Replies to post #211872 on NorthWest Biotherapeutics Inc (NWBO)

[exwannabe]

The 'pack has AWSOME computer math/stat and hacking skills :-)

Ok,back OT. It could have no effect on the ratio of meth/unmeth enrolled. That is 100% an issue with the docs doing the screening.

As far as too predictable, that is what drives the body of math that has well solved this. Insuring the randomization is neither predictable nor able to get far off balance.

I doubt it was UCLA who was in charge. I assume the sponsor (likely with the help f the CRO) specifies it while the randomoization company implements it.

[flipper44]

I guess in the end, it never seemed too important to me until now, but the increase in percentage of methylated patients compared to a normal distribution (aka: this trial had more methylated than historical precedent might have it) could be a very subtle way to hurt this trial.

First, methylated are more subject to psPD. I don't need to explain what potential problem that could increase do I?

Second, methylated for both arms would be expected to live longer than if this trial had a historical randomization ratio of methylated to unmethylated. This would cause late separation in the trial (the critical 36 months Dr. Bosch emphasized), but if the trial was terminated early, would not reflect the potential of the trial. The trick to hurt the trial would be, if intentionally done, pressure NWBO into unblinding back in 2013/2014. Maybe even 2016 or perhaps 2017. Unblinding the trial early would magnify the disparity and reduce the chance of any observed separation between the arms.

[AVII77]

I'm just loosely examining the new info -- aka: the last 31 were truly randomized.

(18 Distribution of final 31 patients): All patients were randomized. You can't just put a group in one arm or the other. NWBO had nothing to do with which patients ended up in which arm. Because of the large number of sites, the final randomization may not have been exactly 2:1 but may have been off by a few people one way or the other.

Flipper, take heart. 31 is not "a few". If Les is full of shit with a statement like that what makes you think he is not full of shit with the rest of his assertion.

On what planet is 31 a "few"?

Amarin just reported on a 8179 patient trial with a "large number of sites".

473 sites

they ended with:
4089 vs 4090

I'm guessing we someday find out the last 31 were indeed treatment.