Advertisement

Estimating the Economic Burden of Acute Myocardial Infarction in the US: 12 Year National Data

Published:February 21, 2020DOI:https://doi.org/10.1016/j.amjms.2020.02.004

      ABSTRACT

      Background

      Acute myocardial infarction (AMI) carries a substantial mortality and morbidity burden. The purpose of this study is to provide annual mean cost per patient and national level estimates of direct and indirect costs (lost productivity from morbidity and premature mortality) associated with AMI.

      Methods

      Nationally representative data spanning 12 years (2003-2014) with a sample of 324,869 patients with AMI from the Medical Expenditure Panel Survey (MEPS) were analyzed. A novel 2-part model was used to examine the excess direct cost associated with AMI, controlling for covariates. To estimate lost productivity from morbidity, an adjusted Generalized Linear Model was used for the differential in wage earnings between participants with and without AMI. Lost productivity from premature mortality was estimated based on published data.

      Results

      The total annual cost of AMI in 2016 dollars was estimated to be $84.9 billion, including $29.8 billion in excess direct medical expenditures, $14.6 billion in lost productivity from morbidity and $40.5 billion in lost productivity from premature mortality between 2003 and 2014. In the adjusted regression, the overall excess direct medical expenditure of AMI was $7,076 (95% confidence interval [CI] $6,028-$8,125) higher than those without AMI. After adjustment, annual wages for patients with AMI were $10,166 (95% CI −$12,985 to −$7,347) lower and annual missed work days were 5.9 days (95% CI 3.57-8.27) higher than those without AMI.

      Conclusions

      The study finds that the economic burden of AMI is substantial, for which effective prevention could result in significant health and productivity cost savings.

      Key Indexing Terms

      INTRODUCTION

      Acute myocardial infarction (AMI) is a common diagnosis in the United States where it is also associated with a high mortality.
      • Boateng S
      • Sanborn T
      Acute myocardial infarction.
      ,
      • Kochanek KD
      • Murphy SL
      • Xu JQ
      • et al.
      Deaths: final data for 2014.
      In 2014, annual AMI-related deaths in the United States numbered 114,019.
      • Kochanek KD
      • Murphy SL
      • Xu JQ
      • et al.
      Deaths: final data for 2014.
      Treatment of AMI is associated with important healthcare resource utilization ranging from hospital or intensive care admission unit, to invasive procedures, pharmacological treatment, and cardiac rehabilitation. The care of patients with AMI usually necessitates a multidisciplinary approach with significant human capital utilization. Consequently, it is purported that AMI will impose a high financial burden both at the individual and societal levels. The financial repercussions of AMI may also be related to lost work productivity and death.
      The economic cost of AMI has been studied with acute coronary syndrome (ACS) in the United States, but not independently.
      • Ghushchyan V
      • Nair KV
      • Page II, RL
      Indirect and direct costs of acute coronary syndromes with comorbid atrial fibrillation, heart failure, or both.
      ,
      • Kolansky DM
      Acute coronary syndromes: morbidity, mortality and pharmacoeconomic burden.
      Direct costs of ACS and related comorbidities including atrial fibrillation (AF) and heart failure (HF) among adults (≥18 years) were analyzed by Ghushchyan et al using data from the Medical Expenditure Panel Survey 1998-2009.
      • Ghushchyan V
      • Nair KV
      • Page II, RL
      Indirect and direct costs of acute coronary syndromes with comorbid atrial fibrillation, heart failure, or both.
      The authors concluded that the adjusted mean direct medical expenditure for ACS only was $7,482 (in 2011 dollar) higher compared with no ACS, AF and HF. Death following AMI is high in United States; however, prior studies have not investigated the indirect cost of premature mortality.
      Unstable angina and AMI are in the same continuum and fall under the umbrella of ACS. Patients with unstable angina (unstable plaque) eventually progress and develop AMI (rupture of unstable plaque and thrombosis of a coronary artery). While it is conceivable from a therapeutic standpoint to lump AMI and ACS together, the economic impact of AMI may be different from that of ACS as a whole, as the former represents a more severe entity with potentially higher economic costs.
      The United States health system faces incredible challenges, with rising healthcare costs ($2.59 trillion in 2010 and $3.33 trillion in 2016, increased by 28.5%) threatening the viability of ongoing health programs.
      • Hartman M
      • Martin AB
      • Espinosa N
      • et al.
      National Health Care Spending in 2016: spending and enrollment growth slow after initial coverage expansions.
      Despite the increase in healthcare expenditures, much remains unknown about how much is spent on each disease conditions.
      • Dieleman JL
      • Baral R
      • Birger M
      • et al.
      US spending on personal health care and public health, 1996-2013.
      In order to identify areas in which value and cost are misaligned, a better grasp of the current cost burden of specific disease states is needed. Disentangling myocardial infarction cost from ACS as whole is in line with this imperative. Prior studies have assessed the financial burden of AMI. These studies however either focused on inpatient costs
      • Afana M
      • Brinjikji W
      • Cloft H
      • et al.
      Hospitalization costs for acute myocardial infarction patients treated with percutaneous coronary intervention in the United States are substantially higher than Medicare payments.
      ,
      • Kauf TL
      • Velazquez EJ
      • Crosslin DR
      • et al.
      The cost of acute myocardial infarction in the new millennium: evidence from a multinational registry.
      or on the impact of comorbid conditions such as diabetes mellitus.
      • Zhou X
      • Shrestha SS
      • Luman E
      • et al.
      Medical expenditures associated with diabetes in myocardial infarction and ischemic stroke patients.
      In order to better inform stakeholders and efficiently influence economic decisions pertaining to AMI, economic analysis would ideally have a national reach and include all age categories. There is a paucity of comprehensive analyses of direct and indirect costs of AMI, at per patient and national levels in all age groups in the United States. In the current study, we estimate the direct medical cost and the indirect cost (lost productivity from morbidity and premature mortality) of AMI using rigorous cost estimation methods and 12-year data (2003-2014) from the largest nationally representative survey of medical costs and mortality statistics in the United States.

      METHODS

      Data Source and Sampling

      Nationally representative data were obtained through the publicly available Medical Expenditure Panel Survey (MEPS). MEPS is an ongoing national household survey for the civilian noninstitutionalized US population. Using computer-assisted personal interviews, MEPS collects detailed data on healthcare use, expenditures, source of payments, health insurance coverage, health status, demographic, employment, access to care and satisfaction.

      Agency for Healthcare Research and Quality (AHRQa). Medical expenditure panel survey. 2014 Full year consolidated data file 2016a, Available at:https://meps.ahrq.gov/data_stats/download_data/pufs/h171/h171doc.pdf. Accessed August 1, 2019.

      ,

      Agency for Healthcare Research and Quality (AHRQb). Medical expenditure panel survey, 2014 medical conditions 2016b, Available at:https://meps.ahrq.gov/data_stats/download_data/pufs/h170/h170doc.pdf. Accessed August 1, 2019.

      In total, we included a sample of 324,869 (weighted 253,235,052) all age observations with and without AMI in the analysis.
      MEPS consists of 3 components (1) the household component (HC), a nationally representative survey on demographic, health conditions, charges and payments drawn from a subsample of households that participated in prior year's National Health Interview Survey; (2) the Medical Provider Component (MPC), which collects data from medical care providers and facilities reported as providing care to persons interviewed in the HC; and (3) the Insurance Component (IC), which collects data on the types and costs of workplace health insurance.

