Fragments of note

…OxyContin abuse kills three times more people than gun homicides yearly.

That is from Scott Alexander, USA only, and here is Scott’s earlier post on guns, follow-up here.

Addendum: Do note the comment from GregS, this comparison may not be correct.  Here is an update.


Is this a meaningful comparison? I remember during the Vietnam war the number of deaths in the war was being compared to deaths from automobile deaths. Does one make the other somehow right or insignificant? If OxyContin kills more people than ISIS or the Taliban, does that mean anything?

It's easier to end a war than to end driving. Most of the "gun deaths" are suicides. It's a bait and switch by gun controllers, the data shows at worst a neutral correlation between gun ownership and homicides. Gun control advocates move the goal posts and say they want to stop gun suicides, but drug legalization replaces guns as an available suicide method. I would bet more people will commit suicide if death by happy pills is widely available.

The lethality rate of attempted gun suicides is much, much, higher than for attempted prescription (or illegal) drug overdoses (85% vs. 2%). While suicide numbers may increase under drug legalization, it suggests firearm law liberalization could be much worse.

If we ban abortion clinics pregnant women will find alternative less safe ways to have an abortion.

If we ban guns suicidal people will NOT find alternative methods to commit suicide?

Hypocrisy smells I.

"This data doesn't conform to my priors. I shall therefore ignore it."

Often times, people learn from the first suicide attempt and don't attempt again. Those using guns don't benefit from the "mistakes" of not killing themselves the first time.

It's about _attempting_ to commit suicide.

Those who use guns are more committed to dying, while the overdoses are a cry for help.

To support Jay's statement, why are suicide rates similar to countries who have strong gun control?

Assuming there's not a miscarriage, pregnancy is guaranteed to produce a child that at minimum will be taken away from you in adoption and more likely be someone you have to take care of, possibly without a father to help, for 18 years.

Suicidal depression is a transient condition (source: me - I've suffered from it a few times) and there is a ton of research showing people who are prevented or talked out of killing themselves are usually out of that condition within a few months.

Among other things, the research implies that the city of San Francisco is terribly negligent by making it easy to jump off the Golden Gate Bridge.

Gun suicides are messy and an inconvenience for everyone. It takes a real egomaniac to blow their own brains out. Suicide by drugs, on the other hand, isn't much different than dying of a heart attack. People should be schooled from infancy that if they feel the need to voluntarily depart, it should be done in as unspectacular a way as possible. The neighbor girl hung herself in the basement. That sort of thing seems to be going out of fashion.

And while you're at it, could we have some PSAs suggesting to those considering murder-suicide that they could solve their problem just as well if they just did themselves first?

It's easier to (legally) get your hands on a gun than drugs that kill you painlessly.

The appropriate comparisons in my view would be to accidental gun deaths (OxyContin is far, far worse), and to gun suicides (as I think it is fair to assume that some not insignificant share of OxyContin deaths that are formally classified as accidental actually involve either suicidal ideation or indifference to living or dying in pursuit of the relief that it provides). The number of OxyContin homicides is negligible.

Indeed, even OxyContin accidents, if we could switch them one for one for most other accidents, are preferable, as there is much more overlap between the person at fault and the victim than for example, in drunk driving accidents.

Legalize hard drugs and this number will multiply to an extent only few people imagine.

I always find it so funny that liberals are pro legalizing drugs but against liberal weapon laws while conservatives are pro liberal weapon laws but basically against free drugs.

Neither side is consistent.

It's consistent with their philosophies.

What? Almost nobody wants to legalize hard drugs.

While Republicans tended to favor legalization the least, support among Democrats and independents didn't reach 20 percent for any drug and was lower for most.

Well then switch "hard drugs" with "drugs" or "marihuana" or whatever. You know what I meant.

I guess I don't get your point. You started off by telling us about all the people who would die if "hard drugs" were legalized.

