Heart Failure Quality of Care Among Asian Patients in the United States

Abstract

Background:

Heart failure (HF) incidence varies among U.S. Asian populations.

Objectives:

The goal of this study was to identify differences in quality of inpatient care among individual groups of Asian patients hospitalized with HF and compared with non-Hispanic White patients.

Methods:

This study included 824 U.S. hospitals from the Get With The Guidelines-Heart Failure registry (2015–2023). It evaluated the odds of optimal medical therapy (OMT) (defined as an angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker or angiotensin receptor-neprilysin inhibitor for those with an ejection fraction [EF] ≤40%; a beta-blocker for those with an EF ≤40%; and mineralocorticoid receptor antagonist for those with an EF ≤35%) at discharge for HF with reduced EF, and length of stay (LOS) >4 days, defect-free care (OMT, follow-up visit scheduling, HF education) at discharge, and in-hospital mortality among patients with any EF. The study cohorts included Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, and Other Asian vs non-Hispanic White patients.

Results:

Among 7,261 Asian patients with individual Asian group identified (mean age range 69.9–78.8 years; 41%−51% female) vs 768,566 non-Hispanic White patients (mean age 74.6 years; 47% female), the odds of receiving OMT were lower in Vietnamese male patients (adjusted OR [aOR]: 0.68; 95% CI: 0.47–1.00). The odds of longer LOS (>4 days) were lower in Vietnamese male patients (aOR: 0.68; 95% CI: 0.50–0.91) and Filipina female patients (aOR: 0.66; 95% CI: 0.46–0.95). The odds of defect-free care were lower in Filipina female patients (aOR: 0.52; 95% CI: 0.34–0.82). In-hospital mortality did not differ compared with non-Hispanic White patients. Among female Asian patients, there was significant heterogeneity in frequency of LOS >4 days and defect-free care across individual Asian groups.

Conclusions:

HF quality of care is heterogeneous, with substantial gaps in certain U.S. Asian groups compared with non-Hispanic White patients.

Central Illustration.

Heart Failure Quality of Care Among Asian Patients in the United States

Patient sample, setting, outcome measures, and primary findings in the analysis of heart failure quality of care in the Get With the Guidelines-Heart Failure registry data of U.S. Asian patients, 2015–2023. DFC = defect-free care; LOS = length of stay; OMT = optimal medical therapy.

Asian populations are the fastest growing racial groups in the United States,^1,2 and burden of heart failure (HF) morbidity and mortality is growing in U.S. Asian populations.^3–5 However, there is evidence of heterogeneity in HF mortality rates among Asian groups according to region of origin. Although HF mortality rates are increasing in all U.S. Asian groups, Asian Indian, Filipino, and Japanese men and Asian Indian and Vietnamese women experience the highest rates of HF mortality.⁶ Among Asian adults in a large U.S. health system, Southeast Asian adults have the highest HF incidence, and Southeast and South Asian adults have the highest HF prevalence.⁷ The burden of upstream HF risk factors, including obesity, diabetes, and hypertension, varies considerably across Asian groups in the United States.^8–10

Reasons for the heterogeneity in HF outcomes across Asian groups remain under investigation. Adverse social risk factors, such as limited English proficiency and unhealthy environmental exposures, may increase the risk for HF among individual Asian groups and may contribute to suboptimal outcomes.¹¹ Asian populations in the United States also have unique experiences of both interpersonal and systemic racism, which can create barriers to care and contribute to worse health outcomes.¹² Prior investigation found variability in quality of care and outcomes after myocardial infarction across U.S. Asian groups.¹³ However, quality of care for HF, such as prescription and adherence to optimal medical therapy (OMT), remains largely unexplored as a potential contributor to contemporary differences in HF outcomes in U.S. Asian populations.

Clinical characteristics and outcomes of Asian patients with HF were previously evaluated in the American Heart Association (AHA) Get With The Guidelines–Heart Failure (GWTG-HF) registry for hospitalizations between 2005 and 2012. This analysis suggested that Asian patients largely had comparable quality of care relative to White patients.¹⁴ However, in these data, Asian patients were aggregated into a single category, masking heterogeneity across Asian groups. The goal of the current study therefore was to assess HF quality of care and outcomes in the 6 largest Asian ethnic groups in the United States (Asian Indian, Chinese, Filipino, Japanese, Korean, and Vietnamese) and the “Other Asian” group in the AHA GWTG-HF registry between 2015 and 2023.

Methods

Study population.

The AHA GWTG-HF registry is a voluntary quality improvement program in the United States. Details of the registry and its data collection methods have been described elsewhere.^15,16 Participating hospitals (n = 824) concurrently or retrospectively abstract chart data for consecutive patients admitted with HF based on the primary admission diagnosis (Figure 1); this may include information from unstructured clinical notes, structured data entry fields, and/or administrative diagnosis codes, among other electronic health record (EHR) locations. Registry participants aged <18 years, without documented left ventricular ejection fraction (LVEF), without information on vital status at discharge, and those transferred to other facilities were not eligible for this analysis.

Figure 1.

GWTG-HF Registry Patient Eligibility

A patient may be excluded from primary analysis for multiple exclusion criteria; the frequencies displayed show the total number of patients who meet each stated criterion. CAD = coronary artery disease; GWTG-HF = Get With The Guidelines–Heart Failure; NHW = non-Hispanic White.

Among patients with confirmed HF, data were included from registry patients between January 1, 2015 (when the disaggregated Asian group was first identifiable on registry case report forms), and March 3, 2023 (the most recent available data in the registry at the time of analysis), to evaluate contemporary quality of care and outcomes. In the primary analysis, we evaluated patients who were specifically identified as Asian ethnic groups: Asian Indian, Chinese, Filipino, Japanese, Korean, or Vietnamese. A category of “Other Asian” was also evaluated, which is a specific and separate Asian subcategory group available on the registry case report form.

Patients who were Asian were identified by self-report, from designation in the EHR, or designated by the hospital staff completing registry case report forms. Patients who identified as Asian were compared with non-Hispanic White patients to identify differences in HF outcomes and quality of care experienced by minoritized groups. In addition, there were patients in the registry identified as Asian race, without further specification of individual Asian subcategory group. This group likely included individuals from all 7 subcategory groups, and thus a secondary analysis was conducted evaluating outcomes separately in this group of Asian patients who did not have individual groups specified.

Patients identified as Hispanic or Black or African American were not evaluated in this analysis because HF care and outcomes for these patients in this registry have previously been reported.^17–20 Patients identified as multiple race or ethnicity groups were not evaluated because of the heterogeneity of this group.

Data from the AHA GWTG-HF registry are available via a proposal submission and evaluation process. IQVIA serves as the data collection and coordination center.

Each participating hospital received either human research approval to enroll cases without individual patient consent under the common rule, or a waiver of authorization and exemption from subsequent review by their Institutional Review Board (IRB). Advarra, the IRB for the AHA, determined that this study was exempt from IRB oversight.

Patient characteristics and outcomes.

We evaluated participant characteristics, including age, insurance status, medical history (diabetes, hyperlipidemia, hypertension, any smoking history, prior coronary artery bypass graft, prior myocardial infarction, prior percutaneous coronary intervention, prior heart failure diagnosis, prior cerebrovascular accident or transient ischemic attack, prior atrial fibrillation or atrial flutter, chronic kidney disease, and valve disease), presentation point of origin (clinic, department room, another health care facility), and HF presenting clinical characteristics (body mass index, systolic and diastolic blood pressures, creatinine level, electrocardiogram QRS duration, and sodium concentration) based on data entered on the registry case report form.

