The placebo effect: more than sugar pills
From the January ACP-ASIM Observer, copyright © 2002 by the American College of Physicians-American Society of Internal Medicine.
By Deborah Gesensway
A few years ago, a nurse at Winthrop-University Hospital on New York's Long Island complained to supervisors that an internal medicine intern had ordered her to give a patient complaining of pain a placebo—a saline injection—rather than an active drug. She felt it was unethical to deceive a patient.
That complaint eventually made its way to the hospital's ethics committee, and it prompted a survey whose results were published in a 1999 issue of the Western Journal of Medicine. In the survey, interns confessed to ordering placebos for patients 10 times over two years.
Perhaps more interesting, however, the interns did not expect the placebo to help their patients. While they acknowledged the growing body of literature on the ability of placebos to make patients feel better, many said that they simply didn't believe that a placebo—something no more powerful than a sugar pill—could reduce real pain.
According to study author Jeffrey T. Berger, FACP, director of clinical ethics at Winthrop, the survey points to the "cognitive dissonance" between what doctors know about placebos and the placebo effect and how they interpret placebos' actions and patient reactions to them.
But the survey—and the attention that it and other journal articles have received during the last few years—raise questions, some of which hone in on that dissonance. If physicians, for example, can help patients by giving them a placebo that could help their pain without deceiving them, should they? And if using placebos seems too drastic, should doctors use what is known as the placebo effect to give their "real" medicines maximum benefit?
The most robust evidence about the efficacy of placebo treatments comes from research on pain relief. In some studies, between 25% and 75% of patients responded to a placebo, and research has shown that an expectation of analgesia can produce real analgesic effects in patients.
While placebos seem to alleviate pain by activating endogenous opioid systems, they have also been shown to reduce some symptoms in common diseases ranging from chronic pain to hypertension and asthma. In the field of psychiatry, researchers are breaking new ground examining how expectation and conditioning make placebos work for many patients. This particular area of study may ultimately help clinicians make smaller doses of potentially toxic drugs more effective.
"Increasing a patient's sense of optimism about a good outcome may actually increase the effect of an active drug," said Arthur Kleinman, MD, professor of psychiatry and medical anthropology at Harvard Medical School. "At the very least, we should be training physicians to get the maximum placebo response from their treatments."
Dr. Kleinman co-chaired an NIH conference, "The Science of the Placebo," in 2000 that examined a substantial, growing body of science about placebos. Researchers now believe that placebos do not work just one way for different conditions. Moreover, they say, it remains difficult to predict which patient may benefit from one.
For clinicians, therefore, determining how to harness the placebo effect for their patients is tricky. Could the Long Island interns, for example, have given their patients placebos in a manner that prompted them to expect pain relief and to relax, causing their bodies to produce endogenous opioids, thereby truly reducing their real pain? Could the fact that patients reported relief mean that placebos worked, not that patients were initially lying about their pain?
Another school of thought views the issue more suspiciously, stating that the placebo effect is nothing more than one of medicine's most persistent myths. An article in the May 24, 2001, issue of The New England Journal of Medicine gave skeptics ammunition.
Researchers analyzed clinical trials in which patients were randomly assigned to groups where they received a placebo or no treatment. They concluded that except for pain, there is "little evidence in general that placebos had powerful clinical effects." Their conclusion appeared to be even more damning, stating that "outside the setting of clinical trials, there is no justification for the use of placebos."
These researchers explained away the oft-cited benefits of placebos by citing several factors: the natural course of illnesses in which some people get better on their own, statistical regression to the mean, and other factors such as changes in diet or exercise during a clinical trial.
Critics contend that the study's broad conclusion was flawed because researchers looked only at placebo use in clinical trials. Others criticized the statistical methodology behind the meta-analysis, particularly the way the authors lumped together all randomized controlled trials rather than looking at diseases for which placebos are considered useful.
"Placebos don't work equally well for everything. We know that," said Walter A. Brown, MD, clinical professor of psychiatry at Brown Medical School and Tufts University School of Medicine, and author of a much-quoted Scientific American article, "The Placebo Effect," which was published in January 1998.
"It would be like studying how well penicillin works for pain, congestive heart failure and bacterial pneumonia," he added. "You would find that it worked pretty well for bacterial illnesses but not for these other things. If you grouped them all together, the final result would be that penicillin doesn't work at all."
Worries about deception
If the evidence about placebos is murky, one thing is clear: Placebo and the placebo effect are not the same thing.
In modern American medical practice, few argue that doctors in regular clinical practice should give patients an inert substance like a sugar pill or a saline injection while lying to patients who think they are getting something more. Bioethicists frown on such deception.
Some worry that when physicians lie to patients, it is often to serve the doctor, not the patient. (The interns at Winthrop-University Hospital, for example, gave fake drugs to patients largely to get them to go away.)
The Winthrop-University Hospital bioethics committee drafted a position that placebo use is ethically acceptable outside of research studies only if the patient is aware of the use in advance and if informed consent is obtained. The policy also notes that "patient deception, even for patient benefit, is generally ethically inappropriate."
Admittedly, the instances in which placebos can be given without patient deception in regular clinical care—not research—are rare. Proponents of placebos, however, regularly cite two examples.
First, physicians and their patients can work together to reduce the amount of an active drug the patient takes by interspersing placebo pills in the medicine vial. A large body of literature about conditioned pharmacotherapeutic responses contends that when people take a pill with a certain look, smell and taste, they become conditioned (think of Pavlov's dogs) to expect a certain effect. If you give a placebo pill that has that same look, smell and taste, patients will often react as if they took the real pill.
