(Total Views: 556)
Posted On: 03/12/2021 8:02:48 PM
Post# of 154090
ctd555 -
I'm happy with your approach - use the experienced deaths rates in each of the 4 groups and multiply those by the "fair share" number of over and under 65 patients in treatment and control.
My interpretation of the results is that we had 75 "junk" patients who weren't given any other Covid drug. I'm OK with being approved as a "post-one-failure drug initially.
With that in mind, a request.
Could you run the same calculations for the "any Covid treatment" and "dexy" groups and summarize for us? I think there is something to learn from both the relative, but especially absolute change in risk. I am also wondering if the P-values look like those from the logistic regression they used.
Even running this on the criticals, while numerically unstable, might give some insight.
Thanks in advance!
(Ohm - I hear your comment but don't think it applies under the null hyposthesis)
I'm happy with your approach - use the experienced deaths rates in each of the 4 groups and multiply those by the "fair share" number of over and under 65 patients in treatment and control.
My interpretation of the results is that we had 75 "junk" patients who weren't given any other Covid drug. I'm OK with being approved as a "post-one-failure drug initially.
With that in mind, a request.
Could you run the same calculations for the "any Covid treatment" and "dexy" groups and summarize for us? I think there is something to learn from both the relative, but especially absolute change in risk. I am also wondering if the P-values look like those from the logistic regression they used.
Even running this on the criticals, while numerically unstable, might give some insight.
Thanks in advance!
(Ohm - I hear your comment but don't think it applies under the null hyposthesis)
(1)
(0)
