A survey of internists and rheumatologists suggests that prescribing active “placebos” is relatively common, even though ethicists generally frown on the use of such therapies, especially if the patient is not informed.
The survey was conducted by five ethicists from the National Institutes of Health; the University of Chicago; and Harvard Medical School, Boston, who said they were interested in exploring physicians' attitudes about placebo treatments because there is little systematic data on the topic (BMJ 2008;337:a1938 [doi: 10.1136/bmj.a1938]).
They surveyed 1,200 randomly selected physicians, half of whom were internists, and half, rheumatologists; 679 physicians (57%) agreed to participate and received $20 for completing the survey. The respondents comprised 334 internists and 345 rheumatologists.
Depending on how the question was asked, 46%–58% of the physicians said they prescribed placebos on a regular basis, and 399 of 642 said it was ethically permissible to do so.
The study was somewhat deceptive, however, said Dr. Roy Altman, a rheumatologist in Agua Dulce, Calif., who was not involved with the study. The authors did not ask the physicians if they were merely going along with patients who were already taking placebos, Dr. Altman said in an interview.
He said that if a patient is taking a placebo that he knows is not harmful, he generally won't stop him or her. “I'm not supporting the practice of giving placebos,” he said, adding, “[but] I think it's something that's a part of medicine and I don't think you can take it away.”
The authors noted that, to avoid using the word “placebo,” they began with broad questions. For example, they asked physicians to respond to a hypothetical scenario in which a dextrose tablet had proven superior to no treatment. Would they recommend it as a therapy for nondiabetics with fibromyalgia? Twenty-four percent of physicians (160 of the 654 who answered these questions) said it was very likely they'd use the sugar pill; 34% (221 of 654) said it was moderately likely. But 31% (205) said it was unlikely and 10% (68) said definitely not.
They were then asked how often they recommend a therapy because they believe it will enhance the patient's experience. Fifty-nine percent (380 of 642) said it was permissible to recommend such a treatment; 31% (197) said it was permissible, but only in rare circumstances; and only 7% (46) said it was never permissible.
Finally, physicians were asked what therapies they had used primarily as a placebo treatment. Placebo was defined as “a treatment whose benefits derive from positive patient expectations and not from the physiological mechanism of the treatment itself.”
Fifty-five percent of the respondents (370 of 679) said they had recommended—but not necessarily prescribed—some type of placebo in the past year. A total of 267 of 648 physicians (41%) prescribed over-the-counter analgesics and 243 of 648 physicians (38%) prescribed vitamins. Sedatives were prescribed by 86 of 652 physicians (13%); the same number prescribed antibiotics. Saline and sugar pills were used by 18 of 623 and 12 of 642 physicians (3% and 2%, respectively).
When asked how these treatments were described to patients, 18% (62 of the 352 who actually prescribed placebos) said they were “medicine.” About 285 physicians said they had not prescribed placebos. Only 18 physicians (5%) said they identify the treatments as a placebo. A large percentage—68%, or 241 of the 352—described the placebo as a medicine not typically used for the condition that might benefit the patient.
An American Medical Association policy on the use of placebos is a bit murky; it advises against their use without the patient's knowledge or if the placebo could cause medical harm. A placebo can be prescribed “only if the patient is informed of and agrees to its use,” but the physician does not have to identify the placebo, or explain its potential effect, according to the policy.
In an interview, Dr. Norman Gaylis said that he believes the use of active placebos is “an inappropriate way to treat a patient.” Giving a patient naproxen, for instance, “has potential for significant side effects to the kidney, so that's not a placebo,” said Dr. Gaylis, a rheumatologist in private practice in Aventura, Fla.
Although the placebo effect is well documented in medicine and can be effective, he said, “the first thing for us is to not harm our patients.”