The art and science of evidence about drugs
It is very difficult to study the relative harms of individual drugs – and there are lots of vested interests
The government claims to want to pursue "evidence-based policy". I've often wondered whether there is any evidence that evidence-based policy is better than its alternative. What alternative? Daily Mail-based policy, of course.
The issue has been in the news lately because of the Nutt affair: the sacking of the chair of the Advisory Council on the Misuse of Drugs, who has been arguing that government policy on drug classification runs foul of the scientific evidence.
I've been following the work of Professor David Nutt even before the newspapers took interest, since sitting on a working party on drug regulation for the Academy of Medical Sciences. As a newcomer, I wanted to find the "facts on drug harm", and read Nutt's A tale of two Es – ecstasy and ethanol (alcohol) – published in 2006. According to the statistics in this paper, there is not a single dimension on which ecstasy is known to be more harmful than alcohol. Nothing. Alcohol causes death in large numbers, brain and liver damage, violence and public disorder. Ecstasy helps white men dance and makes them hug each other. Tragically, it also leads to some deaths, but then, notoriously, so do horse riding and peanuts.
Nutt, of course, has his academic critics. One of them, Professor Andy Parrott, wrote in a letter to the Evening Standard on 3 November: "All recreational drugs cause more harm than benefits … ecstasy [leads to] depression, memory loss and impaired immunocompetence (more coughs and colds)". I see, so that's why you can go to prison for possessing it. It gives you sniffles, just like not drying your hair properly after your swimming lesson.
A lot is known about the harms of alcohol. A lot is known about the harms of drug overdoses and infections. But very little seems known about the health effects of taking a regular, clean dose of many drugs.
It is a very hard area to study. Randomised controlled trials won't pass the ethics committee. Observational studies are little help. Most drug users use several different drugs, including alcohol, so harm cannot be attributed to a particular source. Longitudinal research would be helpful, but addicts, funnily enough, tend to drop out of the research cohort. Drug researchers end up studying only those users they can catch: the ones who turn up in clinics with a crisis. Something of a skewed sample, of course.
Animal experiments are an alternative. There is some great work looking at the types of webs spiders spin under the influence. More relevantly, a paper from 2003 in the leading journal Science claimed that primates injected with a recreational dose of ecstasy developed brain damage. Very worrying. But a few months later, when they couldn't replicate the experiment, the researchers published a retraction, saying that the primates were given crystal meth by mistake, which has a very similar chemical name and was delivered to the lab on the same day. Whoops.
For many drugs there seems little evidence of physical harm. But that shouldn't be confused with the very different claim that there is evidence of little physical harm. So how can Nutt be so confident about relative harms? Isn't there a famous bar chart, showing that heroin and cocaine are the most harmful drugs, with alcohol and tobacco more dangerous than cannabis, ecstasy and LSD? Well, yes there is, and it was published in The Lancet. But because so little hard evidence is available, the methodology was "Delphic": ask the oracle, in this case a collection of drugs experts. In fact, the table made its first appearance in an appendix to a Commons select committee report, where the initiative to calculate drugs harms was welcomed, but the committee noted that this particular attempt "is almost as much an art as a science".
So we have the spectacle of scientists over-claiming the quality of their results to publicise their research, journalists whipping up a moral panic to sell newspapers, politicians doing whatever they think will keep them in power, and the public looking on in semi-horror, as one would on a reality TV show spinning out of control. So, to paraphrase the great Jonathan Aitken, if it falls to me to start a fight to cut out the cancer of bent and twisted science, politics and journalism in our country with the simple sword of truth and the trusty shield of British fair play … forget it. Just pour me a glass of wine, and turn on The X Factor, please.
The Guardian, Tuesday 1 December 2009
• Jonathan Wolff is professor of philosophy at University College London. His column appears monthly