The statistical issues in irritation/sensitization study for transdermal drug products are very challenging. The purpose of irritation/sensitization study is to ensure that, in addition to the bioequivalence between the test product (TEST) and the reference listed drug (RLD) demonstrated by a comparative pharmacokinetic study, the skin irritation/sensitization is not inferior to that of the RLD. In this type of study, irritation and sensitization are assessed based on the skin irritation score (an integer ranging 0-10). The analysis endpoint is the mean skin irritation score obtained by averaging the irritation scores recorded during the evaluation period. In this presentation, we will use historical data and simulation studies to investigate issues in analyzing irritation/sensitization study, such as the appropriate comparison scale, type of outcome (binary or continuous) and best statistical model.