Complementary medicine: the good the bad and the ugly
Honouring me with this year's HealthWatch Award seems a doubly courageous act, began Edzard Ernst, as he accepted the 2005 HealthWatch Award. It is courageous of HealthWatch as I am a researcher of the very subject this organisation often criticises, and it is courageous of me to accept the award as it is unlikely to result in praise from the proponents of complementary medicine (CM).
Courageous or not, it is definitely timely to put the spotlight on CM. Patients love it, the media and many people in power promote it, yet few people seem to understand it. In the following discussion I will try to highlight some of those aspects of CM which, I feel, are currently plagued by confusion, lack of transparency and sometimes even wilful deceit. Using the headings 'good, bad and ugly' inevitably requires a degree of simplification. In reality things are rarely black or white but different shades of gray.
It has always puzzled me how anyone could be for or against something like a medical intervention. Does it make sense to be in favour of appendectomy or anticoagulants? I don't think so! Why then do people hold emotional views on CM? It seems to me that, when it comes to healthcare, likes and dislikes should matter far less than evidence. Healthcare is not a fashion where one might legitimately have this or that opinion, nor should it be confused with religion in which one either believes or doesn't. Medical treatments either demonstrably and reproducibly work or they don't. Therefore reliable evidence on what is effective and safe must always be "good" - to view a trial of spiritual healing, homeopathy, for example, which fails to show that the tested intervention works (e.g. is better than placebo) as "negative" seems ludicrous to me.
Examples include the recent (first ever) trial of shark cartilage for cancer . Its results showed that it has no beneficial effects. Surely this must be good news all around. Sharks will not die needlessly, cancer patients will not attach false hopes to a bogus treatment, money can be saved for effective treatments. The only people who could possibly perceive this finding as "negative" are those involved in peddling bogus cancer cures and swindling desperate patients and their families of their savings.
Whenever we demonstrate that CM does work, the situation could quickly reverse. Examples for this scenario can also be found easily. Compelling evidence now suggests that real acupuncture is better than sham acupuncture for a range of pain-related syndromes, e.g. back pain . If the findings are based on good science, it must be good news: it could help millions who suffer from back pain, particularly as conventional medicine is not very successful in dealing with this problem. Similarly, there are now several systematic reviews of rigorous clinical trials demonstrating that certain herbal medicines are efficacious for certain indications  (see table 1, below). Making more general use of these options could benefit many patients - provided that the risks of these remedies do not outweigh the benefit.
It follows, I think, that finding the truth (arguably this is what science should be about) is always a good thing in medicine. As long as the results are reliable, they can only further our knowledge and will eventually improve healthcare. It also follows, I hope, that the incessant criticism directed at the work of my unit by enthusiasts of CM is based on a profound misunderstanding: we may have shown that this or that form of CM is not effective or not safe, but I fail to see that this was in any way negative for those who, in medicine, matter most: our patients.
In CM, many researchers seem to use science to prove that what they believe is correct. This is not what I was taught. Science is not for proving but for testing. The former approach does not only reveal an unprofessional attitude, it is prone to seriously mislead us all. Emotions and strong beliefs can lead to bias , and bias leads to bad science.
Sadly poor science is rife in CM. Here I could cite hundreds of examples. A recent study of anthroposophy  may suffice. Its aim was "to compare anthroposophic treatment to conventional treatment". Patients elected to consult either an anthroposophic or a conventional doctor. The results showed more favourable outcomes for the former approach and the authors concluded that "anthroposophic treatment... is safe and at least as effective as conventional treatment". Because of numerous sources of bias and confounding, many other conclusions are just as likely. The type of patients who elect to see an anthroposophic doctor may differ in many ways from patients who consult a conventional physician.
This example highlights much of what can be (and frequently is) wrong with CM research. It typifies how the aims of a study can be mismatched with the methodology and how the results may not justify the conclusions. If I had to name the characteristic that I find most disturbing in published CM research it would be this frequent inconsistency. Wishful thinking is of course only human, but the regularity of this incongruence in CM is nevertheless most remarkable.
What follows is, I believe, more than obvious: not only is good science good but bad science is bad. It is not bad because some 'out-of-touch' scientists in the 'ivory towers' think so. It is bad because it leads to wrong decisions in healthcare. Ultimately this will be detrimental to those who we should care for most: our patients.
The bad is bad enough, but the ugly is worse. I define ugly here as directly or indirectly preventing (future) patients from receiving the best available healthcare. I could lament about many aspects of CM that fall into this category: dishonesty, neglect of medical ethics, exploitation of vulnerable patients, political interventions are themes that come to my mind (see table 2, below).
The over-riding principle in all this is, I think, the application or promotion of one standard for conventional medicine and another for CM. Double standards are typified, I fear, in the new and increasingly popular movement (its proponents would probably say 'philosophy') of 'integrated medicine'. Its two basic tenets are that integrated medicine cares for the individual as a whole rather than looking at a diagnostic label ; and integrated medicine uses "the best of both worlds"6. Both claims look superficially convincing and plausible. At closer inspection they are, however, neither . Caring for the whole individual has always been and will always be a hallmark of any good medicine . It is thus not legitimate to adopt it as a main characteristic that differentiates 'integrated medicine' from conventional healthcare - on the contrary, conventional healthcare professionals who work towards optimising patient care must feel insulted by it. Using "the best of both worlds" (i.e. CM and mainstream healthcare) sounds fine until one realises how crucially it hinges on the definition of "best". In modern healthcare, this term can only describe those treatments that demonstrably and reproducibly do more good than harm. But this is precisely what evidence based medicine (EBM) is all about. Either 'integrated medicine' is synonymous with EBM (in which case the term would be redundant) or it applies a different standard for the term "best". Considering what 'integrative medicine' in the UK currently promotes  (see table 3), one has to conclude that the latter applies. This discloses integrative medicine as an elaborate smoke screen for adopting unproven treatments into routine healthcare . In the long run, this strategy will therefore turn out to be detrimental to everybody, including patients and even CM itself.
