Abstract: Hospitalization costs are substantial in chronic obstructive pulmonary disease (COPD). Little is known regarding the effect of long-acting beta2-agonists (LABAs) vs short-acting beta2-agonists (SABAs) on hospital length of stay (LOS). Methods: This retrospective analysis utilized data from the Premier research database. Patients meeting the following criteria were included in the analysis: hospitalized for COPD; aged 40 years or older before hospital admission; and with a primary discharge diagnosis of COPD (ICD-9-CM codes 491.xx, 492.xx, or 496.xx) or a secondary diagnosis of COPD, with a primary discharge diagnosis for another illness (ICD-9-CM codes 460.xx–519.xx), between January 1, 2006, and March 31, 2010. Patients were matched by age, gender, and COPD severity. Comparisons were made for LOS and total hospitalization costs between those treated with nebulized arformoterol (a LABA) vs nebulized SABAs. Results: In total, 1165 arformoterol-treated patients were matched to 2610 SABA-treated patients. Unadjusted mean LOS was similar between the arformoterol and SABA groups (5.7 vs 5.9 days; P = .0506); after adjusting for background differences, arformoterol patients had significantly shorter LOS (4.5 days vs 5.5 days; P < .0001). Mean total hospitalization costs were lower for arformoterol than for SABA ($7004 vs $8638; P < .0001); adjusted total hospitalization costs were also lower in the arformoterol group ($5211 vs $7453; P < .0001). Conclusions: COPD patients treated with nebulized arformoterol experienced shorter LOS and reduced total hospitalization costs compared with nebulized SABA–treated patients. Arformoterol may reduce health care–related costs for hospitalized COPD patients.
Approximately 15 million US adults have been diagnosed with chronic obstructive pulmonary disease (COPD),1 and it is estimated that the number of COPD patients may be much higher when considering underdiagnosis.2 COPD is one of the leading causes of death in the United States1,3 and results in substantial costs and health care utilization. Overall, total medical costs for COPD in the United States in 2010 were estimated at $32.1 billion, with a projected increase to $49 billion by 2020.4
The 2011 Behavioral Risk Factor Surveillance System survey of patients with COPD showed that 18.6% of 13,279 respondents, adjusted for age, visited a hospital or emergency department (ED) in the previous 12 months for COPD symptoms.1 Data from the Healthcare Cost and Utilization Project showed that aggregate hospital costs for COPD as the primary diagnosis were approximately $6 billion in 2008, and mean length of stay (LOS) was 4.8 days.5 An evaluation of costs for COPD–related ED visits; simple inpatient admissions (no intubation or no intensive care); and complex admissions (categorized as intubation/no intensive care, intensive care/no intubation, or intensive care/intubation based on International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9 CM)] Current Procedural Terminology codes, and charge codes) revealed that mean costs in 2008 US dollars were $647 for ED visits, $7242 for simple admissions, and $20,757 for complex admissions. Additionally, mean LOS ranged from 4.5 days for simple admissions to 16 days for intensive care/intubation-related admissions. Although medication treatment patterns were assessed in this study, the potential relationship between inpatient treatments and associated LOS and costs was not evaluated.6
Managing symptoms with long-acting bronchodilators (eg, beta2-agonists, muscarinic antagonists) for maintenance therapy, and short-acting bronchodilators for additional symptom control as needed, is key to minimizing acute exacerbations and preventing hospitalization.7 The efficacy of short-acting beta2-agonists (SABAs) for treatment of acute exacerbations of COPD has been established, and current guidelines from the Global Initiative for Chronic Obstructive Lung Disease (GOLD) recommend SABAs with or without short-acting muscarinic antagonists for this purpose in the ED or inpatient setting.7 Additionally, GOLD guidelines recommend use of nebulizers or spacers in the ED or hospital setting for management of exacerbations.7 Although clinical studies on the use of long-acting bronchodilators such as long-acting beta2-agonists (LABAs) for exacerbations are lacking,7 single center and retrospective studies suggest the potential use of LABAs for treatment of COPD-related hospitalizations. The addition of formoterol to a standardized respiratory therapist–driven protocol reduced the need for SABAs, health care resource utilization, and adverse events in hospitalized COPD and asthma patients.8 Patients treated with the nebulized LABA arformoterol during a COPD-related hospitalization had lower rates of readmission compared with patients treated with nebulized SABAs (neb-SABAs) in the 30 days following hospital discharge (8.7% vs 11.9%).9 An analysis of inpatient and outpatient claims data showed that patients receiving LABA maintenance therapy in the outpatient setting had reduced total health care costs and hospital inpatient days compared with patients receiving SABA therapy.10
Little is known about the effects of specific inpatient treatments on hospital LOS.11 The purpose of the current study was to examine hospital LOS and costs for patients with COPD who received neb-SABA therapy compared with the nebulized LABA arformoterol.
