In recent years, the possible influence of ethnic factors on clinical outcomes for evaluation of efficacy and safety of study medications under investigation has attracted much attention from both the pharmaceutical/biotechnology industry and the regulatory agencies such as the United States (US) Food and Drug Administration (FDA), especially when the sponsor is interested in bringing an approved drug product from the original region such as the US or European Union (EU) to a new region (e.g., Asian Pacific Region). However, the key issues lie on when and how to address the geographic variations of efficacy and safety for the product development.
After a pharmaceutical product has been approved for commercial marketing in one region (e.g., the US or EU) based on its proven efficacy and safety, the pharmaceutical sponsor might seek registration of the product in a new region (Asian Pacific Region). However, the differences in race, diet, environment, culture, and medical practice among regions may have an impact on the extrapolation of the clinical outcomes from the original region to the new region. To address this issue, the International Conference on Harmonisation (ICH) published a guideline entitled
“Ethnic Factors in the Acceptability of Foreign Clinical Data” in 1998. This guideline is known as ICH E5 (ICH, 1998). The ICH E5 guideline provides a general framework for evaluation of the impact of ethnic factors on the efficacy, safety, dosage, and dose regimen.
As indicated in the ICH E5 guideline, a bridging study is defined as a study performed in the new region to provide pharmacokinetic (PK), pharmacodynamic (PD), or clinical data on efficacy, safety, dosage, and dose regimen in the new region that will
allow extrapolation of the foreign clinical data to the population in the new region.
The ICH E5 guideline suggests the regulatory authority of the new region to assess the ability to extrapolate foreign data based on the bridging data package, which consists of (1) information including PK data and any preliminary PD and dose-response data from the complete clinical data package (CCDP) that is relevant to the population of the new region and if needed, (2) bridging study to extrapolate the foreign efficacy data and/or safety data to the new region. The ICH E5 guideline indicates that bridging studies may not be necessary if the study medicines are insensitive to ethnic factors. For medicines characterized as insensitive to ethnic factors, the type of bridging studies (if needed) will depend upon experience with the drug class and upon the likelihood that extrinsic ethnic factors could affect the medicine’s safety, efficacy, and dose-response. On the other hand, for medicines that are ethnically sensitive, bridging study is usually needed since the populations in two regions are different. In the ICH E5 guideline, however, no criteria for assessment of the sensitivity to ethnic factors for determining whether a bridging study is needed are provided. Moreover, when a bridging study is conducted, the ICH guideline indicates that the study is readily interpreted as capable of bridging the foreign data if it shows that dose-response, safety, and efficacy in the new region are similar to those in the original region. However, the ICH does not clearly define the similarity.
Shih (2001) interpreted similarity as consistency among study centers by treating the new region as a new center of multicenter clinical trials. Under this definition, Shih (2001) proposed a method for assessment of consistency to determine whether the study is capable of bridging the foreign data to the new region. Alternatively, Shao and Chow (2002) proposed the concepts of reproducibility and generalizability
probabilities for assessment of bridging studies. If the influence of the ethnic factors is negligible, then we may consider the reproducibility probability to determine whether the clinical results observed in the original region are reproducible in the new region.
If there is a notable ethnic difference, the generalizability probability can be assessed to determine whether the clinical results in the original region can be generalized in a similar but slightly different patient population due to the difference in ethnic factors.
Lan et al. (2005) introduced the weighted Z-tests in which the weights may depend on the prior observed data for the design of bridging studies. Note that other methods such as based on similarity in terms of equivalence and non-inferiority have also been proposed in literature (Chow, Shao, and Hu, 2002; Hung, 2003).
One of the crucial reasons for the ICH E5 guideline to emphasize on minimizing unnecessary duplication of generating clinical data in the new region is that sufficient information on efficacy, safety, dosage and dose regimen has been already generated in the original region and is available in the CCDP. One should therefore borrow
“strength” from the information on dose response, efficacy, and safety from the CCDP in the original region and incorporate them into the analysis of the additional data obtained from the bridging study. Liu, Hsiao, and Hsueh (2002) have proposed a Bayesian approach to synthesize the data generated by the bridging study and foreign clinical data generated in the original region for assessment of similarity based on superior efficacy of the test product over a placebo control. However, the results of the bridging studies using this approach will be overwhelmingly dominated by the results of the original region due to an imbalance of sample sizes between the regions.
In other words, it is very difficult, if not impossible, to reverse the results observed in the original region even the result of the bridging study is not consistent with those of
the original region. However, this issue will occur for any methods for cross-study comparisons if the amount of information is seriously imbalanced between studies. To conquer this problem, Hsiao et al. (2007) then proposed a Bayesian approach with the use of a mixture prior for assessment of similarity between the new and original region based on the concept of positive treatment effect. For both approaches, even if both regions have positive treatment effect, their effect sizes might in fact be different.
That is, their approach could not truly assess the similarity between two regions.
Therefore, in this thesis we propose a Bayesian approach in which similarity between the new and original region will be concluded if the treatment effect for new region is more than a fraction of the treatment effect for original region trial.
This thesis is organized as follows. In Section 2, a Bayesian consistency approach for assessment of similarity is suggested. The method for sample size determination is given in Section 3. A numerical example is presented in Section 4 to illustrate the Bayesian approach. Discussion and final remarks are given in Section 5.