Resource utilization and costs following hospitalization of patients with chronic heart failure in the US

Abstract

Background:

Despite advances in its management and the identification of preventable risk factors, heart failure (HF) is a growing health problem in the US. The objective of this study was to describe treatment patterns, medical resource utilization and costs following hospitalization for chronic HF for patients stratified by age.

Methods:

Patients with at least one hospitalization with chronic HF were identified in a US commercial insurance claims database from 2004–2008. Patients were followed from the 1st day of chronic HF hospitalization (index hospitalization) until disenrollment or end of data availability. Inpatient, outpatient and prescription drug utilization rates were calculated per person per month (PPPM). Costs included payments made by insurers and, where available, patient out-of-pocket payments and sick-leave costs were also calculated. Utilization rates and costs were stratified by patient age.

Results:

There were 7814 patients included in the study. Index hospitalization was the most resource intensive and expensive ($31,023 age <65, $12,426 age ≥65). The rate of outpatient visits was the highest within 3 months following index hospitalization (3.6/PPPM age <65, 4.1/PPPM age ≥65). For the older age group, rate of re-hospitalizations was highest (0.06/PPPM) within 3–6 months following index hospitalization, while the younger group had its highest rate (0.08/PPPM) during the first 3 months following index hospitalization. Prescription dispensing rates were similar between age groups; average reimbursement PPPM for cardiovascular drugs did not exceed $92 (age <65) and $221 (age ≥65), which represents less than 3% of hospitalization costs for both groups.

Conclusions:

Treating chronic HF patients is resource intensive. The greatest burden occurs within 6 months after index hospitalization for both age groups; patients continue to be burdened after hospitalization by high inpatient and outpatient visit rates. Outpatient cardiovascular drug costs account for a small proportion of total healthcare costs.

Keywords::

• Heart failure
• Medical resource utilization
• Follow-up studies
• Cost

Introduction

In 2010, the direct and indirect costs of HF treatment in the US were estimated to be $39.2 billion1. Hospital discharges for HF in the US increased by ∼25% over the past decade1, with HF being the most frequently reported reason for hospitalization among adults aged 65 years and older. Major advances have been made in the spectrum of therapies for heart disease, including the introduction of new drugs, single-tablet combination therapies and new devices, all of which lead to increased physician visits2.

Despite advances in its management and the identification of preventable risk factors, HF remains a major health problem in the US, with high incidence (10 in 1000 people over the age of 65), prevalence (5.8 million adult Americans)3, and mortality (30–40% within the first year following diagnosis)3. While some new interventions may be cost-effective4,5, HF continues to have a marked impact on medical resource utilization and costs, with high rates of re-hospitalization and caregiver needs also contributing to the economic burden of HF6. HF will continue to be a substantial financial and public health concern among the aging population due to the increasing prevalence of diabetes and obesity, two key risk factors for developing HF7,8. Chronic HF, specifically, is a progressive syndrome leading to poor quality-of-life9–11 and high mortality. It is associated with a poorer quality-of-life compared with other chronic diseases, including diabetes, arthritis and high blood pressure12,13.

A better understanding of medical resource utilization following hospitalization for chronic HF (including diagnostics, procedures and hospitalizations) and how this utilization may vary during the course of HF disease progression will provide insight into the most resource intensive components of care for chronic HF patients. Examining resource use and costs over time by age groups is also informative since prior research has shown a greater relative increase in hospitalization for patients less than 65 years old compared to those of 65 years and older over the past three decades14. Characterizing the landscape of care can indicate how enhanced treatment of patients with chronic HF may mitigate future medical resource utilization. The objectives of this study were therefore to describe current healthcare resource utilization among patients with chronic HF in two age groups, under 65 years old and 65 years old and older, to identify key direct cost drivers of managing patients with this disease following their first hospitalization for HF and to describe the timing of these expenses along the patient disease pathway.

Patients and methods

Data source

This retrospective study evaluated claims data from the MarketScan® 2004–2008 Commercial Claims and Encounters and Medicare Supplemental and Coordination of Benefits databases from Thomson Reuters. These databases represent the health services claims of ∼61 million employees, dependents, and retirees in the US, with primary or Medicare supplemental coverage through privately insured fee-for-service, point-of-service or capitated health plans (<1%) and with full Medicare coverage plans (not Medigap). All enrollment records and inpatient, outpatient, ancillary and drug claims were collected. Data are individual-level, de-identified and compliant with the Health Insurance Portability and Accountability Act (HIPAA).

Study population

Patients who met the following criteria were included in this study: at least one claim for an inpatient hospitalization with a diagnosis of chronic HF (International Classification of Disease [ICD-9] codes 428.22, 428.32 or 428.42) as the primary or one of the 14 secondary diagnoses available on the discharge summary claim during their first HF hospitalization (index hospitalization), at least 6 months of continuous baseline insurance enrollment prior to the index hospitalization and no gaps in enrollment subsequent to the index hospitalization. Patients were observed during their index hospitalization until either the end of insurance eligibility or data availability, whichever occurred first. The analysis was restricted to patients with both medical and pharmacy benefits to ensure that complete healthcare utilization data were obtained.

