Tuesday, September 27, 2016

Who is Reading?

Just curious, because I get almost zero feedback. I know very little about who my readers are. If I just e-mailed someone a link or posted to Facebook I’ll sometimes see a spike in page views. In those cases it’s fairly obvious where my readers are coming from. But other than that I have no idea. Drop me a line or a comment. Or not. I’d love to hear from you! Or not! Lurkers welcome! I appreciate the views, input or no input. But I am curious, and I’m particularly curious about regular readers who I don’t personally know.

Monday, September 26, 2016

Age Distribution of Legitimate and Non-Medical Opioid Use

SAMHSA (the Substance Abuse and Mental Health Services Administration) very recently released the 2015 data for its annual drug abuse survey. (The National Survey on Drug Use and Health, NSDUH.) Detailed tables here. This year’s survey, for the first time, includes questions that distinguish legitimate use of prescription drugs from misuse. Previous incarnations only asked about “illicit use.”  The survey asks the user about “Any Use in the Past Year”, “Misuse in the Past Year”, and “Misuse in the past month.” (See table 1.23A in the link above.) This is the first time I’ve ever seen an age distribution for legitimate prescription opioid users, and for all I know it's the first that's ever been published.

There are competing theories about why opioid overdoses are rising. Are the extra deaths simply coming from normal users? Or are they coming from recreational users, who may not have a legitimate medical need for them? Or are they coming from illicit users who *once* had a legal prescription but were cut off by their pain doctor? (I explore several competing explanations here.)  My hope is that I can somehow match up the age distributions of the “users” and the “deaths.” We know the age distribution of prescription opioid deaths, because we know the age of every decedent for whom opioid poisoning was mentioned on the death certificate. The CDC tracks this data, and it’s publicly available. I was curious if this distribution better overlapped the age distribution of illicit users or the age distribution of legitimate users. (Click on the chart to get the best view of it.)

The “Legitimate Use” age distribution agrees best with the “age of mortality” distribution. Legitimate users skew older, and so do the deaths. (I’m assuming here that I can subtract “misuse in past year” from “Any use in past year” to get legitimate use in past year. The “Any Use” age distribution is on there for comparison. Most use is legitimate use; there are ~97 million “any” users and 12 million “misuse” users in the past year.) On the other hand, the “misuse” distribution skews young.  At first glance, this supports the theory that the deaths are coming mostly from legitimate users. But it’s not at all conclusive.

As I mentioned in this post age is a huge risk-factor in drug-related poisonings. The disjoint between the two distributions is pretty extreme for cocaine (click the link above for details). So I could still believe that most of the deaths are coming from illicit users of legal painkillers; the age distributions don’t help answer that question. The users can skew young while the deaths skew old because older people die from drug poisonings at a much higher rate. (I’m very glad I did this previous post, otherwise I might have been misled into thinking that this was some kind of smoking gun in support of the “there are simply more legitimate users who are dying at an unchanged mortality rate” theory.)

The reason I doubt the “illicit use is driving the excess deaths” story is that prescription painkiller abuse isn’t increasing. Total illicit use and even “substance abuse disorder” (people with especially severe drug habits) are basically flat. See page 7 figure 6 and page 26 figure 36 in this document. These certainly aren’t rising enough to explain the roughly three-fold increase in opioid poisoning deaths that’s happened in the past 15 years. I don’t totally discount this explanation, but on its face it’s contradicted by the survey data.

 There are important policy implications here; I'm not just engaging in meaningless number-crunching. If the true explanation for the increase in opioid poisonings is that there are just more normal users, who are overdosing at the same rates as 15 years ago, then I don’t think there’s any cause for alarm. A treatment with a known and acceptable risk has simply been expanded into a bigger population; it shouldn’t worry us that the raw numbers increase. If the explanation is that people are getting cut off from their legal prescriptions and turning to illegal dealers and dangerous routes of administration (injecting and snorting rather than pill swallowing, or resorting to heroin), then the solution is to back off enforcement and ensure that these drugs are readily available. Don't cut these people off of their legal supply and they won't engage in the risky behaviors that are leading to so many deaths. If the explanation is that people are getting addicted to their prescription opioids and their addictions are escalating into full-blown drug abuse, then only in this third case might it make sense to crack down on prescriptions. (And it’s by no means clear that a crack-down is in order; many other criteria would need to be met first.) Hopefully coroners and medical examiners start collecting relevant information for answering this question. Did the decedent have a legal prescription at the time of death? Did they recently have a prescription? Did they have a prescription at any time, and when? Knowing this, at least for a representative sample from a few hospitals, might answer the question once and for all. I’m trying to answer the question with disparate data sets that weren’t compiled for the purpose. This approach offers suggestions, but no clear answers. 

Trump's Popularity

Recently I was having donuts for breakfast at a local donut shop. There was a table populated by a bunch of old guys bullshitting with each other, eating donuts and drinking coffee. I thought to myself, “This is Trump!” None of them expressed a coherent worldview. They weren’t meticulously cost-benefit analyzing their political proposals. No, they were speaking from the “id” and trying to one-up each other on who had the cleverest political comment. If you had a transcript of all such conversations from this one group, you’d find the same people voicing support for opposite policies on different days (like Trump). Suppose you put these men in charge of government. Just sit them at a table, give them coffee and donuts, and turn each of their verbal ejaculations into a law. You would end up with an incoherent, self-contradicting hodgepodge. (Err…more so.) The bureaucrats who administer the law would be paralyzed by their inability to comply with contradicting orders. Watching these old guys talk gave me some insight, I think, into Trump’s popularity. And imagining them directing government gave me some insight into what a Trump presidency would look like. (Or any populist presidency, for that matter.)

Sunday, September 25, 2016

Mostly Harmless Dissociative Anesthetics – More Evidence From a Medical Text

There is an entire chapter in Karch’s Pathology of Drug Abuse (fifth edition) covering “dissociative anesthetics.” These are drugs that put people in a detached state of mind. Some are used as surgical anesthetics. It turns out this is a very safe class of drugs. From the first page of the chapter:

 In 1999, only 93 PCP-related deaths were reported in the Drug Abuse Warning Network (DAWN) survey (SAMHSA, 2000), and even fewer were attributed to ketamine. The Substance Abuse and Mental Health Services Administration (SAMHSA) report from December 2013 cites information from the Drug Enforcement Administration (DEA) indicating that, save for one or two “hot spots” (New York and Chicago), use has declined to negligible levels (SAMHSA, 2013). Neither Salvia nor GHB has ever been mentioned in any federal report dealing with death or morbidity. Clearly, deaths from GHB overdose do occur, but they must be relatively uncommon to rate so little mention. Deaths from PCP derivatives may be more common than deaths from PCP itself.

This can be contrasted with cocaine and heroin, which both cause a few thousand deaths every year. Of course a rate per user might be more meaningful than a raw number, but clearly there aren’t a lot of deaths. PCP has a really terrible reputation as a monster drug, but 93 deaths in a country of 300 million people (admittedly very few being users) is not at all scary.

The emergency room (ER) component of the 2009 DAWN report places the number of PCP-related ER visits at 12/100,000 persons, compared to 0.6/100,000 for GHB and 0.2/100,000 for ketamine (SAMHSA, 2010). No deaths from Salvia or propofol were listed in any of the official drug surveys, although, as mentioned, cases of propofol-related deaths have been reported (Kirby et al., 2009; Han et al., 2013).

The book is full of death statistics, along with stats on ER visits. Often there aren’t enough users of the drug or enough deaths to warrant any statistics-gathering, so the author has to rely on case studies of isolated deaths. Any drug whose mortality rate is quantified by a few case studies rather than population statistics is probably pretty safe.

