Medical Marijuana and the Ecological Fallacy

Some recent studies have shown that states with more medical marijuana availability have lower rates of opioid overdose and young male suicides. This was interpreted as meaning that people who use medical marijuana are at lower risk of overdose and less likely to take their own lives. If you think that constitutes good reasoning, you should also believe that smoking and being exposed to radon reduces your risk of cancer because in the aggregate, those variables are negatively correlated with cancer rates!

The recent medical marijuana studies have fallen into a seductive logical error called the ecological fallacy, which my colleagues and I explain in detail at a post at PLOS Mind the Brain. Once you understand the ecological fallacy, you will realize that many, many news stories about research make claims that just are not true (we give examples in our post).

Perhaps surprisingly, whether medical marijuana availability at the state level correlates with some other state-level indicator actually tells us nothing about how medical marijuana affects individuals. If a state-level correlation with some indicator is positive (e.g., states with more medical marijuana have higher rates of violence) the individual level relationship can still be negative (e.g., medical marijuana use makes people less violent).

Some medical marijuana activists have argued to me that it is inappropriate to point out that the studies in this area are methodologically flawed because doing so harms “the cause”. I don’t sympathize with such Lysenkoism. I believe scientists should seek and report the truth regardless of whether it concords with a political agenda (The scientist’s own or anyone else’s). If we give up on that regulating ideal of scientific inquiry, we can’t really complain when other people deny the evidence of oceanic acidification change or assert that the MMR vaccine causes autism.

The Hows of Marijuana Legalization

weed Americans have expended tremendous energy debating the why/why not question of marijuana legalization. In contrast, little attention has been given to the hows of marijuana legalization, e.g., would a legal industry be for-profit or non-profit? How and at what level would it be taxed? How would it be regulated? The hows matter enormously. Indeed, once they are spelled out, some people who think they are against marijuana legalization realize that they could support it, and some people who think they are for marijuana legalization realize that they don’t want it after all.

The California Blue Ribbon Commission on Marijuana Policy was set up by the Northern California ACLU to dig into the hows of marijuana legalization. The Commission is not itself going to write a marijuana legalization ballot initiative nor is it going to oppose or endorse any that are written by others. Rather, we are a mix of a think tank and a public education enterprise, encouraging the public to consider carefully what marijuana legalization might look like if it were adopted in California. Everyone is welcome to attend the public events of the Commission as well as to send in their thoughts directly through our website.

My fellow commissioner Professor W. David Ball and I were recently on KQED Forum to discuss the Blue Ribbon Commission’s work. Listening to the broadcast will give you a flavor of the issues with which Californians will have to grapple as they consider the 2016 marijuana legalization ballot initiative(s).

Be Skeptical of Illegal Drugs as Miracle Cures (Legal Drugs Too)

Like everyone else, I constantly see headlines that the cure for some dread disease has been discovered. On those occasions when journalists interview me about such stories, I have a habit of dispensing cold water. For example, a few years ago, a small clinical trial seemed to show that anti-depressants helped meth-addicted people to stop using drugs. This is what I said to an excellent health reporter, Erin Allday, about the findings:

“There have been quite a few bombs pharmacologically…those earlier experiences have taught me to be cautious now.”

Being skeptical about miracle cures is simply playing the odds. As my colleague John Ioannidis pointed out in one of the most-read papers in medical history, most medical research findings are wrong. This is particularly true of small studies, which are usually followed by larger studies that disconfirm the original miracle finding (Fish oil pills are a good example).

Lately, currently illegal drugs such as LSD, ecstasy and marijuana have been aggressively hyped as miracle cures for a range of serious disorders (cancer, diabetes, PSTD, alcoholism etc.). You may have heard for example dramatic anecdotes “proving” that high-CBD marijuana cures seizures in children. Sounds great, but as more data were gathered by neurologist Dr. Kevin Chapman “the miracle” took a beating:

Dr. Chapman’s study, which involved a review of the health records of 75 children who took CBD, found that 33% of them had their seizures drop by more than half. However, 44% of the children experienced adverse effects after taking CBD, including increased seizures. Of the 30 patients whose records included the results of brain-wave tests, a less subjective measure of seizure activity, only three showed improvements in those exams.

“It really wasn’t the high numbers we were hoping for,” Dr. Chapman said.

No one who understands medicine will be surprised by this result. It happens every day with initially touted legal miracle cures too (e.g., PROMETA for methamphetamine addiction). Alas, legal or illegal, most flashes in the medical pan are pyrites rather than gold.

Can Medications Turn Problem Drinkers Back Into Moderate Drinkers?

