Chapter 5: Trials and Tribulations

Research into alternative medicine differs from most other kinds of medical research in one significant way: there is a strong pre-existing element of public affection and enthusiasm for this type of health care. The rapid rise of alternative medicine during the last two decades and the increasingly insistent demands for its inclusion as a part of standard medical education have moved it from the realm of the purely therapeutic and placed it squarely into the forum of public policy. Anyone who sets out to scrutinize or test the worth of a particular alternative therapy must therefore be prepared at the outset to deal with a certain degree of lobbying by loyal adherents. Similarly, if the research produces results that are, in the minds of the adherents, equivocal or unfavourable, the researcher can expect a torrent of impassioned and often unapologetically personal criticism and attacks on the national stage.

This was something I had definitely not anticipated at the time I took up my post at Exeter. On several occasions in those early days I had expressed my inclination to begin my research work by focusing on those alternative treatments that were particularly popular in Britain. I had assumed that this would mean things like homeopathy and chiropractic, modalities that internationally occupied a very prominent place in the alternative pantheon. I certainly had never imagined that spiritual healing would be high on the list.

It came as a bit of a surprise, therefore, when I was approached by a group of spiritual healers - actually they came to my home and rang the doorbell - expressing great delight that I was interested in studying their craft. At first I was bewildered. “Why would I do that?”, I asked. “Because you said so - you promised to focus on popular alternative therapies, and we are by far the largest group of alternative practitioners in the UK.”

They were absolutely correct. In the early 1990s, about 14,000 healers were registered with the UK Federation of Spiritual Healers - a total larger than the number of chiropractors, osteopaths, acupuncturists, homeopaths and herbalists put together; in fact, at the time, it was almost equal to the number of physicians in general practice.

Healers believe that there is a universal “healing energy” that emanates from cosmic, divine or other supernatural sources, and that they have the power to channel this energy into the bodies of their patients. Assisted by this channelled energy, they claim, patients are able to heal themselves.

In truth, I had never planned to investigate this area, and even if I had, it would not have been a high priority. I had always been of the opinion that it would be wiser and more productive to concentrate on investigating treatments that were based on some kind of physical or biological hypothesis, and spiritual healing certainly doesn’t come under that rubric. There is no way to explain logically how spiritual healing might work, based as it is on a set of beliefs and assumptions that are essentially metaphysical. That obviously makes it very much harder to design anything approaching a rigorous study. But, although dedicating research efforts towards a scientifically implausible treatment was unlikely to be a fruitful endeavour, the truth was that I had been hoisted by my own petard: after promising that I would concentrate initially on the most popular alternative therapies, I had little choice but to go forward with the design of an investigation of spiritual healing.

I began the process by applying for a grant from the Wellcome Trust, and started discussions with five local healers, trying to figure out how best to design and run a methodically sound study of this mystical therapy. The healers came highly recommended by their professional organization, which was as much of a reassurance as one could hope for that they were as good as they get. I put my initial reservations firmly aside: together, my team and I were determined to give it our best shot. If nothing else, this might be an interesting experience. As it turned out, “interesting” was an understatement.

To my surprise, our grant application was successful, and we received £50,000 to go ahead with planning a study - by no means a huge sum for a clinical trial, but it was a start. The terms of the grant were most unusual: the Wellcome Trust allowed us considerable latitude to design the trial collaboratively, drawing on the wisdom of both the healers and my research team. Not even the type of illness or medical condition that was to be the subject of spiritual healing had been predetermined.

Of necessity, our research had to rely heavily on the input of the healers themselves. They, after all, were the experts, the ones who knew best how they go about treating a certain type of patient to achieve a particular type of result. It has always struck me as important to incorporate such expertise at an early planning stage of a study. Had we not put the therapists themselves at the heart of the planning process, we would have been in danger of conducting a trial which was either entirely useless or which failed to test spiritual healing, and merely debunked it. In addition, as in any clinical trial, we required the input of methodologists with experience in the design, conduct, evaluation and publication of such investigations. If these aspects were neglected, as they often are in alternative medicine research, we might have ended up conducting a trial that failed to answer the specific research question at hand. And finally, of course, we needed to have the medical staff on board, the clinicians with expert knowledge in managing patients suffering from the condition to be treated - but in the early stages, the healers were still uncertain which type of patients they should choose.

After considerable debate, we all agreed that their best option would be to go for patients suffering from chronic pain. Our healers felt that, since they had had ample experience in treating pain and had reportedly previously seen excellent results, they would be able to relieve pain effectively in the context of the study. My team liked the idea, not least because there is never a shortage of pain patients, and we thus felt confident about recruiting sufficient numbers for an adequately sized study. Our colleagues from the local pain clinic were delighted to cooperate by providing us with their expertise in pain management and sending us their chronic pain patients. So, all the essential three areas of expertise were covered.

Next, the healers had to decide the details of the treatment schedule. They determined that no less than eight treatment sessions were required to generate a marked pain reduction. This created a bit of a problem for us: our budget was tight and, of course, the healers wanted to be paid for their services. We agreed to their demand of eight sessions but, as a consequence, there were hardly any funds left to support my team, but somehow we managed nevertheless.

From the outset, I had made it clear to everyone that the aim of our research was to test whether spiritual healing generated effects beyond a placebo response. Quite apart from the fact that this was the issue that interested me most, it also was, more importantly, the question on the basis of which we had obtained the funds from the Wellcome Trust. But what would constitute an adequate placebo intervention in a trial of spiritual healing?

