NorthWest Biotherapeutics Inc (NWBO)

[Doktornolittle]

AVII: Thank you for the links about multiplicity. I really will read/view them all today. As dry a subject as it is, I am curious enough to push through material.

But I do want to comment on the issue of penalty for an unplanned look:

I argue that if Northwest did not call for the unplanned look then there should be no penalty. I argue that for the following reason.

The intuitive reason for this penalty is the following:

If you did have access to all efficacy data as the trial moved along you would see a very noisy signal slowly dropping in noise as it moved along, but with a moving average mean value that remains nearer the center of that noise spread, more-so, of course, as time goes on. So you might guess what the signal is going to converge to by looking at these peak to peak values. If you had that info, you could deliberately pick a point where the noise was near a positive extreme in efficacy, and ask for consideration of early approval at that point.

Northwest does have some data. They know when people cross-over and when they pass away.

But if the Germans or anybody else setup in advance a look at efficacy at the point of full enrollment in the old trial design, and that is the reason for the current look, if there is a look going on, then there should not be a penalty even though it is unplanned by the sponsor or it's dmc. As long as it is a planned look, it does not matter who planned it, because the advantage I describe above would not exist. Maybe there would be rationale for a penalty moving forward, but not for the current consideration of approval.