Chronic obstructive pulmonary disease (COPD) is characterized by persistent, progressive airflow limitation that is not fully reversible.1 Although COPD treatments can reduce symptoms and can slow disease progression, acute periods of rapid symptom worsening beyond typical day-to-day variations (ie, exacerbations) can lead to hospitalization and a change in medication.1 Because exacerbations often occur in clusters, with an increased risk for a second exacerbation within 8 weeks of an initial event,2 treatments that minimize the impact of an active exacerbation and decrease the risk for future exacerbations can improve patient outcomes,1 which would reduce disease-related costs.
Currently in the United States, chronic lower respiratory disease, which includes COPD, is the third leading cause of death.3 In addition to the physical and emotional burdens of COPD, the total medical costs of the disease are staggering. In a recent US study, $32.1 billion was attributed to COPD-related medical costs, with a projected increase to $49 billion by 2020.4 In a large US retrospective claims study, the average cost per COPD-related hospital stay (mean, 4.8 days) was $7500, with acute exacerbations accounting for approximately 63% of all COPD-related hospitalizations (with COPD as the primary diagnosis).5 In another retrospective claims study, all-cause quarterly incremental US healthcare costs for patients with COPD who experienced an exacerbation were nearly double those of patients without exacerbations.6
Roflumilast is an oral, once-daily selective phosphodiesterase (PDE)-4 inhibitor indicated for reducing the risk for COPD exacerbations in patients with severe COPD associated with chronic bronchitis and a history of exacerbations.7 Currently, roflumilast is the only PDE-4 inhibitor approved for this indication. The efficacy of roflumilast as adjunctive therapy has been demonstrated in several well-controlled clinical trials, in which patients with severe COPD had fewer exacerbations with roflumilast.8-10
This study analyzed the effects of early versus delayed initiation of roflumilast treatment after a hospital or emergency department discharge resulting from COPD exacerbation. Moderate-to-severe exacerbation rates, rehospitalization rates, medication use, health services utilization, and medical costs were calculated for patients with a primary or a secondary diagnosis of COPD, adjusting for COPD disease complexity and comorbidities.
In this retrospective, observational cohort study, deidentified data from March 2011 through March 2013 were extracted from Truven Health MarketScan combined commercial and Medicare supplemental healthcare claims databases, which contain healthcare costs and utilizations, as well as inpatient, outpatient, and prescription drug services data on almost 230 million patients since 1995.11 These databases include inpatient and outpatient diagnosis, and drug and procedure codes submitted with each prescription and each professional or facility claim. The coding schemes used in the databases included the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis and procedure codes, Current Procedural Terminology procedure codes, Healthcare Common Procedure Coding System codes, and National Drug Codes universal product identifiers.
The index date was defined as the date of the first hospitalization or emergency department visit with a primary or secondary diagnosis of COPD (ICD-9 codes, 491.xx, 492.xx, 496.xx). Patients with COPD initiating roflumilast treatment ≤30 days after hospital or emergency department discharge were included in the early roflumilast treatment cohort; patients initiating roflumilast 31 to 180 days after discharge were included in the delayed roflumilast treatment cohort. The data were examined for a minimum of 180 days after the index date.
Patients aged ≥40 years at the index date who were continuously enrolled in a health plan during the study period (including 12 months before and ≥6 months after the index date) were included in the analysis if they had ≥1 inpatient hospitalizations with a primary or a secondary diagnosis of COPD or 1 emergency department visit with a primary or a secondary diagnosis of COPD and had been dispensed roflumilast within 6 months after the index date. To reduce bias resulting from confounding respiratory diseases, patients were excluded if they had any conditions other than COPD or asthma that can cause chronic airflow limitation, such as cystic fibrosis, congenital cystic lung malformation, allergic alveolitis, respiratory cancer, pneumoconiosis, pulmonary fibrosis, sarcoidosis, or pulmonary tuberculosis.
The patients were grouped into 1 of 3 COPD complexity categories based on criteria adopted from Pasquale and colleagues and Macaulay and colleagues, which included greater disease complexity (≥1 of the following conditions: nutritional abnormalities; osteoporosis; lung volume reduction surgery; oxygen therapy; COPD-related emergency department visit; or COPD-related outpatient visit), lesser disease complexity (renal failure ICD-9 codes: 585, 586, v45.1, v42.0, v56), or unclassified.12,13
The main outcomes measures of this study were the number, rate, and severity of exacerbations, rehospitalization rates, and healthcare resource and utilization costs. Exacerbations were defined as an inpatient hospitalization or emergency department visit with a diagnosis of acute COPD exacerbation (ICD-9 codes 491.21, 491.22, and 493.22) or emphysema (ICD-9 code 492.8x); or ≥1 medical claims for primary respiratory failure (ICD-9 codes 518.81 and 518.84) and a medical claim for a COPD diagnosis with acute exacerbation at secondary diagnosis position or for emphysema (ICD-9 code 492.8x) at a secondary diagnosis position in an inpatient or an emergency department setting; or ambulatory care of an exacerbation (identified by medical claims for a qualifying diagnosis occurring in a physician office setting or pharmacy claims for qualifying systemic steroids or oral antibiotics frequently used for respiratory infections).
