I think it is extremely clear that something like this needs to have been in effect several years before any conclusions of success can be drawn. Failure would have been potentially much more obvious, but I would say ten years of observation and follow up would be optimal. Since optimal never applies to these things, five years should be the absolute minimum.