Robust Designs in Noninferiority Three-Arm Clinical Trials With Presence of Heteroscedasticity

In this article, we describe an adjusted method to facilitate a noninferiority trial using a three-arm robust design. Because local optimal designs derived by Hasler et al. require knowledge about the ratios of the population variances and are not necessarily robust with respect to possible misspecifications, we adopt a maximin approach. This method requires only the specification of an interval for the variance ratios and yields robust and efficient designs. We demonstrate that a maximin optimal design depends only on the boundary points specified for the variance ratios' intervals and we obtain numerical and analytical solutions which are demonstrated in several examples. The derived designs are robust and very efficient for statistical analysis in noninferiority three-arm trials.