Tests of the efficacy of homeopathy

Tests of the efficacy of homeopathy have taken many different forms, with inconsistent outcomes.

Some studies of homeopathy have been funded by the U.S. National Center for Complementary and Alternative Medicine (NCCAM). In NCCAM's taxonomy of complementary and alternative medicine, homeopathy is a "whole system" that is an alternative to the entire medical mode, although it can be complementary to conventional care. The funded studies include:


 * A study on fibromyalgia, which showed clinical benefits from individually chosen homeopathic remedies as well as objective differences in EEG readings in homeopathic and placebo subjects.


 * A study on homeopathy for mild traumatic brain injury. This pilot study indicate a significant benefit from the homeopathic treatment but requires large-scale, independent replication.

Cochrane Reviews did a meta-analysis of seven randomized controlled trials, three prevention trials (number of participants, $$N = 2265$$) and four treatment trials, $$(N = 1194)$$. The authors considered only two were statistically adequate, but data from the prevention trials showed no effect. Data from the treatment trials showed enough efficacy that further trials were recommended, but that the remedy could not be recommended for first-line therapy. No homeopathic counterarguments to these specific meta-analyses have been published, but homeopaths often consider large trials invalid because they do not reflect the individualized nature of homeopathic treatment.

In vitro
A systematic review of the in vitro evidence of homeopathic high potencies evaluated 67 experiments (mostly performed on basophils) in 75 publications (1/3 of them replications. Nearly 3/4 of them found a high potency effect, including 12 of 18 studies that scored 6 points or more for study quality and which controlled for contamination. Nearly 3/4 of all replications were positive. The authors concluded that even experiments with a high methodological standard could demonstrate an effect of high potencies, but no positive result was stable enough to be reproduced by all investigators.

In vivo
A meta-analysis of 105 animal trials evaluated the use of homeopathic remedies in the treatment of environmental toxicology. Of these, 28 studies were considered high quality, and positive effects from homeopathic doses were reported 50% more often than no effects. The authors also noted, "experiments using the 'high' dilution range (by our definition those preparations diluted beyond the presence of original substance) had higher quality evaluations than experiments in lower dilution ranges, making their validity more likely. Second, our reevaluation of results from these studies using the raw data showed that over 70% had positive effects."

One group of researchers tested the effects of heavy metals on mice who were given homeopathic doses of these toxic substances after exposure. Using Arsenicum album (arsenic oxide) $$6C, 30C$$ and/or $$200C$$ to treat mice who were exposed to crude doses of arsenic, they found that it significantly reduced the cytotoxic effects of arsenic, including the levels of two liver enzymes which are indicators of liver toxicity and both of which are increased as a result of arsenic poisoning.

Randomized controlled trials in humans
Homeopaths assert that homeopathic remedies generally have to be individually prescribed to a sick person based on the totality of symptoms, not just the disease that he/she has been diagnosed with; they therefore argue that many clinical trials are inappropriate tests for homeopathic treatment, but acknowledge that there are sometimes exceptions to the need for individualization (the experience above with Oscillococcinum is one such exception).

Clinical trials are the "gold standard" for efficacy in mainstream medicine, often creating an impasse with homeopaths who do not find their paradigm compatible with large scale randomized controlled trials.

Some individual RCTs have shown no statistically significant effect, while others indicate a positive difference in people receiving homeopathic treatment. There are single RCTs in favor of homeopathy for a number of other conditions, including: respiratory allergies such as hayfever, asthma and perennial allergic rhinitis, acute otitis media, influenza, childhood diarrhea, symptoms of chronic obstructive pulmonary disease, arthritis, fibromyalgia,  vertigo, varicose veins ankle sprain, bronchitis, chronic fatigue and premenstrual syndrome.

Some trials have been performed that partially meet these criteria, and some of these have reported positive effects. These have not been considered as providing compelling evidence, partly because of deficiencies in trial design, but mainly because of the possibility of publication bias - the phenomenon whereby trials that happen by chance to appear to show a positive outcome are more likely to be published than those which are inconclusive or appear to show a negative outcome. Nevertheless, even if homeopaths distrust clinical trials, some have shown, in small groups, enough positive results to justify larger tests. Positive results came from Taylor and colleagues, writing on their placebo-controlled randomized controlled trial of a homeopathic treatment for allergic rhinitis, concluded " this study has failed to confirm our original hypothesis that homoeopathy is a placebo". Two invited commenters were published alongside, and found the study interesting, but of small size. They urged larger trials, and cited the challenge of doing "large trials that really could change thinking," citing work on influenza by Vickers et al. and Ferley et al., on muscle recovery, as having done large-scale trials. The muscle soreness trial showed no significant effect of homeopathic treatment; $$(N =400)$$ but the abstract did not break out the groups. Effects were measured by formal methods of the runners' subjective soreness.

