The American Journal of Managed Care July 2008
Economic Assessment of Initial Maintenance Therapy for Chronic Obstructive Pulmonary Disease
Objective: To compare the effects of initial maintenance therapy with fluticasone 250 µg plus salmeterol 50 µg in a single inhaler versus other inhaled medications on exacerbation risks and treatment costs among chronic obstructive pulmonary disease (COPD) patients.
Study Design: A retrospective observational analysis was conducted by using medical/ pharmacy claims from a large managed care database between January 2000 and February 2004. Patients age 40 years or older with a primary diagnosis of COPD (International Classification of Diseases, Ninth Revision, Clinical Modification code 490, 491, 492, or 496), at least 18 months of continuous eligibility, and an index prescription for fluticasone/ salmeterol combination, salmeterol alone, inhaled corticosteroid alone, ipratropium/ albuterol combination, or ipratropium alone (reference) were identified.
Methods: Time to first COPD-related hospitalization or emergency department (ED) visit was estimated by using Cox proportional hazard models. All-cause and COPD-related treatment costs were estimated by using generalized linear models with a gamma distribution and log link. Multivariable regressions were used, controlling for age, sex, comorbidities, COPD subtype, preindex medications, and hospitalizations and ED visits.
Results: Initial maintenance therapy with fluticasone/ salmeterol combination was associated with a 31% to 56% lower risk of hospitalization or ED visit compared with ipratropium alone, adjusting for baseline characteristics and preindex resource utilization. Fluticasone/ salmeterol combination therapy was related to lower medical costs, higher pharmacy costs, and almost similar total costs in all populations studied.
Conclusion: Fluticasone/salmeterol combination therapy was considered to be costeffective compared with ipratropium alone because it achieved better clinical outcomes with similar or lower treatment costs.
(Am J Manag Care. 2008;14(7):438-448)
Medical/pharmacy claims from a large managed care database were used to compare fluticasone 250 µg plus salmeterol 50 µg in a single inhaler with other inhaled medications for initial maintenance therapy in patients with chronic obstructive pulmonary disease.
Fluticasone/salmeterol combination therapy was associated with a 31%-56% lower risk of hospitalization or emergency department visit compared with ipratropium alone.
Fluticasone/salmeterol combination therapy was considered to be cost-effective compared with ipratropium alone because it achieved better clinical outcomes with similar or lower total treatment costs.According to the Global Initiative for Chronic Obstructive Pulmonary Disease (COPD) guidelines, prevention of exacerbations is an important goal of COPD management.1 However, data from the third National Health and Nutrition Examination Survey suggest that the burden of COPD still is substantial.2 The surveys reveal high rates of COPD-related hospitalizations and emergency department (ED) visits, outcomes that serve as proxies of exacerbations in the real-world setting. In 2000, COPD accounted for 1.5 million ED visits and 726,000 hospitalizations in the United States.2 Hospitalizations for COPD have increased across all age groups since 1990.
Acute exacerbations of COPD have a negative impact on quality of life.3,4 Additionally, exacerbations generate a high demand for healthcare services. Studies have consistently shown that hospitalizations account for the largest portion of healthcare expenditures for COPD.5-10 For example, the Confronting COPD survey, a large-scale international patient survey of disease burden, demonstrated that inpatient hospitalizations accounted for approximately 70% of the total direct costs of COPD in the United States.6 Thus, identification of therapeutic agents that prevent or reduce the risk of hospitalization and ED visits is an important strategy for reducing the overall burden of the disease.
Because airway obstruction is a common feature of COPD, bronchodilators are commonly provided for symptomatic relief. Anticholinergics (eg, ipratropium), short-acting β2 agonists (eg, albuterol), long-acting β2 agonists (eg, salmeterol), or a combination of these drug classes appear to increase lung function (forced expiratory volume in the first second of expiration), decrease frequency of exacerbations, and improve quality of life.11 Inhaled corticosteroid (ie, fluticasone) is used to augment regular bronchodilator treatment for attenuating airway hyperresponsiveness and inflammation. Results of several prospective clinical studies showed a reduction in the rate and/or severity of exacerbations of COPD with the use of inhaled corticosteroid alone or in combination with long-acting β2 agonists.12-16
Moreover, a recent randomized clinical trial, the Towards a Revolution in COPD Health survival study, reported the combination therapy with fluticasone and salmeterol significantly reduced exacerbations by 25% and hospitalizations by 17% compared with placebo among 6000 patients over 3 years.17 However, information about economic benefits of selecting appropriate initial maintenance therapy for COPD remains limited, especially in the real-world setting.
Therefore, the present study was conducted to investigate the effect of various COPD medication regimens on (1) the risk of all-cause and COPD-related hospitalization and ED visits (risk analysis) and (2) all-cause and COPD-related treatment costs (cost analysis) using claims from a large managed care database. Of note, this was the first observational study to investigate the effect of fluticasone/salmeterol combination therapy in a single inhaler on utilization and costs in a managed care population.
Study Design and Sample
This was a retrospective, observational cohort study using medical and pharmacy claims from a large managed care database encompassing more than 30 different managed care plans and 33 million patients across the United States. This study was exempt from institutional review board oversight because no human subjects were involved and all data were HIPAA (Health Insurance Portability and Accountability Act) compliant, with all health plan and personal information identifiers removed to ensure confidentiality.