Phencyclidine (PCP)

I don’t want to understate the risks. Clearly PCP can cause psychotic states, and the book discusses this possibility in some detail. But then again, potential users *know* about this risk, and most people avoid the drug for this reason. By all accounts, uncontrolled psychotic episodes are relatively rare.

Episodes of fatal PCP intoxication from direct toxicity (as opposed to homicides and trauma deaths where PCP is an incidental finding) are uncommon (Noguchi and Nakamura, 1978; Budd and Liu, 1982; Poklis et al., 1990; Li and Smialek, 1996) and, in any event, blood levels in patients dying directly from the effects of PCP overlap with the blood levels seen in victims of accidental deaths (Poklis et al., 1990).

Most cases of sudden death in police custody involve abusers with many years’ history of polydrug abuse who have already undergone left ventricular remodeling. At least two reported deaths thought to involve neck hold restraint actually involved intoxicated PCP abusers (Barton et al., 1981; Mercy et al., 1990). Neck hold restraint produces carotid sinus stimulation and reflex bradycardia.

In other words, even those few deaths attributed to PCP use might be due to other causes, like a history of drug use.

What is known now, but was not known then, is that ~35% of sudden unexplained death cases and ~20% of sudden infant death syndrome cases may be explained by mutations in cardiac ion channels (cardiac channelopathies) (Mizzanti et al., 2014; Wang et al., 2014). Without DNA resequencing, it would be very difficult to say, absent any obvious signs of toxicity (such as rhabdomyolysis), that PCP and not some genetic aberration is the case of death.

This is a common theme in the book: attributing a cause of death is difficult, because people drop dead from undetectable conditions all the time.


The 1999 [DAWN] report contains 21 ketamine mentions, with an insignificant increase in the number of deaths from 3 to 16 deaths reported a year earlier, ranking ketamine in the 71st position on the DAWN drug-related deaths lists in 2000 and accounting for 0.18% of all drug-related deaths reported to the federal government that year (Kissin et al., 2000). According to the DAWN’s 2005 report, there were only 275 ketamine-related ER visits in the preceding year, and none was fatal (SAMHSA, 2006).

“DAWN” = Drug Abuse Warning Network, a government survey based on hospital surveys in select hospitals in representative cities/communities. The point here is that ketamine is a very small driver of drug-related mortality.

Ketamine is classified as a “dangerous drug,” but it has quite an extraordinary safety profile. A 1996 study on the use of ketamine as an anesthetic in the developing world surveyed 122 physicians about their experiences in operating on 12,000 patients. … One unexplained pediatric death occurred during unmonitored recovery, and one adult suffered cardiac arrest after a failed intubation attempt. Apnea, possibly related to ketamine, was reported in 10 patients and laryngospasm in six. The Red Cross in its field hospitals (Lenz and Stehle, 1984) reported similar experiences. Even in the case of a substantial overdose, the main effect seems to be prolonged sedation. Green et al. described nine cases of inadvertent ketamine overdose in children. Three of the children received five times the recommended dose, five received 10 times the ordered dose, and one child was given a dose 100 times greater than ordered (all by intravenous or intramuscular route). All nine experienced prolonged sedation (3–24 h). Except for prolonged sedation, no adverse outcomes were noted (Green et al., 1999).

Take that in for a second. Read it again a few times if you have to. The authors are saying that even a massive overdose (given to children) isn’t all that dangerous. An “overdose” (unless it’s truly massive) just means you spend more time napping. I find this amusing, because so much of the anti-drug rhetoric begins and ends with “Think of the children!” Here is a pretty damn safe drug. Children who are accidentally given a massive overdose turn out okay. Granted some of these were probably in a hospital setting with frequent monitoring, but clearly it's quite non-toxic:

Compared to other anesthetic agents, ketamine appears to possess little intrinsic toxicity. In a published review of 87 ketamine-positive deaths that occurred over a 2-year period, almost all of the positive test results were found in hospitalized patients following surgical procedures or burn treatment patients, and not a single case of death could be attributed to intoxication with ketamine (Gill and Stajic, 2000).
 Ketamine rarely causes death by itself, and if it does, peripheral blood concentrations are likely to be well over 5 mg/L in the case of acute intoxication. Drug interactions and presence of significant natural disease are more likely to be present in a forensic case. As always, the interpretation of the significance of the results cannot be made on tissue concentrations alone.


I recommend reading Illlegal Drugs by Paul Gahlinger on GHB. There is a chapter dedicated to this drug, and it goes much more into the history and sociology of the drug. Still, Karch’s Pathology of Drug Abuse offers some useful details.

GHB-related ER visits increased significantly during the 1990s, but the number of deaths attributable to GHB today has declined, at least within the United States, to a level that is difficult to estimate.

So whatever statistics exist, take them with a grain of salt.

In the past, nearly two-thirds of GHB-related ER visits were due to overdose and one-third to unexpected reactions. Sixty percent of ER visits were said to have involved the use of multiple other drugs, usually GHB in combination with ethanol (76%), cocaine (6%), marijuana (5%), and MDMA (4%) (Woodworth and DEA, 1999).

This is a common theme in drug-related morality. Poly-drug use is the driving factor. Most of these deaths involve multiple substances interacting with each other. A sensible person using a single drug at an appropriate dosage has nothing to worry about.

The body metabolizes GHB at an extraordinary rate:
 The half-life of GHB in humans is known to be approximately 20 min...

The book Illegal Drugs mentions cases of ER patients in a deep coma who wake up an hour later and walk out of the emergency room with no issues. Clearly overdoses happen, but given the high rate of metabolism, even an overdose is unlikely to be fatal. This seems to be a point of contention. The book Buzzed suggests that the fatal dose is within a factor of two of the effective dose. I suspect these “doses” are hard to quantify given that there are so few overdose deaths to extrapolate from. But beware that there is some controversy over the safety of GHB.

GHB users often coingest ethanol. That practice may be dangerous because, in animal studies, high concentrations of either drug affect the metabolism of the other. When large doses of GHB are given, ethanol elimination is reduced (Hoes et al., 1981), an effect that would almost certainly lead to higher ethanol concentrations and probably to increased GHB toxicity. The picture becomes even more complicated when other drugs are used in addition to alcohol, as is often the case (Liechti et al., 2006).

So don’t take GHB with alcohol.

GHB is commonly abused in combination with other drugs acting on the CNS [(Central Nervous System)] (Caldicott et al., 2004). For example, in Sweden, 21 of 23 deaths attributed to GHB were found to involve other substances as well, particularly alcohol and opioids (Knudsen et al., 2010).
Don’t take it with opioids, either. But if you avoid other drugs, and you know what dose you’re taking, you’re probably safe. And just to reiterate, the section on GHB ends with this comment:

Clearly, as with any forensic matter, the context of the case is paramount. Witnesses to ingestion of GHB and other substances will be most helpful. High postmortem blood concentrations alone do not necessarily prove that the drug was taken or that it caused the death. Indeed, death from GHB is relatively uncommon, and when it does occur, it is usually in conjunction with the use of ethanol and/or other CNS depressants.

Salvia Divinorum

This is a plant that’s a member of the mint family. When smoked it causes intense hallucinations and loopy behavior. (Look at some Youtube videos  of people smoking it to see what I mean.) Most users don’t enjoy it and promptly stop using, often ending with the first use.

I’ll start with the punchline, the very last sentence of the section on salvia:

No deaths from Salvia abuse have ever been reported.

This is in spite of millions of users:

 According to the “National Survey on Drug Use and Health Report,” published by SAMHSA in February 2008, an estimated 1.8 million U.S. citizens aged 12 or older reported having used S. divinorum in their lifetime. Nearly one-third of those had done so in the past year.