When we study a drug, the placebo problem is usually quite easy to overcome: we only need to produce a pill that contains no active ingredient but is otherwise indistinguishable from the real thing. In alternative medicine, though, things are typically not quite that simple. Particularly with physical or psychological interventions, such as acupuncture or spiritual healing, there is normally no obvious placebo or sham therapy which is inert (as a placebo must be), and which patients find indistinguishable from the real treatment (which is the second precondition for an adequate placebo).

After further discussions with our five healers, we settled for actors, or “sham healers”, who would masquerade as healers. They would be taught the ritual of a healing session by our healers but, not being trained healers, they would not be emitting any “healing energy” and would thus be therapeutically inert. If they followed the ritual of the healers, our patients would be unable to distinguish them from the real healers. This seemed like a perfect solution to a tricky problem, and everyone was happy to have found it.

Little did we know that a big surprise awaited us. After teaching the ritual of healing to our five actors, the healers decided that our study design was no good after all. With breathless excitement, they reported to have discovered a most remarkable phenomenon: the actors also possessed healing powers! Thus, they argued, a zero difference between the therapeutic effects generated by the healers and those prompted by the actors would, in fact, not indicate that spiritual healing had been ineffective. According to the consensus of our five healers, such a result was likely, and it would merely demonstrate that both types of treatment had been effective. I was baffled; if the healers were to be believed, it seemed as though we had manoeuvred ourselves into a situation where we were attempting to test an untestable hypothesis.

More discussions ensued, and eventually we came up with a new and seemingly watertight concept: we would place the healer in a cubicle adjacent to the patient, so that the patient could not see the therapist. The great appeal of this solution was that it would enable us to include a placebo group where nobody at all was present in the cubicle. With this set-up, there would be no danger of any inadvertent healing powers confounding our conclusions. But then second thoughts began to emerge: healers do not normally sit in cubicles while treating their patients. Might this abnormal situation impede their healing energy? Perhaps the one-way mirrors of the cubicle would deflect their healing energy? In this case, a negative result would again be difficult to interpret, and the healers might find reasons for declaring our trial invalid.

This point is less trivial than it may seem at first glance. When doing research, we want to make sure that any possible result can be interpreted in only one way. If two or more plausible conclusions can be drawn from it, we do not advance the existing knowledge and might even obscure the truth. We felt that it was our responsibility to design the study in such a way as to produce unambiguous results, one of the hallmarks of rigorous science.

After further discussions with our healers, we eventually reached a workable compromise: we agreed to include not two but four different treatment groups in our study. According to the final protocol, patients were allocated at random to receive one of the following four interventions:

This finally looked like a watertight study, and we were all pleased and proud to have solved these difficult methodological issues amicably. Our ideas in planning this study were indeed quite innovative: several investigators have since copied our trial design and have published results that confirm our findings. I was particularly pleased to have reached a consensual solution within a team of people coming from the most diverse backgrounds imaginable. This demonstrated, in my view, that even complex methodological issues can be tackled by individuals with no prior experience in research, provided one set of experts guides the other and real teamwork is allowed to develop.

All that was left to do now was to obtain approval from our local ethics committee and sort out a handful of logistical details; after that, the adventure of conducting this trial could begin.

At that point, I delegated the day-to-day running of the study to two of my co-workers who regularly briefed me on its progress. They worked hard to recruit patients, made sure the treatments were applied as specified in the protocol, motivated patients to show up for return visits, recorded the results and generally kept everyone happy. For the most part of a year, my team was busy running the trial.

During this time, I occasionally lent a helping hand to wheelchair-bound patients who had to navigate the stairs leading to our research laboratory. It was on one such occasion that I got an early glimpse of what the results of this study might show. Some of the patients whom I had previously encountered in wheelchairs I later observed walking the stairs. I was puzzled and asked: “Did I not see you in a wheelchair the other day?” “You certainly did, professor. The healing is doing me so much good that, for the first time in years, I can do without my wheelchair.”

I was astonished. How could this be? I recall being so shocked that I needed to talk it over with Danielle at home. The nature of the trial meant that patients were receiving either healing or placebo - neither they nor I could be sure which. But such dramatic benefit was extraordinary in either of those situations. I awaited the results of our research with increasingly eager anticipation.

Finally the study ended. After the last patients had received their last healing session the randomization code was opened, the statistics were calculated and the results were summarized. What did they show? Was spiritual healing an effective form of pain control? Were we about to surprise the world of science with a sensation?

The findings demonstrated considerable pain reductions in all four groups but no significant difference between them. If anything, the placebo groups had experienced a little more pain relief than the patients receiving real spiritual healing. In other words, the effects patients had experienced during and after the healing sessions were due to a placebo response. Our results demonstrated how impressive this response could be - so effective, in fact, that some patients were able to abandon their wheelchairs entirely. This extraordinary response had occurred in a few patients from all four groups, and the average pain scores did therefore not differ between them. Our conclusion in the final paper was clear: “...a specific effect of face-to-face or distant healing on chronic pain could not be demonstrated...”

Once we presented the results to our five healers and other interested parties, lively and at times emotional discussions ensued. Some people argued that our findings proved how very useful healing truly was: after all, many patients had experienced significant pain reduction. Therefore, they urged us not to publish the findings, particularly not with the conclusions we proposed to draw. If we went ahead with our plan to report that spiritual healing relied entirely on a placebo effect, we would inevitably deprive thousands of patients of the obvious remarkable and clinically important benefit healing had to offer.