Methods
Study Design
This retrospective observational database study employed a matched case-control design. The administrative inpatient hospital claims data were from the Premier health care database (Premier Healthcare Solutions, Inc, Charlotte, NC), which contains comprehensive patient-level information including diagnosis codes, procedure codes, drugs provided, tests performed, and costs associated with the hospitalization on more than 50 million inpatient hospitalizations at over 600 hospitals of various sizes and characteristics. The database is de-identified, pursuant to the Health Insurance Portability and Accountability Act of 1996 and other regulations, and institutional review board review was not required for conducting this research.12
Inclusion and Exclusion Criteria
Patients were included in the analysis if they were hospitalized for an acute exacerbation of COPD, were at least 40 years of age before admission to the hospital, and had a primary discharge diagnosis of COPD (ICD-9-CM codes 491.xx, 492.xx, or 496.xx) or a secondary diagnosis of COPD, with a primary discharge diagnosis of another respiratory condition (ICD-9-CM codes 460.xx–519.xx) between January 1, 2006, and March 31, 2010. Patients were grouped into two cohorts: treatment and control. Patients in the treatment cohort were required to have nebulized arformoterol treatment for at least 80% of the inpatient days between the initiation and last documented use of arformoterol, to ascertain multiple treatment doses and continuity of treatment once initiated. Patients treated with nebulized arformoterol could have received a neb-SABA as a rescue medication to stabilize an exacerbation prior to or concurrent with nebulized arformoterol treatment. The control group included patients who were treated with a neb-SABA on one or more days during the hospitalization and did not receive any nebulized LABA treatment during their entire inpatient stay. Patients in the neb-SABA group could receive other respiratory-related drugs as standard of care (eg, muscarinic antagonists that are delivered via a handheld inhaler device [eg, metered dose inhaler, dry powder inhaler], corticosteroids, or antibiotics). The first hospitalization for a patient within the time period matching the selection criteria was utilized for the analyses.
Matching Procedure
To prevent potential selection bias, neb-SABA patients were matched to arformoterol patients based on four key variables: age (± 3 years), gender, primary vs secondary COPD diagnosis, and All Patient Refined Diagnosis Related Groups (APR-DRGs) classification of severity of illness as a measure of hospitalization severity, ranging from 1 (mild) to 4 (extreme). Matching occurred prior to the exclusion of patients who died during the baseline admission. All of the other background characteristics were statistically controlled in the data analyses. Each patient treated with arformoterol was matched 1:2 to patients treated with a neb-SABA. Exact matching on a large number of variables was not possible; therefore, variables including hospital, patient demographic, clinical, and prior treatment characteristics and illness severity were statistically controlled (Table 1). These background characteristics were identified from information collected during the initial hospitalization, prior to assessing LOS and total hospitalization cost. After matching at the 1:2 ratio, patients’ records that had incomplete or invalid data elements were excluded from analyses.
Patient illness severity level was assessed using the 3M APR-DRG classification system.9 These levels range from 1 (mild) to 4 (extreme) and are calculated after the hospitalization has ended, using age, ICD-9 diagnoses, and procedures performed during the hospitalization.9 Finally, information on all treatments administered during hospitalization that may have reflected differences in complexity of the case were captured with a set of indicator variables for oxygen therapy, respiratory therapy, oxygen saturation tests, anticholinergic treatment, corticosteroid treatment, antibiotic treatment, and other COPD medications.
Outcomes Measures
The outcomes of interest were LOS following treatment initiation and total hospital costs. LOS was defined as the number of days from initiation of beta2-agonist therapy to discharge. Hospitalization cost was defined as the cost of the admission, beginning with the day of initiation of beta2-agonist therapy and ending with discharge from the hospital. Thus, the days before beta2-agonist initiation, which would not have reflected treatment, were not added into the total posttreatment initiation cost of care.
