An Adaptive Design for Case-Driven Efficacy Study When Incidence Rate is Unknown – A Case Study
*Xiaoming Li, Merck Research Laboratories  Ivan S.F. Chan, Merck Research Laboratories  Keaven Anderson, Merck & Co., Inc. 

Keywords: event-driven, adaptive design, vaccine

In vaccine studies the efficacy endpoint usually is a rare disease event, and case-driven design is used. When there is good knowledge on the incidence rate of the endpoint (IR), the study sample size/duration can be determined. However, the IR may not be known in some cases. One option is to have a natural history study first. Another is to start with a large sample size with built-in interim analyses. In light of a potentially long duration of the first option and large up-front investment in the second option, we propose a two-stage adaptive design strategy built on the traditional event-driven design, in which the potential issues on the feasibility, uncertainty in IR and interim analysis strategy are tackled together. The adaptation is based the conditional rejection probability approach. This design will be evaluated and also be compared to a traditional group sequential design.