Replies to post #88106 on Biotech Values

[DewDiligence]

PSDV:

Is this is a mistake or am I interpreting the number incorrectly? The high-dose p value is .026 and this is stat sig.

When two doses are being tested against the control arm, as was the case in this trial, there are effectively two shots on goal. In order to keep the probability of a false positive below 0.05, the aggregate 0.05 of “alpha” must be divided between the two dosing arms, and the conventional way of doing this is to allocate 0.025 of alpha to each arm.

Thus, absent a prior agreement with the FDA to use a different method of alpha allocation (which the PR does not claim), one may assume that the p-value threshold for each arm to be deemed statsig was 0.025. It follows that neither dosing arm in the “A” trial was statsig according to the primary (modified-ART) analysis.