The cost to managed care of managing pulmonary hypertension

Abstract

Objective:

To estimate direct medical costs and resource use for commercially-insured patients within two pulmonary hypertension sub-groups: pulmonary arterial hypertension (PAH) and chronic thromboembolic pulmonary hypertension (CTEPH).

Research design and methods:

Using a retrospective cohort design, subjects (≥18 years) with ICD-9 code 416.0 (PAH or CTEPH) or 416.8 (CTEPH) were identified during 2004–2009 within the MarketScan database. The date of the first observed claim was the index date. Each PAH and CTEPH patient was matched to one-to-five controls without PAH and CTEPH on age, gender, region, and payer type. Patients and controls were continuously enrolled for at least 12 months pre- and 12 months post-index date. Per-patient-per-month costs and resource use were compared using Wilcoxon rank-sum test.

Results:

PAH patients (1647) and controls (6352) were identified (mean age 63 years, 73% female). Total monthly costs before PAH diagnosis were: PAH patients $2064 vs controls $1094. After PAH diagnosis, PAH patients had significantly higher monthly costs and resource use vs controls: Total costs $4021 vs $1533, outpatient visits 1.1 vs 0.8, inpatient visits 0.7 vs 0.2, prescriptions 3.6 vs 2.7, all p-values <0.05. One hundred and forty-six CTEPH patients and 558 controls were identified (mean age 64 years, 54.8% female). Total monthly costs in the period before CTEPH diagnosis were higher for CTEPH patients ($3895) than controls ($1177). After CTEPH diagnosis, CTEPH patients had significantly higher monthly costs and resource use vs controls: Total costs $6198 vs $1579, Outpatient visits 1.2 vs 0.8, inpatient visits 2 vs 0.2, prescriptions 4.2 vs 2.8, all p-values <0.05.

Key limitations:

Identification of PAH is complicated, as there exists no precise ICD-9-CM code for the condition. CTEPH diagnosis was based upon claims data and was not verified clinically.

Conclusions:

CTEPH and PAH patients incurred higher costs and used more resources than controls in the baseline and follow-up periods.

• Pulmonary arterial hypertension
• PAH
• Chronic thromboembolic pulmonary hypertension
• CTEPH
• Cost
• Resource use

Introduction

Pulmonary hypertension (PH) is a complex disorder characterized by elevated pulmonary arterial pressure and increased pulmonary vascular resistance. This condition, if untreated, ultimately leads to right ventricular failure and death. The current clinical classification of PH distinguishes five sub-groups of the disease1. This study examined direct medical costs and resource use for patients with commercial insurance within two of these sub-groups: pulmonary arterial hypertension (PAH) and chronic thromboembolic pulmonary hypertension (CTEPH).

Specific pulmonary vasodilator therapies, including phosphodiesterase inhibitors (PDE5 inhibitors), endothelin receptor antagonists (ERAs), and prostacyclin analogs, are approved by the Federal Drug Administration (FDA) for the treatment of PAH, and current pharmaceutical development programs are being conducted in both PAH and CTEPH sub-groups. Therefore, understanding the economic impact in both of these groups may be of interest to payers.

Pulmonary arterial hypertension

PAH is a progressive condition resulting from restricted flow through the pulmonary arterial circulation and characterized by elevated pulmonary arterial pressures leading to right ventricular (RV) failure1–3. The development of PAH appears to be attributable to multiple pathogenic pathways, including both genetic and molecular factor levels related to the smooth muscle cells, endothelial cells, and adventitia. The basis for current medical treatment is the imbalance in the vasoconstrictor/vasodilator milieu, although consensus is building that PAH also involves an imbalance of proliferation and apoptosis.

PAH typically has poor prognosis with annual mortality rate of ∼15%3. Treatment options for patients with PAH are limited. Currently available treatment options include oral anticoagulants, calcium channel blockers, ERAs, prostacyclin analogs, and PDE-5 inhibitors4–6. Although these treatments have had positive effects on functional capacity and quality-of-life of patients, the disease progression while on therapy is still common and survival remains poor7. However, over the last few years the Food and Drug Administration has approved several new treatment options for those with PAH, including ERAs, prostacyclins, and PDE-5 inhibitors4,5,8.

