Optimal designs for composed models in pharmacokinetic-pharmacodynamic experiments

Abstract

We discuss design issues for pharmacokinetic and pharmacodynamic (PK/PD) models and provide closed form descriptions for locally optimal designs for estimating individual parameter in two frequently used models. We propose standardized maximin optimal designs that remove dependence on the particular parameter of interest by maximizing the minimal efficiency across all parameters. Further, robust designs are proposed to overcome the dependence on the parameters of interest and the nominal values of the parameters. We compare performance of these optimal designs with designs used in four real studies from the pharmacokinetic/pharmacodynamic literature and show that our proposed designs provide definite advantages over those used in practice.