The second trial, on influenza, used the objective endpoint of time to return to normal body temperature as the endpoint. They observed that more people in the active drug group $$(N=237)$$ than the control group $$(N=241)$$ had faster recovery $$(17.1% against 10.3%,  P=0.03).$$ While they could not explain the result, they called for "rigorously designed clinical studies."

Meta-analyses on humans
Some authors of meta-analyses have concluded that the placebo response is an inadequate explanation for the positive responses that have been reported in some trials. For example, one review assessed 105 trials, 81 of them positive; the authors concluded that the evidence on balance is that trials are positive but not sufficient to draw definite conclusions, and they said that “based on this evidence we would be ready to accept that homeopathy can be efficacious, if only the mechanism of action were more plausible”.

In another review, Linde and colleagues analyzed 89 trials and found a mean odds  ratio  of  2.45 (95% confidence interval, 2.05–2.93) in favor of homeopathy. When considering just those trials of “high quality” and after correcting for publication bias, the findings remained significant (means odds ratio of 1.86), however, the main conclusion was that the results “were not compatible with the hypothesis that the effects of homeopathy are completely due to placebo.” The authors also concluded that the higher quality trials were less likely to be positive than those of lower quality, saying  “There is increasing evidence that more rigorous trials tend to yield less optimistic results than trials with less precautions against bias.”

Cochrane meta-analysis of Oscillococcinum
Oscillococcinum is a commercially available homoeopathic remedy manufactured from wild duck heart and liver that is sold as a remedy for influenza, and which is particularly popular in France. A Cochrane review of seven placebo-controlled studies included: but only two of these provided adequate data for full analysis. Two of the trials were unpublished, according to the New Zealand Guidelines Group, which provides evidence-based reporting on complementary and alternative medicine.
 * three prevention trials (number of participants $$(N) = 2265$$)
 * four treatment trials $$(n = 1194)$$.

The review found "promising" evidence that Oscillococcinum is effective in the treatment of influenza and influenza-like illness, but it was not effective in the preventation of influenza and influenza-like symptoms ($$(relative risk (RR) 0.64, 95% confidence interval (CI) 0.28 to 1.43).$$). The review of research found evidence for a small effect on duration of illness (Treatment reduced the duration by 0.28 days $$(95% CI 0.50 to 0.06).$$). This positive result from this homeopathic medicine was felt to justify further research, but the review concluded that the data were not strong enough to make a general recommendation to use Oscillococcinum for first-line treatment of influenza.

The Lancet controversy
In 2005, the Lancet published a meta-analysis of 110 placebo-controlled homoeopathy trials and 110 matched conventional-medicine trials. A press report on the article, which described the Lancet study as performed by Swiss scientists, who concluded that the only effects of homeopathy were in the mind.

There was considerable criticism of the Lancet article from journals in complementary and alternative medicine. In an editorial referring to the Shang article published by The Lancet, the ''Journal of Alternative and Complementary Medicine asked, One cannot but wonder and inquire, in all seriousness, what drove the editors of one of the most highly respected and influential journals of our time to allow through peerreview and thence to publication, a manuscript that fails totally to provide the information necessary for full independent replication or analysis

While Shang was listed as the lead author, the work was done under the Swiss Complementary Medicine Evaluation Programme (Programm Evaluation Komplementärmedizin, PEK). PEK's review board has protested about political interference in the scientific process, according to Peter Fisher, Director of Research, Royal London Homoeopathic Hospital.

The outcome suggested that the clinical effects of homeopathic remedies might all be placebo effects. The Lancet study is notable as a "global" meta-analysis of homeopathy, not an analysis of particular remedies, i.e. it tested the hypothesis that all of the reported effects of homeopathic remedies are placebo effects. If so, then reports of positive effects reflect publication bias (the tendency to publish results when they show a positive effect but not when they are negative), and the magnitude of such effects should diminish with sample size and study quality. They analyzed an equal number of conventional medicine trials similarly; these showed a real effect of treatment, in that the size of the reported effect was independent of sample size, but the trials of homeopathy remedies did not. The study does not prove that homeopathy is never effective, but is consistent with the interpretation that all reported effects are placebo effects. The Lancet subsequently published a selection of critical correspondence.

Several of these published responses remarked (incorrectly) that the researchers evaluated only those studies that met certain criteria for “high quality” scientific investigations. Of the original 110 trials, 21 of the homeopathic studies fit this definition but only nine of the conventional studies did so. The researchers then considered only those trials that had a large number of patients, as trials with a larger sample size have greater statistical power. They thus compared eight homeopathic trials with six conventional medical trials, even though these trials were only matched as being of large size and good quality. The results of this subanalysis was that the effects observed in the homeopathic trials were no greater than the placebo effects observed in the clinical trials. They suggested that no further research on homeopathy is necessary, while advocates of homeopathy assert that almost all of the eight homeopathic trials used only a single homeopathic remedy given to every subject without the typical individualization of treatment that is commonly used in homeopathic practice.

