Hospital Utilization for Coronary Artery Disease, 1997–2014

Abstract

Background 

Coronary artery disease (CAD) is the leading cause of cardiovascular morbidity, mortality, and healthcare costs in the United States. There are few reports on how public health and payment reforms might have influenced inpatient hospital use among patients with CAD.

Objective 

This study describes trends in hospital discharges, hospital charges, and discharge destinations in a national sample of patients with CAD between 1997 and 2014.

Methods 

This was a longitudinal study with descriptive analysis of the Healthcare Cost and Utilization Project of National Inpatient Sample data.

Findings 

During this study period, the total number of discharges was 1 333 996. Patients with CAD between 65 and 84 years old were among the highest users of inpatient hospital services, followed by those in the 45- to 64-year age group. The death rate increased from 5961 to 7217 per 10 000 patients during this time. The mean charge increased more than 5 times, from $9100 to $49 643. There was a large difference in mean hospital charges in urban ($51 666) and rural ($25 548) locations in 2014. Coronary artery disease patients with private insurance paid more than those with Medicaid and Medicare plans. The discharge to home and healthcare costs increased by 4.1% and 4.8%, respectively.

Conclusion and Implications 

Future researchers should use data sets, such as Medicare claims/Medical Expenditure Panel Study, that can provide comprehensive insights into patient-level factors influencing the use of inpatient care services among patients with CAD. Healthcare providers in posthospital settings should be well skilled in providing advanced cardiac rehabilitation and education to patients with CAD.

Coronary artery disease (CAD) involves inflammatory changes and the building of plaques in the coronary arteries, eventually compromising blood supply to the heart.¹ Coronary artery disease is linked with adverse cardiovascular events such as myocardial infarction, cerebral infarction,^2,3 cognitive impairment, dementia,⁴ premature mortality,⁵ and postoperative complications, including prolonged ventilation and intensive care unit stays, atrial fibrillation, and cerebrovascular accidents.⁶ Approximately 15.5 million adults (≥20 years old) in the United States are inflicted by CAD.⁷ Coronary artery disease is one of the most common conditions for emergency department visits in the United States.⁸ The care related to CAD is costly, and it ascribes to $219-billion healthcare cost annually due to either the use of healthcare services or the loss of productivity.⁹ The burden on the healthcare system incurred due to CAD is projected to increase with the rapid growth of the aging population.⁸

In the United States, nationwide public health programs and agendas such as Healthy People 2000 and 2010 were formulated and implemented from the beginning of the 1990s¹⁰ to promote cardiovascular health and curb cardiovascular disease at the population level. At the same time, several payment reforms were initiated in the late 1990s to curtail acute hospital expenditure and promote the use of cost-effective postacute services. In addition, the Affordable Healthcare Act enacted in 2010 expanded healthcare coverage for various chronic conditions, including cardiovascular diseases, to promote cardiovascular health and disease prevention.^11–13 These national policies and strategies can influence the healthcare utilization among patients with CAD.¹¹

Previous studies have shown that the national trends of the emergency department use are decreasing in patients with CAD.¹¹ Wide disparities have been documented in terms of the services and quality of care provided to patients with CAD.¹⁴ However, information on the trends of inpatient hospital utilization among patients with CAD is lacking. Therefore, there is a need to examine the trends of inpatient hospital care utilization for CAD stratified by gender, age group, insurance status, hospital ownership, and geographical location to infer the impact of policy and payment reforms.

This study aims to address this gap and investigate the trends of inpatient hospital use including length of hospital stays, mean hospital charges, and discharge outcomes for CAD and the variation in the inpatient hospital use by gender, age, insurance status, hospital ownership, and geographical location. In this study, the National Inpatient Sample (NIS), a large national data set of inpatient hospital stays, was used to examine trends in hospital utilization for CAD between 1997 and 2014 post public health and payment reforms. This study aims to answer these research questions: (a) how did the trends of inpatient hospital use related to CAD change over 17 years, and (b) how did the trends of costs and discharge outcomes associated with CAD change over 17 years?

