In recent years, cost-effectiveness analysis has become a frequent component of randomized clinical trials. In statistical terms, the major efforts addressed the method for estimating the Incremental Cost Effectiveness Ratio (ICER) and its confidence interval both with parametric and non-parametric methods. The goal of the present work is to briefly present the main non-parametric methods, based on a bootstrap approach. The methods which have been considered were applied to the data of a randomized clinical trial comparing two alternative approaches to treat hepatocellular carcinoma. The example shows that the application of different methods leads to significantly different confidence intervals' estimates, with consequences on the interpretation of the study results.