Drug review procedures and drug development strategies are changing rapidly now due to the "The International Conference on Harmonization of Technical Requirements for the Registration of Pharmaceuticals for Human Use" (ICH). ICH started in 1990 in order to improve the efficiency of the drug review process for promising new drugs. Particularly, the E5 guidelines (for extrapolation of foreign data to new regions in new drug applications) is having a very significant impact on new drug development in Japan.

In several stages of bridging studies, we consider the one to show similar efficacy of the new drug in a new region to the one derived from several phase II/III studies performed outside of the given region. The bridging study has two aspects to be considered ﾐdesign issues and statistical issues. The design issues are related to extrinsic matters, such as different medical practices, local primary endpoints, and culturally different attitudes to particular questionnaires, etc. in the new region. The statistical issues are related to an estimation problem, not to a hypothesis testing problem that usually receives the most attention in the drug review process in regulatory agencies. Regarding the estimation problem, first we have to know the effectiveness of the drug estimated from the large pre-existing outside data. This is related to the meta analysis issue ﾐ fixed effects vs. random effects of the studies or of the centers. Second, in order to calculate the required sample size of the bridging study in a protocol, we have to define the ﾒsimilarityﾓ of the efficacy of the drug estimated from the data in the new region with the estimated one from the outside data. We will investigate the statistical properties related to these issues.

Key words: Bridging study, Similarity, Meta analysis