And this:

An epidemiologic study of 42,179 Canadian adolescents, aged 12–17 years, during the years from 2008 to 2009, found that overall 3.8% of adolescents reported using Salvia in the past year and 6.2% had used the substance in their lifetime. Surprisingly, in this study, the prevalence of 12-month Salvia use was higher than use of cocaine and amphetamine, but lower than use of ecstasy, cannabis, and other hallucinogens.

I think it’s probably a good thing that kids are using savlia *instead of* cocaine and amphetamines, taking it as a given that there is going to be some amount of drug use. Drug policy ought to take this harm reduction approach of directing people to the least dangerous drugs. That’s in contrast to the current strategy of banning everything that’s fun. Luckily, salvia is still legal in most states and countries, though some have banned it. It’s often included on lists of “legal ways to get high.”

Dextromethorphan (DXM)

This is the active ingredient in many cough medications. Some people have figured out that you can get high if you take several times the suggested dose. The effect of a large dose is qualitatively different from the effect of a small dose. (This was the topic of a South Park episode. The helpful pharmacist tells them, “Now, this is the one you want if you *really* want to trip balls.”) Deaths are rare. There were 283 deaths with “antitussives” listed on the death certificate in 2014 according to CDC statistics, but if you look at the death records almost all of these involve multiple substances. Alcohol, heroin, benzodiazepines, and prescription opioids are very common in “antitussive” poisonings; it’s not at all clear that DXM even contributed to any of these deaths. It may well be an incidental finding in most of these cases. DXM is probably pretty safe if it’s all you’re taking, and you aren’t taking a massive dose of it.

A 6-year retrospective study, from 1999 to 2004, of the CPCS showed a 10-fold increase in the rate of DXM abuse cases in all age groups and a 15-fold increase in the rate of abuse cases in adolescents. In 2004, CPCS reported 1382 DXM abuse cases. Extrapolating these figures to the national level suggests there may be millions of users, but without additional information, the conclusion must remain speculative.

It’s not at all clear that this is a big problem, considering the safety profile of DXM. I did know someone who was abusing DXM and it really did cause problems in his life. If you’re spending all your time zonked out, or you try to go into work or operate a vehicle in a zonked state, then obviously there’s a problem. But let’s be clear that there’s nothing inherently addictive about the drug, and it doesn’t seem to cause cumulative organ damage like alcohol or cocaine.

A normal dose of DXM is 15–30 mg. It is claimed that hallucinatory effects may be experienced with doses as low as 100 mg and frank psychosis at higher levels (Roberge et al., 1999; Miller 2005). Ingestion of hundreds of milligrams has been reported, with few or no significant side effects.

Because deaths involving DXM are rare and produce only nonspecific findings, such as pulmonary and cerebral edema, it is hard to say what sort of mechanism would be involved.


This is the drug that killed Michael Jackson. (Although it was inappropriately administered by his doctor.) Its use is not common and not much is known about its incidence and epidemiology:

Nothing has ever been published about the prevalence or epidemiology of propofol abuse or propofol-related deaths. According to a review paper published in 2009, there had been 38 reported deaths, most of which were the result of accidental overdoses (Kirby et al., 2009). Prevalence of abuse is highest among medical professional and allied staff (e.g., nurses and radiologists) as they have the most ready access.

So this one is a big question mark. It looks like most of the users are medical professionals who have access to a legitimate stash of the drug. Clearly it can be fatal, but without knowing the number of total users there’s no way to quantify the risk. (We don’t know what denominator goes with those 38 deaths if we wanted a death rate per 100k users.)

Here’s a common theme in the book: blood concentrations rise after death. So even something that looks like a massive overdose may not be:

As a consequence, blood concentrations measured at autopsy will almost certainly be higher than plasma concentrations during life. Consequently, there is no concentration postmortem that will indicate the likely contribution to death without a thorough review of the circumstances and associated medical and pathologic information (Kranioti et al., 2007; Kirby et al., 2009).

So some of these drugs are almost completely safe, and some of them less so. None of them appear to be quite as deadly as alcohol, tobacco, cocaine, heroin, or prescription drugs in terms of sheer numbers or even rates-per-user. They don’t seem to cause the decades-long addiction problems that these other drugs cause in their users. But they clearly aren’t very popular. Salvia is still legal in many places, but salvia use appears to be mostly a phenomenon of teenagers, perhaps acting on a dare. (Again, watch some of the youtube videos. It looks like they’re doing it for the novelty, but most aren’t really enjoying themselves.) DXM, being the active ingredient of cough syrup, is readily available, even to the kids of South Park. GHB is illegal, but given its good safety profile it should probably be de-scheduled by the DEA and outright legalized. It’s still somewhat popular as a party drug. If it’s keeping people away from heroin or cocaine, or even alcohol, that’s a good thing. Or maybe there’s a yet-undiscovered drug in this class that gives users a mellow high without any nasty side-effects. Perhaps Star Trek's "synthahol" is a yet to be discovered drug within this category. Our drug policy needs to take the approach of substituting safe recreational drugs for the dangerous ones, because the “ban everything/Just Say No” approach isn’t working. The desire to get high doesn’t magically change when you pass a law, but we can at least nudge people toward relatively safe substances. This class of drugs shows some promise in that vein. 

Friday, September 23, 2016

Talk of “Free Will” Is So Much Navel Gazing

There was someone I used to interact with on Facebook a lot. We used to have endless arguments about what “free will” means. I couldn’t use simple words like “choose” or “decide” without being sucked into a tedious discussion about his favorite hobby-horse. It was all pointless because none of these arguments (on the actual topics I *wanted* to discuss) ever hinged on how the “free will” question was answered. Even had I fully explained a working model of human consciousness and decision-making, it wouldn’t have changed my policy views or social outlook on any topic.

I eventually realized that I don’t really care if you have free will or not. I care how you will react if I treat you as a responsible human being versus if you are shielded from the consequences of all your actions. All this endless talk about free will is meaningless navel-gazing. It doesn’t really matter if you’re a human being with agency, feelings, values, etc., or a mindless robot programmed to act like a human being with all those nice things. It doesn’t matter what’s going on under the hood. What matters, to me and to everyone else around you, is whether you behave like a good person or a shitty person. 

Thursday, September 22, 2016

The Implications of Denying Economics as a Science

There are many arguments that proceed as follows. One person will propose a policy (like the minimum wage, worker protection, job safety requirements, licensure requirements, etc.). A second person will suggest that economic theory predicts bad consequences for such a policy, usually compelling enough to render a verdict against that policy. The first person will respond by saying, Well, economics isn’t really a science, so it shouldn’t inform our opinions on such matters.

This is in contrast to another kind of argument, in which one person suggests that economic theory predicts bad consequences and the other person argues that it predicts good consequences. I applaud this kind of discussion. My point in this post is to address the first kind of argument, in which one of the discussants denies that economic reasoning has anything to offer.

 I’ll start with the observation that most economic reasoning is simply common sense disciplined with some mathematics, such that the speaker defines clearly what they mean. This process of clarifying alone can make the thinker realize that an initially plausible idea is utter nonsense once it’s laid bare. Sometimes mere quantification and accounting can make it obvious that a policy is silly. And some things that are dismissed as economic “theory” are unimpeachable truths, like that “when the price of doing X goes up, people do X a lot less” and “when the cost of producing X goes up, people produce less X” (along with the inverse of these statements, which are logically implied by the statements themselves).

