Thanks for your feedback. My best guess is that i
Post# of 153886
(Total Views: 641)
Posted On: 03/14/2021 1:17:47 PM
Re: Borel Fields #82808
Thanks for your feedback. My best guess is that in the "junk patient" grouping is where a lot of the age skew went against LL. Because it doesn't really make sense for that group to do so much worse than placebo given what we know about the drug.
The reason unskewing the age groups was important was because LL had over 3:1 (instead of 2:1) of the 65+ group.
For your request on unskewing the "any treatment" and "dex," impacts will be miniscule because those were nearly 2:1 already. "Any treatment" was 1.97:1, and "dex" was 1.99:1
For the critical subgroup, the company included their analysis for this in the PR. 24.2% relative reduction with LL. Absolute reduction of 8.93%. I don't think they included a p-value, but it was weak at 0.487 due to the small sample size. This could use an unskew because it came in at 2.26:1, which is unfavorable for LL, so I will do that below. This would have resulted in 0.28 fewer deaths overall in the LL group, and an additional 0.74 deaths in the placebo. That wouldn't shift the overall mortality figures much because of course our critical population was relatively small.
Lastly, I included a projection for CD16 mortality using the critical result from CD12. It should be noted of course that the sample size in CD12 was relatively small so it would be unwise to put too much weight into this, I imagine the ultimate mortality figure will differ somewhat. The p-value for 140 patients at a 1:1 ratio (which I believe is what they said) would be weak at 0.442 if the same mortality rate repeats itself. Which I am sure is why they are using days to recovery instead as the PE.
Weak (in terms of p-values) mortality results was ultimately why I sold shares last Monday. I was surprised as I think most of us were. It sounds like the new CD16 design and PE will actually have a solid shot at meeting results.
Hopefully other longs can at least understand where I'm coming from instead of thinking I'm a paid short like Mr. Goosebumps and his 4.6% mortality projection.
The reason unskewing the age groups was important was because LL had over 3:1 (instead of 2:1) of the 65+ group.
For your request on unskewing the "any treatment" and "dex," impacts will be miniscule because those were nearly 2:1 already. "Any treatment" was 1.97:1, and "dex" was 1.99:1
For the critical subgroup, the company included their analysis for this in the PR. 24.2% relative reduction with LL. Absolute reduction of 8.93%. I don't think they included a p-value, but it was weak at 0.487 due to the small sample size. This could use an unskew because it came in at 2.26:1, which is unfavorable for LL, so I will do that below. This would have resulted in 0.28 fewer deaths overall in the LL group, and an additional 0.74 deaths in the placebo. That wouldn't shift the overall mortality figures much because of course our critical population was relatively small.
Lastly, I included a projection for CD16 mortality using the critical result from CD12. It should be noted of course that the sample size in CD12 was relatively small so it would be unwise to put too much weight into this, I imagine the ultimate mortality figure will differ somewhat. The p-value for 140 patients at a 1:1 ratio (which I believe is what they said) would be weak at 0.442 if the same mortality rate repeats itself. Which I am sure is why they are using days to recovery instead as the PE.
Weak (in terms of p-values) mortality results was ultimately why I sold shares last Monday. I was surprised as I think most of us were. It sounds like the new CD16 design and PE will actually have a solid shot at meeting results.
Hopefully other longs can at least understand where I'm coming from instead of thinking I'm a paid short like Mr. Goosebumps and his 4.6% mortality projection.
(10)
(0)