Science

Placebo Effect Is Shown to Be Twice as Powerful as Expected

By DANIEL GOLEMAN Published: August 17, 1993

"HURRY, hurry -- use the new drugs while they still work!" a 19th-century French physician urged his colleagues. He may not have known why faddish drugs work on credulous patients, but the fact that they do has been borne out by scientists studying the power of the placebo to cure. New findings show that the placebo effect -- in which patients given an inactive treatment believe it can cure them -- is most powerful when a trusted physician enthusiastically offers a patient a new therapy. In a study of more than 6,000 patients being given experimental treatments for asthma, duodenal ulcer and herpes, two-thirds improved, at least temporarily, even though rigorous tests later found the treatments medically useless. They were then abandoned. The old rule of thumb among medical researchers was that only about one-third of patients will show some improvement when given a placebo. The results of the new studies reveal the effect to be twice as powerful as was thought. These and other findings that the placebo effect can be far stronger than had been widely assumed are leading some researchers to call for stricter standards for testing new medications. Others are proposing that physicians try to capitalize on the placebo effect in treating their patients in order to marshall the body's own healing powers. "I argue that instead of just trying to control for placebo, we should try to maximize it," said Dr. Frederick Evans, a psychologist at the Robert Wood Johnson School of Medicine in New Brunswick, N.J. "If a doctor believes in what he's doing and lets the patient know that, that's good medicine." While many people think a "placebo" is simply a sugar pill or other medicine with no active ingredients, the term has a broader meaning. The "placebo effect" includes any improvements in a patient not specifically due to a particular ingredient in a treatment, like a drug or surgical procedure. These "nonspecific," or placebo, effects, may be due to causes ranging from a patient reporting relief from symptoms in an unconscious effort to please a well-liked physician, to actual biological improvement. Testing the Placebo Effect To assess the potency of the placebo effect during the burst of enthusiasm for a new medical treatment, researchers re-examined data from initial clinical trials of five procedures which had at first seemed highly promising, and then later were shown to be useless. The procedures included surgical removal of the glomus, a structure near the carotid arteries in the neck, to treat asthma; and gastric freezing for duodenal ulcers. They also included three treatments for herpes simplex virus -- the drug levamisole, organic solvents like ether and exposure of dyed herpes lesions to fluorescent light. "In these studies, the doctors treating were also those evaluating the symptoms, which is what happens in a typical physician's office," said Dr. Alan H. Roberts, a psychologist at the Scripps Clinic and Research Foundation in La Jolla, Calif., who led the research. The results were published in the current issue of Clinical Psychology Review. The physicians, who offered the treatments as part of an early clinical trial and believed in their efficacy, told their patients the various approaches were new and promising. With both physicians and patients having high hopes for a cure, the resulting placebo effect was potent. Because these were very early trials of new drugs, no control groups were used.

Of a total of 6,931 patients receiving one or another of the five treatments, 40 percent were reported to have excellent results, another 30 percent had good outcomes and only 30 percent were reported to have "poor" results, Dr. Roberts and his colleagues found. Yet in later trials, when patients who received the treatments were methodically compared with control groups of patients who received placebos or nothing at all, "the effectiveness disappeared," said Dr. Roberts. Dr. Roberts believes that for relatively mild medical problems, under the best conditions the placebo effect will produce positive results in roughly two-thirds of patients. The effects would not be nearly as strong for serious diseases such as AIDS or cancer, he said: "In the more severe disorders, the placebo effects would be mainly in terms of patient's subjective complaints, not their physical symptoms." Psychological Factors In Dr. Robert's view, the improvements associated with placebos are caused by factors like patients unconsciously exaggerating improvements of their symptoms in order to please their doctors, and doctors who hope for positive results skewing their evaluations of symptoms favorably. The notion that the placebo effect is due to biological changes from patients' hopes being raised is met with skepticism by Dr. Roberts. But other researchers disagree. "Could an enthusiastic physician and a believing patient create a clinical improvement in a patient?" said Dr. Ronald Glaser, a virologist at Ohio State University Medical School. "That question has haunted drug studies. But there may well be a psychological effect with a significant biological outcome, if you extrapolate from data showing that psychological factors like stress can affect viruses like herpes. It's definitely one possible explanation." The herpes virus is one of Dr. Glaser's specialties; with his wife, Janice Kiecolt-Glaser, a psychologist, he has studied the effects of people's emotional swings on the replication of herpes virus. "We've found herpes viruses are responsive to stress, improving or worsening depending on a patient's emotional state," said Dr. Kiecolt-Glaser. "Since herpes virus is quite responsive to psychological influences, the first wave of physicians' enthusiasm could well have a beneficial medical effect." Dr. Roberts is not the first researcher to find that the placebo effect can account for improvements in more than one-third of patients, a ratio proposed in the 1950's by Dr. Henry Beecher, one of the first to do research on the placebo. Patients Most Likely to Benefit "The range for placebo recovery I've seen varies from zero to 100 percent," said Dr. Arthur K. Shapiro, a psychiatrist at Mt. Sinai Medical Center in Manhattan. "Different factors combine to produce the magnitude of a placebo. For example, in my own research with 1,000 patients, those who like their physician most and who were most anxious showed the greatest improvement from placebo." Such findings on the power of the placebo are bringing calls for revising the way in which new treatments are tested. In order to be sure the benefits attributed to experimental treatments are not simply due to placebo, tests of new medications now use a "double-blind" design, in which neither the physician nor the patient knows what medicine is being given, and some patients are given a nonactive treatment. "There is a false sense of security about the scientific tests of drugs, particularly psychiatric drugs," said Dr. Roger Greenberg, a psychologist at the State University of New York Health Science Center at Syracuse. One of the main problems with the standard double-blind test, said Dr. Greenberg, is that patients and physicians alike can very often tell who is getting the active medication and who is getting the placebo, because only the true medication has side effects. This can lead to placebo enhancement of the seeming effectiveness of the medication being tested.

"In instances when researchers have asked patients and physicians to guess whether they were using the active medication or the placebo, the results are sobering -- in one such study, 78 percent of patients and 87 percent of their physicians could tell," said Dr. Greenberg. "That means the so-called 'double-blind' is not really blind." For that reason, Dr. Greenberg proposes that in addition to the medication being tested and the usual inert placebo, tests of new drugs should include an "active" placebo which produces side effects but has no medical consequence. And, in the most rigorous test, a physician other than the one giving the medicines would make the evaluations of improvement. Call for Changes in Clinical Trials In a meta-analysis of 22 studies of antidepressants, Dr. Greenberg and colleagues found that if, in addition to the new drug being tested, some patients were given an older antidepressant as a control, the new drug was only onequarter to one-half as powerful as was reported in studies without the comparison drug, in which the new drug was pitted only against an inert placebo. The research finding, published last year in the Journal of Consulting and Clinical Psychology by Dr. Greenberg and colleagues, concludes that current standard practices for drug testing often exaggerate the potency of new medicine. "In general," said Dr. Greenberg, "the better a study is controlled, the blinder it becomes and the smaller the difference becomes between the real drug and the placebo."