(Total Views: 418)
Posted On: 03/06/2021 10:55:37 AM
Post# of 153964
Two thoughts for this morning. Would appreciate comments:
I think that the 62 patients are a subset of the "critical patient" subset: The fit this definition: "Mechanically ventilated critically ill" --- this is the headline definition that ties to the 24%. Later in the press release, it lists 3 data points for "this critically ill population" - I'm pretty sure they mean the "mechanically ventilated critically ill"
Based on the trial definitions, this seems like a subset of the entire critical population.
2nd thought: I tried to "mock up" the 62 patients and came up with 22 in Placebo arm and 40 in LL arm. I'm guessing 8 dead in Placebo (36%) and 11 dead in LL (28%). This achieves the 24% reduction in death. The p-value on this (I think) is very high at .46 (due to the small "n"
.
I then looked at an entire trial (394 patients) of "mechanically vented critically ill" with 24% reduction. 132 placebo, with 48 dead. 262 LL with 72 dead. The p-value on this (I think) is .0667 Still a bit above .05. (Surprising to me). (Although I think Guru did tell us that the trial was under-powered if we didn't have a larger reduction in death).
For p-value, I'm using the calculator that TechGuru provided for our contest.
I don't think either of the above ideas are good for us in terms of getting a full approval. But, I also agree that we may find some type of EUA (here or abroad) after a few more weeks of politicking. "Shows benefit and does no harm."
I think that the 62 patients are a subset of the "critical patient" subset: The fit this definition: "Mechanically ventilated critically ill" --- this is the headline definition that ties to the 24%. Later in the press release, it lists 3 data points for "this critically ill population" - I'm pretty sure they mean the "mechanically ventilated critically ill"
Based on the trial definitions, this seems like a subset of the entire critical population.
2nd thought: I tried to "mock up" the 62 patients and came up with 22 in Placebo arm and 40 in LL arm. I'm guessing 8 dead in Placebo (36%) and 11 dead in LL (28%). This achieves the 24% reduction in death. The p-value on this (I think) is very high at .46 (due to the small "n"
. I then looked at an entire trial (394 patients) of "mechanically vented critically ill" with 24% reduction. 132 placebo, with 48 dead. 262 LL with 72 dead. The p-value on this (I think) is .0667 Still a bit above .05. (Surprising to me). (Although I think Guru did tell us that the trial was under-powered if we didn't have a larger reduction in death).
For p-value, I'm using the calculator that TechGuru provided for our contest.
I don't think either of the above ideas are good for us in terms of getting a full approval. But, I also agree that we may find some type of EUA (here or abroad) after a few more weeks of politicking. "Shows benefit and does no harm."
(4)
(0)
