Reminds me of a comment I saw on a drug chemist's blog decades ago where they were talking about ACE inhibitors for Parkinsons' Disease - by dicking around with nerve conductivity and acetylcholine reuptake you could "boost" the failing nerves a bit, presumably until they fail altogether.

Anyway much of the discussion was on how hard it is to make suitable drugs, and someone said "We actually have some fairly inexpensive, reliable, and well-characterised ACE inhibitors with a nice long half-life and known efficacy, but the therapeutic window of nerve gas is pretty narrow..."

The folks I know who've been prescribed fentanyl weren't particularly upset about it. It is very potent, and thus measuring the dose properly is very important, but it's not particularly fussy beyond that.

The people you know may be obtaining it from less reliable apothecaries; that can be a real issue.

Fentanyl isn't all that unsafe, aside from the habituation potential.

When your measurement is "that looks about right" when purity is entirely unknown, not much really is all that safe.