Opioids for the masses?

This has long seemed to me an understudied topic, so I was interested to read the job market paper of Angela E. Kilby, who is on the market this year from MIT.  And she does what I like to see in a paper, namely try to figure out whether some practice or institution is actually worth it.

The background is this: “…In the face of concerns that undertreatment of pain was a “serious public health issue,” medically indicated use of these drugs over the past 15 years has increased dramatically, and attitudes have liberalized towards the use of opioids for chronic non-cancer pain.”

When it comes to the increased use of opioids, she finds the following trade-offs:

1. Since 1999, there has been a fourfold increase in drug overdose deaths linked to opiod pain relievers.  In 2013, the number of opiate-linked overdose deaths was 25,117, a higher number than I was expecting.  (But note that most of these can no longer be reduced by the feasible interventions under consideration.)

2. The increased use of opioids seems to pass a cost-benefit test, compared to the passage of a tougher Prescription Monitoring Plan.  With a host of caveats and qualifiers, she measures the pain reduction and other benefits from looser regulation at $12.1 billion a year and the costs of higher addiction rates, again from looser regulation, at $7.3 billion per year.

There is much more to it than what I am reporting, and in general I believe economists do not devote enough attention to studying the topic of pain.


Pain is really hard to measure. This makes it less worth studying, since it is harder to make any robust findings about it, but it also means it is probably underrated in terms of importance.

Pain is a huge issue and it certainly ought to be treated. However, there is really no good evidence to support the long-term use of opioids for chronic pain. There are no RCTs of their use that have lasted more than a few months. However, we do know that the longer people stay on these meds their doses increase and there is good evidence for dose-dependent risk of serious harm.

"With a host of caveats and qualifiers": jolly good. Now how about a stab at error bars on her $12.1 billion a year and $7.3 billion per year? No error bars, no pass in school science labs.

Opioid-related deaths appear to be a growing problem for whites and American Indians, but much less so for blacks and Hispanics. Jason Bayz has graphed much of the relevant data:


It just hasn't happened to them yet. These drug problems tend to start in one place or race and then spread. Sometimes it takes a different form but it gets around.

Could be if prescription painkillers were the vector to heroin.

Mr Sailer,

Isn't that we are already seeing?

Lots of middle aged white people with back pain were given powerful narcotics in the form of Oxycontin. They became dependent on these painkillers but then the government cracked down on pill mills and they couldn't get access to their preferred opioids.

Eventually they turn to heroin to satisfy their addiction.

Judging from the increase in 45-54 year old death rates among whites by state (as collected by blogger Sendil), the White Death might be particularly a plague on the Scots-Irish:


No surprise there. Scots-irish is the dominant ethnicity in Appalachia.

And here are Andrew Gelman's graphs of the White Death:


Easier access to cannabis as a pain killer might help out with this situation.

Are people really dropping dead of vicodin or heroin overdoses because marijuana is illegal?

Looking at birthyear data, the death rate seems to have gone up sharply among white people who turned 18 after about 1968 -- i.e., the people who were introduced to recreational drugs when young.


I don't know of the numbers, but the number of people who took opioids who would have alternatively been prescribed or otherwise taken cannabis for pain cannot be zero. And meanwhile, it's a basically risk-free drug.

It's not that. It's because those people's move into early middle age coincided with the explosion in prescribing of opioid painkilllers for every ailment. They just happened to be the demographic showing up at the doctor's office for typical aches and pains at the same time pain because the fifth vital sign and the industry started to pay more attention to the phenomenon of undertreated pain.

But I don't think you see as big of a jump in poisoning deaths in this century for people who hit age 18 before recreational drug use became a thing at the end of the sixties, even though they presumably have even more aches and pains.

I could be wrong about this -- it's hard to keep the dates straight in your head.

I'm trying to encourage more analysis of the White Death by throwing hypotheses out there for testing and making data conveniently available in different formats.

You're right about the numbers. This is also just a hypothesis, but I think it boils down to the fact that the post rec drug revolution generation never had to learn how to deal with pain without being prescribed opioids. People who are a bit older did learn other coping mechanisms, so they were less likely to become dependent on these painkillers later on. And I do think the majority of those 45+ who overdose started taking opioids due to legitimate pain, rather than with the intent of getting high.

