The changes don't address one issue which seemed odd in original 2 trials: the drug did much better at week 4 than week 12 for both dosage groups - if it were really placebo effect, wouldn't placebo effect be more obvious at week 4 than week 12? I'd think if the drug is really effective, placebo effect should go down rather than up over time. Could the drug work only in short term (4 weeks), but effect wanes in longer term (12 weeks)?
It is very clear that the relative efficacy is waning over time, but I would attribute it to placebo effect. Just not the way you are thinking.
In the control arm it will be a certainty that both severity and frequency will go down given enough time. This placebo effect has nothing at all to do with the patients reaction to the placebo, but just the natural course of the condition.
I would assume that what we are seeing is just the natural decay in symptoms on both arms, which gives less frequent and intense flashes to work on. Thus less ability to detect efficacy over time.
So the 12 (and 24) week targets are just harder to hit and nothing in the design can change this.
Back to the trial changes. The run-in made complete sense. The patients might be enrolling due to a temporary spike, and it would be good to filter that noise out,
I thought the change in the stat analysis from ANCOVA to non-parametric was more interesting, and perhaps questionable. Yes, it did reduce the ability of a few outliers to kill the trial. So if the trial really was well powered this might have been good insurance. Alas it also trades off some power, which turned out to be critical.
Did over-thinking, combined with one of DD's favorite issues, do them in?
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.