|
Abstract:
|
Evaluating the efficacy of a vaccine during a public health emergency presents important logistical challenges. The novel ring vaccination cluster randomized trial design, in which a cluster is defined as the contacts and contacts of contacts of an index case, addressed many of these challenges and was successfully used in Guinea in 2015 to evaluate an Ebola vaccine candidate. One key statistical consideration of this and other vaccine trial designs is determining which cases should contribute to the primary vaccine efficacy outcome. Cases appearing shortly after vaccination may have been infected prior to vaccination or before developing protective immunity. The traditional strategy is to start counting cases after a long fixed, delay period, but this may require dropping a large proportion of the cases observed in the trial. Furthermore, if incidence declines over time within clusters, it may be hard to collect enough countable endpoints. We use an analytical and simulation based framework to provide recommendations on how to select the optimal delay period to maximize study power for the primary vaccine efficacy outcome.
|
