Should doctors recommend homeopathy?BMJ 2015; 351 doi: https://doi.org/10.1136/bmj.h3735 (Published 14 July 2015) Cite this as: BMJ 2015;351:h3735
- Peter Fisher, director of research, Royal London Hospital for Integrated Medicine, London WC1N 3HR,
- Edzard Ernst, emeritus professor, University of Exeter
- Correspondence to: P Fisher , E Ernst
Of all the major forms of complementary medicine, homeopathy is the most misunderstood. Based on the concept of “treating like with like,” homeopathy originated with the German physician Samuel Hahnemann (1755-1843). But similar ideas are found in the Hippocratic Corpus, in the work of Paracelsus, and in the medical traditions of several Asian countries.
Homeopathy is part of a family of toxicological and pharmacological phenomena that are attracting growing interest, characterised by secondary, reverse, or paradoxical reactions to drugs or toxins as a function of dose or time or both. These include hormesis (the paradoxical, stimulatory, or beneficial effect of low doses of toxins), paradoxical pharmacology, and rebound effects.
The controversial element of homeopathy is that some medicines are highly dilute, including “ultra-molecular” dilutions, in which it is highly unlikely that any of the original material is present. This is a major scientific concern and the source of the view that homeopathy “doesn’t work because it can’t work.”
However, recent in vitro research shows repeatable effects (for instance, inhibition of basophil degranulation by highly dilute histamine1) while basic physical research shows that the homeopathic manufacturing process changes the structure of the diluent, including the formation of nanoparticles of silica and gas.2 The physical research is of little clinical relevance but provides a possible mechanism of action for the controversial high dilutions.
Systematic reviews and meta-analyses
A recent review by the Australian National Health and Medical Research Council concluded that “there are no health conditions for which there is reliable evidence that homeopathy is effective.”3 But this report used unusual methods of analysis: the reviewers assumed that a positive trial showing a homeopathic treatment to be effective was negated by a different trial showing a different homeopathic treatment for the same condition to be ineffective. But the fact that one homeopathic treatment for a condition is ineffective doesn’t mean that another is ineffective. In addition, several key meta-analyses were unaccountably omitted.4 5 6
A meta-analysis published in the Lancet in 2005 concluded that there is “weak evidence for a specific effect of homeopathic remedies,” based on the results of just eight trials.7 Remarkably, this meta-analysis gave no hint of the identity of these eight trials, making informed discussion impossible.
Both these reviews are out of line with the other three systematic reviews and meta-analyses of homeopathy for all conditions published in the peer reviewed literature,8 9 10 all of which have come to essentially positive conclusions, as have several systematic reviews and meta-analyses for specific conditions.4 5 6 11
The negative publications have influenced medical and scientific opinion but had little effect on the public popularity of homeopathy. More recent overviews have had more favourable conclusions, including a health technology assessment commissioned by the Swiss federal government that concluded that homeopathy is “probably” effective for upper respiratory tract infections and allergies.12
Readers interested in the clinical evidence can access the CORE-HOM database of clinical research in homeopathy free of charge (www.carstens-stiftung.de/core-hom). It includes 1117 clinical trials of homeopathy, of which about 300 are randomised controlled trials.
Comparative effectiveness research
For practical decisions about homeopathy the most relevant evidence is observational comparative effectiveness research examining effectiveness in real world situations, which the Australian review did not include. Several such non-randomised studies have compared outcomes in patients attending family physicians who do and do not integrate homeopathy into their practice.
The multinational comparative effectiveness studies IIPCOS (International Integrative Primary Care Outcomes Studies) compared outcomes in primary care for patients with acute upper respiratory problems. IIPCOS-I included 456 patients treated by 30 doctors at six clinical sites in four countries (including the UK). At 14 days symptoms had resolved for 82.6% of homeopathic patients compared with 68% of patients receiving conventional treatment; the incidence of adverse events was 7.8% and 22.3%, respectively.13 The IIPCOS-II study included 1577 patients at 57 primary care sites (10 in the UK) in eight countries, with similar results.14
Studies in France and Germany show that GPs who integrate homeopathy in their practice have better outcomes than those who do not, for a range of conditions commonly treated in general practice; costs are equivalent and homeopathic GPs use fewer antimicrobial drugs.15 16
Integration of homeopathy improves patient outcomes
The largest comparative effectiveness study of homeopathy is the EPI3 study. This nationwide French study, coordinated by the department of pharmacoepidemiology at the University of Bordeaux, included 6379 patients, comparing treatment outcomes in musculoskeletal conditions, upper respiratory tract infection, sleep disorders, anxiety, and depression in terms of clinical benefit, medical care and medication, adverse effects, and loss of therapeutic opportunity.
