Trick or Treatment (10 page)

Of course, all of the above statements refer to the average patient, because the actual placebo effect for a particular patient depends entirely on the belief system and personal experiences of that individual. This variability of placebo effect among patients, and its potentially powerful influence on recovery, means that it can be a highly misleading factor when it comes to assessing the true efficacy of a treatment. In fact, the placebo effect is so unpredictable that it could easily distort the results of a clinical trial. Therefore, in order to test the true value of acupuncture (and medicines in general), researchers somehow needed to take into account the quirky, erratic and sometimes strong influence of the placebo effect. They would succeed in this endeavour by developing an almost foolproof form of the clinical trial.

The blind leading the double-blind

The simplest form of clinical trial involves a group of patients who receive a new treatment being compared against a group of similar patients who receive no treatment. Ideally there should be a large number of patients in each group and they should be randomly assigned. If the treated group then shows more signs of recovery on average than the untreated control group, then the new treatment is having a real impact…or is it?

We must now also consider the possibility that a treatment might have appeared to be effective in the trial, but only because of the placebo effect. In other words, the group of patients being actively treated might expect to recover simply because they are receiving some form of medical intervention, thus stimulating a beneficial placebo response. Hence, the straightforward trial design can produce misleading results, because even a useless treatment can give positive results in such a trial. So the question arises: how do we design a clinical trial that takes into account the confusion caused by the placebo effect?

A solution can be traced back to eighteenth-century France and the extraordinary claims of Franz Mesmer. Whilst Mesmer is nowadays associated with hypnotism (or mesmerism), in his own lifetime he was most famous for promoting the health benefits of magnetism. He argued that he could cure patients of many illnesses by manipulating their ‘animal magnetism’, and one of the ways of doing this was to give them magnetically treated water. The remedy was very dramatic, because sometimes the supposedly magnetized water could induce fits or fainting as part of the alleged healing process. Critics, however, doubted that water could be magnetized and they were also dubious about the notion that magnetism could affect human health. They suspected that the reactions of Mesmer’s patients were purely based on their faith in his claims. In modern parlance, critics were suggesting that Mesmer’s remedies were exploiting the placebo effect.

In 1785, Louis XVI convened a Royal Commission to test Mesmer’s claims. This Commission, which included Benjamin Franklin, conducted a series of experiments in which one mesmerized glass of water was placed among four glasses of plain water – all five glasses looked identical. Unaware which glass was which, volunteers then randomly picked one glass of water and drank it. In one case, a female patient tasted her glass and immediately fainted, but it was then revealed that she had drunk only plain water. It seemed obvious that the fainting woman thought that she was drinking magnetized water, she knew what was supposed to happen when people drank such water, and her body responded appropriately.

After all the experiments had been completed, the Royal Commission could see that patients had responded in a similar way regardless of whether the water was plain or magnetized. Therefore, they concluded that magnetized water was the same as plain water, which meant that the term magnetized water did not really mean anything. Moreover, the Commission stated that the effect of supposedly magnetized water was due to the expectation of patients; today we would say that it was due to the placebo effect. In short, the Commission accused Mesmer’s therapy of being fraudulent.

The Royal Commission did not, however, speculate about the widespread effects of placebo throughout medicine, which is why Haygarth’s research on tractors fourteen years later is credited with formally recognizing the role of the placebo effect in medical practice. On the other hand, the Royal Commission did make a major contribution to the history of medicine, because it had designed a new type of clinical trial. The key breakthrough in the Royal Commission’s experiment was that the patients were unaware of whether or not they were receiving the real or fake treatment, because the glasses of mesmerized water and plain water were identical. The patients were said to be
blind
.

The concept of blinding can be applied to entire trials, which are known as
blinded clinical trials
. For example, if a new pill is being tested then it is given to all the patients in the treatment group, while a pill that looks the same but without any active ingredient is given to the control group. Importantly, patients have no idea if they are in the treatment or control group, so they remain blind as to whether or not they are being treated. It is quite possible that both groups will show signs of improvement if both respond to the placebo effect caused by the possibility of receiving the real pill. However, the treatment group should show greater signs of improvement than the control group if the real pill has a genuine effect beyond placebo.

In a blinded trial, it is crucial that both the control group and the treatment group are treated in similar ways, because any variation can potentially affect the recovery of patients and bias the results of the trial. Therefore, as well as receiving pills that look the same, patients in both groups should also be treated in the same location, be given the same level of attention and so on. All these factors can contribute to so-called
non-specific effects
– namely effects resulting from the context of the treatment process, but which are not directly due to the treatment itself. Non-specific effects is the umbrella term that also covers the placebo effect.

It is even necessary to monitor patients in both groups in exactly the same way, because it has been shown that the act of close monitoring can lead to a generally positive change in a person’s health or performance. This is known as the Hawthorne effect, a term that was coined after researchers visited the Hawthorne Plant in Illinois, part of the Western Electric Company. The researchers wanted to see how the working environment affected the plant’s output, so between 1927 and 1932 they increased artificial illumination and then reduced it, they increased room temperature and then reduced it, and so on. The researchers were amazed to find that any change seemed to cause an improvement. This was partly because workers expected that the changes were supposed to bring about improvements, and partly because they knew they were being monitored by experts with clipboards. It is difficult to remove the Hawthorne effect in any medical trial, but at least the effect should be the same for both the treatment group and the control group so that a fair comparison can be made.