The economic costs of PAH are not well studied or understood. The studies of current treatments suggest that cost of treatment is very high9. A 2003 study by Michelakis et al.10 examined the annual per patient cost in Canada of three different categoris of PAH treatments: a PDE-5 inhibitor (sildenafil), an ERA (bosentan), and a prostacyclin (epoprostenol). The authors found that the annual per patient treatment costs with sildenafil was $6000 CAN, treatment with bosentan was 6–7-times higher, and treatment with epoprostenol was ∼15-times that of treatment with sildenafil. A 2007 study found that the annual per patient cost of treatment was €5000–10,000 for sildenafil, €45,000 for bosentan, €70,000–180,000 for iloprost (a prostacyclin), and €230,000 for epoprostenol11. However, both of these studies employed cost estimates based on the assumption of continuous treatment at the recommended dosage level with medication purchased at the official list price, and therefore are not truly reflective of real-world costs and usage patterns.

A more recent study used a retrospective analysis of the data held in a medical claims database to evaluate the economic burden of PAH to third-party payers, focusing on privately-insured patients (under age 65) in the US, which may have biased the analysis and limited its generalizibility12. In any case, this analysis estimated healthcare costs of $4236 per patient-month among PAH patients treated with oral medications indicated for PAH.

Chronic thromboembolic pulmonary hypertension

Chronic thromboembolic pulmonary hypertension (CTEPH) is an uncommon complication of a pulmonary thromboembolic (PE) event, but a leading cause of severe pulmonary hypertension1,3,13. It’s true prevalence is still unclear, likely in part due to lack of awareness of the disease among medical caregivers1,3,13. CTEPH is manifested by intraluminal thrombus organization and fibrous stenosis, completely damaging pulmonary arteries. The result is increased pulmonary vascular resistance, which in turn leads to pulmonary hypertension and progressive right heart failure14–19.

Previous research has found that up to 4% of patients with acute pulmonary embolism (PE) develop CTEPH within 2 years, and this diagnosis is associated with significant morbidity and mortality15–17. Complicating our current understanding of the disease, a considerable proportion of CTEPH patients have no history of acute PE prior to receiving a CTEPH diagnosis, and the vast majority of patients with CTEPH have no identifiable abnormality of the coagulation or fibrinolytic pathways20,21.

The only potentially curative treatment for CTEPH is a pulmonary endarterectomy (PEA)13. However, PEA remains a relatively rare procedure with ∼130 procedures per year, suggesting that most patients are not eligible for this surgery22. There are currently no FDA-approved drugs indicated to treat CTEPH, although certain medications indicated for use in PAH may be efficacious among some selected inoperable CTEPH patients23–25. In addition, anti-coagulation is recommended in CTEPH in order to prevent recurrent thromboembolic events13.

At present, there is even less information on the burden of illness of CTEPH in large commercially-insured US populations than for PAH. Even in the absence of previous research into the matter, one would expect these costs to be high, given (1) the debilitating nature of the disease, (2) the similarities in pathogenesis to PAH, and (3) the similarities in medical therapy to PAH.

The aim of the current study is to estimate direct medical costs and resource use of patients with PAH and CTEPH in a commercial administrative claims database. Diagnoses of PAH and CTEPH have proven to be clinically challenging. However, the reclassification of PAH during the Third World Symposium on Pulmonary Hypertension (Venice classification, 2003), and then subsequently at the Fourth World Symposium (Dana Point classification, 2009) have enabled clinicians to apply consistent criteria to assess patients with PAH and CTEPH.

Our analysis was built on Peacock et al.’s26 approach, which involves an ICD-9-CM-based approach, consistent with the most recent clinical classification. However, in that study it was not possible to verify the diagnostic accuracy of cases because of a lack of availability of clinical data. Further, the Peacock et al. study represents a survey of inexpert diagnoses made using a number of investigatory techniques. We identified PAH and CTEPH in an administrative claims database through the use of ICD-9-CM claims codes and most recent clinical classification recommendations. The primary advantage of this approach is that it reflects real-world practice and presents costs and outcomes of patients being treated for PAH and CTEPH, although it does not allow for the use of clinical measures or specific diagnostic results. Once appropriate subjects were identified, we were able to detail the demographics, comorbidity profile, medical resource utilization, and healthcare costs of patients with PAH and CTEPH as compared to demographically-matched controls. We present here a retrospective database approach to evaluate the economic burden of PAH and CTEPH to third-party payers, focusing on privately-insured patients in the US.