Outcomes

Baseline characteristics

Insurance enrollment information and data from the index hospitalization were used to determine patients’ age, gender, year of hospitalization and insurance type. Claims data from the 6-month period prior to the index hospitalization were used to identify baseline comorbidities. Claims during the baseline period were examined to identify the presence of ICD-9 codes indicating specific cardiovascular or other co-morbid conditions. In addition, based on the presence of ICD-9 codes for claims during the 6-month baseline period, the Charlson comorbidity index (CCI)15 with Deyo modifications16 was calculated based on all medical claims (i.e., both inpatient and outpatient claims were used). Utilization and costs were stratified by patient age (under 65 and 65 and over) at index hospitalization.

Healthcare resource utilization

The number of patients receiving cardiac diagnostics and procedures at the index hospitalization, number of subsequent re-hospitalizations (for both all causes and separately for HF) and number of outpatient visits following discharge from index hospitalization (for both all causes and separately for HF) were determined. HF re-hospitalizations were defined as hospitalizations with at least one 428.xx ICD-9 code (heart failure) in the first 15 diagnoses of the discharge claim as done previously25. Outpatient HF-related visits were defined as any outpatient visit with at least one 428.xx ICD-9 code on the visit claim. Visits to the emergency department resulting in a hospitalization were not counted as separate outpatient visits. All-cause inpatient and outpatient visits during the baseline period were also calculated. Cardiac diagnostics and procedures were identified by claims for Current Procedural Terminology (CPT) codes, Healthcare Common Procedure Coding System (HCPCS) codes and ICD-9 procedure codes (Appendix A). In addition, the frequency of cardiac diagnostics and procedures and lengths of stay (LOS) were calculated for each HF-related re-hospitalization.

Outpatient pharmacy dispensings

Outpatient prescription dispensing claims were analyzed to determine whether a patient received prescription medication. National Drug Codes (NDCs) were used to identify dispensings of cardiovascular disease drugs of interest, listed in Table 3. Inpatient drug utilization and costs were not recorded in the database.

Table 1.  Baseline Characteristics of Chronic HF Patients.

	Chronic HF Patients N = 7814
Age, years (mean ± SD)	73.17 ± 12.98
Age (n,%)
<18 years	2	<0.1%
18–54 years	699	9.0%
55–64 years	1372	17.6%
65–74 years	1555	19.9%
≥75 years	4186	53.6%
Gender (n,%)
Male	4355	55.7%
Female	3459	44.3%
Insurance type (n,%)1
Comprehensive	3182	40.7%
PPO	2836	36.3%
HMO	1283	16.4%
POS	321	4.1%
CDHP	40	0.5%
EPO	10	0.1%
POS with capitation	7	0.1%
Unknown	135	1.7%
Calendar year of initial chronic HF hospitalization (n,%)
2004	741	9.5%
2005	1622	20.8%
2006	1138	14.6%
2007	1569	20.1%
2008	2744	35.1%
Charlson comorbidity index (mean ± SD)2	1.78 ± 1.76
Comorbid conditions (n,%)2
Diabetes mellitus	2046	26.2%
Chronic obstructive pulmonary disease	1441	18.4%
Peripheral vascular disease	635	8.1%
Renal disease	612	7.8%
Connective tissue disease	161	2.1%
Dementia	145	1.9%
Malignant solid tumor	91	1.2%
Malignant lymphoma	72	0.9%
Liver disease	60	0.8%
Ulcer disease	39	0.5%
Hemiplegia	38	0.5%
Leukemia	38	0.5%
AIDS	6	0.1%
Cardiovascular comorbidities (n,%)3
Hypertension	2879	36.8%
Atrial fibrillation	1417	18.1%
Hyperlipidemia	865	11.1%
Cerebrovascular disease	718	9.2%

1. EPO = Exclusive Provider Organization, HMO = Health Maintenance Organization,

POS = Point of Service, PPO = Preferred Provider Organization, CDHP = Consumer-Driven Health Plan

2. Based on a patient's comorbidities, the Charlson comorbidity index predicts the one-year mortality for the patient. The following ICD-9 codes were used for identifying comorbidities: AIDS (042), chronic obstructive pulmonary disease (491, 492, 493.2, 496), connective tissue disease (710, 714), dementia (290, 331.0, 331.1, 331.2), diabetes mellitus (250), hemiplegia (342, 344), leukemia (201.4, 205, 206, 207, 208), malignant lymphoma (200, 201, 202, 203.0, 273.0, 273.3), peripheral vascular disease (440, 441, 442, 437, 437.1, 447.1, 785.4), ulcer disease (531, 532, 533, 534), liver disease (570, 571, 572, 573, 456.0, 456.1, 456.2), renal disease (585, 586, V56, V420, V451), malignant solid tumor (140–171, 174–195, 199, V10)

3. The following ICD-9 codes were used for identifying cardiovascular comorbidities: hypertension (401–405), cerebrovascular disease (430–438), atrial fibrillation (427.31), hyperlipidemia (272.2, 272.4)

Cost

Patients younger than 65 years

Actual amounts paid by insurance companies to providers were recorded in the database. Patients’ out-of-pocket expenses were calculated as the sum of the deductibles, coinsurances and copayments paid by patients directly to providers. Insurers’ and patients’ payments for inpatient visits, outpatient visits and pharmacy dispensings in the outpatient setting were calculated.

The total number of sick-leave days was calculated for each patient based on the number of inpatient and outpatient visits. It was assumed that one inpatient day was equivalent to the loss of a full work day (8 h of labor) and an outpatient visit was equivalent to the loss of a half work day (4 h of labor). Sick-leave cost was calculated for those younger than 65 years old using an average hourly rate of $21.87 17.

