Abstract
Background:
Idiopathic pulmonary fibrosis is a life-threatening condition, and few data concerning the impact on healthcare utilization and associated costs are available. The objective of this study was to describe the burden of illness (comorbidity, healthcare resource utilization, and associated costs) in patients with idiopathic pulmonary fibrosis.
Methods:
Two cohorts (patients with idiopathic pulmonary fibrosis and matched controls) were retrospectively identified from US claims databases between January 1, 2001 and September 30, 2008. Cases with idiopathic pulmonary fibrosis were defined by age of 55 years or older and either two or more claims with a code for idiopathic fibrosing alveolitis (ICD-9 516.3), or one claim with ICD 516.3 and a subsequent claim with a code for post-inflammatory pulmonary fibrosis (ICD-9 515). The prevalence and incidence of pre-selected comorbidities, healthcare resource utilization (hospital, outpatient, drugs), and direct medical costs were assessed in each cohort.
Results:
A total of 9286 patients with idiopathic pulmonary fibrosis were identified. When compared with age- and gender-matched controls, these patients were at significantly increased risk for comorbidities including pulmonary hypertension and emphysema. The all-cause hospital admission rate (0.5 per person-year) and the all-cause outpatient visit rate (28.0 per person-year) were both ∼2-fold higher than in controls. Total direct costs for patients with idiopathic pulmonary fibrosis were $26,378 per person-year; the incremental costs over controls were $12,124 (2008 value).
Conclusions:
Patients with idiopathic pulmonary fibrosis experience increased comorbidity, healthcare resource utilization, and direct medical costs compared to controls.
Introduction
Idiopathic pulmonary fibrosis (IPF) is a highly morbid disease of aging, affecting an estimated 89,000 Americans and approximately one in 1000 American adults over the age of 65 yearsCitation1. Its incidence has been reported to be climbing as the population of older adults increases and as recognition of the condition growsCitation2. IPF remains a relentlessly progressive disease with no widely-accepted therapy in the US other than lung transplantation, and patients often experience severe shortness of breath and disabilityCitation3,Citation4. Median survival from time of diagnosis is ∼3 yearsCitation5,Citation6.
While healthcare utilization has been studied in more prevalent chronic pulmonary diseases like chronic obstructive pulmonary disease and asthmaCitation7, the overall burden of illness (i.e., comorbidity, healthcare resource utilization, and healthcare cost) associated with IPF remains unknown. A comprehensive understanding of the burden of illness in IPF would provide a more accurate assessment of the current unmet medical needs related to this disease, which could help in the prioritization of future research efforts. Given the aging population, the avoidable morbidity and costs associated with IPF could potentially be sizable.
In this study, claims databases were used to investigate the burden of illness associated with IPF by assessing the rates of comorbidity, the rate of healthcare resource utilization, and the associated direct medical costs.
Methods
Data source
This study was performed using two US claims databases: the Commercial Claims and Encounters Database (patients with employer-based health insurance) and the Medicare Supplemental and Coordination of Benefits Database (patients with Medicare supplemental insurance), both MarketScan Research Databases, Thomson Reuters (see supplement). (MarketScan is a registered trademark of Thomson Reuters.)
Study population
Patients were identified from medical claims between January 1, 2001 and September 30, 2008. Patients included in the IPF cohort were aged 55 years and older and had at least two claims on separate days associated with ICD-9 code 516.3 (idiopathic fibrosing alveolitis) or one ICD-9 516.3 code and a subsequent ICD-9 515 code (post-inflammatory pulmonary fibrosis), not restricted to the primary diagnosis. The index date was the date of the second qualifying ICD-9 code (either 516.3 or 515) for each IPF patient. Patients were excluded if they did not have continuous eligibility for 6 months before the first ICD-9 516.3, or had two or more claims for another type of interstitial lung disease after the initial ICD-9 516.3 claim (see eTable 1). Patients were classified as ‘newly diagnosed’ based on the absence of an ICD-9 516.3 or lung transplant code in the 12 months preceding the first 516.3 diagnosis date. The sub-set of ‘newly diagnosed’ cases was required to have at least 12 months of continuous eligibility before their first 516.3 code. The control cohort was constructed by randomly sampling patients from the general population of the same databases who had no claims with ICD-9 516.3 or ICD-9 515 during the study period. Controls were matched by gender, and age at index date of the IPF patients. The index date of the controls was set to correspond to the index date of the matched IPF patient (±3 months).Two secondary IPF cohorts (‘broad’ and ‘narrow’ definitions) were created to allow for sensitivity analysis (see supplement).