Methods

Data collection

Data were derived from a MarketScan® Commercial Claims and Encounters database, a fully integrated, de-identified, individual-level national healthcare claims database comprising privately and publicly insured individuals covered by a variety of fee-for-service, preferred provider organizations, and capitated health plans. These data included health insurance claims across the continuum of care (e.g., inpatient, outpatient, outpatient pharmacy, carve-out behavioral healthcare) as well as enrollment data from large employers and health plans across the US who provides private healthcare coverage for more than 50 million employees, their spouses, and dependents. Complete payment information for all types of healthcare resource consumption is captured, both what the health plan and patient paid and rigorous validation methods are employed to ensure that claims and enrollment data are complete, accurate, and reliable. The database is Health Insurance Portability and Accountability Act (HIPAA) compliant and is the basis of nearly 450 articles published in leading journals since the first article published in 199027.

Data were collected over the period ranging from January 2004 to December 2009. The baseline period comprised a minimum of 12 months prior to the index date (pre-index). Similarly, the follow-up period was a minimum of 12 months after the index date (post-index), up to December 31, 2009.

Design

A retrospective cohort difference-in-difference design was used. This design is used to estimate differences between groups (cases and controls) and between time periods using pre–post-index date time periods.

Subjects

The date of the first observed claim with an ICD-9 code of 416.0 (PAH) or 416.8 (CTEPH) was designated as the index date. All patients had to be 18 years of age or older at the index date and were required to be continuously enrolled in a health plan for at least 12 months pre- and at least 12 months post-index date. Each PAH and CTEPH patient was matched to one-to-five controls without PAH and CTEPH on age (year of birth), gender, region, and payer type, using a frequency matching technique.

PAH patient-specific criteria

PAH patients were eligible9,10 if they had: ≥2 claims with a primary PAH (ICD-9 code 416.0) separated by at least 1 day but within 12 months of each other and ≥1 claim for right heart catheterization (RHC) or echocardiogram occurring before the second claim for any PAH diagnosis (ICD-9 code 416.0), but not more than 12 months prior to the first claim for any PAH diagnosis.

PAH patients were excluded if they had any of the codes appearing in categories 2–5 of the Venice Classification scheme during 12 months pre- and 1 month post-index date1,9,28,29. These exclusion codes encompass pulmonary hypertension with left heart disease, pulmonary hypertension associated with lung diseases and/or hypoxemia, CTEPH, and miscellaneous (i.e., Sarcoidosis, histiocytosis X, lymphangiomatosis, compression of pulmonary vessels). Finally, those with a prescription for a drug having PAH-specific indication within 12 months before the index date were also excluded.

CTEPH patient-specific criteria

CTEPH was defined based upon the presence of pulmonary embolism and pulmonary hypertension9,28,29 as indicated by: ≥2 claims for PH (ICD-9: 416.0, 416.8) with the first claim occurring within at least 6 months after a claim for pulmonary embolism (PE) and ≥1 claim for right heart catheterization (RHC) or echocardiogram before the second PH claim. PE was defined as a thromboembolic obstruction of either proximal or distal pulmonary arteries or non-thrombotic pulmonary embolism (all ICD-9 code 415.1, pulmonary embolism) or a medical event classified as history of venous thrombosis and embolism (including pulmonary embolism) ICD-9 code V12.51, or vena cava filter (CPT-4: 36010, 37620, 75825, 75940; HCPCS: C1880; ICD-9: 38.7). Finally, those with a prescription for a drug having CTEPH-specific indication within 12 months before the index date were also excluded.

Study measures

For subjects in each of the groups, i.e., PAH, CTEPH, and their respective matched controls, we identified age, gender, region of residence based on the four census regions, the type of health plan they were enrolled in, and co-morbid conditions, along with the calculation of the Charlson comorbidity index. All of these variables were measured at the baseline. Further, resources used were calculated as the number of inpatient visits, outpatient visits, and prescription drug claims for each subject. Total costs, and costs for each of the components including inpatient, outpatient, emergency room (ER), and prescription drugs were measured. Both the resources used and costs were measured over the pre-index and follow-up periods. Finally, treatment characteristics and types of diagnostic procedures were indentified over the pre-index and follow-up periods as well. Treatment characteristics were categorized as pharmacotherapy or other therapeutic treatments.