Patients 65 years and older

To avoid under-estimation of costs due to incomplete Medicare reimbursement data captured in this private insurer database, all costs for patients of 65 years and older were estimated using Medicare reimbursement rates published annually by The Centers for Medicare and Medicaid (CMS).

Hospitalization costs were estimated based on Medicare’s Inpatient Prospective Payment System (IPPS). Each hospitalization discharge was assigned a relative weight18 based on its Diagnosis Related Group (DRG) code (or MS-DRG code if the hospitalization was in 2007 or later). To obtain the cost of a hospitalization, the national adjusted operating standardized labor, non-labor and capital base-payment hospitalization rate as published annually by the Federal Register19 was multiplied by the relative weight of the DRG or MS-DRG code. No adjustments for geographic location, hospital teaching status or proportion of low-income patients were made. Physician fees for inpatient visits were identified using HCPCS codes (99234–99236, 99221–99223, 99231–99233, 99238, 99239, 99251–99255, 99261–99263). Costs associated with each HCPCS code were based on Medicare’s Physician Fee Schedule20 for the year in which the hospitalization occurred.

Outpatient costs were estimated based on Medicare’s Outpatient Prospective Payment System (OPPS). Physicians’ costs of office visits and procedures were calculated using Medicare’s Physician Fee Schedule for the year of visit20. For procedures provided in a facility (as opposed to a physician’s office), costs were calculated based on the Ambulatory Payment Classification (APC) mean annual reimbursement21 and added to physician fees.

Costs of outpatient prescription drugs were based on the average wholesale price (AWP)22. Costs were adjusted to reflect the actual amount of drugs supplied. Sick-leave cost was assumed to be zero and therefore not calculated for this age group.

Statistical analyses

Analyses were conducted for the entire population of patients meeting the inclusion criteria. For patient baseline characteristics, the mean and standard deviation were calculated for continuous variables and the frequency and percentages were calculated for categorical variables. For cardiac diagnostics and procedures and outpatient prescription dispensings, the frequency and percentages of patients receiving these services were calculated. The mean and standard deviation were calculated for length of hospital stay and cost measures by hospitalization event for up to four HF-related re-hospitalizations. The rate of inpatient and outpatient visits per person per month (PPPM) was calculated by dividing the total number of visits by the total person-months of follow-up. The average outpatient pharmacy dispensing cost PPPM for patients incurring such costs was calculated as the total cost (for insurers and for patients less than 65 years and as estimated by the AWP for patients 65 years and older) across all patients divided by the total person-months of follow-up, adjusted to 2009 US dollars. Outpatient prescription dispensings and costs and the rate of inpatient and outpatient visits were calculated for each 3–6 month time interval following the index hospitalization for up to 24 months. All analyses were conducted using SAS software version 9.2 (SAS Institute, Cary, NC).

Results

Baseline characteristics

There were 7814 patients with a claim with a diagnosis code for chronic HF at initial hospitalization who met all eligibility criteria; 2073 were less than 65 years old and 5741 were at least 65 years old. Of the 7814 total patients, 5107 had their first chronic HF diagnosis claim during hospitalization, 432 patients had a chronic HF diagnosis code during an outpatient visit prior to their hospitalization for chronic HF and the remaining 2275 had a non-chronic HF diagnosis code during an outpatient visit prior to their first hospitalization for chronic HF. The 7814 patients contributed 11,302.2 person-years of follow-up.

Table 1 displays the demographic and clinical characteristics of the entire study population. There were 4355 males (55.7%) and 3459 females (44.3%). Patients had a comprehensive type of health insurance (40.7%), followed by a Preferred Provider Organization (PPO; 36.3%) and Health Maintenance Organization (HMO; 16.4%).

Table 2.  Cardiac Diagnostics and Procedures by HF Hospitalization.

