Identification of Minimum Effective Dose based on Ratio of a Normally Distributed Data under Heteroscedasticity

Abstract

Efficacy and safety study is of practical importance in modern drug development. It is a key component in evaluating the safety of food additives or pesticides, and assessing the effectiveness and safety of drugs. In most of the various statistical procedures, homogeneity of variances among different dose levels was required. This paper without a need for multiplicity adjustment proposes a stepwise confidence set procedure for estimating Minimum Effective Dose (MED) of drugs based on ratio of population means for normally distributed data under heteroscedasticity. The procedure employed Fieller’s (1954) method and obtained individual confidence intervals for identification of MED. The procedure is applied to a data of an experiment that was published by Ruberg (1989) where the effect of a new compound is measured by an increase in the weight of a particular organ in mice. Simulation study was carried out and results indicate that the procedure controls the family-wise error rate (FWER) strongly. Power of the procedure increases with increasing ratio of means and sample size.