The musculoskeletal cohort included 1153 patients; patients who chose homeopathy had healthier lifestyles, higher levels of education, and more motivation to self care but had more chronic disease than patients attending conventional physicians. Outcomes were similar between groups, but patients who attended homeopathic physicians consumed about half the amount of non-steroidal anti-inflammatory drugs.17 The upper respiratory tract infection cohort of EPI3 yielded a similar result, with reduced use of antibiotics.18
Unbiased decision making
Doctors should put aside bias based on the alleged implausibility of homeopathy. When integrated with standard care homeopathy is safe, popular with patients, improves clinical outcomes without increasing costs, and reduces the use of potentially hazardous drugs, including antimicrobials. Health professionals trained in homeopathy do not oppose the use of conventional treatments, including immunisation (www.facultyofhomeopathy.org/media/position-statements/immunisation). Doctors should recommend the use of homeopathy in an integrated manner.
No— Edzard Ernst
Homeopathy is “a therapeutic method using preparations of substances whose effects when administered to healthy subjects correspond to the manifestations of the disorder in the individual patient.”19 Nobody questions, of course, that some substances used in homeopathy, such as arsenic or strychnine, can be pharmacologically active, but homeopathic medicines are typically far too dilute to have any effect.
One of the most commercially successful remedies, for example, is based on an extract of duck liver in the C200 “potency,” which means it is diluted at a ratio of 1:10 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000. Beyond the C12 “potency” (that is, 12 dilutions at 1:100), the likelihood of a single “active” molecule being present in a homeopathic pill is effectively zero.
The assumptions underlying homeopathy fly in the face of science, and critics have long pointed out that, unless our understanding of the laws of nature is incorrect, homeopathy’s mode of action has no rational explanation.
Evidence fails to show efficacy
Numerous trials have tested the clinical efficacy of homeopathic remedies. Their results depend critically on the study design: uncontrolled studies almost invariably yield positive findings (for example, Spence and colleagues’ observational study20), whereas this is not true for the most rigorous of the 250 or so controlled clinical trials (such as a study in headache by Walach and colleagues21).
The explanation seems obvious: the perceived benefits of homeopathy are caused by non-specific effects. Once these are adequately controlled for in trials, the results tend to show that highly dilute homeopathic remedies are indistinguishable from placebos. Even a former consultant of the Royal London Homeopathic Hospital has agreed, writing, “The great majority . . . of the improvement that patients experience is due to non-specific causes . . . Homeopathy has not been proven to work.”22
To avoid cherry picking, it is advisable to evaluate the totality of the reliable evidence. Most independent systematic reviews of randomised controlled trials have failed to show that homeopathy is effective,23 and reviews with positive conclusions24 usually have serious methodological flaws.25
The most comprehensive, independent, and rigorous evaluation of homeopathy was published earlier this year by the Australian National Health and Medical Research Council.3 In line with many previous assessments, it concluded that “homeopathy should not be used to treat health conditions that are chronic, serious, or could become serious.”
Homeopathy can harm
As the typical homeopathic remedy is devoid of active molecules, it is unlikely to cause serious adverse effects. However, even a placebo can cause harm, if it replaces an effective therapy26; in the words of the Australian report: “People who choose homeopathy may put their health at risk if they reject or delay treatments for which there is good evidence for safety and effectiveness.”3
Yet homeopaths continue to advocate their treatments for many life threatening conditions; some even backed homeopathy for treating patients with Ebola. Others recommend their remedies as a replacement for conventional immunisations.27 With almost no systematic research into such issues, however, nobody can say how often they have caused actual harm to patients; anecdotally, however, I know of several deaths that have occurred in this unnecessary way. The ultramolecular homeopathic remedy might be harmless but the same cannot be said for all homeopaths.26
Because the risk-benefit balance of homeopathy is clearly not positive, even its use as “benign placebo” for self limiting conditions is problematic. In such cases, it would be preferable to reassure patients rather than to deceive them with placebos. In other instances, such an approach seems unnecessary: clinicians administering treatments that are effective with compassion and empathy will also generate a placebo response28—with the additional benefit of a specific therapeutic response. Finally, some claim that homeopathy might be helpful in cases where no specific treatment exists; this argument too is questionable and would apply only to the relatively few patients who cannot be helped even by a symptomatic treatment.
Costs and opportunity costs
In the European Union, the annual expenditure on homeopathic (and anthroposophic) remedies exceeds €1bn (£700m; $1bn).29 Spending by the NHS has declined sharply recently, in line with the evidence supporting homeopathy. The exact spend is unknown but is estimated at £3m-£5m, not including staff or infrastructure. These funds could and should be spent more usefully elsewhere. The notion that the NHS must provide homeopathy for patients who want it is disingenuous: patient choice is, of course, an important principle
, but the choice must be evidence based and should not be confused with arbitrariness.
In summary, the axioms of homeopathy are implausible, its benefits do not outweigh its risks, and its costs and opportunity costs are considerable. Therefore, it seems unreasonable, even unethical, for healthcare professionals to recommend its use.
Cite this as: BMJ 2015;351:h3735
Competing interests: Both authors have read and understood BMJ policy on declaration of interests and declare PF has acted as consultant to Heel and Similasan and given expert witness testimony on behalf of Standard/Hylands.
Provenance and peer review: Commissioned; externally peer reviewed.