The assumption was that the group proportions of the opioid crisis were relatively fixed, meaning that if you cracked down on one class of user then the overall size of the problem would shrink. But that turned out to be terribly wrong. The group proportions weren’t fixed at all.
