Health economics evaluations are affected by uncertainty when estimating their parameters. Therefore, it is important that we use a sensitivity analysis to determine the robustness of these evaluations. Most countries' guidelines recommend using a probabilistic sensitivity analysis (PSA), which enables us to evaluate the uncertainty of multiple parameters at the same time, based on a joint probability distribution. In this article, we first introduce the Monte Carlo simulation and Bootstrap method as PSA methods. Then, we review how various guidelines incorporate the PSA. Finally, we review Japanese health economics studies to determine the level of PSA use in Japan. Guidelines published before 2008 recommend conducting a sensitivity analysis, but do not specify a method. In contrast, both the French, American and English guidelines, which were published after 2011, specifically recommend using a PSA. In Japan, the “Guideline for economic evaluation of healthcare technologies in Japan” recommends conducting a PSA, “if possible”. However, PSA methods are not widely used in Japan. Of 49 Japanese health economics studies based on quality-adjusted life years, only six conducted a PSA (12.2%), although 35 (71.4%) did conduct other types of sensitivity analyses. If PSA methods are accepted as a good way to determine the robustness of an evaluation, then we need to foster their use. This, in turn, means we need specific guidelines on how best to use these methods.