thanks again, TechGuru. I feel like I am slowly le
Post# of 154562
(Total Views: 581)
Posted On: 09/30/2020 8:31:42 PM
thanks again, TechGuru. I feel like I am slowly learning this stat math stuff for human trials.
In particular, I now see the importance of your distinction between between the two proposed cases for 45 trial deaths noted in today's ASM slide.
23/22 comes in at a p value of .006.
24/21 comes in at a p value of .002.
Thats 23 placebo deaths/22 leronlimab deaths versus 24 placebo deaths/21 leronlimab deaths.
.006 is higher than the presumed interim analysis efficacy stop value of .005, and thus does not achieve an efficacy stop.
.002 is lower than the presumed interim analysis efficacy stop value of .005, and thus does achieve an efficacy stop.
So that all makes sense.
I wonder where does the interim analysis efficacy stop criteria of 0.005 come from?
Is that number a standard FDA protocol? Or maybe a statistical measurement that describes the probability of success if the trial is run out to the end?
In particular, I now see the importance of your distinction between between the two proposed cases for 45 trial deaths noted in today's ASM slide.
23/22 comes in at a p value of .006.
24/21 comes in at a p value of .002.
Thats 23 placebo deaths/22 leronlimab deaths versus 24 placebo deaths/21 leronlimab deaths.
.006 is higher than the presumed interim analysis efficacy stop value of .005, and thus does not achieve an efficacy stop.
.002 is lower than the presumed interim analysis efficacy stop value of .005, and thus does achieve an efficacy stop.
So that all makes sense.
I wonder where does the interim analysis efficacy stop criteria of 0.005 come from?
Is that number a standard FDA protocol? Or maybe a statistical measurement that describes the probability of success if the trial is run out to the end?
(0)
(0)