But never mind that. Let’s take economic nihilism as a given and concede that it’s a legitimate intellectual position. What are the implications? Basically, your hands are tied if you can’t reason in any systematic ways about economic policy. How do you predict the effects of a minimum wage if you don’t have a theoretical framework for estimating the effect on the demand for low-skilled workers? How do you render a verdict in favor of the minimum wage if you can’t do this kind of cost-benefit analysis and quantification of the trade-offs? You’d be reduced to simply arguing, “My gut tells me X is a good policy, so I support it.” But then there’s nothing left to discuss. Without a little bit of economic theory, there is no framework for identifying good and bad economic policy. Maybe you think you can appeal to a moral trump card, as in “We should enact the minimum wage because it’s just the right thing to do!” But if the practical effects of minimum wage are massive unemployment for the very poorest workers, then those effects should carry some moral weight. It’s not possible for something to be “Just the right thing to do” or “Just wrong” without any regard to their consequences.

Maybe some people think they can escape this messy business of “theorizing” by appealing to pure empiricism. But this is wrong, because you can’t evaluate data except in the context of some kind of theory. Even by calculating something so simple as a sum or an average, you are positing a theory that that number is somehow relevant to the discussion. Empiricism without theory tends to sloppy reasoning. (Don Boudreaux makes the point beautifully here in one of my favorite posts by him.) You end up making a lot of “post hoc ergo propter hoc” and “correlation implies causation” arguments.

So what are the policy implications of denying economics as a science? They aren’t the implications that the denier wishes to draw: “Now I’m free to basically pick whatever policies make me feel warm and fuzzy inside.” No, you need to set some sort of reasonable default rule for what policy *ought to be* given our radical uncertainty. Maybe democratic popularity? Unfortunately, this leads to incoherent policy, full of self-contradictions. If you put every individual policy up for popular vote, you’d get a hodge-podge of unclear laws and factions of warring bureaucracies. (Oops! I guess I mean you'd get *more* of this.)  Pure populism also tends toward *bad* policy. If you’re ideologically democratic, meaning you think that majority should rule, right or wrong, you shouldn’t dodge the question of what the populace really wants. Do they *really* want their slate of policies? Or do they *really* want economic prosperity? If the latter, then you should subvert the former, unless you think that these things are always in agreement by some obscenely unlikely coincidence.

I think the only policy default rule that makes sense is the libertarian default: don’t initiate violence unless there is overwhelming justification for doing so. Don’t threaten, harass, and arrest business owners for offering a wage that their workers all agree to, just because you think it’s too low. Don’t harass those business owners for failing to comply with a regulatory “check-list” of “thou shalt”s and “thou shalt not”s. At any rate, don’t do any of these things unless such threats and harassment reap overwhelming benefits for society. But you’d damn sure better show your work. And that requires some economic reasoning.

As an aside, I agree with the criticism that much economic reasoning is overly mathematized and abstruse. (See Bryan Caplan’s post here.) But the critiques of these approaches use…economic reasoning! As in, something in the math doesn’t make economic sense. (See Caplan’s post: “You could reply that mathematical economics shows that economic intuition is often wrong.  I beg to differ.  My experience: When mathematical economics contradicts common sense, there's almost always mathematical sleight of hand at work - a sneaky assumption, a stilted formalization, or bad back-translation from economath to English.“)

You can’t escape economics if you want to make an argument about economic policy (or *any* policy for that matter). To do so is to intellectually anesthetize yourself. If you don't have the patience for economic arguments, then you should probably refrain from having any strong policy opinions. 

Wednesday, September 21, 2016

Mostly Harmless Hallucinogens - Evidence From a Medical Textbook

I’ve been reading a textbook called Karch’s Pathology of Drug Abuse. It discusses all the major drug categories, their histories, their chemical structure, their effects, the various kinds of organ damage they cause, etc. There is a chapter on hallucinogens, and it turns out this category of drugs is extremely non-toxic. Below I’ll share some of the passages I’ve highlighted with some commentary. It depends on which particular one you’re talking about, but by-and-large these drugs *won’t* kill you. They have a very good safety profile. Near the beginning of the chapter, the authors say:

“Except for the designer amphetamines, none of the hallucinogens discussed is associated with unique or specific pathologic lesions.”

Translation: Out of an entire chapter on hallucinogens, only one class of these drugs (MDMA/Ecstacy and its analogs) causes physical damage to the body.


“No mescaline-related deaths or emergency room visits have ever been reported in any DAWN survey, and only two case reports describing mescaline-related deaths have appeared in the world literature.”

In other words, mescaline is completely non-toxic. If one of those “mescaline-related deaths” is the same one described in Paul Gahlinger’s book Illegal Drugs, it was “a man who had severe alcoholic liver disease and died from bleeding into the stomach when he vomited.” Out of millions upon millions of uses, they can only find two deaths in the entire world literature, one of which is more appropriately blamed on alcohol than on mescaline. Given the safety profile (and inserting a little bit of Bayesian reasoning), it’s extremely likely that this other death was one where mescaline use was incidental.

“No specific pathologic findings have been identified. A search of the California Poison Control System database (1997–2008) uncovered 31 single-substance exposures to peyote or mescaline. Almost all of these (97%) were intentional and the drug was taken orally, with only individual patients who insufflated mescaline powder. None of these subjects died or even required hospital care (Carstairs and Cantrell, 2010).”

This is the profile of an extremely non-toxic drug. However:

“According to the Drug Enforcement Agency, the hallucinogenic dose of mescaline is about 0.3–0.5 g and the effects last about 12 h.”

So I can see why someone might freak out and call poison control or go to the emergency room, even though they aren’t in any physical danger. This is an issue with marijuana as well. Some critics of marijuana legalization point to the number of ER visits reported in government statistics, but it needs to be remembered that these people aren’t in any physical danger. The peyote/mescaline chapter in “Illegal Drugs” says that basically someone who takes an “overdose” just vomits up the excess, so a fatal overdose is ruled out of the question. (There is apparently a lot of vomiting involved in these mescaline trips.)


“Later in the 1950s, the U.S. Army contracted with a group of researchers at the University of Michigan to perform MDMA toxicity studies. The results of the Michigan study remained classified until 1973, when they were finally released. The studies showed that MDMA was somewhat less toxic than MDA but more toxic than mescaline (Hardman et al., 1973). MDMA was classified as a schedule I drug in 1985.”

This is kind of a subjective judgment, but I appreciate that the author is talking about *relative* risk, not  just taking the line that “anything risky is bad”, as some anti-drug rhetoric does.

“No MDMA-related deaths are listed in the DAWN report for 1999 (Kissin et al., 2000). The emergency room component of the most current DAWN report lists 8621 MDMArelated visits (CI 5,985–11,257). Whether any were fatal is not known (SAMHSA, 2006).”

With probably a million or so users and several million incidents of use, this is pretty stunning.

“In a 2006 U.K. survey, based upon medical death certificates, coroners’ reports in the United Kingdom tended to underestimate the problem of MDMA, MDA, and MBDB deaths occurring from 1994 to 2003 found a total of 394 deaths, although suicides are rare and deaths are almost always accidental. In 42% of the deaths identified, MDMA was the only drug detected. All of the other deaths occurred in polydrug abusers, making causality assessment all but impossible.”

So there is a very real danger with MDMA. It’s possible to overdose. But this is mostly a problem that prohibition has saddled us with, not a problem inherent to the drug itself.

“Once MDMA is produced, it is converted into tablets, each with a recognizable logo (see Figure 4.16). The tablets quickly develop names based upon the imprinted logo. Brand loyalties evolve, and users ask for particular tablets by name (tablets with the Mitsubishi label were once particularly popular, with a very loyal following). Some of the logos often have whimsical themes, ranging from an imprint of McDonald’s Golden Arches to the Rolex trademark symbol, the Mercedes symbol, and even the skull and crossbones. A recognized logo is, however, no guarantee of quality, safety, or purity. Once the makers have developed a following for a pill with a particular logo, they then begin substituting cheaper ingredients in the pill such as methamphetamine or even PMA (a much more dangerous drug than MDMA) (Lora-Tamayo et al., 2004).”