	Index hospitalization		First re-hospitalization		Second re-hospitalization		Third re-hospitalization		Fourth re-hospitalization
	Age < 65 N = 2073	Age ≥65 N = 5741	Age <65 N = 514	Age ≥65 N = 1469	Age <65 N = 210	Age ≥65 N = 512	Age <65 N = 93	Age ≥65 N = 200	Age <65 N = 47	Age ≥65 N = 86
Length of stay, days (mean ± SD)	6.8 ± 11.3	6.7 ± 7.2	6.8 ± 7.1	6.6 ± 6.8	7.3 ± 6.4	7.1 ± 7.3	9.7 ± 9.7	7.6 ± 8.7	10.1 ± 11.0	7.2 ± 6.5
Days since previous hospitalization for HF (mean ± SD)	N/A	N/A	209.9 ± 266.1	283.8 ± 285.2	169.0 ± 196.7	238.4 ± 244.2	135.1 ± 205.2	172.3 ± 186.7	141.7 ± 171.8	118.1 ± 129.1
Diagnostics (%)1
Any diagnostic procedure below	91.4%	87.7%	86.0%	81.4%	85.2%	75.6%	81.7%	74.0%	78.7%	76.7%
Chest x-ray	72.8%	71.6%	68.5%	67.3%	68.6%	65.2%	75.3%	60.0%	66.0%	64.0%
Electrocardiogram	52.9%	55.0%	50.2%	49.1%	52.4%	47.3%	57.0%	49.0%	48.9%	48.8%
Echocardiogram	50.6%	49.4%	37.0%	34.3%	33.3%	31.1%	29.0%	32.5%	38.3%	27.9%
Cardiac catheterization	23.8%	11.5%	15.4%	6.0%	17.1%	5.3%	16.1%	4.5%	6.4%	2.3%
Coronary catheterization (angiogram)	20.3%	9.9%	10.1%	4.8%	10.5%	3.5%	10.8%	4.0%	2.1%	1.2%
Nuclear myocardial perfusion scan	7.2%	6.1%	3.5%	2.6%	2.9%	2.7%	2.2%	3.0%	2.1%	1.2%
Multiple gated acquisition scanning of the heart (MUGA)	7.2%	5.9%	3.7%	2.7%	2.4%	2.7%	3.2%	3.5%	2.1%	1.2%
Stress test	7.0%	5.6%	2.5%	2.7%	2.4%	2.7%	1.1%	3.5%	2.1%	2.3%
Blood test for natriuretic peptide	6.0%	0.8%	5.6%	0.7%	4.3%	0.2%	3.2%	1.0%	0.0%	0.0%
Intravascular ultrasound	0.6%	0.5%	0.2%	0.3%	1.0%	0.0%	1.1%	0.0%	0.0%	0.0%
Magnetic resonance imaging (MRI)	0.5%	<0.1%	0.4%	<0.1%	0.0%	0.2%	0.0%	0.0%	0.0%	0.0%
Positron emission tomography (PET) scan	0.2%	<0.1%	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%
Computerized tomography (CT)	0.0%	<0.1%	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%
Cardiac surgeries and procedures (%)1
Any surgery or procedure below	17.8%	12.0%	12.3%	5.4%	11.4%	4.9%	7.5%	3.0%	10.6%	0.0%
Biventricular pacemaker or cardiac resynchronization therapy (CRT)	7.5%	6.7%	2.9%	2.2%	1.9%	1.8%	2.2%	1.5%	4.3%	0.0%
Stent procedure	5.0%	3.1%	5.6%	2.0%	4.8%	1.8%	1.1%	1.0%	2.1%	0.0%
Coronary artery bypass graft (CABG) surgery	3.5%	1.5%	1.9%	0.7%	1.9%	1.2%	1.1%	0.5%	4.3%	0.0%
Valve surgery	1.7%	0.9%	1.6%	0.3%	1.9%	0.6%	2.2%	0.5%	2.1%	0.0%
Implantable cardioverter- defibrillators (ICDs)	0.5%	0.3%	0.6%	0.5%	0.0%	0.2%	1.1%	0.0%	2.1%	0.0%
Angioplasty (PTCA), laser angioplasty, atherectomy	0.4%	0.3%	0.4%	0.2%	1.0%	0.2%	0.0%	0.0%	0.0%	0.0%
Heart transplantation	0.2%	0.0%	0.0%	0.0%	1.9%	0.0%	1.1%	0.0%	0.0%	0.0%
Ventricular assist device (LVAD)	0.1%	0.0%	0.0%	<0.1%	0.5%	0.0%	0.0%	0.0%	2.1%	0.0%

1. The ICD-9, CPT and HCPCS codes for identifying these diagnostics and cardiac surgeries and procedures are specified in Appendix A.

The mean CCI among the study population was 1.78 (SD = 1.76). The most prevalent comorbidities were diabetes mellitus (26.2%), chronic obstructive pulmonary disease (COPD; 18.4%), peripheral vascular disease (8.1%) and renal disease (7.8%). The majority (69.3%) of patients had some cardiovascular comorbidity in the 6-month period prior to their initial chronic HF hospitalization, with the leading co-morbid disease being hypertension (36.8%), followed by atrial fibrillation (18.1%) and hyperlipidemia (11.1%).

Healthcare utilization—hospitalization

Heart failure-related hospitalization

Within the study population and entire observation period, 10,945 hospitalizations for HF were observed: 2937 were among those under 65 years old and 8008 were among the older age group. Figures 1a and b show the rate of HF re-hospitalizations (i.e., excluding index hospitalization) for each age group. For the younger age group, the rate of HF re-hospitalizations peaked at 0.08 visits per person-month during the first 3 months following index hospitalization and 0.06 visits per person-month for the older age group during the 3–6 months following index hospitalization.

Figure 1.  Hospitalizations over time, by age group.

Table 2 shows treatment and care provided during each hospitalization for HF for the index hospitalization and four subsequent re-hospitalizations (the number of patients experiencing five or more hospitalizations was too low to produce meaningful statistics). For both age groups, the greatest number of diagnostic procedures occurred during the first hospitalization; nearly 90% of patients in both age groups underwent at least one of the 14 listed diagnostic procedures. The most frequent diagnostic procedure for both age groups was chest x-ray followed by electrocardiogram and echocardiogram. During index hospitalization, 17.8% of patients in the younger age group underwent at least one of the 10 cardiac procedures or surgeries displayed in Table 2, while only 12.0% of older patients did so. The most common procedure for both age groups was biventricular pacemaker or cardiac resynchronization therapy (7.5% and 6.7% for younger and older age groups, respectively) followed by stent implantation (5% and 3.1%, respectively) and coronary artery bypass graft (CABG) surgery (3.5% and 1.5%, respectively). The proportion of patients undergoing a diagnostic or medical procedure decreased with each subsequent re-hospitalization for both age groups with the exception of the fourth re-hospitalization for the younger age group. For each re-hospitalization, the younger age group had more cardiac procedures or surgeries on average than the older age group.

Table 3.  Outpatient Prescription Patterns for Cardiovascular Medications.