LVEF during hospitalization was categorized as ≤40% (defined as HF with reduced ejection fraction [HFrEF]), 41% to 49%, or ≥50%. Pre-hospitalization use of an angiotensin-converting enzyme (ACE) inhibitor or angiotensin II receptor blocker (ARB) or angiotensin receptor-neprilysin inhibitor (ARNI), beta-blocker, or mineralocorticoid receptor antagonist (MRA) was assessed. Interventions during the hospitalization (no procedures, coronary intervention, valve intervention, renal replacement therapy, rhythm device) were quantified. Measures evaluated at time of discharge included documented contraindications to medical therapies among patients with HFrEF, discharge laboratory values (creatinine and potassium levels), discharge disposition (home, hospice, another health care facility, or against medical advice), and receipt of influenza and pneumococcal vaccines.

Primary outcomes were measures identified in contemporary guidelines for HF clinical performance and quality metrics,²¹ as well as those specifically under the AHA Target: Heart Failure quality improvement program that were recommended during the study period (2015–2023) for which sufficient data were available.²² Although dozens of quality measures are suggested in the 2020 American College of Cardiology/AHA guidelines and the AHA GWTG program, the Target: Heart Failure program measures are specifically identified to leverage AHA content resources for patients and health care professionals (including educational tools, prevention programs, treatment guidelines, and quality initiatives).

Among patients with HFrEF, the primary pharmacologic treatment measure was the likelihood of OMT at discharge, which was defined as being prescribed all medications in a class for which a patient is eligible and for which there was no documented contraindication. Medication classes qualifying for OMT were as follows: 1) an ACE inhibitor or ARB or ARNI for those with EF ≤40%; 2) a beta-blocker for those with an EF ≤40%; and 3) an MRA for those with an EF ≤35%. Prescription of sodium-glucose cotransporter-2 (SGLT2) inhibitors were not included in the OMT definition: use of SGLT2 inhibitors was added to the Target: Heart Failure quality improvement measures after the study period, and uptake of this class of medications is relatively recent, with little to no reported usage among all included race groups for most of the study period.

Outcomes evaluated for all patients with HF (with any EF) included hospitalization length of stay (LOS) (ie, admission) >4 days, defect-free care (for those alive at discharge), and in-hospital mortality. Defect-free care was defined as meeting all HF care quality measures for which a patient was eligible, including prescription of up to 3 medication classes (as defined earlier for EF ≤40% or ≥35%), a follow-up visit scheduled ≤7 days after discharge, and receipt of HF enhanced education (which included either referral to an HF management program, ≥60 minutes of patient education, or receiving an HF workbook).²² Given varying eligibility based on EF and contraindications, each patient may be eligible for a different number of quality measures. Experiencing each quality measure was based on documentation in the registry case report form; if boxes for a measure were left unchecked on the form (eg, missing response), the outcome was conservatively coded as “not prescribed” (for medications) or “not performed” (for process measures), as aligned with prior analyses in GWTG registries.¹³

In this analysis, only all HF or HFrEF was evaluated. Outcomes were not specifically assessed for patients with HF with preserved EF (HFpEF) because of the evolving nature of clinical guidelines during the study period, as well as correspondingly low use of HFpEF-directed therapies such as SGLT2 inhibitors during the study period.

Statistical analysis.

Participant characteristics are described as mean ± SD, median (25th, 75th percentile), or frequency (percent). Logistic regression via generalized estimating equations was used to compare the likelihood of several HF-related measures in patients in individual Asian groups vs non-Hispanic White patients. Hospitals were specified as the clustering unit with an independence working correlation and robust SEs, ensuring valid inference despite within-hospital patient correlation. Regression models estimated the odds of OMT among patients with HFrEF in Asian groups compared with non-Hispanic White patients, and analyses were stratified according to sex.

For all patients with HF, we estimated the odds of an LOS >4 days, defect-free care, and in-hospital mortality in Asian groups compared with non-Hispanic White patients. Primary analyses of the odds of OMT, LOS >4 days, and defect-free care excluded patients who were discharged to hospice, who left against medical advice, and who were transferred to another facility.

For all logistic regression analyses, model 1 was unadjusted. Model 2 was adjusted for age, LVEF, diabetes history, hypertension history, smoking history, and insurance status to align with the Meta-Analysis Global Group in Chronic Heart Failure risk score accounting for the risk factors for which data were available with minimal missing data.²³

Adjusted ORs (aORs) and 95% CIs were calculated. To report heterogeneity in outcome measures across race and ethnicity groups, 2 overall Wald test P values are provided. First, global Wald test P values are provided for the primary analysis evaluating outcomes in Asian groups compared vs non-Hispanic White individuals. Second, an analysis using equivalent methods was conducted evaluating the odds of each outcome in individual Asian groups compared vs the largest specified Asian group (ie, Filipino); global Wald test P values are reported to indicate heterogeneity in outcome measures specifically between Asian groups themselves. Because prior research found that disparities are experienced by individual Asian groups, suggesting plausibility of the hypothesized differences in HF quality of care across groups, this analysis focused on a modest number of primary outcomes; in addition, because these data are intended to be hypothesis-generating for future research in robust prospective samples, we did not correct for multiple comparisons. Chi-square tests were used to evaluate for statistically significant differences in the proportion of HFrEF patients with documented contraindications to each medication class across race and ethnicity groups.

Covariate data were abstracted from the registry case report form, which has single checkboxes to indicate if any of the aforementioned medical history or clinical conditions are present. If a box for medical history or clinical condition was not marked as present, it was coded to be “no” or “not present” for analysis. Accordingly, there was minimal missing data for analysis, and regression models were fit with complete case data, in alignment with prior analyses in this registry.¹⁵ For primary analyses, an interaction was tested between Asian group identification and U.S. geographic region (Northeast, Midwest, South, and West); most interaction terms were not significant, and thus analyses were not stratified according to region (results not shown).

One sensitivity analysis and several secondary analyses were conducted. The sensitivity analysis evaluated the primary outcomes in the sample of participants from the 347 hospitals in the registry that contributed data from ≥10 Asian patients during the study period. The first secondary analysis evaluated differences in the likelihood of the primary outcomes in a combined East Asian (Chinese, Japanese, and Korean) group of patients and a Southeast Asian (Filipino and Vietnamese) group of patients compared with non-Hispanic White patients, using the same primary statistical methods. The group of Asian Indian patients (who are one of several South Asian groups) were individually evaluated in the main analysis; because other South Asian groups (eg, Pakistani) are not specified in the GWTG-HF registry, a combined South Asian group was not evaluated. In the second secondary analysis, the primary outcomes were evaluated in the sample of aggregated Asian patients who did not have further individual Asian groups specified and thus could not be further disaggregated.

A 2-sided P value <0.05 was considered statistically significant. Analyses were performed with R version 4.2.0 (R Foundation or Statistical Computing) using the AHA Precision Medicine Platform.

Results

Demographic characteristics of patients in the registry are shown stratified according to sex in Table 1 and overall in Supplemental Table 1. There were 7,261 Asian patients with individual Asian group identified, of whom 1,143 were Asian Indian, 961 were Chinese, 1,598 were Filipino, 698 were Japanese, 360 were Korean, 565 were Vietnamese, and 1,936 were Other Asian. Mean ± SD age ranged from 69.9 ± 14.8 years in Filipino patients to 79.4 ± 14.1 years in Japanese patients. The percentage of patients who were female ranged between 40.8% in Asian Indian patients and 51.1% in Korean participants. There were differences across Asian ethnic groups in frequency of cardiovascular risk factors, cardiovascular disease (CVD) history, and prior cardiac interventions. For example, Asian Indian patients had the highest prevalence of diabetes (66%) and hypertension (87%), and Filipino patients had the highest prevalence of hyperlipidemia (64%). Among Asian patients, Filipino patients had the highest frequency of current or former smoking (12%). Frequency of an HF diagnosis before admission ranged from 66% in Chinese patients to 75% in Asian Indian patients. Frequency of prior myocardial infarction ranged from 13% in Japanese patients to 23% in Other Asian patients. Frequency of prior coronary artery bypass graft surgery ranged from 10% in Japanese patients to 25% in Asian Indian patients. There were 768,566 non-Hispanic White patients, with a mean age of 74.6 ± 13.5 years and 47% female; 42% had diabetes, 58% had hyperlipidemia, 79% had hypertension, 72% had a prior HF diagnosis, 20% had a prior myocardial infarction, and 19% had prior coronary artery bypass graft surgery.