Statistical analysis

Descriptive statistics (means and percentages) were calculated for demographic characteristics and clinical data. The Chi-square test for categorical variables and Wilcoxon rank-sum test for continuous variables were used to compare across groups. Per-patient-per-month (PPPM) utilization and costs for inpatient, emergency department, and outpatient visits costs were compared between the two groups (PAH or CTEPH vs matched controls) at baseline (at least 12 month period prior to index date) and follow-up (at least 12 month period post index date) using the paired t-test. Wilcoxon rank sum test was used to compare differences in direct medical costs and resource use between cases and controls in the periods prior to and post index date.

Results

PAH cost and resource use

After applying all the inclusion and exclusion criteria we found 2097 patients with PAH. Of these, 1647 PAH patients were matched with 6352 controls (Table 1). The mean age of both groups was 63 years and 73% were female. The mean number of follow-up months for which data was available was 20.8 for the PAH patients and 23.9 for the control group. The most commonly observed comorbidities were heart failure (14% of the PAH group, 4% of the control group), cardiac dysrhythmia (10%, 5%), respiratory and chest symptoms (9%, 5%), and diabetes mellitus (7%, 4%).

Table 1.  Baseline demographic characteristics of study sample for PAH patients and matched controls.*.

	PAH (n = 1647)	Controls (n = 6352)	p-value*
Age, years, mean	62.57 (15.84)	63.65 (15.39)	0.1
Age group, n (%)
18–30 years	51 (3%)	147 (2%)	0.1
31–45 years	177 (11%)	625 (10%)
46–65 years	711 (43%)	2649 (42%)
66–85 years	612 (37%)	2540 (40%)
86 years or older	96 (6%)	391 (6%)
Female, n (%)	1203 (73%)	4642 (73%)	0.98
Region, n (%)
North East	169 (10%)	683 (11%)	0.62
North Central	510 (31%)	1989 (31%)
South	673 (41%)	2476 (39%)
West	288 (17%)	1182 (19%)
Unknown	7 (<1%)	22 (<1%)
Plan type, n (%)
Comprehensive	575 (35%)	2188 (34%)	0.98
HMO	237 (14%)	966 (15%)
POS	125 (8%)	480 (8%)
PPO	629 (38%)	2426 (38%)
Others	81 (5%)	292 (5%)
Comorbid conditions, n (%)
Heart Failure	229 (14%)	243 (4%)	<0.001
Cardiac dysrhythmias	161 (10%)	293 (5%)	<0.001
Symptoms involving respiratory system and other chest symptoms	146 (9%)	295 (5%)	<0.001
Diabetes mellitus	115 (7%)	280 (4%)	<0.001
Charlson Comorbid Index Score at baseline 1 year period, mean (SD)	2.55 (2.40)	1.56 (2.17)	<0.001

*Based on Chi-square test for categorical variables and Wilcoxon rank-sum test for continuous variables.

HMO, health maintenance organization; POS, point of service; PPO, preferred provider organization.

The mean per-patient per-month medical costs prior to PAH diagnosis were $2064 for the PAH group and $1094 for the control group (Table 2). Subsequent to the PAH diagnosis, the PAH group experienced a significantly larger increase in monthly medical costs and resource use. Total costs for the PAH group increased by $1957, which included an increase of $671 in outpatient medical costs, $1028 in inpatient medical costs, and $258 in prescription costs. Following a similar pattern, PAH group resource utilization increased by 0.32, 0.53, and 0.64 for outpatient visits, inpatient visits, and number of prescriptions, respectively. The control group experienced an increase in medical costs (total costs rose by $439) over the mean 23.9 months that this group was followed. However, the increase in medical costs and resource utilization by the PAH group was numerically larger and reached statistical significance (p ≤ 0.05 for all values) relative to the increases seen in the control population.

Table 2.  Mean per-patient-per-month costs and resource use for PAH patients and matched controls.*.