My opinion, and fortunately that of the rest of my team, was, however, quite different: yes, we had demonstrated the amazing power of the placebo effect, but we had also shown that spiritual healing was nothing more than a placebo. It was important that health care professionals and patients knew the truth. Deliberately refraining from revealing our results would make a mockery of our research. The conclusions of our study were very clear; however, the implications and recommendations that would follow from them were a different matter. The practical implications of our results depended on a range of factors that were not directly related to our study. For instance, an important next step would now be that our trial needed to be replicated by an independent team of researchers. Not to publish our results at all would prevent this crucial process and was therefore not an option. Unpublished science, I had always been taught, is no science.

This debate also had serious ethical implications. Contrary to my critics’ assertions, I had no desire whatever to prevent patients from benefitting from alternative therapies. I did, however, feel duty bound to make sure that, in future, patients would receive the best available treatment - and the best available treatment is not merely a placebo, but a therapy that generates beneficial effects in addition to those induced by a placebo response.

Any effective treatment - effective beyond placebo that is - will generate a specific effect plus a placebo effect, provided that clinicians administer it with sufficient time, dedication, compassion and empathy. These non-specific effects, as experts often call them, are essential elements in health care and, most likely, they were the reason why the patients in our spiritual healing study had experienced less pain. Yet while non-specific effects are clearly useful, they can never be a sufficient reason for deliberately using placebos in place of established treatments. Clinicians who only apply placebos - when there is an effective therapy available - are in effect cheating their patients out of something that is often essential in bringing about a return to good health.

Others criticized our research because they felt that the reductionist approach of a randomized clinical trial is intrinsically incompatible with alternative medicine. Alternative medicine, they felt, is far too holistic, subtle and individual to be assessed by the standard scientific methodology. Again, I disagreed. A randomized clinical trial is the best research tool for providing reliable estimates of the effectiveness of a therapeutic intervention. Our lengthy discussions with the five healers had clearly demonstrated this to be true. The healers had originally wanted us to evaluate the effect of their treatment by measuring the pain of their patients before and after their therapy and subsequently calculating the difference between the two measurements. This would have certainly been much easier to do. We would not have needed to think about what might be an adequate placebo, and we would not have required any control groups, which in turn would have drastically reduced the number of patients we needed to recruit for our trial. Yet this approach would have generated a misleading impression: by disregarding any consideration of the placebo response, it would have wrongly implied that the spiritual healing itself was the sole effective agent in alleviating pain. Remarkably, our healers were able to resolve the methodological issues with us and eventually agree on a study design that would provide a fair and conclusive answer to the research question we had decided to tackle.

They were able to comprehend that all symptoms, particularly pain, can improve over time, often even without therapy. Such symptoms can also respond to the patient’s own expectations, or to a multitude of other circumstances. The crucial point about a clinical trial is that it is a test that should give a reliable result about the value of the treatment per se. A properly planned, rigorously conducted and adequately analysed clinical trial can not only determine whether the observed outcome was caused by the therapy rather than by non-specific effects, but it can also avoid the problem of false positive or false negative findings.

Let me explain:

A false positive finding suggests that a treatment or intervention was effective when in fact it was not. Conversely, a false negative finding is a result that erroneously suggests that a treatment does not work when, in fact, it is effective.

In our spiritual healing study, a false negative result could have occurred for a range of reasons. If, for instance, the group of patients which we recruited had been too small, the statistical tests might have failed to show a significant difference in pain reduction between the groups despite there really being one. Therefore we had to make sure to include a sufficiently large number of patients. But there were many other potential pitfalls as well: if, for example, we had only asked one healer to perform the treatment, this person might have been useless at his job and we might have ended up inadvertently testing the skills of that particular healer rather than testing the effectiveness of the treatment overall. We thus recruited five healers of good reputation within their profession.

If we had chosen patients with a medical condition that the healers had not been confident that they could treat effectively, a negative result would have been a foregone conclusion, and our study would have been little more than a shabby debunking job. For instance, no healer I know claims to treat hair loss. It would therefore have been easy to find 100 bald men and demonstrate how utterly ineffective healers are at re-growing hair. We had to be certain to choose a condition that, in the opinion and experience of the healers, was likely to respond optimally to their therapy. Similarly, if we had used an insensitive method of measuring pain, we might have missed an effect that was, in fact, present. Luckily, we were able to choose from several properly validated means of quantifying pain.

A false positive finding, by contrast, is a result that erroneously suggests that a treatment does work when, in fact, it is not effective. This can happen all too easily in poorly designed clinical trials, particularly if they are open to bias. For instance, if we had allowed the healers or the patients to decide who received healing and who got the placebo, expectation of an effect would most likely have created a false positive impression. Therefore, it was important to allocate patients by chance, i.e. through a process of randomization, rather than by individual choice. A similar error might have arisen if the patient or the researcher evaluating the results had known which patient belonged to which treatment group. Thus we had to “blind” both parties to the treatment allocation. Not only that, but we also had to determine how successful our attempt to blind our patients had been. Blinding, in turn, was only possible once we had developed a placebo treatment that, to the patients, was indistinguishable from the actual healing intervention.

In the end, we had collaboratively designed what is technically called a randomized, placebo-controlled, double blind trial with four parallel groups - and, remarkably, we had done so not because the methodologists had dictated this type of study but because, after many long discussions and debates, we had arrived at a consensus amongst all parties concerned.

