Author + information
- Received October 20, 2017
- Revision received December 11, 2017
- Accepted December 14, 2017
- Published online April 30, 2018.
- Boback Ziaeian, MD, PhDa,b,
- Paul A. Heidenreich, MD, MSc,
- Haolin Xu, MSd,
- Adam D. DeVore, MD, MHSd,e,
- Roland A. Matsouaka, PhDd,f,
- Adrian F. Hernandez, MD, MHSd,
- Deepak L. Bhatt, MD, MPHg,
- Clyde W. Yancy, MDh and
- Gregg C. Fonarow, MDi,∗ ()
- aDivision of Cardiology, David Geffen School of Medicine at University of California, Los Angeles, Los Angeles, California
- bDivision of Cardiology, Veterans Affairs Greater Los Angeles Healthcare System, Los Angeles, California
- cDivision of Cardiology, Veterans Affairs Palo Alto Healthcare System, Palo Alto, California
- dDuke Clinical Research Institute, Durham, North Carolina
- eDivision of Cardiology, Duke University Medical Center, Durham, North Carolina
- fDepartment of Biostatistics and Bioinformatics, Duke University, Durham, North Carolina
- gHeart and Vascular Center, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts
- hDivision of Cardiology, Northwestern University, Chicago, Illinois
- iAhmanson-UCLA Cardiomyopathy Center, University of California, Los Angeles Medical Center, Los Angeles, California
- ↵∗Address for correspondence:
Dr. Gregg C. Fonarow, Ahmanson-UCLA Cardiomyopathy Center, Ronald Reagan UCLA Medical Center, 10833 LeConte Avenue, Room A2-237 CHS, Los Angeles, California 90095-1679.
Objectives The purpose of this study was to analyze cumulative Medicare expenditures at index admission and after discharge by race or ethnicity.
Background Heart failure with preserved ejection fraction (HFpEF) is a growing proportion of heart failure (HF) admissions. Research on health care expenditures for patients with HFpEF is limited.
Methods Records of patients discharged from the Get With The Guidelines-Heart Failure registry between 2006 and 2014 were linked to Medicare data. The primary outcome was unadjusted payments for acute care services. Comparisons between race/ethnic groups were made using generalized linear mixed models. Cost ratios were reported by race/ethnicity, and adjustments were made sequentially for patient characteristics, hospital factors, and regional socioeconomic status.
Results Median Medicare costs for index hospitalizations were $7,241 for the entire cohort, $7,049 for whites, $8,269 for blacks, $8,808 for Hispanics, $8,477 for Asians, and $8,963 for other races. Median costs at 30 days for readmitted patients were $9,803 and $17,456 for the entire cohort at 1-year. No significant differences were seen in index admission cost ratios by race/ethnicity. At 30 days among readmitted patients, costs were 9% higher (95% confidence interval [CI]: 1% to 17%; p = 0.020) for blacks in the fully adjusted model than whites. At 1 year, costs were 14% higher (95% CI: 9% to 18%; p < 0.001) for blacks, 7% higher (95% CI: 0% to 14%; p = 0.041) for Hispanics, and 24% higher (95% CI: 8% to 42%; p = 0.003) for patients of other races. No significant differences between white and Asian expenditures were noted.
Conclusions Minority patients with HFpEF have greater acute care service costs. Further research of improving care delivery is needed to reduce acute care use for vulnerable populations.
- diastolic heart failure
- health care costs
- health care disparities
- heart failure with preserved ejection fraction
- hospital readmissions
Direct medical costs for heart failure (HF) were estimated at $28.5 billion with an average growth rate of 1.1% per year in the United States for 2013 (1). An estimated 6.5 million American adults had HF between 2011 and 2014, based on self-reported data from the National Health and Nutrition Examination Survey (2). By 2030, the prevalence of HF was expected to increase to over 8 million people secondary to shifting age demographics (3). HF prevalence among Medicare beneficiaries was 13.5% in 2015 (4). Medicare Parts A and B combined spending averaged $28,963 per HF beneficiary and was the second most expensive chronic condition behind stroke in 2015 (5). Although improvements in the prevention and treatment of ischemic heart disease have lowered the age-standardized rates of HF with reduced ejection fraction (HFrEF), the proportion of patients with HF with preserved ejection fraction (HFpEF) continues to grow (6,7).