This would obviously not be the case in a legal market. Producers would compete on the safety and purity of the MDMA they are selling, and buyers could be reasonably sure of what they are getting. It’s impossible to have government regulators or third-party auditors testify to the purity of a drug that’s produced and sold in an illegal market. Such third-party assurances, however, are the norm in legal markets.

“The most feared complications of MDMA use are serotonin syndrome (now increasingly referred to as serotonin toxicity), hyperthermia with rhabdomyolysis, and hyponatremic encephalopathy. The first disorder is characterized by the rapid onset of confusion, diaphoresis, diarrhea, increased muscle tone, and cardiac arrhythmias. There may also be shivering, myoclonus, and increased deep tendon reflexes.”

I don’t want to understate the dangers of MDMA use. Serotoninsyndrome sounds scary as hell. But it sounds like this is mostly an effect of multi-drug use, including prescription antidepressants:

“Risks for developing the syndrome are higher in individuals taking SSRI and monoamine oxidase inhibitor (MAOI) drugs. Most adult drug users take multiple drugs, and it is worth remembering that many of these other drugs may have weak, but quite real, SSRI and MAOI activity (methadone, tramadol, dextromethorphan, and meperidine).”

I remember hearing reports that Ecstasy use “turns your brain into Swiss cheese.” I think this passage sheds some light on that claim:

“When intracranial bleeding does occur, it is usually as a consequence of a preexisting malformation such as undiagnosed aneurysm or arteriovenous malformation (Hughes et al., 1993; Auer et al., 2002; Drees et al., 2009). Several reports of cerebral infarction were published in the early 1990s (Manchanda and Connolly, 1993; Hanyu et al., 1996); few new cases have been published recently (Goldstein and Mordish, 2006). Given the very great number of users today and the paucity of new cases, intracerebral infarction or hemorrhage should not be high in the differential diagnosis. Similar considerations apply to the one report of MDMA-related spongiform encephalopathy (Bertram et al., 1999), where the victim may have used other drugs in the past.”

Emphasis mine. So in those cases where MDMA users have some kind of brain lesions or bleeding, it’s usually do to some other cause, perhaps exacerbated slightly by MDMA use.

“Even though there is a wealth of experimental evidence, much of it derived from PET and fMRI scanning of human volunteers, suggesting that MDMA is toxic to serotonergic neurons, there is no clinical evidence that humans ever develop the typical symptoms of 5-HT depletion (disorders of sleep, mood, appetite), and the most recent evidence suggests that depressive symptomology simply does not occur in occasional users (Falck et al., 2008).”

I read this as saying, you can see these anatomical effects in the lab, but you don’t actually see it out in the field. The brain scans are suggestive of a disease that doesn’t actually manifest itself in the population of users.


Tellingly, the section on DMT is very short, because the drug is fairly harmless. An incredibly intense and long-lasting version of the drug is called ayahuasca. Massive quantities of DMT are ingested, and the trip lasts an entire day. And yet deaths are exceedingly rare, with only a single death recorded in the entire world literature. Ryan Grim describes an ayahuasca trip in his book “This is Your Country on Drugs.” He describes the intense and introspective vision quest, in which his conscience convinced him to stop being a journalist. After sobering up, he promptly realized that he liked being a journalist and his “revelation” was a false one. As with other psychedelics, it can give the user false moments of clarity and false insights, but it certainly doesn’t compel them to do anything. According to Grim’s account, there’s a lot of giggling, some barfing, and some false revelations, all done in the presence of a shaman (essentially a “trip-sitter”). But nothing dangerous or permanent. False revelations are simply ignored.

“Only one alleged case of ayahuasca poisoning associated with recreational abuse has ever been reported or studied. Autopsy was performed within 24 h of death and no gross lesions were apparent. Blood levels from various sites were measured (see Table 4.10); however, these levels cannot be reliably used to determine the cause of death for a number of reasons: (1) the decedent, a 25-year-old, could well have suffered from a heritable channelopathy, but no testing was done; and (2) nothing is known about blood levels in the living or the behavior of any of these compounds after death, nor is anything known about the toxicokinetics of smoked DMT;”

“The risk of addiction is thought to be negligible (Gable, 2007). Overall, the risk compared to other drugs seems to be very low. In spite of earlier work suggesting hormonal and blood pressure changes, in the most recent human study, where freeze-dried ayahuasca was given (two doses, 12 h apart, 0.75 mg/kg), the only statistically significant findings were mild decrease in heart rate and blood pressure and a marked rise in growth hormone secretion (Dos Santos et al., 2012).”

This is a remarkably safe drug. There is no reason whatsoever to ban it.


Psilocybin is the active chemical in “magic mushrooms.” Ironically, the biggest danger in consuming these is mistaking a poisonous mushroom for the (very non-toxic) psilocybin-containing variety. This is another danger that isn’t really inherent to the drug itself but is exacerbated by prohibition. No doubt some amateur mushroom-hunters have poisoned themselves, whereas a legal supplier would never sell a toxic mushroom by accident, given reputational risk and the risk of license revocation implied by such an error. Here is what Karch has to say about this problem:

“Identifying wild Psilocybe is difficult and dangerous. Psilocybin-containing mushrooms grow side by side with the poisonous Galerina autumnalis. Galerina species have rust-brown-colored spores, while the spores of Psilocybe species are gray to lilac. Some, but not all, species can be distinguished from poisonous mushrooms by their reaction to room air; when Psilocybe mushrooms are cut, they oxidize and turn blue within 30–60 min. Unfortunately, some poisonous mushrooms can do the same thing. Pathologists are much more likely to encounter cases of mushroom poisoning than they are to encounter psilocybin-associated medical problems!”

Physically, the effects are quite mild:

“After oral doses of up to 15 mg, psilocybin produces no significant alteration in heart rate, blood pressure, or neuroendocrine function, although profound psychological alterations do occur (Gouzoulis-Mayfrank et al., 1999).”

There is no discussion of mortality or long-term health effects, presumably because there just aren’t any. The authors comment on the toxic effects of every other drug discussed in the book, even mentioning when there are only a handful of deaths (or a *single* death, as in the case of ayahuasca/DMT). This is another safe drug. Legalization is a no-brainer. The textbook doesn’t mention this, but my understanding is that users develop such a fast tolerance to psilocybin that it’s impossible to use it on consecutive days. Users have to give it a break for a week or so for it to have an effect again. So it’s  not “habit-forming” in the sense of compelling the user to take more, as in the case of opioids or tobacco.


LSD is powerful even at extremely low doses, but it’s fairly non-toxic.

“In the late 1990s, interest in LSD seemed to renew, but reports of toxicity continued to remain extremely rare. The emergency room component of the DAWN report contains 2006 LSD mentions during the first half of the year 2000 and 2028 in 2009 (compared with 422,896 cocaine-related visits in the same year).”

It’s not terribly surprising that some people having bad trips ended up in the emergency room, but note the lack of death statistics (also tracked by DAWN, and mentioned throughout the book for drugs that are significantly lethal).

“When formal neuropsychological testing is performed, few, if any, sequelae can be attributed to LSD use (or to the use of any other hallucinogen, for that matter) (Halpern and Pope, 2003).”

“Sequelae” being a fancy word for “bad stuff that happens after you take drugs.” This is a fitting summary, on the last page of the chapter on hallucinogens.