To be clear, liberals aren't pro "legalizing drugs." They are, by a fairly small margin, pro legalizing one drug--marijuana--which even a third of Republicans now support.

I doubt the legalisation of marijuana would lead to a serious exacerbation of this problem. In fact, pot legalisation possibly *reduces* violent/property crime

if you want to reduce drug deaths, maintain 21+ drinking age and increase alcohol taxes.

You literally cannot overdose on marijuana. It's a different brain interaction than alcohol.

Leftist stances aren't popular, news at 5....

Have you never heard them complaining about the "war on drugs?" There is a distinct minority on the Left that favors legalizing everything.

A very small minority, and a much larger minority (probably a majority) of libertarians favor that FWIW.

This is one of the issues where the far left and the libertarian right overlap.

No. "The Left" does not want to legalize drugs any more than "The Right" wants to mandate teachers carry guns.

Didn't the opposite happen in Portugal?

Drugs include pot, lsd, speed in hundreds of versions including Ritalin, alcohol, opiates in hundreds of versions including synthetic analogs like methadone,.

"Arms" the word used in the second amendment includes knives, swords, canons, rockets, bombs, landmines, chemical weapons, armed drones, surface to arm shoulder rockets, body armor, and nuclear weapons. Oh, yeah guns, which are mini canons, today with autoloader and rapid fire.

Why can't any American carry a backpack nuclear bomb?

The second amendment says nothing about guns other than in the sense "drugs" includes alcohol, aspirin, etc.

These is a plausible theory, but it is empirically wrong. The experience in Portugal where this was done completely, and in Switzerland and the Netherlands, where it has been done in part, has been that there is harm reduction and not increased harm as a result of treating it as a public health problem instead of a crime.

How did that work out when all those drugs were legal in the U.S.?

The CDC is authorized to do infinite times as much research on drug overdose deaths as gun deaths.

Every police department in America does research on gun deaths...

Is that why we have such great national stats about shootings by police, presumably the gun deaths about which they should have the best data?

Law-enforcement experts long have lamented the lack of information about killings by police. “When cops are killed, there is a very careful account and there’s a national database,” said Jeffrey Fagan, a law professor at Columbia University. “Why not the other side of the ledger?”

except for the ones caused by their guns.

Yes, the "Center for Disease Control" doesn't do research on gun deaths. But gun deaths aren't a disease by any reasonable definition. NASA's probably does an infinite times more research on space related deaths than it does on AIDs related deaths.

Did you the CDC does research on all kinds of other injures? They research and help us better prevent traumatic brain injury, child abuse and other domestic violence, motor vehicle crashes, falls, fires, etc. But not gun violence. That is a bridge too far. Tell me exactly which federal agency is now doing research on gun violence--NASA, HHS, Treasury?

Even the member of Congress who led the charge on banning federal gun violence research now says he regrets it: "If we had somehow gotten the research going, we could have somehow found a solution to the gun violence without there being any restrictions on the Second Amendment," Dickey said. "We could have used that all these years to develop the equivalent of that little small fence."

Do you think maybe we've gone too far?

"Do you think maybe we’ve gone too far?"

I don't really have an opinion one way or another. I don't feel the need to get outraged about every little thing or an impulse to "do something".

The impulsive outrage to ban research seems to be culprit here.

Refusing to pay for something is not the same as banning it.

In this case, when the CDC is not banned from researching anything else, yeah, it is.

Passing a law banning research into some area because you might not like the results seems like a pretty obviously bad idea to me.

That's not what happened though. The CDC started political lobbying (not doing research) in support of gun control using public funds and Congress shut that down.

Has Mr. Dickey ever heard of a thing called "criminology?"

Please stop linking SSC. Your readers are toxic.

Sounds like you need to crawl back into your safe space

This would be a meaningful comparison if one were comparing accidental suicides (or maybe accidental homicides) and ODs.