Relatively few studies have been peformed which have evaluated health care use and expenditures associated with HFpEF. Patients with HFpEF have an observed lower mortality rate but higher readmission rate than HFrEF patients (8). With respect to race/ethnicity, black patients with HFpEF have a higher risk of readmission at 30 days and 1 year when adjusting for patient characteristics, socioeconomic status (SES), and hospital factors (9). Furthermore, SES may also influence health care use. How the uses of health care for patients with HFpEF may differ based on race/ethnicity and SES status are not well described. This study reports differences in Medicare inpatient expenditures by race/ethnicity, using the Get With The Guidelines-Heart Failure (GWTG-HF) registry linked to the Centers for Medicare and Medicaid Services administrative data.
Patients discharged from the GWTG-HF registry between January 1, 2006, and December 1, 2014, were screened. All patients included in GWTG-HF registry were identified by medical providers based on clinically diagnosed HF. Inclusion in the final cohort required age ≥65, eligibility for Medicare Parts A and B fee-for-service benefit during the discharge month, and left ventricular ejection fraction (LVEF) ≥50% on quantitative assessment; or if quantitative LVEF was not available, qualitative assessment of normal or mild dysfunction was included. LVEF criteria were consistent with society guideline definitions (10,11). Patients were excluded if disposition indicated transfer to a hospice facility, or they left against medical advice, or disposition was unknown. Patients were also excluded if regional SES variables were not available. The GWTG-HF registry was linked to Centers for Medicare and Medicaid Services administrative claims data providing use of services, expenditures, and outcomes at index admission and post-discharge.
Baseline patient and hospital characteristics were described by race/ethnic groups. Patient factors included age, sex, medical history, vital signs, body mass index, laboratory test values (i.e., blood urea nitrogen, serum creatinine, serum sodium, hemoglobin, hemoglobin A1c, and lipid panel). Hospital characteristics included region, rural location, teaching status, and size (number of beds). Percentages and median interquartile ranges were reported for categorical and continuous variables, respectively. The Pearson chi-squared test was used to compare categorical variables, and the Wilcoxon rank-sum test was used to compare ordinal categorical variables or continuous variables. Standardized differences were used to describe significant differences between groups. SES was linked using patients’ zip codes, geocoding to the 2015 Area Health Resource File provided through the Health Resources and Services Administration (12). SES variables included median household income, median home values, percentages of patients with high school diplomas, and percentage with ≥4 years of college (13). Patients’ SES variables were assigned values from census data of the year closest to the patients’ year of admission.
Primary outcomes included unadjusted Medicare (Part A) inpatient payments at index hospitalization, at 30 days, and at 1 year. Costs were standardized to 2014 dollars by using the Personal Health Care Hospital Care Index (14). Comparisons of health care use were modeled using generalized linear mixed models with a log link function and robust Poisson error distribution to allow for over-dispersion. Hospital-level random intercepts were used to account for clustering of patients within hospitals. Cost ratios were reported by race/ethnic classification controlling for relevant covariates. Secondary outcomes included the proportional change in variance for 1-year Medicare health care expenditures within the mixed models (15). Models were adjusted sequentially for patient and hospital characteristics, followed by patient SES based on zip code (Online Tables 1 to 3). Control variables were selected based on a review of published reports and prior established models used in GWTG-HF (8,13,16). Cumulative costs at 1 year were averaged over the number of patients at risk and accounting for the competing risk of death and differential length of time observed. Medical conditions coded as missing were imput to not present. Patient covariates with missingness were imputed for generalized linear mixed models (Online Table 4). Multiple imputations with 25 datasets were used to impute other patient covariates. Hospital characteristics were not imputed. Analyses were performed in SAS version version 9.4 software (SAS Institute, Inc., Cary, North Carolina).