TABLE 1.

Demographic and Medical History Among U.S. Asian Patients Admitted With HF

	Asian Indian	Chinese	Filipino/Filipina	Japanese	Korean	Vietnamese	Other Asian	NHW
Male
N	677	525	933	350	176	304	1,017	409,762
Age, y	69.3 ± 12.6	76.8 ± 13.6	67.1 ± 14.6	74.4 ± 14.7	71.4 ± 14.6	71.3 ± 13.9	67.5 ± 15.9	72.6 ± 13.4
Insurance
Medicare	221 (39)	263 (53)	370 (42)	183 (60)	78 (46)	113 (39)	371 (38)	220,117 (59)
Medicaid	101 (18)	117 (24)	283 (32)	80 (26)	33 (20)	79 (27)	261 (27)	77,542 (21)
Private	224 (39)	102 (21)	210 (24)	40 (13)	57 (34)	95 (33)	282 (29)	59,124 (16)
Self-pay or uninsured	27 (5)	11 (2)	14 (2)	0 (0)	1 (1)	2 (1)	41 (4)	9,475 (3)
Prior history
Diabetes	465 (69)	250 (48)	524 (56)	168 (48)	94 (53)	156 (51)	572 (56)	177,673 (43)
Hyperlipidemia	425 (63)	331 (63)	562 (60)	206 (59)	91 (52)	164 (54)	573 (56)	240,587 (59)
Hypertension	572 (85)	424 (81)	756 (81)	278 (79)	147 (84)	247 (81)	831 (82)	320,637 (78)
Smoking history	79 (12)	57 (11)	160 (17)	50 (14)	26 (15)	54 (18)	176 (17)	69,789 (17)
CABG	204 (31)	85 (16)	193 (21)	49 (14)	38 (22)	53 (17)	232 (23)	100,334 (25)
MI	149 (22)	106 (20)	204 (22)	46 (13)	39 (22)	64 (21)	276 (27)	96,996 (24)
PCI	221 (33)	85 (16)	107 (12)	32 (9)	25 (14)	43 (14)	202 (20)	94,921 (23)
HF	510 (76)	359 (68)	662 (71)	240 (69)	115 (65)	210 (69)	712 (70)	298,673 (73)
CVA/TIA	127 (19)	86 (16)	149 (16)	58 (17)	23 (13.)	30 (10)	135 (13)	61,518 (15)
AF/AFL	17 (3)	44 (8)	42 (5)	22 (6)	9 (5)	25 (8)	51 (5)	23,752 (5.8)
CKD	252 (38)	194 (37)	410 (44)	94 (27)	65 (37)	118 (39)	379 (37)	100,365 (26)
Valve disease	85 (13)	114 (22)	159 (17)	66 (19)	34 (19)	63 (21)	148 (15)	74,921 (19)
Female
N	466	436	665	348	184	261	919	358,804
Age, y	73.6 ± 11.8	81.3 ± 12.4	73.7 ± 14.2	84.4 ± 11.5	79.5 ± 12.6	77.2 ± 11.8	74.6 ± 13.6	77.0 ± 13.0
Insurance
Medicare	133 (34)	215 (54)	285 (46)	200 (66)	67 (39)	102 (42)	354 (41)	199,771 (61)
Medicaid	75 (19)	93 (23)	176 (28)	64 (21)	37 (21)	51 (21)	235 (27)	71,986 (22)
Private	159 (41)	86 (21)	148 (24)	39 (13)	66 (38)	88 (36)	240 (28)	50,562 (15)
Self-pay or uninsured	19 (5)	7 (2)	11 (2)	0 (0)	2 (1)	4 (2)	22 (3)	3,695 (1)
Prior history
Diabetes	292 (63)	194 (45)	405 (61)	151 (43)	101 (55)	150 (58)	523 (57)	141,579 (40)
Hyperlipidemia	279 (60)	255 (59)	464 (70)	221 (64)	90 (49)	160 (61)	595 (65)	205,129 (52)
Hypertension	422 (91)	363 (83)	568 (85)	293 (84)	156 (85)	227 (87)	818 (89)	289,689 (81)
Smoking history	9 (2)	11 (3)	33 (5)	17 (5)	9 (5)	5 (12)	41 (5)	41,121 (12)
CABG	79 (17)	21 (5)	60 (9)	23 (7)	18 (10)	35 (13)	116 (13)	42,108 (12)
MI	73 (16)	33 (8)	129 (19)	46 (13)	27 (15)	52 (20)	176 (19)	59,845 (17)
PCI	98 (21)	35 (8)	69 (10)	22 (6)	17 (9)	31 (12)	133 (15)	57,094 (16)
HF	347 (75)	272 (62)	455 (68)	237 (68)	129 (70)	179 (69)	634 (69)	254,221 (70.9)
CVA/TIA	77 (17)	64 (15)	107 (16)	54 (16)	23 (13)	30 (12)	143 (16)	60,920 (17)
AF/AFL	14 (3)	18 (4)	31 (5)	14 (4)	5 (3)	10 (4)	35 (4)	16,044 (4.5)
CKD	153 (33)	140 (32)	304 (46)	79 (23)	53 (29)	113 (43)	325 (35)	73,406 (22)
Valve disease	76 (16)	106 (24)	141 (21)	81 (23)	40 (22)	69 (26)	161 (18)	75,896 (22)

Values are N, mean ± SD, or n (%, rounded to the nearest whole percent). Valve disease defined per the case report form as history of valvular heart disease, prior tricuspid valve procedure, prior transcatheter mitral valve replacement, or prior transcatheter aortic valve replacement.

AF/AFL = atrial fibrillation/atrial flutter; CABG = coronary artery bypass graft; CKD = chronic kidney disease (defined as chronic renal insufficiency with chronic serum creatinine >2.0 or kidney transplant or chronic dialysis); CVA/TIA = cerebrovascular accident/transient ischemic attack; HF = heart failure; MI = myocardial infarction; NHW = non-Hispanic White; PCI = percutaneous coronary intervention.

Admission and hospitalization clinical characteristics across Asian and non-Hispanic White patients are shown stratified according to sex in Table 2; characteristics overall are presented in Supplemental Table 1. In male patients with HF, the prevalence of HFrEF (LVEF ≤40%) ranged from 48% in Chinese patients to 59% in Asian Indian and Japanese patients, and it was 50% in non-Hispanic White patients. In female patients with HF, the prevalence of HFrEF ranged from 27% in Asian Indian patients to 38% in Korean patients, and it was 28% in non-Hispanic White patients.

TABLE 2.

