kyoto27

Dendreon's Provenge: Government Agencies Play Hide and Seek With Facts

The Pharmaco Man said:

“Sorry, but the data are pretty uncompelling--the vaccine doesn't enhance survival in a statistically significant way.”

Pharmaco needs to go back to school:


He asserts that the D9901 study did not attain statistical significance for survival when the final data was released in October 2004. This is an incorrect assertion. The overall survival p value was 0.01, easily beating the p=0.05 threshold. This goes to the crux of the anger of the Provenge advocates. Statistical significance on survival is the "gold standard" by which the FDA judges all oncology treatments, and Provenge met it with flying colors in the 9901 trial.

In addition, the p value for the primary endpoint of cancer progression was p=0.052, barely missing the p=0.050 threshold. For Howard Scher and Maha Hussain, this narrow miss was blown up to be an utter failure. They said that because the p value did not reach 0.050, then all additional data (including survival) should be discarded. Basically, their position was that because the odds of being successful in progression were only 94.8% and not 95.0%, there was no correlation between the two metrics and the 99.0% survival odds should be thrown in the trash.

So Pharma Man, the data does suggest it prolongs survival. And the data also suggests our FDA’s decision to de-rail Provenge had nothing to do with good science and a lot do with politics and money. Period. Next?