Replies to post #5635 on Dendreon Corp fka DNDNQ

[p3analyze]

"I’m astonished that you think it may."

Yes, I thought you would be. I am going back to my favorite Gliadel FDA approval example. They filed based on a significant survival benefit with p-value less than 0.05 via a stratified log-rank test. However, the SPA'd analysis plan said an unstratified log-rank test should be used, and unfortunately that test yielded a p-value of 0.08. FDA rejected the NDA, but six months later, with more deaths occurred, the unstratified log-rank p-value became smaller than 0.05.

The sponsor refiled, and the drug was approved by FDA.

Of course, this would only happen that FDA really were thinking like doctors but not statisticians. Time has changed, and FDA has been pazdurized. Despite this, I still think provenge would have a chance in front of an ODAC if that scenario of yours played out.

[rjmcanoe]

I would be very angry with the FDA if they declined to approve provenge with 56 less deaths. Then after the final 56 deaths the p value dropped below .05 and became statistically significant.
I would then strongly suport DNDN rallys against the FDA and I would support any other pharmaceudical or similiar biotech company whether I was invested in them or not. I will think about what DEW says that after the study is completed additional encouraging data will have no effect on approval. What I am saying is that I would like to explore past FDA decisions on drug approvals to see if a similar event did not change the FDAs decision. It does not seem fair.
Bob