Are RCTs appropriate to homoeopathy?

Barry raises a number of objections to the use of randomised control trials in homoeopathy. Some of these were dealt with earlier, but there are four that deserve further attention.

1. Anthropology offers different modes of evidence to biomedical science.
2. Homoeopathy as practised in the alternative setting is different to homoeopathy in the biomedical setting.
3. Homoeopathy is too holistic to test with RCTs.
4. Without the support of the pharmaceutical industry, there is very little funding for research in homoeopathy.

Anthropology vs. biomedicine

Barry appeals for biomedical evidence to be expanded to include anthropological studies.

Just as the scientific laboratory method and the nature of population statistics have shaped the nature of RCT evidence, so too anthropological method influences what constitutes evidence. Ethnographic research is conducted in everyday real-life settings and so can pay attention to the all-important contextual features of interaction. Reality is seen as ever-changing through a series of processes, formed by interactions and relationships between people, and always affected by the context in which social actions take place. The method utilises an observer situated in the context, not researching from afar.

What would James Lind have thought, I wonder, of the “from afar” comment after sharing a 50-gun warship with his subjects for weeks on end?

Here Barry has confused different types of evidence. Randomised control trials evolved to answer a specific type of question: Does treatment X work? Even within medicine, RCTs are not able to answer everything because they are not suited to all questions. Nobody is about to use an RCT to locate the mutation responsible for Ann Coulter. Anthropological evidence is good at what it does. If you want to know how women work within the tribal heirarchy of the Chimbu, nothing can beat a camera, a notebook, and a ticket to New Guinea. But observational anthropological evidence has its own limitations. It is good for describing human activity, but is not good for assessing its effectiveness. Recording a rain dance does not make one a meteorologist.

This does not mean that anthropological evidence is useless in the biomedical context. Quinine was found to treat malaria only after a Jesuit priest called Agostino Salumbrino observed the Quechua Indians of Peru chewing bark to ward off fevers. Salumbrino was not satisfied with mere observation. Back to Rome he sent a parcel of dried cinchona bark to be tested for its curative powers. (In the seventeenth century, malaria was still a major cause of death and disease in Rome.) Cinchona bark turned out to have anti-malarial properties. Several anti-malarials in common use today are chemically-manipulated variations of the drug discovered by the Quechua Indians.

Anthropological observation can generate valuable medical hypotheses, but it can not substitute for the testing of those hypotheses. Otherwise, observing the rituals of sports fans becomes evidence that wearing the same shirt to matches influences results. It means that observing churchgoers praying to God is evidence ipso facto of God’s existence and at the same time, cataloguing the posts in an atheist email list is evidence of God’s non-existence. It means that N-rays were real when Blondlot was alone and then evaporated into nothingness when Wood entered the lab. This knot of contradictions does not trouble the deconstructive anthropologist.

“Reality is seen as ever-changing,” says Barry, “through a series of processes, formed by interactions and relationships between people…” Whatever people believe is real.

While Barry may regard this as a worthy anthropological principle, I think it neatly cleaves the anti-rationalist school of anthropology from the scientific school. There are quite a few anthropologists who would choke on these words, and it is disingenuous of Barry to use the adjective “anthropological” throughout as if her view was standard among the field.

(It also creates a paradox for her agenda: if reality is ever-changing according to the interactions between people, then the role of RCTs in medicine is best evaluated according to its use by medical practitioners, in which case deconstructive anthropology is doomed to fail on its own terms. This is, by the way, why many left-leaning scientists like Alan Sokal have been so vehemently opposed to the adoption of postmodern relativism by leftists in the humanities and social sciences. By asserting that science and knowledge and the concept of truth are nothing more than political tools wielded by the powerful, it leaves nothing for the powerless. If truth is just a weapon of privilege, then it will always belong to the rich, to the great religions, and to the military. Everyone else had better get their yokes fitted now. There is only one place where the anti-scientific, anti-truth agenda thrives in conjunction with great political power, and that is in the hands of authoritarian dictators. Even among Western governments, the most aggressive scourge of scientific principles has been the Bush administration. When you abandon objective reality, you are not helping the poor and the marginalised, but the plutocrats, the priests, and the generals.)

