Dylan Evans

PLACEBO

The Belief Effect

The accepted treatment of the acute attack consisted in advising the patients to rebreathe their own air using a paper bag; for a more lasting cure they were taught breathing exercises to favour abdominal diaphragmatic breathing over thoracic breathing. The results were sometimes astonishing. One woman I saw used to have attacks in which she fell to the ground in the street and was unable to move for up to two hours. By doing no more than explain the syndrome to her I was able to bring about a complete cure: no further attacks. I could not have wished for a more dramatic demonstration of the truth of the theory. But alas, it seems I was wrong; all the evidence indicates that the hyperventilation syndrome is a chimera. In one well-executed study the hyperventilation provocation test showed no correlation between blood carbon dioxide levels and patients' symptoms and it is not alone in reaching this negative conclusion [1,6,12].

I was reminded of this in reading Dylan Evans's book, for in it he describes a rather similar experience to mine. While working as a psychotherapist in the 1990s he achieved an impressive cure in a patient suffering from panic attacks by supplying a psychoanalytic explanation for them, and he took this to be confirmatory evidence for the theory. Now, however, he questions the cause-and-effect process that seemed so convincing at the time. Perhaps the man's recovery was mere coincidence, or perhaps the happy result was simply due to the patient's belief that Evans's explanation was correct, even though it was not.

This is just one example of the critical approach that Evans brings to his subject. At the outset he effectively demolishes various items of medical folklore about the placebo effect. He makes it clear that what we know, or think we know, about the subject is largely based on misquotation, blind repetition of poorly substantiated claims, and the uncritical reporting of anecdotes. We really have little firm evidence about the matter. Although placebos are very widely used in clinical trials, these seldom include a no-treatment group so it is impossible to know if the improvement ascribed to the placebo might not have happened anyway for other reasons.

In fact, researchers who have investigated placebos have produced widely varying reports of their effectiveness, ranging from zero to 100 per cent, though a figure of about 30 per cent is often quoted and seems to be widely accepted. This derives from a very influential paper by Henry Beecher published in 1955, which Evans regards as having had serious flaws[2]. But Beecher was one of the leading advocates of the need to evaluate treatments by means of randomized placebo-controlled clinical trials, and because of this his view of the placebo as extremely powerful became widely accepted. Many people today believe that almost any medical disorder you like to name may respond to a placebo, but Evans takes this to be a medical myth. There is good evidence that placebos can affect influence pain and inflammation, including peptic ulcers, he says, and they also work for depression and anxiety, but this is pretty much the whole of their effectiveness.

Evans uses his list of conditions that may respond to placebos to construct a hypothesis about how they may work. In outline, he suggests that they do so via nervous system influence on the immune system. His idea is that they damp down the acute phase response and so produce changes both peripherally in the tissues and centrally in the patient's sense of well-being. He develops this idea at some length and tries to set it in an evolutionary context to explain how it may have arisen. This part of the book is, as he admits, somewhat speculative, and I think he has not entirely escaped falling into the trap of choosing his list of responding conditions to fit in with his theory rather than the other way round.

This emerges in his discussion of placebos in relation to cancer. He tells us that there is no evidence that placebos can cure cancer. However, the medical literature does contain a fair number of well-authenticated cases of spontaneous remission of cancer[4,5,7,8,9,11]. Unless one is willing to ascribe these to divine intervention (and not all the cases have occurred in a religious context), there must be some physiological explanation for them. The most popular suggestion is that such recoveries have an immunological basis, though admittedly there are difficulties with this idea and other mechanisms have been suggested[5,7,9]. If it is right, however, it would seem to be a logical extension of Evans's immunological explanation of the placebo effect. I therefore find it rather puzzling that he fails to acknowledge the occurrence of spontaneous recovery in cancer. His reluctance to do so may be due to his justifiable wish to discount over-optimistic claims that "positive thinking" can cure cancer.

In any case, the main interest of the book, for me at least, lies not in Evans's theorizing but in his presentation of the facts about placebos and in the implications of these facts for CAM. His view, not surprisingly, is that all forms of "alternative medicine", as he refers to it, depend on the placebo effect and on the patient's belief. He assigns a central role to belief, but he makes the point that this is not a simple concept and claims that it is possible for people to hold beliefs without realizing it. What matters is how people behave, not what they say, so even animals may hold beliefs of this kind. This is a behaviourist view of belief. "Beliefs are brain processes which cannot be observed directly, but whose existence can usually be reliably inferred from a person's behaviour - which includes, but is not limited to, his verbal behaviour." In this connection he makes the important point, which I agree with, that it is necessary for a patient to interpret a procedure such as acupuncture as constituting a therapeutic intervention if it is to work. Inserting needles into patients for diagnostic as opposed to therapeutic purposes seldom seems to produce relief of pain.

In his discussion of acupuncture he correctly points out that it is exceedingly difficult to devise suitable control treatments for clinical trials of it. Even single-blind trials are almost ruled out by the fact that if they are performed by a skilled acupuncturist he or she will know when a needle is being inserted in the "wrong" place and will probably communicate this to the patient even without meaning to do so. He therefore concludes that when clinical trials appear to show that "real" acupuncture is better than "sham" acupuncture the difference is due to inadequate blinding, and that acupuncture "is probably a pure placebo".

Whether this is true or not, acupuncture is of course a form of physical treatment which involves a social interaction between therapist and patient; the same is true of osteopathy and chiropractic. I agree with Evans that the preliminary manual examination of patients must have a therapeutic effect in itself, independently of the formal treatment methods used, and that there is a similarity here to the grooming behaviour of chimpanzees and other primates[3].