The randomized, placebo-controlled, double blind trial is the design that seems to engender great scepticism and wariness among many proponents of alternative medicine. Had we made a point of insisting from the outset that we were planning to conduct such a study, our healers might have flatly declined to participate. Instead, we debated and re-debated all the issues with them and let them find for themselves the most appropriate, most logical common sense solutions to the problems that emerged while designing the study. All the features of such a study serve the purpose of rendering the results reliable and of making as sure as possible that they are not influenced by bias, and can thus be interpreted clearly and unambiguously. In other words, these features are important for maximizing the likelihood that the results of a trial will actually test whether the observed effect is caused only by the treatment or by some other factors out of our control.

Without randomization, blinding and placebo controls, nobody could have been sure what the cause of the observed effect had truly been. If we had noticed a difference in pain reduction between the four groups of patients, and if the safeguards of randomization and double-blinding, etc. had not been in place, it would have been entirely reasonable to conclude that an observed effect might have been due to the expectation of the patient, or to the encouragement of the healer, or to other phenomena not directly related to the spiritual healing intervention itself. The question of cause and effect - that is, whether the outcome is directly and entirely due to the intervention - can only be determined if a clinical trial is designed to exclude possible causes other than the treatment itself.

Of course, clinical trials are not infallible. Of course, they can be flawed, and, of course, they can produce misleading results or lead to inaccurate conclusions. I have not yet met anyone in any branch of health care who would claim otherwise. The point, though, is that, until we have a better tool for testing the effectiveness of therapeutic interventions, clinical trials are without a shadow of a doubt the best tool we have - or, to borrow a famous aphorism, clinical trials are the worst research tool for determining therapeutic effectiveness of medical treatments, except for all other tools we presently have at our disposal.

Because clinical trials can never be 100 per cent fool proof, it is advisable not to rely on just one single trial when issuing general recommendations. In fact, whenever possible, it is by far better to make sure that, for therapeutic decisions, all the available studies, particularly those which are of high quality, are taken into consideration - and this requires what is termed a “systematic review”.

A systematic review is an exhaustive examination and discussion of all the available published research that has been carried out on a given topic. Conducting a systematic review necessitates digging out all the studies that have ever been conducted on the given subject, often a tedious and time-consuming exercise requiring a considerable amount of skilled detective work. This task must be performed diligently because it is the only sure way to avoid cherry picking - that is, selecting those studies that, for whatever reason, we happen to like while ignoring those that don’t support our position or the hypothesis we’re developing. Next, we must critically assess the quality of the trials that we have identified; that is to say, we must evaluate the soundness of their results and conclusions and weigh the possibility of their having produced a biased, misleading, false negative or false positive result. Finally, we summarize what the best trials tell us. This process of summarizing the findings of multiple studies can sometimes require a statistical approach, in which case we are dealing with what experts call a “meta-analysis”.

Systematic reviews are much more reliable than single trials because, by virtue of the sheer number of different studies from different investigators they involve, they minimize the play of chance as well as the influence of bias. They are therefore generally accepted to be the most trustworthy source of evidence in health care. This is one important reason why, over the years, our research in Exeter focused more and more on this type of research.

In the case of spiritual healing, we conducted a systematic review that we later updated when new evidence had become available. Initially, the evidence seemed mixed and at times contradictory but, once our trial design had been adopted by other investigators and further rigorous studies had emerged, the totality of the reliable clinical trials generated a fairly clear picture which confirmed the findings of our study: “...the evidence [is] against the notion that distant healing is more than a placebo.”

Our healing study was a typical example of the difficulties in designing research in alternative medicine but, at the same time, it also showed that, formidable though they were, the obstacles in the way of high quality trials can be overcome. This study was one of the first that my Exeter team conducted. Over the years, we published around 40 more clinical trials - always trying to incorporate the views and expertise of the therapists - and well over 300 systematic reviews.

With some alternative therapies, designing clinical trials can be fairly straightforward. For instance, when testing treatments that are taken by mouth, we usually do not encounter significant obstacles in controlling for placebo effects. It is relatively easy to “blind” patients and their physicians, and to give placebos to those allocated to the control group. Thus our trials of a weight loss aid, of homeopathic arnica, of Bach Flower Remedies, of a garlic supplement or of a Ginkgo biloba supplement all gave us relatively few problems in that respect.

In other areas of alternative medicine, though, things can be much trickier. We often had to develop innovative concepts and new ideas to solve the methodological problems we encountered. What, for instance, might constitute an adequate placebo for a trial testing the effectiveness of acupuncture? This question had intrigued the field of acupuncture research for quite some time. Several investigators had used needling at non-acupuncture points; others had opted to insert needles only superficially into the skin; others still had used acupuncture points that, by general consensus, were not indicated for the condition in question. Whenever such trials had not generated the desired positive result, acupuncturists had argued that the placebo intervention was not entirely inert, and that therefore the negative result was not valid. Such studies, they were convinced, showed no difference between acupuncture and sham-acupuncture simply because both interventions were effective. For them, paradoxically, each such negative finding was merely a confirmation of the effectiveness, and not of the ineffectiveness, of acupuncture. Due to such arguments, very little progress was being made in answering the crucial question: is the benefit which many patients experience after acupuncture caused by the treatment per se, or is it the result of a range of non-specific effects such as, for instance, the expectations of the practitioner, the patient or both?

To end this frustratingly fruitless dispute, we started a project to develop an acupuncture needle that would only appear to penetrate the skin while, in fact, it retracted into itself like a stage dagger, thus fooling the patient into believing that he was receiving real treatment while in fact he was not. After our PhD student had formally validated this method, we conducted a clinical trial with more than 100 patients who had suffered a stroke. Previous studies had found encouraging results suggesting that such patients would regain their bodily functions faster and more completely if they received acupuncture as an add-on to conventional rehabilitative therapy. If true, this would be an important finding: many of us suffer from a stroke during our later years. For most patients this is a devastating event from which recovery is often slow and incomplete. Any treatment that would improve this distressing situation would be more than welcome. Sadly, using our placebo needle, we found that the sham-acupuncture group in our trial recovered just as fast and as well as the patients who had received real acupuncture. We had little choice but to conclude that acupuncture is not superior to sham treatment, which must have been disappointing to enthusiasts of this treatment.

