Debates on researching SCAM frequently hinge on the tension between theoretical plausibility and empirical testing. The central question is this: should interventions that contradict well-established scientific principles nonetheless be subjected to clinical trials? This issue raises fundamental concerns about the allocation of research resources, the epistemology of medical science, and the boundary between scientific openness and credulity.
In medical research, plausibility refers to the compatibility of a treatment’s purported mechanism of action with established biochemical and physiological knowledge. Treatments such as homeopathy, reiki, bioresonance, etc., etc. are typically judged to be implausible because their mechanisms violate basic physical principles (Ernst, 2010; Offit, 2013). One could argue that conducting randomized clinical trials (RCTs) on such interventions is methodologically and ethically questionable, as the prior probability of efficacy is exceedingly low (Goodman, 1999) and the probability of a positive result obtained with a rigorous trial approaches zero. Framed in Bayesian terms, if a hypothesis begins with a negligible prior probability, even seemingly positive trial outcomes are unlikely to meaningfully change its posterior credibility (Spiegelhalter, 2019).
But insisting that only theoretically plausible hypotheses merit empirical testing risks scientific conservatism. Medical history includes numerous examples – such as the discovery of the antipyretic and antithrombotic effects of aspirin – where therapeutic value was demonstrated before mechanisms were fully understood (Vane, 2000). Advocates of broader testing argue that empirical methods should retain the capacity to surprise theory and that excluding “implausible” ideas a priori risks reinforcing disciplinary dogma (Ioannidis, 2012). This appeal to epistemic humility emphasizes observation as a safeguard against the overreach of theoretical reasoning.
However, defenders of plausibility-based research prioritization contend that such humility must be balanced against the always finite resources for research and the ethical responsibility of researchers. Health research funding is limited, and prioritizing the study of implausible treatments may divert attention from interventions with rational mechanistic foundations and higher expected utility (Sampson, 2005). Additionally, the evidentiary record of clinical trials in SCAM demonstrates a consistent pattern: small, underpowered studies occasionally produce marginally positive results that fail replication, while systematic reviews of rigorous studies yield null or inconclusive conclusions (Shang et al., 2005; Ernst et al., 2011). In such cases, further testing is more likely to perpetuate public misunderstanding than to advance medical knowledge.
So, what is the solution? An epistemologically coherent approach, frequently advocated in evidence-based medicine, is to calibrate evidential standards to plausibility. Highly improbable claims should first demonstrate compelling preclinical signals – biochemical, mechanistic, or reproducible physiological effects – before human trials are considered (Goodman, 1999; Howick, 2011). This proportionality upholds methodological rigor without foreclosing the possibility of genuine empirical discovery. It respects Bayesian reasoning: extraordinary claims require extraordinary evidence.
Science must remain open to the unexpected yet disciplined in method and theory. Medical research is usually at its most productive when it operates between the two extremes—dogmatism that refuses to test unconventional claims, and indiscriminate empiricism that tests everything without theoretical guidance. The prudent path lies in aligning the scope of empirical investigation with scientific plausibility, ensuring openness tempered by rational constraint.
References
- Ernst, E. (2010). “A systematic review of systematic reviews of homeopathy.” British Journal of Clinical Pharmacology, 69(5), 577–582.
- Ernst, E., Pittler, M. H., Wider, B., & Boddy, K. (2011). The Desktop Guide to Complementary and Alternative Medicine. Elsevier.
- Goodman, S. N. (1999). “Toward evidence-based medical statistics. 1: The P value fallacy.” Annals of Internal Medicine, 130(12), 995–1004.
- Howick, J. (2011). The Philosophy of Evidence-Based Medicine. Oxford University Press.
- Ioannidis, J. P. A. (2012). “Scientific inbreeding and same-team replication: Type D personality as an example.” Journal of Psychosomatic Research, 72(6), 408–410.
- Offit, P. A. (2013). Do You Believe in Magic? The Sense and Nonsense of Alternative Medicine. HarperCollins.
- Sampson, W. (2005). “Antiscience trends in the rise of the ‘alternative medicine’ movement.” Annals of the New York Academy of Sciences, 775(1), 188–197.
- Shang, A., et al. (2005). “Are the clinical effects of homeopathy placebo effects? Comparative study of placebo-controlled trials of homeopathy and allopathy.” The Lancet, 366(9487), 726–732.
- Spiegelhalter, D. (2019). The Art of Statistics. Pelican Books.
10 Responses to Testing the Implausible: Clinical Trials of So-Called Alternative Medicine (SCAM)?
Leave a Reply
This site uses Akismet to reduce spam. Learn how your comment data is processed.
Hey Edzard and fellow cultists…
You all might enjoy the recent dialogues that I have had with ChatGPT on the implausiblity of homeopathic medicines. Well, you’re consider dinosaurs of the 19th and 20th centuries.
