I think there isn't enough emphasis in the article on how the data started to look exceedingly terrible right before the information was severely restricted/stopped.

Seems relevant.