“As recently as 1990, it was generally believed that no death had ever been caused by the direct action of LSD; however, several very probable cases have been reported and more than a few have required medical care.”

“Several probably cases” out of decades of use by millions of people is a pretty good safety profile.

“In 1975, “eight patients were seen within 15 min of intranasal self-administration of large amounts of pure d-LSD tartrate powder. Emesis and collapse occurred along with sign of sympathetic overactivity, hyperthermia, coma, and respiratory arrest. Mild generalized bleeding occurred in several patients and evidence of platelet dysfunction was present in all. Serum and gastric concentrations of LSD tartrate ranged from 2.1 to 26 ng/ml and 1000 to 7000 ng/100 ml, respectively. With supportive care, all patients recovered” (Klock et al., 1975). However, the situation has changed drastically since LSD was so popular and now, more often than not, one of the designer amphetamines such as 25I-NBOMe is likely to be detected.”

In the above passage, the author describes eight patients who recovered from a bad acid trip. The last sentence is saying that “LSD” sold today is likely to be something else, possibly something much more dangerous. Once again, this is not an inherent danger of the drugs themselves. The danger is an artifact of the illegal market, where it’s difficult to hold sellers and distributers to account for selling a bad product. It’s another problem that would go away if we allowed people to legally sell these substances in well-labeled dosages, so the user knows what they are taking.

This is a remarkably safe class of drugs. They also have therapeutic value. Some have been used to successfully help patients with drug or alcohol addiction. It would be great to shift people from, say, cocaine or heroin use to psilocybin use. Surely a significant (non-zero, anyway) number of existing drug abusers would make that substitution if it were easier. If we want to implement a harm-reduction approach to the drug problem, this is very low-hanging fruit. Hell, some of it is literally on the ground.

I’ve never personally used any of these. Anyone considering doing so should read anything and everything they can find before doing so. “Karch’s Pathology of Drug Abuse” is sometimes hard to read, but it’s worth doing the basic research if you’re potentially monkeying with your health. Erowid is another great source. Wikipedia is fine, so long as you check a few of the references. “Illegal Drugs”, “Saying Yes”, and “Buzzed” are some good reference books. And this post at Less Wrong is a must read. Stay safe out there, but don't fall for the propaganda. Some of these substances are either completely safe or are safe in comparison to risks that we all take on every day of our lives. Their legal status as "banned" is absurd. 

Friday, September 16, 2016

Why Wait for “Economic Growth” to Raise Your Income?

Something annoys me about how the media reports economic growth statistics. All the major outlets are talking about the recent good news from the Census Bureau. Supposedly middle-income households have seen their wages increase for the first time in decades. (Here’s one such story. Every major outlet has reported on this, as far as I can tell.)

I think this is the wrong way to think about income growth, especially with regard to your personal income. A nation-wide average is not a number that you can personally control with any decision you make. Your contribution to that is probably a rounding error, even if you are an enormously influential policy maker like the Fed chairman or the president. On the other hand, you have *enormous* control over your personal income. Do you want to see *your* personal income rise with double-digit growth for several decades? Then pick a lucrative profession and start hustling!

Do you really expect your earnings to increase without making any kind of change in your life? Well, in that case you’re asking other people to work harder so that your income will grow. You’re essentially asking other people to do the heavy lifting so you can keep doin’ what you’re doin’. I can't help but think this attitude is a little bit selfish. That's not to say you won't share in the growth anyway. You most likely will. Some incredibly productive individuals will invent products that you use daily, that entertain you for hours, that replace several other bulky and expensive products. (Are you reading this on a smartphone? A laptop computer? Then you really are the beneficiary of such individuals.) Your life really will keep improving, probably even if you don’t make adjustments that make you personally more productive. Economic growth is great that way, and I certainly approve of pro-growth policies (lower taxes, less burdensome regulation, lower government spending, etc.). But it’s not something you should feel entitled to. The world doesn’t owe it to you. 

If a very large number of people were hustling, as described above, to improve their individual careers, there would be more good news regarding increasing wages. It would be accomplished by private policy rather than public policy. It’s great to have *both* kinds of policies be “pro-growth,” but individuals should concern themselves with the one they can actually affect. If you’re waiting for the rising tide to lift your boat, you’ll be waiting a very long time. 

People Choose Their Career Paths, and Therefore Choose What Income They Earn

If people can choose their personal income (as opposed to that number being assigned to them by some cosmic roll of the dice), then it makes no sense to worry about income inequality. I'm using "choose" to mean "select from available options"; obviously you can't just pick some arbitrarily high number. I present the following as “Exhibit A” for the claim that people do, in fact, choose their personal income:

(Scroll down until you see the salary ranges.) I know a lot of smart, underemployed young people who could easily knock out at least the first two or three exams and land a decent analyst job. They’d improve their salaries enormously. They’d also massively improve their options (where to live, who would date/marry them, etc.). It might take a few hundred hours of studying to pass exams 1 and 2. For a well-equipped math person, exam 1 probably requires about 100-200 hours of study, similar for exam 2, at least back when I took them a few years ago. That sounds like a lot, but the opportunity cost is low if you’re current career prospects aren’t so good. And you don’t have to keep going if you decide the exams are boring. You can land a job as a well-paid Excel jockey after completing only two or three. (I personally landed my first actuarial job after passing a single exam.)

These are solid middle-class jobs, well above the median for personal income. All it requires is to hunker down and study for a bit. This is by no means “out-of-reach.” It’s within striking distance of a lot of struggling people, and they don’t reach for it. I can only conclude that they don’t wanna. That’s not a moral judgment. I’m not at all saying they *should* want this career path over whatever it is they’re doing now. And that’s actually my point. We *shouldn’t* be second-guessing other people’s career decisions, nor should we be bemoaning their “inadequate” personal income. For whatever reason, this person had decided to spurn a perfectly lucrative career path for whatever it is s/he is doing right now.

I’m not saying that everyone in the world has this option within arms-reach. I’m sure there are some people who just aren’t smart enough to pass the exams, no matter how hard they study. Some people just don’t have a knack for math, and it’s too late to learn (although I personally think this is overblown and most people can improve their math skills more than they realize). Some people have spotty employment histories, or perhaps a history of being fired or arrested. Such a history will follow you no matter how resolutely you decide to shape up and fly right. (You know, starting now. Or maybe tomorrow.) And maybe some people have personality issues that make them difficult to employ. That’s all fine. Not everyone in the world can turn themselves into actuaries with a few months of study, and I’m not claiming everyone should. The actuarial career path isn’t the only option in the world for improving your future prospects. It happens to be the one that was most readily available to me when I was looking for a grad school exit-strategy, but there are all kinds of options like this one at all levels of education. Learn to code at coding bootcamp, or take a bunch of Coursera classes. Or get certified in a skilled trade (HVAC, plumbing, electrician, etc.). I’m using the “become an actuary” example because it’s relevant to a lot of people who I happen to know, people who could improve their outlook if they really tried. But these *kinds* of unexploited career options are available to everyone. If you really thought about it, you could make yourself more employable and get a more lucrative career. It would take some time, but you could do it.

I think it makes a lot less sense to complain about income inequality when you admit that people have a lot of control over that variable, at least in the long run. That’s not to say we’d all have equal incomes if we all tried just as hard. Some people are the beneficiaries of chance, be it the luck of being born talented or the idiosyncratic luck that rewards random people at random times. But it’s not all luck. I see many cases of equally talented, equally endowed individuals, and one decides to zig while the other decides to zag. So one ends up a doctor making $200,000+ per year, and the other lives on his mom’s couch playing video games and earning zero. It makes little sense to complain about the “unjust” distribution of incomes when you zoom in on the details. An individual’s very high or very low annual income typically become understandable the more you know about them. Rarely will you see that someone’s lot in life is something that was simply “done to” them. 

