OBJECTIVES: In preparation for exploring the potential cost-effectiveness of inhaled corticosteroids (ICS) for the treatment of chronic obstructive pulmonary disease (COPD), we explored the costs and outcomes of patients with COPD in the presence of missing data. METHODS: Using data on 751 patients enrolled in the ISOLDE randomized controlled trial that received either fluticasone propionate or placebo, the outcomes studied were COPD exacerbations, successfully treated weeks, rate of decline in FEV1, costs, survival and health status. In line with the original clinical analysis, costs and outcomes were estimated using a hierarchical multilevel model assuming data were missing at random. Further analyses explored the use of quality-adjusted survival techniques for health outcomes as well as multiple imputation methods to address missing data issues. RESULTS: In the base case analysis we estimated the incremental costs of ICS versus placebo to be £900 (95% confidence interval (CI): 100; 1,700) and the additional effect of 0.59 (CI: 0; 1.3) exacerbations avoided and 3 (CI: 1.7; 4.5) successfully treated weeks. We explored the robustness of these results to alternative assumptions in extensive sensitivity analysis. While survival was higher in the ICS group, the difference in mortality rate was not statistically significant (Hazard Ratio: 0.77 CI: 0.52; 1.15). However, there was a significant quality-of-life benefit in favor of ICS. CONCLUSIONS: Despite a lack of significance in survival benefits, joint considerations of quality of life and survival indicate that ICS could be considered potentially cost-effective. Imputation methods can be employed to address missing data issues when the extent of missingness is not too extreme.