Any guess why the study powered at an alpha of 0.01?
Great find! And no concrete idea, only speculations:
a) FDA wanted it more aggressive since the ph ii was post hoc
b) They 'powered it' this way but are really using 0.05 as their final alpha (after reading way too many SAPs in my experience this is exceedingly rare, if it every happens at all.)
c) They are using some alpha for an interim or other endpoint that they aren't telling anyone about. (alpha assignment in trials with multiple endpoints are a complicated maze I haven't figured out yet.)
I'd tentatively vote for #a above - but I am nowhere near certain.
PS - For those non-statisticians reading the thread, this is a big deal since it reduces the OS HRs that can be stat sig to OS HR <=0.83 (not 0.87 as I said previously when I assumed they were using 0.05 as the final alpha).
Market Data 
Markets 
AI
