Replies to post #25051 on Cel Sci Corporation (CVM)

[RobotDroid]

You gotta love Oxford examples of or:
Used to connect different possibilities:
Is it Tuesday or Wednesday today?
You can pay now or when you come back to pick up the paint.
Are you listening to me or not?
The patent was granted in (either) 1962 or 1963 - I can't quite remember which.
It doesn't matter whether you win or lose - it's taking part that's important.
There were ten or twelve (= approximately that number of) people in the room.
He was just kidding - or was he (= but it is possible that he was not)?

A2
used after a negative verb to mean not one thing and also not another:
The child never smiles or laughs.
Compare
nor
More examples
If the product doesn't work, you are given the choice of a refund or a replacement.
Are the photos in colour or black and white?
Are you comfortable or shall I turn the heat down?
You can move the cursor either by using the mouse or by using the arrow keys on the keyboard.
Is it my imagination or is David behaving strangely at the moment?

[georgebailey]

@fosco Yep that’s a big OR. Is it possible the power of the study was a close call in March ? For this to be a variable worth considering we must assume that significant dropouts could force the idmc to order additional patient enrollment to achieve power. At the September meeting is this a realistic potential outcome?
If I was cvm I would have paid the stat consultants to run numbers including the 3rd arm and try to match the event total. Maybe they did that.

[Cotton_Farmer]

@Fosco1 - please confirm results for one additional variation. 1% drop-outs.

[staticmirror79]

Seems to me that if the dropout numbers are reasonable, let's say 15%. And the SOC survival is even close to the SEER data at Y3 and Y5. We should be anywhere from 20-40% increase in survival.

Even if the IDMC simply says "continue" that should tell us we have a reasonable amount of dropouts i.e. not above 15%.

Question:
I understand that we can't have too many dropouts otherwise we don't have enough statistical power to end the study. But how do we know the upper end is 15%?