I just saw yet another bogus statistic about the opioid crisis, and specifically the underutilization of treatment, and I about lost my mind. It was in a New York Times (NYT) editorial, by their Editorial Board, titled: Want to Reduce Opioid Deaths? Get People the Medications They Need. The tagline says “Drugs like buprenorphine could sharply curb the nation’s opioid overdose crisis. But federal laws make it difficult for people who need such medications to get them.” And the opening puts a doozy of a mistruth out there:
A federally funded report released last week came to a striking conclusion: More than 80 percent of the roughly two million people struggling with opioid addiction in the United States are not being treated with the medications most likely to nudge them into remission or prevent them from overdosing. This denial of care is so pervasive and egregious, the report’s authors found, that it amounts to a serious ethical breach on the part of both health care providers and the criminal justice system.
To back up that “80 percent” untreated with medications claim, the NYT links to a report by the National Academies of Science, Engineering, and Medicine (NASEM) which states:
Despite the preponderance of evidence that medications to treat OUD are safe and effective, they remain highly underused in the United States. In 2017, about 80 percent of people who needed OUD treatment did not receive it, amounting to some 1.7 million people (Park-Lee et al., 2017)
In raw numbers then, the NASEM claim is (in rounded numbers) that 450,000 out of 2.1 million people with Opioid Use Disorder (OUD) get any kind treatment. They clearly state on that same page that the number who receives treatment specifically with medications is unknown to them. The NYT turns the any kind of treatment stat into a treatment with medications stat. So that’s one big error already along the chain of misinformation, but it’s not the main error.
The big problem is that there are far more than 450,000 people taking medications for OUD. That means far less than “80 percent of the roughly two million people struggling with opioid addiction in the United States are not being treated with the medications most likely to nudge them into remission or prevent them from overdosing.”
In fact, there are nearly 2 million people taking the medications mentioned by the NYT and NASEM (buprenorphine, methadone, or naltrexone) for OUD. This fact doesn’t necessarily bring the 80% untreated figure down to 0% though, because some of those people wouldn’t count as currently having OUD because some are stabilized for a year or more (makes you wonder why they’re taking meds for a condition they aren’t considered to be experiencing any longer, but that’s beside the point). We’ll parse that 2 million figure later, but I want to get to the main point first.
Here’s the major error made by the NYT, NASEM, and so many others.
The error comes down to a single term used in SAMHSA’s National Survey on Drug Use and Health (NSDUH) data that they cite: specialty treatment.
When the Substance Abuse and Mental Health Services Administration (SAMHSA) collects data on how many people receive addiction treatment, they usually only look at and report on what they call specialty treatment. Here is how they define that and it’s counterpart, nonspecialty treatment, in NSDUH:
Specialty treatment refers to substance use treatment at a hospital (only as an inpatient), a drug or alcohol rehabilitation facility (as an inpatient or outpatient), or a mental health center. This NSDUH definition historically has not considered emergency rooms, private doctors’ offices, prisons or jails, and self-help groups to be specialty substance use treatment facilities; in this report, these other locations are referred to as nonspecialty treatment facilities.
The key here, with the specialty/nonspecialty treatment definitions, is this: when private doctors prescribe buprenorphine for people with OUD (or naltrexone, but that’s a much smaller factor), their patients do not get counted in specialty treatment statistics. The claim that “More than 80 percent of the roughly two million people struggling with opioid addiction in the United States are not being treated with the medications” is based on specialty treatment statistics. Which means that it’s missing a whole group of people with OUD that get buprenorphine through private physicians.
My estimate (originally published and explained on BRI’s website) is that in the year referenced by NASEM, almost 1.6 million people received buprenorphine from private physicians. The estimate is based on figures released by the Department of Health and Human Services (DHHS) . DHHS reported numbers of people with OUD receiving buprenorphine from private physicians (i.e. nonspecialty treatment) in the Federal Register in 2012:
“There is now even more experience with buprenorphine in the treatment of opioid dependence. Since 2002, almost 22,000 physicians have sought and obtained the federal certification to prescribe buprenorphine products. According to the DEA Automated Reports Consolidated Orders System (ARCOS), over 190 million dosage units were distributed to pharmacies in 2010, a more than fourfold increase from the almost 40 million dosage units distributed in 2006. It should be noted that only 1.1 million dosage units were distributed to OTPs during 2010. In addition, almost 800,000 individuals received buprenorphine addiction treatment prescriptions from office-based physicians in 2010, increasing almost fivefold from the 150,000 estimated in 2006. (REF 4).”