The final cohort included 53,065 beneficiaries (Online Tables 5 to 8). The median age of hospitalization was 83 years for whites, 77 years for blacks, 79 years for Hispanics, and 81 years for Asians (Table 1). A higher proportion of women was observed among blacks than among other race/ethnic groups. Additional stratified analyses by sex were made available (Online Tables 9 to 11). Black and Hispanic patients had lower rates of atrial fibrillation than the other race/ethnic groups. Black and Hispanic patients had higher rates of hypertension, diabetes, and median body mass index. Among black, Hispanic, and Asian patients, chronic renal disease and dialysis were more common. Systolic blood pressure was highest among black patients, followed by Hispanic and Asian patients. LVEF rates were similar among the racial/ethnic groups. More black patients were admitted to teaching hospitals.
Median Medicare Part A costs for index hospitalization were $7,241 for the entire cohort, $7,049 for whites, $8,269 for blacks, $8,808 for Hispanics, $8,477 for Asians, and $8,963 for other race (Figure 1, Table 2). Minorities overall had higher hospitalization costs for the index admission. Median costs at 30 days for readmitted patients were $9,803 and $17,456 at 1 year for the entire cohort. Medicare costs among minority patients at index, at 30 days, and at 1 year were higher. Unadjusted cumulative average Medicare costs by race/ethnicity are shown in Figure 2.
After adjustments for patient characteristics, hospital factors, and regional SES, cost ratios for the index admission were not considerably different based on race/ethnicity across all models (Table 3). At 30 days, Medicare costs were greater for blacks with a 9% (95% confidence interval [CI]: 1% to 17%) higher relative cost than for whites after sequential adjustments across all models. At 1 year, Medicare costs were higher for black, Hispanic, and patients of other races than for whites, which was significant across all models. For blacks, unadjusted costs were 27% higher, and adjusted costs (for patient characteristics, hospital factors, and regional SES) were 14% higher than for whites. For Hispanics, unadjusted costs were 19% higher, and costs after sequential adjustment were 7% to 8% higher. For the other race category, unadjusted costs were 40% higher, and sequentially adjusted costs were 24% to 25% higher. Costs among Asians were not significantly different compared with those of whites across any of the models.
The proportional changes in variance (PCV) from the mixed models with sequential adjustments were used to explain hospital-level variations in Medicare expenditures. For the overall cohort, patient factors explained 14.7% of the variances among hospitals and hospital factors added 23.5%, and regional SES 30.4% to the incremental PCV (Table 4). When we evaluated these factors by race/ethnicity, we found that black race and hospital factors contributed less to the observed variations in Medicare expenditures with a larger incremental PCV (57.1%) explained by the addition of regional SES. For Hispanics, both hospital factors and regional SES explained a greater portion of Medicare expenditure variation than patient factors alone.
Rates of acute care service use was higher among minority patients (Table 5). Cumulative length of stay for patients readmitted at 1 year was higher for black, Hispanic, and other patients. Readmission rates at 30 days and at 1 year were also higher for minority patients, along with medical procedure rates.
This study describes the differential expenditures for HFpEF for Medicare Part A payments based on race/ethnicity. We found health care costs at 1 year after an index admission were higher among black, Hispanic, and other race patients when adjusting for patient characteristics, hospital factors, and regional SES. On average, at 1-year, Medicare Part A paid $9,846 more per black beneficiary and $8,265 more per Hispanic beneficiary. The higher use of health care for acute care services indicates an opportunity to reduce disparities in patient outcomes by investing in cost-effective, high-value interventions that would lower the incidence of HF. Manageable comorbidities such as poorly controlled hypertension, diabetes, chronic kidney disease, and elevated atherosclerotic risk are prevalent among race/ethnic minority populations, and improved treatment may help reduce acute care service use among Medicare beneficiaries (17,18).
