ABSTRACT
The purpose of sensitivity analysis in health economic evaluation is to assess the uncertainty of conclusions (such as those based on the incremental costeffectiveness ratio [ICER]) by varying the model inputs. Data from a single clinical trial might be of limited value, especially if it is the only trial that has been carried out for the disease under investigation (or if it is unlikely that a similar trial will be conducted again in the foreseeable future). For example, assuming that the treatment-related adverse event rate is 5% for a particular treatment group, but the reliability of the rate is questionable (because, for example, the trial was unblinded, and hence any assessment of treatment relatedness is biased), one might assess the impact on the ICER if the adverse event rate were larger or smaller than 5%.