"I always find it so funny that liberals are pro legalizing drugs but against liberal weapon laws while conservatives are pro liberal weapon laws but basically against free drugs.

Neither side is consistent"

The difference is that you can't kill someone else simply by consuming a drug. Pointing and shooting a gun on the other hand.

But you can, of course, surrender your cognitive ability to choose not to kill someone simply by consuming a drug.

"The difference is that you can’t kill someone else simply by consuming a drug."

There's literally tens of thousands of alcohol related driving deaths.

indeed, in addition to all the alcohol-fuelled violence that occurs daily. yet we don't outlaw alcohol sales anymore. I think the point is that there are always going to be negative side effects of e.g. alcohol being legal, of teenagers/senior citizens being allowed to drive, etc etc. But society has decided that we shouldn't criminalise these activities. And now that argument has been extended to certain drugs

Alcohol related driving deaths. The real relationship isn't with alcohol, it's with falling asleep at the wheel. Some wage slave gets up at 4:40 am, drives to work, maybe puts in some overtime, stops at the saloon on the way home and after a couple of games of pool, all of a sudden it's midnight and time to hit the highway. He's been awake for 16 hours, worked hard and drank some beer. If he was already home he'd be asleep on the couch, instead he's asleep at the wheel.

But this an alcohol-related problem, and a crime, when he has a head-on with an SUV or runs off the road into a tree. At the same time, somebody that's gone through the same process, minus the drinking, and passes out while driving, then runs over a pedestrian, isn't guilty of anything but a body function. If my kid gets killed by a sober sleeping driver am I supposed to be satisfied because they didn't have any alcohol in their system?

"Alcohol related driving deaths. The real relationship isn’t with alcohol, it’s with falling asleep at the wheel. "

Plenty of alcohol related accidents have nothing to do with being tired and everything to do with drinking large amounts of alcohol.

Falling asleep while driving is an all too common occurrence. According to the National Highway Traffic Safety Administration (NHTSA), each year, more than 100,000 police-reported car accidents result from drivers who fall asleep. Of those, more than 1,600 cause fatalities, and more than 72,000 cause serious injuries. Because not everyone reports all sleep-caused accidents, the actual number is probably much higher.

According to a Farmers Insurance Company survey, "almost three times as many men (15.9 percent) as women (5.8 percent) said they had fallen asleep at the wheel. Those ages 55 to 64 had the highest percentage of any age group surveyed (13.7 percent)."

Stress, overwork, personal problems, long drives, alcohol, and medications are some reasons why drivers fall asleep. While many of those who fall asleep are lucky to escape accidents, there are still thousands who aren't as fortunate. It's usually just a matter of time before a driver who falls asleep at the wheel ends up causing an accident.
Unfortunately, when it comes to falling asleep while driving, proving the driver was negligent is much more difficult. Unless a driver admits he fell asleep at the wheel and crashed into you, you're going to have a tough time gathering the evidence you need to prove it.

No it doesn’t. I believe Scott is confused, not that he’s alone on this one. I think he is comparing gun homicide deaths to *all*drug poisonings, or all prescription painkillers, or something like that. The figures I’m finding are ~10,000 gun homicides, ~15,000 prescription painkiller deaths, and ~39,000 unintentional drug poisonings (overlapping the prescription painkiller deaths, obviously). None of those gives me a 3-to-1 ratio, so I’m not sure what he’s doing.

