Against Opioid Epidemic Skepticism

Recently, I’ve been seeing increasing skepticism about the causes, extent, nature, or even existence of the opioid epidemic in the USA. I usually categorize these somewhere between “reasonable criticisms of a breathless and overhyped media narrative” and “conspiracy theory”, and go on my merry way.

However, a couple of weeks ago Scott at SlateStarCodex gave a “comment of the week” to a particularly, apparently, evidence-based and “lucidly” put case for skepticism on the SSC subreddit. Something about this got under my skin — something in the author’s apparent certitude, the weight of his claims, and the deafening lack of a critical response from a community that is usually razor-sharp when it comes to poorly argued strong opinions. So, below is my criticism of this comment, which I’ve also sent to Scott since I’m worried he’s given a megaphone to a misleading and possibly intellectually dishonest piece.

I recommend reading the original comment and followup thread, to form your own opinions about the issue and the way it’s presented there.

Critique

Overall, I believe the commenter has a worrying tendency to cherry-pick his evidence in support of the thesis that the opioid crisis “has jack shit to do with the medical system or pharmaceutical industry”, whereas a more careful reading even of his own sources would lead to the opposite conclusion. Below I will look at his three main supporting claims in a bit more detail to demonstrate this, but the main evidence that flies in the face of his skepticism are:

1. Prescription opioids have caused hundreds of thousands of deaths and continue to kill more people per year than drunk driving, even though this rate is no longer increasing.
2. Prescription opioids, especially when prescribed in high doses or to at-risk populations, have a high chance (single digit percent) of causing dependency.
3. Much of the fentanyl and heroin crises (which have each recently eclipsed the prescription opioid crisis) are/have been enabled by users beginning on prescription opioids.

None of these necessarily indicate that the points in “Against Against Pseudoaddiction” are wrong, or that opioids have been a net negative for the American people. One could make the normative argument that the chronic pain treated outweighs the accidental deaths caused, or that consenting adults ought to be allowed to make medical decisions like this on their own. However, the idea that the opioid epidemic is unrelated to prescription painkillers is a completely unsupported and dangerous meme to have in our political consciousness.

Cherrypicked, misleading statistics

Below are some of the most egregious cherrypicks, which also make up the core of the evidence for this argument:

1. Fact: The rate of opiate addiction/abuse for long-term chronic pain patients without a prior history of abuse is 0.19%

This claim is based on a metastudy that examined “aberrant drug-related behaviors” and “abuse/addiction” in chronic pain patients prescribed opioids. Some of these studies preselected for patients with no history of drug abuse, leading to the following stats:

• Abuse/Addiction — 3.27% overall, 0.19% when prescreened
• Aberrant drug-related behaviors — 11.5% overall, 0.59% when prescreened

The implication of this is that, in non-vetted groups, opioid-prescribed chronic pain patients have an addiction/abuse rate of 3.27%, which is obviously quite high. The study further indicates that prescreening can be a fairly effective tool, cutting down addiction rates by over 90%; however, this does nothing to illustrate that opioid abuse cannot/does not result from legitimate prescription — just that the subpopulation most vulnerable to this abuse/addiction can be somewhat selected for beforehand.

2. Fact: The death rate from prescription opiates has not budged since 2006. The vast majority of opiate deaths are fentanyl, heroin or methadone.

This one is curious because it directly contradicts the source, rather than cherrypicking from it:

The first graph clearly shows a rise in prescribed opioids from 2000 until 2011, 5 years after the claim above. Further, the second graph shows that, while fentanyl and heroin now each kill more Americans than prescription drugs, (1) this change-over happened recently (2015), and (2) prescription drugs are still killing a lot of people in absolute terms (>4/100,000 or >13,000/year — more than the number of people killed by gun violence per year in the USA (excluding suicides))

3) “Fact: The fatal overdose mortality rate for patients with long-term opiate prescriptions is 17 per 100,000.”