Economic outcomes	PAH (n = 1647)		Controls (n = 6352)		p-value**
	Pre-index period	Follow-up period	Pre-index period	Follow-up period
Utilization (n)
Inpatient visits	0.17	0.70	0.10	0.19	<0.001
Outpatient visits	0.81	1.13	0.73	0.82	<0.001
ER visits	0.14	0.24	0.11	0.15	<0.001
Number of Prescriptions	3.00	3.64	2.52	2.72	<0.001
Costs ($)
Total cost	$2064	$4021	$1094	$1533	<0.001
Inpatient cost	$713	$1741	$298	$519	<0.001
Outpatient cost	$961	$1632	$561	$758	<0.001
ER cost	$28	$55	$18	$29	0.012
Prescription cost	$390	$648	$226	$256	<0.001

*Costs are reported in 2009 US dollars.

**Based on Wilcoxon rank-sum test for estimating the significance in difference-in-difference.

Following the PAH diagnosis, members of the PAH group used more cardiovascular-related prescriptions and had more diagnostic procedures than the members of the control group (Table 3). Elevated rate of cardiovascular-related prescriptions observed included prescriptions for calcium channel blockers (30% in the PAH group vs 20% in the control group), diuretics (50% vs 27%), oral anticoagulants (25% vs 9%), digoxin (11% vs 4%), and ERAs (8% vs <1%). Diagnostic procedures performed at an elevated rate on the PAH group included echocardiography (24% of the PAH group vs 4% of the control group), electrocardiography (38% vs 22%), computerized tomography (11% vs 8%), and right heart catheterization (RHC) (1.5% vs <0.1%).

Table 3.  Treatment and diagnostic characteristics of PAH patients vs matched controls.

	Baseline period		Follow-up period
	PAH (n = 1647)	Controls (n = 6352)	PAH (n = 1647)	Controls (n = 6352)
Pharmacotherapy
Oral anti-coagulants	340 (20.6%)	519 (8.2%)	419 (25.4%)	572 (9.0%)
Diuretics	785 (47.7%)	1709 (26.9%)	817 (49.6%)	1712 (27.0%)
Calcium channel blockers	488 (29.6%)	1261 (19.9%)	489 (29.7%)	1274 (20.1%)
Digoxin	188 (11.4%)	245 (3.9%)	186 (11.3%)	244 (3.8%)
Endothelin receptor antagonists (ERAs)	84 (5.1%)	0 (0.0%)	138 (8.4%)	1 (0.0%)
Prostacyclin analogues	26 (1.6%)	0 (0.0%)	55 (3.3)	0
Phosphodiesterase (PDE-5) inhibitors	74 (4.5%)	159 (2.5%)	175 (10.6%)	154 (2.4%)
Other therapeutic treatments
Vena cava procedure	5 (0.3%)	2 (0.0%)	10 (0.6%)	5 (0.1%)
stress echocardigraphy	7 (0.4%)	30 (0.5%)	2 (0.1%)	22 (0.3%)
Ventilation perfusion	0	0	1 (0.1%)	0 (0.0%)
Pulmonary endarterectomy	1 (0.1%)	0 (0.0%)	0	0
Oxygen therapy	0 (0.0%)	3 (0.0%)	3 (0.2%)	8 (0.1%)
Diagnostic procedures
Echocardiography	268 (16.3%)	303 (4.8%)	396 (24.0%)	259 (4.1%)
Other radiography	3 (0.2%)	17 (0.3%)	2 (0.1%)	14 (0.2%)
Electrocardiography	638 (38.7%)	1440 (22.7%)	619 (37.6%)	1389 (21.9%)
Computerized tomography (CT)	150 (9.1%)	457 (7.2%)	183 (11.1%)	494 (7.8%)
Angiography	6 (0.4%)	30 (0.5%)	18 (1.1%)	20 (0.3%)
Right heart catheterization	6 (0.4%)	0 (0.0%)	24 (1.5%)	2 (0.0%)
Other exercise testing	92 (5.6%)	223 (3.5%)	100 (6.1%)	159 (2.5%)
Magnetic resonance imaging (MRI)	0 (0.0%)	4 (0.1%)	1 (0.1%)	6 (0.1%)

CTEPH costs and resource use

After applying all the inclusion and exclusion criteria we found 161 patients with CTEPH. Of these, 146 patients were matched with 558 controls (Table 4). The mean age of both groups was 64 years and 55% were female. Medical data was available for the pre-index period for a mean of 33.9 months in the CTEPH group and 28.0 months in the control group. The mean number of follow-up months for which data was available was 19.8 for the CTEPH patients and 22.8 for the control group. CTEPH patients were more frequently diagnosed with acute pulmonary heart disease (40% vs 0.6% in the control group), symptoms involving the respiratory system and other chest symptoms (14% vs 4%), and other forms of lung disease (13% vs 2%). Overall, the CTEPH group had a higher comorbidity burden compared to matched controls, as indicated by the Charlson Comorbidity Index Scores (4.02 vs 1.73).