	Baseline period		0–6 months following index hospitalization		6–12 months following index hospitalization		12–18 months following index hospitalization		18–24 months following index hospitalization
	Age <65 N = 2073	Age ≥65 N = 5741	Age <65 N = 2049	Age ≥65 N = 5683	Age <65 N = 1340	Age ≥65 N = 3487	Age <65 N = 890	Age ≥65 N = 2250	Age <65 N = 628	Age ≥65 N = 1550
Medications (n,%)
Beta blockers	53.4%	57.4%	69.2%	59.9%	63.3%	60.7%	62.2%	61.9%	61.0%	61.7%
ACE inhibitors or ARBs	50.8%	52.0%	60.7%	50.8%	55.6%	50.8%	53.8%	51.3%	56.4%	52.3%
- ACE inhibitors	39.7%	37.3%	49.1%	38.4%	43.7%	36.2%	41.6%	36.0%	43.9%	36.6%
- ARBs	13.5%	16.9%	15.8%	14.8%	14.3%	16.1%	14.3%	16.8%	14.2%	17.0%
Loop diuretics	42.0%	52.8%	54.8%	56.9%	47.1%	54.6%	44.9%	54.7%	46.3%	54.8%
Statins	37.6%	43.7%	43.0%	39.7%	41.6%	42.1%	40.8%	41.5%	43.8%	43.1%
Calcium channel blockers	19.2%	24.9%	18.0%	20.1%	16.3%	19.1%	15.8%	18.5%	16.1%	17.9%
Digoxin	14.0%	19.4%	18.9%	20.1%	18.1%	19.6%	19.0%	19.6%	18.8%	19.5%
Thiazide diuretics	11.2%	12.6%	11.2%	9.1%	8.7%	8.7%	8.3%	9.0%	10.5%	8.7%
Potassium sparing diuretics	12.7%	10.2%	19.7%	13.5%	18.0%	11.6%	18.1%	11.7%	18.5%	12.0%
- Aldosterone antagonists	14.1%	10.9%	21.8%	14.4%	20.1%	13.0%	20.3%	13.3%	21.3%	13.9%
- Spironolactone	12.8%	10.3%	20.0%	13.6%	18.3%	11.7%	18.3%	11.9%	18.6%	12.4%
- Eplerenone	1.6%	0.6%	2.0%	0.9%	2.0%	1.3%	2.0%	1.5%	2.7%	1.7%
- Epithelial sodium channel blockers	0.2%	0.3%	<0.1%	0.1%	0.2%	0.1%	0.1%	<0.1%	0.2%	<0.1%
- Amiloride	0.2%	0.3%	<0.1%	0.1%	0.2%	0.1%	0.1%	<0.1%	0.2%	<0.1%
- Triamterene	2.8%	4.1%	1.4%	1.7%	1.2%	1.8%	1.0%	1.4%	0.8%	1.6%
Alpha blockers	6.0%	8.7%	5.8%	7.0%	4.9%	6.2%	5.2%	5.6%	4.1%	5.1%
Vasodilators	3.4%	3.5%	4.9%	5.3%	4.3%	4.2%	4.6%	4.8%	4.6%	4.6%
Aliskiren	0.1%	0.3%	0.1%	0.2%	0.1%	0.2%	0.1%	0.2%	0.2%	0.2%
Medication combinations (n,%)
Aldosterone antagonists and ACE inhibitors	9.2%	4.9%	4.4%	3.1%	3.7%	2.7%	3.9%	2.6%	4.1%	3.4%
ACE inhibitors and ARBs	2.3%	2.2%	4.2%	2.4%	2.5%	1.5%	2.0%	1.4%	1.8%	1.2%
Pharmacy costs per patient per month (mean ± sd)
Cardiovascular drugs
Reimbursement from insurers	$68 ± $69	$162 ± $121	$92 ± $106	$221 ± $464	$80 ± $76	$189 ± $155	$82 ± $77	$185 ± $176	$82 ± $77	$182 ± $129
Patient out-of-pocket	$20 ± $19	N/A	$32 ± $37	N/A	$24 ± $23	N/A	$23 ± $22	N/A	$22 ± $23	N/A
Other drugs
Reimbursement from insurers	$295 ± $472	$447 ± $687	$436 ± $832	$668 ± $2655	$325 ± $534	$492 ± $833	$335 ± $546	$483 ± $808	$386 ± $612	$484 ± $573
Patient out-of-pocket	$50 ± $51	N/A	$82 ± $103	N/A	$56 ± $62	N/A	$55 ± $59	N/A	$59 ± $60	N/A
Total-pharmacy	$398 ± $508	$593 ± $707	$584 ± $869	$848 ± $2715	$452 ± $571	$660 ± $857	$460 ± $587	$649 ± $843	$513 ± $653	$647 ± $597

ACE = Angiotensin-converting enzyme; ARB = Angiotensin II (A-II) receptor blockers

  ICD-9, CPT and HCPCS Codes for Identifying Diagnostics and Cardiac Surgeries and Procedures.

