Mate brilliant as usual.
Minimise placebo, maximise treament effect and optimise timepoint for max difference. Surely P3 should come out well compared with P3.
Doesnt really get any better does it.............or does it?
That SAS data is really interesting, and got me looking closer at it.
When they report from each cohort seems they report the
cumulative pain reduction.
If your whole population average increases with an additional cohort, the average of that cohort must be greater than the population average you end up with. Simple example.....If you have 10 samples averaging 5 and then when you get to 20 samples and it averages 6, it means the average of the second 10 is 7.
So in our figures, cumulative averages ...
n=34 - 44.9%
n=76 - 47.3%
n=89 - 49.6%
....means the three group averages were....
n=34 - 44.9%
n=42 - 49.2%
n=13 - ....wait for it...
63.0%The final 13 subjects bumped up the entire population (89) average more than 2%, hence why the average of that cohort of 13 must be so much higher. I wonder what they did in that last cohort, and if it can be replicated in trials going forwards!
(20min delay)