      National Center for Health Statistics. Health, United States, 2009: with special feature on medical technology. Hyattsville, MS. 2010, Available at:https://www.cdc.gov/nchs/data/hus/hus09.pdf. Accessed August 1, 2019.

      MEPS-HC contains self-reported information, while the MPC contains information on visit dates, ICD-9-CM (International Classification of Disease, Ninth Revision, Clinical Modification) diagnosis codes, procedure codes, charges and payments.

      National Center for Health Statistics. Health, United States, 2009: with special feature on medical technology. Hyattsville, MS. 2010, Available at:https://www.cdc.gov/nchs/data/hus/hus09.pdf. Accessed August 1, 2019.

      MPC is used to validate and supplement information received from the MEPS-HC. We merged MEPS-HC 12-year data from the medical condition files and the full year consolidated files at person level. The pooled data increases sample size and provides robust cost estimations. MEPS do not contain a personal identifier and is a publicly available database, thus, this study qualified for Institutional Review Board waiver.

      Measures

      Dependent Variables

      Direct medical cost

      The dependent variables in this study are total direct medical expenditure, inpatient hospital expenditure, outpatient expenditure, emergency room (ER) expenditure, prescription medicine expenditure, home health care expenditure, dental expenditure, and other medical expenses. Direct medical expenditures in MEPS are defined as a sum of the direct payments including out-of-pocket payments and payments by private insurance, Medicaid, Medicare and other sources during the year.

      Agency for Healthcare Research and Quality (AHRQa). Medical expenditure panel survey. 2014 Full year consolidated data file 2016a, Available at:https://meps.ahrq.gov/data_stats/download_data/pufs/h171/h171doc.pdf. Accessed August 1, 2019.

      Indirect costs

      These are determined using measures of workplace productivity (employment, wage and missed workdays).
      • Ghushchyan V
      • Nair KV
      • Page II, RL
      Indirect and direct costs of acute coronary syndromes with comorbid atrial fibrillation, heart failure, or both.
      ,
      • Libby AM
      • Ghushchyan V
      • McQueen RB
      • et al.
      Economic differences in direct and indirect costs between people with epilepsy and without epilepsy.
      Our analyses of workplace productivity were limited to nonretired working age individuals between 18 and 64 years.
      Employment was defined as “full year employed” if the person had full-time or part-time paid work with no breaks in employment during the interview year regardless of hours worked. The wage was defined as annual wage in 2016 dollars for persons never retired between 18 and 64 years of age.
      • Ghushchyan V
      • Nair KV
      • Page II, RL
      Indirect and direct costs of acute coronary syndromes with comorbid atrial fibrillation, heart failure, or both.
      ,
      • Libby AM
      • Ghushchyan V
      • McQueen RB
      • et al.
      Economic differences in direct and indirect costs between people with epilepsy and without epilepsy.
      In MEPS data, the wage for those who were not employed was recorded as zero. Missed workdays are the number of days the person missed a half-day or more from work due to illness, injury or mental or emotional problems during the study period

      Agency for Healthcare Research and Quality (AHRQa). Medical expenditure panel survey. 2014 Full year consolidated data file 2016a, Available at:https://meps.ahrq.gov/data_stats/download_data/pufs/h171/h171doc.pdf. Accessed August 1, 2019.

      Because half-day and full-days lost were not distinguished in MEPS, all days lost were recorded as full-days, which is consistent with previous studies using MEPS data.
      • Cawley J
      • Rizzo J
      • Haas K
      The association of diabetes with job absenteeism costs among obese and morbidly obese workers.
      ,
      • Bishu KG
      • Gebregziabher M
      • Dismuke CE
      • et al.
      Quantifying the incremental and aggregate cost of missed work days in adults with diabetes.
      Annual lost productivity from morbidity was calculated as lost wages due to AMI. Lost wages were used as a monetary measure of lost productivity from AMI related morbidity due to unemployment and absenteeism.
      • Ghushchyan V
      • Nair KV
      • Page II, RL
      Indirect and direct costs of acute coronary syndromes with comorbid atrial fibrillation, heart failure, or both.
      ,
      • Libby AM
      • Ghushchyan V
      • McQueen RB
      • et al.
      Economic differences in direct and indirect costs between people with epilepsy and without epilepsy.
      Lost productivity from AMI related morbidity represents the values of foregone earnings; therefore, it was calculated as the adjusted wage difference between individuals with and without AMI.
      • Ghushchyan V
      • Nair KV
      • Page II, RL
      Indirect and direct costs of acute coronary syndromes with comorbid atrial fibrillation, heart failure, or both.
      ,
      • Libby AM
      • Ghushchyan V
      • McQueen RB
      • et al.
      Economic differences in direct and indirect costs between people with epilepsy and without epilepsy.

      Independent Variables

      The primary independent variable was the presence of AMI, coded as binary variable. AMI was identified by ICD-9-CM code 410 (acute myocardial infarct), after collapsing of fully specified codes into 3 digits.

      Agency for Healthcare Research and Quality (AHRQb). Medical expenditure panel survey, 2014 medical conditions 2016b, Available at:https://meps.ahrq.gov/data_stats/download_data/pufs/h170/h170doc.pdf. Accessed August 1, 2019.

      We used MEPS pooled panel cross-sectional sample with diagnosis and costs measured within the same year for each individual. Age (<65 years vs ≥65 years), sex, race/ethnicity (Non-Hispanic White, Non-Hispanic Black, Hispanic and Other race), education (<High school, High School graduate, College or more), insurance status, census region (Northeast, Midwest, South and West), income, marital status, Charlson Comorbidities index (CCI) and year category were included as covariates in estimating the adjusted expenditure attributable to AMI. Insurance status was divided into 3 categories any private, public only and uninsured. Private insurance coverage in MEPS is defined as having a major medical plan covering hospital and physician services. “Any private” was assigned if a person had any private insurance coverage including TRICARE/Civilian Health and Medical Program for Department of Veterans Affairs during the year. “Public only” was assigned if a person had only public including Medicare, Medicaid, other public hospital or physician coverage during the year, and uninsured was assigned if a person was uninsured during the year.

      Agency for Healthcare Research and Quality (AHRQa). Medical expenditure panel survey. 2014 Full year consolidated data file 2016a, Available at:https://meps.ahrq.gov/data_stats/download_data/pufs/h171/h171doc.pdf. Accessed August 1, 2019.

