We may be "the most awesomest druggies"...but this paper doesn't show that.

Jun 30 2008 Published by under Drug Abuse Science, Public Health

Greg and Steve (not Adam and Eve) both blogged a recent paper by Degenhardt and (20 other) colleagues in PLoS Medicine.
Steve:

According to a new survey the USA has highest level of illegal cocaine and cannabis use in the world. Thank goodness the War for Drugs is working so well! Ohh... wait... that's the war ON drugs and it's supposed to protect us from ourselves and our nasty drug habits.

Greg asks:

Does the observed age difference (younger cohorts with more drug use) reflect a reporting bias or a reality? It seems that over the last several decades the evidence that younger people are using more drugs is so often reported that all people must be using all drugs by now, but they aren't! Do studies that show declines in drug use get less press, or go unfinished? (Is there a reporting bias or a confirmation bias at work here?)


The full citation of the paper is here:

Toward a Global View of Alcohol, Tobacco, Cannabis, and Cocaine Use: Findings from the WHO World Mental Health Surveys Degenhardt L, Chiu WT, Sampson N, Kessler RC, Anthony JC, et al. PLoS Medicine Vol. 5, No. 7, e141 doi:10.1371/journal.pmed.0050141

Now, this is interesting and all as a rough international comparison. I do like reading drug epidemiology papers. And I happened to see the first author [her National Drug & Alcohol Research Centre site nas some interesting reading] present some of their approach to creating these WHO focused international comparisons of risk...in that case she was talking about how they are trying to assess the costs (health, productivity, longevity, etc) associated with cannabis use for the Global Burden of Disease estimates mentioned in the paper. I have some impression of the staggering difficulty of making international comparisons using epidemiological data that were collected using different methodologies. I'm not here to critique their selection of lifetime-use as a measure.
But one simple thing that Steve and Greg both seem to overlook is that the basic data point is lifetime use. As in "Have you ingested cannabis ever in your lifetime?". And this seriously limits the conclusions that may be drawn vis a vis more serious or involved questions of drug use and drug policy. Steve then goes totally off the rails in terms of conclusions that may be drawn from limited data.
Now Greg's post can be excused because it is quite circumscribed. OTOH, Steve's goes hyperbolic in a tone that jumps from these data limited to lifetime use to questions of public policy that are best focused on more significant use patterns such as those that meet criteria for abuse and dependence. And after that, well....:

According to a new survey the USA has highest level of illegal cocaine and cannabis use in the world. Thank goodness the War for Drugs is working so well! Ohh... wait... that's the war ON drugs and it's supposed to protect us from ourselves and our nasty drug habits.
-snip-
And finally... the most important take home message: Drug Policy has NOTHING to do with drug use. We can put users, dealers, producers, cousins of uncles of friends of users, their dogs, cats, guinea pigs, or the lint under their couches in jail or to death and it will have no effect on whether people use drugs or not. Education however... just maybe that will help. Oh.. that and taxation - Heavy heavy taxation. After all tobacco use has gone drastically down in the last decade with more education and higher drug prices.

This publication and these data do not even remotely give us enough information to make such statements. The alternative hypotheses are obvious enough to make such statements ludicrously naive (or agenda driven, your call). At best we can conclude that Drug Policy (if by that we mean overall strategies of legalization/crimininalization, enforcement, interdiction) has only some influence on drug use (and again let us keep firmly in mind what breadth of "uses" we're talking about- from smoked a joint once four decades ago to this). There are plenty more influences that might be expected to vary from country to country including drug cost/ personal income, severity of legal penalties, education, the type of social influence and fertility of the ground (i.e., for religious-based prohibitions), general personal freedom, success of interdiction, proximity to sources,...... The list goes on and on.
Given the priority of personal liberty in the US, minimal legal penalties for, say, single joint possession, lack of a strong anti-drug use thread in the most-popular religious traditions and many other factors...how does Steve know that our drug use would not in fact be higher than it already is, save for our "Drug Policy"?
I will encourage Readers who are interested to start with the Monitoring the Future publications (overview, vol I, vol II) for a very easy look at how lifetime, annual and past-month drug use statistics relate.

