Abstract: Biosimilars are copies of biological medicines developed after the patent for the originator drug (the reference product) has expired. Extensive clinical trials are required to show the therapeutic equivalence of the biosimilar and its reference product before the biosimilar can be sold on the market. However, even after more than 10 years of experience with biosimilars, there is still uncertainty whether patients can switch between the biosimilar and its reference product without negative effects. One convenient way to assess the impact of switches is to analyze their mixed and self carryover effects: if the products are switchable, there should be no difference between the carryover effects. For p = 3 periods (and the number of subjects is divisible by 8) and for p ≡ 1 mod 4 periods (and the number of subjects is divisible by 4), determine a series of simple designs that efficiently compare the mixed and self carryover effects of two treatments. The proof of the efficiency is not straightforward, because the information matrices of the efficient designs are not completely symmetric.

Key words and phrases: A-optimality, biosimilars, crossover designs.