What is it in a number ??? I am sorry for stres
Post# of 155052
(Total Views: 536)
Posted On: 05/02/2020 11:22:32 AM
What is it in a number ???
I am sorry for stressing so much the p-number (this is not the first post about this). Let me explain why: few years back I had a considerable amount of money invested in a company. After a pivotal trial of more than 2 ½ years the results yielded a p-number of 0.056
The drug was not approved by FDA. After several ins and outs FDA agreed to allow the company to do another trial. This was a pre-revenue company and did not have the funds to carry this out. That was the end of the company and a large part of my investment. Lesson: 0.006 in a p-number matters !!!
We will be approved if we meet this number (p<0.05). Some say it is not carved in stone, but it is
Now, let’s analyze, for sake of argument, the results that Drs. NP & BP heralded in last CC:
To play devil’s advocate I will assume that all the patients where in “trial” conditions (we know that the patients from Montefiori where in their last moments).
We have then, that of a total of 39 patients (Very Critically ill + Critically ill + Severely ill) 27 survived, 4 “saved”, 10 extubated or improving, 13 discharged or improving respectively.
The question is: is this significant from the statistical perspective ?? or, would we be approved if we repeat results like these in the Critical-Severe trial??? The answer is that this depend on the placebo arm.
So, utilizing a binomial probability distribution (normally used when the outcomes are binary, that is survival vs non-survival). We have that, if the placebo arm shows a survival rate of less than 46% (that is 21 patients or less survive) with 27 patients we will have a p-number of 0.0387. If we had 26 patients surviving our p-number would be 0.07574 and we would not be approved.
The point here is: we know that the survival rate of intubated patients is much smaller than 47% !! meaning that our p-number would be very small.
To stress this fact let me, again, playing devil’s advocate, take into account ONLY the Very Critically ill and Critically ill (total of 22). Of these 14 survived (again I am counting the patients in very bad shape from Montefiori where the doctors treated them with Leronlimab as last resort).
How many patients in a placebo arm (in similar conditions) should we have for our 14 survivors be statistically significant (p<0.05)??
The answer is: if less than 9 patients survive (59% deaths) in the placebo group with our 14 patients we will have a p=0.027 and will (or, shall I say, should) be approved. We KNOW that the number of deaths for intubated patients is much larger than 59% !!!.
So, the data released by Drs. NP and BP is VERY meaningful, from previous interview I had already made some calculations and that is why I committed more money to CYDY friday morning.
We have meaningful results and, if the trend continues we will be approved. We need to move swiftly in the recruitment. And needless to say, at the end of the day the market will be HUGE.
I am sorry for stressing so much the p-number (this is not the first post about this). Let me explain why: few years back I had a considerable amount of money invested in a company. After a pivotal trial of more than 2 ½ years the results yielded a p-number of 0.056
The drug was not approved by FDA. After several ins and outs FDA agreed to allow the company to do another trial. This was a pre-revenue company and did not have the funds to carry this out. That was the end of the company and a large part of my investment. Lesson: 0.006 in a p-number matters !!!
We will be approved if we meet this number (p<0.05). Some say it is not carved in stone, but it is
Now, let’s analyze, for sake of argument, the results that Drs. NP & BP heralded in last CC:
To play devil’s advocate I will assume that all the patients where in “trial” conditions (we know that the patients from Montefiori where in their last moments).
We have then, that of a total of 39 patients (Very Critically ill + Critically ill + Severely ill) 27 survived, 4 “saved”, 10 extubated or improving, 13 discharged or improving respectively.
The question is: is this significant from the statistical perspective ?? or, would we be approved if we repeat results like these in the Critical-Severe trial??? The answer is that this depend on the placebo arm.
So, utilizing a binomial probability distribution (normally used when the outcomes are binary, that is survival vs non-survival). We have that, if the placebo arm shows a survival rate of less than 46% (that is 21 patients or less survive) with 27 patients we will have a p-number of 0.0387. If we had 26 patients surviving our p-number would be 0.07574 and we would not be approved.
The point here is: we know that the survival rate of intubated patients is much smaller than 47% !! meaning that our p-number would be very small.
To stress this fact let me, again, playing devil’s advocate, take into account ONLY the Very Critically ill and Critically ill (total of 22). Of these 14 survived (again I am counting the patients in very bad shape from Montefiori where the doctors treated them with Leronlimab as last resort).
How many patients in a placebo arm (in similar conditions) should we have for our 14 survivors be statistically significant (p<0.05)??
The answer is: if less than 9 patients survive (59% deaths) in the placebo group with our 14 patients we will have a p=0.027 and will (or, shall I say, should) be approved. We KNOW that the number of deaths for intubated patients is much larger than 59% !!!.
So, the data released by Drs. NP and BP is VERY meaningful, from previous interview I had already made some calculations and that is why I committed more money to CYDY friday morning.
We have meaningful results and, if the trend continues we will be approved. We need to move swiftly in the recruitment. And needless to say, at the end of the day the market will be HUGE.
(12)
(0)