8 responses so far

  • Dunc says:

    This publication and these data do not even remotely give us enough information to make such statements. [...] At best we can conclude that Drug Policy [...] has only some influence on drug use

    Hmmm.... I'd say that this data does not allow us to make any definitive statements as to whether policy has an impact or not. Still, it is generally believed that the burden of proof rests with the person making the positive claim, which in this case would be the claim that our current policies have a beneficial effect. It seems to me that the hypothesis that drug policy has no appreciable impact on patterns of drug use is the null hypothesis.
    It is, of course, possible that drug use would be higher without the current policies. However, I am not aware of any evidence that this is the case. Furthermore, I believe there is evidence that increasing the severity of the law-enforcement side of the policy does not help (assuming you define "help" as "reduce the incidence of drug-related harms"), and it doesn't rest on problematic cross-country comparisons - we can examine changes in problematic drug use over time within the same locality and see how (or if) it correlates with changes in policy. I hate to dredge up the Prohibition example again, but it certainly does seem relevant here...

  • Greg Laden says:

    This is definitely a case where anyone interested in this needs to read the original paper. The analysis has a number of interesting dimensions but also a lot of limitations.
    Also, I'm convinced that that there is a large shift in reporting bias from younger vs. older individuals in the US, and that this bias may be utterly different elsewhere.

  • DrugMonkey says:

    Also, I'm convinced that that there is a large shift in reporting bias from younger vs. older individuals in the US, and that this bias may be utterly different elsewhere.
    Since you are talking about the US, I'll direct you to the Monitoring the Future data, links above. See Volume II.
    First take the 30-day, annual and lifetime use figures across age groups- Figures 4-3 p100 for Marijuana and 4-7 p104 for Cocaine.
    Then look at prevalence of past-year use in Figures 5-3a p152 for Marijuana and 5-9 p160 for Cocaine.
    does this convince you that there is a logical relationship between prevalence rates reported at a given fixed year and the present lifetime estimates for given age cohorts?
    The US cohort effects are not huge but if you look at the 35 yr olds vs 45 yr olds in 2006, these represent some of the biggest cohort effects. track back to when these groups were 20 yrs old and you will see that this maps nicely onto an interval in which annual prevalence changed significantly (i.e., from the early 80s to the early 90s).
    So I don't know that we have reporting bias here in the US epidemiological numbers (for MtF datasets, of course, it is the survey with which I am most familiar). I should mention that the MtF press releases and executive summary type comments are plenty forthright about trumpeting declines in drug use from year to year (even when I look at it and think it looks like the usual year-to-year wobble for a given drug, frankly).
    Getting back to the more impressive looking age cohort effects in Fig 2 of the PLoS Med paper under discussion, since it is the entire dataset (I think?) well sure this representation could be concealing interesting differences between countries. I don't see where this is evidence of reporting bias though...

  • DrugMonkey says:

    oh, let's drag up Prohibition Dunc. Because this site and the included figure seems to imply that good data on consumption during Prohibition are not available. Also that the immediately surrounding numbers suggest alcohol consumption per capita declined sharply during the presumed political hoopla/activism leading up to Prohibition and then only increased gradually after Prohibition (which one of the links below attributes to a series of economic measures post-repeal).
    also see:
    http://www.drugtext.org/library/articles/craig102.htm
    http://encarta.msn.com/encyclopedia_761564677_2/prohibition.html ("no striking evidence of a crime wave")
    http://www.cedro-uva.org/lib/levine.alcohol.html
    So the Prohibition argument seems to be that you can get a whopping reduction in recreational drug use by draconian policy. This argues that perhaps it just has to be a really oppressive approach and/or be tied to whatever other societal influences led up to the political will for Prohibition. But don't try to argue from this that prohibitory policies are ineffectual.
    jumping back to the PLoS paper, the comparisons across countries are imperfect because I don't see many countries on there with truly repressive and punitive legal consequences for drug use, although by some accounts Japan might qualify.
    It all adds up to a slightly different interpretation. You seem to be saying that US drug policy pretty repressive and it doesn't clearly work. A different analysis might suggest that in fact US drug policy is not all that repressive and that it might work better if it really posed a threat to casual users, was successful in interdicting the supply, etc.
    To return to my main point, these data in the paper really do not support over excited policy conclusions. One can attempt to use the paper this way but all one does is repeat one's pre-existing positions and attempt to paste the paper onto those biases. It isn't convincing.