      Marital status was coded into 3 groups married, nonmarried (Widowed/Divorced/separated), and never married. Income level was defined as a percentage of the federal poverty level and grouped into 4 categories poor (<125%), low income (125% to less than 200%), middle income (200% to less than 400%) and high income (≥400%). CCI was adapted from D'Hoore,
      • D'Hoore W
      • Bouckaert A
      • Tilquin C
      Practical considerations on the use of the Charlson comorbidity index with administrative data bases.
      where the ICD-9 codes were compatible with the 3-digit collapsed ICD-9 codes included in MEPS data. The CCI is an aggregate of 17-weighted conditions including myocardial infarction, congestive HF, peripheral vascular disease, dementia, cerebrovascular disease, chronic pulmonary disease, connective tissue disease, ulcer disease, mild liver disease, hemiplegia (weight 1), moderate or severe renal disease, diabetes, any tumor, leukemia, lymphoma (weight 2), moderate or severe liver disease (weight 3) and metastatic solid tumor (weight 6). For the purpose of this analysis, we have excluded AMI while computing the CCI because the condition is the primary independent variable. Each condition is identified using ICD-9-CM codes keeping in mind that MEPS is a cross-sectional survey. The CCI represents the best available indicator of medical comorbidity in MEPS that affect cost.
      • D'Hoore W
      • Bouckaert A
      • Tilquin C
      Practical considerations on the use of the Charlson comorbidity index with administrative data bases.
      The CCI was grouped into 3 categories 0, 1 and ≥2. Year category included the following: 2003/06, 2007/10 and 2011/14.

      Mortality Data

      Deaths in the report are based on information from all resident death certificates filed in the 50 states and the District of Columbia. Cause of death statistics presented in this report are identified using the International Classification of Diseases, Tenth Revision (ICD-10) and more than 99% of deaths occurring in United States are believed to be registered.
      • Kochanek KD
      • Murphy SL
      • Xu JQ
      • et al.
      Deaths: final data for 2014.
      The National Vital Statistical Report (NVSR) presents detailed data on mortality patterns among residents of the United States by variables such as age, sex, Hispanic origin, race and cause of death. We calculated the 12-year deaths (2003-2014) attributed to AMI, obtained from the NVSR by age category.

      Statistical Analyses

      Chi square (χ²) tests were used to compare demographic characteristics between individuals with AMI and those without AMI. A 2-part econometric model was used to estimate the adjusted excess medical expenditure for patients with AMI (the comparison group was those without AMI).
      • Manning WG
      • Mullahy J
      Estimating log models: to transform or not to transform?.
      The 2-part model accounts for situations with excess zero expenditure. The 2-part model accounts for situations with excess zero expenditure and rightward skewed MEPS cost data, in accordance with prior work.
      • Belotti F
      • Deb P
      • Manning WG
      • et al.
      Twopm: two-part models.
      In our 12-year pooled panel cross-sectional data, the proportion of zero costs for the general population was as follows: Total expenditure 9.1%, Inpatient expenditure 91.4%, Outpatient expenditure 21.0%, Medication expenditure 28.5%, ER expenditure 83.8% and Home health expenditure 97.3%. First we used a probit model to estimate the probability of observing a zero versus positive medical expenditure, and then for those with positive medical expenditures, a generalized linear model (GLM) with gamma distribution and log link was used to estimate the medical expenditures associated with AMI.
      • Manning WG
      • Mullahy J
      Estimating log models: to transform or not to transform?.
      ,
      • Belotti F
      • Deb P
      • Manning WG
      • et al.
      Twopm: two-part models.
      GLM was used to address the positive skewness of the expenditures and disenfranchise ourselves from the normality and homoscedasticity assumptions therefore avoiding bias when retransforming the log values into the dollar.
      • Belotti F
      • Deb P
      • Manning WG
      • et al.
      Twopm: two-part models.
      The “margins, dydx(*)” post estimation command in Stata provides the marginal (or incremental) estimates of cost for all independent variables (AMI, age, sex, race/ethnicity, marital status, education, insurance, census region, income, CCI and year category) from the first (probit) and second (GLM gamma distribution and log link) part of the 2-part model.
      • Belotti F
      • Deb P
      • Manning WG
      • et al.
      Twopm: two-part models.
      The Park test was used to examine the model fit and verified the use of a gamma distribution with a log link as the best–fitting GLM for consistent estimation of coefficients.
      Missed workdays in a given year are count data, thus a negative binomial regression model was used for the adjusted model. A logistic regression model was used to estimate the probability of full year employment. After fitting the logistic regression models taking the survey design into account, the F-adjusted mean residual goodness-of-fit suggested no evidence of lack of fit.
      • Archer KJ
      • Lemeshow S
      Goodness-of-fit test for a logistic regression model fitted using survey sample data.
      For GLM with family gamma, log link was used for the adjusted analysis of annual wage.
      Our analysis combines 12-years of MEPS data. Thus, the medical expenditures and wages were inflated to the recent estimate of 2016 dollar value using the consumer's price index obtained from the Bureau of Labor Statistics.

      Bureau of Labor Statistics, CPI Inflation Calculator. Available at: https://www.bls.gov/data/inflation_calculator.htm. Accessed July 10, 2019.

      Regression analyses were performed to control for age, sex, race/ethnicity, marital status, education, health insurance, census region, income, CCI and time trend. Estimates that were statistically significant at the P < 0.05 level are discussed in the paper. All analyses were performed at the person-level. In order to generalize our findings to the US population, our study accounted for the complex sample design (sampling weight, variance stratum and primary sampling unit) in all analyses using Stata ver.14.
      StataCorp
      Stata: Release 14. Statistical Software.
      Of note, the weighting process included an adjustment for nonresponse or missing data.

      Agency for Healthcare Research and Quality (AHRQa). Medical expenditure panel survey. 2014 Full year consolidated data file 2016a, Available at:https://meps.ahrq.gov/data_stats/download_data/pufs/h171/h171doc.pdf. Accessed August 1, 2019.

      To estimate lost productivity from premature mortality associated with AMI, we used the net present value (PV) of future productivity adopted from Grosse et al.
      • Grosse SD
      • Krueger KV
      • Mvundura M
      Economic productivity by age and sex: 2007 estimates for the United States.
      The PV of one life based on the recommended 3% discount rate adopted from Grosse and colleagues was converted into 2016 dollar using gross domestic product deflator.

      U.S. Bureau of Economic Analysis, Gross Domestic Product: Implicit Price Deflator [GDPDEF], retrieved from FRED, Federal Reserve Bank of St. Louis, Available at: https://fred.stlouisfed.org/series/GDPDEF. Accessed June 5, 2019.

      The number of deaths from selected causes by age was obtained from the National Vital Statistics Report.
      • Boateng S
      • Sanborn T
      Acute myocardial infarction.
      ,
      • Hoyert DL
      • Heron MP
      • Murphy SL
      • et al.
      Deaths: final data for 2003.
      • Miniño AM
      • Heron MP
      • Murphy SL
      • et al.
      Deaths: final data for 2004.
      • Kung HC
      • Hoyert DL
      • Xu JQ
      • et al.
      Deaths: final data for 2005.
      • Heron MP
      • Hoyert DL
      • Murphy SL
      • et al.
      Deaths: final data for 2006.
      • Xu JQ
      • Kochanek KD
      • Murphy SL
      • et al.
      Deaths: final data for 2007.
      • Miniño AM
      • Murphy SL
      • Xu JQ
      • et al.
      Deaths: final data for 2008.
      • Kochanek KD
      • Xu J
      • Murphy SL
      • et al.
      Deaths: final data for 2009.
      • Murphy SL
      • Xu JQ
      • Kochanek KD
      Deaths: final data for 2010.
      • Kochanek KD
      • Murphy SL
      • Xu JQ
      Deaths: final data for 2011.
      • Murphy SL
      • Kochanek KD
      • Xu JQ
      • et al.
      Deaths: final data for 2012.
      • Xu JQ
      • Murphy SL
      • Kochanek KD
      • et al.
      Deaths: final data for 2013.

