Replies to post #19786 on Biotech Values

[iwfal]

And I suppose more accurately, examples of FDA rejection due to reliance on pre-specified Cox analysis? (In the last DNDN SEC filing, "Cox" analysis is discussed four times, and every time it is described as "pre-specified Cox multivariate regression analysis", presumably emphasizing that it was not post hoc.)

Several comments:

a) "pre-specified" is not black and white. For instance, pre-specified multivariate Cox with 20 covariates doesn't mean much because of the effect that Dew and I have been discussing - that just by lowering the noise the p value improves. Dendreon didn't do anything quite that bad in 9901, but it does look a little like data mining analysis. 9902a, with its smaller subset of covariates as found in 9901 is substantially cleaner, but probably not 'pre-specified' with the same formality.

b) Finding FDA rejections is hard(!) because the FDA doesn't publish the results (except via AC minutes) and the companies notoriously don't tell the whole truth. But I will say I've never seen an approval with that much improvement due to a Cox Regression - supporting or primary.

[AlohaDan]

presumably emphasizing that it was not post hoc

It was not post hoc. It was prespecified.