Comorbidities and death
Comorbidities were chosen for inclusion a priori based on their perceived relevance to the IPF population. At least one of the ICD-9 codes listed in eTable 2 was required to identify a comorbid condition as present. Inpatient deaths were based on hospital discharge-status disposition data. Outpatient deaths are not captured.
Resource utilization and costs
Healthcare resource use profiles and associated direct medical costs include all-cause claims (i.e., irrespective of associated diagnostic code) for hospitalizations (including procedures, tests, and drugs), outpatient care, and outpatient pharmacy services (additional information on the codes are detailed in eTable 3). The costs included in this analysis represent the covered direct medical costs. The total direct medical costs include all covered medical and pharmacy claims; costs are also presented separately by major category (medication claims, outpatient services, hospitalizations). Claims for selected drug therapies that are recommended or frequently prescribed in IPF were assessed. Cost estimates were reported in 2008 US dollars (see supplement).
Statistical analyses
Analyses were performed using SAS (v9.1). All results presented compare the IPF and control cohorts, unless otherwise specified. Baseline characteristics and prevalent comorbidities were determined from claims made during the 6-month period preceding the index date (pre-index period). Comorbidity prevalence estimates were compared between IPF patients and controls using relative risks and 95% confidence intervals. Incident comorbidities were determined from the claims made after the index date (post-index period) in patients who did not have a claim for the relevant comorbidity during the pre-index period. Incident comorbidity and inpatient death rates were calculated per person-years at risk and compared between IPF patients and controls using rate ratios and 95% confidence intervals. Proportions and rates were compared using Chi Squared tests.
Healthcare resource utilization profiles and direct medical costs (i.e., covered costs whether paid by patient or insurer) incurred in the post-index period were recorded. Because total follow-up differed across patients, resource utilization and direct medical costs were standardized to per person-year.
Results
Patient characteristics and selected healthcare resource utilization pre-index
A total of 9286 IPF cases were identified (). Sixty-eight per cent of the IPF cohort was classified as newly diagnosed. The median length of follow-up of the IPF cohort was 1.03 years (interquartile range 0.36–2.13 years). The mean age of the cases and controls at index date was 74 years, with most patients greater than 65 years old (79% cases and 81% controls). More than half were male (55%) in each cohort. Only 7% (512 of 7013) of the cases and 8.6% of the controls (609 of 7092) who were employees were categorized as currently working at the index date (the employment status of dependents was not known). The secondary ‘broad’ IPF cohort contained 14,339 patients and the secondary ‘narrow’ IPF cohort 3768 patients; their characteristics and other findings were similar to the primary IPF cohort (see eTable 4).
Table 1. Baseline demographics and selected healthcare utilization pre-index.a
Thirty-six per cent of IPF patients had a claim for corticosteroids during the 6-month pre-index period. This was substantially greater than controls (8.4%). Oxygen therapy also appeared more common in the IPF population (27%) compared to controls (2.4%). Claims for outpatient prescription of azathioprine or cyclophosphamide, pulmonary rehabilitation, and lung transplantation were rare. Thirty-five per cent of the IPF cohort had at least one hospitalization during the pre-index period compared to 10% of controls, and 14% of the IPF cohort had at least one hospitalization for IPF.