I don’t have the full text of the article (behind a paywall), but the abstract does not come close to supporting this view. The study followed ~10,000 chronic pain patients who received three or more prescriptions for opioid painkillers, and recorded 51 overdoses and 6 deaths (60/100,000). Furthermore, when they break their data down by amount prescribed an obvious and troubling pattern emerges:

• 1–20mg/day — 0.2% annual overdose rate
• 50–99mg/day — 0.7% annual overdose rate
• 100+mg/day — 1.8% annual overdose rate(!)

Without access to the full text I cannot deeply critique the study’s methodology; however, if we do take it at its word, the picture is radically different from this author’s claims — even patients with relatively low total daily quantities prescribed had an annual overdose rate of tenfold the number he’s quoting (and, it would appear, 1/10 of these is fatal).

In aggregate, these sources — the exact sources that the author quoted — paint the picture that prescription opioids (1) have caused and are causing deaths at an alarming rate; (2) without proper vetting, opioid prescription for chronic pain can cause addiction/abuse in ~3% of the population (compared with 0.2% with vetting), and (3) the likelihood of overdose increases linearly or superlinearly with the amount prescribed, from 0.2% — 2% per year. None of this, to me, supports the notion that prescription opioids are harmless or unrelated to the present overdose epidemic.

An Incomplete Picture

Further, this line of argumentation misses one of the key points of the opioid epidemic theory — namely, that the deaths attributed to heroin or fentanyl in large part represent users who began their addiction with prescription drugs and, as these got harder to obtain or more expensive, moved on to cheaper, more potent, more deadly options. This is fundamentally a difficult thing to prove, but this source (for instance) points out:

A study of young, urban injection drug users interviewed in 2008 and 2009 found that 86 percent had used opioid pain relievers nonmedically prior to using heroin, and their initiation into nonmedical use was characterized by three main sources of opioids: family, friends, or personal prescriptions (Lankenau et al., 2012). This rate represents a shift from historical trends. Of people entering treatment for heroin addiction who began abusing opioids in the 1960s, more than 80 percent started with heroin. Of those who began abusing opioids in the 2000s, 75 percent reported that their first opioid was a prescription drug.

In missing the linkage between prescription opioid abuse and heroin (and, one assumes, fentanyl) addiction/overdose, the author misses one of the key reasons that prescription drugs are implicated in the ongoing crisis.

Small Red Flags

I am not an expert in this space by any means (I’m a software developer), so I don’t mean to gatekeep and say that only experts ought to be allowed into public discourse. However, CPlusPlusDeveloper shows himself at a couple of points to be unfamiliar with the basics of the topic at hand. For instance:

• He exclusively uses the word “opiate” rather than “opioid”. A casual Google search reveals that the former refers only to poppy-derived opioids, whereas the latter is the superset of drugs that operate in a morphine-like way. “Opiate epidemic” is not a term I’ve ever seen used in a legitimate publication:

• He mentions “cartefentanyl”, which isn’t a thing. Carfentanyl, on the other hand, is a synthetic opioid 100x more potent than fentanyl.

In general, I don’t like to dismiss people’s arguments based on missing details or unfamiliarity with the nuances of a topic; however, in a case where someone with no credentials is going against a fairly unified field (in this case, a field unified against the monied interest, which is no mean feat), I would expect that person to hold a high standard of detail-orientation to ensure they actually understood the issues at play. That was not evident from these low-level features, which was my first smell that this comment could be inaccurate.

Conclusion

With this context, the original comment is clearly either deliberately misleading or is mistaken — sentences like

America does have an opiate crisis, as can be seen in overdose rates that exploded from 2010–2019. However 100% of that increase was due to illicit opiates.

can be seen to be both misleading (opioid overdose rates were steadily climbing well before this, so 2010 is an arbitrary selection to prove his point), and incomplete (most studies contend that these illicit opioids had such a firm foothold due to the inroads made by prescription opioids). And sentences like

Frankly the people at Purdue Pharmaceuticals should receive the Nobel Prize in Medicine, as far as I’m concerned. They’ve done more to alleviate human suffering in America, then every single political and journalist put together. There is definitely an opiate crisis in the United States, but it has jack shit to do with the medical system or pharmaceutical industry.

seem less like innocuous contrarianism.