All Times EDT
Keywords: Bayesian hierarchical model, Local treatment effect, Heterogeneity, Consistency assessment
A multi-regional clinical trial, conducted in different regions under one protocol, is a gold standard in global drug development. In addition to positive benefit-risk in the overall population, evaluation of the local treatment effect and assessing regional consistency are important components for labeling discussions with the local health authorities. However, relatively small sample sizes in some regions lead a lower precision for the local treatment effect estimates. One strategy is to borrow the results from other regions to a specific region while estimating the region-specific treatment effect. Commonly used methods are complete pooling of data from specific regions and Bayesian hierarchical model. For example, pooling patient data from all Asia region to support effect seen in Japanese patients provided the medical practice is consistent. However, such approach ignores heterogeneity between different regions and cause an unrealistic estimate. On the other hand, Bayesian hierarchical model considers the between-region heterogeneity and provide shrinkage estimates for region-specific treatment effect. But shrinkage estimator is often criticized to provide too optimistic borrowing for outliers. In the talk, a robust meta-analytic approach for estimating region-specific treatment effects will be discussed. While improving the precision, the proposed estimate is less influenced by other regions and provide a reasonable estimation of treatment effect. The proposed methodology will be illustrated by simulation study and a real-life example.