Exacerbation severity was categorized as either moderate (ie, the use of oral corticosteroids and/or oral antibiotics, including those frequently used to treat respiratory infections) or severe (ie, COPD-related rehospitalization or death [discharge status of “died” or ICD-9-CM code of 798] that occurred after the index COPD hospitalization and within the study period). This algorithm was chosen based on previous studies in which claims data were used to classify exacerbation severity in the hospital setting.14,15
Hospitalization rates were examined at 60 days, 90 days, and 180 days after hospital or emergency department discharge. The use of maintenance drugs and other respiratory medications included long-acting beta-2 agonists, long-acting muscarinic antagonists, theophylline, inhaled corticosteroids, short-acting beta-2 agonists, and short-acting muscarinic antagonists.
Healthcare resource utilization included inpatient care (eg, rehospitalizations) and outpatient care (ie, physician office visits, emergency department procedures, COPD-related services, and other outpatient visits). For the inpatient costs, medications, physician services, procedures and laboratory tests, and general hospital services were evaluated.
The inpatient utilization variables included the length of stay, the number of procedures or laboratory tests performed, the number of medication orders written, and the total dose of COPD medication administered. Comorbidities were defined using the Charlson Comorbidity Index with the Romano adaptation.16,17
To evaluate the benefits of early roflumilast treatment after a hospital or an emergency department discharge, univariate and modeling analyses were used. The baseline patient characteristics (including age, sex, and comorbidities) for each cohort were summarized using descriptive statistics, and between-group comparisons were made between those characteristics using t-tests for continuous variables and chi-square testing for categorical variables. The comorbidity burden was calculated using the Charlson Comorbidity Index with Romano adaptation16,17 and was compared between the 2 cohorts. The proportion of patients with exacerbations, COPD-related rehospitalizations, and all-cause hospitalizations were compared between the 2 cohorts using chi-square testing. The annualized exacerbations and the number of exacerbation episodes between the 2 groups were compared using t-tests.
Based on previous publications,1,13 univariate and bivariate analyses were used to select variables with a cutoff for significance of P <.10 in the bivariate analysis to be included in the multivariate regression analysis. Factors such as age, sex, comorbidities (ie, asphyxia and respiratory arrest, respiratory infection, nutritional abnormalities, pneumonia, and osteoporosis), COPD complexity, and insurance plan type were used as covariates to examine the effect of early roflumilast treatment in the model development. Logistic regression modeling was used to predict the subsequent exacerbations using the above factors.
The generalized linear models were adjusted for baseline age, sex, COPD complexity, nutritional abnormalities, insurance plan type, and preindex comorbidities, and costs with gamma distribution and log-link function were used to compare the between-group healthcare costs. Utilization and costs were aggregated at the patient level and were presented by COPD-related and all-cause utilization. The average annual utilization per group for each type of service used was determined. All costs were adjusted to 2013 pricing levels using the medical care component of the Consumer Price Index.
A total of 995 patients were included in the analysis, 280 in the early roflumilast treatment group and 715 in the delayed roflumilast treatment group (Figure). Overall, the demographics and baseline clinical characteristics were similar between the cohorts, with the exceptions of COPD complexity, sex, and nutritional abnormalities (Table 1).
Compared with the delayed roflumilast treatment group, a larger proportion of patients in the early roflumilast treatment group had greater COPD complexity at baseline and a significantly greater proportion were male (P = .05; Table 1). A significantly larger percentage of patients in the early roflumilast treatment group had comorbid nutritional abnormalities compared with the patients in the delayed roflumilast treatment group (P = .01; Table 1). There were no significant between-group differences by preindex medication use, by US region (ie, Northeast, North Central, South, West, or unknown), or by payer status (eg, HMO, preferred provider organization [PPO], point of service [POS], and indemnity health plan).
