(Total Views: 1195)
Posted On: 11/25/2020 9:57:09 AM
Post# of 154115
Back to the numbers.
We don't know what efficacy criteria the board will be using, however, a good approach is to use the O'Brien-Fleming criteria which calls for a p-value around <=0.0151 for our interim .
How this translates to number of survivals in both arms?? Please refer to the tables below for two possible cases, 18 and 20 more deaths respectively (I am assuming that the deaths increase is proportionally slightly smaller from the mid-point)
So, for the case of 18 more deaths (in addition to the 45 we already had) we will need a 33/30 death ratio (Vyrologix/SOC) or 45% improvement for a value of 0.01.
Once again, I believe that it would be criminal not to approve a drug even with an improvement of, say, "only" 30% (a 36/27 split). At this improvement rate 75000 lives would have been saved. Yes 75000 more Americans would be alive today . Now FDA, please don't tell me this is not good enough !!!
Waiting anxiously for the readouts !!! The time has come once again ...
We don't know what efficacy criteria the board will be using, however, a good approach is to use the O'Brien-Fleming criteria which calls for a p-value around <=0.0151 for our interim .
How this translates to number of survivals in both arms?? Please refer to the tables below for two possible cases, 18 and 20 more deaths respectively (I am assuming that the deaths increase is proportionally slightly smaller from the mid-point)
So, for the case of 18 more deaths (in addition to the 45 we already had) we will need a 33/30 death ratio (Vyrologix/SOC) or 45% improvement for a value of 0.01.
Once again, I believe that it would be criminal not to approve a drug even with an improvement of, say, "only" 30% (a 36/27 split). At this improvement rate 75000 lives would have been saved. Yes 75000 more Americans would be alive today . Now FDA, please don't tell me this is not good enough !!!
Waiting anxiously for the readouts !!! The time has come once again ...
(21)
(0)
