Consumers are not the only ones experimenting with complementary and alternative medicine these days. After shunning CAM for most of the past century, many conventionally trained physicians are now working to assess the worth of nonconventional remedies. Alas, the mission is not as straightforward as it sounds.

Western science has developed powerful methods for testing pills and devices, but CAM therapies can pose unique challenges. Appraising them will require not only new studies but, in some cases, more sophisticated ways of designing studies.

The gold standard in scientific medicine is a randomized, placebo-controlled, double-blind trial. If researchers wanted to test a new pill for heart failure, they would assign patients at random to receive either the new pill or a placebo pill that contained no active medicine. Neither the doctors nor the patients would know which subjects were getting the actual treatment. After all the results had been collected, the investigators would determine whether the severity of heart failure was lower in the group that received the new medicine.

Now imagine conducting a similar study of acupuncture. Unlike a pharmaceutical drug, acupuncture technique varies among practitioners. Will participants in the trial receive Chinese acupuncture, or will they get the Japanese or Korean variety? Which acupuncture points will therapists target on the patients’ bodies? How far will they insert their needles? Will they twist the needles or apply electrical currents, or will they simply apply physical pressure?

The challenges don’t end there. To separate the “acupuncture effect” from the placebo effect, the researchers will have to devise a sham procedure that is indistinguishable from the real one. Should patients in the placebo group have real needles inserted at a slight distance from the true acupuncture points? At points used to treat some other condition? Or should the therapists use sham needles that don’t really penetrate the skin? And how will the acupuncturists avoid conveying unconscious clues about whether they’re giving real treatment? Should the patients be blindfolded? These dilemmas are not unique to acupuncture. Imagine studying massage therapy for low-back pain, yoga for fibromyalgia or hypnosis to help people quit smoking. Each remedy would pose new problems in study design.

Herb studies are less daunting, but they, too, present challenges. To design a rigorous echinacea study, researchers would have to settle on one species of the herb (three are in widespread use). They would also have to use plants of a specified age, and decide how to prepare and store them. A liquid extract might have different effects from dried, crushed leaves.

In many cases, we will need to compare CAM therapies not only with placebos but also with conventional therapies. Can you imagine designing a double-blind, placebo-controlled study comparing acupuncture with pain pills? It could require designing placebos for both of the active treatments, and administering four interventions in different combinations. Experts are now racing to develop new study designs, and their efforts could open new avenues to good medicine.

---

Drs. Ted Kaptchuk, David Eisenberg and Anthony Komaroff,
Harvard Medical School   

Dec. 2, 2002, © 2002 Newsweek, Inc.