Methods

This is a retrospective longitudinal study. The data for this study came from the NIS, which is funded by the Agency for Healthcare Research and Quality. The NIS is part of the Healthcare Utilization Project (HCUPnet) databases and is publicly available containing information on regional and national inpatient utilization, access, charges, quality, and outcomes based on the billing data presented by hospitals and statewide data organizations across the United States.^15,16 The NIS is nationally representative of all community hospitals, and it includes all the patients from each hospital, irrespective of their insurance payment status.¹⁷ The inclusion criteria for this study are admission to inpatient hospitals with the diagnosis of CAD (International Classification of Diseases, Ninth Revision code 414.00). The search was restricted between 1997 and 2014 to maintain the consistency of coding for CAD as per International Classification of Diseases, Ninth Revision standards. This study does not include data from 2015 onwards because the revised coding for CAD as per International Classification of Diseases, Tenth Revision standards (introduced in 2015) does not match the International Classification of Diseases, Ninth Revision standard.

Variables

The operational definitions of constructs of interest for this study are as follows: (a) length of stay (LOS) is the number of nights spent in the hospital to receive care after being admitted to the hospital unit for CAD; (b) the charge is the amount of US dollars charged for using hospital services, excluding physician services; (c) discharge is the unit of analysis for HCUPnet, and discharges are number of cases with CAD who were admitted and discharged from the hospital; (d) discharge status is the discharge disposition post hospital stay, which can be either home, short-term hospital, nursing home, home healthcare, or leave against medical advice; and (e) inpatient mortality represents the number of patients who died in the hospital after admission from the emergency department per 10 000 admitted patients.

Data Analysis

An online query system available on the HCUPnet website was used to evaluate the trends of inpatient service utilization among patients given a diagnosis of CAD between 1997 and 2014.^16,17 The unit of analysis was the discharge.^16,18 Trends data were extracted, aggregated, cleaned, and examined for trends in the Excel spreadsheet. Data on number of discharges, LOS, hospital charges, inpatient hospital mortality, and discharge destination were stratified by each year. A descriptive summary and trends of number of discharges, LOS, hospital charges, inpatient hospital mortality, and discharge destination between 1997 and 2014 are described in the Results section.¹⁶

Results

Table 1 shows the utilization of inpatient hospital services for CAD, stratified by year from 1997 to 2014. During this period, an estimated total of 1 333 996 discharges were reported for the patients given a diagnosis of CAD. The total discharge for CAD was 107 625 in 1997, followed by a dip in 1999 (102 625). The annual discharges increased gradually and peaked in the year 2001 (n = 115 855); then, there was a gradual decline in the total discharge in 2014 (27 020). As shown in Figure 1, the trends of inpatient service use for CAD by age groups, remained consistently higher for 65–84 years, followed by 45–64 years, 85 years and above, and 18–44 years. The highest cases admitted and discharged were between 65 and 84 years old (n = 707 492), followed by 45 to 64 years old (n = 445 713), 85 years and older (n = 113 722), and 18 to 44 years old (n = 48 687). During this period, the total number of cases admitted and discharged were men (n = 758 171) relative to women (n = 541 871). The trends of inpatient hospital service use for CAD stratified by insurance type show that the total number of cases admitted and discharged for CAD remained consistently higher for Medicare beneficiaries and lower among those with unspecified payer status, as shown in Figure 2. The total number of cases admitted and discharged for CAD was greatest among Medicare beneficiaries (n = 862 574) and lowest among those with unspecified payer status (n = 29 531).

FIGURE 1: Acute care use for coronary artery disease stratified by age groups between 1997 and 2014.

TABLE 1 - Inpatient Service Use for Coronary Artery Disease Stratified by Year, 1997–2014

Variable	Year
	1997	1998	1999	2000	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	2013	2014
Total no. discharges	107 247	107 245	102 626	111 285	115 855	109 841	94 745	91 516	76 549	70 787	61 814	57 132	51 931	41 260	40 489	36 320	30 335	27 020
Mean LOS, d	3.2	3.06	2.9	2.9	2.8	2.9	2.7	2.7	2.8	2.7	2.7	2.8	2.8	2.7	2.7	2.8	2.9	3.3
Mean charges in US $	9100	9731	10 425	10 500	11 881	14 151	15 069	16 496	18 063	20 120	22 220	26 920	27 492	29 908	33 509	36 860	44 606	49 643
In-hospital deaths (death rate per 10 000)	5962	6392	5487	5785	5427	4809	4411	4709	3951	5107	4821	7235	4885	5801	7219	7296	9725	7217
Routine discharge	73 587	73 206	69 175	75 218	79 534	76 323	65 755	62 558	51 357	47 264	43 395	40 412	36 952	28 962	28 944	27 010	22 150	19 640
Another short-term hospital	20 118	21 713	22 166	24 034	23 693	21 712	17 974	17 533	14 212	13 245	9819	8074	7835	5704	5061	3515	2840	2295
Another institution (nursing home, rehabilitation)	6456	6104	5473	5704	6304	5699	5221	4970	4795	4437	3997	3672	2897	2588	2503	2125	1890	1755
Home healthcare	4921	4223	4050	4406	4181	3918	3865	4559	4563	4108	3415	3575	3018	2938	2912	2795	2670	2530
Against medical advice	1468	1282	1165	1263	1410	1551	1414	1368	1284	1336	849	946	916	757	723	605	485	550

