Ben Goldacre, Saturday 17 October 2009, The Guardian
For those with the finances to try to silence their critics, this has been a week of spectacular own goals. Trafigura has loudly advertised the report on the dumping of toxic waste in Africa by taking out a super-injunction through Carter-Ruck. And on Wednesday Simon Singh, the science writer being sued by the British Chiropractic Association, won his right to an appeal.Briefly, Singh was sued by the BCA over an article in the Guardian in which he criticised chiropractors for claiming they can treat children’s colic, sleeping and feeding problems, ear infections, asthma and prolonged crying, by manipulation of the spine. He said these interventions were “bogus”, with “not a jot of evidence”. Before this case most people hadn’t really noticed chiropractors. Now the internet is awash with reviews of the evidence and its flaws, so this is a good time to revisit the evidence.
Richard Brown, vice-president of the BCA, writes in the British Medical Journal (the article is open-access) that there is “substantial evidence for the BCA to have made claims that chiropractic can help various childhood conditions”. He provided references to 19 academic papers. These have now been examined and effectively demolished in a response by Edzard Ernst, professor of complementary medicine at the Peninsula medical school, Exeter. I am happy to see that someone else has done this work, and I will now walk you through his summary.
Four of the BCA’s references do not contain data about chiropractic treatment for the conditions being discussed, and so are irrelevant. Eight refer to types of evidence that can tell us little about whether a treatment works: questionnaire surveys, case reports and so on. That’s 12 of the 19 references out already.
Among the rest, we are given a small pilot study, by Mercer and Nook, but it is unpublished, and the short conference summary that is available does not have enough information for anyone to know whether their methods were sound.
We are given a trial by Wiberg et al, in which 50 children with colic got either chiropractic or dimethicone: they found less crying in the chiropractic group, but because the babies and parents could not be blinded – they knew which treatment they were getting – even the researchers themselves felt this improvement might be due to a placebo response, or other non-specific effects unrelated to the chiropractic treatment, such as extra attention.
Hawk et al have a systematic review of various types of spinal manipulation for non-spinal conditions and look at the treatment of asthma, otitis and colic. Systematic reviews examine trials and summarise them. They give a positive conclusion for the treatment of asthma, but this relies on a study about osteopaths manipulating the ribs, so that’s not relevant.
They are also positive about colic, to be fair, but, for evidence, they rely on the flawed and unblinded study by Wiberg, described in the previous paragraph. This review therefore adds nothing.
There is a Cochrane review looking at various treatments for bed-wetting, and Cochrane reviews are high-quality summaries of the evidence, as you will know. This one found two trials of chiropractic, which were not of high quality – in fact, the authors described them as “weak evidence”. So that’s not good.
Browning did a trial – which was published after Singh’s article – comparing spinal manipulation with occipitosacral decompression. Both are dubious treatments, and the trial found no difference between them. So both may be effective, or both may just be equally ineffective; either way, no prize.
But most interesting are the studies which the BCA chose not to mention: three randomised controlled trials and two systematic reviews, arguably the strongest evidence, were omitted. The BCA is aware of these studies, because it has specifically commented on other work based on them. So why did they not explicitly reference them? I don’t know. But they are negative papers, which failed to show that chiropractic was effective; while quoting irrelevant papers, questionnaires, case studies and weak trials, the BCA ignored these higher-quality studies, with their unflattering results.
There are huge, endless debates to be had on our libel laws, on the risks they pose to the public by stifling access to information, and on the changes that could be made. But, for today, know this: there is no good evidence that chiropractic is effective for the conditions claimed by the BCA, for the reasons you now know. Shout it from the highest rooftop, when you tell your friends about Trafigura, because, until the law changes, the strongest disincentive to this effect is a very close examination of the companies involved.