I have more thoughts on the homeopathy matter than fit in follow-up notes, so here goes.
First, David Gorski recalls the 1994 Pediatrics report on childhood diarrhea treated with tailored homeopathic remedies for each patient. There is more to the story than has been written. I am certain much of this will get back to the authors, but others may benefit from knowing how this group of homeopaths operate.
I recall the paper well, because it was the first journal report that I deconstructed and published (Pediatrics, Oct 1995) as a regular article. I think it was the first time the journal had published a formal rebuttal outside the Letters section. The head of pediatric pharmacy at Valley Medical Center, San Jose, brought the paper to me and asked what I thought if it. Bill London of National Council against Health Fraud and I spent six months discussing it and working over the details.
The paper had so many flaws, that one letter could not contain them. It had five or six end points, several arithmetical errors, graphs with missing data, only one end point reached consensus signficance (barely.)
Each case received a remedy tailored to the age, condition, duration of symptoms, appearance and odor of the stool, the recall of the parent or relative about stool frequency (which depended on how often the child’s diapers were changed, and a number of other qualities, typical of a homeopathic approach to diagnosis. The remedies given were not based on etiology, but based on personal observations, We saw that the remedy was chosen at a snapshot in time, depending on all those factors which varied from hour to hour. So the remedies depended on the time at which the child was brought in for examination and were unchanged throughout the duration of illness. That made no sense at all. Besides, the specific remedies had no data behind them for proof of efficacy, and were chosen on basis of charts and computer references.
One could hardly find anything about the paper that would lend credibility to its conclusion that suggested homeopathy “worked“ better than placebo. The results in our opinion demonstrated nothing more than the variations in the clinical trial system (noise.)
The editors received dozens of complaining letters about its publication, and was embarrassed enough to depart from tradition and publish ours, which was four pages long. Although we made one technical error about the statistical analysis, the other criticisms were obvious and valid.
The current SBM discussion regarded the ethics of the study. (beginning with the one written about by Harriet Hall.) I must admit that ethics were not in our consciousness at the time we discussed the paper, so we did not comment on ethics. One wonders if standard therapy was withheld; but the authors were careful to have selected only children with mild diarrhea – less than 6 stools/day – who did not require rehydration fluids. Sicker children were excluded or removed from the trial and treated appropriately. I’d have to check on the informed consent – I don‘t recall. But unless it stated that there was no evidence that homeopathy works, any consent would have failed the criterion of informed. .
But more followed and the story got more interesting. We had pointed out that despite the absence of clear efficacy there was no assurance that the homeopathic remedies were not adulterated by antibiotics or anti-diarrheals. So the authors repeated the study, but next time in Nepal instead of Nicaragua. No reason was given for the change. This time an independent agent checked the remedies for adulteration and none was found. They published the results of the second study in the Journal of Alternative and Complementary Medicine (2000). Word got around that it was submitted to standard medical journals and was rejected. When I saw the second study, the most apparent problem I saw was that the end points “positive” in the Nicaragua study were negative in Nepal and vice versa. I interpreted the studies as having disconfirmed each other. Nevertheless, the homeopaths claimed that the Nepal study confirmed the Nicaragua trial. As I recall also, the homeopathic group weighed more (was significantly older) than the control group; older children have shorter courses of diarrhea illness than younger. A letter to the “alt” journal would have been wasted effort. Unfortunately, the journal is abstracted on Medline.
If that was not enough to bother you, the authors went further. And here is where the two strains of recent discussions meet – the Hall/Gorski/Atwood et al discussion – homeopathy/ethics and clinical trial/meta-analysis. Jacobs combined the borderline data from the two studies plus the pilot study preceding the Nicaragua trial into a meta-analysis. They obtained a more significant P value. Not mentioned was that the respective data from the trials conflicted with one another. This MA they published in Pediatrics Infectious Disease Journal I 2003. I took the results of the MA to a well known authority on MAs at Stanford whose course I had been auditing. His response? A combined blank stare and shrug. Then a look like…don’t you really know? I did, of course.