Maybe they completely reversed the causality, it's a demand shock not a supply shock. There are less users because they died, and they died pretty fast compared to previous opioid users. As demand diminished there was over supply and to maintain their margins provider had to lower the supply. QED.

As it's a pretty simple hypothesis to test and that it was not maybe imply that the conclusion is politically motivated. Supply-shock imply that something was done and it worked, but that the problem solved itself is not as palatable for someone politically motivated like an administration.