My Back Hurts - Poke Me with Toothpicks!

thelondonpaper, the only newspaper I read regularly, and one which I am generally willing to defend as being relatively thoughtful and intelligent, especially for a free publication, has joined the bandwagon for proclaiming that acupuncture works based on a recent study.

Apart from a few commendable nods toward scepticism, and a fence sitting conclusion that "the jury's still out", the overall tone of the article was, in my opinion, weighted in favour of the position that acupuncture is at least worth a try. In fact it ends with an entirely useless "case study":

There was only one thing for it. We sent our promotions manager Sarah Cox to an acupuncturist to relieve her chronic back pain...

Which, in turn, concludes:

My scepticism gave way to belief

My back and shoulders seemed looser. The dull ache had eased. I’ll definitely do another session.

Don't get me wrong, this was by no means the worst reporting of the Cherken et al. study (I still love ya, thelondonpaper, with all your wacky joined up lowercase letters!), but it would still give the impression to someone unaware of the science that it is worth shelling out money ("A one-hour session with Steve Kippax costs £80", the article helpfully informs us) for a treatment with no really compelling evidence.

Perhaps I should explain what this study actually said and how this, and most other papers, got it wrong.

The study randomised 638 adults with lower back pain into 4 groups: individualised acupuncture (an acupuncturist chooses the pin points based on a consultation), standardized acupuncture (pin points are chosen from text book positions corresponding to the teachings of traditional chinese medicine), simulated acupuncture (a placebo which uses non-penetrating toothpicks, but looks and feels the same to the patient and practitioner), or no acupuncture. All subjects continued to receive any standard care as well as the treatment associated with their group.

Placebo controlled trials like this are designed to test one thing: does this modality work better than placebo. If it does, it is good evidence that it is a worthwhile treatment with a real medical effect. If it performs as well as placebo, then its effects can be put down to those of placebo itself, and the null hypothesis can be accepted.

In this case, the two acupuncture groups scored remarkably similarly to the placebo acupuncture group. All of these performed better than the group that received no acupuncture, as would be expected.

What this shows is that acupuncture, individualized or standardised, performs no better than placebo, and this study can therefore be seen as evidence against its efficacy.

However, this is not the conclusion that the authors arrived at when they stated:

Although acupuncture was found effective for chronic low back pain,
tailoring needling sites to each patient and penetration of the skin appear to be unimportant in eliciting therapeutic benefits.

"Penetration of the skin appear(s) to be unimportant"!? In acupuncture!? Acupuncture - the therapy entirely charactrised by penetration of the skin with needles?

What they're saying is, because placebo acupuncture (poking with toothpicks, need I remind you) is as good as actual acupuncture, this must have therapeutic effect too!

This is exactly like trialling a new heart disease pill against a placebo sugar pill and then, on finding that they offer the same level of benefit, stating that sugar must actually have clinically useful effects for heart disease aswell!

This is not the proper conclusion of this trial! The proper conclusion is that acupuncture performs no better than placebo, and therefore can be assumed to be offering nothing but placebo benefit.

This doesn't mean that the placebo effect can't be quite impressive. It certainly can be. The placebo effect is an incredibly interesting, subtle and potentially powerful effect. The more invasive and extravagent the treatment, the greater the placebo effect. And acupuncture is up there for invasiveness and extravagance. And it's quite expensive, which always helps.

That's all this study shows that acupuncture is: placebo. To claim anything else of it is to speak above the level of evidence given.

So, this is the letter I've sent to thelondonpaper's letters page:

I'd like to point out the recent study you mentioned yesterday does not show that acupuncture "is good for your back". The study was designed to test whether there was any benefit of real acupuncture over placebo acupuncture (using not-penetrative tooth-picks, unbeknownst to the patient). It found that there was not. Any benefit observed owing to the acupuncture conditions was entirely explainable by the placebo effect of a novel and theatrically invasive intervention. To claim that because receiving acupuncture fared as well as being poked with tooth picks shows that both have clinical benefit is a misinterpretation of the evidence.
Tim, London

By the way, the quote I used ("is good for your back") was in the print version, but isn't in the online version, linked to above.