Admission and Hospitalization Clinical Characteristics of U.S. Asian Patients With HF, Stratified According to Sex

	Asian Indian	Chinese	Filipino/Filipina	Japanese	Korean	Vietnamese	Other Asian	NHW
Male
N	677	525	933	350	176	304	1,017	409,762
Point of origin
Clinic	48 (7)	38 (7)	77 (8)	22 (6)	10 (6)	19 (6)	51 (5)	15,257 (4)
ED	441 (65)	355 (68)	604 (65)	200 (57)	108 (61)	188 (62)	687 (68)	166,726 (41)
Another health care facility	44 (7)	28 (5)	72 (8)	30 (9)	14 (8)	23 (8)	66 (7)	28,029 (7)
BMI, kg/m2	26.7 ± 5.	25.5 ± 5.6	28.5 ± 6.7	27.5 ± 7.3	25.6 ± 5.2	24.6 ± 4.3	27.3 ± 6.6	31.3 ± 8.4
Systolic BP, mm Hg	143 ± 32	138 ± 26	141 ± 29	137 ± 28	138 ± 29	138 ± 30	144 ± 31	137 ± 28
Diastolic BP, mm Hg	79 ± 18	77 ± 18	82 ± 20	80 ± 19	79 ± 18	79 ± 19	83 ± 21	78 ± 18
Creatinine, mg/dL	1.6 (1.1–2.8)	1.6 (1.2–2.3)	1.7 (1.2–2.7)	1.6 (1.2–2.5)	1.6 (1.1–3.0)	1.5 (1.1–2.6)	1.6 (1.2–2.5)	1.4 (1.0–1.9)
ECG QRS duration, ms	115 ± 32	116 ± 22	110 ± 29	117 ± 32	115 ± 32	106 ± 27	113 ± 32	123 ± 35
Sodium, mEq/L	136 ± 5.2	137 ± 4.8	137 ± 4.6	137 ± 4.9	137 ± 4.9	136 ± 5.3	137 ± 4.7	138 ± 4.6
Ejection fraction
LVEF #40%	402 (59)	253 (48)	486 (52)	205 (59)	97 (55)	155 (51)	553 (54)	205,454 (50)
LVEF 41%−49%	67 (10)	45 (9)	84 (9)	29 (8)	20 (11)	37 (12)	103 (10)	46,486 (11)
LVEF $50%	208 (31)	227 (43)	363 (39)	116 (33)	59 (34)	112 (37)	361 (36)	157,822 (39)
Female
N	466	436	665	348	184	261	919	358,804
Point of origin
Clinic	19 (4)	39 (9)	46 (7)	14 (4)	8 (4)	14 (5)	31 (3)	10,462 (3)
ED	343 (74)	287 (66)	455 (68)	187 (54)	122 (66)	177 (68)	667 (73)	151,742 (42)
Another health care facility	27 (6)	30 (7)	66 (10)	26 (8)	15 (8)	12 (5)	54 (6)	25,613 (7)
BMI, kg/m2	27.1 ± 6.5	25.3 ± 6.8	27.7 ± 8.4	24.9 ± 6.6	25.2 ± 6.5	24.8 ± 5.1	26.6 ± 7.2	31.8 ± 10.3
Systolic BP, mm Hg	151 ± 33	142 ± 27	147 ± 31	143 ± 28	147 ± 32	150 ± 30	150 ± 32	143 ± 29
Diastolic BP, mm Hg	77 ± 17	74 ± 17	77 ± 19	75 ± 18	80 ± 19	77 ± 19	78 ± 18	76 ± 18
Creatinine, mg/dL	1.3 (0.9–2.3)	1.3 (0.9–2.0)	1.5 (1.1–2.8)	1.2 (0.9–2.1)	1.2 (0.9–2.2)	1.4 (1.0–2.7)	1.4 (0.9–2.3)	1.1 (0.9–1.6)
ECG QRS duration, ms	100 ± 28	102 ± 29	102 ± 29	101 ± 28	105 ± 34	101 ± 29	100 ± 28	110 ± 33
Sodium, mEq/L	136 ± 5.4	136 ± 6.3	136 ± 5.5	136 ± 5.6	137 ± 5.9	135 ± 6.5	136 ± 5.4	138 ± 4.8
Ejection fraction
LVEF #40%	124 (27)	129 (30)	189 (28)	106 (31)	69 (38)	72 (28)	272 (30)	101,648 (28)
LVEF 41%−49%	29 (6)	38 (9)	55 (8)	25 (7)	15 (8)	24 (9)	24 (9)	33,943 (10)
LVEF $50%	313 (67)	269 (61)	421 (63)	217 (62)	100 (54)	165 (63)	165 (63)	223,213 (62)

Values are N, frequency (%), mean ± SD, or median (Q1-Q3).

BMI = body mass index; BP = blood pressure; ECG = electrocardiogram; ED = emergency department; LVEF = Left ventricular ejection fraction; other abbreviations as in Table 1.

Prehospitalization medication use and in-hospital treatments are shown stratified according to sex in Table 3 and overall in Supplemental Table 2. Among male patients, prehospitalization use of an ACE inhibitor, ARB, or ARNI ranged from 40% in Japanese patients to 51% in Asian Indian patients and 40% in non-Hispanic White patients; use of a beta-blocker ranged from 61% in Vietnamese patients to 72% in Asian Indian patients and 64% in non-Hispanic White patients; and use of an MRA ranged from 7% in Japanese patients to 12% in Korean patients and 15% in non-Hispanic White patients. Among female patients, prehospitalization use of an ACE inhibitor, ARB, or ARNI ranged from 41% in Asian Indian, Japanese, and Vietnamese patients to 45% in Chinese patients and 38% in non-Hispanic White patients; use of a beta-blocker ranged from 56% in Korean patients to 72% in Asian Indian and Filipina patients and 64% in non-Hispanic White patients; and use of an MRA ranged from 3% in Filipina patients to 11% in Chinese patients and 13% in non-Hispanic White patients. Most patients did not have coronary, valve, or rhythm device procedures or renal replacement therapy while admitted, with variation in frequency across groups.

TABLE 3.

Prehospitalization Medication Use and In-Hospital Treatments Among U.S. Asian Patients With Any HF, Stratified According to Sex

	Asian Indian	Chinese	Filipino/Filipina	Japanese	Korean	Vietnamese	Other Asian	NHW
Male
N	677	525	933	350	176	304	1,017	409,762
Prehospitalization medication use
ACE inhibitor, ARB, or ARNI	215 (51)	151 (44)	264 (42)	92 (40)	70 (52)	110 (46)	363 (48)	65,122 (40)
Beta-blocker	306 (72)	220 (64)	424 (67)	143 (62)	84 (62)	147 (61)	476 (63)	103,465 (64)
MRA	41 (10)	33 (10)	49 (8)	17 (7)	16 (12)	22 (9)	99 (13)	24,599 (15)
In-hospital treatments
No procedure	256 (64)	267 (77)	412 (66)	153 (77)	81 (58)	165 (67)	506 (64)	162,808 (72)
Coronary interventiona	20 (5)	4 (1)	9 (1)	3 (2)	5 (4)	3 (1)	22 (3)	5,319 (2)
Valve interventionb	2 (1)	2 (1)	1 (0.2)	1 (1)	1 (1)	1 (0.4)	6 (1)	1,019 (1)
Renal replacement therapyc	32 (8)	19 (6)	81 (13)	15 (8)	19 (14)	31 (13)	101 (13)	11,098 (5)
Rhythm deviced	12 (3)	2 (1)	8 (1)	1 (1)	3 (2)	5 (2)	18 (2)	5,684 (3)
Female
N	466	436	665	348	184	261	919	358,804
Prehospitalization medication use
ACE inhibitor, ARB, or ARNI	111 (41)	127 (45)	196 (44)	87 (41)	56 (44)	87 (41)	300 (44)	53,210 (38)
Beta-blocker	197 (72)	175 (62)	322 (72)	127 (60)	70 (56)	128 (60)	451 (66)	89,498 (64)
MRA	18 (7)	31 (11)	12 (3)	10 (5)	10 (8)	20 (9)	54 (8)	17,745 (13)
In-hospital treatments
No procedure	179 (69)	206 (75)	329 (70)	171 (82)	100 (70)	147 (69)	507 (69)	154,018 (78)
Coronary interventiona	3 (1)	2 (1)	4 (1)	1 (1)	2 (1)	4 (2)	15 (2)	2,839 (1)
Valve interventionb	1 (0.4)	1 (0.4)	1 (0.2)	0 (0)	0 (0)	1 (1)	2 (0.3)	638 (0.3)
Renal replacement therapyc	23 (9)	14 (5)	61 (13)	10 (5)	14 (10)	35 (16)	92 (13)	7,717 (4)
Rhythm deviced	4 (2)	5 (2)	3 (1)	4 (2)	2 (1)	3 (1)	13 (2)	3,041 (2)

Values are N or n (%). Frequencies and percentages are among patients with any ejection fraction, with denominators excluding patients with missing data for each metric.