Homeopathy is easier than acupuncture to investigate scientifically because it is possible to provide convincing placebos for double-blind trials. At least one meta-analysis has concluded that homeopathy is, on average, more effective than a placebo[14]. Evans rejects this on the grounds that, if homeopathy works, we must abandon most of modern physics and chemistry, which he finds too high a price. This is really another version of David Hume's argument against miracles. Homeopaths will naturally object that Evans is expressing an unreasonable prejudice here, but he is undoubtedly right in holding that research of this kind is unlikely to settle the arguments. "Despite their appearance of absolute impartiality and objectivity, even the most rigorous statistical analyses allow much scope for personal interpretation and subjective opinion." Indeed.

It is not only CAM that Evans considers to be dependent on the placebo effect; he reaches a similar conclusion about psychotherapy. He points out that all the numerous psychoanalytic schools and their offshoots have about the same success rate, which suggests that their different theoretical assumptions are irrelevant. More controversially, he adopts much the same attitude to supposedly more scientific forms of treatment such as cognitive and behavioural psychotherapy: even if these are based on correct theories of how the mind works, this does not prove that they produce their effects in the way they claim rather than by making use of the patients' belief in their efficacy.

Perhaps the most important question of all is that taken up by Evans in his final chapter, entitled "The Witchdoctor's Dilemma". Even if many treatments do rely on the placebo effect for their success, does this matter? Perhaps we should simply accept the fact and seek to maximize the role of suggestion deliberately. We sometimes hear this argument advanced by advocates of CAM, and it does have a good deal of validity so far as the patient is concerned. Why should one care how one has been made better provided one has?

There is however an obvious ethical and practical dilemma for the practitioner. Is it ever justifiable to deceive patients, even for their own good? And this aside, if you do not yourself believe in a treatment, can you continue to use it effectively or will your skepticism convey itself to your patients and inhibit the placebo effect? Although Evans raises these questions he does not really answer them, though he does cite the fascinating trial done in 1965 at Johns Hopkins University, in which 15 patients suffering from anxiety were told that they were having sugar pills yet 13 improved. The outcome of this trial is certainly difficult to interpret for anyone, like Evans, who regards belief as the critical element in the placebo effect, although a possible explanation is that the patients may not really have believed that the doctors would do anything so silly as to give anxious people mere sugar pills and therefore concluded that they were being lied to. As Evans remarks, this story clearly illustrates how difficult it can be to put the principle of informed consent into practice.

This is a well-researched book that merits reading by anyone who is seriously interested in unorthodox medicine or psychotherapy. Some may find its conclusions unwelcome or uncomfortable. It has to be acknowledged, however, that although clinical experience strongly suggests that some complementary treatments work, convincing objective evidence for their efficacy has frequently proved disappointingly elusive. The fact that different versions of the various CAM therapies all seem to have about the same success levels is also worrying. And although it is possible to provide plausible-sounding explanations for some forms of treatment such as acupuncture we have to recognize that these explanations could be wrong. For all these reasons it is hard to exclude the possibility that much if not all of CAM works because of the placebo effect. Whether this ultimately matters is perhaps partly a philosophical question, but if we do in the end find ourselves forced to conclude that these treatments work only because the patients (and the therapists) believe in them, we shall need to confront the difficult dilemma that Evans posits at the end of his book.

References

1. Bass C. Hyperventilation syndrome: a chimera?. Journal of Psychosomatic Research 1997;42(5):421-6.
2. Beecher HK. The powerful placebo. Journal of the American Medical Association 1955;159:1602-6.
3. Campbell A. Acupuncture and the placebo response. Complementary Therapies in Medicine 2000;8(1):43-46.
4. Heim ME, Kobele C. Spontaneous remission in cancer. Onkologie 1995;18:388-392.
5. Hercbergs A, Leith JT. Spontaneous remission of metastatic lung cancer following myxedema coma. Journal of the National Cancer Institute 1993;85(16):1342-1343.
6. Hornsveld HK, Garssen B, Dop MJ, van Spiegel PI, de Haes JC. Double-blind placebo-controlled study of the hyperventilation provocation test and the validity of the hyperventilation syndrome. Lancet 1996;348(9021):154-8.
7. Kappauf H, Gallmeier WM, Wunsch PH, Mittelmeier H-O, Birkmann J, Buschel G, Kaiser G, Kraus J. Complete spontaneous remission in a patient with metastatic non-small-cell lung cancer. Case report, review of literature, and discussion of possible biological pathways involved. Annals of Oncology 1997;8(10):1031-1039).
8. Kappauf HW. Unexpected benign course and spontaneous recovery in malignant disease. Onkologie 1991;14(SUPPL. 1):32-35).
9. Kleef R, Jonas WB, Knogler W, Stenzinger W. Fever, cancer incidence and spontaneous remission. Neuroimmunomodulation 2001;9(2):55-64).
10. Linde K, Clausius N, Ramirez G, Melchart D, Eitel F, Hedges LV, Jonas WB. Are the clinical effects of homeopathy placebo effects? A meta-analysis of placbo-controlled trials. Lancet 1997;350:834-843.
11. Mastall H. Spontaneous remission of lung metastases of a primary uterus carcinoma during immune therapy. Zeitschrift fur Onkologie 1997;29(3):87-88).
12. Morgan WP. Hyperventilation syndrome: a review. American Industrial Hygiene Association Journal 1983;44(9):685-9.
13. Park L.C., Covi L. Nonblind placebo trial. Archives of General Psychiatry 1965;12:336-345.
14. Troosters T, Verstraete A, Ramon K, Schepers R, Gosselink R, Decramer M, Van de Woestijne KP. Physical performance of patients with numerous psychosomatic complaints suggestive of hyperventilation. European Respiratory Journal. 14(6):1314-9, 1999.

18 March 2003