The study population consisted of patients ≥40 years of age with a diagnosis of COPD less than 1 year prior to initial treatment with fluticasone 250 μg and salmeterol 50 μg in a single inhaler, salmeterol alone, inhaled corticosteroid alone, ipratropium and albuterol in a single inhaler, and ipratropium alone (index medications). COPD was identified by International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code 490.xx, 491.xx, 492.xx, or 496.xx. Patient data were included in the study if the first pharmacy claim for the index medications occurred between January 2, 2001, and August 12, 2003 (index date). The time frame of the analysis is summarized in Figure 1. For the risk analysis, patients were required to have 18 months of continuous eligibility (12-month preindex period and 6-month postindex period). The cost analysis included a subpopulation who had a 12-month postindex period in order to calculate annual costs.
Patients were not permitted to receive any COPD treatment other than oral corticosteroids, short-acting β2 agonists, or theophylline during the 12-month preindex period, and patients could not initiate treatment with any other COPD medications within 60 days of the index date. Additionally, patients without a pharmacy benefit through their health plan were excluded.
The risk analysis assessed the time to first all-cause hospitalization, ED visit, or a combined end point of either of them, and time to first COPD-related hospitalization, ED visit, or a combined end point of either of them as defined by the primary diagnosis. Patients without a hospitalization or ED visit were censored when they lost eligibility or changed health insurance plans. Additionally, patients who remained eligible at completion of the database in February 2004 also were censored.
The cost analysis evaluated all-cause and COPD-related medical, pharmacy, and total costs during the 12-month postindex period. All cost outcomes were expressed as annual costs per patient. Costs of COPD-related care were identified by the primary diagnosis codes of COPD (ICD-9-CM code 490.xx, 491.xx, 492.xx, or 496.xx), COPD-related medical procedures (ie, spirometry test, oxygen therapy, pulmonary rehabilitation), and COPD-related medications (index medications plus oral corticosteroids, short-acting β2 agonists, theophylline, and oral antibiotics).
Differences in demographic and utilization characteristics measured during the preindex period were assessed by using Wilcoxon’s rank sum tests for continuous variables and c2 χtests for categorical variables. Because of the large number of comparison tests (n = 56), the alpha level was adjusted by using a Bonferroni correction procedure (ie, 0.05/56 = .001).
For the risk analysis, Cox proportional hazard analyses on the time to the first all-cause/COPD-related hospitalization and/or ED visit were performed comparing the fluticasone/ salmeterol combination, salmeterol, inhaled corticosteroid, and ipratropium/albuterol combination cohorts with the ipratropium cohort.18 The model was adjusted for underlying differences in age (continuous), sex (female as reference), comorbidities (asthma [ICD-9 code 493.xx], congestive heart failure [CHF; ICD-9 code 428.xx], upper respiratory tract infection [ICD-9 code 460.xx, 462.xx, 464.xx, 466.xx, or 487.xx], and lower respiratory tract infection [ICD-9 code 480.xx, 481.xx, 482.xx, or 483.xx]), COPD subtype (chronic bronchitis [ICD-9 code 491.xx] or emphysema [ICD-9 code 492.xx]), number of preindex prescriptions for respiratory medications (oral corticosteroids, short-acting β2 agonists, theophylline), number of preindex prescriptions for all medications, and number of preindex all-cause hospitalizations and ED visits. The hazard ratio (HR) and 95% confidence interval (CI) were calculated for each variable of interest. The proportional hazards assumptions were confirmed through a graphical approach (Kaplan-Meier survival curve) and Cox significance test. The final model was selected after examining interaction terms with age, asthma, and CHF diagnosis.
For the cost analysis, annual all-cause and COPD-related costs were compared across study groups. Unadjusted costs were summarized using means, SDs, and 95% CIs. Then, multivariate generalized linear models with a log link and gamma distribution were used to estimate treatment costs, controlling for differences in demographic and utilization characteristics during the preindex period.19 Adjusted costs were reported with 95% CI of cost differences (vs ipratropium), which were obtained by the bootstrap method with 1000 replications.20
Stratified analyses for patients with and without an asthma diagnosis were conducted to examine whether observed treatment effects were modified by this diagnosis. All statistical analyses were conducted with SAS version 8.2 (SAS Institute Inc., Cary, NC) or STATA version 9.2 (StataCorp LP, College Station, TX).
A total of 12,381 patients met the established inclusion and exclusion criteria for the risk analysis (Figure 2), including 1832 patients in the fluticasone/salmeterol combination cohort, 1099 patients in the salmeterol cohort, 3940 patients in the inhaled-corticosteroid cohort, 3388 patients in the ipratropium/albuterol combination cohort, and 2122 patients in the ipratropium cohort. Demographic and utilization characteristics during the preindex period for the 5 cohorts are shown in Table 1. Several statistically significant differences were observed between treatment groups. For example, patients in the ipratropium cohort were significantly older (mean age = 63.3 years), had a higher frequency of comorbid CHF (21.4%), and had a higher mean [SD] number of preindex ED visits (2.3 [6.3]) and hospitalizations (0.8 [1.4]) compared with the other cohorts. A comorbid diagnosis of asthma was significantly more common in the fluticasone/salmeterol combination (40.5%), salmeterol (29.7%), and inhaled-corticosteroid (39.6%) groups compared with the group receiving ipratropium alone (20.8%). A diagnosis of chronic bronchitis was more common than emphysema across all cohorts. The mean number of preindex prescriptions for oral corticosteroids ranged from 0.2 to 0.4 across cohorts. Compared with the ipratropium cohort, preindex utilization of short-acting b2 agonists was significantly higher in the fluticasone/salmeterol combination, salmeterol, and inhaled-corticosteroid cohorts and was significantly lower in the ipratropium/albuterol combination cohort.