^aIncludes PCI, PCI with stent, or CABG.

^bIncludes transcatheter mitral valve replacement, tricuspid valve procedure, cardiac valve surgery, or transcatheter aortic valve replacement.

^cIncludes ultrafiltration, dialysis or ultrafiltration unspecified, or dialysis.

^dRhythm device includes cardiac resynchronization therapy-pacing only, implantable cardioverter-defibrillator, pacemaker, or cardiac resynchronization therapy with implantable cardioverter-defibrillator.

ACE = angiotensin-converting enzyme; ARB = angiotensin II receptor blocker; ARNI = angiotensin receptor-neprilysin inhibitor; MRA = mineralocorticoid receptor antagonist; other abbreviations as in Table 1.

Discharge measures, including medication contraindication frequency among HFrEF patients, prescription of OMT for HFrEF, discharge laboratory values, and discharge disposition, are summarized according to sex in Table 4 and overall in Supplemental Table 3. The proportion of HFrEF patients with contraindications to each medication class was significantly different across race and ethnicity groups (Figure 2). Rates of OMT on discharge across groups are shown in Figure 3. The adjusted odds of receiving OMT on discharge in all Asian subjects and separately, in individual Asian groups, compared with non-Hispanic White patients is shown in Figure 4 and listed in Supplemental Table 4. Prescription of beta-blockers was generally highest for those eligible (male patients: 95%−98%; female patients: 93%−97%), followed by ACE inhibitors, ARBs, or ARNIs (male patients: 84%−90%; female patients: 66%−94%); prescription of MRAs was lowest (male patients: 37%−53%; female patients: 31%−57%). Among Asian patients, the rate of receiving OMT ranged from 31% in Japanese and Vietnamese patients to 58% in Asian Indian male patients (40% in non-Hispanic White male patients), and from 30% in Korean patients to 54% in Asian Indian female patients (38% in non-Hispanic White female patients). Relative to non-Hispanic White patients, the adjusted odds of receiving OMT at discharge were lower in Vietnamese (aOR: 0.68; 95% CI: 0.47–1.00) male patients.

TABLE 4.

Discharge Measures among U.S. Asian Patients With HFrEF, Stratified According to Sex

	Asian Indian	Chinese	Filipino/Filipina	Japanese	Korean	Vietnamese	Other Asian	NHW
Male
N	677	525	933	350	176	304	1,017	409,762
Contraindications among patients with HFrEF
ACE inhibitor, ARB, or ARNI	381 (65)	249 (54)	460 (56)	179 (56)	70 (45)	121 (46)	429 (46)	151,919 (46)
Beta-blocker	42 (6)	68 (14)	75 (8)	46 (14)	17 (10)	31 (11)	89 (9)	35,471 (9)
MRA	305 (46)	168 (33)	263 (29)	123 (36)	43 (26)	94 (33)	271 (28)	84,470 (23)
Discharge prescription of OMT among patients with HFrEF without a contraindication
OMTa	221 (58)	98 (43)	158 (35)	58 (31)	33 (38)	43 (31)	194 (38)	74,263 (40)
ACE inhibitor, ARB, or ARNI	87 (89)	73 (84)	149 (84)	66 (87)	33 (89)	60 (90)	208 (86)	61,333 (83)
Beta-blocker	360 (98)	208 (95)	409 (96)	162 (96)	81 (98)	120 (95)	459 (96)	168,552 (96)
MRA	68 (43)	47 (44)	104 (41)	35 (37)	29 (53)	37 (49)	148 (52)	52,775 (53)
Discharge laboratory values
Creatinine, mg/dL	1.5 (1.1–2.6)	1.7 (1.1–2.3)	1.7 (1.2–2.7)	1.5 (1.1–2.2)	1.5 (1.1–3.6)	1.5 (1.1–2.6)	1.5 (1.1–2.4)	1.3 (1.0–1.9)
Potassium, mEq/L	4.2 (3.9–4.5)	4.0 (3.7–4.3)	4.0 (3.7–4.3)	4.1 (3.7–4.4)	4.0 (3.6–4.4)	4.0 (3.7–4.3)	4.1 (3.8–4.4)	4.0 (3.7–4.3)
Discharge disposition
Home	573 (85)	429 (82)	808 (87)	273 (78)	146 (83)	242 (80)	864 (85)	308,396 (75)
Hospice	11 (2)	16 (3)	24 (3)	18 (5)	4 (2)	14 (5)	21 (2)	17,645 (4)
Health care facility or Against medical advice	77 (11)	61 (12)	70 (8)	48 (14)	19 (11)	35 (12)	106 (10)	71,568 (18)
Female
N	466	436	665	348	184	261	919	358,804
Contraindications among patients with HFrEF
ACE inhibitor, ARB, or ARNI	217 (55)	194 (53)	286 (48)	165 (53)	71 (43)	92 (38)	332 (40)	104,901 (37)
Beta-blocker	29 (6)	64 (16)	63 (10)	46 (14)	23 (13)	19 (8)	70 (8)	26578 (8)
MRA	168 (37)	122 (30)	151 (23)	113 (34)	53 (29)	72 (29)	220 (25)	55755 (17)
Discharge prescription of OMT among patients with HFrEF without a contraindication
OMTa	64 (54)	44 (38)	64 (36)	40 (44)	18 (30)	22 (32)	91 (36)	35,646 (38)
ACE inhibitor, ARB, or ARNI	32 (91)	30 (75)	53 (79)	19 (66)	26 (87)	31 (94)	94 (80)	31,981 (83)
Beta-blocker	113 (97)	103 (93)	161 (96)	82 (97)	53 (93)	63 (94)	233 (96)	84,450 (96)
MRA	27 (47)	17 (31)	35 (41)	15 (35)	13 (38)	17 (57)	57 (48)	23,277 (50)
Discharge laboratory values
Creatinine, mg/dL	1.2 (0.9–2.0)	1.3 (0.9–2.0)	1.5 (1.0–2.8)	1.3 (0.9–2.1)	1.3 (0.9–2.4)	1.3 (1.0–2.4)	1.3 (0.9–2.3)	1.1 (0.9–1.6)
Potassium, mEq/L	4.2 (3.8–4.5)	4.0 (3.7–4.4)	4.00 (3.6–4.4)	4.0 (3.7–4.4)	4.1 (3.7–4.4)	4.1 (3.8–4.4)	4.0 (3.7–4.4)	4.0 (3.7–4.3)
Discharge disposition
Home	409 (88)	316 (73)	575 (87)	251 (72)	145 (79)	230 (88)	742 (81)	246,283 (69)
Hospice	10 (2)	27 (6)	25 (4)	22 (6)	12 (7)	9 (3)	32 (4)	17,768 (5)
Health care facility or against medical advice	38 (8)	76 (17)	53 (8)	64 (18)	25 (14)	16 (6)	118 (13)	85,233 (24)

Values are N or frequency (%) with denominators excluding patients with missing data for each metric or median (Q1-Q3).

^aOptimal medical therapy (OMT) defined as being prescribed all medications in a class for which a patient is eligible; percentage represents patients prescribed the medication class(es) among the group of patients eligible for that/those class(es). Medication classes qualifying for optimal triple therapy are: 1) ACE inhibitor or ARB or ARNI; 2) beta-blocker; and 3) MRA. Eligibility criteria for ACE inhibitor/ARB/ARNI and beta-blocker classes are an ejection fraction #40% with no documented contraindication. Eligibility criteria for the MRA class was an ejection fraction #35% with no documented contraindication.