      RESULTS

      A total of 42,12,846 (1.7%) people in the United States verified the diagnosis of AMI per year. AMI was more frequent among individuals aged 65 and older compared to younger individuals (1.7% versus 0.77%, P-value < 0.001). Individuals with AMI were more likely to be men, Non-Hispanic Whites, married or nonmarried (widowed/divorced/separated), high school graduates, publicly insured, Midwest or South region residents, poor and have comorbid conditions (Table 1).
      TABLE 1Weighted population characteristics with and without acute myocardial infarction (AMI).
      VariablesAMI (%)No-AMI (%)P Value
      N (n)4,212,846 (4,769)249,022,206 (320,100)
      Age category
       Age <6540.885.5<0.001
       Age ≥6559.214.5
      Sex
       Male64.146.7<0.001
       Female35.953.5
      Race/ethnicity
       Non-Hispanic White80.568.4<0.001
       Non-Hispanic Black9.011.1
       Hispanic6.813.7
       Others3.76.8
      Marital status
       Married58.052.9<0.001
       Non-married36.020.3
       Never married6.026.8
      Education category
       <High school9.819.7<0.001
       High school graduate46.233.9
       College or more44.046.4
      Insurance
       Private57.069.6<0.001
       Public38.520.4
       Uninsured4.510.0
      Census region
       Northeast18.418.2<0.001
       Midwest25.422.6
       South39.036.3
       West17.222.9
      Income category
       Poor income21.517.3<0.001
       Low income17.713.4
       Middle income29.930.3
       High income30.939.0
      Charlson Comorbidity Index
       036.277.0<0.001
       111.88.3
       ≥252.014.7
      N, weighted sample size; n, unweighted sample size; %, weighted percentage.

      Annual Mean Unadjusted Direct and Indirect Costs by AMI Status

      Relative to individuals without AMI ($4,822 [$4,735-$4,908]), individuals with AMI had nearly 4 times higher overall mean unadjusted direct medical expenditure ($18,739 [$17,692-$19,787]) per year (Table 2). Patients with AMI had nearly 7 times higher mean inpatient expenditure ($8,346 [$7,599-$9,093]), more than double outpatient expenditure ($4,152 [$3,823-$4,482]) and nearly 4 times higher prescription expenditure ($3,976 [$3,737-$4,215]) compared to those without AMI. Additionally, patients with AMI had higher ER ($689 versus $192) and home health ($988 versus $168) expenditure relative to individuals without AMI. The annual hospital inpatient expenditure of AMI accounts for most (45%) of the overall mean direct expenditure. Compared to individuals without AMI ($31,566 [$31,045-$32,087]), individuals with AMI had 50% lower wages ($15,712 [$14,161-$17,264) in terms of indirect costs. Individuals with AMI had twice as many missed workdays per year (8.3 [6.81-9.78]) as those without this condition (4.0 [3.91-4.09]).
      TABLE 2Annual unadjusted mean of direct and indirect costs with and without acute myocardial infarction (AMI) (2016 $).
      CostsAMI

      Mean
      95% CINo-AMI

      Mean
      95% CIP Value
      Direct costs ($)
      Total costs18,73917,692-19,7874,8224,735-4,908<0.001
      Inpatient8,3467,599-9,0931,2591,212-1,305<0.001
      Outpatient4,1523,823-4,4821,6671,634-1,701<0.001
      Medications3,9763,737-4,2151,0941,064-1,124<0.001
      Emergency Room689601-777192187-198<0.001
      Home Health988813-1,164168142-194<0.001
      Others585514-656438429-448<0.001
      Indirect costs
      Wage ($)25,90423,614-28,19537,06036,489-37,631<0.001
      Missed workdays (n)10.88.70-12.964.14.07-4.27<0.001

      Adjusted Excess Direct Medical Expenditure of AMI

      The overall excess direct medical expenditure of AMI participants was $7,076 (95% confidence interval [CI] $6,028-$8,125) compared to their counterparts (Table 3). The excess inpatient expenditure of AMI was $3,636 (95% CI $3,048-$4,223) while the excess outpatient expenditure was $932 (95% CI $713-$1,151) and the excess medication expenditure was $1,072 (95% CI $935-$1,208). Similarly, the excess ER expenditure was $395 (95% CI $311-$478) when AMI diagnosis was present. Finally, we extrapolate the estimated overall direct medical expenditure of AMI for the total civilian noninstitutionalized US population. At the national level, the overall adjusted excess medical expenditure of AMI to the US population was $29.9 billion compared to populations without AMI in a year.
      TABLE 3Two-part regression model: adjusted excess direct healthcare cost by acute myocardial infarction (AMI) status.
      Regression
      For regressions I-VII, the primary outcome is total, inpatient, outpatient, medication, emergency room, home health and other expenditures, respectively. Each regression controls for age, sex, race/ethnicity, marital status, education, health insurance, census region, income, Charlson Comorbidity Index and time trend (2003/06, 2007/10 and 2011/14).
      VariablesExcess Cost95% CIP Value
      ITotal cost
      No-AMI (Ref.)
      AMI7,076
      Level of significance P < 0.001.
      6,028–8,125<0.001
      IIInpatient
      No-AMI (Ref.)
      AMI3,636
      Level of significance P < 0.001.
      3,048–4,223<0.001
      IIIOutpatient
      No-AMI (Ref.)
      AMI932
      Level of significance P < 0.001.
      713–1,151<0.001
      IVMedications
      No-AMI (Ref.)
      AMI1,072
      Level of significance P < 0.001.
      935–1,208<0.001
      VEmergency room
      No-AMI (Ref.)
      AMI395
      Level of significance P < 0.001.
      311–478<0.001
      VIHome health
      No-AMI (Ref.)
      AMI179
      Level of significance P < 0.001.
      107–252<0.001
      VIIOthers
      No-AMI (Ref.)
      AMI72
      Level of significance P < 0.01,
      14–1300.014
      a Level of significance P < 0.01,
      b Level of significance P < 0.001.
      c For regressions I-VII, the primary outcome is total, inpatient, outpatient, medication, emergency room, home health and other expenditures, respectively. Each regression controls for age, sex, race/ethnicity, marital status, education, health insurance, census region, income, Charlson Comorbidity Index and time trend (2003/06, 2007/10 and 2011/14).

