The Best Clinical Trials Fail to Demonstrate Effectiveness
Yet acupuncture might still work via other mechanisms. Just because we don’t understand how it works it does not mean it is ineffective. The only way to find this out is to conduct controlled clinical trials. Several hundred such studies are now available. Their results are mixed but generally they demonstrate that, if placebo-effects are rigorously controlled for, the clinical outcomes are not different than those for sham acupuncture. Two exceptions might exist: certain pain syndromes and nausea/vomiting due to various causes. For all other conditions, there are no reliable data to show that acupuncture is effective.
Sign up for the OV Daily Newsletter
OV Social
My postdoctoral advisor when I was at Cornell is an experienced experimental scientist and a deeply skeptical thinker, including regarding non-standard medicine. He suffered from fairly severe back pain and was talked in to trying acupuncture. He had to convince the acupuncturist to stop spouting the ridiculous explanations of why it worked, but the truth, for him, is that it does reduce his pain. He now goes about once a month, and his standard of living is improved. Placebo effects are considerably stronger for people who have positive expectations. I have a hard time believing that he is fooling himself, but regardless, I am glad that his suffering is eased.
Disproof is impossible, right?
If you reduce the effective n of these studies, the statistical significance drops (of course). There could be a segment of the population for whom acupuncture works (as with some other treatments). It is also highly unlikely that acupuncture can do all of the things that have been ascribed to it. Hypothetically, if acupuncture worked only for pain, and only for one third of the locations ascribed to it, the studies to date could be missing a large positive signal. Alternately, the target could be remarkably small (a specific nerve?) such that even 'good' practitioners hit it only 10% of the time and 'bad' ones only 1%.
I am glad that the studies continue, and I am not willing to close the books yet. It seems to me that the risks are low enough that people should be encouraged to try it if they have the inclination. Furthermore, medical clinicians should avoid carelessly denigrating the practice, so that their patients are willing to be forthcoming.