ABSTRACT
Statisticians have developed and promoted many methods for designing and analyzing dose response studies. This chapter summarizes some meta-analyses of dose response studies and describes parametric Emax dose response model. It presents the design of a dose response trial for a continuous endpoint following a successful proof of concept study comparing a single high dose to placebo. The chapter discusses planning of a phase II dose response study using Bayesian methods to quantitatively assess the properties of potential designs and analyses. Model-based Bayesian and maximum likelihood estimators are evaluated. Simple pairwise comparisons of dose group sample means are reported as a standard benchmark. The example is based on a small-molecule compound in phase II development for the treatment of a degenerative chronic disease. The design and various analyses were evaluated over a factorial collection of model parameters yielding different dose response curves.
