Estimating the Cost-Effectiveness of Fluticasone Propionate for Treating Chronic Obstructive Pulmonary Disease in the Presence of Missing Data

Authors


Andrew Briggs, Public Health & Health Policy, University of Glasgow, 1 Lilybank Gardens, Glasgow G12 8RZ, UK. E-mail: a.briggs@clinmed.gla.ac.uk

ABSTRACT

Objectives:  To explore the cost-effectiveness of fluticasone propionate (FP) for the treatment of chronic obstructive pulmonary disease (COPD), we estimated costs and quality-adjusted life-years (QALYs) over 3 years, based on an economic appraisal of a previously reported clinical trial (Inhaled Steroids in Obstructive Lung Disease in Europe [ISOLDE]).

Methods:  Seven hundred forty-two patients enrolled in the ISOLDE trial who received either FP or placebo had data available on health-care costs and quality of life over the period of the study. The SF-36-based utility scores for quality of life were used to calculate QALYs. A combined imputation and bootstrapping procedure was employed to handle missing data and to estimate statistical uncertainty in the estimated cumulative costs and QALYs over the study period. The imputation approach was based on propensity scoring and nesting this approach within the bootstrap ensured that multiple imputations were performed such that statistical estimates included imputation uncertainty.

Results:  Complete data were available on mortality within the follow-up period of the study and a nonsignificant trend toward improved survival of 0.06 (95% confidence interval [CI]–0.01 to 0.15) life-years was observed. In an analysis based on a propensity scoring approach to missing data we estimated the incremental costs of FP versus placebo to be £1021 (95% CI £619–1338) with an additional effect of 0.11 QALYs (CI 0.04–0.20). Cost-effectiveness estimates for the within-trial period of £17,700 per life-year gained (£6900 to ∞) and £9500 per QALY gained (CI £4300–26,500) were generated that include uncertainty due to the imputation process. An alternative imputation approach did not materially affect these estimates.

Conclusions:  Previous analyses of the ISOLDE study showed significant improvement on disease-specific health status measures and a trend toward a survival advantage for treatment with FP. This analysis shows that joint considerations of quality of life and survival result in a substantial increase in QALYs favoring treatment with FP. Based on these data, the inhaled corticosteroid FP appears cost-effective for the treatment of COPD. Confirmation or refutation of this result may be achieved once the Towards a Revolution in COPD Health (TORCH) study reports, a large randomized controlled trial powered to detect mortality changes associated with the use of FP alone, or in combination with salmeterol, which is also collecting resource use and utility data suitable for estimating cost-effectiveness.

Ancillary