HFrEF = heart failure with reduced ejection fraction; other abbreviations as in Tables 1 and 3.

Figure 2.

Proportion of Patients With HFrEF With a Documented Medication Contraindication

P values represent chi-square tests to evaluate differences in proportion with contraindications across race and ethnicity groups. Data presented in Supplemental Table 3. ACEi = angiotensin-converting enzyme inhibitor; ARB = angiotensin II receptor blocker; ARNI = angiotensin receptor-neprilysin inhibitor; HFrEF = heart failure with reduced ejection fraction; MRA = mineralocorticoid receptor antagonist.

Figure 3.

Proportion of Patients With HFrEF Who Received Optimal Medical Therapy

Optimal medical therapy was defined as being prescribed all medications in a class for which a patient is eligible and for which there was no documented contraindication. Medication classes qualifying for optimal medical therapy were: 1) an ACEi or ARB or ARNI for those with an ejection fraction (EF) ≤40%; 2) beta-blocker for those with EF ≤40%; and 3) MRA for those with EF ≤35%. Data are presented in Supplemental Table 3. Abbreviations as in Figure 2.

Figure 4.

Heart Failure Quality of Care in U.S. Asian Patients Compared With Non-Hispanic White Patients

ORs represent likelihood of experiencing quality of care metric compared with non-Hispanic White patients, adjusted for age, left ventricular ejection fraction, diabetes history, hypertension history, smoking history, and insurance status. *CIs that exclude 1.0 indicate statistically significant differences compared with non-Hispanic White patients. ORs and 95% CIs are listed in Supplemental Table 4 (optimal medical therapy), Supplemental Table 6 (length of stay >4 days), and Supplemental Table 8 (defect-free care). **P values represent Wald test P values for overall heterogeneity in outcomes specifically among individual Asian groups.

Hospitalization quality of care measures stratified according to sex are summarized in Table 5 and overall in Supplemental Table 5; adjusted odds of outcomes are shown in Figure 4 for LOS >4 days and defect-free care. Corresponding data for adjusted odds are summarized in Supplemental Table 6 (LOS >4 days in any HF), Supplemental Table 7 (LOS >4 days in HFrEF), Supplemental Table 8 (defect-free care in any HF), and Supplemental Table 9 (in-hospital mortality in any HF). Relative to non-Hispanic White patients with any HF, the adjusted odds of an LOS >4 days were lower in Vietnamese male patients (aOR: 0.68; 95% CI: 0.50–0.91) and Filipina female patients (aOR: 0.66; 95% CI: 0.46–0.95). Based on the Wald test global P value, there was statistically significant heterogeneity (P < 0.01) in covariate-adjusted LOS >4 days between Asian groups among women. Relative to non-Hispanic White patients with any HF, the adjusted odds of defect-free care were lower in Filipina female patients (aOR: 0.52; 95% CI: 0.34–0.82). In-hospital mortality among all Asian groups was similar compared with non-Hispanic White patients. Based on the Wald test global P value, there was statistically significant heterogeneity (P < 0.01) in covariate-adjusted defect-free care between Asian groups among women.

TABLE 5.

Hospitalization Quality of Care and Outcomes Among U.S. Asian Patients With Any HF, Stratified According to Sex

	Asian Indian	Chinese	Filipino/Filipina	Japanese	Korean	Vietnamese	Other Asian	NHW
Male
N	677	525	933	350	176	304	1,017	409,962
In-hospital mortality	15 (2)	18 (3)	28 (3)	11 (3)	7 (4)	13 (4)	26 (3)	11,942 (3)
Length of stay >4 d	239 (41)	179 (39)	287 (35)	116 (40)	60 (40)	87 (33)	353 (39)	136,341 (41)
Referral to cardiac rehabilitation	50 (11)	24 (7)	49 (8)	13 (5)	9 (7)	26 (11)	112 (14)	30,309 (13)
Referral to outpatient HF management program	337 (53)	233 (47)	365 (41)	103 (31)	71 (43)	130 (47)	413 (42)	151,813 (42)
Provision of $60 min of HF education	358 (56)	239 (48)	369 (42)	71 (22)	74 (46)	160 (57)	437 (45)	161,910 (45)
Received HF workbook	112 (18)	92 (18)	107 (12)	26 (8)	40 (25)	75 (27)	274 (28)	91,815 (25)
Any postdischarge follow-up visit scheduled	534 (81)	416 (82)	725 (80)	247 (73)	116 (69)	221 (76)	704 (71)	287,614 (73)
Postdischarge follow-up visit scheduled, within #7 d of discharge	370 (56)	321 (64)	584 (65)	178 (53)	92 (56)	173 (60)	510 (52)	214,385 (55)
Smoking cessation counselinga	75 (97)	48 (89)	145 (95)	40 (93)	26 (93)	43 (86)	147 (87)	67,577 (88)
Defect-free HF careb	111 (24)	48 (14)	57 (9)	19 (7)	13 (13)	20 (11)	95 (13)	35,129 (12)
Influenza vaccinationc	301 (45)	280 (54)	483 (52)	192 (55)	95 (55)	149 (50)	475 (47)	185,240 (48)
Pneumococcal vaccinationc	362 (54)	344 (66)	623 (67)	233 (67)	119 (68)	198 (66)	592 (59)	242,779 (62)
Female
N	466	436	665	348	184	261	919	358,804
In-hospital mortality	9 (2)	17 (4)	12 (2)	10 (3)	2 (1)	6 (2)	27 (3)	9,409 (3)
Length of stay >4 d	161 (38)	129 (37)	188 (33)	106 (40)	57 (37)	100 (42)	330 (42)	108,559 (41)
Referral to cardiac rehabilitation	34 (11)	16 (6)	34 (8)	12 (5)	8 (6)	16 (8)	81 (11)	20,071 (10)
Referral to outpatient HF management program	186 (42)	138 (34)	248 (38)	74 (22)	67 (38)	116 (47)	300 (34)	122,959 (38)
Provision of $60 min of HF education	241 (54)	162 (40)	271 (42)	71 (23)	70 (41)	124 (50)	361 (42)	137,713 (43)
Received HF workbook	73 (17)	76 (19)	60 (9)	27 (9)	41 (24)	69 (28)	206 (24)	79,240 (25)
Any postdischarge follow-up visit scheduled	356 (78)	300 (72)	509 (78)	239 (71)	120 (66)	196 (77)	601 (68)	244,510 (71)
Postdischarge follow-up visit scheduled, within #7 d of discharge	231 (51)	228 (55)	391 (61)	180 (54)	93 (52)	160 (63)	436 (49)	182,772 (54)
Smoking cessation counselinga	6 (60)	11 (92)	38 (95)	15 (94)	8 (80)	4 (80)	42 (80)	42,690 (88)
Defect-free HF careb	25 (10)	20 (8)	19 (5)	14 (6)	9 (9)	9 (7)	47 (9)	16,444 (7)
Influenza vaccinationc	224 (49)	251 (58)	358 (54)	207 (60)	91 (50)	150 (58)	478 (53)	170,443 (51)
Pneumococcal vaccinationc	249 (54)	281 (65)	507 (77)	265 (76)	103 (56)	172 (66)	584 (64)	229,496 (66)

Values are N or frequency (%).

^aFrequency (%) of patients who received smoking cessation counseling, eligible to achieve this metric (denominator) is patients who have a history of smoking.

^bDefect-free HF care is defined as experiencing as many of 5 quality of care measures that a patient is eligible for; each patient may be eligible for a different number of these quality measures. A patient who is eligible for all 5 measures and experiences all 5 is considered as having received “defect-free care.” A patient who is eligible for only 2 measures and experiences these 2 is also considered as having received “defect-free care.”