Homoeopathy in different settings

Barry argues that homoeopathy trials fail because the system of homoeopathy is quite different when performed in alternative clinics than when it is performed in biomedical centres such as general practice or hospital.

Homeopathy as conducted by a homeopathically trained biomedical GP in an NHS practice was so different from the practice of lay homeopaths as to be unrecognisably the same therapy (Barry, 2005). The GP prescribed homeopathic remedies for biomedical diagnoses, collecting little information from patients. The professional homeopaths prescribed on the totality of symptoms, many nonmedical, which necessitated much more complex consultations.

I have no difficulty at all in accepting this fact. A biomedically trained doctor is always likely to assess a health problem from a different point of view to a lay healer, even if he or she has rejected the biomedical model. A lapsed Catholic is not the same as a freethinker. The problem, though, is overstated. If homoeopathy is different in different settings, then it should be tested in different settings.

But this leads on to Barry’s next point: that homoeopathy is so incompatible with biomedical thinking that it simply cannot be tested within the biomedical model.

Is homoeopathy too different to test with an RCT design?

The nature of Homeopathy precludes the straightforward administration of clinical trials to measure it. Based on the principle of treating like with like, homeopathic remedies are developed from substances in the natural world. A picture of the symptoms of these substances is catalogued by ‘proving’ the effects on healthy volunteers. A much-diluted form of the remedy is then administered to patients suffering with a picture of symptoms that is closest to that particular remedy. Each remedy picture includes multiple physical symptoms in multiple body locations, diverse psychological and emotional states, and aspects of behaviours that are not part of biomedical diagnoses. For example symptoms that improve on violent motion, particularly dancing (irrelevant to a biomedical diagnosis) is one of the keynote aspects of the symptom picture for Sepia (Vermeulen, 2000). Different individuals with the same biomedical diagnosis will be prescribed different remedies, as their symptom and personality picture will likely be different in each case. So two important aspects of homeopathy, individual prescribing and attention to non-biomedically recognised ‘symptoms’, problematise the use of RCT methodology.

Barry is trying to say that homoeopathy is beyond the power of biomedical science to evaluate. But she seems unwilling to contemplate the fact that biomedical therapists are also faced with complex problems, unknown mechanisms of action, individual differences, and subtle measurement problems. Nothing illustrates this better than the management of pain.

The physiological mechanism of pain is still poorly understood despite volumes of research. There is no objective tool for measuring pain. The perception of pain varies with different personalities and, even in the same person, with changing circumstances. Different practitioners have different approaches. In other words, every objection that Barry raised for homoeopathy applies equally to the study of pain. Biomedical researchers, however, did not abandon scientific techniques. Instead they worked to adapt the techniques to the problems of pain.

Measuring the subtle

While researchers prefer big, bold outcome measures like mortality, it is not always possible to deal in such discrete, easily measured outcomes. With pain, there is no way of counting, no easily observeable measure, and no machine that can tell how much pain a person experiences. So a number of different scales were developed to deal with the problem. Many of these scales were borrowed from psychological research, which often faces the same obstacles.

One of the best-known scales is the Likert scale. You will have seen this scale almost any time you have been asked to fill out a survey form. It looks something like this:

My pain level today is…

no pain

1 2 3 4 5 6 7

excruciating

The Likert Scale can be adapted quite freely. For instance, when researchers study pain in children, they often use a “faces” scale that can be answered by children from the age of three.

Some critics feel that the Likert scale limits responses to unrealistic preset possibilities, and therefore compresses the information from the survey data. So along came the Visual Analogue Scale, a simple 10-centimetre line on which the subject can put a cross wherever they like, and the answer is then measured in millimetres.

This certainly gives the respondent a great deal of freedom. However, most research has shown that the responses tend to cluster at a few common points so that the data is very similar to that from the Likert scale above. And the Visual Analogue Scale has a big disadvantage: up to 10% of adult respondents fail to answer correctly. Some people just don’t get it. For the Likert scale, only about 3-5% can’t manage it.

The debate and research into different measurement scales is extensive. There is of course no perfect way of measuring anything, not even hard physical constants. But even subtle things are measurable. The Likert scale and the Visual Analogue Scale can measure anything that can be put into words.