The organization that integrates scientific information for the UK National Health Service has recommended that the drug nalmefene can be prescribed to alcohol dependent patients for the purpose of reducing their drinking but not necessarily stopping it. Nalmefene, like the better known medication naltrexone, is an opiate antagonist that is intended to make the experience of drinking less rewarding. This reduction in subjective reward could make it more likely than a drinker will stop at 2, 3 or 4 drinks rather than going on to 5, 6, 7 or many more drinks.

Many people will find this concept strange, as they think that the purpose of opiate antagonists (indeed all alcohol-related medications) is to help problem drinkers continue pre-existent abstinence (e.g., Keep dry after leaving a detox center). Sometimes they do just that. However, from the point of view of behavioral extinction, one could argue that its actually a good thing when a person on an opiate antagonist decides to drink. They will then experience drinking behavior as less rewarding, which should over time reduce the behavior and also the craving to drink.

My colleagues and I, led by Dr. Natalya Maisel, conducted a meta-analysis of this question focused on the medication naltrexone, which we contrasted with a different medication prescribed to problem drinkers that is not an opiate antagonist (acamprosate). The entire research synthesis is available for free here, but the key finding for present purposes is that while acamprosate was more strongly associated with maintaining abstinence, naltrexone was better at limiting the number of drinks on drinking days.

Does this mean that every problem drinker can return to moderate drinking if they just take naltrexone or nalmefene? No. Problem drinkers who have more physical dependence symptoms (e.g., needing a drink in the morning, tremors, cravings), less social capital and worse mental health are generally unlikely to return to moderate drinking no matter what treatment they get. However, if dependence isn’t too advanced and the problem drinker has significant social and psychological resources available to support their efforts, they may indeed be able to go back to non-problem drinking, and medication can help with that. Even then, it’s a decision to make in consultation with a health professional, and the taking of the medication should be accompanied by additional counselling.

Prohibition and Marijuana Potency

At Washington Post Wonkblog, Christopher Ingraham notes that the potency of available marijuana has been increasing:

Retail outlets in Denver and elsewhere advertise strains that contain 25 percent THC or more. As legalization opponents are forever fond of saying, this isn’t your daddy’s weed.

There is extensive evidence available to support Christopher’s observation about rising marijuana potency. But I am not sure that his causal explanation for the potency increase comports with the facts. He attributes it to prohibition and believes therefore that legalization will reverse the trend:

As prohibition eases and legal markets open up, growers now have the breathing room to select for traits beyond high THC content. Demand from new users looking to experience a social high, rather than four hours of couch lock, will likely drive this. The end result may be a resurgence of milder strains of weed that are more akin to fine wines than to bathtub gin.

There are a couple analytic problems here. First, over the past 5-10 years, legalized marijuana markets (medical and recreational) have expanded dramatically and marijuana enforcement has dropped sharply, but as the data Christopher himself presents show, potency went up rather than down during that time. Second, if one looks at other legal industries that sell psychoactive drugs, they typically do anything they can to give customers as much potency as possible for every dollar they take in.

For example, the British alcohol industry has been warring with the government from at least 1751 (e.g., The Gin Act) to the present day (e.g., minimum unit pricing) over the former’s efforts to churn out cheap, high potency alcoholic beverages. Even if one restricts analysis to the “fine wine” analogy that Christopher evokes, it has to be said that over the past quarter century, the alcohol content of wine sold in the U.S. has crept up from the 8-10% range to 12-14%.

For drug producing industries, higher potency products are a good. They lower transport costs because the product has less mass and weight. Consumers like getting more bang for the buck. Finally, higher potency products are more addictive, and addicted customers are the best customers. I can’t stress strongly enough that all these advantages accrue to both legal and illegal drug sellers. There is nothing in the legalization of drug-producing industries that eliminates these financial incentives and because they can operate openly, such legal sellers actually may be better at achieving a saturation of lower cost, high-potency products in a market than are sellers constrained by prohibition.

People who enjoy the occasional low potency marijuana cigarette tend to assume that sellers have a big economic interest in catering to them. But about 90% of the weed being sold in Colorado is used by people who smoke every day or almost every day. There just isn’t as much money in the “fine wine” set of occasional marijuana users as there is in daily, physically dependent users of high-strength pot.

A legal market thus doesn’t inherently bring us low potency. Rather, it brings us a fight between regulators who generally want low potency products to dominate the market and sellers who generally want high potency products to dominate the market. Sometimes the regulators win, sometimes the industry does. When the industry wins, higher potency products are more available than they are under prohibition. When the regulators win, industries are forced against their economic interests to produce more of the lower potency products that Christopher and I agree pose less risk to consumers.

