Since we presented a new statistical method for sample-size calculation assuming constant cost in a cost-effectiveness trial, the use of the method has been under evaluation these two years. OBJECTIVE: To evaluate the new method using retrospective evidence in terms of applicability and validity in real cases, and then clarify the advantage and the limitation for users of the method. METHOD: We searched and reviewed the literature, published in English journals, on pharmacoeconomics trials, which reported definite research design with respect to clinical effects and the associated costs. Subsequently we checked whether or not the candidate articles satisfy the required conditions in theory to apply our method, and eventually identified the articles that can show the advantage and the limitation of our model. RESULTS: Three articles were extracted: one for beneficial and two for not beneficial. In the beneficial case, our method was able to reduce the sample-size favorably from 3000 to 28 subjects regarding the same cost-effectiveness ratio given in the article. One of not-beneficial cases showed no change of sample size with about 100 subjects in both calculations whether new or classical. In another case of not-beneficial, the sample-size newly estimated was still practically large with 743 subjects since the original trial theoretically required a large value of sample-size with 26,252 subjects according to a very small effect-size of the trial. The evaluation on the validity of assumption for constant cost was left for future investigation due to lack of raw data in cost.  CONCLUSION: The new method showed that it could be beneficial for reducing the sample-size required in a cost-effectiveness trial, if we carefully estimate the expected cost-effectiveness ratio attained by the trial to be designed in advance.