Probability of asthma exacerbations

A post-hoc analysis was carried out to estimate the probability of an asthma exacerbation using a database of a population-based study by Dennis et al that reported that 43% (95% CI, 36.3-49.2) of patients with current asthma symptoms reported requiring an ED visit or hospitalization in the last year (28). Patients with 5-59 years of age with a physician diagnosed asthma or rhinitis were selected. The frequency of nocturnal symptoms (i.e., <2 nights/month, 2/month, 1-3 times/week and every night) was used to classify patients among different severity categories. A previous diagnosis of asthma or rhinitis by a physician was established as a selection criterion for a better approximation of patients that were undergoing pharmacological treatment and thus reflect the baseline scenario of routine care with ICS in Colombia.
Two logistic regression models (one for patients with asthma and other for patients with asthma and rhinitis) were constructed using patient demographics and categories of nocturnal symptoms frequency as predictors and the history of an ED visit/hospitalization as outcome (yes/no). Coefficients were converted to probabilities and the estimated parameter for the “1-3 times/week” category in the nocturnal symptoms variable (using the “<2/per month” as reference level) was assumed to reflect the baseline probability of an exacerbation in patients with moderate persistent asthma. The resulting probabilities were converted to rates and re-expressed as 3-month probabilities for inclusion in the model.
A 75.4% of reduction in the proportion of patients that reported unscheduled medical visits over a 9-month period reported by El-qutob et al. was used in the model to reflect the effect of SCIT+ICS in the reduction in the probability of asthma exacerbations (29). This parameter was applied to the estimated baseline probabilities of an asthma exacerbation of 0.331 and 0.465 for AA and AA+AR patients, respectively (Table 1).

Health care resource utilization and costs

The considered costs were cost of medications, medical services (outpatient visits and specialized care) and ambulatory services (i.e., laboratory/image procedures). Costs per year were calculated by multiplying individual costs inputs with age-specific medication doses and medical services frequency considered to be appropriate to achieve disease control based in local and international clinical guidelines (18,30,31) (Table 1). For GINA Step 3 and GINA Step 2 states the recommended average daily dose of ICS (medium and low dose per day) for patients between 6-11 and >=12 years in the GINA report were used, respectively (5). The number of days with salbutamol use per month was obtained from a study by Sánchez et al and were multiplied with the daily dose of salbutamol used in the analysis (32). Frequency of outpatient, specialized care, ED visits and hospitalizationsper year was obtained from a study by Florez et al (3).
The administration of nasal ICS in addition to loratadine was allowed for patients with AA+AR according to the Allergic Rhinitis and its Impact on Asthma (ARIA) guidelines (30). For the GINA Step 3 and GINA Step 2 states, medium and low doses of nasal ICS were used, respectively. Loratadine was withdrawn in the GINA Step 2 state and no use of AR medications were allowed in the remission state. It was assumed that the reduction of AR medications was positively associated to the reduction of AA medications. Disaggregated cost inputs according to age group, health state and type of exacerbation are displayed in the Supplementary material 1.
A complementary scenario analysis was conducted by evaluating the use of SCIT as add-on therapy with a low dose of ICS + long acting β2 agonists (LABAs) (i.e., formoterol/budesonide or salmeterol fluticasone) and salbutamol. This scenario was conducted to account for other controller strategies recommended in the Step 3 of the GINA report that are frequently used in the clinical practice. A similar effectiveness between BDP and formoterol/budesonide or salmeterol/fluticasone was assumed. LABA medications were discontinued in the GINA Step 2 and AR medications were considered to be administered under the same scheme adopted in the base-case scenario. Costs and HCRU parameters were validated by clinical experts in the research staff. Unit cost parameters for all medications were obtained from the Drug Price Information System of the Colombian Ministry of Health and Social Protection (SISMED). Costs for the SCIT and other medical services were obtained from the Social Security Institute medical fee manual of 2001+30% as recommended by Colombian guidelines for economic evaluations (33,34).

Utility values

Previous EuroQol-5D utility values reported by Szende et al. for intermittent, mild, and moderate severity levels in Hungary were used (35). Utilities were assigned to the asthma without medication, GINA Step 2 and GINA Step 3 states in the model, respectively. Disutility associated to exacerbations were obtained from a previous study by Lloyd et al. that reported changes in baseline EuroQol-5D utility values in patients from the UK (36).