Medical use of opioids not driver of overdoses

Below is a series of Twitter postings by @JohnTuckerPhD from April 22. where he lays out arguments, with excellent graphs, showing there’s no correlation between prescribing and overdosing.

I promised to send @raj_mehta the data behind our discussion of opiate prescribing-related harms yesterday. Apologies to those who have seen this data ad nauseum, but I am going to post it all here together.

The “pain pt becomes addicted, overdoses and dies” narrative is probably happening, but its at best a very minor component of the OD epidemic. The demographics of Rxing and ODing are very, very different. 

No correlation of OD or opioid related ER visits with Rxing by state:

 

States that have reduce opioid Rxing have not seen a decrease in ODs relative to those that have not. While not statistically significant, reduced Rxing shows a weakly positive correlation with increased ODs:

 

The age demographics are all wrong too. Express Scripts has summarized long-term opiate Rxing by age group. These are related to the OD rates of the same age groups in the graph below.

Those over 65 get most of the Rxs, but have the lowest OD rate among adults.

 

The central role of diversion, not medical use in the OD epidemic is further supported by longitudinal analysis. Those who were > 50 y.o. in 1999 have seen almost no increase in OD rates in spite of getting most of the opioid Rxs.

 

This data fits with the results of a Cochrane review, which found the use of opioids was associated with minimal risk of addiction in well-selected chronic pain patients.
http://www.cochrane.org/CD006605/SYMPT_opioids-long-term-treatment-noncancer-pain

And I guess we can add to this the CDC’s recent data that Rxing of opioids dropped 10% last year, but overdoses increased 30%.

How long can we continue to ignore the data?