Efficient designs for the estimation of mixed and self carryover effects

Abstract

Biosimilars are copies of biological medicines that are developed by a competitor after the patent for the originator drug has expired. Extensive clinical trials are required to show therapeutic equivalence between the biosimilar and its reference product before a biosimilar can be sold on the market. However, even after more than 10 years of experience with biosimilars in Europe, there is still some uncertainty if the patients who are already taking the reference product can switch between the biosimilar and its reference product. One convenient way to assess the impact of switches is the analysis of mixed and self carryover effects: if the products are switchable, there should not be any difference in the carryover effects. This paper determines a series of simple designs which are highly efficient for the comparison of the mixed and self carryover effects of two treatments. The proof of efficiency is not straightforward because the information matrix of the efficient designs is not completely symmetric.