Table 4.  Baseline demographic characteristics of study sample for CTEPH patients and matched controls.*.

	CTEPH (n = 146)	Controls (n = 558)	p-value*
Age, years, mean	63.8	64.3	0.709
Age group, n (%)			0.9904
18–30 years	2 (1%)	8 (1%)
31–45 years	11 (8%)	39 (7%)
46–65 years	72 (49%)	266 (48%)
66–85 years	53 (36%)	210 (38%)
86 years or older	8 (6%)	35 (6%)
Females, %	81 (55.5%)	305 (54.7%)	0.859
US region, n (%)			0.965
North east	12 (8%)	50 (9%)
North central	61 (42%)	241 (43%)
South	54 (37%)	193 (35%)
West	19 (13%)	74 (13%)
Plan type, n (%)
Comprehensive	57 (39%)	210 (38%)	0.921
HMO	13 (9%)	62 (11%)
POS	8 (5%)	31 (6%)
PPO	60 (41%)	226 (40%)
Others	8 (5%)	29 (5%)
Comorbid conditions, n (%)
Acute pulmonary heart disease	58 (40%)	3 (0.6%)	<0.001
Heart failure	20 (14%)	22 (4%)	<0.001
Symptoms involving respiratory syst/chest	20 (14%)	33 (6%)	0.002
Other lung diseases	19 (13%)	9 (2%)	<0.001
Pneumonia organism nos	17 (12%)	0 (0%)	<0.001
Charlson Comorbid Index (CCI) Score for the baseline 1 year period	4.02	1.73	<0.001

*Based on Chi-square test for categorical variables and Wilcoxon rank-sum test for continuous variables.

HMO, health maintenance organization; POS, point of service; PPO, preferred provider organization.

The mean per-patient per-month medical costs prior to the index point were $3895 for the CTEPH group and $1177 for the control group (Table 5). Subsequent to the index point, the CTEPH group experienced a significantly larger increase in monthly medical costs and resource use. Total costs for the CTEPH group increased by $2303, which included an increase of $1007 in outpatient medical costs, $1035 in inpatient medical costs, and $261 in prescription costs. Following a similar pattern, CTEPH group resource utilization increased by 0.24, 1.41, and 0.78 for outpatient visits, inpatient visits, and number of prescriptions, respectively. The control group experienced an increase in medical costs (total costs rose by $402) over the mean 22.8 months that this group was followed. However, the increase in medical costs and resource utilization by the CTEPH group was numerically larger and reached statistical significance (p ≤ 0.05 for all values) relative to the increases seen in the control population.

Table 5.  Mean per-patient-per-month costs and resource use for CTEPH patients and matched controls.*.

Economic outcomes	CTEPH		Controls		p-value**
	Pre-index period	Follow-up period	Pre-index period	Follow-up period
Utilization
Inpatient visits	0.56	1.97	0.10	0.17	<0.001
Outpatient visits	0.93	1.17	0.76	0.83	0.0375
ER visits	0.25	0.44	0.09	0.12	0.4559
Number of prescriptions	3.40	4.18	2.64	2.82	0.0035
Costs
Total cost	$3895	$6198	$1177	$1579	<0.001
Inpatient cost	$2347	$3382	$258	$507	<0.001
Outpatient cost	$1186	$2193	$658	$780	<0.001
ER cost	$46	$89	$20	$30	0.9111
Prescription cost	$362	$623	$260	$292	0.0337

*Costs are reported in 2009 US dollars.

**Based on Wilcoxon rank-sum test for estimating the significance in difference-in-difference.

Table 6 shows the main treatment and diagnostic characteristics of CTEPH patients and matched controls for the baseline and the follow-up periods. In both the pre-index and follow-up periods, CTEPH patients had higher prescription and procedure use relative to controls. Notably, prescription usage was elevated for oral anticoagulants (57% in the CTEPH group vs 12% in the control group), diuretics (47% vs 29%), digoxin (9% vs 3%), and ERAs (3% vs 0%). Diagnostic procedures performed at an elevated rate on the CTEPH group included echocardiography (23% of the CTEPH group vs 5% of the control group), electrocardiography (43% vs 25%), and computerized tomography (17% vs 8%). In addition, a vena cava procedure was performed on 34% of the CTEPH group in the pre-index period, and on 25% of the CTEPH group in the post-index period. None of the control group had this procedure performed at any point during the study.