Overall, early initiation of roflumilast was associated with significantly fewer and less severe exacerbations. Compared with patients in the delayed roflumilast treatment group, fewer patients in the early roflumilast treatment group had any exacerbation after the index date (73.6% with early roflumilast treatment vs 81.4% with delayed roflumilast treatment; P = .007; Table 2). Fewer patients in the early treatment group had moderate (P = .013), severe (P = .002), and moderate and severe (P = .002) exacerbations after the index date compared with the delayed treatment group (Table 2).
During the follow-up period, the patients in the early roflumilast treatment cohort had a lower mean number of exacerbations of any severity (P = .002) and of moderate severity (P = .001) compared with those in the delayed roflumilast treatment cohort (Table 2). There were no significant between-group differences observed in the mean number of severe or of moderate and severe exacerbations or in the time to first exacerbation after the index date (P >.05; Table 2). Similarly, when the exacerbation rates were annualized per patient, no significant differences were observed between the early roflumilast treatment and the delayed roflumilast treatment groups (P >.05; Table 2).
Significantly fewer patients in the early roflumilast treatment group were rehospitalized for COPD-related causes at 60, 90, or 180 days after discharge than the patients in the delayed roflumilast treatment group (P = .018, P = .006, and P = .05, respectively; Table 2). Similarly, a smaller percentage of early roflumilast treatment patients were rehospitalized for all causes at 60 or 90 days (P = .025, P = .015), with a similar trend observed at 180 days (P = .068; Table 2). No deaths on discharge were recorded in either cohort.
Overall, the patients in the early roflumilast treatment group had a lower risk for exacerbation than those in the delayed treatment group (Table 3). Patients who initiated roflumilast treatment early after hospitalization had a 39% lower risk for an exacerbation than patients who delayed their treatment (odds ratio [OR], 0.608; 95% confidence interval [CI], 0.435-0.851; P = .004; Table 3).
Annualized Utilization and Costs
The patients in the early roflumilast treatment group had fewer COPD-related hospitalizations (P = .012) and procedures and therapies (P = .016), as well as fewer outpatient visits per patient (P <.001) and emergency department visits per patient (P = .035), than those in the delayed roflumilast treatment group (Table 4). Compared with the patients in the delayed roflumilast treatment group, the patients in the early treatment group had significantly lower total COPD-related costs (P = .014) and lower costs for outpatient services (P = .04; Table 4).
Early roflumilast treatment was associated with fewer all-cause hospitalizations and emergency department visits, as well as fewer procedures and therapies (P <.01; Table 4) and fewer emergency department and outpatient visits per patient (P <.01 for all; Table 4). Significantly lower all-cause total and outpatient services costs were observed in the early roflumilast treatment group versus the delayed roflumilast treatment group (P ≤.01 for all; Table 4). The mean all-cause total costs in the early roflumilast treatment group also were significantly less than in the delayed treatment group ($41,973 vs $56,084, respectively; P = .002; Table 4).
When adjusted for age, sex, insurance plan type, and COPD complexity and comorbidities, the early roflumilast treatment group had a 32% COPD-related annualized cost reduction and a 26% all-cause annualized cost reduction compared with the delayed treatment group (P <.001 for both; Table 5). The factors that affected the COPD-related costs were a diagnosis of pneumonia in the preindex period (P = .0001) and having an HMO versus a PPO health insurance plan (P = .043; Table 5).
The factors that affected the all-cause costs were age (P = .023), a diagnosis of pneumonia in the preindex period (P = .007), COPD complexity (lesser vs unclassified; P = .041), and having a POS versus a PPO health plan (P = .029; Table 5).
Risks for COPD-Related and All-Cause Hospitalizations
Using a logistic regression model and adjusting for demographics, COPD complexity, comorbidities, and insurance plan type, the patients in the early roflumilast treatment group had a 42% lower risk for COPDrelated hospitalization (OR, 0.578; 95% CI, 0.403-0.829; P = .003) and a 37% lower risk for all-cause hospitalization (OR, 0.633; 95% CI, 0.461-0.869; P = .005) than the patients in the delayed roflumilast treatment group.
This is the first study to examine the effects of the timing of roflumilast treatment initiation on exacerbation rates and healthcare resource utilization and costs among patients who were discharged from hospitals and emergency department visits as a result of COPD exacerbations. The results from this study demonstrate that the patients receiving roflumilast within 30 days of discharge had significantly fewer future exacerbations, rehospitalizations, and emergency department visits than the patients receiving roflumilast 31 days to 180 days after discharge. The patients who initiated roflumilast treatment early also had a 39% lower risk for a second COPD exacerbation than those who delayed treatment.
