The latest in an occasional series looks at homeopathy.
What is homeopathy?
It is a system of medicine based on the principle that ‘like cures like’. The German physician Samuel Hahnemann founded homeopathy in the 18th century after hearing claims that cinchona could cure malaria. Hahnemann took cinchona bark and decided that the symptoms he noted in the following weeks were similar to those associated with malaria.
He concluded that “that which can produce a set of symptoms in a healthy individual, can treat a sick individual who is manifesting a similar set of symptoms.” The principle of like curing like, and the discipline of homeopathy, came into being based on this conclusion. The ideas of Succussion and High Dilutions of remedies followed.
What are these high dilutions?
Homeopathic remedies are diluted in stages by a factor of 10 or 100. Remedies above 24X or 12C are unlikely to contain a single molecule of the active ingredient. While some homeopaths tend to prescribe so-called ‘lower potency’ remedies that are at or below 24X or 12C and may therefore contain a trace of the original active ingredient, others claim that higher dilutions are more potent – despite them not containing any trace of the active ingredient. To back up this claim, they invoke ‘the memory of water’.
What does the memory of water mean?
It has been claimed that the water in homeopathic remedies is somehow able to retain a memory of the original active ingredient that has been diluted out of existence. In the 1990s an experiment (using antibodies that were then diluted out of existence) was reported to have found that water had a ‘memory’ – but only when the experimenters were aware of which test tubes originally contained the antibodies and which did not. When the experiment was repeated with blinding, the results could not be replicated.
In 2005, researchers found that “liquid water essentially loses the memory of persistent correlations in its structure within 50 fs.” 50 fs means 50 femtoseconds – and a femtosecond is one thousandth of one billionth of one second one thousandth of one trillionth (or one millionth of one billionth) of a second*. This is rather less time than would be needed to take a homeopathic remedy after its manufacture.
Is homeopathy explained by quantum physics?
I doubt it. Some advocates have claimed that they can explain how homeopathy works using quantum theory. However, several bloggers have taken a closer look at some of those making these claims and examined the basis for the claims. Here, it is noted of a paper on homeopathy and quantum theory that “Having misunderstood and misrepresented quantum theory, Milgrom now goes on to do the same for weak quantum theory (WQT).”
Does homeopathy work?
No. There are several ways in which we can try to judge whether a treatment works. Prof Ernst described these methods in a talk he gave on alternative medicine. Ernst highlighted four ways of attempting to find out if something works: plausibility; the test of time; ask the patient; do the research.
Plausibility is a poor guide to the efficacy of a medical treatment. Blood-letting was (in its time) seen as a plausible treatment because of the belief that health was affected by the status of the four humours, yet it did not work**. Homeopathy is implausible (some would argue that efficacy in terms of the higher dilutions, which contain not a single molecule of the original ingredient, is impossible rather than implausible), but the implausibility of homeopathy is not the final nail in its coffin.
The test of time is also a poor guide – blood-letting was used for hundreds, if not thousands of years despite being ineffective. An appeal to tradition is a logical fallacy – there is nothing inherently good about an idea simply because it has been held for some time. Old ideas can be bad ideas and the longevity of an idea, whether old or new, gives us no clue as to its truth. That homeopathy has lasted for some time should not lead us to assume that it may be of use in treating patients.
Asking the patient is an unreliable way to discover the truth about a treatment. What Ernst refers to as the perceived therapeutic effect covers: the specific effect of therapy; true placebo-effect; clinician-patient interaction; the natural history of the condition; regression towards the mean; social desirability (e.g., the desire to please the person treating you may make you more inclined to claim a treatment worked); concomittent therapies; other context effects. While homeopathy may lack the first factor – a specific beneficial effect on the patient – some or all of the other factors will be present and will influence the patient’s view of the treatment.
We are then left with research. Homeopaths claim that their remedies will make you better. This is something that can be tested. Not only is it possible to conduct a double-blind, randomised, placebo-controlled trial on homeopathic remedies, it is possible to conduct such a trial on individualised homeopathy. There are some trials that show that homeopathy has an effect above placebo, but these tend not to be as well-designed as those that show no effect over placebo. Anyone can cherry-pick a single trial with a result that suits their argument, but if you look at the available literature and pick out the best-quality trials you will have a more reliable view of the evidence.
