I mentioned, when I wrote challenging the things that ‘everybody knows’, I had come across a new publication from the United Nations Office on Drugs and Crime (UNODC) that seemed to have been more leaked that published though I’m not sure if ‘leak’ is the right word for an organisation placing information on its own website without telling anyone it is there.
I said I would take more of a look at ‘Cannabis: A Short Review’1 and I’ve been trying to but it really is a pretty dispiriting read. It is almost as if Ben Goldacre never wrote ‘Bad Science’. In fact, if you’ve only recently read the book I recommend this UNODC discussion paper as a useful exercise to check out real life examples of the misdirection techniques discussed by Dr. Goldacre.
I’m going to willingly expose myself to the accusation of cherry-picking because I don’t expect to be able to engage your attention while I go line by line through this 29 page document with its 129 references so you can believe me or not when I say that I beginning by writing about the first two points that I selected at random and not discarding any that don’t make my case.
I’ll come back to the issue of cherry-picking but I want to start with the business of giving a reference that turns out not to fully support the claim you have made for it. On the effects of smoking Cannabis sativa, marijuana, the UNODC review says;
‘Animal lungs exposed to cannabis smoke developed abnormal cell growth and accelerated malignant transformation, to a greater extent than those exposed to tobacco. 45’
The point of giving a reference in a paper is not to enable people to find more detailed information – most readers are far too busy for that. Some will take the time to read the full citation within the document and a very, very few will seek out the referenced work. The majority, however, will see that a claim has a superscript number attached to it and accept that as evidence for the claim. That is not necessarily so.
In this case reference 45 is ‘NIDA, Research Report Series: Cannabis Abuse, 2010’. This is a document that NIDA (the USA’s National Institute on Drug Abuse, part of the National Institutes of Health, itself part of the US Department of Health and Human Services) revises from time to time and the latest version is from July 2012.2
Far from supporting the implication, created by use of terms like ‘abnormal cell growth’ and ‘malignant transformation’, that smoking marijuana is more likely to cause cancer than smoking tobacco, the NIDA document says;
‘However, while several lines of evidence have suggested that marijuana use may lead to lung cancer, the supporting evidence is inconclusive.’
‘Thus, the combined evidence from animal studies plus the limited human data available seem to warrant additional research on the impact of marijuana on the immune system.’
So, far from supporting the statement made by UNODC, this reference demonstrates that the claim was, at least, over-simplified and, more probably, unjustified.
Turning now to the matter of cherry-picking the UNODC says;
‘Examining drug-tested federal employees, researchers found that those who tested positive for cannabis on a pre-employment urinalysis test had 55 percent more industrial accidents, 85 percent more injuries, and a 75-percent increase in absenteeism compared with those who tested negative for cannabis use.61’
Reference 61 turns out to be exactly the same as reference 45.
What the NIDA document says is;
‘For example, a study among postal workers found that employees who tested positive for marijuana on a pre-employment urine drug test had 55 percent more industrial accidents, 85 percent more injuries, and a 75-percent increase in absenteeism compared with those who tested negative for marijuana use.’
Although NIDA gives references (18 in a 12 page document) it doesn’t offer a reference for that statement.
That could be because the paper that cites those numbers3 concludes;
‘Although these are unfavorable outcomes, the effects of drug use on job-related factors do not seem to be dramatic. It seems likely that claims of the devastating effects of drugs in the workplace used to justify preemployment drug testing are exaggerated.’
Of course, if it were interested in presenting an unbiased view the UNODC might have referenced ‘An evaluation of preemployment drug testing’4 that concludes;
‘No significant associations were detected between drug-test results and measures of injury and accident occurrence.’
There is another example of cherry-picking and I confess I went looking for this one because I noticed mention of Portugal.
Whilst the UNODC mentions and debunks the report from ‘the libertarian Cato Institute’5 and praises what the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) has to say on the subject6, it makes no mention of the Hughes and Stevens paper7 generally accepted as the most objective study of the topic to date. (See previous blog entry)
Because I said I didn’t want to go on so long that I lose your attention I’ll just briefly mention that the document trots out the discredited COMT research and leave you to decide whether to read my previous piece on the problem of people citing research that supports their case and hoping no-one notices that it has been disproved.
So this new UNODC document adds nothing of value to the consideration of this subject. That should be no surprise since right at the end UNODC acknowledges the ‘valuable contribution’ of Kevin Sabet.
The most dispiriting part of it all is my usual reaction. If cannabis is really as harmful as this document would have you believe how can there be any moral justification for leaving its manufacture, promotion and distribution in the hands of criminals?
How can it be right that the only piece of paper required to get access to this substance by anyone, of any age, is a banknote of the appropriate denomination of their country’s currency?
Cannabis: A Short Review UNODC 28th August 2012
2. Research Report Series: Cannabis Abuse NIDA July 2012
3. Costs and Benefits of Preemployment Drug Screening. Zwerling, Ryan and Orav JAMA 267(1): 91-3 (1992). (Abstract on readabstracts.com)
4. An evaluation of preemployment drug testing Normand J, Salyards SD, Mahoney JJ. J Appl Psychol. 1990 Dec;75(6):629-39.
5. Drug Decriminalization in Portugal: Lessons for Creating Fair and Successful Drug Policies. Greenwald, G. Cato Institute 2009.
6. Drug Policy Profiles – Portugal European Monitoring Centre for Drugs and Drug Addiction (EMCDDA)
7. A resounding success or a disastrous failure: Re-examining the interpretation of evidence on the Portuguese decriminalisation of illicit drugs Hughes & Stevens Drug and Alcohol Review Volume 31, Issue 1, pages 101–113, January 2012
You can send comments via the contact page but please be sure to say what blog entry you are commenting on.