This retrospective database study used data from the QuintilesIMS Real-World Data Adjudicated Claims Database (formerly known as PharMetrics Plus). The database contains adjudicated medical and pharmacy claims for more than 150 million US health plan members from 2006 onwards. The data are patient-level, longitudinal, and representative of the US commercially insured population. The database covers 90% of US hospitals, 80% of all US doctors’ offices, and relate to 85% of Fortune 100 employees. Inpatient and outpatient diagnoses are recorded as International Classification of Diseases (ninth and tenth revisions) Clinical Modification (ICD-9/10-CM) codes. Data include inpatient and outpatient procedures, dates of service, retail and mail-order claims, inpatient stay, and provider specialty. Amounts allowed and paid by health plans are available for all services provided along with the dates of service for all claims. Other data elements available include demographic information (patient age, gender, and geographical region), product type (e.g., health maintenance organization and preferred provider organizations), payer type (e.g., commercial and self-insured), and start and stop dates of health plan enrollment.
Patients were selected for the study if they were treated with an AC+FV AVHC or a non-antistatic VHC (control VHC) between January 1, 2010 and August 31, 2015 (the “selection window”). The date of the first VHC device within the selection window was defined as the index date. All patients had to be treated with a long-term controller medication [inhaled corticosteroid (ICS) or fixed-dose combination of ICS and long-acting β2 agonist (ICS/LABA)] MDI within 60 days before or 60 days after the index date. All patients had at least 12 months of continuous healthcare coverage before the index date (“preindex period”), were diagnosed with asthma (ICD-9: 493.x) on index date or during the preindex period, and had at least 30 days of continuous healthcare coverage after the index date (“post-index period”). Patients were excluded if they had incomplete age, gender, or payer information; had a diagnosis of COPD (ICD-9: 490.x, 491.x, 492.x, 494.x, 495.x, and 496.x) on the index date or during the preindex period; or discontinued their long-term control MDI before index. In order to ensure that patients were newly treated with an AC+FV AVHC or control VHC, patients treated with AC+FV AVHC were excluded if they had evidence of non-AC+FV AVHC use (antistatic or control VHC) at any time during the study or AC+FV AVHC use during the preindex period. Patients treated with control VHC were excluded if they had evidence of any antistatic VHC use at any time during the study, or had control VHC use during the preindex period. Patients were categorized into mutually exclusive cohorts (AC+FV AVHC cohort or control VHC cohort) based on the type of VHC at the index date.
In order to reduce pretreatment confounders between the cohorts, patients in the AC+FV AVHC cohort were propensity score (PS) matched to patients in the control VHC cohort at a 1:1 ratio using a greedy matching algorithm without replacement [18]. Patients were matched on baseline age categories (0–2, 3–5, 6–12, 13–17, 18–34, 35–44, 45–54, 55–64, 65–74, 75 or older), gender, geographic region (Northeast, Midwest, South, West), Charlson comorbidity index (CCI) category (0, 1, 2, 3, 4+) [19], comorbid conditions [attention deficit hyperactivity disorder (ADHD), allergic rhinitis, anemia, anxiety, cancer, bronchopneumonia, cardiac disease, cerebrovascular disease, depression, diabetes, hypertension, obesity, other vascular diseases, pneumonia, pulmonary hypertension, and respiratory infection], medication history [ICS, short-acting β2 agonist (SABA), LABA, fixed-dose ICS/LABA and oral corticosteroids (OCS)], type of MDI (ICS or fixed-dose ICS/LABA), history of respiratory support including oxygen therapy and mechanical ventilation, history of severe exacerbations, history of moderate exacerbations, and history of all-cause hospitalizations. Patients without appropriate matches were not included in the analysis.
Treatment outcomes included number of patients with asthma exacerbations (moderate to severe); number of exacerbations per patient at 1, 6, 9, and 12 months after the index date; and time (days) to the first exacerbations. These measures were reported for patients with at least 12 months of post-index follow-up. Exacerbation severity was defined according to the literature [20,21,22]. A moderate exacerbation was defined as an ED visit claim (not leading to a hospital admission) with an asthma diagnosis, or an OCS prescription fill within 30 days of a physician visit with an asthma diagnosis; a severe exacerbation was defined as an inpatient admission claim with an asthma diagnosis. Furthermore, the incidence rate (IR) of exacerbations (either moderate or severe) was measured for patients with at least 30 days of post-index follow-up. The IR was defined as the number of patients with an event divided by total time spent at risk, which was from the index date to the end of enrollment, the end of the study period, discontinuation of index treatment (at least 60 days between the end of the supply and the next prescription fill), or the occurrence of an event—whichever occurred first during the variable follow-up. Asthma-related healthcare resource use measures included hospitalizations, outpatient visits, ED visits, laboratory tests, and ancillary and other services based on medical claims with an asthma diagnosis as well as prescriptions for an asthma medication. Asthma-related costs associated with these resources were also reported. Healthcare resource use and cost measures over 12 months of follow-up were reported for patients with at least 12 months of post-index follow up. The use of a fixed follow-up period was required to ensure a fair comparison of these measures between the two cohorts.
Baseline patient characteristics were reported using descriptive statistics, with frequency (n, %) for categorical measures and mean, standard deviation (SD), median, minimum, and maximum values for continuous variables. Treatment outcomes, healthcare resource use measures, and related costs in the AC+FV AVHC cohort and matched control VHC cohort were compared using bivariate chi-square tests for proportions and Student’s t test for means. The numbers of moderate-to-severe exacerbations per patient in the AC+FV AVHC and control VHC cohorts were also compared using rate ratios, defined as the number of moderate-to-severe exacerbations per patient in the antistatic VHC cohort divided by the number of moderate-to-severe exacerbations per patient in the control VHC cohort. The proportions of patients without an exacerbation in the AC+FV AVHC cohort and the control VHC cohort were compared using a Kaplan–Meier estimation of time to first occurrence of a moderate-to-severe exacerbation. In addition to the t test of differences in the mean costs of the two matched sample, a regression analysis was performed in the unmatched population to confirm the results of the bivariate analysis of healthcare costs in the matched population. In the regression analysis, the marginal effect of AC+FV AVHC on per-patient costs for moderate-to-severe exacerbations was measured using generalized linear models (GLMs) with a log link and gamma family distribution, along with covariate adjustments to control for potential confounders. All inferential statistical analyses were conducted assuming a two-tailed test of significance and a alpha level set a priori at 0.05. All analyses were conducted using SAS 9.4 (Cary, NC, USA).
No institutional review board (IRB) review was required for this retrospective cohort analysis using HIPAA-compliant de-identified patient data.