  • Dunc says:

    Guess I may need to reconsider that argument then... Every day's a school day here. Although I notice that you're referring to "a whopping reduction in recreational drug use" whilst I said the goal should be to "reduce the incidence of drug-related harms". I believe our previous discussions have indicated I do not subscribe to the belief that recreational drug use is necessarily problematic in and of itself.
    I also notice that your first links states that "it is true that prohibition provided a major boost for organized crime", whilst your third mentions that "[a]dulterated and even poisonous alcohol was sold". Now, I can find some studies and data for cirrhosis, but nothing (epidemiological, anyway - there are case studies) for methanol or ethyl glycol poisoning, which would be the two I would expect to see from back-street spirits.
    The third link also states that "except for the first three years, the per capita consumption of alcohol has been greater under prohibition than during the war period" (contemporary quote attributed to Warburton) and "[w]hatever public health benefits prohibition achieved in terms of reducing consumption, alcohol regulation in the 1930s and early 1940s accomplished them as well. Further, this occurred despite the fact that the post-prohibition regulatory system had little or no public health focus", whilst "other nations achieved even greater reductions in per capita consumption than the United States -- without the negative consequences of prohibition".

    So the Prohibition argument seems to be that you can get a whopping reduction in recreational drug use by draconian policy.

    Except that, as you just stated yourself, that whopping reduction started well before the policy started and continued after it ended. Given that, I find it difficult to understand how you conclude it was the policy that made the difference.

    But don't try to argue from this that prohibitory policies are ineffectual.

    I am not arguing that prohibitory policies are ineffectual - I am arguing that there is insufficient evidence of their effectiveness to conclude that they are effective. And we still seem to have a slight disagreement as to what "effective" actually means in this context.

    jumping back to the PLoS paper, the comparisons across countries are imperfect because I don't see many countries on there with truly repressive and punitive legal consequences for drug use

    Cross-country comparisons are also imperfect because it's very difficult to control for the differences in social and cultural context (which I strongly suspect play a far greater role in determining the extent and nature of drug-use in a society than the legal framework).

    To return to my main point, these data in the paper really do not support over excited policy conclusions. One can attempt to use the paper this way but all one does is repeat one's pre-existing positions and attempt to paste the paper onto those biases.

    My complete agreement with this point was part of my first comment. In fact, I went so far as to say that this paper doesn't support any policy conclusions at all, and made absolutely no reference to it whatsoever in my off-hand comment on Prohibition.
    Anyway, I do not particularly want to turn this into an extended debate on the rights and wrongs of the Prohibition era. Sorry I mentioned it.

  • dude says:

    I am going to address the argument above about the effect of a draconian drug policies. THEY DONT WORK PERIOD. here's why though. Go to jail, get a felony drug charge. You cant get jobs because you have to put that you have a felony on your record. What do you do to make money, DEAL DRUGS. Also, hard drugs are easier to smuggle than non hard drugs. Meaning if you have someone in jail for weed who experiments with hard drugs and all they can do in prison is addictive hard drugs, they come out of prison with a worse drug problem than when they went in. Also they are already preconditioned to use drugs as an escape as they go to prison for drugs and prison sucks so it makes sense to use drugs in prison to escape. Coming out of prison with a nasty habit and few job prospects, what do you do, deal drugs. This is why the drug war has created a cycle of utter failure. It is creating more dealers than exist naturally. Send people to rehab instead and make sure there's no felony on there record and you can change lives and prospects. The current system encourages drug dealing economically in that people get shut out of the mainstream economy so they turn to the underground economy. The drug war has failed and that is a statistically proven fact. But aside from studies think of it from a common sense perspective, talk to people involved in this horrendous cycle, and look at the situation on the street and you will realize that policy must change or the problem will get worse period. Apologists for the drug war are by definition ignorant of reality and terrible at thinking of the consequences of the system--i.e. what do you do when you cant get a job, make money from crime idiots. Also the drug wars strategy of attacking supply raises the price of drugs because there isnt the volume to meet demand making it all the more appealing to follow this path. If you think the drug war is good or has a positive effect on drug use YOU ARE AN IDIOT and need to learn some critical thinking skills and a dose of reality.
    I doubt there is much of a confirmational bias. Americans use a lot of drugs period. Talk to druggies that travel, Americans party harder when it comes to drugs, and some of these characters have shot down Irish and Germans when it comes to both drinking and drug tolerance. Screw the stereotypes when it comes to drinking or drugs in other cultures, the american hard user can party all them to the floor or ER. An american hardcore alcoholic can generally destroy other countries alcoholics. Sorry if this offends people in other countries but americans manage to be a little bit more dysfunctional in this regard.

