LveGun-TakeCannoli, It's counterintuitive to th
Post# of 153778
(Total Views: 433)
Posted On: 04/05/2020 7:05:55 PM
LveGun-TakeCannoli,
Our results will be compared against the outcomes of the "placebo" arm. If you take out a negative outcome from this arm this will make it more difficult for us to have a significant improvement (p number). So, this will bias the results against us. From the humanitarian perspective I have nothing against them doing so. But statistically speaking wit will not be good for us, unless, of course they count that patients as "failed".
No, lets suppose that during the trial some patients are so ill that need to be taken to a different treatment. This might be Lero or another drug, and if Lero's patients are already above statistical significance against what is currently being achieved. The trial is then terminated. As I mentioned for this to happen there is need for monitoring, which is not possible with double-blinded tests. However, once a patient has been "taken out" of the trial (put into another treatment) it might be possible to "see" which arm was he/she in. It depends on the design of the trial.
Yes, ideally the number of patients should remain the same. Less patients less statistical relevance, however FDA takes into account few patients dropping out for whatever reason. The placebo arm provides the probability of success (or failure). The Leronlimab Arm provides the test to the null theory (the drug does not work). Say at the end 80 patients (out of 114) "finish" the two weeks in "placebo" arm. And 20 are alive (others left trial for other reasons). The probability of success is 0.25 (0.75 failure). Then the Leronlimab arm will use this to determine the significance of the results.
For the good of the patients I really hope the numbers are higher !!, these are just examples.
Quote:
It's counterintuitive to think we'd move patients over to Leronlimab and that would bias results against Leronlimab. I think this is a high probability given the severity of these patients and doctors being unwilling to risk their patients' lives with near certain death knocking on their doors. Plus, they need the beds/ventilators for the next severe patient.
Our results will be compared against the outcomes of the "placebo" arm. If you take out a negative outcome from this arm this will make it more difficult for us to have a significant improvement (p number). So, this will bias the results against us. From the humanitarian perspective I have nothing against them doing so. But statistically speaking wit will not be good for us, unless, of course they count that patients as "failed".
Quote:
Are you saying patients on Leronlimab will be taken off Leronlimab (because they're doing better) and returned to placebo? How do the doctors even know what arm the patient is in if it's double blinded? Would love a little clarity here. If I was a family member, I'd have a fit.
No, lets suppose that during the trial some patients are so ill that need to be taken to a different treatment. This might be Lero or another drug, and if Lero's patients are already above statistical significance against what is currently being achieved. The trial is then terminated. As I mentioned for this to happen there is need for monitoring, which is not possible with double-blinded tests. However, once a patient has been "taken out" of the trial (put into another treatment) it might be possible to "see" which arm was he/she in. It depends on the design of the trial.
Quote:
In the long run, does this undermine our data when science looks back at the statistics? Wouldn't a smaller sample size lead to lower statistical relevance/study power? Or, does the world just throw out statistics when the control group's patients improve greatly vs placebo?
Yes, ideally the number of patients should remain the same. Less patients less statistical relevance, however FDA takes into account few patients dropping out for whatever reason. The placebo arm provides the probability of success (or failure). The Leronlimab Arm provides the test to the null theory (the drug does not work). Say at the end 80 patients (out of 114) "finish" the two weeks in "placebo" arm. And 20 are alive (others left trial for other reasons). The probability of success is 0.25 (0.75 failure). Then the Leronlimab arm will use this to determine the significance of the results.
For the good of the patients I really hope the numbers are higher !!, these are just examples.
(1)
(0)