Overall health care expenditures are concentrated among patients with chronic conditions. A study of Medicare beneficiaries in the top decile for costs noted a much higher prevalence of HF (44.4%) and a higher representation of black and Hispanic beneficiaries compared with patients in the lower deciles (19). The top decile patients in Medicare consume 73.0% of all the acute care service spending (19). Many of these hospitalizations and emergency department visits may be preventable with improvements in outpatient management. Vulnerable populations such as minorities with HFpEF are observed to be high users of acute care services (9,19). The differential expenditures for minority patients we observed at 30 days and 1 year are primarily driven by higher admission and readmission rates for minority patients (9). Interventions and systems of care tailored to vulnerable populations are needed to decrease the observed disparities in outcomes and use of acute care services.
Medicare is known to spend nearly as much on post-acute care services and readmissions within 30 days of discharge as at initial hospitalizations (20). Between 1994 and 2009, post-acute care service spending doubled for HF (21). In the Olmsted County cohort study, patients with HFpEF were observed to have 23.6% higher lifetime medical costs than HFrEF patients when controlling for other comorbidities and patient factors (22). We describe similarly high cost burdens for patients with HFpEF at the year after an index admission, with marked variation with respect to race/ethnicity.
Using sequential adjustments, we have attempted to demonstrate what factors may be driving differences in Medicare expenditures based on race/ethnic classification. We found that among minority groups, regional SES explained greater differences in Medicare expenditures at 1 year than it did among whites. Differences in expenditures adjusted for patient factors alone did not explain much of the variation in Medicare expenditures. This suggests that the disparities we observed based on race/ethnic categorization may largely be driven at the hospital or regional level. Communities in low-SES regions may lack the resources for high-quality acute care services and may be doubly disadvantaged with poor quality outpatient care, thus increasing the risk of using acute care services repeatedly. Variability in the effectiveness of clinical care likely contributes to differences in acute care use. Patients with HF do better when aware of their disease process and self-management strategies (23,24). Perhaps a less consistent and effective effort is made when engaging race/ethnic minority patients to ensure understanding of the care plan upon discharge or outpatient follow-up. Understanding how regional SES might be influencing strains on care networks and increased use of costly acute care services requires further investigation.
Adequate control of hypertension and volume status are critical to clinical management of HFpEF. Additional therapies have not shown effectiveness in altering the natural disease course or improving survival for patients with HFpEF (10). Despite limited treatment strategies, a combination of known behavioral and conventional cardiovascular risk factors predispose to the risk of developing and exacerbating symptoms of HFpEF (25). Recent research highlights the potential to reduce cardiovascular health care use with improvements in lifestyle related cardiovascular risk factors (26). For race/ethnic minorities and black patients especially, targeting known risk factors such as hypertension, diabetes, and obesity may be critical to both preventing and managing HFpEF (27,28).
This analysis used Medicare Part A administrative data to calculate costs. Costs for Medicare Part B used data related to outpatient visits, inpatient physician payments, and testing, and ambulance services were not included. Medicare Part D was also not included to capture costs related to outpatient prescription drug services. If patients switched to Medicare Advantage plans during the period observation, these costs were not captured for the year following index admission. Overall, these limitations underestimate the true total Medicare expenditures on patients with HFpEF. Prior research using Medical Expenditure Panel Survey data has shown that approximately 40% of the total direct health care expenditures on patients with HF relates to outpatient care, prescription medications, and other services (29). Medicare Part A payments were left as unadjusted costs. The Medicare payment calculator includes adjustments based on geographic factors related to regional differences in facility/provider input costs. Although the generalized linear models used for cost ratios included adjustments for regional SES, these do not mirror adjustments made by the Medicare payment formula. If minority patients are predominately seen in urban hospitals with higher average Medicare payments, there are expected to be residual differences in spending related to Medicare’s geographic payment adjustments. The SES factors used decreased the measured disparity in Medicare expenditures. Additional unobserved SES factors not included in the statistical models may further explain differences based on race/ethnicity. There are also potential selection biases because the GWTG-HF registry is based on voluntary hospital participation. However, prior studies have suggested that Medicare beneficiaries enrolled in this registry are representative of the U.S. Medicare population (30).