Prevalent and incident comorbidities
Most of the selected comorbidities were more prevalent in IPF patients compared to controls during the pre-index period (). Compared to controls, patients with IPF were at increased risk for having pulmonary hypertension (relative risk [RR] = 15.56), emphysema (RR = 7.11), pulmonary embolism (RR = 6.97), chronic bronchitis (RR = 5.19), and pulmonary infection (RR = 4.20).
Table 2. Prevalence of selected comorbidities.a
The occurrence of selected comorbidities during the post-index period (incident comorbidity) was also more common in IPF patients compared to controls (see eFigure 1 and eTable 5). The highest rates were again seen for pulmonary hypertension, emphysema, chronic bronchitis, pulmonary embolism, and pulmonary infection (rate ratio = 4.49, 3.06, 2.51, 2.24, and 2.21, respectively).
Mortality
Inpatient mortality for IPF patients was 52.6 deaths per 1000 person-years. This was significantly higher than in controls (14.8 per 1000 person-years, rate ratio = 3.64, 95% CI: 3.12–4.26).
Healthcare resource utilization and direct costs, post-index date
Healthcare resource utilization during the post-index period was higher in patients with IPF compared to controls (). The all-cause hospital admission rate (0.5 per person-year), days in hospital (3.1 per person-year), and the all-cause outpatient visit rate in the IPF cohort (28 per person-year) were each ∼2-fold higher than in controls. Of note, ICD-9 516.3 was not the discharge diagnosis recorded for most hospital admissions in the IPF cohort; in the first year post-index, only 5% of discharges had ICD-9 516.3 as the principal diagnosis assigned. Pulmonary function testing and CT scanning were performed in most patients with IPF during the observation period, while claims for the 6-minute walk test and arterial blood gas measurement were fairly rare (although more common than in controls). Forty-seven per cent of the IPF cohort had at least one claim for corticosteroids during the post-index period, compared to 17.9% of the controls. Azathioprine or cyclophosphamide outpatient prescription claims were much less frequent, recorded by 10% of the IPF cohort and 0.4% of the controls. Since drug claims do not have a diagnosis assigned, the reason these medications were prescribed is not known.
Table 3. Selected healthcare resource utilization rates (post-index per person-year).a
Average direct costs for IPF patients and controls are shown in . The annual direct cost post-index date was ∼2-times higher for IPF patients than controls, with an incremental annual cost estimated at $12,124. The incremental annual costs in the ‘broad’ and ‘narrow’ IPF secondary cohorts were $9493 and $15,423, respectively (see eFigures 2 and 3). Overall drug costs accounted for about a quarter of the costs, and any-cause inpatient care accounted for one-third of the total costs in each of the three IPF cohorts.
Figure 1. Direct medical costs per person-year (2008). Overall direct medical costs for any medication claims, inpatient services, and outpatient services are calculated for patients with IPF and age- and gender-matched controls.
Discussion
This study describes the comorbidity, incremental healthcare resource utilization, and additional healthcare costs associated with the diagnosis of IPF using a large and geographically diverse population of patients in the US with employer-based insurance. The results demonstrate increased rates of comorbidities such as pulmonary hypertension and emphysema, and show that resource utilization and costs are substantially higher in patients with IPF than in matched controls. Assuming an estimated 89,000 Americans are affected by IPF each yearCitation1, an incremental cost of $12,000 per person-year translates into an annual aggregate incremental cost of slightly over $1 billion dollars. The per-person annual costs may be lower than expected for an orphan disease, likely because there is no approved pharmacological therapy for IPF in the US, non-drug therapies such as pulmonary rehabilitation and lung transplantation are under-utilized, and drugs sometimes used to treat IPF, such as corticosteroids or azathioprine, are inexpensive.
