OBJECTIVES: To assess uncertainty in a cost-utility analysis (CUA) of adjunct entacapone treatment with levodopa among patients with Parkinson’s disease (PD). The purpose of the study was to apply probabilistic sensitivity analysis in the comparison of alternative treatment strategies using second-order simulation methods. METHODS: Two treatment alternatives of PD, i.e. levodopa with or without entacapone, were compared in a cost-utility analysis. A Markov model was constructed based on data from phase III clinical trials of entacapone and a naturalistic health economic study of PD. Second order simulation and bootstrap methods were employed to provide understanding of the uncertainty due to sampling variation. Cost and utility parameters were drawn from empirical distributions. Parametric distributions were used in the generation of transition probabilities. RESULTS: Using a bootstrap sample size of 200 and 1000 patients, joint distribution of the mean incremental costs and mean incremental utilities were calculated and displayed in the cost-utility plane. The results for a bootstrap sample of 1000 patients were all clustered in the quadrant IV that includes situations in which entacapone treatment yields gain in QALYs and cost savings. However, there was more variation with the sample of 200 patients. 85.4% of the bootstrap replications were in quadrant IV. 12.1% of the joint distribution fell into quadrant III indicating cost savings at the expense of loss in QALYs. Gain in QALYs at extra costs resulted in 2.1% of the observations. Only 0.4% of the simulated results indicated less QALYs and increased costs. CONCLUSIONS: The simulation methods used provided valuable information on the sensitivity of the results of the CUA. The probabilistic sensitivity analysis used in this study strengthened confidence in the conclusions that entacapone as an adjunctive treatment to levodopa is both cost saving and increases the quality of life of PD patients.