We had thus generated what is often called a “negative” result, i.e. a finding that failed to show that the tested intervention worked as claimed by its proponents. But the word “negative”, in colloquial parlance at least, has pejorative connotations. Why is it considered “‘negative” if medical experts and stroke patients discover that a treatment, in this case acupuncture, does not speed up recovery from a stroke? As I see it, far from being negative, it is actually the opposite, enabling everyone concerned to direct limited health resources towards treatments which demonstrably work instead of wasting money and time on those that don’t. Seen from this perspective, the notion of such a result being inherently negative is misleading. The term “negative” should perhaps be reserved for a clinical trial that proves to be untrue and misleads people, thus hindering progress.

Clinical trials have always been my favourite research tool. In alternative medicine, however, they tend to be both methodologically challenging and extremely expensive to conduct. Grants for medical research are always highly competitive but the shortage of funding is particularly marked in alternative medicine research. Eventually a lack of funds led to our having to reconsider clinical trials as our primary research approach. Whenever possible we did, of course, continue with clinical trials, but now we began honing our expertise in conducting systematic reviews, eventually publishing more than 300 such projects. This figure dwarfs the number of our clinical trials, and today we are internationally best known not for the 40 clinical trials but for the abundance of our systematic reviews.

* * *

Evaluating the effectiveness of alternative treatments is, of course, important - but not as important, in my view, as the investigation of their safety. I had already made this point unmistakably clear in the mission statement and I was determined to adhere to it.

Proponents of alternative medicine were often irritated by what they saw as unnecessary and potentially damaging lines of enquiry. They were quick to point out that many more patients are being harmed on a daily basis by mainstream pharmaceuticals than by alternative treatments. As a general rule, this is probably true, and I certainly never challenged the veracity of this assertion. What I vehemently disagreed with was the implication that, simply because pharmaceuticals often have serious side effects, the risks of alternative medicine should be spared scrutiny. This seemed to me to be a total non sequitur. If we do not systematically investigate alternative treatments, we cannot be sure exactly how safe they are, or indeed whether any safety issues exist. And if we fail to recognize existing problems, we cannot address them. As I see it, there is no more important issue in clinical medicine than patient safety, and for any treatment to be adopted based only on unproven assumptions is inherently unethical.

Moreover, the absolute risk of any treatment is in many ways not a useful concept for estimating the value of a therapy. Any responsible therapeutic decision needs to take into account not only the risks of a certain treatment but also its anticipated benefit. If the benefit of a given intervention is small, uncertain or non-existent, then even a relatively minor risk might tilt the risk–benefit balance into the negative. If a herbal remedy, for instance, shortens the symptoms of a common cold by a few hours but, at the same time, causes severe allergic reactions in a sizable percentage of patients, no responsible health care professional would recommend it for general use.

In this context, there was a further argument to consider, one that I had quietly hoped might get alternative practitioners more firmly on my side. If their assertions about the safety of alternative medicine were correct, my research would only confirm them. Surely, I thought, this must be an aim worth fighting for, and an argument that would get the support of the alternative medicine lobby. The logic seemed to me to be unassailable; but did it win over the alternative practitioners? The short answer, I am afraid, is no.

Over the years, I had to get used to an incessant drumbeat of hostile criticism, aggression and even personal attacks. Many people, regardless of their background, status, knowledge or expertise, felt they knew so much better than I did. This often bordered on the absurd: sometimes it seemed that a critic’s conviction in pressing the point was inversely proportional to his understanding of the issues at hand. I had studied my subject in depth, had earned two higher degrees, had years of clinical experience, hundreds of publications and even a dozen scientific awards to my name, yet none of this prevented others who had no qualifications, expertise or experience whatsoever from forcefully and often insultingly expressing their ill-informed opinions about research in general or my work in particular. A constant stream of letters, telephone calls and emails reminded me in no uncertain terms that I was an impostor, unqualified, wilfully wrong, misguided, stupid, corrupt and so on. At meetings with proponents or practitioners of alternative medicine, I was frequently left in no doubt that many people vehemently disagreed with me, and more than once the level of aggression became such that I felt concerned about my safety.

This was an entirely new experience for me, I have to admit, and something of a culture shock. In Vienna, I had been Herr Professor, someone whom nobody would dream of criticizing, at least not openly. On my becoming the Professor of Complementary Medicine at the University of Exeter, however, the climate had drastically changed. Now every Tom, Dick and Harry seemed to want to correct my concepts, reinterpret my results or question my background, attitude, integrity or intellect.

Initially, I thought this tsunami of criticism was quite refreshing in a perverse sort of way, coming as it did as an agreeable contrast to the stuffy atmosphere of Vienna. Since I was advocating critical thinking, there was, of course, no reason why I should be exempt from criticism myself. Indeed, in the beginning I was often able to see the funny side of it: for instance, getting vehemently attacked by homeopaths for not having any formal qualifications in homeopathy is hilarious considering that, in the UK (and in many other countries), no qualifications whatever are required to practise or research homeopathy.