^cInfluenza vaccination defined as vaccine given during this hospitalization during the current flu season, or received before admission during the current flu season but not during this hospitalization; pneumococcal vaccination defined as vaccine given during this hospitalization or received in the past not during this hospitalization.

Abbreviations as in Table 1.

The sensitivity analysis of the primary outcomes, evaluated only in the 347 hospitals that contributed ≥10 Asian patients to the registry, showed similar findings as the primary analysis (Supplemental Table 10). There were lower odds of LOS >4 days among Filipino and Vietnamese patents, lower odds of OMT in Vietnamese patients, and lower odds of defect-free care in Filipino patients, all compared with non-Hispanic White patients. In the first secondary analysis comparing patients from an aggregated East Asian and aggregated Southeast Asian group vs non-Hispanic White patients (Supplemental Table 11), LOS >4 days and defect-free care was less likely in Southeast Asian patients compared with non-Hispanic White patients. In the second secondary analysis of 21,671 aggregated Asian patients who did not have further individual Asian groups specified (characteristics listed in Supplemental Table 12), men with HFrEF had higher odds of OMT at discharge (aOR: 1.17; 95% CI: 1.01–1.35; P = 0.03), men (aOR: 0.79; 95% CI: 0.73–0.87; P < 0.01) and women (aOR: 0.80; 95% CI: 0.73–0.88; P < 0.01) with any HF had lower odds of LOS >4 days, and women with any HF had lower odds of defect-free care (aOR: 0.81; 95% CI 0.67–0.99; P = 0.04), compared with non-Hispanic White patients. There were no other statistically significant ORs.

Discussion

In this national quality of inpatient care registry of patients with HF in the United States, there were important differences in HF quality of care and process measures among specific groups of Asian adults compared with White adults (Central Illustration). Among HFrEF patients, there was clinically meaningful variation among Asian groups in rates of OMT prescribed at discharge. Specific groups of Asian patients with HF were less likely to have a longer LOS (>4 days) and receive defect-free care accounting for both HF management process measures and pharmacologic therapy; there was significant variation across individual U.S. Asian groups in these measures, particularly among women. Inpatient mortality rates for the assessed hospitalization were reassuringly low for all groups, with no differences experienced by Asian patients. These data align with observed differences experienced by groups of U.S. Asian patients for quality of care and outcomes for myocardial infarction,¹³ and they suggest that differences in HF quality of care may contribute to disparities in CVD experienced by Asian populations in the United States.

Given the strong evidence for, and well-established clinical guidelines recommending the use of ACE inhibitors, ARBs, or ARNIs, beta-blockers, and MRAs for HFrEF,²⁴ the lower likelihood of receiving OMT at discharge in specific Asian groups is worrisome. Furthermore, OMT is one component of defect-free care, which also accounts for HF education and hospital follow-up within 7 days scheduled at the time of discharge. Overall rates of OMT and defect-free care were low for patients in all groups, signaling that evidence-based quality improvement efforts are necessary for all patients, although specific Asian groups were observed to have a lower likelihood of experiencing defect-free care compared with White patients.

The observed differences may arise in part due to differences in the medical history of patients across groups, with variation noted in the prevalence of prior CVD risk factors, coronary artery disease and associated interventions, and valve diseases. In addition, there were significant differences in frequency of contraindications of OMT across groups. Clinician- and health system–level barriers likely also contribute to the observed differences in OMT and defect-free care rates experienced by several groups of Asian patients. Medication reporting on the GWTG-HF registry case report form reflects prescription of these medication classes and not medication adherence, and clinical inertia is a well-documented cause of suboptimal guideline-directed medical therapy rates for all patients with HFrEF.²⁵ Provision of HF education and scheduling follow-up appointments are driven by clinician referral during admission.

The observed differences may also arise in part due to patient-level social factors such as health literacy or English language barriers, which may compound clinician time constraints that lead to inadequate explanation of the purpose of OMT among such patients. Although English proficiency is only one potential factor, limited English language proficiency is present in approximately 48% of Vietnamese, 42% of Chinese, 38% of Korean, 22% of Japanese, 20% of Filipino, and 18% of Asian Indian individuals in the United States according to the 2017 to 2021 American Community Survey.²⁶ Therefore, strategies to overcome language barriers are likely necessary to optimize HF care.²⁷ Furthermore, it remains to be clarified in this group of patients the extent to which financial barriers contribute to differences in OMT prescription, given the relatively high rates of health insurance in all groups of patients in the registry and availability in the United States of generic forms of medications for each of the three major medication classes evaluated. In addition, regional and health system variations in OMT prescription may be present, which was not evaluated in this analysis due to relatively limited sample sizes in Asian groups at regional levels.

LOS is another HF care metric that is evaluated in the GWTG program. In a prior study from this registry, Black, Hispanic, and Indigenous American and Pacific Islander patients were observed to have longer LOS compared with non-Hispanic White patients, whereas patients who were Asian (in aggregate) did not have differences in LOS compared with White patients.¹⁷ Registry data indicate that patients with HF who have longer LOS tend to have more comorbidities and higher disease severity.²⁸ Our analysis showed heterogeneity in LOS in Asian groups that was masked with aggregation of patients into a single Asian category, with Vietnamese male patients and Filipina female patients experiencing a lower likelihood of an LOS >4 days compared with non-Hispanic White patients. Reasons for the lower likelihood of an LOS >4 days are likely multiple. Shorter LOS may be favorable in some instances (eg, with less severe HF presentation) or may be indicative of suboptimal care and premature discharge. The observed lower likelihood of OMT and defect-free care in certain Asian groups suggests that patients in these groups may experience premature discharge with suboptimal HF management. A post hoc analysis of the EVEREST (Efficacy of Vasopressin Antagonism in Heart Failure Outcome Study with Tolvaptan) trial among participants in international settings showed that a shorter LOS was associated with a higher risk of HF readmissions within 30 days of discharge.²⁹ Although there were reassuringly no differences in rates of in-hospital mortality in Asian patients compared with White patients, whether the shorter LOS in specific Asian groups compared with White patients results in worse postdischarge outcomes for these patients remains to be investigated.

Some of the observed differences may be related to variation across hospitals in care practices and adherence to guideline-directed HF management. Such variability may potentially be harmful or beneficial for HF quality of care. For instance, the ORs for likelihood of experiencing OMT or defect-free care among Asian Indian patients are >1, although these ORs are not statistically significant. The higher likelihood of experiencing these quality metrics among Asian Indian patients is plausible, as enhanced attention in certain hospitals or regions for optimizing HF management in recognition of the disproportionate burden of CVD experienced by South Asian individuals may potentially result in more favorable outcomes for patients in this group.^30–32 However, whether larger patient samples confirm these findings remains to be evaluated. Although data from Asian regions are limited, our findings also suggest differences in OMT rates compared with Asian populations outside the United States. In the ASIAN-HF (Asian Sudden Cardiac Death in Heart Failure) registry of >5,000 adults with symptomatic HFrEF between 2012 and 2015 in both inpatient and outpatient settings from 46 centers across 11 countries in Asia, the mean prescription rates were 77% for ACE inhibitors or ARBs, 79% for beta-blockers, and 58% for MRA with variation across countries.³³ Although the GWTG-HF registry data generally indicate more favorable use of ACE inhibitors or ARBs and beta-blockers among Asian American patients compared with counterparts in Asian countries, MRA use was generally lower. Notably, the medication use data from the ASIAN-HF registry data are older than our analysis from 2015 to 2023, and contemporary medication use data in Asian countries are needed.³⁴

The available data collected in this registry limit the extent to which upstream determinants, including social risk factors, contribute to observed quality of care differences. However, variation in a range of social determinants of health likely underlie many of the observed differences among Asian groups¹¹; these include the lower odds of OMT and LOS >4 days observed among Vietnamese male patients compared with non-Hispanic White patients and the lower odds of LOS >4 days and defect-free care among Filipina female patients compared with non-Hispanic White patients. In addition to health literacy and language accessibility, examples of such upstream factors may include the influences of educational attainment, place of birth and immigration status, and social support, which each have been shown to influence health outcomes across diverse Asian and other predominantly immigrant populations.³⁵ Differences in acculturation as measured by years lived in the United States, culturally influenced health behaviors (including dietary patterns and physical activity practices), and cultural norms related to health care and medication use may also contribute to the observed differences in HF presentation, care, and outcomes across Asian groups by influencing the prevalence of HF risk factors, medication acceptability and adherence, and access to care.¹¹ Although the GWTG-HF registry data preclude direct investigation of the roles of these factors in HF care across Asian groups, many of these factors have been linked to differences in cardiovascular health and outcomes, including among Vietnamese³⁶ and Filipino³⁷ populations in the United States.

