Replies to post #4978 on Dendreon Corp fka DNDNQ

[DewDiligence]

I think CEGE’s CEO is just playing word games by saying that the interim analysis will be based on the HR. Since the interim p-value follows directly from the interim HR, there is nothing unusual about CEGE’s interim analysis, IMO.

[io_io]

<<<"Today, the CEGE CEO said that their 2Q08 interim look at the fully enrolled 600 patient Vital 1 trial in asymptomatic AIPC will be looking at Hazard Ratios since it is an event based trial.">>>


He will indeed be looking for a certain HR, and there is a simple explanation.

It doesn't really matter what the Design is, or what p-value is needed, but what has happened is that he has asked the statistical designer to "back-calculate" what HR he needs, because when he think of the strength of the drug over control, he thinks in terms of a HR. And now that they know both the design and the final number of patients, the statisticians can readily give him what he needs.

So suppose he thinks (based upon the Phase I and Phase II) his drug gives him an average drug HR of say 1.40 over control. It is average, because he doesn;t really know.

And so for the interim look, the target HR might be 1.50, and it's worth a look because it gives him a fighting chance, and if successful then he saves a year of losses and $50m in trial costs.

And so for the final look, the target HR might be 1.33, and this is the backup plan, the CEO knows he has a very good chance in the long run, even if it costs a lot more money/dilution, etc.


So you can see he wants to know the HRs as they are what he can relate to. He can't take a p-value from a smaller phase I and phase II to his pivotal trials, but the HR (in theory !) should tend to be roughly the same every time - or maybe even a little better, to get lucky - or hopefully not too much weaker, to avoid the worst.

[walldiver]

I don't know if HR will be as relevant as the p value. As I recall, the HR for 9902A wasn't bad, something like 0.77-0.79 if calculated the same way that every biopharma company in the world except DNDN does it. Taxotere's pivotal Phase III that got it approval in AIPC had a similar HR, but Tax's p value was <0.01 while the Provenge p value in 9902A was 0.33.