Dosage unit distribution figures for other years is not publicly available, but the DEA’s figures on number of grams of buprenorphine can be used to estimate the other years. For both 2006 and 2010, the number of patients receiving buprenorphine from private physicians is roughly half of the number of grams distributed in those years. For example, in 2010, about 1.5 million grams of bupe were distributed, while “almost 800,000 individuals received buprenorphine… from office-based physicians.” In 2016, DEA’s ARCOS Reports shows that almost 3 million grams of buprenorphine were distributed – using more exact percentages based on this and other data would mean that the private physician bupe patient pool had grown to almost 1.6 million in 2016.
To bolster this estimate, I refer to the University of Maryland’s Center for Substance Abuse Research, who reported that:
The number of patients receiving a prescription for Subutex or Suboxone [brand name versions of buprenorphine] from U.S. outpatient retail pharmacies increased from slightly less than 20,000 in 2003 to more than 600,000 in 2009.
Again, the DEA’s figures on grams of buprenorphine distributed that year (1.225 million) indicates that the number of patients getting it through pharmacies is roughly half the number of grams distributed for the year.
Further, an extensive clinical review submitted to the FDA on an injectable form of buprenorphine included a figure that also tracks with my simple estimate formula:
Approximately 10.7 million prescriptions from outpatient retail pharmacies were dispensed and approximately 1 million patients received a dispensed prescription for buprenorphine tablets or films during 2012.
Again in 2012, the number of grams distributed, just over 2 million, is about twice the number of patients (1 million) known to pick up their buprenorphine from pharmacies rather than “specialty treatment” sources.
And now let’s note that none of these patients are counted as having received medications in the NSDUH Report (Park-Lee 2017) cited by the National Academies opioid report, which was then bastardized and cited by the New York Times. Those folks are referring to specialty treatment. I know this because their claim that 80% aren’t getting treatment doesn’t appear in the Park-Lee citation anywhere, but probably comes from a sister report: the NSDUH 2016 Detailed Tables. It shows that 453,000 people with OUD received treatment at Specialty Facilities that year (Table5.35A). Another table shows that this number represents 21.1% of people with OUD (Table 5.35B). This is probably where the NASEM Opioid Report authors they came up with the “about 80 percent of people who needed OUD treatment did not receive it, amounting to some 1.7 million people” claim. The 1.7 million figure would suggest this since Table 6.43A of the same report shows 2,144,000 people 12 and older with OUD. Thus 2,144,000 minus 453,000 = 1,691,000, or “some 1.7 million.”
Now let me give you a rundown of how many buprenorphine patients show up in specialty treatment figures versus the private physician (nonspecialty treatment) figures cited above:
2006: 2,042 specialty treatment versus 150,000 nonspecialty treatment
2009: UNKNOWN specialty treatment versus 600,000 nonspecialty treatment
2010: 27,456 specialty treatment versus 800,000 nonspecialty treatment
2012: 39,223 specialty treatment versus >1,000,000 nonspecialty treatment
2016: 61,486 specialty treatment versus 1,577,000 nonspecialty treatment
[Those figures on OUD patients receiving buprenorphine in specialty treatment come from SAMHSA’s annual N-SSATS reports.]
SAMHSA’s various reports never include the numbers of nonspecialty buprenorphine patients I cited above. Based on that comparison of figures above, you can see that anyone who cites SAMHSA (and their NSDUH or N-SSATS reports) to support the claim that only a tiny fraction of people with OUD get treatment with medications, is making a grave error. It could be cleared up if we simply acknowledged the definition of “specialty treatment” in the data, and remain aware that it doesn’t include people who get their buprenorphine from private doctors.
And that brings up another important point. The activist/journalists who write countless stories in the NYT and other popular online news sources extolling the virtues of buprenorphine and screaming hysterically that nobody is taking it, know better. They know this distinction, and I know they know it. They know that the big push behind buprenorphine was based on the idea that it could be safely prescribed from doctors offices, and provided a smoother more accessible option than dealing with tightly regulated methadone clinics. In fact, the NYT knows this too.
