Replies to post #78656 on Biotech Values

[CrazyPete]

If you believe in statistic there is 97% chance it will be stat significant.

That's false. The chance of the trial giving a significant result is almost certainly much lower than 97%, and depends on many other things besides the phase II P value.

[iwfal]

Belimumab ph ii results in SLE:

The phase two they run in seropositive SLE patients was stat sig (P=0.03 if I remember correctly).

0.044. But a clearly post hoc subgroup of 'seropositive' where not only is it a subgroup, but one which in which there is no standard definition of seropositive (hence they have to define it in the papers). So very unreliable.

I give the chance of them being stat sig as very low - just because they are clearly chasing a very post hoc subgroup. And not only that but are using a convoluted endpoint. The only positive they have going for them is a large trial size.

Much more interesting than the likely fail in the phase 3 for Belimumab will be the secondary endpoints. (Speaking of which I need to see if anything further has been released on the Rituxan SLE trials).

[microcapfun]

HGSI/lupus
>>The phase two they run in seropositive SLE patients was stat sig (P=0.03 if I remember correctly). If you believe in statistic there is 97% chance it will be stat significant.<<

You don't remember correctly. Their Phase 2 trials were not in seropositive SLE patients. When post-facto they defined "seropositive SLE" and threw out the patients that weren't ... they STILL didn't have statistical significance for their original endpoint, only trends. So (post-facto again) they made up a new endpoint (that the FDA later signed off on for their Phase 3 trials) and only then did the Phase 2 data - with the post-facto patient population and the post-facto endpoint - show statistical significance.

I'm not saying the Phase 3 trials will necessarily fail, but certainly the probability of success is way less than 97%. More like 25% is my guess.

HGS also presented data of a continuation of the Phase 2 trial. The data looked good, but the patients who choose to continue getting treated are obviously self-selecting themselves. If you are in that continuation trial and you get a flare ... you're probably going to leave the trial, right? So those who remain are going to have less symptoms and that's going to make Lymphostat B look good. But so would koolaid, under those conditions ...

micro