Testing for Similarity of Dose Response in Multiregional Clinical Trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacometrics

Background:

• Ensuring subgroup dose-response similarity is crucial for multiregional clinical trials.
• Assessing dose-response curves across diverse populations requires robust statistical methods.

Purpose of the Study:

• To develop and validate statistical tests for assessing dose-response similarity between subgroups and the full population.
• To provide methods for comparing single or multiple subgroups against the overall trial population.

Main Methods:

• A parametric framework is employed for dose-response modeling.
• Two novel bootstrap tests are developed to assess curve similarity.
• Maximal deviation between dose-response curves serves as the primary metric for similarity assessment.

Main Results:

• The proposed bootstrap tests are proven to be statistically valid.
• Simulation studies demonstrate favorable finite sample properties of the tests.
• The methodology is illustrated using a practical case study.

Conclusions:

• The developed bootstrap tests offer powerful tools for evaluating dose-response similarity in clinical trials.
• These methods enhance the reliability of subgroup analyses in multiregional studies.
• The study provides a validated framework for subgroup comparisons in pharmaceutical development.