If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Correspondence: Kinfe G. Bishu, PhD, Section of Health Systems Research and Policy, Medical University of South Carolina, 135 Rutledge Avenue, Room 1233, Charleston, SC 29425-0593
Section of Health Systems Research and Policy, Medical University of South Carolina, Charleston, South CarolinaDivision of General Internal Medicine and Geriatrics, Department of Medicine, Medical University of South Carolina, Charleston, South CarolinaCharleston Health Equity and Rural Outreach Innovation Center (HEROIC), Ralph H. Johnson VA Medical Center, Charleston, South Carolina
Section of Health Systems Research and Policy, Medical University of South Carolina, Charleston, South CarolinaDivision of General Internal Medicine and Geriatrics, Department of Medicine, Medical University of South Carolina, Charleston, South Carolina
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.
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.
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.
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.
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
Hospitalization costs for acute myocardial infarction patients treated with percutaneous coronary intervention in the United States are substantially higher than Medicare payments.
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.
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.
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.
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
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.
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.
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.
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.
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.
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,
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.
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.
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).
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.
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.
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.
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.
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.
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.
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.
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.
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.
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).
Variables
AMI (%)
No-AMI (%)
P Value
N (n)
4,212,846 (4,769)
249,022,206 (320,100)
Age category
Age <65
40.8
85.5
<0.001
Age ≥65
59.2
14.5
Sex
Male
64.1
46.7
<0.001
Female
35.9
53.5
Race/ethnicity
Non-Hispanic White
80.5
68.4
<0.001
Non-Hispanic Black
9.0
11.1
Hispanic
6.8
13.7
Others
3.7
6.8
Marital status
Married
58.0
52.9
<0.001
Non-married
36.0
20.3
Never married
6.0
26.8
Education category
<High school
9.8
19.7
<0.001
High school graduate
46.2
33.9
College or more
44.0
46.4
Insurance
Private
57.0
69.6
<0.001
Public
38.5
20.4
Uninsured
4.5
10.0
Census region
Northeast
18.4
18.2
<0.001
Midwest
25.4
22.6
South
39.0
36.3
West
17.2
22.9
Income category
Poor income
21.5
17.3
<0.001
Low income
17.7
13.4
Middle income
29.9
30.3
High income
30.9
39.0
Charlson Comorbidity Index
0
36.2
77.0
<0.001
1
11.8
8.3
≥2
52.0
14.7
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 $).
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.
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).
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.
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.
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.
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 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 1Annual aggregate cost of acute myocardial infraction (in 2016 $).
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
Hospitalization costs for acute myocardial infarction patients treated with percutaneous coronary intervention in the United States are substantially higher than Medicare 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.
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.
Recurrence of acute myocardial infarction in patients discharged on clopidogrel and a proton pump inhibitor after stent placement for acute myocardial infarction.
30054-9&id=gr1.jpg){kind=link}