  • adude says:

    So there was the comment about how do people not know how without our policies drug abuse wouldnt be worse. The answer is the economics STUPID. When people cant get jobs because of convictions you reduce opportunity, which results in crime to make money and drug abuse to escape reality. Our system actually promotes more drug abuse, as oppose to a small reduction of it which would be worse without the policy. Sorry author, but steve is right and there are methodologies that can be used to prove his point, your are just too wrapped up in statistical studies, epidemiological data, etc. instead of looking at the economic side. Drug use is a consumption and production issue and is also economic. There are the economic tools of analysis to prove that the policy is actually counter productive. Your concerns are with the wrong datasets. Look at how much money is made and spent on drugs, and you will have your answer. Don't rely on statistics about who does or doesnt use (people may not answer honestly, not enough people may be samples) instead FOLLOW THE MONEY.

  • Jim says:

    For a policy analyst, there may be six especially interesting statistics in the PLoS article, all based upon essentially the same social survey methods as described in prior "ESEMED" reports on this set of studies in Europe.
    *
    If you are interested, go to the PLoS article, and take a look at Table 3 and inspect the Netherlands, Belgium, and Germany values for the cumulative incidence proportions for young adults age 22-29 years old in each country. These epidemiological estimates, all subject to well-known survey participation and reporting biases, are from three neighboring countries in which the biases most likely operate pretty much at the same levels and in the same degree. The analysis for Table 3 is restricted to 22-29 year olds so that issues of forgetting, generation/cohort differences, etc. are constrained.
    *
    By age 21, an estimated 35% (+/-7%) of the 22-29 year olds in Netherlands had started to smoke cannabis. (Note that all of the 22-29 year olds in the survey had already passed entirely through the interval of risk of starting to smoke cannabis by age 21.) This is a country where national policy is in conformity with a US-enforced UN psychotropic drug convention. But, some 10-20 years back, in a rejection of the international convention, individual communities were allowed to try to segregate the cannabis market from the markets for cocaine and other drugs, via local cannabis cafes and shops allowing on-premise and (to some extent) off-premise consumption of small amounts o cannabis.
    *
    Now, let's turn to Table 3's corresponding estimate for Belgium, which some years back implemented a policy that the 1972 US President's Commission on Marihuana and Drug Use ("Shafer Commission") called 'partial prohibition' in that people are allowed to grow small amounts of cannabis for personal consumption and to smoke it in the home, but not publicly. Whereas in the Netherlands, about 1 in 3 young adults age 22-29 had started to smoke cannabis by age 21, the corresponding value is for Belgium's 22-29 year olds is 22% (also +/-7%), about 1 in 5, with a partial prohibition policy in place while these young adults were growing up.
    *
    Now let's turn to Germany where until recently the policy in effect was in conformity with the US policy of total prohibition, with some jurisdiction to jurisdiction variation of the type we see in the US (e.g., compare police action against cannabis in San Francisco, CA, or Eugene Oregon and Ann Arbor Michigan or Madison Wisconsin versus other places). Back before many of Germany's communities moved in the direction of a partial prohibition policy, what was the result on 22-29 year olds? An estimated 41% (+/-5%) of the 22-29 year olds had started to smoke cannabis (about 2 in 5).
    *
    Is this evidence of adverse effects of the Netherlands and Belgium policies in place while these 22-29 year olds were growing up?
    *
    One suspects otherwise. If anything, there is no variation across these three countries in the recent experience of 22-29 year olds. If there is variation across the countries, then the evidence tends to support the idea of a greater risk of onset of cannabis smoking before age 22 among citizens of Germany (with the most US-like policies in place while they were teenagers), and lower risk of cannabis onset in the Netherlands and Belgium (with 'partial prohibition' and limited availability policies).
    *
    So what about the segregation of the cannabis market from the cocaine market?
    *
