Replies to post #707 on Dendreon Corp fka DNDNQ

[io_io]

"We know that 2 of the 5 from 9901 would have shown up, but how many of the 20?"

Seeing as they didn't know the prognostics thru all the years of emdical history, how many do you expect to show up with SIGNIFICANCE in a 98-person trial ??


"To be clear, I don't think the reputation of any of the presenting doctors is on the line in any way. Higano and Small are free to present any accurately run analysis because it is an academic exercise for them."

They are not going to mis-lead people according to the theory you have painted - that 9902a only gets significant if you use borrowed covariates. These people DO depend on their reputation.


"They are not there to communicate likelihood or confidence of FDA approval, nor are they there to characterize DNDN's relationship with the FDA."

They made no such attempts ?

[ocyanblue]

<>So, if DNDN had gone by the same procedure as done for 9901, they would have found just three significant covariates (plus treatment)<

We don't know that, do we? We know that 2 of the 5 from 9901 would have shown up, but how many of the 20?>

Is this a visceral response? You sounded like you missed the rest of my post?

<To be clear, I don't think the reputation of any of the presenting doctors is on the line in any way. Higano and Small are free to present any accurately run analysis because it is an academic exercise for them.>

If they did not reveal the stat sig covariate, the analysis was not accurate. And if that wasn't, their reputation would be on the line.

<They are not there to communicate likelihood or confidence of FDA approval, nor are they there to characterize DNDN's relationship with the FDA.>

My comment was not about the likelihood of FDA approval. As scientists, they/we should aim for accuracy in their/our data and analyses first. Yes?