Advocates of homeopathy assert that this study's emphasis on certain large homeopathic trials lacked external validity, that is, the trials were not consistent with the homeopathic methodology, thereby reducing the meaning of the study's conclusions. Some insight may be gained from a branch of mainstream medicine in which treatment must be far more individualized than any plausible drug trial: surgery. A trial lacks internal validity when the difference in the control and treatment groups cannot be explained only by the specific different treatment. Paradis gives a hypothetical example of a trial of wound irrigation with an antibiotic. The treatment group has irrigation from a prefilled, mechanically operated fluid dispenser, while the control group's irigation is done with a hand-operated syringe filled with saline. The potential differences are obvious; there can even be more differences in full surgery, due to the differences in anatomy, body fat, and other individual variations.

External validity is also called "research validation": can the results of controlled trials be seen in community practice? Purists will say that external validation cannot make up for a lack of internal validity; the external experience is essentially separate from the flawed trial. Arguably, many retrospective clinical studies are based purely on reviewing "external" cases, but an attempt is made to find cases with as much similarity as possible. From the perspective of mainstream medicine, it may appear as if homeopaths want to insist on as much individualization as possible, frustrating even retrospective techniques.

External validity, used with care, does guide surgical development; many number of surgical techniques seemed promising at first, but, with a larger data base, proved ineffective, while others gained more credibility. For there to be better understanding between homeopathy and mainstream medicine, there must be mutual willingness to converge on a model where a degree of individual treatment is accepted as part of the homeopathic approach, but the need to have some statistical rigor has to be accepted as part of the approach of evidence-based medicine.

Other Meta-analyses
Several meta-analyses evaluating the homeopathic treatment of specific diseases has also found positive results. These studied childhood diarrhea, respiratory allergies , and postoperative ileus.

In one meta-analysis of 6 trials $$(N=556)$$, there were six RCTs, which used different treatments that were difficult to form into a meaningful pool. The authors observed that they used standardized treatment, which probably differs from the homeopathic standard of practice. They observed "There has been only a limited attempt to measure a 'package of care' effect (i.e., the effect of the medication as well as the consultation, which is considered a vital part of individualised homeopathic practice)." The conclusion was that there was not enough useful data, and strongly recommended, in addition to RCTs, collecting "observational data to document the different methods of homeopathic prescribing and how patients respond. This will help to establish to what extent people respond to a 'package of care' rather than the homeopathic intervention alone." In conventional medicine, there have been comparisons where the actual treatment was the same, but the variable was the type of encounter (e.g., short or long, physician specialist vs. generalist, nurse-educator rather than physician).

Another study, conducted by an outspoken critic of homeopathy (Ernst) identified 17 useful articles. Six, however, re-analyzed the same major meta-analysis. This major study observed some differences between homeopathic treatment and placebo, but also could find no specific evidence of efficacy for any specific condition. The authors of this meta-analysis defined "evidence" in this context as having at least three independent researchers finding positive results from a specific treatment. Within one year of publication of this meta-analysis, a third independent trial was conducted testing Oscillococcinum in the treatment of influenza. Also, a third trial in the treatment of childhood diarrhea was published (the lead researcher for these trials were the same individual, but the prescribing homeopaths for each trial were different clinicians.) Even though these newer studies helped homeopathic research pass this threshold, the authors of the 1997 meta-analysis still concluded,  "Further research on homeopathy is warranted provided it is rigorous and systematic."

The meta-analysis of 17 papers concluded "Until more compelling results are available, homeopathy cannot be viewed as an evidence-based form of therapy," but strongly recommended formal study "individualized, empathetic and time-intensive approach most homeopaths adopt to healthcare yields good clinical results. This emphasizes the importance of the therapeutic encounter and is in accordance with a wealth of information in this area". In other words, there may be useful information, not being studied, about the methods by which patients interact with homeopath.

Are the right things being analyzed?
An observation from some authors was that a key element missed in many studies was a systematic observation of the the interaction between homeopath and patient, in contrast with the interaction between patient and conventional health care worker. In conventional medicine, there is concern within the U.S. healthcare economics area about even collecting minimal data on the quality of encounters. One policy study stressed the need not to understand the nature of the interaction, but even to get basic demographics on the encounter; especially in programs servicing low-income populations, there may be little more than billing data. There are a number of trial programs to use nurses and other providers, at a lower cost than physicians, to do case management of chronic disease. Case management interactions may be as long as some of those with homeopaths. It is not unreasonable to note that a number of nursing associations call for more holistic patient interactions than is typical for conventional medicine, and to wonder if there most important lessons from current homeopathy, which are not being studied.