DR. EISENBERG: Ken, your point is well taken, and I think you point out how complex it is to design studies that provide an answer with which one can be satisfied.

Certainly, things taken orally that we can control with a credible sham placebo pill are the easiest. I think once you get into provider–dependent interventions, again, the methodologists have been very creative.

In the area where we think modality is the key to the efficacy, we try to standardize the intervention as best as possible. Even there, just to stick with it for a moment, you get into a conundrum. Do you use the same point for nausea in all patients, or do you let 10 qualified acupuncturists from five different Asian traditions pick their own, hoping there will be some overlap?

The NIH has, in a rather remarkable way, opted to fund projects to follow each of those trajectories. Some studies will standardize the acupuncture points or the massage points or the chiropractic approaches, where herbs for irritable bowel will leave it up to the practitioner to diagnose and treat, as is their normal practice.

We end up with three arm studies. The study for irritable bowel in JAMA was one of these studies that used that paradigm.

There is also tremendous creativity in how to appropriately blind and create controls that are credible for some of these techniques.

There are actually, I think, fairly good examples of sham placebo devices that can be used in patients who are naive to acupuncture, which are spring loaded and go through a tube.

The person who is naive to real acupuncture, when punctured by this sham acupuncture needle, which is on a spring, feels the needle touching the skin but it never touches the skin; that is one level of control. Whether that is satisfactory for all questions about acupuncture is another question.

I guess what I am trying to say, without appearing or being overly defensive, is that each study has its own limited number of questions it can ask and answer authoritatively.

That is precisely why this field needs people who are trained in clinical epidemiology, who understand the modalities as they are used—not just the rules of evidence, and why we need universities that are willing to bring in clinicians from the different professions to ask how this is actually practiced, before devising a randomized trial that meets reductionistic requirements, which has little or no bearing to the way you and your colleagues actually practice it.

This is not easy. It is very messy science. If you extend that to herbs, they are the messiest of all. Even if you can give a credible sham foultasting elixir that tastes like the foul-tasting elixir of herbs, the consistency



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement