Replies to post #4072 on Dendreon Corp fka DNDNQ

[p3analyze]

why does it matter- how do you know that minimally symptomatic patients would necessarily take up a large percentage of the enrollment knowing that it was only allowed for part of the trial? You don't and they won't.

Even if it is a non-ignorable proportion, how do you know it would necessarily be unbalanced between the two arms? You don't, it would not likely be.

Even if it were not balanced, how do you know that minimal symptomatic would necessarily be a prognostic factor of overall survival? You don't, and it wasn't

If you multiply the above three answers together, the most likely outcome would be the minimally symptomatic thing is not going to affect the treatement effect on overall survival.

Mind you that it is effect modification but not confounding that one should worry about. And you have been hacking on confounding (via your conspiracy theory on how DNND would apply halabi) which is not terribly relevant.

[iwfal]

There must be 50 ways that DNDN could game the Halabi output without anyone ever knowing.

Agreed that this is theoretically true to some degree (although it would be a lot tougher to game than the Cox Regression because Halabi is pretty explicit)


But I may be willing to cut them some slack due to Halabi being tougher to game and Dendreon has been right on some things - e.g. the vote at the AC implies that they getting their manufacturing up and running was not a complete game. No strong opinion either way at the moment.

Think of all the things that you guessed cynically about Dendreon and were wrong. This NOT meant to be a bash - only that Dendreon is no CTIC.