(Total Views: 551)
Posted On: 01/08/2021 3:46:27 PM
Post# of 148919
Hello havasu78:
A major point as we consider that the DSMC recommended continuing to a 75$ enrollment point for another interim analysis. Folks continue to perceive that because the trial was not halted at the 50% interim analysis, nor was it halted at the 75% enrollment interim analysis, folks assume that thus, we must not have achieved a p value of less than 0.05 and it appears that it has been supported, that for a trial to end, the data needs to be much more compellint, in fact, that the p value needs to be near 0.0001 extremely better than for EUA at completion of a trial. We would need above 60% reduction is deaths vs placebo to get such a p value.
We do not need 50% death reduction to get p<0.05 and thus to get an EUA.
Mid 30 I think will do it, 34% plus. I'm great at math, but have not seen an understandable equation for calculating p values.
Using our old eIND data that yielded a death rate (not reduction) but a death rate of 14.5% for s/c folks who got leronlimab, this would lead us to what many consider an unreasonably high, 63% reduction in death vs placebo today in our current s/c trial. I've triple checked that math and obtained confirmation from a stat guy her on IH.
If we had 50% death "reduction" at our 50% enrollment interim analysis or at our 75% enrollment interim analysis, that would not have been enough for the trial to be halted for compassionate reasons, you need a p value <0.0001 to halt a trial.
Thus we may have had a high 50% reduction in death rate at our 50% and/or our 75% interim data analysis.
A major point as we consider that the DSMC recommended continuing to a 75$ enrollment point for another interim analysis. Folks continue to perceive that because the trial was not halted at the 50% interim analysis, nor was it halted at the 75% enrollment interim analysis, folks assume that thus, we must not have achieved a p value of less than 0.05 and it appears that it has been supported, that for a trial to end, the data needs to be much more compellint, in fact, that the p value needs to be near 0.0001 extremely better than for EUA at completion of a trial. We would need above 60% reduction is deaths vs placebo to get such a p value.
We do not need 50% death reduction to get p<0.05 and thus to get an EUA.
Mid 30 I think will do it, 34% plus. I'm great at math, but have not seen an understandable equation for calculating p values.
Using our old eIND data that yielded a death rate (not reduction) but a death rate of 14.5% for s/c folks who got leronlimab, this would lead us to what many consider an unreasonably high, 63% reduction in death vs placebo today in our current s/c trial. I've triple checked that math and obtained confirmation from a stat guy her on IH.
If we had 50% death "reduction" at our 50% enrollment interim analysis or at our 75% enrollment interim analysis, that would not have been enough for the trial to be halted for compassionate reasons, you need a p value <0.0001 to halt a trial.
Thus we may have had a high 50% reduction in death rate at our 50% and/or our 75% interim data analysis.
(4)
(0)
Scroll down for more posts ▼