a community to quality. http://healt hcare .utah.
edu/quali ty/conso rtium .php.
16. ICD9Data.com Heart Failure 428. http://www.icd9d ata.com/2014/
Volum e1/390-459/420-429/428/defau lt.htm.
17. University HealthSystem Consortium. UHC risk adjustment meth-
odology. Chicago, IL: UHC; 2014.
18. University HealthSystem Consortium. UHC clinical data base/
resource manager user manual. Chicago, IL: UHC; 2015.
19. StataCorp LP. Generalized estimating equations: xtgee. http://
www.stata .com/featu res/gener alize d-estim ating -equat ions/.
20. Escarce JJ, Kapur K. Chapter10: access to and quality of health
care. In: Tienda M, Mitchell F, editors. Hispanics and the future
of America. Washington DC: National Academies Press; 2006.
21. Bui AL, Horwich TB, Fonarow GC. Epidemiology and risk profile
of heart failure. Nat Rev Cardiol. 2011;8(1):30–41.
22. Lan T, etal. Social determinants of cardiovascular diseases. Public
Health Rev. 2012;33(2):601 – 22.
23. Melton KD, etal. Heart failure in Hispanic Americans: improv-
ing culture awareness. School of Nursing Faculty Publications
of Purdue University. http://docs.lib.purdu e.edu/cgi/viewc onten
t.cgi?artic le=1026&conte xt=nursi ngpub s.
24. Calvillo-King L, etal. Impact of social factors on risk of read-
mission or mortality in pneumonia and heart failure: systematic
review. J Gen Intern Med. 2012;28(2):269–82.
25. Rahimi K, etal. Risk prediction in patients with heart failure.
JCHF. 2014;2(5):440–6.
26. Ross JS, etal. Statistical models and patient predictors of read-
mission for heart failure: a systematic review. Arch Intern Med.
2008;168(13):1371–86.
27. Yang H, etal. Clinical prediction of incident heart failure
risk: a systematic review and meta-analysis. Open Heart.
2015;2(1):e000222.
28. Gardetto NJ. Self-management in heart failure: where have
we been and where should we go? J Multidiscip Healthc.
2011;4:39–51.
29. Hubbard T, McNeill N. Improving medication adherence and
reducing readmissions. http://www.nacds .org/pdfs/pr/2012/nehi-
readm issio ns.pdf.
30. Damiani G, etal. Influence of socioeconomic factors on hospital
readmissions for heart failure and acute myocardial infarction in
patients 65 years and older: evidence from a systematic review.
Clin Interv Aging. 2015;10:237–45.
31. Giamouzis G, etal. Hospitalization epidemic in patients with heart
failure: risk factors, risk prediction, knowledge gaps, and future
directions. J Card Fail. 2011;17(1):54–75.
32. Tschihashi M, etal. Medical and socioenvironmental predictors
of hospital readmission in patients with congestive heart failure.
Am Heart J. 2001;142(4):E7.
33. Desai AS, Stevenson LW. Re-hospitalization for heart failure:
predict or prevent? Circulation (Special Report). 2012;126:501–6.
34. Voigt J, etal. A reevaluation of the costs of heart failure and its
implications for allocation of health resources in the United States.
Clin Cardiol. 2014;37(5):312–21.
35. Joint Commission. Specification manual for Joint Commission
National Quality Core Measures. https ://manua l.joint commi ssion
.org/relea ses/archi ve/TJC20 10B/Heart Failu re.html.
36. Ranasinghe I, etal. Readmissions after hospitalization for heart
failure, acute myocardial infarction, or pneumonia among young
and middle-aged adults: a retrospective observational cohort
study. PLOS Med. 2014;11(9):e1001737.
37. Yancy C, Strong M. The national history, epidemiology, and prog-
nosis of heart failure in African Americans. http://www.medsc ape.
com/viewa rticl e/46811 9.
38. Centers for Medicare and Medicaid Services (CMS). Innovation
models. https ://innov ation .cms.gov/initi ative s/index .html#views
=model s.