      Adjusted Indirect Costs of Lost Productivity Attributable to AMI

      After adjusting for demographic variables, comorbidities and time trend, annual wages for patients with AMI were $10,166 (95% CI −$12,985 to −$7,347) lower and annual missed workdays were 5.9 days (95% CI 3.57-8.27) higher than those without AMI (Table 4). Likewise, patients with AMI had 46% (odds-ratio, 0.56; 95% CI 0.469-0.623) lower odds of employment compared to those without AMI. At the national level, the estimated indirect cost of lost productivity from morbidity due to AMI for nonretired population aged 18-64 was estimated at $14.6 billion per year.
      TABLE 4Adjusted productivity cost estimates associated with acute myocardial infarction (AMI) (aged 18-64 years).
      Regression
      For regressions I-III, the primary outcome is wage, missed work days and employed full time, respectively. Each regression controls for age, sex, race/ethnicity, marital status, education, health insurance, census region, income, Charlson Comorbidity Index and time trend (2003/06, 2007/10 and 2011/14). Regression I-reported marginal effect from GLM, family gamma and log link. Regression II- reported the marginal effects from negative binomial model. Regression III-reported odds-ratio from logistic regression model.
      VariablesCoefficients95% CIP Value
      IWage
      No-AMI (Ref.)
      AMI−10,166
      Level of significance P < 0.001.
      −12,985 to −7,347<0.001
      IIMissed workdays
      No-AMI (Ref.)
      AMI5.9
      Level of significance P < 0.001.
      3.57-8.27<0.001
      IIIEmployed full time
      No-AMI (Ref.)
      AMI0.54
      Level of significance P < 0.001.
      0.469-0.623<0.001
      a Level of significance P < 0.001.
      b For regressions I-III, the primary outcome is wage, missed work days and employed full time, respectively. Each regression controls for age, sex, race/ethnicity, marital status, education, health insurance, census region, income, Charlson Comorbidity Index and time trend (2003/06, 2007/10 and 2011/14). Regression I-reported marginal effect from GLM, family gamma and log link. Regression II- reported the marginal effects from negative binomial model. Regression III-reported odds-ratio from logistic regression model.
      The total number of deaths related to AMI was 16,02,926 in the 12-year period (2003-2014) and the average number of deaths per year was 133,577 (Table 5). Extrapolating this with a previously published estimate of the total PV of life, we found that the total value of lost productivity from premature mortality due to AMI over the 12-year period was $485.573 billion, with an average per year of $40.5 billion.
      TABLE 5Cost of mortality due to acute myocardial infarction (AMI) by age group.
      Age, yearsTotal deaths 2003-2014
      Derived from the National Vital Statistical Report (2003-2014).1,24-34
      Present value of one life (2016 $)
      Present value (PV) of one life adopted from Grosse and colleagues22 and converted into 2016 using GDP deflator.23
      Present value of total mortality 2003-2014 (in billion 2016$)
      The total PV of mortality for AMI from 2003 to 2014 was $485.573 billion, with average per year $40.5 billion.
      <1981,357,7150.133
      1-4281,357,7150.038
      5-14751,579,2200.118
      15-246961,845,5361.284
      25-344,7181,808,6398.533
      35-4427,9501,472,03241.143
      45-54109,6931,020,564111.948
      55-64217,717562,298122.421
      65-74294,079269,31779.200
      75-84446,694139,33862.241
      ≥85501,178116,75358.514
      Total1,602,926-485.573
      a Derived from the National Vital Statistical Report (2003-2014).
      • Boateng S
      • Sanborn T
      Acute myocardial infarction.
      ,
      • Hoyert DL
      • Heron MP
      • Murphy SL
      • et al.
      Deaths: final data for 2003.
      • Miniño AM
      • Heron MP
      • Murphy SL
      • et al.
      Deaths: final data for 2004.
      • Kung HC
      • Hoyert DL
      • Xu JQ
      • et al.
      Deaths: final data for 2005.
      • Heron MP
      • Hoyert DL
      • Murphy SL
      • et al.
      Deaths: final data for 2006.
      • Xu JQ
      • Kochanek KD
      • Murphy SL
      • et al.
      Deaths: final data for 2007.
      • Miniño AM
      • Murphy SL
      • Xu JQ
      • et al.
      Deaths: final data for 2008.
      • Kochanek KD
      • Xu J
      • Murphy SL
      • et al.
      Deaths: final data for 2009.
      • Murphy SL
      • Xu JQ
      • Kochanek KD
      Deaths: final data for 2010.
      • Kochanek KD
      • Murphy SL
      • Xu JQ
      Deaths: final data for 2011.
      • Murphy SL
      • Kochanek KD
      • Xu JQ
      • et al.
      Deaths: final data for 2012.
      • Xu JQ
      • Murphy SL
      • Kochanek KD
      • et al.
      Deaths: final data for 2013.
      b Present value (PV) of one life adopted from Grosse and colleagues
      • Grosse SD
      • Krueger KV
      • Mvundura M
      Economic productivity by age and sex: 2007 estimates for the United States.
      and converted into 2016 using GDP deflator.

      U.S. Bureau of Economic Analysis, Gross Domestic Product: Implicit Price Deflator [GDPDEF], retrieved from FRED, Federal Reserve Bank of St. Louis, Available at: https://fred.stlouisfed.org/series/GDPDEF. Accessed June 5, 2019.

      c The total PV of mortality for AMI from 2003 to 2014 was $485.573 billion, with average per year $40.5 billion.

      Total Aggregate Cost Attributable to AMI

      The total annual cost of AMI in 2016 dollars was estimated at $84.9 billion per year, including $29.8 billion in excess direct medical expenditures; $14.6 billion in lost productivity from morbidity and $40.5 billion in lost productivity from premature mortality (Figure 1). The proportion of the indirect cost of AMI due to lost productivity from morbidity and premature mortality comprised 64.9% of the total cost.
      FIGURE 1
      FIGURE 1Annual aggregate cost of acute myocardial infraction (in 2016 $).

