Benchmarking to evaluate heart failure mortality by race and ethnicity in the US from 2018 to 2023

INTRODUCTION

Despite increasing options for prevention and treatment of heart failure (HF),[1] HF-related mortality increased from 2012 to 2021.[2] At the same time, racial and ethnic differences in HF incidence,[3–6] access to lifesaving therapies,[7–10] and HF mortality have been observed.[3, 5]

Central to reversing these trends is examining the current state and setting ambitious improvement goals. One strategy endorsed by the 2022 HF guidelines to do this is benchmarking, the practice of comparing performance to a quality goal.[1] Typically, a disparities approach has been used to examine racial and ethnic variation in care, in which comparisons between perceived “advantaged” and “disadvantaged” populations, typically compared to a White population, are performed.[3–10] There is, however, an opportunity to apply an equity lens to benchmark HF outcomes by setting a goal that is not tied to a single population. An equity benchmark is aspirational by definition, used to quantify variation in outcomes but does not define the care processes to achieve the goal. Instead, it sets an outcome goal recognizing that structural inequities and differential needs exist that will require tailored strategies across populations and contexts. For example, the processes needed to implement evidence-based HF care may differ between rural and urban settings. The approach removes the assumption that outcomes are optimal in any population and is consistent with aspirational goals outlined in programs such as Healthy People 2030, which aims to set measurable improvement goals for policymakers and healthcare providers. Importantly, an equity benchmark sets a higher goal that is achievable as it is being accomplished. Once identified, gaps in outcomes can be addressed through targeted policies and programs such that people receive context-specific interventions to achieve the desired outcome.[11]

This study applies an equity benchmark approach to evaluate variation in HF mortality by race and ethnicity to characterize trends relative to an achievable benchmark goal and provide policymakers and providers with a framework to understand variation and track progress towards improved outcomes.

Methods

This population-based cohort study used the Centers for Disease Control and Prevention's Wide-ranging Online Data for Epidemiologic Research (CDC WONDER) multiple-cause-of-death data that includes death certificates for U.S. residents from 2018 to 2023.[12] HF-related deaths were identified using International Classification of Diseases, 10th Revision codes (ICD-10 codes I11.0, I13.0, I13.2, and I50)[2, 5] if HF was listed as the underlying or a contributing cause among adults aged ≥25 to 84 years old.[2, 5] The starting year of 2018 was selected as it was the first year a single-race format was recorded.

The primary outcome was annual age-adjusted HF mortality rates per 100,000 population. Racial and ethnic categories were grouped according to the National Institutes of Health recommendations: American Indian or Alaska Native (AI/AN), Asian, Black or African American, Hispanic or Latino, Native Hawaiian or Other Pacific Islander (NH/PI), or White.[13] States were ranked by age-adjusted HF mortality. The equity target was defined, using previous methodology, as the annual age-adjusted HF mortality of the state with the 5^th lowest value by year (10^th percentile).[11] The target was selected as an ambitious yet achievable goal that is already attained in the US, while minimizing the influence of outlier state performances, consistent with recommendations from the Agency for Healthcare Research and Quality, to provide concrete milestones for feasible improvement.[14] The target was allowed to vary annually to enable comparisons with the best currently achievable outcomes rather than a historical target that is more likely to be affected by broader healthcare changes (e.g., coding, COVID-19 pandemic). Age-adjusted mortality was selected, given known racial and ethnic differences in age of HF diagnosis and the associations of age on outcomes.[15]

Linear regression was performed to evaluate trends in HF mortality.[15] Trends in HF mortality and absolute and relative differences from the equity benchmark (gaps) were determined. Analyses were conducted utilizing Stata/MP version 19.0 (StataCorp).