Medicare costs for race/ethnic minority patients with HFpEF are greater at index admission and at 1 year after an index admission, suggesting these patients are more vulnerable to use of acute care services. Interventions to prevent readmissions and lower the incidence of HFpEF among all beneficiaries, but especially minority patients, are needed to reduce health care spending on acute care services. A significant portion of the disparity in use relates to differences at the hospital level and regional SES. Further research that investigates the differences in hospital performance in lower SES regions and outpatient care networks are needed to understand differences in the use rate of acute care services.
COMPETENCY IN MEDICAL KNOWLEDGE: Vulnerable minority patients with HFpEF are at greater risk for repeat admission after an index hospitalization. Hospital and regional factors increase the risk of acute care service use. Ensuring safe transitions of care out of the hospital and quality outpatient management are important for minimizing the risk of acute hospitalizations and improving patient outcomes.
TRANSLATIONAL OUTLOOK: Further qualitative work evaluating the delivery of hospital and community care is needed to develop interventions that reduce the hospitalization burden for patients with HFpEF, particularly for minority populations.
Dr. Ziaeian was supported by the American College of Cardiology Presidential Career Developmental Award and American Heart Association 2015 Young Investigator Database Research Seed Grant. The Get With The Guidelines-Heart Failure program, provided by the American Heart Association, is sponsored in part by Amgen Cardiovascular and previously by Medtronic, GlaxoSmithKline, Ortho-McNeil, and the American Heart Association Pharmaceutical Roundtable. Dr. DeVore has received research support from the American Heart Association, Amgen, and Novartis; and has consulted for Novartis. Dr. Hernandez has received research support from AstraZeneca, Bayer, Luitpold, GlaxoSmithKline, Merck, Novartis, Portola Pharmaceuticals, and Verily; and has received honoraria from Amgen, Bayer, Boehringer-Ingelheim, Boston Scientific, Myokardia, Novartis, and Sanofi. Dr. Bhatt is a member of the advisory boards of Cardax, Elsevier Practice Update Cardiology, Medscape Cardiology, and Regado Biosciences; sits on the Board of Directors of Boston VA Research Institute and Society of Cardiovascular Patient Care; is Chair of the American Heart Association Quality Oversight Committee; a member of the Data Monitoring Committees of Cleveland Clinic, Duke Clinical Research Institute, Harvard Clinical Research Institute, Mayo Clinic, Mount Sinai School of Medicine, and Population Health Research Institute; has received honoraria from American College of Cardiology (Senior Associate Editor, Clinical Trials and News, ACC.org), Belvoir Publications (Editor in Chief, Harvard Heart Letter), Duke Clinical Research Institute (clinical trial steering committee), Harvard Clinical Research Institute (clinical trial steering committee), HMP Communications (Editor in Chief, Journal of Invasive Cardiology), Journal of the American College of Cardiology (Guest Editor; Associate Editor), Population Health Research Institute (clinical trial steering committee), Slack Publications (Chief Medical Editor, Cardiology Today’s Intervention), Society of Cardiovascular Patient Care (Secretary/Treasurer), WebMD (CME steering committees), Clinical Cardiology (Deputy Editor), NCDR-ACTION Registry Steering Committee (Chair), and VA CART Research and Publications Committee (Chair); research funding from Amarin, Amgen, AstraZeneca, Bristol-Myers Squibb, Chiesi, Eisai, Ethicon, Forest Laboratories, Ironwood, Ischemix, Eli Lilly, Medtronic, Pfizer, Roche, Sanofi, and The Medicines Company; royalties from Elsevier (Editor, Cardiovascular Intervention: A Companion to Braunwald’s Heart Disease); is site co-investigator for Biotronik, Boston Scientific, and St. Jude Medical (now Abbott); a trustee of American College of Cardiology; and has performed unfunded research for FlowCo, Merck, PLx Pharma, and Takeda. Dr. Fonarow has received research support from U.S. National Institutes of Health; and has consulted for Amgen, Janssen, Medtronic, Novartis, and St. Jude Medical. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- Get With The Guidelines-Heart Failure
- heart failure
- heart failure with preserved ejection fraction
- heart failure with reduced ejection fraction
- socioeconomic status
- Received October 20, 2017.