Diagnostic test/Procedure	ICD-9 procedure codes	CPT	HCPCS
Angioplasty (PTCA)/Laser angioplasty/Atherectomy	00.66, 36.03, 36.09	92982, 35481, 35491, 92995	C1885
Biventricular pacemaker or cardiac resynchonization therapy (CRT)	00.50, 00.53, 37.80	33211, 33213, 33214, 33224–33226	C1785, C2619
Blood test (specifically for natriuretic peptide)		83880
Cardiac catheterization	37.21, 37.22, 37.23	93501–93533
Chest x-ray	87.44, 87.49	71010, 71015, 71020, 71021, 71022, 71023, 71030, 71034, 71035, 71090, 0175T
Computerized tomography (CT)	87.41	0144T, 0145T, 0146T, 0147T, 0148T, 0149T, 0150T, 0151T
Coronary artery bypass graft (CABG) surgery	36.10, 36.11, 36.12, 36.13, 36.14, 36.15, 36.16, 36.17, 36.19	33503, 33504, 33505, 33510–33516, 33533–33536, 33548, 33517–33523, 33530	G8158, G8159, G8161, G8162, S2205, S2206, S2207, S2208, S2209
Coronary catheterization		93555
Echocardiogram	37.28, 88.72	93303–93350	G0366, G0367, G0368, G0403, G0404, G0405
Electrocardiogram	89.51, 89.52, 89.54	93000, 93005, 93010, 93040, 93041, 93042
Heart transplantation	37.51	33945
Implantable cardioverter- defibrillators (ICDs)	00.51, 00.54	33216, 33217
Intravascular ultrasound	0.24	37250, 37251, 75945, 92978, 92979
Magnetic resonance imaging (MRI)	88.92	71550, 71551, 71552, 75552–75564
Multiple gated acquisition scanning of the heart (MUGA)	92.05	78472, 78473, 78478, 78483
Nuclear myocardial perfusion scans	92.05	78460, 78461, 78464, 78465, 78491, 78492
Positron emission tomography (PET) scan		78491, 78492, 78459
Stent procedure	00.55, 00.63, 00.64, 00.65, 36.06, 36.07	92980	G0290, C1874, C1875, C1876, C1877
Stress test	89.41, 89.43, 89.44	93015, 93016, 93017, 93018	C8928, C8930
Valve surgery	35.0, 35.1, 35.2	33400–33478
Ventricular assist device	37.65, 37.66	33975, 33976, 33979, 0048T

Average LOS increased with subsequent re-hospitalizations for the younger age group and reached an average of 10.1 days for the fourth re-hospitalization (Table 2). The average LOS for the older age group varied from 6.7–7.6 days for the index through to the fourth re-hospitalization. The number of days between subsequent re-hospitalizations generally decreased with each additional hospitalization for both age groups. Compared with the older age group, the number of days since the prior hospitalization was lower for the younger age group through the third re-hospitalization.

All-cause hospitalization

Figures 1a and b show the rate of hospitalizations for any diagnosis during the 6-month baseline period and through 24 months following the index hospitalization. The rate of hospitalizations was 0.05 and 0.03 visits per person-month during baseline for the younger and older age groups, respectively. The rate of hospitalizations was highest during the 3-month time period and 3–6 month time period following index hospitalization for the younger and older age groups, respectively. During follow-up through 24 months after the index hospitalization, the rate of re-hospitalizations did not fall below that of the baseline period for either age group. HF-related re-hospitalization accounted for the majority of all-cause re-hospitalizations for both groups during follow-up.

Healthcare utilization—outpatient visits

HF-related outpatient visits

The rate of HF-related outpatient visits (Figures 1a and b) increased from 0.14 visits per person-month at baseline to 0.43 during the 3 months following index hospitalization discharge for the younger age group and increased from 0.17 to 0.65 visits per person-month during this time for the older group. HF-related outpatient visits did not fall below baseline through the end of follow-up.

All-cause outpatient visits

The rate of outpatient visits for any reason through 24 months following index hospitalization discharge is displayed in Figures 2a and b. There were 2.2 and 2.4 outpatient visits per person-month during baseline for the younger and older age groups, respectively. This increased to 3.6 and 4.1 during the 3-month time period following index hospitalization discharge for the younger and older age groups, respectively. The rate for both age groups then decreased but remained elevated relative to the baseline period for at least a year.

Figure 2.  Outpatient visits over time, by age group.

Outpatient pharmacy dispensing

Dispensing of selected drug groups are presented in Table 3. The dispensing of these drugs appears similar during the baseline period for both of the age groups. For the younger age group there is a noticeable increase from baseline to 0–6 months follow-up in the dispensing of loop diuretics (from 53.4 to 69.2%), angiotensin-converting enzyme (ACE) inhibitors and angiotensin II (A-II) receptor blockers (ARBs) (from 50.8 to 60.7%) and potassium sparing diuretics (from 12.8 to 20.0%). There was less variability in the dispensing of drugs between baseline to 0–6 months follow-up for the older age group. The proportion of patients receiving the displayed sub-groups of cardiovascular disease drugs changed little during the 24 months of follow-up after index hospitalization discharge for both age groups.

Cost of chronic HF

The average direct and indirect (sick-leave) medical costs per HF hospitalization for patients less than 65 years, excluding patients with non-reported costs, and the average Medicare payment for patients 65 years and older are displayed in Figure 3. For patients less than 65 years, the average total direct cost for the initial hospitalization for chronic HF was $31,998 (SD = $50,336). Direct medical costs for the first through fourth HF-related re-hospitalizations for this age group were $22,048 (SD = $29,006), $23,946 (SD = $52,952), $24,839 (SD = $38,776) and $24,518 (SD = $64,993), respectively. The amount paid by insurers accounted for more than 90% of total direct costs, with patient out-of-pocket cost accounting for the remainder. Total indirect (sick-leave) cost was estimated to be less than $1800 per hospitalization, accounting for 3.6% of total costs during the index hospitalization and increasing to 6.7% of total costs during the fourth re-hospitalization. For patients of 65 years and older, the average Medicare payment for the initial hospitalization for chronic HF was $12,427 (SD = $11,303). Medicare payments for the first through fourth HF-related re-hospitalizations for this age group were $10,040 (SD = $8771), $9606 (SD = $6735), $9132 (SD = $6818) and $9298 (SD = $6033), respectively.

