Analytical approach:
The effectiveness and cost data were from one randomised controlled trial. The time horizon was nine months (the follow-up period of the trial). The authors stated that the perspective of the UK NHS was adopted.
Effectiveness data:
The effectiveness data were from the General Practice Research Framework Nasal Steroids for Otitis Media with Effusion (GNOME) trial (Williamson, et al. 2009, see 'Other Publications of Related Interest' below for bibliographic details). This was a placebo-controlled, double-blind, randomised trial of 217 children aged four to 11 years, with a history of otitis media and tympanometry-confirmed bilateral OME. The patients were recruited from 76 Medical Research Council General Practice Research Framework practices in the UK, with 105 patients randomised to mometasone furoate and 112 to placebo. The primary outcome measure was the proportion of children who were cured of bilateral OME, using tympanometry criteria (at least one ear with an A or C1 type recording) at one month.
Monetary benefit and utility valuations:
The utilities were elicited using the Health Utilities Index (HUI) versions two and three and the European Quality of life (EQ-5D) questionnaire. All questionnaires were completed by parents at baseline, and at three and nine months. The base-case analysis used the HUI3. The utility scores were adjusted to correct for an imbalance in baseline utility values between the study arms, by subtracting the baseline scores from those at three and nine months; each measurement was the increase (or decrease) in utility from baseline.
Measure of benefit:
The measure of benefit was quality-adjusted life-years (QALYs) gained.
Cost data:
The direct costs included consultations with community health-care providers, medications, investigative tests, hospital out-patient care, and hospital in-patient stay. The resource use data were collected in two ways; the research nurses retrospectively extracted information from children’s primary care records, and parents completed resource use questionnaires at three and nine months. In the base case, information from primary care records was used. The resource use was valued using NHS Reference Costs, local provider tariffs, the British National Formulary, and other national unit cost sources. All costs were reported at 2006 to 2007 prices, in UK pounds sterling (£).
Analysis of uncertainty:
Missing data were imputed using several methods to avoid bias from complete-case analysis. Bootstrap replications of 1,000 to 5,000 simulations were generated for the imputed datasets and cost-effectiveness acceptability curves were generated for positive incremental net benefits. Independent-sample t-tests were used to test for differences in resource use, costs, utility scores, and QALYs between treatment groups. The expected value of perfect information (EVPI) was estimated to assess the theoretical maximum that society could spend on further research. Sensitivity analyses were conducted: using the resource responses from parents; analysing complete cases; using HUI2 and EQ-5D utilities; and not adjusting for baseline utility. Post-hoc subgroup analyses were undertaken, including by age, gender, and atopy.