Accurately identifying patients with IPF in large claims databases is important because miscoding of other forms of interstitial lung disease as IPF is possible. Misclassification could lead to over-estimation or under-estimation of IPF cases and burden of illness. IPF cases in this study were required to have an age of 55 years, based on recent findings that patients over age of 55 years with modest amounts of lung fibrosis almost universally have IPF as the underlying cause of their diseaseCitation8. Further, cases were required to have more than one relevant diagnostic code for lung fibrosis, to protect against a single miscoded event. Also, patients with two or more medical claims for another type of interstitial lung disease after the initial ICD-9 516.3 claim were excluded. Importantly, analyses using more loosely defined (‘broad’) and more rigorously defined (‘narrow’) IPF cohorts with matched controls demonstrated similar results to the primary analysis, increasing confidence in the main findings.
Pulmonary infection, coronary artery disease, diabetes, and heart failure were the most prevalent comorbidity codes in the IPF cohort, and all were significantly more common in IPF than in controls. Pulmonary embolism and lung cancer were also increased in patients with IPF. Several of these conditions have been associated with IPF in the literature, but never in a single study using large healthcare databasesCitation9–11. These findings highlight the importance of a comprehensive approach to the diagnosis and management of comorbidities in IPF.
Although less commonly coded, pulmonary hypertension and emphysema were markedly more prevalent in patients with IPF compared to controls. There is a clear association between these two conditions and IPF; whether these are true comorbidities or represent unique clinical phenotypes of IPF is unclearCitation12,Citation13. The frequency of claims for these two conditions was low (2.7% and 8%, respectively) compared to recent clinical studies that have actively looked for these conditions in IPF cohortsCitation14,Citation15. Although it is important to recognize that the estimated prevalence of comorbidities in this study may be affected by many factors (e.g., under-diagnosis and under-reporting), the relative risk of these comorbidities in IPF cases compared to controls should be unaffected, because under-diagnosis and under-reporting would affect both groups equally.
This study found increased healthcare resource utilization and direct costs in IPF patients compared to matched controls. IPF remains a disease without a widely-accepted therapy, and there are limited pharmacologic management options in the US, all of which involve drugs without a specific approved indication for IPF. Claims associated with corticosteroids and azathioprine were more common for patients with IPF than controls, although the reason these medications were prescribed is not available in these databases. IPF patients in this study experienced all-cause hospitalization rates as well as outpatient visit rates 2-fold higher than controls; in-hospital mortality was 3-times higher. Since only inpatient deaths were recorded in these databases, this study surely under-estimates mortality, as deaths occurring outside the hospital cannot be assessed.
There are limitations to conducting a study using healthcare claims databases. As mentioned previously, there is potential for miscoding of claims for patients with alternative diagnoses. Coding errors may be due to chance or may be influenced by reimbursement rates and other factors. In claims databases, there is also the potential that medical care is delivered but billed to a different insurance provider. Only claims completely billed to other insurers are missed in these databases; all claims covered at least in part by the employer are included in this analysis. If a patient switched out of the health plan provided by their employer (e.g., to change jobs or to switch to their spouse’s plan), subsequent healthcare claims would not be included. About half the patients in each cohort were lost to follow-up during the first year post-index date. As the medical history of each patient was obtained through the claims records from the duration of enrollment in the health plan, the date for a first diagnosis with IPF cannot be determined with certainty. To account for the difference in follow-up, each subject’s contribution was weighted by their duration of follow-up in computing cost rates, using time as the denominator. This study did not explore how a patient’s pattern of care and associated costs evolved over time following diagnosis. Any medical services not covered by health insurance such as over-the-counter drugs or most forms of alternative healthcare are also not included. In addition, medication utilization cannot be directly linked to a particular diagnosis. The Medicaid population and the Medicare population without supplemental insurance were not included in these two databases which represent a sample of the US population with employer-sponsored insurance. This may limit the generalizability of the findings. Because of the limitations of claims databases, future studies should validate these findings, especially the definition of IPF, based on chart reviews or prospectively designed studies.