Cannabis policy for conservatives: the unpalatable vesus the disastrous

Josh Barro’s report in the New York Times starts with the debate about the Washington D.C. grow-and-give system and opens out into the broader question of whether it’s possible to create a system of cannabis controls that:

(1) Allows adult access;

(2) Substantially eliminates the harms from organized illicit business;

(3) Minimizes arrest and incarceration;

(4) Minimizes the increase in heavy use and use by minors.

Those seem to me to be the four key objectives in designing a cannabis policy. (Of course supporters of the current laws don’t regard #1 as an objective at all; tax revenue is relatively unimportant substantively, though a winning argument politically.)

Prohibition is a disaster in terms of illicit markets and law enforcement. Decriminalization might reduce the number of arrests, but wouldn’t reduce the extent of the illicit traffic (currently $40B/yr.) or do much incarceration (about 40,000 behind bars at any one time).  No one seriously proposes mounting the sort of enforcement effort that would be required to shrink the illicit trade back to the level of 20 years ago.

So prohibition is no longer a viable option, and the question is what to do instead. There’s no reason to think that commercialization on the alcohol model will have acceptable results in terms of heavy use and use by minors. “Grow and give” is one among a family of options for non-commercial legalization, alongside state-monopoly retailing, cooperative and other not-for-profit production, or production by public-benefit corporations.

Barro quotes David Frum, an adviser to an anti-legalization group, as praising grow-and-give as an “elegant” approach to finding a middle way, but doubting that it’s sustainable in the face of lobbying pressure. I share that doubt. But if a possibly workable option such as grow-and-give is politically unsustainable, what does that say about trying to hold the line on an increasingly unworkable  prohibition?

If the voters are given a choice between the current system and commercial legalization, it’s increasingly clear that they will choose to treat cannabis the way we treat alcohol. So if, like Frum (and me), you’re worried about the bad   consequences of commercialization, you ought to be working hard at building a political coalition to support some non-commercial option, even if (like Frum but not me) you would really prefer prohibition.

Twenty years ago, the anti-pot forces made the historic blunder of resisting the development of cannabis-based medications, leaving the political field clear for the “medical marijuana” bamboozlement.  Now they’re doubling down on that mistake, resisting non-commercial legalization and paving the way for the very Big Marijuana they most fear.

When the predictable upsurge in problem use arrives, the prohibition forces will have the gloomy consolation of being able to say “I told you so.” But what will console the victims of the avoidable increase in cannabis use disorder, and their parents?

John Kenneth Galbraith once defined politics as “the art of choosing between the disastrous and the unpalatable.” I appreciate that Frum finds the thought of people getting stoned unpalatable. But I hope that he, and his allies, will figure out – before it’s too late, if it isn’t already too late – that the results of allowing cannabis to be pushed the way alcohol is pushed are likely to be disastrous.

Standing athwart history yelling “Stop!” is picturesque, but not productive.

The Marijuana Analyst Class vs. the Marijuana User Class

Pot Use by Education

The above chart was created by Carnegie Mellon University Professor Jonathan Caulkins using data from the National Survey on Drug Use and Health. If you want to read Jon’s intriguing analysis of how the concentration of marijuana use among less educated Americans affects the marijuana market, you can check out one of my recent Washington Post Wonkblog pieces. Here, I want to reflect on a different aspect of these data, namely the orange band in the chart. These college-educated people account for only 1/6 of the marijuana market, but they account for almost all of “marijuana analyst class”.

Who is in the orange bar? Obviously I am, along with every other university professor and think tank scholar who analyzes drug policy. So too are almost all state and federal elected officials and political appointees. Throw in as well almost all journalists who write about or do television stories about pot. Throw in as well almost all health care professionals. Marijuana legalization activists are also heavily drawn from the orange bar, as are DEA and FBI agents. People who consume drug policy analysis on the Internet are disproportionately orange bar members too. What this means practically is that the public debate about marijuana is shaped almost exclusively by people who don’t live and work in the world where 5/6 of the drug is actually consumed. Orange banders can compensate for their insular position somewhat through research (e.g., I spend a lot of time in correctional facilities), but even then they still face significant risk of being out of touch with realities on the ground. Three examples come to mind.

(1) The reality of marijuana possession arrests: The risk of being arrested for marijuana possession in the U.S. is in general extremely low. But when it happens, it is rarely orange banders who get the cuffs slapped on them. Outside the orange band, arrestees usually can’t afford good lawyers who can get marijuana charges reduced or dropped, i.e., the “same” arrest hurts more as one moves down the income scale.