      DISCUSSION

      In this nationally representative analysis of the cost of AMI, we have estimated the total annual cost of AMI in 2016 dollars to be $84.9 billion, of which, nearly half being incurred by lost productivityfrom premature mortality. The financial burden associated with conditions sharing similar risk factors and complicating AMI such as HF, and cerebrovascular disease were reported to be generally high, probably as a result of increasing sophistication of acute care (thrombectomy, thrombolysis and angioplasty) and long-term deleterious effects on vital organs such as life-threatening arrhythmia, and permanent neurological deficits.
      • Prabhakaran S
      • Ruff I
      • Bernstein RA
      Acute stroke intervention: a systematic review.
      ,
      • Heidenreich PA
      • Albert NM
      • Allen LA
      • et al.
      Forecasting the impact of heart failure in the United States: a policy statement from the American Heart Association.
      For example, an annual surplus of $7,482 in 2011 dollar was spent for each individuals with ACS, compared with those with no ACS (AF or HF).
      • Ghushchyan V
      • Nair KV
      • Page II, RL
      Indirect and direct costs of acute coronary syndromes with comorbid atrial fibrillation, heart failure, or both.
      This is comparable to our estimate.
      In this study, the difference in direct medical expenditure between AMI and non-AMI individuals was primarily driven by inpatient expenditure. Several factors could account for this observation. AMI patients tend to have a longer hospital stays, frequently require admissions to intensive care units and may need sophisticated and often expensive medical procedures. Acute complications are also common among those afflicted by AMI and may include mechanical complications such as ventricular free wall ruptures, ventricular septal ruptures, acute heart failure and life threatening arrhythmia, all potentially resulting in excess cost.
      • Anderson JL
      • Morrow DA
      Acute myocardial infarction.
      ,
      • Kinjo K
      • Sato H
      • Nakatani D
      • et al.
      Osaka Acute Coronary Insufficiency Study (OACIS) Group. Predictors of length of hospital stay after acute myocardial infarction in Japan.
      The higher direct medical expenditure among AMI individuals could also be ascribed to the observation that in our sample, AMI individuals tend to be older and have multiple comorbidities and therefore prone to experience more severe disease and poor outcomes.
      • Leal MF
      • Souza NF
      • Filho H
      • et al.
      Acute myocardial infarction in elderly patients: comparative analysis of the predictors of mortality. The elderly versus the young.
      ,
      • Fang J
      • Alderman MH
      • Keenan NL
      • et al.
      Acute myocardial infarction hospitalization in the United States, 1979 to 2005.
      Medication and outpatient expenditure also contributed to the burden of direct medical expenditure among AMI individuals with nearly $1,000 higher mean yearly per person expenditure compared with persons with no AMI. A typical medication regimen after myocardial infarction, usually initiated prior to hospital discharge, includes at least 4 agents such as antiplatelet drugs, angiotensin converting enzyme inhibitor, a statin and a beta-blocker. While some antiplatelet drugs such as aspirin and clopidogrel are generic and relatively low cost, there are also several newer and more expensive agents that are increasingly prescribed.
      • Esteve-Pastor MA
      • Ruíz-Nodar JM
      • Orenes-Piñero E
      • et al.
      Temporal trends in the use of antiplatelet therapy in patients with acute coronary syndromes.
      Furthermore, some patients will require additional pharmacological therapy such as antiarrhythmic drugs or diuretics.
      In addition to residual angina, AMI survivors often suffer from HF. Angina and HF may decrease mobility and have a negative impact on quality of life and vocational potential. Nearly 1 of 4 AMI survivors will experience a recurrence within a year, further contributing to missed workdays and productivity loss.
      • Evanchan J
      • Donnally MR
      • Binkley P
      • et al.
      Recurrence of acute myocardial infarction in patients discharged on clopidogrel and a proton pump inhibitor after stent placement for acute myocardial infarction.
      Finally, the psychiatric and psychologic burden of AMI, often manifested by depression and anxiety, is another contributor to the indirect cost incurred due to AMI.
      • Thombs BD
      • Bass EB
      • Ford DE
      • et al.
      Prevalence of depression in survivors of acute myocardial infarction.
      Another important finding of this study is that uninsured patients represented only 4.5% of individuals with AMI. This likely results from a combination of at least 2 factors – first, uninsured patients tend to be young with a proportion of uninsured of 30% in the age group 19-25 year olds, 28.3% in 26-34 year olds and 22.0% in 35-44 year olds as compared with 14.4% in the age group 55-64 year olds, as reported in 2011.

      The Henry J Kaiser Family Foundation. The uninsured: a primer. Key Facts about Health Insurance on the Eve Health Reform, 2016. Available at:https://kaiserfamilyfoundation.files.wordpress.com/2013/10/7451-09-the-uninsured-a-primer-key-facts-about-health-insurance.pdf. Accessed June 5, 2019.

      Second, AMI is more common in advancing age – the proportion of AMI patients was 59.2% for ≥ 65 years versus 40.8% for <65 years.
      Based on the estimates presented herein, the total cost of AMI accounted for 2.5% of the total 2016 National Health Expenditures for United States.
      • Hartman M
      • Martin AB
      • Espinosa N
      • et al.
      National Health Care Spending in 2016: spending and enrollment growth slow after initial coverage expansions.
      An estimated 4.2 million people were diagnosed with AMI and nearly 133,500 annual AMI-related deaths per year were identified in the current study. As demonstrated in this study, these grim epidemiological features engender considerably high economic cost. The observation that more patients with AMI were publicly insured suggests that the financial burden of AMI will be borne primarily by taxpayer's contributions, further stressing the need for rapid and innovative approaches to curb these estimates. Such interventions would ideally include prevention strategies, clinical management, rehabilitation and psycho-social support. They will implicate all stakeholders including epidemiologists, clinicians, families and policy makers. Finally, the estimated AMI-related mortality cost was substantial, averaging $40.5 billion per year or nearly half of the total aggregated AMI cost in 2016. This was even more substantial in the working age population. Our estimates were made from the National Vital Statistical Report (NVSR, 2003-2014) based on the PV of one life adopted from Grosse et al.
      • Grosse SD
      • Krueger KV
      • Mvundura M
      Economic productivity by age and sex: 2007 estimates for the United States.
      We believe that these are nationally representative cost estimates because data from the NVSR are obtained from all the 50 states and the District of Columbia. Furthermore, standardization efforts require that death be reported using ICD codes across the nation. The looming cost associated with AMI mortality is probably the reflection of the significant mortality rate associated with AMI. Despite a decline in cardiovascular mortality, a considerable number of Americans continue to die from AMI with most of them dying before arriving to the hospital. For example, in 2013 it was estimated that one American dies of a coronary event about every 84 seconds.
      • Mozaffarian D
      • Benjamin EJ
      • Go AS
      • et al.
      Executive summary: heart disease and stroke statistics–2016 update: a report from the American Heart Association.
      In all, our estimates reinforce the notion that beyond preventive measures which are paramount, efforts to curb these costs should also aim at improving early recognition of AMI and prehospital management.
      Although the results presented and the nature of the MEPS did not allow us to gauge the impact of prevention strategies in minimizing the costs associated with AMI, these measures in general have been proven to mitigate healthcare costs related to chronic medical conditions. For example, compared to less healthy individuals, relatively healthy patients do not require long hospital stay or extensive procedures, which may translate into health care cost saving.
      • Kauf TL
      • Velazquez EJ
      • Crosslin DR
      • et al.
      The cost of acute myocardial infarction in the new millennium: evidence from a multinational registry.
      ,
      • Seo H
      • Yoon S-J
      • Yoon J
      • Kim D
      • et al.
      Recent trends in economic burden of acute myocardial infarction in South Korea.
      In a study of the impact of a government-directed regional cardiovascular center (RCVC) project on the length of stay (LOS) and medical costs due to acute myocardial infarction, the implementation of prevention strategies and treatment of CVD resulted in shorter LOS for patients with AMI hospitalized in a RCVC and the total medical costs decreased by 797 dollars compared with patients hospitalized in a hospital not designed as a RCVC.
      • Kim A
      • Yoon SJ
      • Kim YA
      • et al.
      The burden of acute myocardial infarction after a regional cardiovascular center project in Korea.
      Prior studies have quantified the economic impact of specific therapeutic and preventive measures in reducing the cost of acute myocardial infarction.
      • Mark DB
      • Hlatky MA
      • Califf RM
      • et al.
      Cost effectiveness of thrombolytic therapy with tissue plasminogen activator as compared with streptokinase for acute myocardial infarction.
      • Naidoo B
      • Stevens W
      • McPherson K
      Modelling the short term consequences of smoking cessation in England on the hospitalisation rates for acute myocardial infarction and stroke.
      • Zeng W
      • Stason WB
      • Fournier S
      • et al.
      Benefits and costs of intensive lifestyle modification programs for symptomatic coronary disease in Medicare beneficiaries.
      Similarly, the barriers and their economic impact on AMI have been identified.
      • Rahimi AR
      • Spertus JA
      • Reid KJ
      • et al.
      Financial barriers to health care and outcomes after acute myocardial infarction.
      Altogether, those studies have helped model cost-saving strategies vital to creating policy-forming recommendations. The current study is an econometric analysis, which has provided important benchmark estimates on the direct and indirect cost of acute myocardial infarction. The findings from this study could renew interests in more granular evaluations of specific interventions aiming at curbing the economic cost of acute myocardial infarction. For example, further studies could evaluate the impact of those interventions across all age groups, focusing on individual in the most productive age groups. Building on our findings, it would be expected that the evaluation of financial benefits of intensive lifestyle modifications or specific endovascular interventions on AMI also include indirect cost components.
      This study has limitations. First, the direct medical expenditure for institutionalized population such as those of nursing home residents, military personnel and people in prison are not included in this analysis as these individuals are not surveyed in MEPS. Second, payments for over-the-counter drugs are not collected in MEPS, which may also underestimate the direct medical expenditure. Third, we were unable to capture other indirect costs like presenteeism, lost tax revenues, pain and stress. Fourth, procedural costs with a potential of higher expenditures associated with AMI such as coronary artery bypass graft surgery and percutaneous coronary interventions could not be computed as the diagnoses for these procedures are not provided in MEPS data. Fifth, we did not capture the indirect cost of wage-based lost productivity for working people aged ≥65 years and for early retirement. Altogether, our estimates are likely to be conservative. Sixth, we used pooled cross-sectional data and thus we cannot infer causality. Seventh, AMI diagnosis was based on ICD-9-CM codes which may be subject to underreporting and sampling error. ICD-9-CM codes of 410 included ST-elevation myocardial infarction and non-ST-elevation myocardial infarction are reported as one condition in MEPS, therefore limiting our ability to estimate each cost separately. However, we used an up to date 12-year nationally representative database, a novel 2-part model that addresses the statistical biases associated with cost analyses as observations with zero-costs, right skewed and heteroscedasticity problems are frequent.
      • Manning WG
      • Mullahy J
      Estimating log models: to transform or not to transform?.
      ,
      • Belotti F
      • Deb P
      • Manning WG
      • et al.
      Twopm: two-part models.
      Unlike previous studies, our analysis spans all age groups and quantifies both the indirect and direct cost of AMI in the United States.

