Statistics and Its Interface

Volume 5 (2012)

Number 4

An adaptive design for case-driven vaccine efficacy study when incidence rate is unknown

Pages: 391 – 399



Keaven M. Anderson (Merck Research Laboratories, North Wales, Pennsylvania, U.S.A.)

Ivan S. F. Chan (Merck Research Laboratories, North Wales, Pennsylvania, U.S.A.)

Xiaoming Li (Gilead Sciences, Seattle, Wash., U.S.A.)


In many vaccine efficacy studies where the endpoint is a rare infection/disease event, an event-driven design is commonly used for testing the hypothesis that study vaccine lowers the risk of the event. Uncertainty of the incidence rate has a large impact on the sample size and study duration. To mitigate the risk of running a potentially large, long-duration efficacy trial with an uncertain event rate, we propose a two-stage adaptive design strategy with interim analyses to allow evaluation of study feasibility and sample size adaptation. During Stage I, a modest number of subjects will be enrolled and the feasibility of the study will be evaluated based on the incidence rate observed. If the feasibility of the study is established, at the end of Stage I a formal interim analysis will be performed, with a potential sample size adaptation based on the conditional rejection probability approach. The operating characteristics of this design are evaluated via simulation.


adaptive design, vaccine efficacy, event-driven, unknown incidence rate

Full Text (PDF format)