Remember that the statistical improvement was limited to one...

Remember that the statistical improvement was limited to one domain of cognition only, and one that their a priori analysis was not designed to detect. It was discovered only in post-hoc analysis (i.e data-trawling)

"Statistical significance" refers to a less than 1 in 20 chance that the result was obtained by chance ("p<0.05").

Thus, if you analyse 40 potential outcomes, the odds are good that at least one will show 'statistical significance.'

That's why post-hoc analyses are often suspect. The data on improvement to which you refer were released several weeks (?if not months) after the initial results were announced, suggesting they trawled the data to find a positive outcome.

The patients in the PBT2 trial thus DIDN't "get better" in any global sense....they demonstrated "significant" (with all the caveats that implies) improvement in one limited domain of a broader cognitive assessment. OVERALL, the PBT2-treated patients still got worse.

Having said that, I do think it's wrong for the market (and the scientific community in general), to focus on "improvement" as the only positive outcome.....as I've said before on this board, the brain has limited capacity to repair itself once damaged, and a slowing in the rate of decline is the most positive likely outcome for any "disease-modifying" drug. It's wrong, therefore, for the market to punish drugs that fail to show improvement. I believe the science, and neurological reality, is something that the market fails to understand...and not unreasonably, as it's very complex!