Monday, September 5, 2016

When a "Drug Overdose" Isn’t

Below are some passages I highlighted in my copy of Karch’s Pathology of Drug Abuse, 5th edition, specifically from the chapter on cocaine. I’ll give away the punchline: this is another post arguing that drug overdose death statistics are probably overstated. I wrote about that here with some passages from the chapter on opioids. Without too much throat-clearing, let's dig in:

[A] handful of important pharmacokinetic studies have been published. In one of these, an uncontrolled clinical study, plasma cocaine concentrations in 111 symptomatic cocaine users were studied and an attempt made at correlation of plasma concentrations with clinical symptoms (Blaho et al., 2000). The study was unsuccessful to the extent that no correlation between dose and effect in real cocaine abusers was demonstrated. Failing to find a difference is, presumably, a result of cocaine tolerance, a process that begins to emerge after the first dose (Howell and Ezell, 1990; Mendelson et al., 1998). Using human volunteers, Van Dyke et al. (1982) showed many years ago that tachyphylaxis occurs after even single dose of cocaine.

In other words, cocaine users rapidly develop a tolerance to the drug, so it’s impossible to specify a threshold for “overdose.” (Tachyphylaxis is a fancy word for diminishing responses to successive doses of a drug.)

The results of these studies suggest that (1) chronic users may consume multigram quantities of cocaine with relative impunity; (2) a variety of cocaine metabolites previously thought to be insignificant (EME, norcocaine) are, in fact, formed in fairly substantial quantities; and finally, (3) the half-life of cocaine may be much variable than had previously been thought. All of these variables are partly a function of the route of administration. 

This would make it impossible to tell if a dead body with very high blood levels of cocaine are actually overdoses or heavy users who die of unrelated causes.

Although the plasma cocaine level approximately doubled following the second cocaine administration, the ratings of positive drug effects, heart rate, and blood pressure did not increase after the second cocaine administration. Clearly, acute within-session tolerance develops during repeated administration of intranasal cocaine (Foltin and Haney, 2004).

This passage illustrates just how rapidly cocaine tolerance forms in a user. You might have the image of a coke-fiend chasing a high and consuming a Tony Montana-esque quantity of cocaine. I’ve heard such descriptions of cocaine users, although I suspect these are mostly exaggerations. At any rate, if the users are rapidly developing a tolerance to the drug’s physiological effects, they are also developing a tolerance to the toxic effects.

[H]ighly accurate measurements are seldom of any help in differentiating deaths due to drugs from those deaths where the presence of drugs is simply an incidental finding or, as is increasingly recognized, an interaction between two drugs to produce a third and even more toxic drug. Why are such precise measurements of so little value? Because, except for episodes of massive overdose (as might occur in a drug mule) where the mechanism of death is perfectly clear, most cocaine-related deaths occur in chronic drug users in whom death is a consequence of neurochemical and anatomic changes induced over a period of months or even years.

So here he just comes out and says that the tissue concentrations of cocaine can’t be used to determine the cause of death. I wonder how many medical examiners understand this point, and how many just mark the cause of death a cocaine overdose when it’s a handy and convenient answer? The number of cocaine-related deaths in 2014 was 5,992, according to CDC figures. (It actually went up, from 4,494 in 1999 to a peak of 8,591 in 2006, then back down to 5,992 in 2014, just to give you a sense of what it’s done over the past 15 years.) I’m curious if most of these are really cocaine overdoses, or if most of these are mislabeled and the numbers are tracking with usage but not necessarily with mortality?  As in, “There are more corpses with cocaine in them in 2006, fewer in 2014, but we really have no idea if cocaine was the cause of death in the vast majority of these deaths.” Then again:

Long-term cocaine users have changes in their hearts (Karch et al., 1998) and in their brains (Volkow et al., 1993) that favor the occurrence of sudden death. If the abuser is truly unlucky, he or she may be heterogeneous for one of several abnormal hERG ion channels (Guo et al., 2006) or polymorphic catecholamine receptors (Ghimire et al., 2012). These changes explain why, in both the living and the dead, it is absolutely impossible to correlate a specific blood or plasma concentration with a specific type of toxicity or even speculate whether cocaine-related toxicity occurred at all (Jenkins and Goldberger, 1997; Karch et al., 1998; Blaho et al., 2000).

So there is a countervailing force here that can lead to *understating* the number of cocaine deaths. Someone could have cumulative organ damage from a cocaine habit, but have no cocaine in their system at all. They could have gotten totally clean after a long period of cocaine abuse, but the damage to their hearts and brains can still kill them later. I suspect this is a small countervailing force because the number of long-term cocaine abusers is so small, but obviously this is a giant question mark.

Given these realities, the accurate certification of a drug-related death requires knowledge of (1) the decedent’s past medical history, (2) an account of what happened at the scene, (3) a thorough postmortem examination, and (4) the results of DNA testing to rule out hereditary forms of heart disease or even myocarditis; the cost of such testing is rapidly decreasing to affordable levels. Though it will still be some time until it becomes widely available.

My impression is that most autopsies aren’t this thorough, and Karch reinforces this impression with the following passage:

The definition of just what constitutes a “complete” autopsy is not as simple or clear as it once was. Does the examination of one section of myocardium constitute a “complete” examination of the heart? When a young person dies suddenly, is it proper to refer to their heart as normal when cardiac ion channels have not been measured? In unexplained cases of drug death, should the examination be considered complete if the P450 metabolizer status is not determined? The answer to all of these questions is “no.” As a consequence, cause of death determinations are often based solely on toxicologic measurements, significantly raising the chances for a missed or incorrect diagnosis. Because cocaine is so widely used, cocaine-related deaths present a special set of problems.

Clearly the author thinks that typical autopsies are not adequate to determine the cause of death in a suspected cocaine overdose.

I thought the following passage was a good explanation of why cocaine users feel compelled to keep taking massive doses of the drug, “chasing the high” as it were:

There is seldom any reason for clinicians to measure plasma concentrations of cocaine or its metabolites. Unlike alcohol intoxication, where specific blood concentrations can generally be related to specific physiologic and psychological states, cocaine blood concentrations do not relate to symptoms (Karch et al., 1998; Blaho et al., 2000), not even in the laboratory setting. Accordingly, treatment must be based on the patient’s symptoms, not on the plasma level of cocaine.

When cocaine is given to volunteers, correlations can be drawn between the degree of mood elevation and peak blood levels, but only when cocaine concentrations are rising. If blood concentrations are falling, the exact same concentration that resulted in a “high” when concentrations were rising can be associated with a dysphoric reaction when concentrations are falling. Cardiovascular effects and feelings of euphoria decline more rapidly than do cocaine blood concentrations (Javaid et al., 1978), but the “rush” experienced by cocaine users follows a different time course than the cardiovascular changes.

Emphasis added. Given this, you can easily imagine someone taking more and more cocaine during one session, chasing the state of rising blood concentrations. Given the rapid onset of cocaine tolerance, it’s not clear that this is necessarily dangerous, unless the user takes a truly massive dose.

At one time, cocaine blood concentrations of more than 5 mg/L were thought to be uniformly fatal (Wetli and Mittlemann, 1981). With more experience, it has become apparent that isolated postmortem blood concentrations cannot be used to determine the cause of death at all. Tolerance on a massive scale occurs, and cocaine concentrations well in excess of 5 mg/L can be encountered in traumarelated deaths where the presence of cocaine is clearly an unrelated finding (Pagel et al., 1994; Shannon et al., 1996). For example, one case report described a man who was shot while drinking in a bar. Prior to being shot, the man’s behavior was said to have been normal. When he was autopsied several hours later, after having undergone extensive attempts at resuscitation, including aggressive fluid replacement, multiple blood specimens showed a blood cocaine concentration of 30 mg/L (Howell and Ezell, 1990).