(2) Ill effects of marijuana use: “I smoke marijuana once a week and I am a top software engineer — marijuana is harmless”. I hear this kind of orange band talk all the time. An occasional marijuana user who has an education, a good job and lots of social capital is indeed likely to be fine, but outside the orange band life can be quite different. If you are teenager in a low-income neighborhood on the cusp between just managing to graduate high school or becoming a drop out, the memory, learning and concentration impairments from a daily marijuana habit (a more common use pattern outside the orange band) can profoundly change your life for the worse.

(3) The economics of marijuana: It’s remarkable that when marijuana prices are debated, the default assumption is that the high-priced sinsemilla favored by orange banders is the benchmark. But 80% of the marijuana market is the lower-cost, commercial grade marijuana (mostly from Mexico) popular outside the orange band. Because an ounce of the former can be priced the same as a half pound of the latter, assuming that the orange band product is the norm can lead to wildly overstated estimates of the price of marijuana and the overall income of the marijuana industry.

p.s. Chart notes for policy wonks/data nerds. The time series begins in 2002 because NSDUH was redesigned in 2001, making comparison to prior years inaccurate. Individuals with “some college” should not be equated with “university students”. Only 7% of this group is currently enrolled in a course, and even that 7% includes people in trade schools and community colleges. Finally, these NSDUH data do not include “blunts”, which would add about 6% more total use days each year but only minimally affect the distribution of marijuana use by educational level.

Another Step Forward for Expanding Access to an Overdose Rescue Drug


The photo above is one that brings me great joy. In January of 2011, I returned to my home state of West Virginia, which has the highest opioid overdose death rate in the U.S. Two friends who are fellow health professionals – Senator Dan Foster and House Delegate Don Perdue – had asked if I could suggest legislation that would respond to the epidemic of addiction that was destroying the well-being of the state. I testified in the Senate and the House and also briefed Governor Tomblin. One of my recommendations was to expand access to the overdose rescue drug naloxone. That’s what the Governor signed into law yesterday.

The signing of the bill into law, after variants of it died in four consecutive sessions, is a credit to Governor Tomblin and to dedicated advocates in the State Legislature (Most particularly, the physician/Senator to the Governor’s left, Ron Stollings, who championed this bill yet again this year). It’s also a credit to people around the state, including county officials, police officers and doctors who contacted the legislature to express support. Congratulations as well to West Virginia recovery advocate Sean Hughes, who went public with his compelling story of how his life was once saved by naloxone. All of them make me even prouder of my state than I was already, if that is possible.

About half of the states have now have passed laws to expand access to naloxone, and the future of expansion looks brighter than ever. When I wrote the naloxone access expansion section of President Obama’s first national drug control strategy in 2009 (reversing prior US policy), most of the people I had to win over didn’t know what I was talking about. Today the medication and its possibilities have a much higher profile and level of popular understanding, and with that has come substantially increased interest in access expansion. Expanding access to naloxone can’t stop our overdose epidemic by itself, but it can certainly save lives in every state that is currently considering following West Virginia’s example.

Violent Crime and Imprisonment

Dana Goldstein at the Marshall Project has created a useful interactive graph showing who is in prison and how we might build further on the de-incarceration trend which started five years ago. Goldstein also echoes a point that Mark Kleiman and I have made here many times: It’s a myth that prisons are full of non-violent drug offenders.

The chart below presents Bureau of Justice Statistics data on state prisons, which is where over 90% of U.S. prisoners reside. Violent crime has consistently been the leading cause of imprisonment, and most state prison inmates are serving time for a violent offense. Importantly, the data reflect current controlling offense only and thus understate the proportion of prisoners who engage in violence: Many inmates currently serving time for a non-violent offense have prior convictions for violent crimes.

Corrections in the United States_0442512_2[1]

These data make de-incarceration more complex in at least two ways , which is perhaps why so many people don’t want to believe them.

First, the noble ongoing efforts to reduce the size of the prison population should take substantial care to protect public safety as violent offenders are released. Mass dumping of violent offenders into communities with no monitoring and no services would be dangerous for them, for their families, and for their neighbors. Further, if it leads to released prisoners committing high-profile acts of violence, it could also choke off political support for continued de-incarceration.

Second, even assuming the best of all policy worlds in which reducing incarceration continues to be a priority, the U.S. is probably too violent of a society to ever shrink its prison population to a Western European level. The proportion of the U.S. population that is serving time for violent crimes is larger than the proportion of the Western European population that is serving time for all offenses combined.