[Clinical trials in dermatology. Statistical support]

To determine the efficacy or safety of new topical treatments, more or less standardized studies must be performed. This article presents several examples of statistical analyses of the data currently seen in the dermatology area, and some guidelines for study sample size determination. More importantly, it indicates, for as soon as phase I studies, the need for preliminary informations such as development strategy, historical data, literature references or pilot studies, to determine or adapt appropriate designs and sample sizes. This need can be filled by close collaboration among clinicians, statisticians and sponsor.