In summary, our results demonstrate a high burden of illness of IPF in terms of increased comorbidity, incremental healthcare resource utilization, and additional direct costs compared to matched controls. To address the unmet needs and reduce the clinical burden of IPF, effective interventions and management plans are needed.
Transparency
Declaration of funding
This study was sponsored by Actelion Pharmaceuticals Ltd., Switzerland.
Declaration of financial/other relationships
HRC has served as a consultant to Actelion in the past. HRC received no payment for work on this project, and currently serves as a consultant for Fibrogen, Gilead, InterMune, and Onyx. He also has received grants and contracts from Boehringer Ingelheim and Genentech. AW and SL have disclosed that they are employed by United BioSource Corporation, a company that received funding from Actelion to conduct this study. DCH has disclosed that he is an employee of Hayflinger Analytic Services, a company that also received funding from Actelion to conduct this study. DMR and EH have disclosed that they are employees of Actelion Pharmaceuticals LTD.
Supplementary Material
Download PDF (300.2 KB)Acknowledgments
Statistical evaluations were performed by AW and SL. HRC and AJW contributed equally to the writing of this article.
References
- Raghu G, Weycker D, Edelsberg J, et al. Incidence and prevalence of idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2006;174:810-6
- Olson AL, Swigris JJ, Lezotte DC, et al. Mortality from pulmonary fibrosis increased in the United States from 1992 to 2003. Am J Respir Crit Care Med 2007;176:277-84
- American Thoracic Society. Idiopathic pulmonary fibrosis: diagnosis and treatment. International consensus statement. American Thoracic Society (ATS), and the European Respiratory Society (ERS). Am J Respir Crit Care Med 2000;161:646-64
- Ryerson CJ, Berkeley J, Carrieri-Kohlman VL, et al. Depression and functional status are strongly associated with dyspnea in interstitial lung disease. Chest 2011;139:609-16
- Nicholson AG, Colby TV, du Bois RM, et al. The prognostic significance of the histologic pattern of interstitial pneumonia in patients presenting with the clinical entity of cryptogenic fibrosing alveolitis. Am J Respir Crit Care Med 2000;162:2213-7
- Bjoraker JA, Ryu JH, Edwin MK, et al. Prognostic significance of histopathologic subsets in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1998;157:199-203
- Priest JL, Cantrell CR, Fincham J, et al. Quality of care associated with common chronic diseases in a 9-state medicaid population utilizing claims data: an evaluation of medication and health care use and costs. Popul Health Manag 2011;14:43-54
- Fell CD, Martinez FJ, Liu LX, et al. Clinical predictors of a diagnosis of idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2010;181:832-7
- Nathan SD, Basavaraj A, Reichner C, et al. Prevalence and impact of coronary artery disease in idiopathic pulmonary fibrosis. Respir Med 2010;104:1035-41
- Gribbin J, Hubbard R, Smith C. Role of diabetes mellitus and gastro-oesophageal reflux in the aetiology of idiopathic pulmonary fibrosis. Respir Med 2009;103:927-31
- Harris JM, Johnston ID, Rudd R, et al. Cryptogenic fibrosing alveolitis and lung cancer: the BTS study. Thorax 2010;65:70-6
- Nathan SD, Noble PW, Tuder RM. Idiopathic pulmonary fibrosis and pulmonary hypertension: connecting the dots. Am J Respir Crit Care Med 2007;175:875-80
- Cottin V, Nunes H, Brillet PY, et al. Combined pulmonary fibrosis and emphysema: a distinct underrecognised entity. Eur Respir J 2005;26:586-93
- Shorr AF, Wainright JL, Cors CS, et al. Pulmonary hypertension in patients with pulmonary fibrosis awaiting lung transplant. Eur Respir J 2007;30:715-21
- Mejia M, Carrillo G, Rojas-Serrano J, et al. Idiopathic pulmonary fibrosis and emphysema: decreased survival associated with severe pulmonary arterial hypertension. Chest 2009;136:10-5