Evidence on the Atlas has a review date, not just a publication date. When one passes, the claim is flagged to its responsible office and — where policy requires it — withheld from the public page until it is re-verified.
This is why a figure here can be trusted to be current rather than merely to have been true once.