Replies to post #5641 on Dendreon Corp fka DNDNQ

[walldiver]

<<Shots on goal.

Sodium chloride will be stat sig for prostate cancer if you take enough looks at data sets.

I really don't want to get into the debate of why they changed the SPA, but it is CLEAR that the 360 event data will not do the trick.>>

Just to be clear, you mean that the 360-event data won't get FDA approval if the 304-event data has already failed.

[p3analyze]

Exwannabe, suppose provenge is saline. Agreed, 1 out of 100 times it will produce a hazard ratio like 9901/02a, 5 out of 100 times it will show a statistically significant HR.

Now what I would like to ask you to tell me what's the probability that you do four trials in a row with the following outcome:
a) the first repeats the HR of 9901/02a,
b) the second shows a p-value of value of 0.042 at 360 deaths, but only demonstrated a p-value at 304 deaths at 0.05.

Pa (Probability of a)=0.011, and Pb=0.05*0.9=0.045 (Note, the 0.9 is the conditional probability of showing a p-value of 0.042 at 360 deaths if the p-value at 304 deaths is 0.05).

Therefore the probability of observing both a and b all happening is 0.01*0.045=0.00045.

Now I ask you to dump the saline in the IV bag of yours, and fill it up with provenge. Suppose IMPACT does play out like the scenario b above, are you telling me given that there is less than a 5 out of 10-thousand chance that Provenge is essentially saline, therefore it is saline? So as a treating oncologist, you will prefer to continue to treat patients with Docetaxel, which has 1 out 100 chance to behave like saline (p=0.01)?