A few weeks ago, Martin over at Aardvarchaelogy, Steve Novella, and I speculated about how alternative medicine modalities might evolve and what the selection pressures on them might be. We all agreed that, to some degree, there is definite selection pressure for remedies that do no harm but that also do no objective good either. In other words, there is selection pressure for placebos.
Obviously, the evolution analogy is imperfect, but there is also another possible explanation for the persistence of something like homeopathy, which is, in essence, no more than water and thus nothing more than a highly complex, ritualized placebo. We could look at classical operant conditioning concepts instead, as Christian at Med Journal Watch suggests. Basically, he likens alternative medicine (like homeopathy) to one big Skinner box, in which the reward is randomly distributed. As all students of classical psychology know, the random “reward” or reinforcement for a behavior produces the strongest conditioning of all. (It’s the same principle that makes playing the slot machine so compelling.) As Christian puts it:
Just like pigeons, humans are programmed in detecting patterns in a complex world. Such a pattern has been detected by Samuel Hahnemann who, using china-bark (a herbal remedy against malaria) as a healthy person, later has been suffering from symptoms similar to malaria. Based on this “similia principle”, he invented a totally different kind of medicine, homeopathy. His fever – the random pellet (a bad one of course). Taking china bark in self-experiment – the random behaviour.
Time has gone by. Today, millions of users are taking plain solvent, sold as homeopathic remedies. Again, we may use the Skinner box as a model to explain their behaviour. The worsening of symptoms is analogue to the condition without food pellet. The betterment is analogue to the pellet. And we all know that the whole life is based on rhythms – by the way a fundament of many woo theories (their arguments can be turned against them). That is, on a more or less random basis, symptoms come and go, wellbeing waxes and wanes. The best example is malaria, Hahnemann’s first similia paradigm, causing periodic fever attacks.
Now, feeling worse, our user will take some drops of solvent or globules of sugar containing no effective substance – the analogue of the nonsensical behaviour of the pigeon. He will continue to do so until he feels better, which is the analogue of the random pellet that inevitably will be given. Next time when feeling worse, he will not just take the remedy but he will take it “because it had helped me before” – and here we have the analogue of Skinner’s reinforcement of a behaviour element.
Such a model may explain why such remedies are so resistant to logic and science. Certainly, it’s worth considering. One thing that appears to be missing (or at least not explicitly included in the discussion) is the concept of regression towards the mean. In other words, the reinforcement delivered (improvement of symptoms) is probably not random, because homeopathy users are more likely to take their remedies when they are feeling the worst, a condition which almost inevitably moves back towards the mean, leading to the impression that homeopathy improved the patient’s symptoms.
Unfortunately, comments don’t appear to be enabled at Med Journal Watch; so leave your thoughts on this concept here.