A fair amount of research in acupuncture has appeared, much of it of rather dubious quality. There are certain intrinsic difficulties confronting anyone who wants to do clinical research of this kind, much of it centred on the choice of a suitable control procedure. To the mildly interested onlooker with no particular interest in acupuncture it might seem that it would be enough to compare ‘real’ acupuncture with needling at non- acupuncture points and a number of trials have used this method. However, we know that, owing to the phenomenon of ‘diffuse noxious inhibitory control’, there will be some therapeutic response from inserting a needle anywhere at all, even to a depth of only 2 or 2 mm.
There is another difficulty as well. Suppose that 20 patients are treated with ‘real’ acupuncture by Dr A, an experienced acupuncturist, and another 20 with ‘placebo’ acupuncture by Dr B, who knows nothing about acupuncture. We cannot validly compare the outcomes, because we cannot assume that the two acupuncturists possess equally potent personalities in terms of their placebo effects. And if, to get round this problem, we allow Dr A to treat both groups, a different question arises. Dr A cannot be neutral about what he is doing; whether he wishes to or not, he will probably convey something of his own beliefs and prejudices to his patients (at least, we cannot assume that he will not). Dr B, on the other hand, may not produce the same therapeutic response as Dr A, and his treatment would not be regarded as valid acupuncture by most medical acupuncturists.
Even if, for the sake of argument, we grant that Dr A can maintain a completely impassive attitude that gives nothing away, we still are not out of the wood. Difficulties now arise concerning the acupuncture itself. How far away from the ‘correct’ site are the placebo needles to be placed? How deeply are they to be inserted? How much stimulation is to be given, and is teh chi to be obtained? In short, there is a wide range of possible variations in the way in which acupuncture is performed, and this makes comparison of different types of needling almost impossible.
In view of all these difficulties many researchers have concluded that placebo acupuncture is not a real possibility, and that a better idea is to compare acupuncture with other kinds of treatment. Two fairly comprehensive reviews of acupuncture research have appeared. Lewith and Machin (1983) reviewed 32 papers and concluded that the response rate is about 30 per cent for placebo, 50 per cent for ‘sham’ acupuncture, and 70 per cent for ‘real’ acupuncture. Most of the published trials, they believe, would not be capable of detecting differences of this order, hence one cannot necessarily conclude from these trials that acupuncture is merely a placebo treatment. They suggest that instead of comparing acupuncture and placebo it would be better to analyse the time for which a patient obtains relief from a given treatment.
Vincent and Richardson (1986) reviewed 40 studies. They, too, found serious shortcom- ings in most of them, but they concluded that there is good evidence for the short-term effectiveness of acupuncture in relieving a number of kinds of pain (mainly headache and backache but also a number of other painful disorders). They found the short-term response rate to be 50 to 80 per cent – higher, that is, than the expected placebo response rate of 30 to 35 per cent. The good initial response was not so well maintained, however, unless patients received booster treatments at intervals; this, of course, accords with what is found in ordinary clinical practice. No conclusions could be drawn about whether certain points are more effective than others.
Like Lewith and Machin, Vincent and Richardson are sceptical about the value of double-blind trials in acupuncture. They argue that single-blind trials are adequate ‘provided efforts are made to monitor independently the impact of non-specific effects and/or ensure that they do not vary between groups’. They also make a plea, which I would certainly endorse, for authors of research papers to give as much information as possible about what they actually did (number of sessions, duration and frequency of stimulation, whether teh chi was sought, method of point selection and so forth).
In spite of the difficulties that attend the carrying out of good research in acupuncture it is very important that it be done. Acupuncture has come and gone a number of times in the West, and there is no guarantee that it will not fall into disuse yet again. That would be a pity, for it has a valuable, if limited, contribution to make to medicine. At present it is more than half-way across the gulf that separates charlatanry from science in the minds of doctors, but if it is to complete the transition medical acupuncturists will have to produce evidence that it works.
By itself, however, that will not be enough. It will also be necessary produce some kind of rational account of how it might work. This will have to be a physiological explanation. In the case of pain a basis exists for explaining at least the short-term effects of acupuncture, but it is more difficult to account for the long-term effects and still more difficult to find a framework to accommodate the role of acupuncture as a treatment for non-painful disorders. However, I do not believe that it is impossible to do so.
Lewith, G.T. and Machin, D. (1983). On the evaluation of the clinical effects of acupuncture. Pain, 16, 111- 127.
Vincent, C.A. and Richardson, P.H. (1986). The evaluation of therapeutic acupuncture: concepts and methods. Pain, 24, 1-13; 15-40.