(Total Views: 1011)
Posted On: 05/19/2020 1:36:16 PM
Post# of 148902
When the rubber meets the road.
So, we are in the midst of a P2 trial with 75 patients. The outcome of this trial will be very important for CYDY, first of all, will launch it into the list of companies with a veritable claim to a COVID drug, but, more importantly, will give CytoDyn national and international exposure.
It will the end of the beginning, and the SP will likely reflect it rapidly.
The rules:
I simulated a possible set of patients with the objective of finding out what to expect to get statistical significance (p-value<0.05).
This type of tests is called an Unpaired test used to compare the mean between two independent groups. In this case the placebo and the one with Leronlimab
So, the question is: how much have the patients with Leronlimab have to improve as compared with the patients with placebo at the end of day 14th? Or, better: How much less have the mean of the results with Leronlimab than that of placebo patients to be to achieve p<0.05?.
Remember that every patient is the addition of 4 conditions (fever, myalgia, dyspnea and cough) and some might or might not have some of these. For example, a patient might have a severe form of all four and will have a score of 12 at the beginning of the trial, at day 14 if it remains the same will have 14, but, let’s assume that it improves from severe to moderate in one and to mild in another one, the score will be then 11. We add all the results and divide by 75 to obtain the mean of each group.
In running a simulation of outcomes for normal distribution and a 2-tailed test I arrived to the conclusion that patients with Leronlimab have to have a mean at least 1.4 points less than placebo (approximately). I calculated this with a variance of 3.52 for the placebo group and 2.21 for the Leronlimab group.
In any case, without doing mathematical parlance (unfortunately can't share tables and probability distribution plots with you as have been unable to post them in this board) what does this mean? If all the patients improve in one point, for example, fever goes down one notch for all the patients, all other conditions being equal, is this good enough ?? No , the means difference would be 1 only (p=0.1486). What about if it goes one point for all in fever and one point for one half of the others (25 patients) in, say, cough ?? yes, the difference on means would be 1.5 (more than 1.4)
These are just approximations as I am assuming results and using a “synthetic” variance (3.4) and means around 7.4 for Placebo, as there is no way to know what it is before we read the results. This is just to form an idea of what to expect.
In any case, this trial will simply test the difference between the means of the two cohorts. I am convinced Leronlimab mean will be MUCH lower that the placebo and our results will be outstanding. For example, what happens if our mean is 2 points less than the control group ?? p=0.00604. Will not be surprised with results like this and even better.
The time has come ... FDA will not be able to "hide" any more, and as Dr. Patterson well said:
So, we are in the midst of a P2 trial with 75 patients. The outcome of this trial will be very important for CYDY, first of all, will launch it into the list of companies with a veritable claim to a COVID drug, but, more importantly, will give CytoDyn national and international exposure.
It will the end of the beginning, and the SP will likely reflect it rapidly.
The rules:
Quote:
Clinical Improvement as assessed by change in total symptom score (for fever, myalgia, dyspnea and cough) [ Time Frame: Day 14 ]
Note: The total score per patient ranges from 0 to 12 points. Each symptom is graded from 0 to 3. [0=none, 1=mild, 2=moderate, and 3=severe]. Higher scores mean a worse outcome.
I simulated a possible set of patients with the objective of finding out what to expect to get statistical significance (p-value<0.05).
This type of tests is called an Unpaired test used to compare the mean between two independent groups. In this case the placebo and the one with Leronlimab
So, the question is: how much have the patients with Leronlimab have to improve as compared with the patients with placebo at the end of day 14th? Or, better: How much less have the mean of the results with Leronlimab than that of placebo patients to be to achieve p<0.05?.
Remember that every patient is the addition of 4 conditions (fever, myalgia, dyspnea and cough) and some might or might not have some of these. For example, a patient might have a severe form of all four and will have a score of 12 at the beginning of the trial, at day 14 if it remains the same will have 14, but, let’s assume that it improves from severe to moderate in one and to mild in another one, the score will be then 11. We add all the results and divide by 75 to obtain the mean of each group.
In running a simulation of outcomes for normal distribution and a 2-tailed test I arrived to the conclusion that patients with Leronlimab have to have a mean at least 1.4 points less than placebo (approximately). I calculated this with a variance of 3.52 for the placebo group and 2.21 for the Leronlimab group.
In any case, without doing mathematical parlance (unfortunately can't share tables and probability distribution plots with you as have been unable to post them in this board) what does this mean? If all the patients improve in one point, for example, fever goes down one notch for all the patients, all other conditions being equal, is this good enough ?? No , the means difference would be 1 only (p=0.1486). What about if it goes one point for all in fever and one point for one half of the others (25 patients) in, say, cough ?? yes, the difference on means would be 1.5 (more than 1.4)
These are just approximations as I am assuming results and using a “synthetic” variance (3.4) and means around 7.4 for Placebo, as there is no way to know what it is before we read the results. This is just to form an idea of what to expect.
In any case, this trial will simply test the difference between the means of the two cohorts. I am convinced Leronlimab mean will be MUCH lower that the placebo and our results will be outstanding. For example, what happens if our mean is 2 points less than the control group ?? p=0.00604. Will not be surprised with results like this and even better.
The time has come ... FDA will not be able to "hide" any more, and as Dr. Patterson well said:
Quote:
We don’t have to wait too much longer to the validation on where we stand in the COVID world.
(21)
(0)Scroll down for more posts ▼
