There’s an interesting new study in the Lancet from Professor David Nutt (erstwhile chief UK drugs advisor prior to being sacked last year for disagreeing with the previous administration), which has been publicised as stating that “alcohol is more harmful than heroin”, as evidenced by the pretty-looking graph published across all the usual media outlets:

The detail behind the headline is rather more complex, and illustrates neatly how the general public and the mass media simply don’t understand how to interpret information.
The key thing to understand is how they got to the end point of this graph; what was the methodology used?
The study aims quantify the harm posed by various drug on both users and on wider society (eg through criminal behaviour to acquire funds to access the drugs), and so place each drug on the same unified scale. The methodology they used to do this is Multi-Criteria Decision Analysis, something I touched upon in a previous blog entry as being a very powerful method of problem-solving.
In order to perform MCDA, you need to have a panel of people making judgements on a series of pre-determined scenarios. The make-up of the panel therefore is the most important determinant of the results generated. In this case, the panel were the members of the government’s Independent Scientific Committee on Drugs, plus another 2 invited experts.
That makes a total of 25 culturally-homogenous scientists.
However expert these members are in the scientific literature, I would strongly question whether 25 people from a narrow professional group constitutes a broad enough panel to make decisions about overall harm of each drug on society as their definition of harm may not be that which a wider panel might use. Furthermore, since all the members of the panel work together regularly, they are likely to have all developed broadly consistent opinions with each other over time, something which will have been regularly reinforced through mutual agreement.
This “crowd convergence effect” or “groupthink” means that in effect, the panel contains far fewer than 25 different opinions, more likely only 2-3 mildly competing views at most. For all practical purposes, this study is therefore based on two or three people’s personal (if unusually well-informed) opinion. That strikes me as methodological thin ice on which to base public policy.
And that’s without the further problem of considering that the availability and stigma of each drug on the list varies wildly, which complicates ability to determine the actual underlying health and societal effects. For instance, because heroin is illegal, to acquire it automatically requires criminal activity. And because alcohol is legal, it is very easily available , multiplying its effect on society disproportionately. This means that making judgements around the harm is the very essence of a wicked problem, where cause and effect are intertwined to such as extent as to make disentanglement impossible.
In short, while the study has a useful role to play in terms of opening up the discussion for public, people should be very wary of assigning too much power to its findings.
Also in the News…
On a completely different note, below is a short collection of links to some other interesting news stories that grabbed my eye over the last few days, and that I haven’t had the time or inclination to make a full blog entry on:
Like this:
Be the first to like this post.