|Home | About | Journals | Submit | Contact Us | Français|
I have probably heard them all—the often weird and wonderful arguments against applying the principles of science to complementary/alternative medicine (CAM)1. Specifically, the arguments against testing the efficacy of therapeutic approaches in CAM have worried me frequently and profoundly. Here I address eight arguments, all of which find expression in many different ways. I hope that the views expressed, even though doubtless biased, will stimulate constructive debate.
Understandably, this is a favourite argument of clinicians versus researchers. On the surface it looks patient-centred and politically correct. It may also be entirely correct when applied to the patients we try to help today. Clinicians must see their primary responsibility towards their present patients while researchers have a responsibility towards patients of the future. Thus researchers aim at refining or critically evaluating existing treatments, defining optimal treatments for given conditions and developing new therapies. The most efficient way of doing this is to conduct research.
If medicine as a whole had adhered to the above argument during the past three hundred years, we would still indulge in blood-letting and purging to the detriment of our patients. The principle of merely doing the best one can, while not advancing therapeutics for future healthcare, stands in the way of progress. I believe argument No. 1 to be a non-argument, perpetuated by failure to recognize the different roles and perspectives of clinicians and researchers.
This pertains to many (but not all) branches of CAM. Obviously, traditional use and experience can teach us important lessons. However, experience and science are items with different (not superior/inferior) values. Clinical medicine is founded on experience and to ignore it would be entirely foolish; few scientists would make this mistake. Unfortunately, experience can also be seriously misleading (the history of medicine is littered with examples). Patients can get better despite rather than because of our therapeutic interventions, and dogmatic adherence to experience has killed thousands.
Experience enables us to formulate hypotheses, but to test them requires rigorous science. As powerful as personal or collective experience often seems, it may lead us to draw wrong conclusions—the plural of anecdote is anecdotes, not data. Argument No. 2 prevents us recognizing the unique value of both experience and science. The latter quite simply cannot exist without the former, and they complement each other in the true sense of the word.
CAM claims to be holistic while science is accused of being reductionistic1. Clinical trials are, according to this argument, good for measuring what is measurable; but when we are dealing with the subtle effects of CAM, the clinical trial is thought to be inappropriate. When this argument is presented, we can reasonably ask what exactly the postulated subtle effects are. The answer may be ‘re-establishing a balance’ (e.g. yin and yang), or boosting some energy, or increasing wellbeing, or enhancing patient satisfaction, or stimulating the immune system. In other words, the problem that we face here is not one of trial methodology per se but one of outcome measures.
Once this has been recognized, a solution usually emerges. For some of the outcome measures, validated methods of quantification do exist (e.g. wellbeing or patient satisfaction). Others, such as re-establishing a balance, seem to present greater difficulty. The problem, however, is only apparent, not real. Let us ask how the patient (or the therapist) recognizes that the goal of re-establishing the balance has been achieved. ‘Well, the patient feels better’ is usually the answer. If this is the postulated effect, it too can be quantified. If a validated method of measurement does not exist, a method can (and should) be developed. Thus, argument No. 3 can invariably be shown to be wrong.
This argument neglects several facts at once1. First, the methodology of single case (n of 1) trials is established and is well suited to individualized approaches2. Such studies can be conducted with all the scientific rigour of other clinical trials; for instance, they can adopt randomization, double-blinding or placebo controls2. Secondly, the argument merely means that each individual requires a treatment that differs from that of the next patient. This can be accommodated through simple modifications of the standard clinical trial. Examples are abundant in homeopathy3. While these usually only allow individualization within a single diagnostic category, one could go a step further and abandon diagnostic categories completely4.
The claim that ‘my holistic approach helps patients’, which underlines argument No. 4, can invariably be reformulated as the basis of a testable hypothesis. This argument arises simply from failure to recognize the adaptability of clinical trial methodology.
This (technical) argument is correct and false at the same time. True, CAM comprises many treatment modalities to which the first part of the argument applies—e.g. massage, aromatherapy, hypnotherapy. The same is true for large parts of orthodox medicine—surgery, psychotherapy. And this is precisely the reason why the placebo-controlled, double-blind trial is not the ‘gold standard’ for efficacy testing of therapeutic interventions.
Placebo controls and blinding reduce bias and are thus desirable features. However, in many areas of clinical medicine the desirable does not coincide with the achievable. The gold standard is therefore the randomized clinical trial (RCT)6, which is possible for all complementary/alternative treatments. For all of the above-mentioned CAM modalities, RCTs have been published7. Thus argument No. 5 is built on misunderstandings about the desirable and achievable rigour in clinical trials and about the gold standard for efficacy testing of therapeutic interventions.
The proposal here is that the treatment lacks therapeutic effects and a short-term RCT will yield a false-negative result. A clinical trial quantifying long-term benefits may be not feasible; who, for instance, would pay for a trial of ginseng to prevent cancer if it required a 20-year treatment phase? Yet the hypothesis that ‘my therapy prevents illness’ is testable. One could, for example, consider doing an epidemiological case—control study: patients who have used ginseng for a long time could be compared in terms of say, development of cancer, with non-users; or individuals with cancer could be compared with healthy people in terms of their use of ginseng. Such investigations have been done9. Thus No. 6 is not a persuasive argument against conducting rigorous research. If the highest scientific rigour in efficacy testing is unobtainable, this is no valid argument for no rigour at all.
It is not easy to find counter-arguments that might convince people who believe in the above statement. In my view it implies that CAM is not really medicine, and if so doctors are indeed not competent to deal with it (if it is religion we should hand it over to priests).
This argument closes the circle and goes straight back to argument No. 1—politically correct, patient-centred and disarmingly convincing at first sight. On closer inspection, it turns out to be ill-conceived and regressive. If patients are helped by treatment X, which is entirely devoid of risks but not better than a placebo, these patients are obviously benefiting from a placebo effect. And there is nothing wrong with that. However, instead of perpetuating the myth that treatment X has specific effects, we should be honest and endeavour to understand its non-specific (placebo) effects. The aim here is to arrive at a state of knowledge and know-how about placebo-effects where not just patients using treatment X profit from the power of placebo but all patients seeking medical treatment under similar conditions. Argument No. 8 is ill-conceived because it prevents us from capitalizing on non-specific effects and thus from improving patient care on a broader scale.
The above is a highly personal statement. I look forward to hearing arguments that I have failed to mention—and views that differ from mine—on the sense and non-sense of applying the rules of science to CAM.