      CONCLUSIONS

      Every year, roughly $84.9 billion in costs are incurred by AMI (approximately $20,176 per patient/year in 2016 $) in the United States. Hospital inpatient stay was the major contributor to the direct medical expenditures while lost productivity from premature mortality represents the majority of the indirect cost of AMI. This study finds that the economic burden of AMI is substantial, for which effective prevention could result in significant health and productivity cost savings. Future research will focus on developing more refined prediction tools to improve cost-saving preventive interventions.

      AVAILABILITY OF DATA AND MATERIALS

      Authors used publicly available data from the Medical Expenditure Panel Survey (MEPS). The authors did not have direct contact with survey participants. Data are available at: https://meps.ahrq.gov/mepsweb/data_stats/download_data_files.jsp

      AUTHOR CONTRIBUTIONS

      K.B.: study concept and design, acquisition of data and statistical analysis, data interpretation, drafting of the manuscript and critical revision of the manuscript for important intellectual content. A.L., E.K., S.S., A.S. and M.H.: study concept and design, data interpretation, drafting of the manuscript and critical revision of the manuscript for important intellectual content. WM: study concept and design, data interpretation, and critical revision of the manuscript for important intellectual content. PM: study concept and design, acquisition of data and data interpretation, critical revision of the manuscript for important intellectual content and study supervision. All authors read and approved the final manuscript.

      GUARANTORS

      The guarantors of this study are KB and AL. We take full responsibility for this work.

      REFERENCES

        • Boateng S
        • Sanborn T
        Acute myocardial infarction.
        Dis Mon. 2013; 59: 83-96
        • Kochanek KD
        • Murphy SL
        • Xu JQ
        • et al.
        Deaths: final data for 2014.
        Natl Vital Stat Rep. 2016; 65: 1-121
        • Ghushchyan V
        • Nair KV
        • Page II, RL
        Indirect and direct costs of acute coronary syndromes with comorbid atrial fibrillation, heart failure, or both.
        Vasc Health Risk Manag. 2015; 11: 25-34
        • Kolansky DM
        Acute coronary syndromes: morbidity, mortality and pharmacoeconomic burden.
        Am J Manag Care. 2009; 15: S36-S41
        • Hartman M
        • Martin AB
        • Espinosa N
        • et al.
        National Health Care Spending in 2016: spending and enrollment growth slow after initial coverage expansions.
        Health Aff (Millwood). 2018; 37: 150-160
        • Dieleman JL
        • Baral R
        • Birger M
        • et al.
        US spending on personal health care and public health, 1996-2013.
        JAMA. 2016; 316: 2627-2646
        • Afana M
        • Brinjikji W
        • Cloft H
        • et al.
        Hospitalization costs for acute myocardial infarction patients treated with percutaneous coronary intervention in the United States are substantially higher than Medicare payments.
        Clin Cardiol. 2015; 38: 13-19
        • Kauf TL
        • Velazquez EJ
        • Crosslin DR
        • et al.
        The cost of acute myocardial infarction in the new millennium: evidence from a multinational registry.
        Am Heart J. 2006; 151: 206-212
        • Zhou X
        • Shrestha SS
        • Luman E
        • et al.
        Medical expenditures associated with diabetes in myocardial infarction and ischemic stroke patients.
        Am J Prev Med. 2017; 53 (Supplement 2): S190-S196
      1. Agency for Healthcare Research and Quality (AHRQa). Medical expenditure panel survey. 2014 Full year consolidated data file 2016a, Available at:https://meps.ahrq.gov/data_stats/download_data/pufs/h171/h171doc.pdf. Accessed August 1, 2019.

      2. Agency for Healthcare Research and Quality (AHRQb). Medical expenditure panel survey, 2014 medical conditions 2016b, Available at:https://meps.ahrq.gov/data_stats/download_data/pufs/h170/h170doc.pdf. Accessed August 1, 2019.

      3. National Center for Health Statistics. Health, United States, 2009: with special feature on medical technology. Hyattsville, MS. 2010, Available at:https://www.cdc.gov/nchs/data/hus/hus09.pdf. Accessed August 1, 2019.

        • Libby AM
        • Ghushchyan V
        • McQueen RB
        • et al.
        Economic differences in direct and indirect costs between people with epilepsy and without epilepsy.
        Med Care. 2012; 50: 928-933
        • Cawley J
        • Rizzo J
        • Haas K
        The association of diabetes with job absenteeism costs among obese and morbidly obese workers.
        J Occup Environ Med. 2008; 50: 527-534
        • Bishu KG
        • Gebregziabher M
        • Dismuke CE
        • et al.
        Quantifying the incremental and aggregate cost of missed work days in adults with diabetes.
        J Gen Intern Med. 2015; 30: 1773-1779
        • D'Hoore W
        • Bouckaert A
        • Tilquin C
        Practical considerations on the use of the Charlson comorbidity index with administrative data bases.
        J Epidemiol. 1996; 49: 1429-1433
        • Manning WG
        • Mullahy J
        Estimating log models: to transform or not to transform?.
        J Health Econ. 2001; 20: 461-494
        • Belotti F
        • Deb P
        • Manning WG
        • et al.
        Twopm: two-part models.
        Stata J. 2015; 15: 3-20
        • Archer KJ
        • Lemeshow S
        Goodness-of-fit test for a logistic regression model fitted using survey sample data.
        Stata J. 2006; 6: 97-105
      4. Bureau of Labor Statistics, CPI Inflation Calculator. Available at: https://www.bls.gov/data/inflation_calculator.htm. Accessed July 10, 2019.