In a study by Wilkinson and colleagues, the patients who sought earlier treatment for COPD exacerbations (with oral corticosteroids and/or antibiotics) had shorter recovery times (P <.001) and a reduced risk for emergency hospitalizations (P = .04).18
In our study, the early initiation of roflumilast treatment was also associated with a $7273 lower total annualized COPD-related cost and with significantly reduced COPD-related services, resulting in a $14,111 lower annualized all-cause total cost.
Our data were collected during the implementation of the 2010 Affordable Care Act (ACA). The goals of the ACA were to expand the coverage of health insurance, incentivize high-quality healthcare, and make prescription drugs more affordable.19,20 In 2012, the Centers for Medicare & Medicaid Services (CMS) initiated the Hospital Readmissions Reduction Program, with the aim of penalizing hospitals that have excessive hospital readmissions through the reduction of their Medicare payments.21 This program was expanded in 2014 to include COPD-related exacerbation readmissions.21
Furthermore, although the ACA aimed to reduce costs, prescription drug costs have been increasing. In 2016, a new Covered Outpatient Drugs final rule was issued by the CMS, which aims to ensure that prescription drugs are affordable and accessible to Medicare and Medicaid beneficiaries.20 As a whole, these systemwide healthcare changes might have had an effect on our results, and future studies evaluating the use of roflumilast after the implementation of the ACA are warranted. Our study could serve as a snapshot of the early ACA implementation landscape and could be compared with more recent claims data that include a similar patient population and study length.
Because this study uses claims data that are designed primarily for billing purposes, the severity of COPD could not be classified according to standard definitions, because of inadequate documentation that is inherent in any retrospective analysis, and the predicted risk for exacerbation was adjusted only by the observed variables and the available data.
Spirometric measurements, such as forced expiratory volume in 1 second, and other outcomes are not available to confirm a diagnosis of COPD or disease severity or to assess the number of exacerbations or other outcomes during follow-up.
In addition, because asthma and COPD can be difficult to differentiate, miscoding errors can occur. Thus, a diagnosis of asthma was not an exclusionary criterion, nor was it assessed as a comorbidity, because excluding all patients with asthma would have greatly reduced our sample size.
Furthermore, not all COPD exacerbations can be detected in claims data, because medical care is not always sought, and it is possible that the patients with severe asthma were misdiagnosed as having COPD.
The imbalance in the number of patients within the early roflumilast treatment group (N = 280) versus the delayed roflumilast treatment group (N = 715) may reflect prescribing practices (eg, patients receiving earlier treatment with roflumilast might have had greater disease complexity, as is shown in Table 1 [70.4% of the early roflumilast treatment group had greater disease complexity compared with 60.6% in the delayed roflumilast treatment group; P = .004]). Because our information comes from insurance claims data, the distribution of patients into each treatment group could not be predicted nor can the reasons for the disparity be confirmed; however, a multivariate analysis was used to control for disease complexity.
Roflumilast was approved by the US Food and Drug Administration in 2011, placing our study period within the first 2 years that roflumilast was available on the market. Thus, patients receiving treatment with roflumilast within the time frame of the study might have had different results from patients who are currently receiving the drug, because of changes in prescribing patterns as roflumilast becomes established in the COPD pharmacopoeia.
As previously mentioned, the adoption of the ACA during the data collection time frame of this study might also have affected our results.
In addition, because roflumilast is the only approved oral PDE-4 inhibitor for COPD, the effects of the timing of roflumilast could not be compared with that of other PDE-4–inhibiting drugs, and thus, it is unknown if early treatment with other PDE-4 inhibitors could confer the same benefit.
Early treatment of COPD-related exacerbations with oral corticosteroids and/or antibiotics also has been shown to improve health outcomes.18 However, because patients in the present study would also have received treatment with oral corticosteroids and/or antibiotics for COPD exacerbations, it is not possible to compare the effects of early versus late roflumilast treatment with that of oral corticosteroids and antibiotics.
Previous evidence has shown that early treatment of COPD exacerbations with corticosteroids or antibiotics improves patient outcomes.18 This current analysis of patient claims data suggests that the early initiation of roflumilast treatment after a hospitalization or an emergency department visit for a COPD exacerbation is beneficial in lowering the risk for subsequent exacerbations, as well as reducing COPD-related and all-cause healthcare utilization and costs. In addition to reducing disease-related costs, minimizing the impact of an active exacerbation and decreasing the risk for future exacerbations may lead to improvements in patient outcomes. The timely use of roflumilast could represent a valuable treatment in a select group of patients with severe COPD that is associated with chronic bronchitis and a history of exacerbations.