Results

Between 2018 and 2023, there were 1,406,976 deaths with HF as a contributing cause in the US, resulting in a crude mortality rate of 106.1 (106.0 – 106.3) per 100,000 in the population. Age-adjusted HF mortality was stable, but inequities were observed for most racial and ethnic groups compared to the equity benchmark (Figure A and Table). The HF equity goal similarly did not increase (slope 2.6 per 100,000, 95% CI [-2.6, 7.7]. Mortality was consistently higher for AI/AN, Black, NH/PI, and White populations, with the largest gaps for Black and NH/PI populations. Asian and Hispanic populations had consistently lower mortality compared to the benchmark.

Figure. Temporal trends in age-adjusted heart failure mortality rates from 2018–2023 by age and race and ethnicity.

The figure shows the yearly adge adjusted heart failure mortality per 100,000 adults for (A) all ages, (B) ages 25 to 44, (C) ages 45 to 64, and (D) ages 65 to 84 years stratified by race and ethnicity. The solid blue line respresents the health equity benchmark defining the goal yearly HF mortality. Abreviations: AI/AN, American Indian or Alaska Native; NH/PI, Native Hawaiian or Other Pacific Islander

Table. Trends in mortality rates and equity gaps in heart failure from 2018–2023 by race and ethnicity stratified by age.

The table shows the rate of change in age-adjusted HF mortality, gap from the equity benchmark in 2018 and 2023, and the slope of the gap.