Naturally, such measurements are less precise and rigorous than, say, a haemoglobin test, but on the other hand, they are actually easier and cheaper to measure. If homoeopathy provides subtle spiritual benefits, then researchers should be able to observe them. For all the limitations of the Likert and Visual Analogue scales, they are still far superior to the uncritical observational approach advocated by Barry.

The holism of homoeopathic treatment

One of Barry’s more interesting points is that homoeopaths outside the biomedical setting tend not to treat single symptoms or diseases and instead focus on multiple factors and prescribe widely varying therapies even for the same presenting problem. This does indeed raise a challenge to biomedical methodology, but it is not insurmountable. After all, whatever the process that homoeopaths use to come to their decision, the final result is a prescription for treatment. And whenever the question is “Does treatment X work?”, the best trial design is the RCT. Let’s see how it could work in the homoeopathic setting.

First, the patient sees the homoeopath. The consultation takes place as usual, and at the end of the consultation, the homoeopath writes a prescription. The patient then delivers the prescription to a homoeopathic supplier who makes up the prescription and at the same time makes up a placebo mixture using the same solvent in the same concentration as the treatment. Both bottles are then passed to a researcher whose sole responsibility is to randomise the bottles so that the patient either receives the prescribed treatment or the placebo. While the patient waits for the prescription, he or she fills out the baseline questionnaire. Then the patient receives the bottle with instructions, goes home, and returns after a suitable interval to repeat the questionnaire. Once all the data is collected, the randomisation is revealed, and the data analysed to see if there was a significant difference between the placebo and the treatment.

Now the advantage of this design is that it does not matter what the homoeopath has prescribed. There is no need for a standardised prescription for a standardised biomedical diagnosis. There are problems with this approach, of course, in that it may be that some homoeopathic treatments work and some don’t, and by mixing them all into the same trial the results will be unimpressive. But the answer to this is to examine specific problems and specific treatments in the homoeopathic setting — a strategy that Barry dismisses a priori as reductionist.

Interestingly, the British Medical Journal paper on homoeopathy discussed in the previous instalment did one thing very, very well and that was its randomisation. The authors went to great pains to randomise and double-blind the data, and almost all the correspondents, even the skeptics, praised this aspect of the study.

Barry’s objections to RCTs in homoeopathy are all surmountable. What Barry is essentially doing is raising objections to the RCT methodology in order to excuse homoeopathy from scrutiny. For instance:

…[T]he Royal London Homeopathic Hospital have (sic) very little success recruiting patients for trials of homeopathy. Their patients refuse to consider the possibility of being in a placebo arm. Such patients are very committed to taking homeopathy and have often been on a long waiting list.

Exactly the same problem faces biomedical researchers. Patients in public hospitals have to wait sometimes years for treatment. Patients with terminal diseases for which there is no established cure don’t want to be in the placebo arm. So what do biomedical researchers do? They offer inducements. Early access to new drugs, free treatment, and jumping the queue are all ways of motivating people to enrol in studies. If the problem with homoeopathy patients is the waiting list, then offer to treat them immediately if they consent to participate in the trial and promise that even the placebo patients will receive their proper prescription at the end of the trial. That way, even if they are in the placebo arm, they will still end up with their homoeopathic treatment earlier.

The money problem

Another of Barry’s objections is that most alternative medicines are not attractive to pharmaceutical companies as they are not patentable.

‘It takes a lot of gold to meet the gold standard of the clinical trial’ (Hess, 1998, p. 17), and in alternative medicine there are rarely powerful pharmaceutical interests, with the exception of herbal products, and therefore few trials.

Now this is undoubtedly true, but it doesn’t address the central point. Pharmaceutical companies are not in the business of researching unpatentable treatments. They should not be expected to be. The people who should be paying for research into alternative therapies are the people who make money from it: the alternative therapists.

The most obvious rejoinder to this is that drug companies have billions of dollars to spend on research every year and the alternative health economy simply doesn’t have access to that kind of spending power. And that rejoinder, as persuasive as it appears on the surface, is dead wrong. In fact, the alternative health industry may even be larger than the pharmaceutical industry.