For the first 10 years or so of my tenure at Exeter, I did my best to avoid open confrontation no matter how exasperated I was with all the mud-slinging. Later, though, when the criticism had turned into constant meddling and overt aggression, threats, and official complaints to my university peers, I slowly began to tire of it. To be taken seriously, criticism requires a basis in logic and fact; if it has neither, it does not deserve the title.

Eventually I began to stand up to the barrage, politely explaining the facts and the science behind them. I did this in publications, in lectures, in discussions, in newspapers, on the radio, on television and on the internet. And I did this not once or twice; I did it over and over again, dozens if not hundreds of times. Inevitably, though, about 15 years into my Exeter job, my patience wore thin and I began to feel that, at least on some occasions, it was probably best to show more determination or even teeth. I became more outspoken and often did not mince my words, but that only seemed to provoke more vitriolic attacks.

Slowly but surely I became resigned to the fact that, for some alternative medicine zealots, no amount of explanation would ever suffice. To them, alternative medicine seemed to have mutated into a religion, a cult whose central creed must be defended at all costs against the infidel.

If a researcher takes his job of critically evaluating alternative medicine really seriously, it is highly likely that he will be perceived as the great Satan by those who make their living from it. Conversely, if someone conducting research into alternative medicine is liked by the majority of alternative practitioners, he is almost certainly not doing his job correctly. Intriguingly, this concept is today known as “Ernst’s Law”.

* * *

In the late 1990s we carried out what was then the largest prospective study ever to investigate the safety of an alternative therapy. We persuaded 78 acupuncturists across the UK to report all adverse effects that occurred in a total of about 32,000 acupuncture sessions. While adverse effect surveillance is, of course, mandatorily and routinely performed in conventional medicine, there is no such oversight or monitoring in the world of alternative medicine; in fact, our project was a complete novelty in alternative medicine. Even though it was merely a research project and not an attempt to implement compulsory safety monitoring into everyday acupuncture practice (a task that even today has not been realized), it was nevertheless greeted with fierce resistance by the UK acupuncture establishment, and we had a major struggle keeping enough acupuncturists on board to see this project through to its end.

The results of our study, for the first time in the history of acupuncture, demonstrated beyond reasonable doubt that acupuncture, as practised in the UK by our study participants, was a relatively safe treatment. Only about 10 per cent of the patients enrolled in the study experienced adverse effects, and the nature of these problems was mostly mild. Several years later, I was involved in an even larger German study which confirmed that, in Germany, the safety profile of acupuncture is very similar to that in the UK. Today, we globally have data from several hundred thousand patients and, essentially, the message is still the same: acupuncture, as practised by Western-trained and experienced acupuncturists, causes mild adverse effects in approximately 10 per cent of patients, and serious complications are very rare.

The world of acupuncture should have been delighted. Not only had we completed a safety study that arguably should long since have been done by acupuncturists themselves, but we had also generated a reassuring bottom line. Yet most acupuncturists seemed to think that our project had been a waste of time: in their view, it only had alerted people to the fact that acupuncture was not entirely free of side effects.

Other investigations by our team and other researchers had shown that acupuncture could, in fact, occasionally cause very severe harm. These complications were caused mostly through needles introducing infection such as HIV or hepatitis B or C into the body, or injuring a vital organ (most frequently the lungs). Such complications could result in severe health problems, even deaths. So the somewhat confusing message was that acupuncture is relatively safe, but sometimes it can cause serious problems. The most likely explanation for this apparent contradiction is that the acupuncturists volunteering to participate in our prospective studies were mostly British and German doctors who were, of course, generally well-trained. To put it bluntly, they had learnt where the lungs are located within the body, and what the sterility of a needle involves; therefore they were not prone to puncturing a lung or infecting their patients with hepatitis. On the other hand, traditional acupuncturists, particularly those in Asian countries, were often completely devoid of any medical training or education. The conclusion to be drawn from this insight is clear and, I think, important: the safety of acupuncture depends on who administers the treatment and where the treatment is given. The average acupuncturist in rural China, for instance, might be less well educated and trained than his UK or US counterpart, and might therefore be more likely to use non-sterile needles or inadvertently puncture an internal organ. In other words, when considering the safety of an alternative treatment modality, we need to think not just of the therapy itself but also of the therapist, the setting and the wider context in which the treatment is administered.

The lessons from this line of research might seem obvious but they also are important: some alternative therapies are relatively safe; however, the same may not always be true for the practitioners who administer them. The systematic, careful and critical evaluation of the risks and the benefits of any treatment is neither unnecessarily alarmist nor of purely theoretical, academic interest. Far from it: it is prudent and necessary to establish exactly what risks might be involved, especially in a setting where there is no standardization of practitioner training and no regulatory oversight of the treatment. Far from being a vendetta motivated by animus against alternative medicine, investigating the risks of alternative therapies constitutes valuable research aimed at ensuring consumer safety.

As a team, we were proud to have constructively contributed to showing that acupuncture, as practised by UK doctors, does not significantly harm patients. This, I felt, represented an important step towards achieving exactly what my chair was created for - namely, to encourage the assimilation of appropriate treatments into mainstream clinical routine, in line with the dictum that there is no such thing as alternative medicine; there are just treatments that work, and those that don’t. Those that work will find their way into the standard armamentarium of medicine, while those that don’t are destined to remain in the realm of quackery.

Some alternative treatments seem intrinsically much more likely to cause serious harm than acupuncture. Spinal manipulation as practised by chiropractors, for instance, turned out to be a much more hazardous therapy. Data published by chiropractors had shown conclusively that about every second patient consulting a chiropractor would experience mild to moderate adverse effects, such as an increase in pain lasting normally 1–3 days. Chiropractors insisted that such reactions were a necessary first step on the path to getting better. But was this really true? The adverse effects could be strong enough to seriously interfere with the quality of life of the patient, and it seemed entirely legitimate - even necessary - to investigate whether these risks might outweigh any possible benefit of spinal manipulation.