Table 6.  Treatment and diagnostic characteristics of CTEPH patients vs matched controls.

	Baseline period		Follow-up period
	CTEPH (n = 146)	Controls (n = 558)	CTEPH (n = 146)	Controls (n = 558)
Prescription medication use
At least one prescription, n (%)
Oral anti-coagulants	66 (45.2%)	61 (10.9%)	83 (56.8%)	65 (11.6%)
Diuretics	75 (51.4%)	147 (26.3%)	68 (46.6%)	163 (29.2%)
Calcium channel blockers	36 (24.7%)	107 (19.2%)	31 (21.2%)	102 (18.3%)
Digoxin	9 (6.2%)	23 (4.1%)	13 (8.9%)	19 (3.4%)
Endothelin receptor antagonists (ERAs)	2 (1.4%)	0 (0.0%)	5 (3.4%)	0 (0.0%)
Diagnostic procedures
At least one procedure, n (%)
Echocardiography	28 (19.2%)	35 (6.3%)	33 (22.6%)	27 (4.8%)
Other radiography	0 (0.0%)	1 (0.2%)	0 (0.0%)	1 (0.2%)
Electrocardiography	66 (45.2%)	134 (24.0%)	63 (43.2%)	139 (24.9%)
Computerized tomography (CT)	36 (24.7%)	43 (7.7%)	25 (17.1%)	42 (7.5%)
Angiography	1 (0.7%)	7 (1.3%)	2 (1.4%)	5 (0.9%)
Other exercise testing	11 (7.5%)	22 (3.9%)	4 (2.7%)	19 (3.4%)
Stress echocardigraphy	2 (1.4%)	2 (0.4%)	0 (0.0%)	1 (0.2%)
Therapeutic procedures
At least one procedure, n (%)
Vena cava procedure	49 (33.6%)	0 (0.0%)	37 (25.3%)	0 (0.0%)
Pulmonary endarterectomy	0	0	1 (0.7%)	0 (0.0%)

Discussion

To our knowledge, this study is the first to assess the utilization of healthcare resources for patients with pulmonary hypertension using a large, commercial US database that includes patients both younger and older than 65 years of age. Additionally, the analysis of patients with CTEPH is a novel population for which scant data regarding healthcare expenditures exists. Not surprisingly, our data suggest significantly increased utilization of healthcare resources for both PAH and CTEPH patients compared to matched controls. Importantly, these data indicate an urgent need for further studies into the determinants of increased healthcare utilization and strategies to reduce the cost burden in these disease states.

PAH findings

The study demonstrates that PAH patients utilized considerably more healthcare resources and incurred significantly higher costs both before and after their PAH diagnosis as compared with matched controls. Our analysis indicates that the net increase in total costs of PAH patients was more than 4-times ($1957 vs $439) those of controls (Table 2). More narrowly, the net increase in outpatient, inpatient, and prescription costs were considerably higher for the PAH patients as compared with the controls. The control group also saw an increase in costs over the follow-up period (mean 24 months), however it was much more modest, and much of this increase is likely attributable to a combination of typical increases in medical spending with increasing age and healthcare cost inflation. In the light of substantially higher costs to treat PAH patients, new treatments that could result in reductions in outpatient visits or inpatient stays may offer important net economic benefits.

Somewhat surprisingly, utilization rates of current PAH-indicated therapies were low among PAH patients. Contributing factors to this finding could possibly include the NYHA functional class composition, which could not be measured in this sample, clinical prescribing practices, and the time frame of the study. Utilization of ERAs and prostacyclin analogs was comparatively higher among PAH patients than controls, but rates of PDE-5 use were similar, likely due to the fact that PDE-5s have other approved indications beyond PAH. Importantly, many insurance companies require a RHC demonstrating PAH prior to approving coverage for PAH therapy. Since only a few individuals had RHCs, insurance approval may not have been obtained for a large majority of patients in the PAH group.

