(Total Views: 584)
Posted On: 01/08/2021 4:11:58 PM
Post# of 148902
Quote:
But for the logic to hold true, does it not require that CD12 report at least 50% mortality reduction?
No, the first thing te FDA will look at is ORR (overall response rate) or in this case the OS (overall survival rate or mortality rate) which shows what the level of clinical significance is. What level over placebo/ SOC the FDA is looking for we don't know exactly. With standard of care it may be as low as a 20% benefit.
Quote:
Many posters have offered statistical analysis here as to the mortality reduction which would be achieved for p<0.05.
At the interim, the mortality reduction for p<0.05 was around 50%.
At 390, the mortality reduction for p<0.05 was around 33%.
p value (probability value) just tells you whether the likelihood of a particular outcome is true. What is considered statistically significant is p =< .05. Which means there's a 5% or less chance that the outcome is wrong and a 95% chance that it's right. It does no good to have clinical significance if the results might be untrue.
The larger the patient population the more likely it is that the outcome is true. Where clinical significance comes into the picture is that the larger the percentage of people with a positive outcome the less patient population you need to prove a statistically significant probability value. With a fairly small population trial a statistically significant outcome will also show good clinical significance thus it's importance.
You could have an outlandishly large trial with 1 million patients and have a massively good statistical significance but clinical significance would be so small as to be useless. But that's one of the reasons they figure out the correct trial size beforehand.
Quote:
But since I am clueless on these statistics, I now believe that efficacy (in other words mortality reduction) is a variable component in the "power" function equation
The power function basically tells you what the odds are that if the trial was run again you'd come up with the same results.
To everyone - if I am incorrect in any of my explanations please feel free to correct me.
(5)
(0)
Scroll down for more posts ▼