Source: HCUPnet, Healthcare Cost and Utilization Project. Rockville, MD: Agency for Healthcare Research and Quality. https://hcupnet.ahrq.gov/. For more information about HCUP data, see http://www.hcup-us.ahrq.gov/.

Abbreviation: LOS, length of stay.

As shown in Table 1, the mean length of hospital stay remained consistent around 3 days between 1997 and 2014. In 2014, the mean LOS was the longest for those aged 65 to 84 years (3.5 days), whereas the mean LOS was shortest for those 18 to 44 years old and older than 85 years (approximately 3 days for both age groups). In 2014, the Medicaid beneficiaries (4.2 days) experienced the longest stay in the hospital, whereas the LOS was shorter (2.7 days) for those with unspecified payer status. The length of hospital stays among patients with CAD in rural and urban locations remained relatively similar (3–3.5 days) during this time.

The mean hospital charges increased remarkably during this period. The mean hospital charge increased more than 5 times from $9100 in 1997 to $49 643 in 2014 (see Table 1). The mean hospital charges were comparably higher among those with private insurance relative to those with unspecified payer status, as shown in Table 2. Similarly, the mean hospital charges remained relatively higher for the private (for-profit) hospital in comparison with government and not-for-profit hospitals, as shown in Table 2. Finally, the hospitals in the urban location observed higher mean charges compared with those in the rural location. Both urban and rural hospitals experienced a sharp rise in mean hospital charges in the year 2001 ($76 044 and $39 811, respectively), followed by a decline in 2002 ($17 132 and $7339, respectively); the mean charges gradually increased until the year 2014 ($51 666 and $25 548, respectively, in the rural and urban locations).

FIGURE 2: Acute care use for coronary artery disease stratified by insurance type between 1997 and 2014.

TABLE 2 - Trends in Mean Charges for Patients With Coronary Artery Disease Stratified by Year, 1997–2014

Variable	Year
	1997	1998	1999	2000	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	2013	2014
Insurance type
Medicare	8684	9217	9755	10 083	10 083	13 745	14 405	16 009	17 567	21 622	26 274	26 274	26 636	29 805	32 022	36 557	44 288	48 706
Medicaid	10 156	10 242	10 695	12 511	12 511	15 095	16 490	16 630	17 608	22 096	25 260	25 260	25 890	28 422	32 925	36 348	43 345	50 520
Private insurance	9920	10 994	12 080	11 337	11 337	15 590	16 856	17 926	20 230	24 774	29 902	29 901	32 063	32 006	39 959	38 972	47 122	54 511
Uninsured	8634	9170	9644	8523	8523	11 585	13 973	18 659	16 912	19 391	24 101	24 101	22 844	27 612	32 798	30 593	42 594	45 876
Other	9511	10 067	10 427	11 466	11 467	12 871	15 522	14 298	16 370	21 371	26 851	26 851	28 285	27 506	35 075	42 280	42 994	43 509
Hospital ownership
Government	7284	7350	8154	9780	18 004	10 185	11 263	11 676	14 299	13 356	17 929	23 124	18 602	24 427	24 060	28 223	36 448	42 366
Private, not-for-profit	9048	10 285	10 249	10 254	81 763	14 834	15 556	16 914	18 498	20 467	21 778	26 942	27 237	28 803	32 769	34 781	41 640	45 494
Private, for-profit	10 739	9615	12 965	12 103	16 087	15 072	16 841	19 379	19 109	25 802	28 529	30 241	33 564	38 483	42 597	49 648	59 927	68 973
Geographical location
Rural	4912	5124	5795	6132	39 811	7339	7769	9853	9254	10 367	12 888	13 598	14 463	15 585	17 234	29 339	24 891	25 550
Urban	11 036	11 808	12 392	12 740	76 044	17 132	18 045	19 640	21 650	23 530	25 180	30 635	30 823	33 418	376 912	34 600	48 000	51 666

Source: HCUPnet, Healthcare Cost and Utilization Project. Rockville, MD: Agency for Healthcare Research and Quality. https://hcupnet.ahrq.gov/. For more information about HCUP data, see http://www.hcup-us.ahrq.gov/.