See for yourself:
https://danaullman.substack.com/publish/post/186004709 (part 1)
https://danaullman.substack.com/p/dialogue-with-chatgpt-on-compelling (part 2)
@DUllman
You are just producing intellectual garbage. That is why ChatGPt also produces garbage. It is called GIGO principle (Garbage In, Garbage Out).
Just for the record again Dana because you seem suddenly to be so keen on Chat GPT, I asked it if you are a credible source regarding homeopathy:
Short answer: No—Dana Ullman is generally not considered a credible scientific source on Homeopathy in mainstream medicine or evidence-based research.
Why his credibility is questioned
1. Advocacy rather than neutral research
• Ullman is a long-time promoter and practitioner of homeopathy.
• He writes books, articles, and opinion pieces advocating it rather than conducting independent, peer-reviewed clinical research.
2. Limited presence in peer-reviewed medical literature
• Most of his work appears in books, alternative medicine outlets, or advocacy websites, not major medical journals.
• Credible medical sources typically rely on systematic reviews, randomized controlled trials, and publications in established journals.
3. Conflict with scientific consensus
• Major scientific and medical organizations—including the National Health Service, World Health Organization, and National Health and Medical Research Council—have concluded that homeopathy lacks reliable evidence of effectiveness beyond placebo.
4. Criticism from science and skepticism groups
• Ullman has been criticized for misrepresenting or selectively citing studies to support homeopathy.
What experts usually consider credible sources instead
For evaluating homeopathy scientifically, researchers typically rely on:
• Systematic reviews or meta-analyses in journals like The Lancet or BMJ
• Reports from organizations such as:
• Cochrane
• National Institute for Health and Care Excellence
Bottom line
Dana Ullman is a prominent advocate, not an impartial scientific authority. His writings may reflect the pro-homeopathy viewpoint, but they aren’t generally treated as reliable evidence in mainstream medicine.
@Dana Ullman
AI chatbots tend to
– make stuff up
– say things that please the human on the other end of the ‘conversation’.
So yes, I guess that such a hallucinating compulsive adulation machine is your perfect soulmate, always telling you what you already believe and want to hear. It must give you almost the same level of gratification as hearing yourself talk.
The overwhelming majority of scientists and doctors tend to disagree.
Exactly. Dana’s knowledge of how generative AI works would seem to be on a par with their knowledge of science.
But generative AI can make great artificial companions, agreeing with everything one says without question – if one wants such a companion of course.
Try getting that inconsequential bumwash published in a reputable journal, Dana. We’ll wait. Have you asked Chat GPT if you’re a credible voice regarding homeopathy? We have. And we’ve shown you the answers. In the meantime, we’ll enjoy watching you howling your inanities into the void. Nobody is paying you any attention you sad little man.
Hip hip hurray! The cultists here have again not disaapointed.
My comment above quoted the precise words of ChatGPT…and instead of providing ANY critique of what was said, the cultists chose to simply do their typical ad homs against me.
One of the stupidest here chose to claim that ChatGPT hallucinates…and of course, it occasionally does that…BUT then, one would expect this same person to be specific on what statement was an hallucination. NOPE. No one can expect something with substance like this.
And BTW, I recently asked to analyze Ernst’s claim that he asked “an AI” to review my contributions to medical science and if it is typical for an AI to avoid referencing any contributions listed in PubMed, THIS is what ChatGPT said:
“Standard LLM behavior does not automatically search PubMed”
I now challenge all of the above cultists to ask about my contributions and include reference to PubMed.
It is so much fun proving the sheer cultish behavior of the “regulars” here. I have no expectation that any of them will acknowledge any errors of judgment on their part. Instead, my contributions here are to the silent majority here…the people who don’t comment but who read what is written…and the cultists here have shown themselves to maintain a piss-poor scientific attitude.
And so, I now await a REAL review and critique of what ChatGPT has said TWICE. Put up, or shut up!
And while you’re there, I am still waiting for anyone here to review and critique my newest contribution to the PubMed literature: Rockefeller, the Flexner Report, and the American Medical Association: The Contentious Relationship Between Conventional Medicine and Homeopathy in America: https://www.cureus.com/articles/370572-rockefeller-the-flexner-report-and-the-american-medical-association-the-contentious-relationship-between-conventional-medicine-and-homeopathy-in-america#!/
You don’t make it clear what the actual prompts you used with chat gpt, and you don’t give any details about whether you used this prompt free of any other context.
But I took the following directly from your headline and used it in a fresh chat gpt session with no prior context:
“Are there plausible scientific explanations for how homeopathic medicines can work?”
And i get the following in response:
“Short answer: there’s no scientifically supported mechanism showing that homeopathic remedies work beyond placebo, and the proposed explanations conflict with well-established chemistry and physics. However, several hypotheses have been proposed to try to explain how they could work. None have convincing evidence.”
Real danger of using AI is that it is easy to fall in love with it, as it seems Dullman is smitten by i
ChatGPT.
Don’t go cheating on ChatGPT with other AIs, Dully!
Perplexity, Grok, Gemini, and Claude have already submitted letters of protest. 😀