A new study has subjected the acupuncture treatment for back pain to a proper controlled randomized testing:
In the largest experiment on acupuncture for back pain to date, more than 1,100 patients were randomly assigned to receive either acupuncture, sham acupuncture or conventional therapy. For the sham acupuncture, needles were inserted, but not as deeply as for the real thing. The sham acupuncture also did not insert needles in traditional acupuncture points on the body and the needles were not manually moved and rotated.
After six months, patients answered questions about pain and functional ability and their scores determined how well each of the therapies worked.
In the real acupuncture group, 47 percent of patients improved. In the sham acupuncture group, 44 percent did. In the usual care group, 27 percent got relief.
The sham acupuncture was introduced to measure the placebo effect: the psychological effect that might be caused by the act of treatment itself, even if the treatment has no real medical benefits. Studies which evaluate the effectiveness of medications often assign the control group a sugar pill or something similar to measure the placebo effect. Thus, one way of interpreting the results is that the real acupuncture treatment was no better than the placebo treatment, and that the whole effectiveness of acupuncture must be a placebo effect.
But why are the two needling treatments better than the usual care? It makes no sense that the placebo effect would be so large for acupuncture and much smaller or nonexistent for the usual care treatment.
I'm not sure what's going here, but one aspect of Chinese medicine might have been ignored here, and that is the difficulty in giving "sham" acupuncture. Ted J. Kaptchuck talks about this in one of the Appendices to his book The Web That Has No Weaver:
Many researchers worry that needling at "non-acupuncture" points is not analogous to a dummy inert pill as a placebo control. The concern is that needling anywhere in the body (at both real acupuncture sites or non-acupuncture sites) may have physiological effects... These responses can modulate pain and might produce effects beyond what might be expected from an inert sham control.
Kaptchuck points out that this is a particularly difficult problem when acupuncture is used just for pain control as is the case in the new back pain study.