Focus on Alternative and Complementary Therapies
www.pharmpress.com/fact
Focus Alternat Complement Ther©2005 Pharmaceutical Press
Focus Altern Complement Ther 2007; 12: 85–7
Therapies that are funded by the NHS should be shown by RCTs to be efficacious, but many treatments in CAM have not passed this test. Although RCTs are invaluable for testing the efficacy of drugs, it is claimed that, when applied to complex non-pharmaceutical interventions such as acupuncture, they may generate false-negative results.1 We believe that this argument is flawed.
Two fundamental requirements must be met so that the efficacy of any orthodox or CAM therapy becomes testable in RCTs. First, there must be a clearly defined desired effect (DE) of the intervention (usually this is an improvement in some aspect of physical or psychological health). Second, there must be a hypothesis that a clearly defined therapy will significantly increase the chance of achieving the DE. The purpose of the trial is to test this hypothesis.
A different view is advanced by Paterson and Dieppe, who maintain that three assumptions underlie the design of all RCTs – preconditions which, according to the authors, cannot be met for some complex CAM therapies.1 Therefore, these authors argue, RCTs cannot be applied to these areas of CAM. These assumptions are:
These three assumptions will be examined in turn.
Paterson and Dieppe claim that it is necessary to have a prior diagnosis before recruiting volunteers to an RCT, because this diagnosis is needed to determine eligibility for the trial.1 This is not true. It is possible to design an RCT of treatments for patients who have vague complaints such as tiredness, nervousness or weakness. The trial participants may not have a diagnosis (in the sense that the cause of their complaint is known) but there is a DE of treatment – namely to render participants less tired, or nervous, or weak. An RCT can compare the effects of (for example) aerobic exercise, homoeopathy, bed rest, hypnosis or no treatment. If the allocation to the interventions is truly random, if the treatment groups are of adequate size, and if the assessment of the extent to which the DE is achieved is made by a person blind to the treatment allocation, this is a valid way to assess the relative efficacy of the various treatments. But at no stage has a ‘biomedical diagnosis’ been required.
It is also possible to adjust the treatment during an RCT to clinical changes that may occur in trial participants during the study. Trialists can review the diagnosis as the clinical presentation alters and adjust their treatment accordingly, provided such adjustments have been made in the trial protocol.e.g. 2 In such situations, the diagnostic process takes place not just before but also during the trial. Contrary to what Paterson and Dieppe claim, such adjustments do not preclude an RCT.
The terms ‘characteristic effects’ and ‘incidental effects’ will be unfamiliar to most readers, but this nomenclature is central to Paterson and Dieppe’s claim that it is impossible to achieve a valid control group for a trial of complex CAM treatments like acupuncture. Characteristic effects (more commonly called specific effects) are defined by Paterson and Dieppe as ‘therapeutic actions that are theoretically derived, unique to a specific treatment, and believed to be causally responsible for the outcome’.1 Traditional Chinese acupuncture is based on the theory that qi (a form of energy) flows round the body along defined pathways. Qi has two opposite components, yin and yang, which should be in balance to maintain good health. The acupuncturist, by placing needles in critical points, restores the balance, and hence health improves.3 Therefore benefit from needling is a characteristic effect of acupuncture.
Incidental effects (more commonly called non-specific effects) are, according to Paterson and Dieppe, ‘the many other factors that have also been shown to affect outcome, such as the credibility of the intervention, patient expectations, the manner and consultation style of the practitioner, and the therapeutic setting’.1 The reason why acupuncture is deemed a complex therapy is that it does two things: it balances the flow of qi, and it also provides a different ambience, i.e. non-specific effects, compared with that provided with conventional treatments.
Even if specific and non-specific effects were inseparable, as Paterson and Dieppe argue, this would in no way preclude an RCT. One could, for instance, conduct a study of acupuncture in its usual ambience vs. a standard intervention or no treatment at all. Such studies would obviously not be placebo-controlled and double-blind, but they could certainly be randomised.
Paterson and Dieppe exaggerate the differences between CAM and conventional medicine. They seem to assume that ‘biomedical doctors’ are uncaring robots and that non-specific factors are unrelated to ‘biomedical’ theory. Wherever good medicine is practised, there will also be an abundance of ‘the diagnostic process and aspects of talking and listening’.1 These and other non-specific factors are characteristic of any type of good medical practice and are supported by sound medical theory.
Paterson and Dieppe base their theory about the indivisibility of specific and non-specific effects on interviews with users of acupuncture. Such qualitative data are always wide open to interpretation and seem woefully insufficient for drawing the wide-ranging conclusions about trial methodology which these authors propose.
The existence of the specific effects of needling in acupuncture has been tested in many RCTs. The placebo used may be a needle that does not actually puncture the skin (although it appears to do so), or a needle that is not inserted in the correct position. The majority of these trials suggest that the specific effects are not associated with significant benefit.4,5 Therefore, by the standards applied to mainstream medicine, acupuncture should not be eligible for NHS funding. This is not a trivial matter. According to the Smallwood report, ‘Of the three million acupuncture treatments received in the UK in 1998 one million were provided on the NHS at a total cost of nearly £26 million, more than the total of all other CAM therapies combined.’6
Those who promote the availability of acupuncture on the NHS use a rather subtle argument. They claim that the value of acupuncture does not reside entirely, or even mainly, in the specific effects (needling), but in the non-specific effects that are integral to the therapy. This explanation is reminiscent of the mythical Wizard of Oz, who was actually devoid of magical powers, but who caused miraculous DEs to occur to his supplicants because his non-specific effects were so persuasive.
If it becomes accepted wisdom, the redefinition of the nature of acupuncture (or other CAM or complex interventions) by Paterson and Dieppe raises several difficult questions.1 First, is it ethical to inform patients that their illness is due to an imbalance of yin and yang, which can be corrected if, by careful study of the patient’s qi, the acupuncturists can use skilfully placed needles to divert it back into the correct pathways? The use of ambience to enhance the confidence of the plaintiff goes back at least to the oracles at Delphi, or to Hippocrates. It persists today in the consultant whose letterhead bears an address in Harley Street to show that he is among the medical elite. But it is no longer acceptable for a healer to invoke what Paterson and Dieppe call a ‘non-biological theory base’ (or, more bluntly, a demonstrably untrue theory) to enhance the potency of their placebo therapy. Basic ethical standards today require that we tell the truth to our patients.
Second, it is not possible to evaluate the value for money of the £26 million spent by the NHS (at 1998 rates) on a therapy in which efficacy largely depends on a delusion, and which cannot be evaluated by the criteria used for other types of therapy. There can only be one set of standards in medicine. If we accept the concepts of Paterson and Dieppe,1 should we also allow a degree of deception within other areas of health care, e.g. drug therapy?
Third, at a time when we try to promote evidence-based health care, what harm is done to the patient who is asked to respect treatments based on sound scientific research, but also treatments based on a mythical life-energy flow in which even many acupuncturists cannot believe? The placebo effect is indeed a valuable aid to any therapist, but it is tarnished if it is achieved by dishonest means. Even non-mystical therapies come with the benefit of a placebo effect. We do not need to give placebo treatments to our patients to benefit from a placebo response, and we do not need to tarnish the placebo effect in order to profit from it.