The 60s were sex, drugs and rock & roll? Or the 60s were war and drugs? Drugs are also used for non entertainment purposes such as staying awake. Up to 2011 non-hispanic whites were overrepresented on the US war veteran population. see figure 1 https://www.census.gov/prod/2012pubs/acsbr11-22.pdf As of 2013, there are still alive 7+ million Vietnam war veterans. About 2.5+ million fought on the Vietnam front http://www.va.gov/vetdata/docs/quickfacts/Population_slideshow.pdf Wartime experiences take years out of life of veterans but the causation/mechanism is not clear, stress? http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2743276/

@Steve Sailer: white mortality may come from: a) party drugs, b) exposure to hazardous materials + free tobacco + war amphetamines & opioids + stressful experiences , c) both. The weigh on options a & b is open to discussion. But putting 100% to option a seems to ignore war veterans.

A Vietnam War connection should be investigated. I vaguely recall that smoking heroin (but not injecting it) was associated to some extent with serving in Vietnam.

There was a big spike in death rates among 45-54 year old white men in 1999-2005, but has since diminished. The death rates of middle-aged white women lagged but then rose through 2013.

Gelman breaks out sex and year, so that could be helpful in testing the Vietnam connection hypothesis. My offhand impression without checking is that the break point would have been in the middle of the Vietnam War, but that wouldn't be a dealbreaker for your theory. I've seen it said that the Vietnamization era was more demoralizing for American troops than the Westmoreland era that preceded it.

On the other hand, the effect seemed to last beyond the Vietnam War.

We could use a multiple regression model.

It would also be helpful to replot the graphs with, say, the age the subjects turned 19 on the horizontal graph.

What would Vietnam have to do with it? The demographic being considered (45-54 year olds) were too young to have been soldiers in Vietnam. In fact they missed the Hippie era entirely (except as children). This is the group that came of age during the Carter/Reagan years and the Disco era. At age 48 I am in this age group-- Vietnam is just a piece of history to me, nothing to do personally with me or anyone my age I know.

Shouldn't be too hard to analyze how marijuana legalization in some states has impacted opioid overdose deaths...

Has recent enough data come out yet? Would be interesting to see. I have read plenty of anecdotes of people who have replaced opioid painkillers with medical marijuana. This will have an impact on opioid deaths. The question is how much.

"In a quirk of the 1971 controlled substances legislation, lower amounts of hydrocodone per pill (15 milligrams or less) or those lower-potency pills combined with another pain-killing drug, such as acetaminophen, were placed in the less-restrictive Schedule III."

Opioid deaths occur mainly due to aceteminophen poisoning, which mostly occurs because combination products are on a more permissive schedule, and thus the only effective pain killer millions of people are given access to.

That is a big problem, but way more people are killed by the opioid in those combination products than the acetaminophen that destroys the liver. It's not even close.

Of course, combination hydrocodone products have now been rescheduled to schedule II--the same as oxycodone and morphine. My prediction is that it won't make any difference. In fact, by putting Vicodin and Oxycontin on the same regulatory level, the change may inadvertently drive up prescribing of more potent opioids, like methadone.

It seems difficult to get easily digestible numbers for prescription opiod versus heroin deaths pre 1999. I recall seeing a graph of heroin versus prescription as addicts' first exposure to opiates that went back at least to the early 90s and it was practically a 1:1 equivalence. As prescriptions went up, heroin went down.

Obviously that's not deaths, but it seemed to suggest that maybe the reporting about prescription overdoses is leaving out some context.

It may be just one more step in the dechristianization of the US white. Pain traditionally had a salvific role in religion. So there has been a sudden change in values.

Utah has one the top three overdose rates among states. Read into that what you will.

Mormons struggle with moderation? (see: porn and candy consumption)

High altitude: http://mic.com/articles/104096/there-s-a-suicide-epidemic-in-utah-and-one-neuroscientist-thinks-he-knows-why#.4Jvg6RvTH

@daguix: that's emotional pain like the loss of a loved one. Physical pain as a sequel from accidents or wartime events is a little bit more difficult to rationalize through religion.