Creating identical conditions for the control and the treatment groups effectively blinds the patients to whether or not they are receiving the treatment or the placebo. Yet it is also important to blind whoever is administering the treatment or the placebo. In other words, even the doctors treating the patients should not be aware of whether they are giving a sugar pill or an active pill. This is because a doctor’s demeanour, enthusiasm and tone of voice can all be affected by knowing that he or she is administering a placebo, which means that the doctor might unconsciously give hints to patients that the medicine is merely a placebo. Such leaking of information, of course, can jeopardize the blinding of the patient and the overall reliability of the clinical trial. The consequence would be that patients in the placebo control group would suspect that they were receiving a placebo and would then fail to exhibit a placebo response. Perversely, patients receiving the real treatment would have no such qualms and would exhibit a placebo response. Hence, the trial would be unfair.

If, however, both the patient and the doctor are unaware of whether a placebo or a supposedly active treatment is being administered, then the trial results cannot be influenced by the expectation of either. This type of truly fair trial is said to be
double-blind
. Including some of the points made in Chapter 1, we can now see that a well-conducted trial ideally requires several key features:

1. A comparison between a control group and a group receiving the treatment being tested.
   
2. A sufficiently large number of patients in each group.
   
3. Random assignment of patients to each group.
   
4. The administering of a placebo to the control group.
   
5. Identical conditions for the control and treatment groups.
   
6. Blinding patients so that they are unaware to which group they belong.
   
7. Blinding doctors so that they are unaware whether they are giving a real or a placebo treatment to each patient.
   

A trial that includes all these features is known as a randomized, placebo-controlled, double-blind clinical trial, and it is considered to be the highest possible standard of medical testing. Nowadays, the various national bodies responsible for authorizing new treatments will usually make their decisions based on the results obtained from such studies.

Sometimes, however, it is necessary to conduct trials that are closely related to this format, but which do not involve a placebo. For example, imagine that scientists want to test a new drug for a condition that is already treated with a partly effective existing drug. Point 3 indicates that the control group receives only a placebo, but this would be unethical if it deprived patients of the partly effective drug. In this situation, the control group would receive the existing drug and the outcome would be compared against the other group receiving the new drug – the trial would not be placebo-controlled, but there would still be a control, namely the existing drug. Such a trial should still adhere to all the other requirements, such as randomization and double-blinding.

These sorts of clinical trials are invaluable when conducting medical research. Although the results from other types of trial and other evidence might be considered, they are generally deemed to be less convincing when it comes to the key question: is a treatment effective for a particular condition?

Returning to acupuncture, we can re-examine the clinical trials of the 1970s and 1980s – were these trials of high quality and were they properly blinded, or is it possible that the reported benefits of acupuncture were due merely to the placebo effect?

A good example of the type of acupuncture trial that took place during this period was one conducted in 1982 by Dr Richard Coan and his team, who wanted to examine whether or not acupuncture was effective for neck pain. His treatment group consisted of fifteen patients who received acupuncture, while his control group consisted of another fifteen patients who remained on a waiting list. The results would have seemed unequivocal to fans of acupuncture, because 80 per cent of patients in the acupuncture group reported an improvement, compared to only 13 per cent of the control group. The extent of the pain relief in the acupuncture group was so great that they halved their intake of painkillers, whereas the control group reduced their intake of pills by only one tenth.

Comparing the acupuncture group against the control group shows that the improvement due to acupuncture is much greater than can be explained by any natural recovery. However, was the benefit from acupuncture due to psychological or physiological factors or a mix of the two? Did the acupuncture trigger a genuine healing mechanism, or did it merely stimulate a placebo response? The latter possibility has to be treated seriously, because acupuncture has many of the attributes that would make it an ideal placebo treatment: needles, mild pain, the slightly invasive nature, exoticism, a basis in ancient wisdom and fantastic press coverage.

So Dr Coan’s clinical trial, along with many of the others conducted in the 1970s and 1980s, suffered from the problem that they could not determine whether acupuncture was offering a real benefit or merely a placebo benefit. The ideal way to find out whether acupuncture was genuinely effective would have been to give a placebo to the control group, something that seemed identical to acupuncture but which was totally inert. Unfortunately, finding such a placebo proved difficult – how can you create a therapy that appears to be acupuncture but which is not actually acupuncture? How do you blind patients to whether or not they are receiving acupuncture?

Placebo control groups are easy to arrange in the context of conventional drug trials, because the treatment group can, say, receive a pill with the active ingredient and the placebo control group can receive an identical-looking pill without the active ingredient. Or the treatment group can receive an injection of the active drug and the placebo control group can receive an injection of saline. Unfortunately, there was no similarly obvious placebo replacement for acupuncture.

Gradually, however, researchers began to realize that there were two ways of making patients believe that they were receiving real acupuncture, when they were in fact receiving fake acupuncture. One option was to needle patients to only a minimal depth, as opposed to the centimetre or more that most practitioners would use. The purpose of this superficial needling was that it seemed like the real thing to patients who had not previously experienced genuine acupuncture, but according to the Chinese theory it should have no medical benefit because the needles would not reach the meridian. Therefore researchers proposed studies in which a control group would receive superficial needling, while a treatment group would receive real acupuncture. Both groups would receive similar levels of placebo benefit, but if real acupuncture has a real physiological effect then the treatment group should receive a significant extra benefit beyond that received by the control group.