COST-EFFECTIVENESS ANALYSIS AND SENSITIVITY ANALYSIS, A GUIDE FOR A PROBABILISTIC APPROACH: INTRODUCTION TO PROBABILISTIC SENSITIVITY ANALYSIS (PART I)

Abstract

This the first of two articles exploring the context of sensitivity analysesb in economic evaluation (Part I) and how it is carried out in Microsoft Excel (Part I). Cost-Effectiveness Analysis (CEA) and its variants have high level of uncertainty inherent to decision analytic models and data. Therefore, it is mandatory to assess robustness of results by means of sensitivity analysis. There are two widely applied methods in CEA: one-way deterministic analysis, which assess uncertainty of one single parameter each time; and probabilistic sensitivity analysis, which is capable to take into account uncertainty of all parameters at one time, varying parameter values through statistical distributions. In this first paper, we describe the concept of probabilistic sensitivity analysis in a pragmatic approach in the context of economic evaluation and illustrate its rationale behind the two most used statistical distributions, Beta and Gamma curves. We also explain how hyperparameters of each curve is calculated.