Sample Size Allocation in Multiregional Dose-Finding Study Using MCP-Mod

Abstract

Simultaneous global drug development with multiregional studies is becoming a common strategy for increasing efficiency of the development process. In particular, multiregional dose-finding studies can be informative to identify inter-ethnic difference in dose-response relationships early in development. An application of MCP-Mod to multiregional studies is discussed in this article. We consider sample size allocation to one specific region and provide three methods for demonstrating consistency in the dose-response relationship between the entire population and one specific region: two methods use contrast statistics to show consistency in dose-response signals, and the third method uses the maximum absolute difference between two dose-response curves to show consistency in the dose-response profiles. The proposed methods do not require studies to have sufficient power to detect a truly consistent dose-response relationship in a confirmatory manner, but rather to allow for quantitative design considerations that can ensure such a relationship is observed at the end of the study with acceptable probability. The three methods are illustrated through an anti-anxiety drug example, resulting in a recommended proportion of 20% for subjects in one specific region. The illustration indicates the recommended proportion could vary depending on the total study sample size.

Keywords:

• Consistency
• MCP-Mod
• Multiregional dose-finding study
• Sample size allocation

Supplementary Materials

Supplementary materials include additional figures for illustration.

Acknowledgments

We would like to thank Richard Croy for many helpful discussions. We are also grateful to two referees and an associate editor for their valuable comments, which improved the presentation of this article.