Race/Ethnicity	Slope 95% CI	P-value	Gap 2018 (95% CI)*	P-value	Gap 2023 (95% CI)	P-value	Gap Slope 95% CI*	P-value
All ages
HF Health Equity Goal	2.6 [−2.6, 7.7]	0.237
AI/AN	3.0 [−5.4, 11.4]	0.373	43.7 [38.1, 49.3]	0.000	40.2 [34.6, 45.8]	0.000	0.5 [−4.1, 5.0]	0.799
Asian	1.0 [−0.6, 2.6]	0.148	−26.9 [−28.4, −25.4]	0.000	−34.8 [−36.3, −33.3]	0.000	−1.6 [−5.9, 2.7]	0.373
Black (Non-Hispanic)	4.3 [−2.5, 11.1]	0.152	59.3 [57.5, 61.1]	0.000	66.6 [64.8, 68.4]	0.000	1.7 [−2.8, 6.3]	0.352
Hispanic	2.0 [−0.7, 4.7]	0.109	−21.7 [−133.4, 90.0]	0.723	−23.5 [−135.2, 88.2]	0.725	−0.6 [−4.0, 2.9]	0.666
NH or Other Pacific	4.8 [−2.6, 12.1]	0.148	64.6 [51.4, 77.8]	0.001	71.0 [57.8, 84.2]	0.000	2.2 [−0.5, 4.9]	0.086
White (Non-Hispanic)	3.0 [−0.2, 6.3]	0.058	22.8 [21.6, 24.0]	0.000	24.0 [22.8, 25.2]	0.000	0.5 [−2.4, 3.4]	0.678
Ages 25 to 44 years
HF Health Equity Goal	0.3 [0.2, 0.4]	0.003
AI/AN	1.6 [−0.7, 3.8]	0.126	8.5 [5.5, 11.5]	0.005	11.4 [8.4, 14.4]	0.002	1.2 [−0.9, 3.4]	0.182
Asian	0.1 [−0.1, 0.3]	0.162	−1.1 [−2.3, 0.1]	0.157	−2.2 [−3.4, −1.0]	0.008	−0.2 [−0.3, −0.0]	0.020
Black (Non-Hispanic)	0.8 [−0.2, 1.7]	0.080	9.8 [8.5, 11.1]	0.000	11.3 [10.0, 12.6]	0.000	0.5 [−0.3, 1.3]	0.178
Hispanic	0.3 [−0.0, 0.7]	0.070	0.0 [−1.2, 1.2]	1.000	−0.1 [−1.3, 1.1]	0.820	0.0 [−0.2, 0.3]	0.905
NH or Other Pacific	2.8 [−3.5, 9.2]	0.250			26.0 [21.2, 30.8]	0.003	2.5 [−3.8, 8.9]	0.293
White (Non-Hispanic)	0.3 [−0.0, 0.6]	0.056	0.9 [−0.3, 2.1]	0.213	0.6 [−0.6, 1.8]	0.219	−0.0 [−0.2, 0.2]	0.943
Ages 45 to 64 years
HF Health Equity Goal	1.0 [−0.2, 2.2]	0.087
AI/AN	7.6 [0.6, 14.6]	0.039	40.6 [32.9, 48.3]	0.001	67.2 [59.5, 74.9]	0.000	6.6 [0.2, 13.0]	0.046
Asian	0.4 [−0.2, 1.0]	0.111	−10.1 [−14.7, −5.5]	0.013	−12.2 [−16.8, −7.6]	0.031	−0.6 [−1.4, 0.2]	0.123
Black (Non-Hispanic)	4.0 [−2.1, 10.1]	0.146	59.9 [55.1, 64.7]	0.000	72.5 [67.7, 77.3]	0.000	3.0 [−2.4, 8.4]	0.199
Hispanic	1.3 [−1.0, 3.6]	0.201	2.1 [−2.5, 6.7]	0.419	4.1 [−0.5, 8.7]	0.334	0.3 [−1.6, 2.2]	0.708
NH or Other Pacific	7.7 [−3.0, 18.3]	0.117	69.0 [52.3, 85.7]	0.001	99.5 [82.8, 116.2]	0.001	6.7 [−3.8, 17.1]	0.151
White (Non-Hispanic)	2.4 [0.5, 4.4]	0.026	12.2 [7.7, 16.7]	0.006	18.9 [14.4, 23.4]	0.007	1.4 [0.1, 2.8]	0.043
Ages 65 to 84 years
HF Health Equity Goal	6.9 [−0.2, 14.0]	0.053
AI/AN	−2.1 [−30.9, 26.8]	0.851	144.5 [114.6, 174.4]	0.001	88.4 [58.5, 118.3]	0.004	−9.0 [−31.5, 13.5]	0.328
Asian	4.7 [−2.5, 11.9]	0.146	−130.0 [−143.6, −116.4]	0.000	−142.8 [−156.4, −129.2]	0.000	−2.2 [−3.8, −0.7]	0.016
Black (Non-Hispanic)	14.4 [−9.5, 38.4]	0.170	191.2 [177.0, 205.4]	0.000	228.1 [213.9, 242.3]	0.000	7.5 [−10.3, 25.3]	0.308
Hispanic	4.8 [−10.3, 19.8]	0.430	−20.3 [−33.8, −6.8]	0.042	−29.9 [−43.4, −16.4]	0.037	−2.2 [−11.2, 6.8]	0.539
NH or Other Pacific	2.1 [−16.8, 20.9]	0.776	209.9 [142.9, 276.9]	0.004	159.3 [92.3, 226.3]	0.005	−4.9 [−19.6, 9.9]	0.413
White (Non-Hispanic)	11.7 [−1.9, 25.3]	0.076	103.1 [90.6, 115.6]	0.000	121.8 [109.3, 134.3]	0.000	4.7 [−3.1, 12.5]	0.168

^*The equity benchmark was defined annually as the state with the state as the 5^th lowest value by year. The gaps were calculated as the absolute differences from the equity benchmark in each year. Abreviations: AI/AN, American Indian or Alaska Native; NH/PI, Native Hawaiian or Other Pacific Islander

There was variation in inequities by age. The largest inequities were observed for younger AI/AN, Black, and NH/PI. For example, there was a 290% increase in age-adjusted mortality for Black people aged 25 to 44 years compared to the equity benchmark in 2023. For White populations, gaps to the benchmark widened for people aged 45 to 64 years, and were present and persistent for those aged 65 to 84 years, with a 36.3% increased mortality in 2023. Asian adults aged 65 to 84 years old were the only population to have a relative decrease in mortality compared to the equity benchmark.