I am convinced that CM has much to offer. In the past 12 years, we have identified numerous CM interventions that generate more good than harm . Many more therapies need scientific testing and some of them will turn out to be useful. The only way to find out is to conduct rigorous research. Poor science will inevitably mislead us. And double standards are detrimental for everyone. In a nutshell, good science is good, bad science is bad and increasing the risk of patients not receiving the best available healthcare is ugly.
E Ernst, MD, PhD, FRCP, FRCPEd
Professor of Complementary Medicine
Peninsula Medical School, Universities of Exeter & Plymouth
Table 1: Systematic reviews suggesting efficacy of herbal medicines
Andrographis: Upper respiratory tract infection
Cranberry:Urinary tract infection
Devil's claw: Osteoarthritis, back pain
Ginkgo: Intermittent claudication, dementia
Ginger: Morning sickness
Hawthorn: Chronic heart failure
Horse chestnut: Chronic venous insufficiency
Kava: Anxiety, menopausal symptoms
Nettle: Benign prostatic hyperplasia
Peppermint: Abdominal pain, non ulcer dyspepsia, IBS
Saw palmetto: Benign prostatic hyperplasia
St John's Wort:Depression
Yohimbe: Erectile dysfunction
Data extracted from reference 3
Table 2: Preventing patients from receiving the best available healthcare
Administering unsafe treatments
- Asian herbal mixtures are sometimes contaminated with toxic heavy metals
- Upper spinal manipulation has been repeatedly linked to arterial dissection followed by stroke
Using invalid diagnostic techniques
- Iridology has been frequently tested and not found to be reliable.
- 'Live blood analysis' is used without evidence that it is valid
Not using CM that has been shown to do more good than harm
- Saw palmetto is effective and safe for BPH, but in the UK it is hardly used
- St John's wort is effective for depression, but in the UK it remains under-used
Misleading consumers through irresponsible advice
- Millions of web sites, hundreds of books, weekly columns in the print media, and even a UK government-sponsored patient-guide all fail to provide responsible advice
- The scarce research funds by the DoH were not used for studying efficacy and safety as recommended by the 'Lord's Report'. Despite the lack of reliable data, the 'Smallwood Inquiry 2005' recommended that large sums of money could be saved if more homoeopathy was used in the NHS.
Unethical behaviour in clinical practice
- A survey showed that the majority of UK chiropractors fail to adhere to their own ethical code (e.g. regarding informed consent).
- A 'Dr Foster' study demonstrated that many CM practitioners fail to comply with 5 very basic "best practice criteria".
Unethical behaviour in research
- Despite the wide-spread use of CM, funds for researching issues such as safety and efficacy of CM remain largely unavailable
Table 3. Selected statements from a recent (government-sponsored) patient guide (a)
...the risk of a stroke [after upper spinal manipulation] is between 1 and 3 in 1 million manipulations.
There are many published estimates that suggest much higher incidence figures. However, due to extreme under-reporting, the risk remains undefined.
Acupuncture is being increasingly used for people trying to overcome addictions...
A Cochrane review fails to demonstrate efficacy of acupuncture for this indication
Craniosacral therapists treat a wide range of conditions from acute to chronic health problems...
There is no trial evidence at all to suggest that craniosacral therapy is effective
Healing is used for a wide range of... conditions. Research has shown benefit in many areas, including healing of wounds, ... migraine or irritable bowel syndrome...
The best evidence available to date fails to demonstrate effects beyond a placebo response
"Homoeopathy is most often used to treat chronic conditions such as asthma"
A Cochrane review fails to demonstrate efficacy of homoeopathy for asthma
(a) Its aim was to "give [you] enough information to help you choose a complementary therapy that is right for you"
(b) The guide does not contain anything else by way of evidence on effectiveness (but was commissioned by the DoH to provide such evidence)
(c) Evidence extracted from reference 3
1. Loprinzi CL, Levitt R, Barton DL, Sloan JA, Ahterton PJ, Smith DJ et al. Evaluation of shark cartilage in patients with advanced cancer. Cancer 2005; 104: 176 - 82.
2. Manheimer E, White A, Berman B, Forys K, Ernst E. Meta-analysis: acupuncture for low back pain. Ann Intern Med 2004; 142: 651-63.
3. Ernst E, Boddy K, Pittler MH, Wider B. The desk top guide to complementary and alternative medicine. 2nd Edition. Edinburgh; Mosby. 2006.
4. Ernst E, Canter PH. Investigator bias and false positive findings in medical research. TRENDS in Pharmacological Sci 2003; 24: 219-21.
5. Hamre HJ, Fischer M, Heger M, Riley D, Haidvogl M, Baars E et al. Anthroposophic vs. conventional therapy of acute respiratory and ear infections: a prospective outcomes study. Wien Klin Wochenschr 2005; 117: 256-68.
6. Rees L, Weil A. Integrated medicine. BMJ 2001; 322: 119-20.
7. Ernst E. Disentangling integrative medicine. May Clin Proceed 2004; 79: 565-6.
8. Calman K. The profession of medicine. BMJ 1994; 309: 1140-3.
9. The Prince of Wales's Foundation for Integrated Health: Complementary Healthcare: a guide for patients. 2005.
10. Smallwood C (led by). The Role of Complementary and Alternative Medicine in the NHS. An investigation into the potential contribution of mainstream complementary therapies to healthcare in the UK, 2005.