Nah. Religion is pretty big on physical pain as well.


I guess I always assumed that the huge boom in 'chronic pain' and opioid use/abuse in white working-class America was purely a manifestation of their own existential crisis as a culture, after their fall from the lower middle class. I'd further posit you could correlate it with the earlier boom in crack use among the black working class after the collapse of the manufacturing sector.

A multiple regression model of the sociology of the White Death could be revealing about larger trends. Or it could just be a fairly technical issue.

I don't know, so I encourage further investigation.

One of saddest parts of this whole problem is the lack of access to effective treatment.

Medication assisted treatment (e.g. methadone, buprenorphine) is by far the most effective intervention for addiction to opioids, but it is offered to only a small share of people entering treatment. This is primarily due to the outdated 12 mentality which still dominates most treatment centers. The line is that one is not truly in recovery if they are taking any kind of opioid agonist--as if it's cheating or something. Most of these treatment centers receive federal funds, by the way. Another problem is that not enough docs are waivered to prescribe buprenorphine. Even though a physician can prescribe as much oxycodone as he wants, he needs a special waiver to treat opioid addicts, and even then may only oversee the treatment of a patient panel of up to 100 people. The final reason is that most insurance companies and Medicaid are very stingy with coverage of methadone and buprenorphine. Even though there is strong evidence that putting addicts on these meds actually saves money by keeping them out of the ER, they often limit coverage to inadequately low doses and limit treatment to 6 months or a year in duration.

Sorry, should say "12 step mentality."

From Wikipedia on Methadone: "Side effects are similar to that of other opioids. The number of drug-poisoning deaths in the United States involving methadone increased from 784 in 1999 to 5,518 in 2007 but declined to 4,418 in 2011."

Buprenorphine also causes overdose deaths, though apparently fewer: http://bmjopen.bmj.com/content/5/5/e007629.full

Can you point to some evidence that these drugs have better outcomes than a 12 step program?

This does not seem clear cut at all. And also, there are millions of these prescriptions issued every year; it is not as if they are unused. Evidence would be needed that they are appropriate for people who are not now receiving them.

"I believe economists do not devote enough attention to studying the topic of pain. - See more at: http://marginalrevolution.com/marginalrevolution/2015/11/opioids-for-the-masses.html#comments." There really is an obvious rejoinder to this. It's rather silly, but I'll bite: Because they're too busy causing pain.

I really appreciate the job market papers, they're fascinating. How about a word of advice on what to study for a prospective phd student? Not what's hot or fashionable, but what areas of economics do you (pl) see as most fertile for new study? Inquiring minds want to (get hired as economists eventually) know.


For the death statistics to actually make sense to me, I'd have to see a breakdown of how many could be classified as euthanasia.

I am a supporter of prescription monitoring programs, but having been able to skim the paper, I would be somewhat cautious about concluding PDMPs are the primary driver of the outcomes the author finds. There is a lot more going on in states that this analysis can't really account for, though it is a great effort.

PDMPs are usually introduced at the same time that opioid misuse is being addressed on multiple fronts in a state: prevention, physician education, Medicaid reforms, etc. Further, in almost all states with new PDMPs, the programs are voluntary. Doctors aren't actually required to use the databases, and as a consequence participation rates are abysmally low--sometimes like only 10% of prescribers even bother to register to use the system. Some states do require docs to participate, but that usually comes a few years after the program is started.

Ah, the market for pain.

A lot of medical research in non-opioid pain relief which may alter the cost/benefit calculus of pain.

I love how error bars are becoming more common in social science.

Another factor against pain medication--this is well known--is that doctors who are labeled as "Dr. Feelgood" (think Michael Jackson's doctor) end up disbarred, or worse, put in prison, for over-prescribing pain medication. So malpractice premiums for such doctors have soared, and consequently there's less hillbilly heroin Rx'd.

Bonus trivia: the pill-prescribing doctor who killed Elvis Presley was a Greek-American. I remember with joy reading this, and feeling proud, I don't know why, when in Greece, and the girls all crying over Elvis' death (what year was that, 1976? or 77? The same year a young Tyler Cowen won the NJ State chess championship). A perverse pleasure inside me was satisfied that it was a Greek that killed him. Anyway he would have died sooner or later since he was overweight and out of control.

