Replies to post #138337 on Biotech Values

[TastyTheElf]

He'd get sued and rightly so - since he has previously said (I've seen it in transcripts) that the boundary was OBF.

This is silly. You yourself have said that OBF means something very specific. In that sense, there is no "OBF" when you don't have equally spaced looks. Since the looks were 50%, 75%, and 100% here, we're already outside OBF in the sense you've propounded.

What I believe Kirkman meant by that was that the mathematical formula used to generate the boundaries is the formula that is used to generate the boundaries in an OBF setup. It is perfectly reasonable to use those formulas at whatever information timeframes you like, using Lan DeMets to calculate the alpha spent at each look.

This being the case, there is no reason why an SPA couldn't use that formula, the Lan DeMets alpha spend provision, and any parameters for calculating the boundaries desired.

It is not true that there are no other trials that have used the OBF formula in a nonstandard way. This one did:

WIZARD trial

Regards,
TGW