Figure 3.  Cost per hospitalization.

Mean cost of HF-related outpatient visits was highest during the baseline period for both age groups ($752 and $502 for the younger and older age groups, respectively) and decreased through the 24 months of follow-up. During the follow-up period, mean cost per visit ranged from $668 in the 3 months following index hospitalization to $224 during the 18–24 month period among the younger age group and from $351 to $278 among the older age group.

The average PPPM cost of cardiovascular disease drugs varied from $88–$124 (Table 3) for the younger age group. Patients’ out-of-pocket expenses for cardiovascular drugs accounted for ∼25% of the total cost of cardiovascular drugs. For the older age group, the average PPPM cost of cardiovascular disease drugs varied from $162–$221. Throughout the study period, cardiovascular drugs for both age groups accounted for less than 30% of total pharmacy costs.

Discussion

The retrospective claims analysis presented here summarizes treatment, utilization and costs among a cohort of 7814 patients hospitalized with chronic HF in the US between the years 2004–2008. This study is the first to look at the continuum of healthcare utilization among patients with chronic HF stratified by insurance coverage (commercially insured vs Medicare insured) and how these patients were diagnosed and managed in inpatient and outpatient settings following the initial hospitalization. This study also measured costs incurred by the patient, including out-of-pocket and sick-leave costs.

The results show that chronic HF is a resource-intensive condition and that the greatest utilization occurs in the first 6 months following initial hospitalization for chronic HF. The high resource utilization rate during this time is mainly reflected in high inpatient and outpatient visit rates during this time. Of note, both the all-cause inpatient and outpatient visit rates remained elevated for up to 18 months following the initial hospitalization for chronic HF compared with the 6-month baseline time period. The vast majority of all-cause inpatient visits following discharge from the initial chronic HF hospitalization were HF-related.

The rates and timing of re-hospitalizations differ between the two age groups. As can be seen in Figure 1, among the younger age group, re-hospitalization rate is the highest (0.08 per person-month) during the first 3 months following index hospitalization, while the rate is the highest (0.06 per person-month) during the 3–6 month time period for the older age group. In addition, the overall re-hospitalization trend differs between the two groups; rates among the younger group exhibit a steady decline of up to 50% in re-hospitalization rate in the 3–24 months follow-up and the re-hospitalization rate decreases only slightly among the older age group during the 6–24 months follow-up. Data from our study presented in Table 2 show that the number of days between re-hospitalizations is substantially lower in the younger age group compared to the older age group, resulting in the re-hospitalization rates peaking at different times. The younger age group also tends to undergo more surgeries and medical procedures during hospitalizations compared to the older age group.

In both age groups, mean cost of index hospitalization was higher than the mean cost of any of the four subsequent HF-related re-hospitalizations. While many factors contribute to the overall hospitalization cost, surgeries and medical procedures are an expensive cost item in any hospitalization. As can be observed in Table 2, the proportion of patients undergoing surgeries or medical procedures decreased with each additional HF re-hospitalization, hence providing one plausible explanation for the more expensive index hospitalization. In addition, in both age groups the index hospitalization for HF and subsequent HF-related re-hospitalizations were long and costly relative to inpatient visits for other conditions within the US23. Data on hospital use for all patients from the 2008 Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample (NIS) indicate that the average LOS per inpatient visit (all cause) was 4.6 days and the average cost was $9170. In our study, the average LOS for a HF-related hospitalization was at least 1.5-times longer and as much as 2.5-times as costly (among patients <65 years).

Differences in hospitalization costs between older and younger age groups in our study are a direct result of different reimbursement methods. Patients younger than 65 years are typically covered by commercial health insurance which reimburses hospitals on a fee-for service basis. A cost tariff exists for each diagnostic and procedure provided during a hospital stay. On the other hand, patients of 65 years and older are covered by Medicare, which reimburses providers based on an inpatient prospective, fixed cost basis (IPPS). Evidence found in the literature suggests that Medicare’s payments are generally less than those of commercial health insurers14.

Our findings of large differences in hospitalization costs between the two age groups are similar to findings from other studies. Whellan et al.24 estimated HF hospitalization costs among the elderly using 2001–2004 Medicare’s inpatient claims and reported a mean cost of $12,719 (SD = $21,087) per hospitalization. Wang et al.25 analyzed commercial claims data to calculate the cost of HF-related hospitalization among patients of 18–64 years and reported a mean cost of $25,325 (SD = $31,406). To identify our study population we used the primary and 14 secondary diagnoses on the patient’s hospitalization claim. Because chronic HF is more likely to be the underlying factor contributing to the hospitalization (as opposed to acute HF which by itself may be the primary reason for the hospitalization), it is often reported as a secondary diagnosis rather than as a primary diagnosis24. Including both primary and the 14 secondary ICD-9 diagnoses available from the discharge claim increased the probability that our analysis captured all patients hospitalized with chronic HF. As a result, as shown by Wang et al.25, our cost estimates may be higher than if we had used only the primary ICD-9 to identify HF-related hospitalizations.