- Revision received December 11, 2017.
- Accepted December 14, 2017.
- 2018 American College of Cardiology Foundation
- Benjamin E.J.,
- Blaha M.J.,
- Chiuve S.E.,
- et al.
- Heidenreich P.A.,
- Albert N.M.,
- Allen L.A.,
- et al.
- ↵Centers for Medicare and Medicaid Services. Medicare Chronic Conditions Dashboard: Region Level. Comparison of Geographic Areas by Chronic Conditions, 2015. Available at: https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Dashboard/chronic-conditions-region/cc_region_dashboard.html. Accessed July 8, 2017.
- ↵Centers for Medicare and Medicaid Services. Use/Spending State Level: All Beneficiaries, 2007-2015. Chronic Cond. 2017:1. Available at: https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Chronic-Conditions/Downloads/CC_Util_Spend_State.zip. Accessed July 8, 2017.
- Gerber Y.,
- Weston S.A.,
- Redfield M.M.,
- et al.
- Steinberg B.A.,
- Zhao X.,
- Heidenreich P a,
- et al.
- Ziaeian B.,
- Heidenreich P.A.,
- Xu H.,
- et al.
- Yancy C.W.,
- Jessup M.,
- Bozkurt B.,
- et al.
- Ponikowski P.,
- Voors A.A.,
- Anker S.D.,
- et al.
- ↵U.S. Department of Health and Human Services. Area Health Resource File. Available at: https://datawarehouse.hrsa.gov/topics/ahrf.aspx. Accessed December 28, 2017.
- Eapen Z.J.,
- McCoy L.A.,
- Fonarow G.C.,
- et al.
- ↵Agency for Healthcare Research and Quality. Medical Expenditure Panel Survey: Using Appropriate Price Indices for Analyses of Health Care Expenditures or Income Across Multiple Years. Washington DC: U.S. Department of Health and Human Services. Available at: https://meps.ahrq.gov/about_meps/Price_Index.shtml. Accessed December 14, 2017.
- Merlo J.,
- Yang M.,
- Chaix B.,
- Lynch J.,
- Råstam L.
- Vivo R.P.,
- Krim S.R.,
- Liang L.,
- et al.
- Carnethon M.R.,
- Pu J.,
- Howard G.,
- et al.
- Chandra A.,
- Dalton M.A.,
- Holmes J.
- Dunlay S.M.,
- Shah N.D.,
- Shi Q.,
- et al.
- Shah S.J.
- Aaron K.J.,
- Colantonio L.D.,
- Deng L.,
- et al.
- Mensah G.A.,
- Mokdad A.H.,
- Ford E.S.,
- Greenlund K.J.,
- Croft J.B.
- Sharma A.,
- Colvin-Adams M.,
- Yancy C.W.
- Echouffo-Tcheugui J.B.,
- Bishu K.G.,
- Fonarow G.C.,
- Egede L.E.
- Curtis L.H.,
- Greiner M.A.,
- Hammill B.G.,
- et al.