Thread by @StefanKertesz

1/A new @bmj_latest paper is seen by some as showing that 2 prior reports were incorrect in reporting major drops in opioid prescribing. Rather, this fine paper shows a problem in how US Rx reductions were achieved @mollyjeffery, @jsross119 @nilaydshah1
bmj.com/content/362/bm…

2/the headline I would offer is: “large, well-documented US opioid Rx reductions have not involved changing a long-term American habit of sloppy short-term prescriptions, but instead reflect changes in care for a small number of sick people on long-term prescriptions”

3//Here’s the report. Key endpoint is important (% of persons receiving an opioid Rx by quarter). No major decline found in 3 distinct insurance groups: bmj.com/content/362/bm…

4/to accept the conventional spin of “no drop in opioid prescriptions”, you have to reject 3 commercial reports as erroneous. Here’s one, from CDC’s excerpt of QuintilesIMS, 18.2% drop from peak in Rx by 2016, and 48% reduction in hi-dose Rx’s since peak of 2008, 10 years ago

5/Here’s another, from IQVIA showing 12% overall drop in 2017, and 16.1% drop in high-dose in 2017 alone:
iqvia.com/gated-form-pag…

6/ In reality, there is no contradiction, once you see that the BMJ paper’s headline result concerns persons who receive ANY opioid prescription, and most people who receive ANY opioid receive that Rx for short durations, at low dose.

7/In fact, many of these short-term opioid prescriptions are flagrantly stupid (or unnecessarily long), never consumed, harmful to the patient or ripe for diversion. Witness the report that ¼ of ankle sprains result in opioid Rx. washingtonpost.com/national/healt…

8/So how did total opioid Rx prescribed go down so MUCH in the US, if the % of persons receiving opioid Rx’s did not? By changing the care of long-term recipients. The BMJ paper itself tell us long-term recipients were a tiny percentage of the overall recipients:

9/Here, as in a prior publication in @AnnalsIM we find that 5% of opioid receiving patients account for 60% of morphine milligram equivalents
annals.org/aim/article-ab…

10/And indeed the @bmj_latest authors do confirm that it’s in those high MME users that reductions were prominent, although they state it a little opaquely, “use was somewhat less concentrated in the top percentiles of opioid users over time”

11/Of course we already knew that, because when you add together the CDC/Quintiles and more recent IQVIA reports, one finds that total high-dose (>90 MME) prescribing dropped >50% since 2008.

12/It’s much easier for insurers to disapprove, & regulators to target, high doses than for anyone to regulate ankles in ER's. I’m NOT saying that dose escalations were wise or always helpful in the first place. Just that high dose, long term patients are the easy targets now.

13/But note: “easy targets” for regulatory action are also the people many studies show us are sick with multiple conditions, where comprehensive care would be the more humane approach to care, whether they are on opioids now, or not ncbi.nlm.nih.gov/pmc/articles/P…

14/To repeat, the biggest achievable drop in MME, at the system level, involves big changes to the care of small number vulnerable patients, a step sometimes tolerated and sometimes traumatic, as we wrote last week: nature.com/articles/s4139…

15/ In our recent paper in @addictionjrnl, we call this the “arithmetic problem” where the perceived need to make a number drop quickly to please policymakers is in tension with our moral obligations as caregivers onlinelibrary.wiley.com/doi/abs/10.111…

16/In sum, new report doesn't contradict the prior ones. It shows us HOW we are making prescriptions drop. Personally, I think we could make better choices. huffingtonpost.com/entry/governme…