Statistical Methods
Baseline differences between patients treated with arformoterol or neb-SABA during the initial admission were first assessed using the Wilcoxon rank-sum test for continuous measures and chi-squared tests for categorical variables. Descriptive measurements included medians and standard deviations for continuous measures. Quantities and percentages of patients were used to describe categorical distributions (eg, APR-DRG severity of illness).
Due to the highly skewed distribution of the outcomes, the adjusted outcome analysis was performed using gamma regression with the log-link function, which recognizes the skewed, positive nature of LOS (counts of days) and the inherent distribution of cost (US dollars) data. The regression models compared estimated mean LOS and mean total hospitalization cost for the arformoterol-treated patients with the neb-SABA–treated patients, while controlling for the previously described variables. Statistical significance was set to P < .05, and all analyses were completed using SAS version 9.2 (SAS Institute Inc., Cary, NC).
Results
Demographic and clinical characteristics of the hospitalized patients are presented in Table 1. In total, 1165 arformoterol patients met all the inclusion criteria and were matched to 2610 neb-SABA patients. Although patients were matched on age, gender, and APR-DRG severity of illness, several meaningful differences were observed in other background characteristics. Significant differences between groups were observed in the race/ethnicity distribution (P < .0001), overall admission types (P < .0001), and use of macrolides (P < .0001), budesonide (P < .0001), and other corticosteroids (P < .0001). Because of significant differences in background characteristics, unadjusted differences in the outcome variables may not be meaningful.
The analysis of LOS and hospitalization costs revealed favorable outcomes in patients treated with arformoterol compared with patients treated with neb-SABA. The unadjusted mean LOS was 5.7 (± 4.3) days in the arformoterol group and 5.9 (± 4.5; P = .05) days in the neb-SABA group. After adjustment for background differences, the mean LOS was significantly lower for patients in the arformoterol group compared with patients in the neb-SABA group (4.5 [3.9-5.1] days vs 5.5 [4.9-6.3] days; P < .0001). Both unadjusted and adjusted total hospitalization cost were significantly lower in the arformoterol group vs the neb-SABA group: the unadjusted cost for the arformoterol group was $7004 (± $7594) compared with $8638 (± $8286) for the neb-SABA group (P < .0001); the adjusted cost for the arformoterol group was $2242 lower than that of the neb-SABA group (P < .0001; Table 2).
The regression analysis revealed that several covariates were significantly associated with longer LOS: APR-DRG severity of illness (P < .0001); use of respiratory (P = .002) or any antibiotics (P = .005); use of budesonide (P = .001) and other corticosteroids (P = .048); and days on oxygen (P < .0001). Likewise, significant covariates of hospital costs were APR-DRG severity of illness (P < .0001); use of respiratory antibiotics (P < .0001); use of anticholinergic agents (P < .00001); use of budesonide (P = .012); and days on oxygen (P < .0001; Table 3). Of note, 135 patients died during the index hospital stay: 54 (6.2%) patients treated with nebulized arformoterol and 81 (4.7%) patients treated with neb–SABA. This suggests that arformoterol-treated patients were sicker and had more severe illness.
Discussion
In this retrospective observational study of inpatient hospital stays from a large number of hospitals across the United States, patients treated with nebulized arformoterol had significantly shorter LOS and lower total hospitalization costs than matched patients treated with a neb-SABA. Observed findings were supported by regression modeling, with adjustments for patient and hospital characteristics. Using arformoterol instead of a neb-SABA to treat hospitalized patients with COPD may be associated with shorter LOS and reduced hospital costs after treatment initiation.
Most COPD patients are not hospitalized; however, inpatient costs are the largest driver of direct health care costs associated with COPD.13 Furthermore, disease progression and severity have been closely linked to LOS and associated costs.14 A review of the mean yearly costs of treating COPD in the United States in 2006 revealed that the average patient incurred $3943 in direct expenses, whereas those who were hospitalized incurred $15,093 and those requiring intensive care unit (ICU) treatment during hospitalization incurred $43,461.15 Therefore, treatments that can reduce the recurrence or LOS for hospitalizations would have a significant impact on total costs of COPD.
