|
Abstract:
|
For some trials simple but subtle assumptions can have a profound effect on the size of the trial. The design of a vaccine lot consistency trial is a case in point. Standard sample size formulas rely on only one component of variation, namely, the variation within lots. The other component, the variation between lots, is either unintentionally ignored or assumed to be equal to zero. But variation between lots, however small, will be inherent in the manufacturing process. Using data from a published lot consistency trial we demonstrate that when the between-lot variation is only 0.5% of the total variation the relative increase in the sample size is nearly 400% when compared to the assumption that the lots are identical.
|
