On evaluation of consistency in multi-regional clinical trials.

In recent years, multi-regional clinical trials (MRCT) that conduct clinical trials simultaneously in Asian Pacific region, Europe, and the United States have become very popular for global pharmaceutical development. The main purpose of multi-regional clinical trials is to shorten the time for pharmaceutical development and regulatory submission, and approval around the world. In practice, however, clinical results observed from some regions (sub-population) may not be consistent with the results from other regions and/or all regions combined (entire population). The inconsistency observed may be due to ethnic differences in different regions, differences in medical practice, time points of assessment, or by random chance due to small sample size for the region. Some regional regulatory agencies require consistency evaluation between local country results and overall results. However, the challenge is there is no detailed guidance on the definition of 'consistency' and methodology to evaluate it. Therefore, the questions are: how to evaluate consistency and what statistical methods are appropriate to be used for consistency evaluation? In this article, several statistical tests for consistency (similarity) between clinical results observed from a specific sub-population and the entire population are proposed. These methods are compared through extensive simulation. As most published articles discussed consistency evaluation for superiority situations, we have discussed consistency evaluation for non-inferiority situation in this article through a simulated example concerning consistency in some countries. Recommendations of the statistical methods to be used for consistency evaluation are given. Other aspects that should be considered for consistency evaluation are also provided.