"In theory, patients can go on an 80% reinforcement schedule, meaning that 80 out of 100 pills are real," explained Robert Ader, PhD, professor of psychosocial medicine and director of the division of behavioral and psychosocial medicine in the department of psychiatry at the University of Rochester. "The patient gets 20% less drug, and conditioning effects make up the balance."
A second nondeceptive use for placebos in a clinical setting can occur when a clinical trial has one subject, known as an "n of 1" clinical trial. Jeffrey Burack, ACP-ASIM Member, assistant professor of bioethics and medical humanities at the University of California, Berkeley, and a physician at East Bay AIDS Center, offered the following example: A doctor tells his patient he is unsure whether a particular medicine is helping. So the doctor and the patient agree that they will experiment with two sets of pills, one real and one placebo. The patient will blindly take one type of pill for two weeks, then the other for another two weeks and keep track of symptoms. At the end, they will compare notes and determine whether the active medicine helped.
"This is perfectly justifiable, ethically," Dr. Burack said. "A placebo is being given to answer a clinical question that the physician and patient share. There's no deception involved."
Even if you resolve the ethical issues of replacing "real" drugs with placebos, however, practical considerations remain: Placebos are difficult to get. "I don't have bottles of sugar pills in my office, and I can't get a pharmacist to give them," Dr. Burack said. "What are you going to write on the prescription? How is the bill going to be submitted? What is the patient's insurance going to do?"
If placebos themselves are rare, the placebo effect may be common. That's because patients do not necessarily have to receive an inert substance to experience a placebo response during their clinical care.
Howard Brody, MD, PhD, professor of family practice and philosophy at Michigan State University, has argued that doctors should do more to enhance the placebo effect as a valuable patient care tool. Dr. Brody, author of the 2000 book, "The Placebo Response," explained that the placebo effect can be achieved by invoking the power of the therapeutic environment, exploiting patient expectations of being helped, and cultivating positive thinking and patients' feelings of control. In other words, he said, physicians can view the placebo effect as anything they do during their patient interactions to help cure disease or alleviate symptoms through something other than a biochemical effect.
Stephen E. Straus, MD, director of the NIH's National Center for Complementary and Alternative Medicine, explained that conventional American medicine all too often focuses only on the negative aspect of placebo use in clinical practice, while discounting positive aspects of the therapeutic encounter. "We are neglecting one of the most powerful elements in the therapeutic interaction: harnessing an individual's will to remain well and get better," he said.
For Dr. Straus, placebos and the placebo effect can be useful in addressing the illness part of disease. He suggested thinking of them as one way to help patients live with—rather than cure—their disease. Harnessing the mechanisms by which placebos work, he said, may play the biggest role in helping patients deal with subjective complaints that often accompany disease, including mood, energy level, pain and nausea.
Even the May 2001 New England Journal of Medicine article, which largely discounted the utility of placebos, found a significant difference between the placebo group and the no-treatment group in trials that tracked "subjective outcomes" (where patients self-report symptoms), such as studies of anxiety and insomnia.
Research has also shown that for patients who strongly prefer one treatment over another, placebos may have a legitimate place in a clinician's treatment arsenal. A study in the journal Spine (July 2001), for example, concluded that "specific expectancies about specific therapy were a very high predictor of getting better," explained Michigan State's Dr. Brody.
"When I give a patient a prescription, I'm creating an expectation in the patient about what it will do," Dr. Brody said. "On some conscious or unconscious level, I'm also reminding the patient of every single pill he or she took in the past. I evoke both conditioning and expectancy simultaneously."
Rhode Island's Dr. Brown pointed to several things physicians can do in ordinary clinical practice to enhance the placebo response, none of which require patient deception. Display diplomas and medical instruments and wear a white coat, because the "theater" of a professional medical office has been shown to give patients confidence that they are in the hands of a professional who knows how to relieve their symptoms. Give all patients a diagnosis and a prognosis whenever possible so they feel as if they have some control over their symptoms and the course of their illness. Write down specific directions on official prescription pads, even if you're telling them to take an over-the-counter medication for a set amount of time.
Dr. Brown said he thinks there is a place for physicians to deliberately treat some people with placebo pills. He pointed to research showing that 25% of patients with moderate hypertension get into the normal range of blood pressure with placebo
"I don't think it would be crazy to offer a placebo treatment to somebody who really doesn't like the idea of taking a diuretic," he explained. "You could say, 'Here are some pills. They don't have any active drug in them, but a good percentage of people get better with them. We don't know exactly why.'"
Many patients would be interested in that type of experiment, he predicted. "They don't care what's in the medicine they get as long as they get better," Dr. Brown said. "The idea of getting an inactive treatment that may stimulate your body's own healing processes is very appealing to a lot of people."
Although Michigan State's Dr. Brody said he wouldn't go as far as Dr. Brown, he did say that many physicians discount what is often a strong placebo effect in "real" drugs they prescribe for patients.
"When I give patients an antihypertensive, in my mind they're getting a chemical that will interact with their biochemistry," he explained. "If they get better, that's what explains it. Every medication probably has a placebo component that is partly—if not largely—responsible for the effect that we see in the patient. And that's the scary part."
Deborah Gesensway is a freelance writer in Glenside, Pa.
Internist Archives Quick Links
Have questions about the new ABIM MOC Program?
One Click to Confidence - Free to members
ACP Smart Medicine is a new, online clinical decision support tool specifically for internal medicine. Get rapid point-of-care access to evidence-based clinical recommendations and guidelines. Plus, users can easily earn CME credit. Learn more