The Excellence trial RSBQ results appears similar to the Avatar trial, in that interim responses (at least at week 4) were higher than end-of-study, so if we had used AUC then the endpoint may have been "stat-sig" (like Avatar). But AUC is not an appropriate endpoint gauge for a larger trial, so we dropped it for Excellence and went back to the normal at-week-12 score. It failed for at-week-12, as Avatar likely did. (The company didn't provided data or graphs for Avatar, either. Ah, Missling.) CGI just outright failed, unfortunately.
It looks like the company is correct that drug results (for RSBQ) were clearly positive and just barely not stat-sig, and would have been stat-sig if not for unusually strong placebo results. The company is going to try to explain away the placebo results, but that's a tough row to hoe with regulators. Nevertheless, Rett is a tough indication to study, which is why we as a company tried to improve the RSBQ endpoint's accuracy. Acadia navigated it successfully, even with similar difficulties.
The results do not appear to be an absolute failure, but a very near miss. We may have more Rett testing to do, though it looks like the company's first strategy is to explain away the placebo effects. Definitely worth a try.