Previous studies have found that maintenance treatment with LABAs lengthens time to hospitalization in patients with COPD compared with patients treated only with SABAs, and that arformoterol reduces the rate of hospital readmissions.11,16 While long-acting bronchodilators (including LABAs) are well established as maintenance treatment for COPD symptoms, few studies address the effect on outcomes when used in an inpatient setting during exacerbation. In a pilot study by DiMarco and colleagues aimed at investigating pharmacodynamic responses,17 patients with a COPD-related exacerbation treated with the long-acting bronchodilators tiotropium and/or formoterol demonstrated improved lung function, suggesting that these treatments are effective at bronchodilation even during an exacerbation episode. Likewise, Drescher and colleagues12 showed that initiating tiotropium in patients with an acute exacerbation of COPD as early maintenance while in the hospital resulted in a shorter LOS compared with patients who were treated prior to tiotropium becoming part of the hospital protocol.17 Conversely, a larger retrospective study by Lindenauer and colleagues18 found no difference in LOS when patients were treated with long-acting bronchodilators during an acute exacerbation.
Demographic characteristics such as age, gender, and race were generally similar between our study and previous studies. There are important differences to note between these studies and the present study. The DiMarco study was notably small (N = 21). The Drescher study was larger and examined tiotropium, an antimuscarinic, as opposed to a LABA, like arformoterol, such as in the current study. Drescher and colleagues also reported higher usage of respiratory antibiotics compared with our study. Finally, the Lindenauer study utilized a large number of data claims to reach their conclusions; however, this study combined all long-acting bronchodilators into one group, which raises the possibility of varying effectiveness of different drugs. Determination as to whether arformoterol or nebulized delivery has an advantage in this setting compared with other LABAs or antimuscarinic drugs may be an area of future interest.
Typically, SABAs are administered as rescue therapy to stabilize the patient upon admission to the hospital for an acute exacerbation, followed by LABA therapy after stabilization. The use of LABAs upon admission potentially offers advantages over SABA therapy. The longer duration of action of LABAs (12-24 hours) compared with SABAs (4-6 hours) confers sustained bronchodilator effects and less frequent administration, thus lower potential for adverse events that could increase LOS (eg, arrhythmias).7
Although this study provides a unique contribution to the literature by focusing on the costs of COPD from the hospital perspective in usual clinical care, some limitations exist. Retrospective cohort studies of claims data may suffer from selection bias due to unobservable variables, such as treatment preferences that may impact the study results and conclusions. The data were obtained from a large database comprising inpatient data from 602 hospitals. Due to incomplete or invalid data elements in patient records that were excluded from analyses, the cohorts were a slightly lower ratio than the planned 1:2 ratio. Clinical characteristics were identified from the initial admission that may have differed between the two cohorts. These included ICU stay and admission type (ED vs other). Although many background variables in the data were controlled by matching or multivariable regression, some important variables were not available for analysis. This database lacked smoking status and clinical measurements to indicate severity of COPD based on spirometry or other lung function test results. These may have differed between patients treated with arformoterol or neb–SABA. We also had no information on patient treatment regimens prior to hospitalization, so we were unable to characterize a patient’s maintenance COPD drug therapy in terms of concomitant medications and adherence to them. These study findings may not be generalizable given the recently updated GOLD 2017 guidelines, which highlight a trend towards precision medicine, using treatment strategies that include dose adjustment (ie, treatment escalation and de-escalation) based on levels of symptoms and patient risk for exacerbation.19 We were also unable to account for potential differences in cost if patients began treatment later in the course of their hospital stay compared with the matching counterparts. Finally, administrative databases are limited by the potential inaccuracies of diagnosis and procedure codes. The administrative data used in this study were collected for purposes other than research.
Conclusion
This comparative study of the nebulized LABA arformoterol vs nebulized SABAs found that patients treated with arformoterol had shorter hospitalizations and reduced total hospitalization cost relative to matched patients treated with SABAs. Initiating hospitalized COPD patients with LABAs may help reduce health care costs and prevent greater deterioration in lung functioning. The finding that arformoterol treatment is associated with lower LOS and hospitalization costs in COPD compared with nebulized SABA treatment suggests that long-acting bronchodilators may be more appropriate as a maintenance treatment for hospitalized patients with COPD.
