Homeopathy doesn’t work. It never works. It cannot work. When it appears to work, the apparent effect can invariably be traced to some cognitive bias or error of perception. So can it be ethical to continue to research the homeopathy’s (in)effectiveness? I think a case can be made that it can.

My colleague and friend Tim Caulfield has written a compelling Policy Options piece on why conducting any further research on homeopathy would be unethical. Caulfield rightly points out that to do scientific tests on homeopathy is to do science on the absurd: “[H]omeopathy,” he writes, “has the same scientific plausibility as using Harry Potter’s wand to treat illness.”

Caulfield’s argument is strong. He rightly points out that expert consensus in the field of human-subjects research ethics suggests that research done on humans must be grounded in good science. You shouldn’t be testing things on people based on a whim or a fantasy. Doing so is generally a waste of resources, and often a waste of research subjects’ time. Homeopathy hasn’t even a shred of scientific plausibility, so there’s no reason to think that further testing of it will advance the cause of human health. To say that homeopathy will never be proven effective isn’t to have a closed mind; it is simply to have a grip on reality.

Caulfied is also right to worry that, in encouraging scientific testing of something as scientifically absurd as homeopathy, we may be granting it dignity it doesn’t deserve. If someone firmly believed that, say, a “lucky rabbit’s foot” could cure cancer, and if we agreed to conduct a rigourous scientific study of the question, we would only be promoting that individual’s delusions, not promoting our understanding of healthcare.

But I’d like to offer a couple of points that should count toward keeping the door open with regard to scientific testing of homeopathy.

My first point has to do with the assumption Caulfield makes that trials of homeopathy always waste resources and time. This is probably true in general, but not true all the time. And a trial of a homeopathic preparation used in addition to some other standard treatment would involve virtually no risks and not waste any appreciable time or resources. If a clinician wanted to conduct a randomized controlled trial in which half of a large group of cancer patients were given a homeopathic preparation in addition to standard treatments, in order to see whether one group did better, in terms of morbidity and mortality, than the other, I think that should be permitted. We already know that homeopathy cannot work, but not everyone is convinced, and convincing them would be truly useful.

My second point has to do with the risk of telling people that they are not allowed to conduct the kinds of scientific testing that would, if successful, lend credence to homeopathy. If such testing were to be forbidden, that would provide proponents of homeopathy a handy rebuttal to charges that their belief in homeopathy is unfounded. When faced with the charge that homeopathy lacks scientific credibility, a proponent would then be able to say, “We’d love to provide evidence, but they won’t let us do the tests!” It would be bad to give them access to such an argument.

So could we still, ethically, conduct scientific studies of homeopathy? Yes.

Of course, two important ethical provisos would apply. The first is that any proposed test should be truly rigourous. The trial should among other things be large, double-blind, and randomized, so that it didn’t merely add to the long list of methodologically suspect trials and case-studies that fans of homeopathy currently point to in support of their beliefs. And second, any ethical trial would have to be candid with participants about the extant scientific evidence in favour of homeopathy, in order not to foster false hope. And such candour would be difficult for those already committed, as too many already are, to the notion that homeopathy is anything other than a fantasy.