(Total Views: 612)
Posted On: 02/11/2021 12:42:25 AM
Post# of 155256
Mtruong,
Thanks. Yes you are correct, my mistake (65 instead of 67). So we have:
203/65=31.23% and, our placebo: 130*0.3123=40.6 will make it 40 (rounding against us).
Therefore, our arm (VX) will have: 87-40=47 deaths !!!! This means a p-value of 0.006914 and a reduction in mortality of 41.3%. If we round up we have 46 deaths with p-value 0.0033 and 43.9% mortality reduction.
Another way of calculating is with their actual placebo death rate:
20 deaths over 67 patients: 29.85% which will extrapolate to 38.8 (rounding to 39).
And our arm (VX) will have: 87-39=48 deaths. This means a p-value of 0.01377and a reduction in mortality of 38.5.%.
I personally trust better the first method as it has a much larger number of patients (larger sampling space) and does not assume that the percentage of deaths was HIGHER in the AV arm (33.1%) than in the SOC side. (29.9%). Taking their total number of deaths is more significant imho.
In any case, in either case we are sitting pretty if I can say so.
OK, time to go and try to sleep ... (12:42 P.M. here)
Thanks. Yes you are correct, my mistake (65 instead of 67). So we have:
203/65=31.23% and, our placebo: 130*0.3123=40.6 will make it 40 (rounding against us).
Therefore, our arm (VX) will have: 87-40=47 deaths !!!! This means a p-value of 0.006914 and a reduction in mortality of 41.3%. If we round up we have 46 deaths with p-value 0.0033 and 43.9% mortality reduction.
Another way of calculating is with their actual placebo death rate:
20 deaths over 67 patients: 29.85% which will extrapolate to 38.8 (rounding to 39).
And our arm (VX) will have: 87-39=48 deaths. This means a p-value of 0.01377and a reduction in mortality of 38.5.%.
I personally trust better the first method as it has a much larger number of patients (larger sampling space) and does not assume that the percentage of deaths was HIGHER in the AV arm (33.1%) than in the SOC side. (29.9%). Taking their total number of deaths is more significant imho.
In any case, in either case we are sitting pretty if I can say so.
OK, time to go and try to sleep ... (12:42 P.M. here)
(8)
(0)