References
1. Centers for Disease Control and Prevention. Chronic obstructive pulmonary disease among adults—United States, 2011. MMWR Morb Mortal Wkly Rep. 2012;61(46):938-943.
2. Mannino DM, Homa DM, Akinbami LJ, Ford ES, Redd SC. Chronic obstructive pulmonary disease surveillance—United States, 1971-2000. MMWR Surveill Summ. 2002;51(6):1-16.
3. Kochanek KD, Murphy SL, Xu J, Arias E. Mortality in the United States, 2013. NCHS Data Brief. 2014:1-8.
4. Ford ES, Murphy LB, Khavjou O, Giles WH, Holt JB, Croft JB. 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(1):31-45.
5. Bollu V, Karafilidis J, Colosia A, Bennett L, Hanania N. Comparison of efficacy and safety outcomes in randomized trials of long-acting and short-acting ß2-agonists for chronic obstructive pulmonary disease: a review. J Pulmon Resp Med. 2013;3:1-5.
6. Dalal AA, Shah M, D’Souza AO, Rane P. Costs of COPD exacerbations in the emergency department and inpatient setting. Respir Med. 2011;105(3):454-460.
7. Global Initiative for Chronic Obstructive Lung Disease (GOLD). Global Strategy for the Diagnosis, Management and Prevention of COPD. https://goldcopd.org/. Updated 2016. Accessed November 1, 2016.
8. Colice GL, Carnathan B, Sung J, Paramore LC. A respiratory therapist-directed protocol for managing inpatients with asthma and COPD incorporating a long-acting bronchodilator. J Asthma. 2005;42(1):29-34.
9. Averill RF, Goldfield NI, Muldoon J, Steinbeck BA, Grant TM. A closer look at all-patient refined DRGs. J AHIMA. 2002;73(1):46-50.
10. Ejzykowicz F, Bollu VK, Rajagopalan K, Hay JW. Health care use and costs among medicare patients with chronic obstructive pulmonary disease treated with short-acting beta agonists or long-acting beta agonists. J Clin Pathways. 2016;2(3):31-38.
11. Bollu V, Ernst FR, Karafilidis J, Rajagopalan K, Robinson SB, Braman SS. Hospital readmissions following initiation of nebulized arformoterol tartrate or nebulized short-acting beta-agonists among inpatients treated for COPD. Int J Chron Obstruct Pulmon Dis. 2013;8:631-639.
12. Drescher GS, Carnathan BJ, Imus S, Colice GL. Incorporating tiotropium into a respiratory therapist-directed bronchodilator protocol for managing in-patients with COPD exacerbations decreases bronchodilator costs. Respir Care. 2008;53(12):1678-1684.
13. Strassels SA, Smith DH, Sullivan SD, Mahajan PS. The costs of treating COPD in the United States. Chest. 2001;119(2):344-352.
14. Sullivan SD, Ramsey SD, Lee TA. The economic burden of COPD. Chest. 2000;117(2 suppl):5S-9S.
15. Dalal AA, Christensen L, Liu F, Riedel AA. Direct costs of chronic obstructive pulmonary disease among managed care patients. Int J Chron Obstruct Pulmon Dis. 2010;5:341-349.
16. Bollu V KJ, Colosia A, Bennett L, Hanania N. Comparison of efficacy and safety outcomes in randomized trials of long-acting and short-acting ß2-agonists for chronic obstructive pulmonary disease: a review. J Pulmon Resp Med. 2013;3:1-5.
17. Di Marco F, Verga M, Santus P, Morelli N, Cazzola M, Centanni S. Effect of formoterol, tiotropium, and their combination in patients with acute exacerbation of chronic obstructive pulmonary disease: a pilot study. Respir Med. 2006;100(11):1925-1932.
18. Lindenauer PK, Shieh MS, Pekow PS, Stefan MS. Use and outcomes associated with long-acting bronchodilators among patients hospitalized for chronic obstructive pulmonary disease. Ann Am Thorac Soc. 2014;11(8):1186-1194.
19. Global Initiative for Chronic Obstructive Lung Disease (GOLD). Global Strategy for the Diagnosis, Management and Prevention of COPD. https://goldcopd.org. Updated 2017. Accessed March 21, 2017.