Australian consumers spend twice as much on alternative therapies as they do on pharmaceuticals. In the UK, one in four people regularly uses alternative therapies and their total spending in 1994 was about a billion pounds. In the US where mainstream medicine is expensive by international standards, the ratio is not so dramatic, but even there, personal spending on alternative medicine is more than the total spent out-of-pocket on hospitalisation. And in 1997, Americans visited alternative therapists almost twice as often as primary care physicians.

The alternative health market is not the struggling pauper it likes to think it is. Worldwide it is a $20 billion dollar a year industry. Actually, that was in 1996, and it has been increasing by 15% per annum, which means it should be worth around $81 billion this year if that trend has been maintained.

Every objection that Barry raises to RCTs in homoeopathy has its corollary in biomedical research, including limited funding. The essential problem is not that homoeopathy is beyond the power of the RCT — and even in instances where RCTs are inappropriate, there are other scientific tools available. The essential problem is that homoeopaths by and large don’t want to test their effectiveness. And why would they? They are involved in a multi-billion dollar industry in which there is virtually no accountability.

This would not matter all that much if the alternative health market had the courage of its convictions and genuinely worked outside the biomedical model…or at the very least, if used to treat recognised medical conditions, had the honesty to admit that the paucity of evidence for its effectiveness and safety. But this is not the case. Alternative therapists often claim to be able to treat cancer and asthma and any number of biomedical conditions. Even when they don’t treat biomedical conditions, they often co-opt biomedical terminology, sometimes in nonsensical ways. Chelation therapy, liver detoxification, ozone therapy, the blood group diet, and “candida in the blood,” are all perversions of biomedical theory that make no sense within the biomedical model and have no antecedents in traditional ethnic medicine. Even homoeopathy, which sits well outside the biomedical model, is often used to treat biomedical diagnoses no matter what Barry claims about its spiritual aspirations.

What evidence, then?

Given that Barry feels that the standard biomedical evidence is incompatible with a fair assessment of homoeopathy, it is reasonable to ask what she does feel is good evidence? Here are two examples, the first from a patient and her homoeopath:

One user judged the fact that her husband sought counselling for his depression as evidence for the effectiveness of her own homeopathic treatment. Homeopathy helped her come to terms with difficulties in adjusting to her second pregnancy and this, in turn, had linked effects of changing the communication between her and her partner, leading to a whole sequence of behavioural changes in the family. She explained this as working through quite physical responses (such as nausea) to remedies. She did not separate out physical and emotional effects. Her homeopath accepted this, unquestioningly, as matter-of-fact evidence of homeopathic therapeutic effect.

I’m sure he did. Even more telling is Barry’s own standard of evidence — not on the level of individual treatment and response, but on the political level of funding for alternative therapies within the NHS. This is what she finds compelling:

…[A]lternative evidence may prove useful in what David St. George, an NHS consultant, has called the potential for holistic transformation of the NHS through a synthesis of science and spirituality into a new paradigm. As he puts it, ‘Perhaps alternative and complementary therapists can help the NHS to break out of its own prison’ (St George, 2004, p. 38). St George’s vision was incredibly well received by the 100 delegates at the Diversity & Debate in Alternative and Complementary Medicine in July 2004, who comprised a high proportion of alternative therapists and social scientists researching alternative medicine. Anthropological and other qualitative forms of evidence may prove a political tool to assist in this enterprise of transformation.

Barry’s preference for New Age jargon at self-congratulatory meetings over biomedical evidence as the basis of funding decisions should be seen for what it is: not a defence of poor, maligned alternative therapies, but a “political tool” (Barry’s words) for a multi-billion dollar industry to gain access to even more money, in this case from the public health purse, while excusing itself from the rigorous appraisal that biomedical therapies have to go through.

Even without further financial burdens, the NHS has severe shortages. It is not uncommon for patients who have had a heart attack to find themselves on a medical ward instead of a coronary care unit. I do wonder how much consolation Barry would expect such a patient to feel on discovering that there are not enough trained coronary care nurses in the hospital, but plenty of well-meaning homoeopaths offering spirit water.