The total number of routine discharges from the hospital for CAD cases decreased from 1997 (n = 107 247) to 2014 (n = 27 020). The post-discharge destinations show the total number of cases referred to short-term hospitals fell from 20 118 in 1997 to 2295 in 2014. Similarly, the total number of cases going to rehabilitation and nursing homes declined from 6456 in 1997 to 1755 in 2014. The number of cases receiving home health care decreased from 4921 in 1997 to 2530 in 2014. The overall number of inpatient hospital deaths rose from 5962 in 1997 to 7217 per 10 000 in 2014, with fluctuation in between. In 2014, the mortality rate was higher for males (7717 per 10 000 persons) in comparison to females (6968 per 10 000 persons). The percentages have been replaced with numbers to correspond with the information in the Table 1. In terms of mortality, the mortality rate fluctuated between 1997 and 2014. The in-hospital deaths gradually decreased from 5961 in 1997 to 3951 per 10 000 persons in 2005. This was followed by a substantial increase in the death rate to 9725 per 10 000 in 2013, and the inpatient death rate fell to 7217 per 10 000 in 2014. In 2014, the mortality rate was higher for men (7717 per 10 000 persons) in comparison with women (6968 per 10 000 persons).

Discussion

There was a wide variation in the inpatient hospital utilization for CAD across the United States by gender, age, insurance, hospital ownership, and geographical location. The change in the utilization of inpatient care may be due to the implementation of public health reforms for the promotion of cardiovascular health at the population level and the implementation of payment reforms.

During this period from the year 1997 to 2014, the total number of discharges declined for CAD cases. This finding is consistent with past studies, which showed decreasing trends of inpatient hospitalization^11,19 and discharge in patients with CAD.¹⁹ A greater number of patients admitted and discharged for CAD were older and middle aged, and men. Most patients with CAD who were admitted and discharged during this time had Medicare and private insurance. The overall inpatient deaths due to CAD increased during this time. This fall in annual discharge could be due to the overall decreasing trends in the prevalence of CAD in the United States.²⁰ The decrease in the overall prevalence of cardiovascular disease might be due to the implementation of nationwide public health programs and policies targeted at preventing cardiovascular diseases at the community level.²¹ Healthy People 2000 was first initiated in 1990 by the US Department of Health and Human Services with the objectives of reducing cardiovascular risk factors at the population level, expanding the health services for early detection and management of cardiovascular disease, and reducing health disparities.²¹ In 2000, this program was revised and followed by another nationwide program, the Healthy People 2010.¹⁰ In addition to this, in 2002, the Centers for Disease Control and Prevention sponsored the Heart Disease and Stroke Prevention Program to prevent cardiovascular risk factors and the onset of cardiovascular disease and mitigate the adverse outcomes in 29 US states.²² These findings indicate that health programs implemented for the primary prevention of cardiovascular diseases may successfully prevent and reduce CAD-related hospital use and influence inpatient deaths.

