The trial readouts are approaching. We have a bet
Post# of 154117
(Total Views: 1754)
Posted On: 06/18/2020 9:48:49 AM
The trial readouts are approaching. We have a better idea of what to expect from the qualitative perspective and the number of patients.
This is it !!!, as Dr. Patterson said: "we will see soon where we stand in the COVID world".
There are two things stand at the front of us: The FDA decision ( approval or P3 with more patients) and the p-value which will be used by them to reach the former.
imo we have to hit a homerun in our P2 trail to get FDA to approve us, this means a p-value less than 0.01 (maybe some of you will disagree with this and this needs to be even lower) I have made some simulations with 90 patients and assuming a Normal distribution, which I am sharing:
https://drive.google.com/file/d/1kFIJ4DUN1IKJ...sp=sharing
This is my initial state of patients at day 0, I made both, placebo and Leronlimab arms the same. I used 46.7%, 63.3%, 43.3% and 16.7% for the percentage of patients with fever, cough, dyspnea and myalgia respectively. This following studies published (with a good number of patients) noting that this numbers vary wildly, for example, China reports more than 80% with fever and so on.
Then, I assumed that some placebo patients will worsen and some on Leronlimab will improve. The results are below:
https://drive.google.com/file/d/19LG5Q0oSsrJK...sp=sharing
Then in analyzing the data I obtain a p-value of 0.0094 (assuming a two-tailed test) that would get the FDA thinking ….
I am sharing this to illustrate what kind of results we would expect to have a potential approval. Basically, my simulation tells us that if the placebo group worsens from an average of 3.067 to 3.567 after 14 days and Leronlimab improves from the same value to 2.75 we would have a pretty nice result.
This is not difficult to achieve if Leronlimab works the way it is supposed to... frankly I expect a lower average in our cohort.
Once again, I have made several assumptions and this is ONLY for illustrative purposes. If you disagree with the data it is fine as I have no way of knowing the shape of patients before of after the trial, just made some "educated" guesses. Same regarding the model.
_____________________________________________________
Only for math fans:
Average at day zero: 3.067 (both cohorts)
Average at day 14 placebo: 3.567
Average at day 14 Leronlimab: 2.750
Standard Deviation Placebo day 0 and day 14 : 1.337 - 1.326
Standard Deviation Leronlimab day 0 and day 14 : 1.337 - 1.348
t=2.6543
Degrees freedom=88
Standard error difference=0.308
95% confidence of means [0.21-1.43]
This is it !!!, as Dr. Patterson said: "we will see soon where we stand in the COVID world".
There are two things stand at the front of us: The FDA decision ( approval or P3 with more patients) and the p-value which will be used by them to reach the former.
imo we have to hit a homerun in our P2 trail to get FDA to approve us, this means a p-value less than 0.01 (maybe some of you will disagree with this and this needs to be even lower) I have made some simulations with 90 patients and assuming a Normal distribution, which I am sharing:
https://drive.google.com/file/d/1kFIJ4DUN1IKJ...sp=sharing
This is my initial state of patients at day 0, I made both, placebo and Leronlimab arms the same. I used 46.7%, 63.3%, 43.3% and 16.7% for the percentage of patients with fever, cough, dyspnea and myalgia respectively. This following studies published (with a good number of patients) noting that this numbers vary wildly, for example, China reports more than 80% with fever and so on.
Then, I assumed that some placebo patients will worsen and some on Leronlimab will improve. The results are below:
https://drive.google.com/file/d/19LG5Q0oSsrJK...sp=sharing
Then in analyzing the data I obtain a p-value of 0.0094 (assuming a two-tailed test) that would get the FDA thinking ….
I am sharing this to illustrate what kind of results we would expect to have a potential approval. Basically, my simulation tells us that if the placebo group worsens from an average of 3.067 to 3.567 after 14 days and Leronlimab improves from the same value to 2.75 we would have a pretty nice result.
This is not difficult to achieve if Leronlimab works the way it is supposed to... frankly I expect a lower average in our cohort.
Once again, I have made several assumptions and this is ONLY for illustrative purposes. If you disagree with the data it is fine as I have no way of knowing the shape of patients before of after the trial, just made some "educated" guesses. Same regarding the model.
_____________________________________________________
Only for math fans:
Average at day zero: 3.067 (both cohorts)
Average at day 14 placebo: 3.567
Average at day 14 Leronlimab: 2.750
Standard Deviation Placebo day 0 and day 14 : 1.337 - 1.326
Standard Deviation Leronlimab day 0 and day 14 : 1.337 - 1.348
t=2.6543
Degrees freedom=88
Standard error difference=0.308
95% confidence of means [0.21-1.43]
(15)
(0)