This is the part of the story where the question changes shape. For most of this series we have asked what homeopathy means, where it came from, and how people experience it, questions where many perspectives can each hold a piece of the truth. Here we ask something narrower and harder: when you measure it, does it work? We owe you those measurements straight, the strongest case its advocates make, the findings that cut against it, and an honest sort of where each piece really sits.
A measurement question deserves a measurement answer. “Does this remedy outperform a placebo for this condition” is not really a matter of worldview or emphasis; it is a matter of what carefully designed trials find when they look. So this section does two things at once, and tries to do both honestly: it gives the case homeopathy’s advocates make its fullest, fairest hearing, and it reports the mainstream scientific verdict without softening it.
How could it work?
The deepest objection to homeopathy is not clinical but physical. Beyond a certain point of dilution, around the twelfth “centesimal” step, the laws of chemistry say that no molecule of the original substance is likely to remain in the preparation at all. Many homeopathic remedies are diluted far past that point. To mainstream science, a medicine containing none of its stated active ingredient cannot have a specific pharmacological effect, and the European Academies’ Science Advisory Council, among others, has called the underlying claims implausible.
Homeopathy’s defenders answer not by denying the dilution but by proposing that the preparation carries information rather than substance. The “memory of water” idea associated with Jacques Benveniste suggested water might retain an imprint of a diluted molecule, though independent teams were unable to replicate the findings. The structured-water research of scientists such as Gerald Pollack explores ordered phases of water at surfaces, intriguing but not established as a mechanism. The nanoparticle work of Chikramane and colleagues at IIT Bombay reported in 2010 that measurable nanoparticles of the source material can persist even in extreme dilutions.
The nanoparticle finding is real, peer-reviewed, and worth taking seriously, because it shifts the conversation from “there is literally nothing there” to “there is something measurable there.” But showing that trace particles persist is not the same as showing they produce the specific, symptom-matched clinical effects the system claims. The fair summary is the careful one: the mechanism is unexplained, not demonstrated impossible, and unexplained is also not the same as demonstrated true.
The strongest case the trials can make
Positive trials do exist, and it would be dishonest to pretend otherwise. The most homeopathy-favorable recent synthesis, the systematic review programme led by Robert Mathie, reported in 2014 that individualised homeopathy was roughly one and a half to two times more likely than placebo to produce a beneficial effect, while its own authors rated the quality of the underlying evidence as low. Individual trials have reported benefits in childhood diarrhoea, allergic rhinitis, and fibromyalgia, among others.
These results are genuine positive signals, but they tend to be small, methodologically uneven, and the decisive problem is replication: a promising small trial is an invitation to a larger, tighter one, not a conclusion in itself, and the larger, tighter confirmations have largely not followed.
The mainstream verdict
The most-cited comparative meta-analysis, published in The Lancet in 2005 by Shang and colleagues, concluded that the clinical effects of homeopathy were compatible with placebo, and prompted an editorial titled “The end of homoeopathy.” Homeopathy’s advocates contest that paper with some force, since its conclusion rests on a small subset of the trials it examined.
The most exhaustive single assessment is the 2015 review by Australia’s National Health and Medical Research Council, which examined more than 1,800 studies and concluded that there are no health conditions for which there is reliable evidence that homeopathy is effective. It went further on safety, warning that homeopathy should not be used to treat conditions that are chronic, serious, or could become serious, because of the risk of forgoing treatment that works.
So here is the bottom line, stated as plainly as a reader deserves: by the standards mainstream evidence-based medicine applies, homeopathy’s effectiveness for specific named conditions is unproven, and the prevailing scientific position is a skeptical one.
Cuba and leptospirosis: a striking story, honestly weighed
In 2007 and 2008, Cuba’s Finlay Institute administered a homeopathic nosode to roughly 2.3 million people during a leptospirosis threat, reporting a sharp fall in cases. Held honestly, this cannot carry the weight often given to it: it was a population-level intervention with no control group, published by the very institute that designed and ran it. Even the UK’s Society of Homeopaths has described the evidence for this kind of homeopathic prophylaxis as largely anecdotal, and does not endorse it as a substitute for vaccination.
The largest modern body of homeopathy research
India has produced the most substantial single body of modern clinical research into homeopathy, through the government-funded Central Council for Research in Homoeopathy. Much of it is observational or smaller in scale, and has not yet delivered the large, pre-registered, independently-run replications that change scientific consensus. That is the honest place to leave it: promising and under-resourced, an open file rather than a closed case.
Where each kind of evidence sits
Robust clinical evidence: Little about homeopathy’s effectiveness for specific conditions sits firmly here by mainstream standards. What does belong here is narrower and real: well-prepared remedies are physically safe, and the unhurried, whole-person consultation has measurable value to patients in its own right.
Emerging or promising research: Individual positive trials and the Indian research programme show signals worth taking seriously and worth funding further. These are not proof, and are not presented as proof.
Traditional or experiential knowledge: Two centuries of detailed clinical observation across dozens of countries is real knowledge of a particular kind. It is not trial-grade evidence and shouldn’t be dressed as such, but it isn’t nothing either.
What the evidence will and won’t carry
Two things are true at once, and a reader is better served holding them together than choosing one and discarding the other. Positive signals exist; the research has been under-funded and under-built for a century; the deeply individualised method is genuinely awkward to fit into trial designs built for one-drug-one-disease testing; and the mechanism is unexplained rather than proven impossible.
