Reply: Topical Silicone Gel versus Placebo in Promoting the Maturation of Burn Scars: A Randomized Controlled Trial-The Pivotal Role of Statistics.

摘要

We agree with Dr. Mangano that an accurate experimental design is essential for drawing valid conclusions from any clinical trial and that the required sample size should be calculated and fixed before starting a trial. We did perform an a priori sample size calculation but unfortunately this was not mentioned in the article. Based on data from previous studies, a difference in redness scores at follow-up of approximately the size of the score's standard deviation was anticipated. Comparing the follow-up measurements under the two conditions using a standard t test would then require 36 scars to be treated and evaluated (18 per treatment) to achieve 80 percent power when tested two-sided at a significance level 5 percent.