These factors warrant more detailed investigation. The extent to which institutional and systemic discrimination and racism contribute to observed differences in quality of HF care remains to be investigated. They are hypothesized, however, to also be a potential contributor, given the role this structural determinant plays in health status and outcomes across a range of cardiovascular and other health conditions.³⁸

Study limitations.

Although this analysis represents a national-level evaluation of HF quality of care for a patient group persistently underrepresented in clinical research, there are several limitations. First, the relatively small number of Asian patients with disaggregated identification captured during the approximately 8-year study period (7,261 patients with individual Asian group specified across 824 hospitals) suggests potential underidentification, misclassification, and/or misrepresentation of Asian patients in clinical data across hospital systems. This persistent misrepresentation may arise due to several compounding factors, including incomplete coding of race and ethnicity in EHR data, and lack of category options for patients to self-identify as Asian race or with individual Asian groups. This is evidenced in part by the large number of additional Asian patients categorized only as “Asian” without an individual Asian group specified, which remains common practice in many EHR systems. These health system–level factors exacerbate structural deficiencies of numerous clinical registry and surveillance programs, many of which are not designed to identify disparities due to aggregation of racial and ethnic groups. When considered together, Asian patients comprise a greater proportion of patients represented in the GWTG-HF registry (approximately 2.5%)¹⁴ compared with the proportion of hospitalized HF patients who are Asian in the US-based National Inpatient Sample (1.9%)³⁹ but slightly less than that represented in estimates of HF prevalence in US-based EHR-based data (2.9%).⁷ Such variation likely reflects differences across clinical settings in the consistency and reliability of identifying patients who are Asian.

Furthermore, the identification of Asian patients was not exclusively based on self-report. Race and ethnicity identification from identification listed in the EHR may be subject to misclassification.⁴⁰ Many Asian patients who may identify in a specific group may consequently have been uncategorized in an individual group or categorized as Other Asian. Modest Asian patient sample sizes, especially among female patients, may be underpowered to detect some differences in HF quality of care in Asian groups compared with non-Hispanic White patients.

Second, although we were able to statistically account for hospital-level clustering, in these data there were more clusters (824 hospitals) than there were patients in several individual Asian groups. Although this reflects the regional distribution of Asian populations, small variation in the outcomes can exacerbate estimated disparities and associated SEs, which may obfuscate racial and ethnic differences in outcomes and site-level variation in quality of care. Local investigation of HF quality of care differences within hospitals and health systems may support ensuring that patients receive equitable treatment. Third, observational data assume accurate and comprehensive completion of registry case report forms. For instance, missing data for body mass index (47%) and blood pressure (41%) limited our ability to additionally account for differences in these clinical factors as covariates in regression models to explore their contribution to differences across race and ethnicity groups. Fourth, although this registry represents clinical care in >800 U.S. hospitals, these data are not necessarily representative of all Asian patients in the United States. Fifth, the COVID-19 pandemic in 2020 and 2021 may have affected the quality and completeness of registry data, but these variations are not expected to differentially affect U.S. Asian patients. Sixth, data on several important social determinants of health (eg, household income, English language proficiency, place of birth, U.S. nativity) were not available, and thus their role in accounting for the observed differences in quality of care could not be evaluated. Seventh, SGLT2 inhibitor therapy was not evaluated because for much of the study period, this was not yet guideline-recommended therapy for HFrEF. Future investigation should evaluate a more comprehensive OMT definition incorporating SGLT2 inhibitor use, given prior data showing lower use among Asian patients compared with White patients with HFrEF in the GWTG-HF registry in 2021 to 2022⁴¹ and lower use among Asian patients compared with White patients with diabetes in national EHR-based data.⁴² Similarly, given the evolving nature of guidelines during the study period, medication prescription and outcomes were not specifically assessed in this study among patients with HFpEF. Finally, postdischarge follow-up data were not available, and thus this analysis focused only on hospitalization measures. Although in-hospital mortality did not differ for individual Asian groups, whether the differences in quality of care resulted in differences in longer term outcomes post-hospitalization, including re-hospitalization and mortality, remains to be investigated.

Conclusions and Future Directions

There is variation in HF quality of care and outcomes among Asian groups in the United States. Certain groups of Asian patients with HF may experience lower likelihood of OMT, defect-free care, and LOS >4 days compared with non-Hispanic White patients. Future investigation to identify the multilevel individual-, health system–, and community-level determinants of differences in HF care among Asian groups is necessary to inform clinical HF implementation and quality improvement strategies tailored for these communities. More broadly, the findings from this study plus the limitations of this analysis and the GWTG-HF registry itself for evaluating equity in HF for Asian populations in the United States underscore the importance of modernizing national-level CVD data infrastructure.

Several strategies may support the appropriate representation of U.S. Asian populations in health research⁴³ and enhance participation and representation in clinical trials.⁴⁴ Alongside ensuring more complete race and ethnicity data for all patients in EHR systems, implementing the capture of individual Asian ethnicity group data in the EHR should be a minimum standard for all health systems, particularly because Asian individuals are among the most misclassified in health care databases,⁴⁵ and most Asian patients in this registry do not have individual ethnicity group identification. National registries such as the AHA GWTG registry program and American College of Cardiology National Cardiovascular Data Registry program should allow for more comprehensive identification of patients in individual Asian groups across all the available CVD registries, as well as provide participating health systems with tools to identify and validate individual Asian group categorization among registry participants. Beyond registry data, although self-identified individual Asian group information has been collected over the last several years in many federal health surveys such as the National Health and Nutrition Examination Surveys, National Health Interview Survey, and Behavioral Risk Factor Surveillance System, access to these data is costly and often possible only through federal Research Data Centers; this therefore presents considerable barriers to comprehensive national-level investigation of CVD and other diseases among U.S. Asian populations.⁴⁶ Addressing the limitations of current data systems across local and national levels is necessary to enhance the visibility of heterogenous Asian populations, the accuracy of data for these groups, and ultimately to promote equity in cardiovascular surveillance and research.

Abbreviations and Acronyms

ACE

angiotensin-converting enzyme

AHA

American Heart Association

aOR

adjusted OR

ARB

angiotensin II receptor blocker

ARNI

angiotensin receptor-neprilysin inhibitor

CVD

cardiovascular disease

EF

ejection fraction

EHR

electronic health record

GWTG-HF

Get With the Guidelines–Heart Failure

HF

heart failure

HFpEF

heart failure with preserved ejection fraction

HFrEF

heart failure with reduced ejection fraction

IRB

Institutional Review Board

LOS

length of stay

LVEF

left ventricular ejection fraction

MRA

mineralocorticoid receptor antagonist

OMT

optimal medical therapy

SGLT2

sodium-glucose cotransporter-2