Imagine an equivalent scenario in the arena of drug therapy: a medication of doubtful effectiveness causing mild to moderate adverse effects for two or three days in every second patient. Would we recommend such a drug? With an adverse event profile like this, it would surely be quickly withdrawn from the market, or at the very least be compelled to carry a prominent warning to prescribers and patients alike.

In addition to these relatively minor, transient problems, hundreds of very serious complications have, over the years, been documented following chiropractic spinal manipulation. The most serious problems involved damage to arteries that supply the brain. Such damage is technically called an “arterial dissection” and can result in a stroke or even death. The explanation for these disasters is quite straightforward: extreme rotation and extension of the neck may overstretch an artery running alongside the upper spine, causing the wall of the blood vessel to rupture, resulting in a stroke.

Whenever I plucked up the courage to publish on this important subject, the chiropractic fraternity, not just in the UK but also across the globe, reacted with fury. Instead of engaging in a scientific debate, they filed complaints with my peers and tried to silence me in every way they could think of. Instead of looking at the evidence impartially, they flatly denied that neck manipulation might cause a stroke. Instead of conducting proper research to fill in the existing knowledge gaps, they ran ad hominem attacks against me and claimed my research was flawed. Instead of recognizing that I was merely doing what a responsible scientist in my position had to do, they claimed I was waging a war against the chiropractic profession.

Alternative practitioners in general remained deeply sceptical that research into the risks and safety of alternative medicine was a crucially important contribution to the evidence. Instead they argued that my research focus on risks amounted to an obsessive personal crusade against alternative medicine; that my conclusions were deliberately alarmist; that I was simply trying to deflect attention away from the real dangers, which were to be found exclusively among the conventional drugs of mainstream medicine; that my research was in any case flawed; that I had an axe to grind; or that I was in the pocket of Big Pharma. When our research indicated quite clearly that the reported risks were only the tip of a much bigger iceberg, alternative medicine extremists regularly redoubled their attacks, on occasion becoming openly threatening.

Things went from bad to worse when I started to investigate what I called the indirect risks of alternative medicine. Even if one particular alternative therapy was entirely devoid of risk, I argued, we still have to consider its indirect risks, by far the most important of which is the possibility that a patient will avoid an effective mainstream treatment in favour of an ineffective alternative option.

The point here is essentially very simple: even the most ostensibly harmless intervention becomes positively life-threatening if it is employed instead of an effective therapy for a serious illness. The very term “alternative medicine” illustrates this potential danger well, implying, as it does, that there is a fundamental equivalence, an interchangeableness, between standard and unconventional approaches. The extent of such substitution is, of course, difficult to investigate and reliable information is therefore scarce - with one important exception: the attitude of some alternative practitioners towards vaccination.

When Hahnemann developed homeopathy about 200 years ago, he was adamant that it should be used as an alternative to and a replacement for conventional medicine. He even went so far as to decry as “traitors” those practitioners who used homeopathy alongside mainstream treatment. In Hahnemann’s time this attitude might have made sense to some degree - many conventional therapies at that period were not just ineffective but actually more dangerous than the diseases they sought to cure. Today, this is no longer the case, and most homeopaths therefore use their remedies as a complement rather than an alternative to mainstream medicine - most, but by no means all. We only need to go on the internet to find innumerable statements claiming homeopathy to be effective for every serious and treatable condition imaginable - anything from AIDS to cancer, tuberculosis or cholera. Whenever a patient is persuaded by such statements and opts to employ homeopathy instead of effective interventions, serious harm is almost unavoidable.

In 1995, I raised for the first time the concern that some UK homeopaths were advising parents against immunizing their children, claiming that their homeopathic “nosodes” are better and safer. At the time, we had published a small survey of homeopaths listed in the Exeter Yellow Pages, demonstrating that none of the non-medically qualified homeopaths recommended immunizations. Such results were interesting but not necessarily generalizable to the wider national population because our study was far too small to draw conclusions regarding the extent of anti-vaccination sentiment across the UK. Nevertheless, we published our findings prominently in the hope that it would trigger some discussion and perhaps even spur the country’s non-medical homeopaths to reconsider their stance on the issue of vaccination.

In 2002, at the time of the MMR (measles, mumps and rubella vaccine) scare promoted by Andrew Wakefield’s now discredited “research”, we decided to broaden our initial pilot study to the national level. In order to accomplish this, we had obtained the email addresses of a sizable number of UK chiropractors and homeopaths, and sent a request to these practitioners. In it, a fictional mother, concerned about the conflicting press reports regarding the safety of the MMR vaccine, asked for advice on MMR vaccination for her one-year-old child. After the replies had come in, we wrote to each respondent again and explained that the query had not been genuine but was, in fact, part of a research project. At the same time, we offered all participants the chance to withdraw their answers. The study had, of course, been approved in advance by our local ethics committee, and debriefing the respondents in this way was part of the approved protocol.

In total, we managed to contact 168 homeopaths, of whom 72 per cent responded, and 26 per cent withdrew their answers after the debriefing. We also contacted 63 chiropractors, of whom 44 per cent responded initially and 27 per cent later withdrew their responses. Our analyses showed that very few homeopaths and only a quarter of the chiropractors would advise their patients in favour of the MMR vaccination. Almost half of the homeopaths and nearly a fifth of the chiropractors would recommend against immunizing.

