Biotech Values

[iwfal]

I'd like to see more evidence for this claim. Showing that one subgroup does much better than another doesn't imply that the drug is actually making the inferior subgroup worse.

In all cases I cited I used only examples with results from randomized trials vs SOC and strong results (i.e. reasonably low p values - to avoid the problem of post hoc exploration) - so in all cases the pfs vs placebo in one subgroup was HR clinically meaningfully greater than 1 (in many cases MUCH greater than 1), while in a different subgroup the pfs vs placebo was HR<<1.

As for the detailed links - I provided some already (MetMab and Gleevec), and Iressa is common knowledge. As for the Erbitux see this link.

I'll leave you to find the Vertibix data.

FWIW