Replies to post #5627 on Dendreon Corp fka DNDNQ

[iwfal]

I respectfully disagree.

Program-survival bias — as opposed to the various other biases in phase-2 trials — would pertain even if phase-3 studies were exact clones of the phase-2 studies in trial design and size.

It is amusing to me how often these discussions come down to semantics. What is the precise definition of PSB? Of "power analysis"? (pertinent to a different conversation I had with Toos on the IV board), ...

[walldiver]

Sorry, I'll have to disagree as well. I stated that program survival bias is a plausible theory, but it's less likely to occur if your nonpivotal Phase 2/3 trials are randomized. In other words, randomizing your earlier trials gives you a better idea if your drug works. This is something that responsible biotechs should do, instead of data-mining your Phase 3 (first randomized trial) in order to find a subgroup that saw success.