Modern Approaches to Clinical Trials Using SAS: Classical, Adaptive, and Bayesian Methods

7.2 Methodology

7.2.1 Original MCP-Mod Approach

In this section we first provide a short, technical description of the original MCP-Mod approach [2] for a single, normally distributed efficacy endpoint. Afterwards, we describe a recently developed variant of MCP-Mod that extends the core methodology to dose-finding studies for general parametric models (e.g. for binary, count, or time-to-event endpoints and including certain longitudinal modeling approaches), see [4].

The MCP-Mod procedure is implemented in five steps; see also Figure 7.1 High-Level Overview of the Original MCP-Mod Procedure :

1. At the trial design stage, identify several candidate dose-response shapes, which are likely to represent the underlying true dose-response.