Editorial support was provided by Lynn M. Anderson, PhD, and Kristen A. Andersen, PhD, of Prescott Medical Communications Group, Chicago, IL, and was funded by AstraZeneca Pharmaceuticals.
This study was supported by Forest Laboratories LLC, an affiliate of Allergan.
Author Disclosure Statement
Dr Lee is a consultant to Forest Laboratories. Dr Sun and Ms Mocarski were employees of Allergan (Legacy Forest Laboratories) at the time this study was conducted.
1. Global Initiative for Chronic Obstructive Lung Disease. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. Updated 2016. http://goldcopd.org/global-strategy-diagnosis-management-prevention-copd-2016/. Accessed December 7, 2015.
2. Hurst JR, Donaldson GC, Quint JK, et al. Temporal clustering of exacerbations in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2009; 179:369-374.
3. Heron M. Deaths: leading causes for 2010. Natl Vital Stat Rep. 2013;62:1-96.
4. Ford ES, Murphy LB, Khavjou O, et al. Total and state-specific medical and absenteeism costs of COPD among adults aged ≥18 years in the United States for 2010 and projections through 2020. Chest. 2015;147:31-45.
5. Wier LM, Elixhauser A, Pfuntner A, Au DH. Overview of hospitalizations among patients with COPD, 2008. Statistical brief #106. February 2011. Agency for Healthcare Research and Quality, Rockville, MD. www.ncbi.nlm.nih.gov/books/NBK53969/pdf/Bookshelf_NBK53969.pdf. Accessed December 7, 2015.
6. Yu AP, Yang H, Wu EQ, et al. Incremental third-party costs associated with COPD exacerbations: a retrospective claims analysis. J Med Econ. 2011;14:315-323.
7. Daliresp (roflumilast) tablets [prescribing information]. Wilmington, DE: AstraZeneca Pharmaceuticals; November 2015.
8. Rennard SI, Calverley PMA, Goehring UM, et al. Reduction of exacerbations by the PDE4 inhibitor roflumilast—the importance of defining different subsets of patients with COPD. Respir Res. 2011;12:18.
9. Calverley PMA, Rabe KF, Goehring U-M, et al; for the M2-124 and M2-125 Study Groups. Roflumilast in symptomatic chronic obstructive pulmonary disease: two randomised clinical trials. Lancet. 2009;374:685-694. Erratum in: Lancet. 2010;376:1146.
10. Martinez FJ, Calverley PMA, Goehring U-M, et al. Effect of roflumilast on exacerbations in patients with severe chronic obstructive pulmonary disease uncontrolled by combination therapy (REACT): a multicentre randomised controlled trial. Lancet. 2015;385:857-866.
11. Truven Health Analytics. Putting research data into your hands with the MarketScan Databases. 2015. http://truvenhealth.com/markets/life-sciences/products/data-tools/marketscan-databases. Accessed May 29, 2015.
12. Pasquale MK, Sun SX, Song F, et al. Impact of exacerbations on health care cost and resource utilization in chronic obstructive pulmonary disease patients with chronic bronchitis from a predominantly Medicare population. Int J Chron Obstruct Pulmon Dis. 2012;7:757-764.
13. Macaulay D, Sun SX, Sorg RA, et al. Development and validation of a claims-based prediction model for COPD severity. Respir Med. 2013;107:1568-1577.
14. Mapel DW, Dutro MP, Marton JP, et al. Identifying and characterizing COPD patients in US managed care. A retrospective, cross-sectional analysis of administrative claims data. BMC Health Serv Res. 2011;11:43.
15. Rothberg MB, Pekow PS, Lahti M, et al. Antibiotic therapy and treatment failure in patients hospitalized for acute exacerbations of chronic obstructive pulmonary disease. JAMA. 2010;303:2035-2042.
16. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373-383.
17. Romano PS, Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: differing perspectives. J Clin Epidemiol. 1993;46:1075-1079; discussion 1081-1090.
18. Wilkinson TMA, Donaldson GC, Hurst JR, et al. Early therapy improves outcomes of exacerbations of chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2004;169:1298-1303.
19. Reineck LA, Kahn JM. Quality measurement in the Affordable Care Act: a reaffirmed commitment to value in health care. Am J Respir Crit Care Med. 2013;187:1038-1039.
20. Centers for Medicare & Medicaid Services. CMS finalizes ACA rule to save taxpayers billions. Press release. January 21, 2016. www.cms.gov/Newsroom/MediaReleaseDatabase/Press-releases/2016-Press-releases-items/2016-01-21.html. Accessed February 22, 2016.
21. Braman SS. Hospital readmissions for COPD: we can meet the challenge. J COPD Found. 2015;2:4-7.