Reposting some useful information from "that other board" regarding unblinding & data analysis.
"Hi Everyone, I am long time reader of this board but just created my account to clarify some of the technicalities regarding the unblinding process. I am statistical programmer, worked on numerous topline reports and was part of 3 successful BLAs in my short career of 10 years.
01/13 is the last patient last visit day which is not the same as unblinding. As Rockleo pointed out the sites will have 24hrs to enter the data in to the EDC system. Data monitoring and Data cleaning will be performed and queriers will be raised and the sites will have to respond to those queries. There is vendor data to be finalized as well. The safety labs, cytokine panels etc all come from the vendors. Each and every data point has to be reconciled and the database has to be locked. Pandemic or No pandemic the FDA is very strict about Data Integrity. If there are any mistakes during this process, the database has to be unlocked, corrected and locked. Trust me nobody wants to go through that.
Once the database is unlocked, unblinding will be authorized and statistical analysis is done. There are multiple steps in this process. The raw data from the EDC and vendors need to be standardized into SDTM and ADaM formats for analysis. The FDA only accepts SDTM and ADaM datasets and every step has to be validated. The programming to do this should already be in place and if Amarex is competent the whole process should not take more than a couple of days. Usually the sponsor which in this case Cytodyn has a dedicated team working on this producing the Topline report.
In short if AMAREX is on top of everything, we should be able to hear something sometime by mid to end of next week which is my most optimistic estimate. If they are bit sloppy it will add an additional week or two.
The FDA knows this clearly and it might be the reason for the authorization of Open label cohort.
Respect for all the senior members of this board. Rockleo, Black-ops and many others.
One more point i would like to make is regarding the no of deaths.
At 50% DSMC review we were told 45 deaths were on the trial but it was not broken down to how many deaths were within 28 days. The DSMC came back saying we would like to do a review at 75% and also want to look at Day 42. This is because some of the deaths were post Day 28 and they wanted to include these as well which might push us towards statistical significance.
If there was no separation from Placebo in terms of efficacy, the trial would have been stopped by the DSMC and everyone knows that.
For every trial, there will be a safety and pharmacovigilance group monitoring the SAEs and deaths. They have to report the unblinded case reports of the patients who died to the FDA immediately with in 24 hrs. The FDA will stop the trail if they see anything unusual.
Now we know there are 87 deaths and we know there will be more. It is sad to say this but the more no of deaths, the better chance to have a positive outcome.
Does everyone think Nader gave out that number because someone posted a random question and he was innocently answering that. Wrong. I was asking for this number prior to every conference call which was never answered. Cytodyn/Nader puts it out there only when they want to do it. They know it is very favourable to the drug and that's the reason they gave out that number in the last conference call.
Being said that, all of the above point to Clinical significance. Whether it shows statistical significance is what remains to be seen. the FDA knows that hence the billing codes were given and the open label cohort was initiated.
In this pandemic nobody is looking for statistical significance. For an EUA the FDA needs a well conducted trial with a decent sample size. For CD10 the sample size is very low. We only had 45 subjects with a total symptom score >=4. As Rockleo pointed out, you cannot treat subjects who don't have enough symptoms and expect them to show reduction in symptoms. All treatments which received EUA had completed their trials with enough sample size which we are on the verge of doing."
"Hi Everyone, I am long time reader of this board but just created my account to clarify some of the technicalities regarding the unblinding process. I am statistical programmer, worked on numerous topline reports and was part of 3 successful BLAs in my short career of 10 years.
01/13 is the last patient last visit day which is not the same as unblinding. As Rockleo pointed out the sites will have 24hrs to enter the data in to the EDC system. Data monitoring and Data cleaning will be performed and queriers will be raised and the sites will have to respond to those queries. There is vendor data to be finalized as well. The safety labs, cytokine panels etc all come from the vendors. Each and every data point has to be reconciled and the database has to be locked. Pandemic or No pandemic the FDA is very strict about Data Integrity. If there are any mistakes during this process, the database has to be unlocked, corrected and locked. Trust me nobody wants to go through that.
Once the database is unlocked, unblinding will be authorized and statistical analysis is done. There are multiple steps in this process. The raw data from the EDC and vendors need to be standardized into SDTM and ADaM formats for analysis. The FDA only accepts SDTM and ADaM datasets and every step has to be validated. The programming to do this should already be in place and if Amarex is competent the whole process should not take more than a couple of days. Usually the sponsor which in this case Cytodyn has a dedicated team working on this producing the Topline report.
In short if AMAREX is on top of everything, we should be able to hear something sometime by mid to end of next week which is my most optimistic estimate. If they are bit sloppy it will add an additional week or two.
The FDA knows this clearly and it might be the reason for the authorization of Open label cohort.
Respect for all the senior members of this board. Rockleo, Black-ops and many others.
One more point i would like to make is regarding the no of deaths.
At 50% DSMC review we were told 45 deaths were on the trial but it was not broken down to how many deaths were within 28 days. The DSMC came back saying we would like to do a review at 75% and also want to look at Day 42. This is because some of the deaths were post Day 28 and they wanted to include these as well which might push us towards statistical significance.
If there was no separation from Placebo in terms of efficacy, the trial would have been stopped by the DSMC and everyone knows that.
For every trial, there will be a safety and pharmacovigilance group monitoring the SAEs and deaths. They have to report the unblinded case reports of the patients who died to the FDA immediately with in 24 hrs. The FDA will stop the trail if they see anything unusual.
Now we know there are 87 deaths and we know there will be more. It is sad to say this but the more no of deaths, the better chance to have a positive outcome.
Does everyone think Nader gave out that number because someone posted a random question and he was innocently answering that. Wrong. I was asking for this number prior to every conference call which was never answered. Cytodyn/Nader puts it out there only when they want to do it. They know it is very favourable to the drug and that's the reason they gave out that number in the last conference call.
Being said that, all of the above point to Clinical significance. Whether it shows statistical significance is what remains to be seen. the FDA knows that hence the billing codes were given and the open label cohort was initiated.
In this pandemic nobody is looking for statistical significance. For an EUA the FDA needs a well conducted trial with a decent sample size. For CD10 the sample size is very low. We only had 45 subjects with a total symptom score >=4. As Rockleo pointed out, you cannot treat subjects who don't have enough symptoms and expect them to show reduction in symptoms. All treatments which received EUA had completed their trials with enough sample size which we are on the verge of doing."
16