Fortunately, I don’t have to do this. Some scientists are willing to spend their time trawling through all the research into homeopathy and producing a systematic review of the literature. One such paper (Shang et al) can be found here.
Biases are present in placebo-controlled trials of both homoeopathy and conventional medicine. When account was taken for these biases in the analysis, there was weak evidence for a specific effect of homoeopathic remedies, but strong evidence for specific effects of conventional interventions. This finding is compatible with the notion that the clinical effects of homoeopathy are placebo effects. [My bold.]
There have been some criticisms of the Shang paper, but these have tended to be insubstantial and have been dealt with by several people. An example of the criticisms and the way they have been dealt with is available here.
Another review, by Linde and Melchart, seemed more positive:
The results of the available randomized trials suggest that individualized homeopathy has an effect over placebo. The evidence, however, is not convincing because of methodological shortcomings and inconsistencies.***
However, the same author (Linde, this time with other colleagues) later noted that:
Studies that were explicitly randomized and were double-blind as well as studies scoring above the cut-points yielded significantly less positive results than studies not meeting the criteria. In the cumulative meta-analyses, there was a trend for increasing effect sizes when more studies with lower-quality scores were added.
In other words: the poorer the quality of the trial, the better homeopathy looks. Design rigorous trials that eliminate bias, and the effects of homeopathy begin to disappear. If you want to show an effect of homeopathy, your best bet is to include the poorest-quality trials – those most likely to yield spurious results due to biases that have not been eliminated.
Believe it or not, there is actually a systematic review of systematic reviews of homeopathy:
Eleven independent systematic reviews were located. Collectively they failed to provide strong evidence in favour of homeopathy. In particular, there was no condition which responds convincingly better to homeopathic treatment than to placebo or other control interventions. Similarly, there was no homeopathic remedy that was demonstrated to yield clinical effects that are convincingly different from placebo. It is concluded that the best clinical evidence for homeopathy available to date does not warrant positive recommendations for its use in clinical practice.
OK – even if homeopathy is implausible, is not explained by quantum physics, and has not been shown to work better than placebo, shouldn’t we harness the power of the placebo effect by using homeopathy?
In my opinion, we should not. While we may all derive some benefit from something that is comforting (for example, a nice warm blanket or a mug of tea), prescribing pills and potions that have no beneficial effect for pretty much anything other than pain is in my view unethical – and any hope provided by homeopathy is a false hope.
Here, I looked at the case for the “powerful placebo” and found no good evidence that there was such a thing as the ‘powerful placebo effect’ – as Hróbjartsson and Gøtzsche put it in their Cochrane Review:
We found no evidence that placebo interventions in general have clinically important effects. A possible moderate effect on patient-reported continuous outcomes, especially pain, could not be clearly distinguished from reporting bias and other biases. We suggest that placebo interventions should not be used outside clinical trials, also for ethical reasons, as the use of placebo often involves deception.
Note that even in pain (the area where a claim for a powerful placebo effect is most plausible), the authors state that, far from being powerful, there is only a possible moderate effect. This moderate effect cannot be clearly distinguished from reporting bias and other biases and is therefore in doubt.
As to the ethics of prescribing alternative medicine such as homeopathy for its placebo effect, here, David Colquhoun writes that: “The whole trend in medicine has been to be more open with the patient and to tell them the truth. To maximise the benefit of alternative medicine, it is necessary to lie to the patient as much as possible.”
With regard to ethics and homeopathy, Gimpy has detailed examples of the issues.
*Correction: I made an error in describing a femtosecond as one thousandth of one billionth. Thanks are due to Col for pointing this out in the comments section.
**Correction: bloodletting is effective for haemochromatosis and therefore is probably more useful than the entire discipline of homeopathy. Thanks to Jo Brodie for pointing this out in the comments.
***Correction: it is worth noting that the authors also state that “when the analysis was restricted to the methodologically best trials no significant effect was seen”. Thanks to Mojo for pointing this out in the comments below.