Bridging is the use of clinical data from one region to register drugs in another region. Efficacy and safety may differ between populations. Bridging requires extrapolation of evidence and a model linking the distributions in different populations. Absent such a model, complete data would be required in each region.
Regulators face a dilemma when they use data from other regions. If they require substantial research within their own country, the cost of drug approval may exceed its economic value. If they do not, they may approve ineffective or harmful drugs. Related issues arise in global drug trials.
We recommend a regional strategy; countries with similar ethnicity and environment develop evidence for efficacy and safety to be applied simultaneously in the several countries. Empirical Bayes methods allow pooling of evidence with allowance for variation within and among regions. The empirical Bayes framework guides the design of multinational trials and the determination of region-specific and total sample sizes, as well as a basis for interim analysis and adjustment of sample sizes. The approach is illustrated by application to a study of antiviral therapy.
|