Very few of the prescription drug poisonings are overdoses of a single drug. When thorough studies of “OxyContin deaths” have been conducted (usually after a few scare-stories citing flubbed statistics tabulated with poor methodologies), they find that almost none of the deaths are attributable to OxyContin alone. The same is true of other prescription painkillers. It’s almost always a combination of these drugs and alcohol, benzodiazepines, anti-depressants, and other drugs. If you look at the SAMSHA and Monitoring the Future drug abuse surveys, abuse of prescription painkillers is roughly flat over the last 15 years, where we’ve seen a dramatic increase in drug poisonings, and in fact rates are *down* over the past 4 or 5 years, when prescriptions of painkillers have leveled off. Abuse rates for other pharmaceuticals is also flat. This narrative where “Doctors hand out opioids like candy, and we just can’t help ourselves!” just isn’t true. These are mostly drug interactions caused by unwary doctors and patients; the solution is better education, not prohibition. Surely there are *some* recreational users/abusers who are taking a bad mix of drugs, but they probably aren’t responsible for the recent increase. You should see that in the survey data (SAMSHA and MTF). “Proper” prescription opioid use is increasing, it’s just bumping into other drugs that people are already taking. Also, sometimes the medical examiner will just write down whatever was on the toxicology report without really investigating what the true cause of death was. It’s possible that more people are simply dying with something in their system (consistent with the increasing numbers of prescriptions), but that most of these deaths aren’t actually caused by the painkillers themselves.

Thanks for the heads-up, because I overlooked this post by Scott. I’ve been looking at CDC mortality data and drug use surveys to try to confirm/discredit the standard narrative explaining the recent overdose surge. Still not quite sure what’s going on, but the standard story is problematic.

Is it a bit Coasian to ask whether the opioids or the benzodiazepines should be blamed for the drug poisonings?

+1 Great Comment.

It's amazing to see how many smart economists are being easily hoodwinked into 2016's version of Reefer Madness.

Thanks! I’m also a bit frustrated by how credulous even smart, numerate people are of these mortality figures. For one thing, nobody knows how many of these are actually drug poisoning deaths and how many are simply people who drop dead (say, from a heart palpitation or something else that’s hard to spot) when they happen to have elevated levels of opioids (or whatever) in their system. Steven B Karch, who literally wrote the book on this topic (Pathology of Drug Abuse), suggests that it’s actually really hard to tell if the cause of death was an overdose. You can’t go by toxicology alone, but that appears to be what the coroners and medical examiners are doing. Radley Balko did some great reporting on this a few years ago:

‘"I don't know where they got their numbers," Karch says of the CDC estimates. "There's no peer review of those figures. You follow the footnotes, and it looks like they're getting the information from medical examiners. But it doesn't say how the medical examiners are concluding that these were overdoses--if, say, they're just relying on toxicology results." Asked if that's usually how overdoses are diagnosed Karch says, "That fits my experience."’

Another problem, even taking the CDC’s numbers as a given, is that their cause of death codes are too coarse to make some of the claims that they make. There’s a multiple cause of death code for the categories “Other Opioids”, “Other synthetic Narcotics”, “Methadone”, and “Other and unspecified narcotics.” There isn’t one for OxcyContin, nor do these categories cleanly split into “prescription” and “illicit”. “Other synthetic Narcotics” actually includes a lot of fentanyl poisonings, which is one of the reasons heroin overdose deaths have been rising (and spiked *hugely* in the past year). Heroin dealers have been spiking their stuff with fentanyl. If you follow the footnotes in the CDC's reports, it looks like their just adding up these homogeneous categories. There certainly isn’t a cause of death code for each drug.

Also to further support your comment: the demographics with the largest increase in opioid deaths (middle aged whites, particularly women) neatly line up with the groups with the largest increase in benzo use. Much more so than opiate use itself. This is fundamentally a sleeping pill, not pain pill, problem.

Yeah. The number of ALL opioid painkiller deaths was <20,000 last year. Oxycodone is a subset of that figure. Assuming you're correct about the number of gun homicides, this guy seems to way off.

Two caveats on the rest of your post:
1) It is correct that the number of Rx overdose deaths (not sure about rates) had in fact leveled off the past few years (actually declining one year), but then they reversed course and increased significantly in 2014. At the same time, heroin ODs have been quickly marching upward the past few years.

