ABSTRACT
This chapter introduces the original MCP-Mod approach for normally distributed data as well as its generalization for general parametric models. The MCP-Mod approach is a multistage procedure that combines the advantages of the multiple comparisons and modeling approaches. The chapter illustrates the DoseFinding R package, which provides a convenient interface to the general approach adopted here. The findings of the dose–response analysis may then form the quantitative basis for selecting the dose for the Phase III program. The chapter describes the individual steps of the MCP-Mod approach for normally distributed data, and then an extension to more general distributions. It discusses practical considerations when implementing MCP-Mod. These are grouped into three sets of topics: general considerations on when to apply MCP-Mod, considerations at the trial-design stage, and considerations at the trial-analysis stage. The chapter focuses on dose and candidate model selection at the trial design stage, as well as various miscellaneous topics.