HF-related outpatient visits accounted for a small fraction of all-cause outpatient visits and the rates and time trends were similar between the two groups. As can be expected, the older age group had higher all-cause visit rates. The cost of HF-related outpatient visits was the highest during baseline in both groups and declined after index hospitalization. This may be due to decreased utilization of outpatient services and an increase in services obtained in the inpatient setting. The relatively high costs observed in the 12–18 months time period in the younger age group and in the 18–24 months time period in the older age group are associated with a few outliers that have a strong influence on the costs given the small sample size for these time periods.

Despite available evidence from clinical trials reporting the beneficial clinical effect of several pharmacotherapy agents26, including ACE inhibitors, ARBs and beta-blockers, our study depicted prescription dispensing rates of only 50–60% for these drugs. Other studies that measure utilization have found the following utilization rates: Stafford et al.27 assessed the use of ACE inhibitors among patients with congestive HF using two independent national outpatient visit samples and documented 28–38% utilization rates in the early 2000s; Cleland et al.28 conducted an international survey among primary care physicians and reported 60% utilization of ACE or ARBs and 34% utilization of beta blockers during the year 2000. Medication dispensing rates captured in any claims data represent prescriptions actually filled by patients that resulted in a claim. Dispensing rates are affected by physician prescribing rates, whether patients fill their prescriptions and whether there is a claim for the prescription filled. It must also be noted that the database for our study does not capture inpatient medication dispensing, so actual total medication use may be higher than calculated from outpatient dispensing data. It should be further noted that the dispensing of a drug does not imply that it was taken as indicated.

While patients’ adherence to treatment was not measured in the current study, this can also affect dispensing rates. Fonarow et al.29 collected nationally representative data from medical records of patients diagnosed with HF or prior myocardial infarction and left ventricular dysfunction treated in outpatient cardiology practice settings. They found that ACE inhibitors or ARBs were prescribed in 80% of eligible patients (with variation across clinics ranging from 5.9–96.3%) and beta blockers were prescribed in 86% of eligible patients (with variation across clinics from 8.6–100%). Several studies30,31 have documented only partial treatment compliance among HF patients who are prescribed these medications.

There were low average PPPM costs for cardiovascular disease drugs that varied from $88–$124 among patients younger than 65 and from $182–$221 among those aged 65 years and older during the follow-up period. The reported PPPM cost is higher for the older age group and this may be due to the use of AWP to estimate costs. While the AWP is a widely used benchmark for drug costs it does not reflect discounts or rebates provided to third party vendors (e.g. pharmacy benefit management organizations). Despite this potential over-estimation of pharmacy costs among the elderly, PPPM accounted for a small fraction of overall treatment cost for the patient, regardless of age, accounting for less than 3% of the average cost of a HF-related hospitalization in both age groups.

This study has several limitations which should be considered. First, as with all claims database analyses, ICD-9 codes were used to identify diagnoses; these codes may not reflect confirmed clinical diagnoses and lack information to assess severity of illness (e.g., NYHA classification was not in the database). In addition, categorizing patients by type of HF (e.g., acute, chronic, reduced ejection fraction) is problematic as patients may have multiple HF diagnoses over time (e.g., diagnoses for both preserved and reduced ejection fraction). Moreover, medical services obtained outside of a patient’s plan are not captured in a claims database, so resource utilization and costs in this study may be under-estimated. Pharmacy dispensing may not represent all drugs prescribed or used by the patient (e.g., samples received at a doctor’s office or hospital). In addition, we did not distinguish between medication eligible and non-eligible patients as we did not have access to clinical data and therefore did not estimate medication dispensing according to eligibility. Therefore, the presented rates of prescription use may appear lower compared to other studies that were based on clinical data and determined eligibility. Due to the fact that costs provided in the dataset for patients of 65 years and older may be incomplete, we estimated medical costs for this population based on a published payment schedule, so some deviation from actual reimbursed amounts is plausible. In addition, we were unable to estimate the impact of chronic HF on long-term disability since this information was not captured in the data. Calculated sick-leave costs reflect the costs associated with visits to the doctor or hospital and do not capture all work-loss days associated with HF. Thus, the costs reported under-estimate the true burden of illness. Finally, mortality information is not captured in this claims database.

Conclusion

Despite the limitations mentioned above, this study provides a good estimate of healthcare resource utilization and costs associated with hospitalization for chronic HF and subsequent management of chronic HF. The results are especially insightful in highlighting that the greatest burden occurs within the first 6 months after the first HF hospitalization and that the burden continues after this hospitalization in terms of increased inpatient and outpatient visits. Furthermore, these data represent a geographically diverse US population and the findings are therefore generalizable to the US at large. Given the growing population of persons of 65 years and older in the US and the increasing incidence and prevalence of chronic HF, this analysis provides important data on the resource utilization demands and costs that will be incurred by growing numbers of chronic HF patients. Given that resource utilization peaks and remains high after the first hospitalization for chronic HF, new medications and interventions that have the potential to prevent hospitalization and re-hospitalization for chronic HF may have a significant impact on resource utilization.

Transparency

Declaration of funding

Funding for this research was provided by Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA.

Declaration of financial/other relationships

CK, AE-L, MHL and MSD are employees of Analysis Group, Inc., which has received research grants from Novartis Pharmaceuticals Corporation. JM was an employee of Analysis Group, Inc. when this analysis was conducted. SHO is an employee of Novartis Pharmaceuticals Corporation.