        • StataCorp
        Stata: Release 14. Statistical Software.
        StataCorp LP, College Station, TX2015
        • Grosse SD
        • Krueger KV
        • Mvundura M
        Economic productivity by age and sex: 2007 estimates for the United States.
        Med Care. 2009; 47: S94-103
      5. U.S. Bureau of Economic Analysis, Gross Domestic Product: Implicit Price Deflator [GDPDEF], retrieved from FRED, Federal Reserve Bank of St. Louis, Available at: https://fred.stlouisfed.org/series/GDPDEF. Accessed June 5, 2019.

        • Hoyert DL
        • Heron MP
        • Murphy SL
        • et al.
        Deaths: final data for 2003.
        Natl Vital Stat Rep. 2006; 54: 1-119
        • Miniño AM
        • Heron MP
        • Murphy SL
        • et al.
        Deaths: final data for 2004.
        Natl Vital Stat Rep. 2007; 55: 1-119
        • Kung HC
        • Hoyert DL
        • Xu JQ
        • et al.
        Deaths: final data for 2005.
        Natl Vital Stat Rep. 2008; 56: 1-120
        • Heron MP
        • Hoyert DL
        • Murphy SL
        • et al.
        Deaths: final data for 2006.
        Natl Vital Stat Rep. 2009; 57: 1-134
        • Xu JQ
        • Kochanek KD
        • Murphy SL
        • et al.
        Deaths: final data for 2007.
        Natl Vital Stat Rep. 2010; 58: 1-134
        • Miniño AM
        • Murphy SL
        • Xu JQ
        • et al.
        Deaths: final data for 2008.
        Natl Vital Stat Rep. 2011; 59: 1-126
        • Kochanek KD
        • Xu J
        • Murphy SL
        • et al.
        Deaths: final data for 2009.
        Natl Vital Stat Rep. 2011; 60: 1-116
        • Murphy SL
        • Xu JQ
        • Kochanek KD
        Deaths: final data for 2010.
        Natl Vital Stat Rep. 2013; 61: 1-117
        • Kochanek KD
        • Murphy SL
        • Xu JQ
        Deaths: final data for 2011.
        Natl Vital Stat Rep. 2015; 63: 1-117
        • Murphy SL
        • Kochanek KD
        • Xu JQ
        • et al.
        Deaths: final data for 2012.
        Natl Vital Stat Rep. 2015; 63: 1-117
        • Xu JQ
        • Murphy SL
        • Kochanek KD
        • et al.
        Deaths: final data for 2013.
        Natl Vital Stat Rep. 2016; 64: 1-118
        • Prabhakaran S
        • Ruff I
        • Bernstein RA
        Acute stroke intervention: a systematic review.
        JAMA. 2015; 313: 1451-1462
        • Heidenreich PA
        • Albert NM
        • Allen LA
        • et al.
        Forecasting the impact of heart failure in the United States: a policy statement from the American Heart Association.
        Circ Heart Fail. 2013; 6: 606-619
        • Anderson JL
        • Morrow DA
        Acute myocardial infarction.
        N Engl J Med. 2017; 376: 2053-2064
        • Kinjo K
        • Sato H
        • Nakatani D
        • et al.
        Osaka Acute Coronary Insufficiency Study (OACIS) Group. Predictors of length of hospital stay after acute myocardial infarction in Japan.
        Circ J. 2004; 68: 809-815
        • Leal MF
        • Souza NF
        • Filho H
        • et al.
        Acute myocardial infarction in elderly patients: comparative analysis of the predictors of mortality. The elderly versus the young.
        Arq Bras Cardiol. 2002; 79: 363-374
        • Fang J
        • Alderman MH
        • Keenan NL
        • et al.
        Acute myocardial infarction hospitalization in the United States, 1979 to 2005.
        Am J Med. 2010; 123: 259-266
        • Esteve-Pastor MA
        • Ruíz-Nodar JM
        • Orenes-Piñero E
        • et al.
        Temporal trends in the use of antiplatelet therapy in patients with acute coronary syndromes.
        J Cardiovasc Pharmacol Ther. 2018; 23: 57-65
        • Evanchan J
        • Donnally MR
        • Binkley P
        • et al.
        Recurrence of acute myocardial infarction in patients discharged on clopidogrel and a proton pump inhibitor after stent placement for acute myocardial infarction.
        Clin Cardiol. 2010; 33: 168-171
        • Thombs BD
        • Bass EB
        • Ford DE
        • et al.
        Prevalence of depression in survivors of acute myocardial infarction.
        J Gen Intern Med. 2006; 21: 30-38
      6. The Henry J Kaiser Family Foundation. The uninsured: a primer. Key Facts about Health Insurance on the Eve Health Reform, 2016. Available at:https://kaiserfamilyfoundation.files.wordpress.com/2013/10/7451-09-the-uninsured-a-primer-key-facts-about-health-insurance.pdf. Accessed June 5, 2019.

        • Hartman M
        • Martin AB
        • Espinosa N
        • et al.
        National Health Care Spending in 2016: spending and enrollment growth slow after initial coverage expansions.
        Health Aff (Millwood). 2018; 37: 150-160
        • Mozaffarian D
        • Benjamin EJ
        • Go AS
        • et al.
        Executive summary: heart disease and stroke statistics–2016 update: a report from the American Heart Association.
        Circulation. 2016; 133: 447-454
        • Seo H
        • Yoon S-J
        • Yoon J
        • Kim D
        • et al.
        Recent trends in economic burden of acute myocardial infarction in South Korea.
        PLoS One. 2015; 10e0117446https://doi.org/10.1371/journal.pone.0117446
        • Kim A
        • Yoon SJ
        • Kim YA
        • et al.
        The burden of acute myocardial infarction after a regional cardiovascular center project in Korea.
        Int J Qual Health Care. 2015; 27: 349-355
        • Mark DB
        • Hlatky MA
        • Califf RM
        • et al.
        Cost effectiveness of thrombolytic therapy with tissue plasminogen activator as compared with streptokinase for acute myocardial infarction.
        N Engl J Med. 1995; 332: 1418-1424
        • Naidoo B
        • Stevens W
        • McPherson K
        Modelling the short term consequences of smoking cessation in England on the hospitalisation rates for acute myocardial infarction and stroke.
        Tob Control. 2000; 9: 397-400
        • Zeng W
        • Stason WB
        • Fournier S
        • et al.
        Benefits and costs of intensive lifestyle modification programs for symptomatic coronary disease in Medicare beneficiaries.
        Am Heart J. 2013; 165: 785-792
        • Rahimi AR
        • Spertus JA
        • Reid KJ
        • et al.
        Financial barriers to health care and outcomes after acute myocardial infarction.
        JAMA. 2007; 297: 1063-1072