NorthWest Biotherapeutics Inc (NWBO)

[MI Dendream]

Ahhh and you will wave your hands and speak of powering. This trial was powered to detect a very small difference at a p value of 0.02. Powering is a forward looking calculation which asks the question how many patients are required to predict the ability to detect a difference at a certain p value given a certain variance. If a study is powered to detect a two month difference at p=0.02 then it is also powered to detect a narrower difference at p=0.05.

You are correct, it is an assumption that the alpha was split with OS, Otherwise, the alternative is even less good for your side. If the endpoints are independently powered at full alpha, you have the ability to statistically measure a very tiny difference. Splitting the alpha up is the best case you could have hoped for.

This is a phase 3 registrational trial, EX. The alpha started at 0.05.