Further driving home the point that blood concentrations can’t be used to determine the cause of death.

The following is good to keep in mind when thinking about the topic of overdose deaths:

Most cocaine-related deaths can be attributed to the occurrence of a lethal arrhythmia.

Unfortunately, lethal arrhythmias happen for reasons unrelated to cocaine. So it’s hard to tell if cocaine was the cause or just incidental.

The last section of this chapter is actually titled “When Is Cocaine the Cause of Death?” and it reiterates the point, repeated throughout the chapter, that it’s a hard question to answer.

It is distressing to see how often deaths caused by cocaine are misclassified. Misconceptions about cocaine-related deaths persist, partly because death certification practices are not standardized and partly because there is no reproducible method by which to make the decision.

I wish this was better quantified, some thing like “50% of ‘drug related’ deaths are misclassified.” Obviously nobody knows what the real number is. But it would be nice if the CDC, when reporting on drug-related fatalities, adjusted for this misclassification problem, or at least acknowledged it.

The following is mentioned in the footnote of a figure:

Except in the case of massive overdose, cocaine-related deaths are always multifactorial.

In other words, it’s never *just* cocaine. It’s always some combination of cocaine and other drugs, or cocaine interacting with a heart condition.

Consider the example of a young cocaine abuser who dies from smoldering myocarditis that has been present for many weeks. If microscopic examination is not performed, or, even worse, if no autopsy was performed except to draw blood for toxicology testing with “a blind stick,” even if the level of cocaine present is trivial, there is a very good chance that death will be attributed to the drug.

Read “smoldering myocarditis” to mean “an unrelated heart issue.” (It’s just a random condition he chose to make his point, as far as I can tell.)  The author is pointing out that sometimes the toxicology is the sole piece of evidence that medical examiners use to determine the cause of death, even though this is completely inappropriate.

In a 2014 study of 245 U.S. college athletes who had died suddenly, with no abnormalities detected at autopsy, nearly a third of the individuals had fully normal autopsies. At about the same time, researchers in Denmark reviewed nearly 439 cases where the decedent was aged 1–49 years (Bjarke et al., 201). As might be expected, nearly 70% of the deaths were the result of coronary artery disease. But in 135 of the U.S. cases, complete autopsy disclosed no anatomic abnormality, very strongly suggesting an underlying mutation (Maron et al., 2014).

In other words, people drop dead for no apparent reason, and sometimes even a thorough autopsy can’t find any underlying anatomical problems. I was glad to see the authors reference some actual studies of this phenomenon, because this is exactly what I suspect is happening in a large number of so-called “drug overdose” cases. This is precisely what I suspect is overstating the overdose numbers.

Suppose a person with a heritable cardiac channelopathy is found dead and a modest amount of [cocaine metabolite] BZE is found in autopsy blood. For many, the temptation to classify the death as drug related may be too strong to resist if they are unaware that a potentially lethal mutation is present.

All of this together reaffirms something I've suspected for a while now. When it’s hard to determine the cause of death, toxicology offers up an easy answer. The medical examiner can fill out the death certificate and go to lunch. I don't know how big a deal this is, but a serious effort should be made to correct for this problem.

There is much, much more in this chapter. It confirms my suspicion that drug overdose deaths are inaccurately measured, and probably are overstated. I hope that the authors don’t mind my generously quoting from their book. Certainly these passages are relevant to a very important public discussion about the drug overdose problem, and certainly these considerations should inform any policy response. 

(If the authors of Karch's Pathology of Drug Abuse see this blog post and want me to cut some of the quotes, I’d be happy to paraphrase instead of quoting. But I insist that the basic point underlying these quotes is relevant to a public discussion, so some form of the above needs to be publicly available. Karch was interviewed for Radley Balko's series on the opioid epidemic at the Huffington Post, where he is quoted as making the above pionts. So surely his intention is to inform the public on this point.) 

Friday, September 2, 2016

How Much Do Labor Unions Affect Wages?

Below I have some excerpts from Deirdre McCloskey’s book Bourgeois Dignity, just in time for Labor Day. I find it obnoxious the way that some commentators attribute far too much explanatory power to labor unions. There’s a bumper sticker that says, “Enjoy weekends? Thank your local union.” Sorry, but this is just implausible. The workers got richer because they became more productive. Their greater earnings led them to purchase, among other things, more leisure time. That is why wages have been rising and the average work-week has been shortening for the past two centuries. It’s plausible that labor unions, in some select cases at certain times, made these advances happen faster than they would have otherwise, but it’s completely unrealistic to credit the labor movement with all the gains since the mid-19th century as some glib writers do.

The struggle by unions for higher wages and better working conditions has inspired many good songs and many good people. Clearly, say the songs, we gotta' go down and join the union; we need to fight for Harry Bridges and build the CIO; and we need to be a union maid when she fought for higher pay. "She'd show her card / To the National Guard, / And this is what she'd say:/ `Oh, you cant scare me: / I'm stickin to the union."' But the fight to get a higher share of the pie for union members leaves out nonunion members, who were always a high percentage of the workforce in the United States. If the size of the pie is pretty much fixed-the underlying assumption of classical economics and of union logic-then higher pay for auto workers implies lower pay for auto mechanics, not something to be celebrated in songs by Woody Guthrie and history books by progressive historians. The economist H. Gregg Lewis painstakingly estimated the effect of unions 1967-1979 on wages of their members relative to comparable nonmembers and found it to be, as an upper bound, rather small: some 14 percent. My uncle, who was himself a union electrician in Michigan, paid union wages for his employees gladly, he said, because he passed the expense on to the companies and hospitals and school buildings that he wired. Oh, good. But who pays that? Other workers. There's no one else to pay it. As Pogo famously said, "We has met the enemy, and he is us."

So, essentially unions succeeded in raising the pay of their own member by at most 14%, and that came at the expense of other workers. Meanwhile, the average wage in the developed world rose by something more like 1600% to 10,000% (a factor of 16 to 100, depending on how you account for changes in quality and other considerations).

And if, as was on the contrary the case, the pie was in fact exploding, then a shift in the size of the slice going to workers as a whole (setting aside that it would in fact go only to the minority of union members) would anyway play a small role in the betterment of even those workers. Workers in the United States and elsewhere grew radically better off from 1800 on, and 1900 on, and even from 1970 on because of the Bourgeois Deal, not because they went down and joined the union. How do I know? Answer: the nonunion people shared nearly equally in the gain, even though they were paying as consumers for the fancier wages for union electricians in Michigan. The maximum of 14 percent of "concessions" extracted by bargaining or strike doesn't come close to accounting for the great magnitude of rise in the real wage. There's not enough profit-usually 10 or 15 percent of national income-to raise the level of the rest of national income by a factor of even two, that is, 100 percent, much less the factor of eighteen it in fact rose after 18oo. Expropriating 15 percent of national income claimed by the wretched profiteers teers and transferring it to the 85 percent earned by us workers raises our income only 18 percent. That's a long, long way from 1,700 percent.

Deirdre McCloskey is making the point that labor unions can only rearrange portions of a relatively fixed pie; nothing they do is likely to grow the pie. They can negotiate with the capitalists to redistribute some of the profits from capitalists to laborers, but they can’t increase the total productivity available for distribution. People love to romanticize the 19th century labor struggle. So much blood was shed, it's irresistible to conclude that some tremendous benefit must have come from it. Alas, a little bit of numeracy and some back-of-the-envelope math implies otherwise.