Moreover, certain types of marijuana do a superlative good job of providing non-opiod relief from chronic pain and aren't psychoactive, so in the same way you or a co-worker might have some Tylenol if they're feeling physical pain while at the office, you or a co-worker could use marijuana to manage chronic pain, without their mental state being noticably altered.

That's not to say marijuana cannot be abused nor is it a silver bullet, but in the middle of an opioid epidemic, holy shit, maybe something many people are illegally using to manage chronic pain without causing an epidemic; maybe it might be worth studying? The vast majority of users continue to be productive members of society; judging by the number of states that have legalized marijuana, we should be in the middle of a marijuana crisis compared to opioids, yet no such crisis exists. Yet the DEA classifies marijuana as having zero medical benefit, which makes it almost impossible, practically speaking, to study properly.

"an average of 1,826 fewer doses" is what percent?

From eyeballing exhibits 2 and 3 in the paper, it looks like pain medication doses dropped 5-10%. That seems like a lot in only a couple of years.

5 patients dosed once per day... Not a lot?

Yes, but without knowing what the total number of prescriptions is... it could be 50%, or 0.005%

By the way, props to your nick :)