Being exposed to something like the Beers list, or ACB, or MARS, does not necessarily lead to practice changes, it's been pretty stable.
As for the patterns, we can keep it pretty simple
Patterns in Outpatient Benzodiazepine Prescribing in the United States
As said, primary care definitely among the increase, psychiatry remains pretty stable in usage despite increase knowledge of risk and adverse outcomes. I've got a few regarding prescribing in the 65+, but those are on my home computer at the moment.
Thanks for the paper, I appreciate to being exposed to data. Obviously a big paper as well but some initial thoughts:
A) ~80% of benzo scripts were written by non-psychiatrists. Roughly half came from PCPs
B) significant majority of scripts are continuation rather than new starts. You would expect this of course but it is also totally consistent with the idea I have been pointing out, namely that many psychiatrists are trying to manage the legacy of the benzo-happy days of yesteryear rather than lots of new starts.
C) how does AMCS identify physicians to survey? It it is based on insurance paneling or registration of any kind you are missing a big chunk of the specialty, introducing a systematic sampling error. This is much less true for most other specialties.
D) there is a drop in the rate at which psychiatrists in general are prescribing benzos per visit. Not as big as either of us would like I am sure but it is still there. I don't see the info I'd need to calculate effect size but maybe I am overlooking it.
E)Psychiatrists are still responsible for the same proportion of total benzo scripts more or less but given the rate of starting them is dropping, again consistent with the managing chronic benzos started a while back picture. You'd also almost hope that psychiatrists would be responsible for a disproportionate percentage of these scripts regardless given the indications for these meds, just like you'd kind of hope GI docs are more responsible for prescribing Linzess or whatever.
F) data's not in the paper but I would love to see a breakdown of prescribing patterns by years in practice. Unfortunately clinical medicine is like science in that enduring change often comes through generational shift rather than particular practitioners remaining flexible. I would put cash on the table betting that a disproportionate share of those scripts are coming from docs who trained before the harms of benzos were obvious.
G) the way in which they assessed indication, while of course reasonable given limitations of data, is going to make it very difficult to actually figure out how appropriate any of those prescriptions are. I agree with you that the answer to the question of how many BZDs should be prescribed is probably "less", it is also not zero and not a trivially small number either.
I will need to digest this further. I hope this convinces you I am engaging with this in good faith.
