This retrospective, observational study compared cohorts of patients with SA-EP before and after initiating mepolizumab. Patients who initiated mepolizumab during the selection period (February 1, 2016, to March 31, 2019) were identified using the Canadian mepolizumab PSP and linked to the ICES database. Individual patient study outcomes were obtained for the 12 months pre- and post-mepolizumab initiation.

The study comprised one main cohort, the overall PSP population, and two subsets of the overall PSP population: (1) the provincial drug coverage subset (used to track utilization of related medications such as OCS/inhalers independently of reimbursement of mepolizumab through the ODB program); and (2) a second subset of patients who were adherent to treatment (≥ 9 mepolizumab prescriptions within 12 months post-index date).

The ICES captures all health services provided by the Ontario public system, and its data repository consists of de-identified, record-level, coded and linkable health data sets for as many as 13 million Ontarians. The records in ICES data reflect Ontarians’ day-to-day interactions with the healthcare system. These include: (1) physician claims submitted to the Ontario Health Insurance Plan; (2) medical drug claims based on prescriptions submitted to and reimbursed by the ODB program; (3) discharge summaries of hospital stays and ED visits; and (4) claims for home care, long-term care, and more.

The primary objectives of this study were: to describe the demographics and clinical characteristics of real-world patients receiving mepolizumab; to compare asthma-related outcomes before and during mepolizumab use; and to compare asthma-related costs before and during mepolizumab use. The secondary objective was to assess real-world asthma-related outcomes and associated costs for patients who were adherent to mepolizumab treatment (≥ 9 prescriptions in 12 months).

Asthma-related real-world outcomes of interest were: asthma exacerbations (defined as outpatient or ED visit with a diagnosis of asthma and at least one dispensing of systemic corticosteroids ± 5 days after the encounter or exacerbations resulting in hospitalization/inpatient hospital admissions with a primary diagnosis of asthma); physician visits (general practitioner [GP], specialist, and other outpatient visits); ED visits; inpatient hospitalizations; OCS use (provincial drug coverage subset only); and short-acting β-agonist (SABA) use (provincial drug coverage subset only).

Real-world outcome costs were: GP cost (asthma-related); specialist cost (asthma-related); ED cost (all-cause); hospital admission cost (all-cause); total cost (excluding all drug costs); drug cost (provincial drug coverage subset only); total cost (including drug cost; provincial drug coverage subset only). When reimbursed by the ODB program after March 2018, the cost of mepolizumab treatment was included in the provincial drug coverage subset costs. Due to resource intensity weighting, costs cannot be accurately estimated from specific hospital admission reasons; thus, any ED visit and inpatient hospitalization costs are not asthma-specific. Cost per service was collected from ICES, which collates data from the Ontario Health Insurance Plan database and covers all publicly funded healthcare billing.

The Canadian mepolizumab PSP was used to identify patients with mepolizumab utilization, who were then linked to the ICES database. Using deterministic linkage, the patients’ unique Ontario health card numbers connected PSP patients to the provincial ICES dataset which allowed tracking of patients’ healthcare system interactions through both the PSP and the healthcare system. The remaining patients were identified by probabilistic linkage using date of birth, sex, and postal code and linked to a corresponding patient in the ICES Ontario asthma cohort.

Patients were included in the study if they had provided informed consent, received ≥ 1 injection of mepolizumab within the selection period, initiated mepolizumab within the selection period, were identified within the ICES database, were ≥ 18 years of age at index, and were active within the ICES database for 1 year pre- and post-mepolizumab index date. Additional inclusion criteria for the provincial drug coverage subset included ≥ 1 ODB asthma claim in both the 12-month pre- and post-mepolizumab analysis periods, and ≥ 1 ODB asthma claim in the 3 months before the pre-mepolizumab period and 3 months after the post-mepolizumab period.

We defined patients adherent to mepolizumab treatment as individuals within the PSP population who received ≥ 9 mepolizumab injections over a 1-year categorization period. This subgroup was further stratified into those who had an eosinophil level ≤ 300 cells/μL (PSP population only), those with ≥ 2 baseline exacerbations and eosinophil level ≥ 300 cells/μL (provincial drug coverage subset only), and those with ≥ 3 baseline exacerbations (provincial drug coverage subset only), where baseline eosinophil counts were recorded as the median eosinophil count in the 12-month pre-mepolizumab period. Additional subgroups were considered but were not evaluated due to low patient numbers.

For the sample size calculation, the minimum detectable reduction in mean exacerbations between pre- and post-mepolizumab initiation periods was used (see Supplementary Table 1, Additional file 1). For all analyses, missing data were reported as their own category, and no imputation was performed. All analyses were conducted using statistical analysis system (SAS) version 9.3 or higher.

Descriptive analyses were performed using the number of patients and percentage for categorical variables, and the mean (standard deviation [SD]) and median (interquartile range [IQR]) for continuous variables. As healthcare resource utilization (HCRU) outcomes and costs data were not continuous and bounded at zero, a distribution other than the normal was assumed. For HCRU, negative binomial distribution was used where variance > mean, or Poisson where mean ≤ variance; gamma distribution was used for cost. Corrections for multiple comparisons were made using a baseline significance level of α = 0.05.

The mean number of events observed in a category was presented for all patients and patients adherent to mepolizumab treatment, and the presence or absence of an event was compared pre- and post-mepolizumab initiation using McNemar’s test. Differences in HCRU counts between pre- and post-mepolizumab initiation (in the overall PSP population and the provincial drug coverage subset) were compared using an adjusted, generalized, linear mixed model with a negative binomial distribution. The model was used to control for covariates and model the HCRU count as the dependent variable, with patient as a random factor to account for the paired nature of the data. Mean costs were compared pre- and post-mepolizumab initiation for all patients (in the overall PSP population and the provincial drug coverage subset) using an adjusted mixed gamma model with patient as a random factor.

This study complied with all laws regarding patient privacy. No direct patient contact or primary collection of individual human subject data occurred. Patients enrolled or previously enrolled in the Canadian mepolizumab PSP program were given the opportunity to provide informed consent that allowed their patient-level data to be shared with ICES. This was conducted by the program coordinators via email, phone, or in person. All executed versions of the consent form and study protocol were approved by an Institutional Review Board (IRB; Pro00034156). These were first approved in May 2019 and continually reviewed every 12 months by the IRB.