Efficacy and safety results have been previously presented (6
Of the 2509 patients enrolled, 2143 were randomly assigned; 1067 to budesonide-formoterol and 1076 to salmeterol-fluticasone. (In Canada, 163 patients were enrolled and 146 patients were randomly assigned.) The baseline characteristics of the patients are shown in . There were no significant differences in age, sex, time since diagnosis or previous use of asthma medications between groups.
Baseline patient characteristics
Budesonide-formoterol used as maintenance and reliever therapy was more effective than salmeterol-fluticasone plus salbutamol, as measured by the primary outcome. The risk reduction for the time to the first exacerbation was 25% (hazard ratio 0.75; 95% CI 0.61 to 0.93; P=0.0076). There was a 22% reduction in the mean number of exacerbations with budesonide-formoterol (hazard ratio 0.78; 95% CI 0.66 to 0.91; P=0.0025). The annualized exacerbation rates were 0.24 events/patient-year and 0.31 events/patient-year, respectively ().
Clinical and resource use trial outcomes
The budesonide-formoterol group required a mean daily dose of 562 μg (maintenance) plus 91 μg of (as-needed) budes-onide, while patients in the salmeterol-fluticasone group used a mean of 583 μg of fluticasone (maintenance only). Expressed as equivalent beclomethasone dipropionate doses, this represented 1019 μg/day (maintenance and as-needed) in the budesonide-formoterol group and 1166 μg/day (maintenance only) in the salmeterol-fluticasone group (6
). At the end of the study, 68% of patients in the budesonide-formoterol group used a budesonide maintenance dose of 800 μg daily (delivered dose of 640 μg) while 31% received 400 μg daily (delivered dose of 320 μg). In the salmeterol-fluticasone group, 58% received the fluticasone dose of 500 μg daily, 27% received 1000 μg and 14% received 200 μg. It was assumed that the proportion of different strength of inhalers used throughout the trial corresponded to that reported at the end of the trial. Use of as-needed medication was lower in the budesonide-formoterol group (0.58 versus 0.93 inhalations/day; P<0.001).
Descriptive statistics for maintenance and as-needed asthma medication were also derived from collected inhalers (). Estimates of mean daily steroid dose based on collected inhalers were in agreement with the mean daily doses calculated from prescriptions and reported as-needed use.
Descriptive statistics for study medication derived from information on inhalers
For both treatment arms, relatively few patients used asthma medication other than study medication and oral steroids for exacerbations. Use of oral steroids due to severe asthma exacerbations was reported in 12% of budesonide-formoterol patients (1980 days of total use) and 14% of salmeterol-fluticasone patients (2978 days of total use). None of the differences in other drug uses between treatment groups was greater than 1%.
No clinically important differences between the two treatment groups were observed with respect to adverse events. Both budesonide-formoterol and salmeterol-fluticasone were well tolerated.
Health care resource use including hospitalization, ER visits, physician visits (general practitioner and specialist), other health care visits and home care visits is reported in . Sick leave and caregiver time identified as ‘days unable to perform usual activity’ are also reported ().
Summary of total health care resource usage throughout the study*
Unit costs for resources are presented in .
Although the physician panel identified the most commonly used concomitant medications within a given class, the cost of concomitant medications (with the exception of prednisone, oral 5 mg tablets, and the occasional use of additional salbutamol in the budesonide-formoterol group) was not included in the analysis. All other concomitant medication use occurred with a low frequency (1% or less), and the majority of use occurred in the salmeterol-fluticasone group. Therefore, the decision to forego inclusion of costs of concomitant medication favoured the salmeterol-fluticasone group.
In the base case analysis, total costs were summed for the budesonide-formoterol treatment arm and the salmeterol-fluticasone treatment arm, and divided by the number of patient-years for each arm. From the health care perspective, drug costs represented 94% of the total costs in the budesonide-formoterol group and 92% of the total costs in the salmeterol-fluticasone group. Mean costs per patient-year are represented in .
Total costs and mean cost per patient-year*
An incremental cost-effectiveness ratio was calculated by dividing the incremental cost by the incremental benefit. From both the health care perspective and the societal perspective, budesonide-formoterol was the dominant treatment strategy compared with salmeterol-fluticasone, meaning that its use resulted in fewer exacerbations at a lower cost (). It was not appropriate to present incremental cost-effectiveness ratios in cases of equivalency or dominance. In the case of equivalency, the incremental difference would be zero, resulting in an undefined number; in the case of dominance (more effective, less costly), the double negative (numerator and denominator) becomes a positive result and leads to an incorrect interpretation of findings.
Incremental cost-effectiveness ratio: Base case analysis
Several sensitivity analyses were conducted (). In the first sensitivity analysis, the annualized rate of exacerbations and the costs (excluding budesonide-formoterol and salmeterol-fluticasone) were varied using the upper and lower boundary of the 95% CI of the hazard ratio for the budesonide-formoterol arm (costs and consequences in the salmeterol-fluticasone group remained constant). Budesonide-formoterol remained dominant from both the health care and societal perspectives at the lower boundary. At the upper boundary, the incremental differences in costs and consequences, although still favouring budesonide-formoterol, were extremely small (0.02 fewer exacerbations per patient-year, and $6 less per patient-year from the health care perspective compared with $59 less per patient-year from the societal perspective). Consequently, this result at the upper boundary was judged to be equivalent.
Summary of univariate sensitivity analysis results
In the second sensitivity analysis, all salmeterol-fluticasone 500 μg/50 μg inhalers were assumed to be salmeterol-fluticasone 250 μg/50 μg and assigned a cost of $98.92 per claim (reduced from $138.52 per claim). In a third scenario, the cost assigned to absenteeism was set at the minimum wage ($7.45 per hour for Ontario), instead of the average wage. In a fourth scenario, the cost of hospitalization was calculated using the average asthma-related hospitalization costs from the Ontario Case Costing Initiative (14
) multiplied by the number of hospitalizations (reported in the clinical trial), in effect reducing the impact of hospitalization in this analysis. Regardless of the changes, budesonide-formoterol continued to be the dominant strategy from both societal and health care perspectives.