2) Less than 1/3 of prescription drug overdoses involve a benzodiazepine. So it is not the case that all or even most of the overdoses are a combination of opioids and benzos. Though I agree that this combination is common, even frequent, in overdoses, it just isn't the bulk of them. This also implies that most overdoses aren't simply soccer moms trying to take their drugs as prescribed. It is a problem of abuse and misuse--even if the people abusing the drugs started off with a legitimate medical need.

Thanks for your comment. You are correct that not all of these are benzodiazepine interactions, but many of them are. It may well be more than the ~30% figure given in the CDC link; I don't think they actually know. Anyway, a lot of other drugs interact with painkillers. I'm currently looking at how every category of drug that the CDC has a "multiple cause of death" code for. Almost all of them are going up by factors of 4 to 10. ("Antiallergic and antiemetic drugs" up by a factor of 4.5, "other antiepileptic and sedative-hypnotic drugs" up by a factor of 10, the dreaded "antitussives" up by a factor of 8, admittedly some from a meager start and to a still-meager end.) Maybe we're having an abuse epidemic in every single category of drug. Or maybe the rise of pain killers has caused an epidemic of interactions in numerous other drug categories. Or maybe we're just doing more thorough tox screenings and are more willing to stop with that piece of information in a cause-of-death investigation than we were 15 years ago. Like I said, I'm not quite sure what's going on.

Abuse of these substances is surely a problem, but it's no more a problem than it was 15 years ago according to the SAMSHA and MTF drug surveys. I guess it's possible that the surveys just aren't picking up a trend of increasing illicit use of these substances, but that's a hard story to tell. I definitely recommend downloading some of the mortality data from the CDC's Wonder database and taking a peek at the drug use surveys. Hopefully someone smarter than me can figure out what's actually going on.

What is a person to make of the fact that nearly 1 in 4 pregnant women enrolled in Mediciad (45% of pregnancies nationwide) had a prescription filled for an opioid? That sounds suspiciously like doctors overprescribing opioids to me.

What is a person to make of the fact that hydrocodone has been either the top or the second most prescribed drug in all the land for years? It is difficult to square the number of prescriptions well over 100 million filled per year for hydrocodone alone with the argument that it is not extremely widely prescribed.

I wrote the post that Kevin Drum was quoting and that Alexander then drew from. My comparison was that total drug overdoses are three times total homicides. (Not just Oxy and not just gun homicides.)

Here's the post:

And here's an earlier post with the methodology and spreadsheets I pulled from CDC WONDER. Drug ODs are tricky; I used the collection of codes the CDC itself used in a report about drug overdoses.

Noted. I misunderstood the linked article. I've corrected my links. Sorry for the mistake.

Compare overdoses to suicides, not to homicides... obviously.

Good thing we have those doctors watching over us at $200 per 3 minute visit. Watching out for our safety. Preventing abuse.

Stand up to organized crime. Legalize drugs, treat addicts.

It might be cheaper to legalise drugs and give addicts a kick up the arse, and advice to get a grip. And possibly no less effective.

"Stand up to organized crime. Legalize drugs, treat addicts. "

Well you can attempt to stand up to organized crime, but they have the guns, lawyers and personnel on their side. And that's just the Federal government.

This is a tough crowd. You economists are drug death deniers. (Okay, I'm teasing you there.)

Gun homicide deaths=11,208
Motor vehicle deaths=35,369
Drug overdose deaths=46,471 (4x gun homicides)
U.S. 2013

The point is did you f-ing know-it-alls know drug overdose deaths were so high? My bet is you had no f-ing idea. Heartless bastards! (Semi kidding.)

The point is, if you're going to prioritize problems, the number of deaths is a good starting point.

My son, now clean, was a heroin addict. He once remarked off-handedly that the hard part was thinking of what to say to the parents at the funerals. OMG, he's been to so many funerals that's become a problem for him!?

"The point is, if you’re going to prioritize problems, the number of deaths is a good starting point." Can't really dispute this.

Comments for this post are closed