Discussion

In this study of US HF mortality from 2018 to 2023, age-adjusted mortality was generally stable across racially and ethnically diverse populations stratified by age, as was an equity benchmark. There were, however, persistent inequities compared to an equity benchmark across all ages for AI/AN, Black, and NH/PI populations, with the largest relative increases among younger adults and the largest absolute differences among older adults. Inequities emerged later in life for White people. Asian and Hispanic populations had age-adjusted mortalities at or better than the equity benchmark.

A primary finding of this analysis is that the observed increase in overall HF mortality from 2012 to 2021 has stabilized, with mortality lower than peak levels across racial, ethnic, and age intersections.[2, 3, 5, 16] In taking an equity approach, this study builds on prior disparities-based analyses that have used a disparities lens, such as the HF Heart Failure Epidemiology and Outcomes Statistics, which found overall the highest mortality for non-Hispanic Black individuals.[3, 5, 16] There are compelling reasons to utilize an equity framework. The most cogent is the need to set a single, ambitious, achievable goal that can be tracked over time for policymakers and healthcare professionals to improve population health. An equity approach further removes differentiation between “advantaged and “disadvantaged” populations that often assume outcomes in a White population are optimal.[3, 5, 11] As seen in this study, older White adults have an increasing burden of HF mortality, a trend that would not have been identified if the White population were the reference. Furthermore, goals for improvement for AI/AN, Black, and NH/PI populations would be less ambitious. By framing outcomes relative to a concrete, achievable benchmark rather than those of any single population, the equity framework makes worsening inequities easier to detect and generates more actionable information for understanding the etiologies and the impact of interventions.

The choice of an equity benchmark merits discussion. For some conditions, national goals have been set, such as through the Healthy People 2030 program. HF is included as part of a goal to reduce admissions and as a component of the overall improvement in cardiovascular health, aiming to reduce deaths from heart disease and stroke. However, reducing HF-specific mortality is not stated. As such, the equity benchmark was selected as the state with the 10^th mortality percentile based on prior work and with a recognition that it is trackable and achievable. Selecting a comparator as the state with the top 10th percentile of mortality is a methodology that is recommended by the Agency for Healthcare Research and Qualityand creates an aspirational benchmark to target for improvement.[11, 14]

Our findings have implications for policymakers and providers. Overall, the largest relative gaps occur earlier in life (ages 25 to 45) while the largest absolute gaps occur in older adults. For younger Black and AI/AN populations, there is a ~3-fold higher mortality. Prioritizing an understanding of the context-specific risk factors for younger adults is a necessary next step towards improving these trends. The persistence of large inequities among AI/AN, Black, and NH/PI populations across the lifespan and the later development among older White adults suggests advancements in risk factor modification and therapies for HF are not being delivered equitably. While this retrospective analysis cannot explore why, differences in an individual’s or community’s circumstances likely impact their access to best care.[17] With an equity benchmark defined, there is an opportunity to investigate drivers of low and high mortality that can be used to inform the development of context-specific interventions or implement evidence-based interventions, similar to the benchmark-driven approach used in the Healthy People initiative in the US that has achieved success.[18] Specifically, the benchmark provides a concrete tool for assessing whether interventions at any level (e.g., national or state policy) are widening or closing gaps in care.

This study has limitations owing to the retrospective use of death certificate information, including misattribution of deaths. Second, it is possible that changes in coding practices evolved over the study period and could impact trends. Third, the study design does not permit inference about the determinants of mortality inequities. Last, age-adjusted rates were used for comparison, as is standard practice to understand relative mortality, but this approach loses information about the absolute burdens of disease.

In conclusion, HF mortality has stabilized but remains inequitable, with persistent inequities among AI/AN, Black, and NHPI populations and emerging inequities among older White adults. These findings highlight the potential for policymakers and clinicians to use equity benchmarks to track outcomes and asses whether interventions narrow these gaps.