NorthWest Biotherapeutics Inc (NWBO)

[Doktornolittle]

For those with more statistics background than myself: A clue is that the increase in PFS events is 2.25 fold, the square of 1.5. The noise goes as the square root of the number of samples. A 2.25 fold increase in samples lower the noise to 2/3 of what it was.

So you have to increase the samples a lot to lower the noise a little. One reason to start the trial with a cautious low number of patients. You can easily waste a lot of money if you try to reduce the noise by just a little more than what you need.

I say for those with more background than myself because that is as far as I can go with this.