Cool story, bro.

Yeah and the good docktor was railroaded (Wikipedia):

The competence and ethics of two of the centrally involved medical professionals were seriously questioned. Before the autopsy was complete and toxicology results known, medical examiner Dr. Jerry Francisco declared the cause of death as cardiac arrhythmia, a condition that can be determined only in someone who is still alive.[290] Allegations of a cover-up were widespread.[289] While Presley's main physician, Dr. Nichopoulos, was exonerated of criminal liability for the singer's death, the facts were startling: "In the first eight months of 1977 alone, he had [prescribed] more than 10,000 doses of sedatives, amphetamines and narcotics: all in Elvis's name." [THAT IS 41 DOSES A DAY, SO ELVIS MUST HAVE BEEN SELLING? -RL] His license was suspended for three months. It was permanently revoked in the 1990s after the Tennessee Medical Board brought new charges of over-prescription.[243]

Amidst mounting pressure in 1994, the Presley autopsy was reopened. Coroner Dr. Joseph Davis declared, "There is nothing in any of the data that supports a death from drugs. In fact, everything points to a sudden, violent heart attack."[243] Whether or not combined drug intoxication was in fact the cause, there is little doubt that polypharmacy contributed significantly to Presley's premature death.[290]

The author repeatedly caveats that the welfare calculation is rough and incomplete. She is right to do so. The $12 billion in benefits of reduced regulation only counts medical costs and lost wages. It makes no attempt to monetize the value of the pain reduction that opioids can provide. As a result, the paper dramatically underestimates the benefits of reduced regulation and more pain relief.

In these kinds of calculations, monetized QALY gains are usually an order of magnitude more than lost wages and medical costs. But we do not need this assumption. Given that the non-QALY benefits and costs of opioid use are the same order of magnitude, this means that the QALYs they give us through better pain relief are effectively free.

. Since 1999, there has been a fourfold increase in drug overdose deaths linked to opiod pain relievers.

With everythning else that is going on in the world, I am wondering why I should care. If someone ODs on prescription opiods that is their own damn fault (unlike ODing on a street product of unknown purity). Probably mostly suicides. Why is this any of the government's concern? Were these people potential military recruits? Are we in a war? Does it matter to anyone other than the family ?

I hope that the people close to you personally are aware of your lack of ability to empathize or care about people dying.

25,000 Americans die a year from opioid overdose? And 12,000 are murdered every year by drunk drivers?

But Americans are scared of terrorists.

But what if the terrorists are drunk?

People can be concerned about all three of those things you know.

Drunk terrorists abusing opioids?

Haha... that one got an audible laugh from me.

Less dangerous that way.

Almost no is 'murdered' by drunk drivers. People die in accidents. An irreducible share of people will. People also suffer accidents with their prescription painkillers and people commit suicide. None of this justifies open borders or importing potentially problem people because it suits the interests of said people or the random preferences of libertarian intellectuals or prog-trash intellectuals.

Yeah it's not like Syrian terrorists are currently in the Middle East killing thousands of people each year. Ohh wait...

Opioid overdose happens to Other People, not me.

I feel safe in my large metal wheeled cage because I can control it.

Terrorists kill people in cafes and there's nothing we can do to stop it.

Yeah, it's reasonable to be more fearful of terrorism than drunks/drug addicts.

People should simply have the right to buy their own medication as they see fit.

Problem solved. And yes, I am serious. Doctors should be medical advisors, not paternalistic commanders. The decision should always, without exception, lie with the client.

Look around town and you'll find plenty of people happy to let you make the choice of what pain medicine to use--and they'll even sell it to you!

Hardly. I'm not that street-wise.

But I'm sure you're technically right in the sense that I could manage to procure medications that would not legally be available to me, if I invested some criminal energy.

And yet, it is a pathetic excuse for a law if it has to be defended by the argument that we can break it.

Fake pills spiked with fatal concentrations of fentanyl? No thanks.

I don't see any problem with this.