The mean hospital charges increased drastically during this period. Our findings suggest that soaring health care costs may discourage the use of inpatient hospital services for CAD in the early stage when the disease is manageable. As a result, CAD patients may be admitted to hospitals in chronic, severe stages requiring more advanced treatment. The aggressive treatment provided at the critical phase may be more expensive, but may not improve health outcomes in CAD patients. Thus, the rise in inpatient deaths, parallel with the escalation of hospital expenditure, suggests that surging health care costs may not guarantee an improvement in health care quality and outcomes. These findings corroborate with several past studies that show hospital care costs are substantially rising for cardiovascular diseases.^23,24 The rise in the inpatient hospital charges might be due to inflation and other patient factors, such as the presence of coexisting comorbid conditions and emotional distress, and hospital readmission.^23–25 The increase in inpatient hospital charges may be responsible for an early discharge to posthospital facilities. The mean inpatient hospital charges were substantially higher among those with private insurance relative to those with Medicare and Medicaid payer status. The inpatient hospital charges under Medicare and Medicaid payment systems are regulated by the Centers for Medicare & Medicaid Services, which has enacted payment reforms and fees for various in-service procedures and care.²⁶ In contrast, private health plans negotiate with the hospitals to set prices for in-service procedures and care through the process of contract negotiation, which might have attributed to greater hospital charges for those with private insurance plans.²⁶ In addition, there was a huge difference in inpatient hospital charges for CAD cases in hospitals located at rural and urban locations. This finding may suggest the differences in the availability of treatment and care options for CAD in rural and urban hospitals. A systematic study showed that patients with stroke who receive inpatient hospital care in urban locations have more access to advanced treatment options, including thrombolysis therapy, evidence-based care, diagnostic interventions, and skilled care providers.²⁷ Similarly, the mean hospital charges remained relatively higher for private/for-profit hospitals compared with government and not-for-profit hospitals, which is consistent with the existing studies.²⁸ This variation in hospital charges can be due to differences in charge rates for diagnostic procedures and treatment provided in government versus proprietary hospitals.²⁹

These findings indicate several policy and clinical implications. The findings suggest that public health programs targeted against cardiovascular disease might be effective in reducing CAD cases, which in turn, reduce inpatient hospital utilization. Future policies should emphasize secondary and tertiary prevention of cardiovascular disease, including prevention of cardiovascular adverse health outcomes. The payment reforms should also focus on regulating diagnostic procedure charges that can also contribute to inpatient hospital charges. The existing payment reforms seem to encourage CAD cases to be discharged early from the hospital to postacute facilities to curtail inpatient expenditure and LOS. This means that patients with CAD discharged to posthospital settings might require more care and support; thus, healthcare providers at posthospital settings should be well equipped in providing cardiac rehabilitation and other care support. At the same time, it is essential that the quality of care is not compromised during the hospital stay. Therefore, proper measures should be in place to evaluate whether the hospitals have adhered to care guidelines set by the regulating bodies while delivering care to patients with CAD.

Future research should use data sets, such as Medicare claims/Medical Expenditure Panel Study, that can provide richer insights into patient-level factors influencing the use of inpatient hospital services for patients with CAD. If possible, future studies should follow the panel of patients with CAD using a longitudinal study design to examine individual characteristics associated with inpatient hospital utilization. Future research should also possibly account for the covariates such as sociodemographics, health beliefs, and comorbidities that might affect inpatient hospital utilization among patients with CAD.

Although this study used large representative data to examine the difference in the utilization of inpatient hospital services for CAD, the findings should be considered carefully in light of limitations. First, findings are based on crude estimates of variables and may not indicate exact figures. This is a descriptive study, and no statistical analyses were performed to infer relationships among variables or control for confounders; thus, the causal inference could not be established. As the unit of analysis is discharge, specific patients with CAD were not examined. Therefore, despite being a longitudinal study, outcomes were examined among different patients, and a cohort of patients with CAD was not followed longitudinally. Because of a large number of missing values, the analyses of LOSs, mean hospital charges, and discharge outcomes could not be stratified by each age group, gender, insurance type, and hospital locations. Finally, the HCUPnet data only includes data from community hospitals; thus, the findings from this study may not represent all hospitals in the US.

Conclusion

The total charges for CAD rose substantially despite the consistent length of hospital stay and a decreasing number of discharges between 1997 and 2014. During this time, the percentage of routine discharge to homes without care, home healthcare, and rehabilitation facilities increased, whereas the discharge to other short-term hospitals reduced. Given the shortened LOS in hospital and raised hospital charges, it is important to promote cardiac rehabilitation care in posthospital settings.

What’s New and Important

• Acute care utilization in CAD was highest among men, middle-aged and older adults, and Medicare and Medicaid beneficiaries.
• A total number of discharges decreased, meanwhile hospital charges increased substantially by 445.5% during this period. This figure raises a critical issue regarding the affordability and accessibility of inpatient hospital care use among patients with CAD.
• The total LOS remained consistent; however, more patients were being discharged to postacute care settings, indicating a pressing need to train healthcare providers working in community settings about advanced cardiac rehabilitation skills.
• More studies are needed to understand how inpatient hospital service utilization varies by individual characteristic and comorbidities.