CTEPH findings

Similar to our findings in PAH patients, this study demonstrates that CTEPH patients utilized considerably more healthcare resources and incurred significantly higher costs as compared with matched controls. For these patients, the increase in the cost of care was even more pronounced than in the PAH analysis, with a net increase in total costs of CTEPH nearly 6-times that observed in the controls, and a relative net increase in outpatient, inpatient, and prescription costs ranging from 4–8-times higher than observed in the control group. The substantially higher costs of treating CTEPH patients suggest that new treatments that reduce outpatient visits and inpatient stays may offer noteworthy net economic benefits.

This study has several limitations, perhaps the most important of which is accurately identifying subjects for inclusion. The identification of PAH is complicated, as there exists no precise ICD-9-CM code for the condition. However, the Venice and Dana Point Classifications provide a roadmap which can be implemented to identify PAH patients in administrative claims data, as demonstrated in Peacock et al.26. Additionally, the diagnosis of CTEPH was based upon claims data and was not verified clinically. Satisfaction of the CTEPH inclusion criteria of PH was based upon either RHC or echocardiography, thereby potentially misclassifying subjects as cases. Similarly, multiple diagnosis codes were used to define PE and might also have led to misclassification. Further, since no sub-type specific ICD-9 codes are available, PAH and CTEPH were not categorized into sub-types.

Since the study sample is drawn from a commercial administrative claims database, the results from our analysis may not be generalizable to a non-commercially insured population. Our analysis did not adjust for baseline comorbidities which may overstate the net costs attributable to PAH or CTEPH. However, we did match the cases to controls on key demographic variables that likely accounted for the significant differences across the cohorts, as the goal was to estimate the incremental burden of disease by teasing out PAH vs non-PAH and CTEPH vs non-CTEPH driven costs. To account for differences, we adopted the Wilcoxon rank sum analysis of this quasi-experimental design with matched controls because it is transparent and a more conservative statistical test. This design has been found to be a robust and appropriate approach30. Furthermore, the claims data do not capture relevant information for any treatments the patients received outside of the system. Also, in the CTEPH analysis, other medications specifically used for therapy were not assessed (e.g., PDE5I, inhaled prostacyclins, intravenous prostanoids), which may have under-estimated the true difference in cost burden between the CTEPH and control groups.

Because of the high mortality rate associated with PAH, severe PAH subjects may have been excluded, potentially under-estimating the costs for these disease groups. Further, the subjects who are diagnosed with a chronic condition are generally under more surveillance and, hence, incur higher costs relative to controls, which may have over-estimated the incremental costs in our study cohorts. It is also possible that at random more controls for less expensive cases and fewer controls for more expensive cases were selected, which may have led to an under-estimation of the costs in the control group. Given the strict eligibility criteria, some cases could only be matched to one control and some to as many as five. However, because of the orphan disease status of PAH/CTEPH, the intent was to include as many cases as possible. Therefore, we did not eliminate cases that had fewer than five controls. Finally, since administrative claims data are captured for billing and reimbursement purposes, our study contains limitations associated with any other claims data, including uncertain diagnostic coding validity, upward adjustment of claims, and lack of clinical information.

Conclusion

These results suggest that both CTEPH and PAH patients used more resources than controls in the baseline and follow-up periods. Total costs doubled for PAH patients after a PAH diagnosis and increased substantially for CTEPH patient after its diagnosis. The authors believe that a comparison of the different treatment modalities of PAH patients and CTEPH patients to assess the effectiveness of currently available therapies may be an important part of future research into understanding how to manage the cost of care of pulmonary hypertension. Further research is also needed to identify factors related to higher utilization rates.

Transparency

Declaration of funding

This study was funded by Bayer HealthCare Pharmaceuticals, the US-based pharmaceuticals business of Bayer HealthCare LLC, a subsidiary of Bayer AG.

Declaration of financial/other relationships

Dr Qayyim Said and Dr Bradley Martin have disclosed that they have a consulting relationship with Bayer HealthCare LLC, sponsor of this study. Dr Vijay N. Joish and Mr Charles Krelick have disclosed that they are employees of Bayer HealthCare LLC. Dr Stephen Mathai has disclosed that he has received consulting fees from Bayer and United Therapeutics and unrestricted research grants from Gilead and the NIH.

Acknowledgments

The authors would like to thank Bridget Banas for providing medical writing assistance, which was paid for by Bayer HealthCare LLC.