I've read a bunch of the research done on acupuncture. Some of it is laughable. For instance, some of the studies I read separate people into three groups -- a control group with no treatment, a group that gets "real" acupuncture on "real" acupuncture points, and a group that gets "sham" acupuncture on "sham" points.
That starts with an unproven assumption -- that acupuncture points exist, have identifiable properties and can be mapped. And an assumption that if the acupuncturiest doesn't exactly needle the correct spot there will be no effect.
And yet, people claim good results from acupuncture. More reliable are studies that have groups of people with similar symptoms, such as back pain, divided into groups that get massage, acupuncture for another group, chiropractic for another group. Even there you have problems removing bias -- how do you make sure the technician does the same thing the same way identically for each person? And the results are pretty subjective, often before and after questionnaires about pain level.