What happened next is amazing and perhaps even unprecedented in the recent history of medical research. This seemingly innocent and insignificant research project almost cost me my job. After receiving the debriefing email, several homeopaths complained to my university peers and to our ethics committee, claiming the research was unethical because it had been conducted on non-consenting human research subjects (the homeopaths and chiropractors) who had been misled about the nature of the enquiry. In view of these complaints, our ethics committee got cold feet and took the most remarkable step of withdrawing their previous approval; not only that, they forbade us to use the results in any way.

However, at this stage of the project I had already submitted our findings as a short report for publication in the British Medical Journal, and I flatly refused to comply with those ridiculous demands. The article was thus published only days after this storm had started blowing.

My Exeter peers were not amused by my disobedience, decided to conduct an official investigation and ordered me to attend several interrogatory sessions. For several weeks, I thought they might find me guilty of conducting unethical research and condemn my actions which, in the worst-case scenario, could have meant disciplinary action against me. Even the mildest reprimand would have been devastating to the credibility of my research team as a whole. The homeopaths who had filed the complaint were only waiting to use such news to discredit me once and for all. Fortunately, after several highly unpleasant exchanges, I managed to convince my peers that, considering the lively public and medical debate about the risks or benefits of the MMR vaccination, a swift publication of these findings had been in the public’s best interest. Eventually, it was decided that no disciplinary action, not even a reprimand, was called for.

I felt strongly then - and I still do today - that our research was not in the least unethical. In order to discover the extent and the effects of irresponsible behaviour, researchers, like police, sometimes have no choice but to conduct undercover investigations of this nature; and, whenever necessary, I continued doing so.

In a similar project, I collaborated with a London-based journalist who, pretending to be a cancer patient, visited five practitioners who offered alternative cancer cures. He personally made appointments, supplied all the necessary details and asked for advice regarding treatment of his highly treatable, and curable, form of lymphatic cancer. In the course of these five consultations, he received recommendations for a myriad unproven treatments. Had he followed the advice of these charlatans, some of whom were medical doctors, he would almost certainly have lost his life - not to mention being relieved of at least £26,000 on the way out. Studies of this nature may seem unfair and threatening to the practitioners concerned, but they are nevertheless immensely important from the perspective of public health; in fact, they can save lives.

Without a doubt, many patients choose to follow advice from irresponsible promoters of alternative medicine and forsake standard medical care entirely. Others may start by consulting alternative practitioners, and only much later, after their disease has failed to respond to quack remedies, seek proper medical care. Avoidance or delays of this kind can also be fatal, allowing a serious disease to progress beyond the point of no return.

Years after our undercover investigation into alternative practitioners’ attitudes towards immunization, my friend Simon Singh repeated a similar, but much more public, sting. He recruited a volunteer to consult several homeopaths requesting their advice regarding malaria protection. The results indicated that most homeopaths would not recommend effective preventive measures but instead advised patients to use “homeopathic immunization” for which there is not a shred of evidence. His investigation was subsequently replicated twice by the television BBC Newsnight programme, with essentially the same results.

The inability of alternative practitioners to learn from mistakes is a phenomenon that never ceases to astound and worry me; it implies that in ten years from now homeopaths and other dogmatists might still be endangering public health with the same scientifically unsound and unproven notions that have been their proud and unwavering mainstay for 200 years. And quite probably, when they are caught with egg on their faces, they will still pretend to be right and turn their fury on the researchers who exposed them.

All too often, the public is hoodwinked into believing that everything “natural” must also be safe. This familiar marketing strategy of alternative medicine may be good for business but it is most certainly very bad for public health. Similarly, consumers are led to believe that, whether or not they actually do good, alternative practitioners cannot do much harm. Even mainstream healthcare professionals who do not necessarily accept every nonsensical claim made by alternative medicine enthusiasts tend to trivialize the issue by saying these therapists might be a bit odd but are essentially harmless.

The indisputable fact, however, is that recommendations to forgo effective mainstream medical treatments in favour of bogus cures and quackery does cost lives - not just occasionally but regularly. It is obvious why this message is not popular with those who promote alternative medicine and earn a comfortable living from it. To me, however, it is equally clear that this message needs to be told and re-told, over and over again, in order to minimize the harm that charlatans inflict on the unsuspecting and often all too trusting public.

Over the years, our team conducted research into most aspects of alternative medicine. While our principal focus had always been the safety and effectiveness of alternative therapies, we also tackled a range of other topics in this area: for instance, we investigated the reasons why consumers and patients elect to use alternative medicine; we researched how many patients are tempted to go down this route; we asked conventional practitioners about their attitude towards and knowledge of alternative medicine. As the unit grew, so did our output of published articles, books and lectures. Our reputation attracted many foreign researchers who travelled to Exeter to join the team for a few months or years. This broadened our perspective and led to a multitude of international collaborations. In turn, this activity further increased our productivity and strengthened our international position as a centre of excellence.

At least two independent assessments identified my unit as the most productive alternative medicine research team in the world. Later, some particularly mendacious commentators would claim that the sheer number of our published papers - today they amount to well over 1000 - was in itself proof of fraudulent activity. It goes without saying that this is not true. The team was highly motivated; we enjoyed our work, as a unit, and we were both very industrious and highly successful. That such success attracts envy is perhaps understandable. That it also attracts defamation is, I think, an inevitable result of the fact that many of our findings challenged many of the treasured beliefs of those for whom alternative medicine functions more as a religion than as method of health care.