When I go through the numbers in the Saturday pres
Post# of 154083
(Total Views: 385)
Posted On: 03/07/2021 6:35:44 PM
When I go through the numbers in the Saturday press release, I get an overall mortality reduction of 18.5%. They needed around 30% mortality reduction to get p<0.05.
384 total patients, 256 tx,128 placebo, 30 placebo deaths, 49 tx deaths, all my extractions from Saturday PR. Saturday's PR does not confirm 256 tx, 128 placebo so room for error there but probably not much error.
What happens with the age adjustment. 33% tx vs. 23% placebo over 65. 10% of the 256 tx arm patients experienced an excess death rate due to their age as compared to the placebo.
That's 25.6 patients. How many excess deaths are in the tx arm as a result of the age correlation?
If in the placebo arm, these patients were less than 65 years of age and they died at the rate of 12%.
But in the tx arm, these 25.6 patients were over 65 and died at the rate of 42%.
(42%-12%)*25.6=0.3*25.6=7.68 excess deaths in the treatment arm.
Removing these 7.68 deaths from the treatment arm yields a mortality reduction of 31.2%, which is right on the margin of p<0.05. I don't know the exact number of mortality reduction needed to hit p<0.05.
If this math is correct, the age adjustment nearly puts them over the top.
Of course, statistical manipulation can do many things. If they are cherry picking on the age adjustment then its meaningless.
384 total patients, 256 tx,128 placebo, 30 placebo deaths, 49 tx deaths, all my extractions from Saturday PR. Saturday's PR does not confirm 256 tx, 128 placebo so room for error there but probably not much error.
What happens with the age adjustment. 33% tx vs. 23% placebo over 65. 10% of the 256 tx arm patients experienced an excess death rate due to their age as compared to the placebo.
That's 25.6 patients. How many excess deaths are in the tx arm as a result of the age correlation?
If in the placebo arm, these patients were less than 65 years of age and they died at the rate of 12%.
But in the tx arm, these 25.6 patients were over 65 and died at the rate of 42%.
(42%-12%)*25.6=0.3*25.6=7.68 excess deaths in the treatment arm.
Removing these 7.68 deaths from the treatment arm yields a mortality reduction of 31.2%, which is right on the margin of p<0.05. I don't know the exact number of mortality reduction needed to hit p<0.05.
If this math is correct, the age adjustment nearly puts them over the top.
Of course, statistical manipulation can do many things. If they are cherry picking on the age adjustment then its meaningless.
(1)
(0)