The majority of opiod overdoses are actually opiod+alcohol/benzo/SSRI overdoses. But we all know that healthy, well-socialized people drink and drugs are bad mmkay. So if someone snorts a small bump of heroin and drinks a liter of Jack Daniels, then it's quite obviously a heroin overdose.

Has anyone of the previous posters actually suffered from long-term chronic pain? For a three year stretch, I had a form of arthritis which at times affected my job (white collar financial type) performance to the extent that I was sure I'd get fired if my condition went on another day. While not the only medication I was taking, a daily regimen of hydrocodone dulled the pain enough so that I could get through the day. And even then it was a tremendous struggle. I'm now in remission and take no medication for pain save the occasional ibuprofen. The point being, opioids have an important place in the treatment of chronic pain. Yes, they have many drawbacks and side effects to the individual and to society but until something better comes along (and it will eventually), have a bit of compassion on those afflicted. I know this is a econ blog, but I see none here. Just a bunch of people theorizing from the outside looking in. Just my view from the inside looking out.

Speaking as one who has tried to engage those with exmtree views about opioid regulation on many disappointing occasions, I would urge oura0community of professionals concerned about patient care to avoid pursuing any effort to negotiate on the details of a request to FDA.a0 We should not buy in to the proposition that opioids need relabeling by the FDA.a0 The public health problem of prescription drug abuse should be addressed by clinician education, smart regulation that does not pose high risk of worsening the existing problem of poorly treated and uncontrolled chronic pain (as the REMS are trying to do), better use of current law enforcement approaches to eliminate pill mills, and more research into the safety and effectiveness of opioid drugs.a0 The PROP petition is inappropriate because 1) FDA should follow the science only and there is no science to support the demands, 2) FDA should not change its standard approach to labeling for one class unless it plans to apply a new stardard to all other drug classes without evidence of long-term safety and efficacy, e.g., antidepressants, neuroleptics, NSAIDs, and many others, 3) the likelihood that the change will adversely affect patients is very high because non-experts will look to the label as reflecting a standard of care and withhold therapya0 from patients who could benefit, and some payers will push the costs of treatment onto patients by claiming that off-label use does not deserve coverage.a0In my view, the proper response of FDA to the PROP petition is simply to reject ita0as unsupported by the science and practice of medicine.a0 Thankfully, this is what the counter-petitions and lettersa0to the docketa0are saying. a0a0The PROP petition is a tactic and it is wonderful that our community is responding to FDA with the analysis it needs to push back.a0 Unfortunately, it may have trouble doing so.a0 To prevent the next step, which would be legislation to force FDA to make label changes, or to do worse, I would suggest that our community make clear several points:a0 First, the pronouncements and proposals of PROP repeatedly demonstrate very little understanding of pain medicine.a0 Most egregious is thea0lumping together of a very diverse patient population with chronic pain into a group called chronic non-cancer pain .a0 Pain specialists know that subgroups, like lowa0back pain patients with substance use disorder, advanced multiple sclerosis patients, and elderly patients with disabling joint pain, differ greatly, and we know that there are mediators of individual risk (such as history of substance abuse) that vary within each pain population.a0 We know that the term cancer pain is vague when referring to the millions of patients who are long-term survivors.a0 The approach taken by PROP does not worry itself with these nuances.a0 Second, the pronouncements and proposals of PROP demonstrate no critical evalution of the risks associated with chronic opioid therapy.a0 There are many egregious examples that could be cited, but perhaps the most is the statement that there is a dose, e.g., 100a0morphine equivalent mg, which when traversed, imparts unacceptable risk to the individual.a0 This conclusion suggests that we know something about opioid molecules that we do not.a0 It originates from population-based claims data,a0which in the realm of science should be viewed as hypothesis-generating, not evidence.a0 PROP does not feel the need to note that these findings should never be considered dispositive.a0 Third,a0PROP proposals are unbalanced, in that they never address the potential unintented consequences of increased pain. a0PROP ignores history that even includes a 2001 position statement endorsed by DEA and 21 professional organizations (www.deadiversion.usdoj.gov/pubs/advisories/newsrel_102301.pdf).a0PROP has filled a vacuum and I am happy to see pain specialists and palliative medicine stepping up to ensure that science is respected and patients' voices are heard.a0

Comments for this post are closed