In 2018, the NYT published an interview with:
Dr. Elinore McCance‐Katz, the Trump administration’s director of the Substance Abuse and Mental Health Services Administration, helped pioneer opioid addiction treatment with buprenorphine in clinical trials in the 1990s. She also helped create the training for doctors who want to prescribe it.
In which they asked “Do you think the number of primary care providers who prescribe buprenorphine is growing fast enough?” And she responded:
What I will say is that we have a lot of work to do. When we first thought about this treatment, it was really thought about as a way to integrate treatment of opioid use disorders, opioid addiction, into primary care. Because we know that many, many people with opioid problems have other medical problems, and sometimes they have psychiatric problems, too. Psychiatry has had much more uptake on this than has primary care. But the idea was you eliminate the stigma by just having them be another patient in the waiting room. That was the hope.
So they know there has been a major push to give opioid treatment with buprenorphine outside of addiction treatment settings. So, particularly, do many of the folks who write for the NYT, and so it’s sad to see them let these bogus statistics propagate.
What about my 2 million figure?
The numbers I listed above for 2016 add up to about 1,640,000 people on buprenorphine. Additionally, there were 354,443 people receiving methadone in specialty treatment settings that year. I don’t have tight source for this at the moment, but there were probably about 10,000 on Naltrexone. This puts us right at about 2,000,000 people taking the various medications for OUD.
However, and this is important, not all of those medicated people would be counted in the 2,144,000 people with OUD in 2016 reported by SAMHSA. The reason for this is that SAMHSA counts people who have fit the diagnosis for Substance Use Disorders over the past 12 months when collecting this data. Some portion of my estimated 2 million would not count as having OUD over the past 12 months, because they are stabilized, not using opioids, and/or at least not exhibiting the symptoms listed in OUD diagnostic criteria. But just how many of them are stabilized is not clear.
What I know is this – the highest, most generous estimates for retention for more than 12 months on methadone is 75%, and for buprenorphine it is 50%. Trials suggest that even among patients retained for a year success rates are low. So where does that put us? This would be much more of a guessing game to figure out, but I think it’s safe to say that less than 50% of that 2 million is stabilized. This would mean that more than 1 million people who currently fit the diagnosis for OUD are being treated with the medications buprenorphine (Suboxone, Subutex), methadone, or naltrexone (Vivitrol). That is to say, contrary to both the New York Times editorial and The National Academies report, approximately 50% of people with opioid use disorders ARE being treated “with the medications most likely to nudge them into remission or prevent them from overdosing.”
That raises an important question: why do fatal overdose rates keep rising? I don’t think these medications are working. To that point, I’ll just say that back in 2002, there were just over 200,000 people on medications for OUD, and about 1.7 million people with OUD. At that point, the annual fatal opioid overdose rate was a quarter of what it is now.
For your takeaway message, please remember that “specialty treatment” statistics don’t come close to accurately representing how many people with OUD receive treatment, let alone treatment with medications. Feel free to send this along to anyone posting such statistics about OUD medications. They need to be made aware of the error they are making, assuming they’ve made it in earnest.
NOTE: Even though NASEM’s 80% figure refers to 2017, the citation is for 2016 data, so I used 2016 data in my analysis. I think they made an error when they said 2017, because they publication date for the data they based it on was 2017.
NOTE: I realize my estimate may not be perfect, but it’s far more accurate than the claims made by the NYT and NASEM. If the DHHS and DEA would release more current figures, then I wouldn’t need to estimate. I believe their numbers of patients are sourced from a private company that tracks prescriptions, and charges a lot of money for their information. I can’t afford to buy this information, but certainly, the government can and should if they’re going to be the authority on these matters.
NOTE: I personally think my loose estimate that more than 50% of the 2 million of the MAT patients currently fit the diagnosis for OUD is low. But I erred on the side of giving bupe & methadone the benefit of the doubt in effectiveness. But even with this low estimate, there are still far more people with OUD receiving meds than the NYT, NASEM, or anyone else claiming underutilization of OUD meds would ever care to admit.
NOTE: there are far more problems in the NASEM report than I mentioned here. It would’ve been even more exhausting if I’d gone into those. But I may tackle them at a later date.
