A Wake Up Call for Employers: One-Third of Opioid Scripts Are Being Abused

Castlight, a health benefits platform provider focused on self-insured employers, published a report last week on the opioid crisis.  The authors were able to take a unique look at the problem through the lens of current data from self-insured employer clients (vs. latent data from public sources).

Lots of interesting data in the report, but here's the headline:

1 out of every 3 opioid prescriptions is being abused.   

I had three reactions, in the following order:

First, I knew that number would seem astronomically large to most people ("Seriously, one-third of all opioid scripts are being abused?  How can that be?")  Fact is there are more sad opioid statistics than most people realize.  It is the disease of not listening.  While it makes for admittedly depressing cocktail party conversation, it is a predictable interchange.  People know there's an issue... they just don't realize how broad and deep it goes.

Second, I personally thought that number seemed low.  While I recognize PRIUM's data is somewhat skewed by our focus on chronic and sub-acute pain (vs. acute pain), our physician consultants conclude that approximately 70% of the the medications we review are not medically necessary based on evidence based guidelines.  I recognize that "lack of medical necessity" and "abuse" are two different phenomenon, but when it comes to opioids specifically, the former tends to lead to the latter. So I thought 1/3 was low.

And that led me to my third reaction: How did Castlight define "abuse"?  They're looking at de-identified diagnosis and prescription data.  I wondered what methodology they used to identify opioid abuse.

Page 12 of the report details their approach:
Excluding cancer diagnoses and hospice care, Castlight defined abuse as meeting both of the following conditions:

1. Receiving greater than a cumulative 90-day supply of opioids; AND
2. Receiving an opioid prescription from four or more providers over the 5 year period between 2011 and 2015.  

Let's acknowledge that this is, at best, a proxy for abuse.  Might there be patients who are defined as "abusers" in the Castlight data who are not, in fact, opioid abusers?  Is it possible that a patient could receive opioid scripts from 4 or more docs over 5 years and not be an abuser?  Of course it's possible.  

But I think the Castlight approach is actually quite conservative.  Using a cut off of 4 prescribers likely leaves out a material number of patients who are abusing opioids but happen to secure their prescriptions regularly from as few as a single provider.  By the way, Castlight doesn't capture work comp data.  So we know (unfortunately) that 1/3 statistic is low.  

A wake up call for self-insured employers?  Hopefully. 

Michael 

On Twitter @PRIUM1