    Let's look at the corresponding values for 22-29 year olds in relation to onset of cocaine use by age 21. Here's what the study showed: in Netherlands, with arguably the most liberal cannabis policies in the world, and with a deliberate policy effort to segregate the cannabis and cocaine markets, the estimated proportion having tried cocaine by age 21 was one percent (+/- 0.6%), about 1 in 100. In Belgium, it was found to be 0.6 percent (=/- 0.4%, very imprecise statistically, but still close to 0), closer to 0 or 1 in 100. In Germany the same proportion is 6% (+/-2% or 3%), some six times greater if we pay attention to the ratios of the point estimates, about 1 cocaine onset for each 16-17 young people in Germany.
    *
    Combine the Netherlands and Belgium values to stabilize the precision of the estimates, and you get a clear sense that some of these communities might have achieved a bit of success in the effort to segregate the cannabis market from the cocaine market. Hint: it wasn't the folks in Germany.
    *
    Now I can shoot holes in this argument, but I like to set it up so that the evidence is gathered in an even-handed fashion, and then the evidence is evaluated in an even-handed fashion. We can make an argument that the 22-29 year old people of Germany are more likely to report their illegal drug use accurately, as compared to 22-29 year olds in the Netherlands or Belgium, but I'd be inclined to be suspicious about this kind of argument, and my first guess is that under-reporting is more common when the legal consequences are more serious (i.e., in Germany). I could go on and on, in relation to the study methods and the possible confounding factors, etc. etc., as well as anyone can.
    *
    But I'm going to stop here with a note about some concerns expressed by my colleagues in the low countries. They are not at all sure it's a good idea to allow cannabis shops to operate in the vicinity of primary and secondary schools. They have some concerns about cannabis smoking among underage kids, perhaps more dope-smoking by teenagers than would have been true if the cannabis shops were not allowed to operate (although these PLoS statistics might serve to suggest something else). Some of them worry about cannabis smoking and precipitation of psychoses among kids with pre-existing vulnerabilities to schizophrenia. Some of them worry about cannabis causing car crashes that otherwise would not occur. But none of them have told me that they think the clock should be turned back to the days when cannabis shops didn't exist and when the cannabis dealers and cocaine dealers all were essentially one and the same.
    *
    I still think there might be an important lesson in these numbers, with a contrast that no one ever has seen before.
    *
    But then I have an obvious bias, and it won't take Sherlock Holmes to figure out what that bias might be. It's not a bias about conclusions from the evidence, that's for sure. I pretty much let the chips fall as they may when it comes to the evidence.
    *
    My main bias can be summarized by saying that we don't have experimental data from a proper randomized trial to compare the responses of young people under deliberately varied policy regimes. That's the kind of evidence I prefer. But in the policy analysis arena, we generally have to make do with what we can learn from observational studies; it's hard to randomize policy instruments of this type. This situation makes me very cautious about policy pronouncements, so I won't make any policy pronouncement of any definitive character. I'll repeat what was said by a National Academy of Sciences and National Research Council task panel a few years back: "America's Drug Policy: What We Don't Know Keeps Hurting Us" (available free online at http://www.nap.edu/openbook.php?isbn=0309072735
    *
    I also will refer the interested reader to the Shafer Commission report and its recommendation of a 'partial prohibition' approach (akin to the current Belgium approach) way back in 1972, and I'll wonder why we spend so much money on cannabis law enforcement and incarceration in this country (USA) and relatively little money on prevention and treatment of the folks who get into trouble when they start smoking dope (and yes these folks do exist).
    *
    Here's the link to the Shafer Commission report from the early 1970s, more ignored than honored:
    http://www.druglibrary.org/schaffer/Library/studies/nc/ncmenu.htm
    *
    In the land of the free and the home of the brave, it might be wise to solve these problems locally, and to have the federal government step back from its curiously draconian federal approach when there is no evidence that supports the federal pre-emption of local rule on this front. It will be interesting to see what will happen if and when the Native American communities decide to allow legal consumption of cannabis and peyote or mescaline within the real estate boundaries now occupied by their gambling casinos. One suspects that there might be no federal pre-emption or jurisdiction in that context, just as there is neither state nor federal authority to block tribal entrepreneurial activity in relation to gambling revenues. We shall see.
    *
    Goodcheer!
    J

Leave a Reply