Simply participating in a medical-research trial sometimes improves a person’s health. That’s why investigators use placebos—inert pills or other dummy treatments—and make both study participants and staff unaware of whether a person is receiving an active treatment or not. Anywhere from 10 to 100 percent of the people taking placebos in trials see their symptoms wane. In such a test, a drug is considered to be beneficial only if it can beat the placebo.
Many studies suggest that problems like pain and depression respond particularly well to placebos. Blood pressure, cholesterol concentration in the blood, and heart rate are also affected by placebos, as are warts. On average, about a third of people taking placebos in studies report a benefit.
Researchers have typically measured this placebo effect for use in their statistical analyses, but recently they’ve become increasingly interested in understanding the effect. A conference last November at the National Institutes of Health in Bethesda, Md., highlighted that shift in attitude.
“The placebo effect has had a very pejorative meaning in the past. Instead of people trying to capitalize on it, they’ve discarded it,” says meeting co-organizer Linda W. Engel of the NIH National Center for Complementary and Alternative Medicine in Bethesda, Md.
“Only since about 1960 have people really recognized the placebo effect as a phenomenon, and it has generally been seen as a hindrance,” agrees Shepard Siegel of McMaster University in Hamilton, Ontario. “Only recently, as exemplified by the NIH conference, have people begun to look at [the placebo effect] as a potentially interesting effect,” he adds.
Scientists at the conference reported findings suggesting that response to a placebo may rest on both the meaning that a person associates with a particular treatment and his or her previous experience with medication.
Subscribe to Science News
Get great science journalism, from the most trusted source, delivered to your doorstep.
Placebos provide “an intriguing domain to ask broader questions about mind-brain relationships,” says Arthur Kleinman of Harvard Medical School in Boston. Those issues include, he says, “how culture affects the success of treatment, why some diseases respond better to placebos than others, and how stress and mood affect biologic interactions in the body.”
Some of the improvement seen among people taking placebos in drug trials may be spontaneous and reflect random fluctuations in their disease symptoms, says Kleinman. People participating in a trial may also be more likely to report positive results from whatever treatment they’re given because they think that’s what the researchers want to hear.
But other factors are at work, too. People getting placebos in clinical trials still get counseling about their disease and other care, says Daniel E. Moerman of the University of Michigan-Dearborn. “It’s not like nothing happened to them,” he says. “Words aren’t inert.”
Psychological studies have suggested that the “meaning” of a treatment matters, Moerman says. For example, people receiving a placebo that they’re told is morphine report more pain relief than people given placebos masquerading as aspirin. Likewise, placebos that people think are brand-name aspirins reduce headache pain more effectively than placebos disguised as generic aspirin.
These studies show that people’s perceptions of their treatment play an important role in healing, Moerman says. Few studies have compared active treatment with both placebo treatment and no treatment at all, he says. However, among the studies with such a design, most have shown that a group receiving a placebo is likely to improve faster than an untreated group, he says.
In the March 2000 Medical Anthropology Quarterly, Moerman analyzed dozens of publications on ulcer, hypertension, and depression treatments.
He found that the strength of the placebo effect varied from country to country. For example, placebos were more effective in healing ulcers in Germany than in Brazil. However, Germans were less likely than their counterparts in other countries to benefit from placebos administered during trials of drugs to reduce high blood pressure. Participants in all the studies that Moerman reviewed were about equally likely to benefit from placebos administered in trials of antidepressants.
The findings are tentative and unexplained, says Ted J. Kaptchuk of Beth Israel Deaconess Medical Center in Boston, but they should be explored further. Some researchers have suggested that any treatment, active or not, works better when doctors tell the patient, “This is powerful medicine.” For a physician to say that about a placebo, however, conflicts with the modern ethical mandates to always give patients the unvarnished truth, says Anne Harrington of Harvard University.
A conditioned response
The placebo response is both a conscious response to the meaning of therapy and a conditioned response to treatment from a health professional, some researchers contend. In other words, people not only think that a drug or procedure will benefit them, but over time, they unconsciously associate the taking of medicine with relief.
“The taste and sight of a drug can elicit a therapeutic effect,” says Siegel. Proof of such conditioned responses, he says, is that “you can get something that looks like a placebo effect in a rat.”
Robert Ader of the University of Rochester in New York has shown positive responses to placebos in experiments on animals with an autoimmune disease. Once a week, Ader gave mice in each of four groups with a saccharin solution to drink. That treatment was followed in one group by injection of saline and in a second group of the drug cyclophosphamide at a dose that prolonged the animals’ survival.
To see whether a placebo could elicit a conditioned response from the mice, Ader followed the sweet drink in the third group with injections of cyclophosphamide or saline in a random pattern such that each mouse got two doses of the active drug in a month. This group showed the same decrease in mortality as the mice getting weekly doses of drug.
A fourth group was given the same schedule of cyclophosphamide and saline injections, but on different days of the week than the sugar injections, to prevent any association between the treatments. The mortality of this group matched that of the group getting no active drug. Thus, the half-dose of cyclophosphamide by itself was inactive, Ader says. But he concludes a conditioned response in the third group permitted this smaller amount of drug to reduce mortality.
He has preliminary, unpublished evidence that the same response occurs in people. “We have every reason to believe that one can use conditioning to effectively reduce drug dosage while maintaining the desired response to therapy,” Ader says.
In cases where an active drug has serious side effects, for example, such a placebo use might enable doctors to reduce the drug dosage without eliminating the treatment’s beneficial effect, he says. However, if patients are aware they are getting a placebo, it may weaken the effect, Ader says.
It isn’t always easy to differentiate between the effects of conscious expectations and conditioning on the placebo effect, says Irving Kirsch of the University of Connecticut in Storrs. In fact, he says, “many of the same responses can be explained by both mechanisms.”
Italian researcher Fabrizio Benedetti of the University of Torino has shown that both mechanisms can play a role in placebo effects in people. In his studies, Benedetti exposes volunteers to narcotics or other painkilling treatments before looking at a placebo effect.
He and his colleagues have reported that applying a placebo—a gel that actually contained no painkiller—to a person’s hand reduced sensitivity to pain administered to their hand but not to their feet. These findings show that the placebo effect depended on the expectations that people had for the gel, says Benedetti.
In contrast, he showed in a different study that placebos for narcotic painkillers slowed people’s breathing, even in volunteers who didn’t know that this is a typical side effect of narcotics. The phenomenon suggests that conditioning also plays a role in placebos’ effects, says Benedetti.
Interested in the underlying physiological effect of placebos in the body, he’s found that people’s responses to placebos can be blocked by an antinarcotic drug called naloxone. It blocks the effect of both narcotics and endogenous opioids, or natural painkillers.
Moerman calls Benedetti’s studies “incontrovertible physical evidence” of a placebo effect.
As another approach to detecting physical responses underlying either conditioned or conscious placebo effects, scientists at the NIH meeting called for new research using brain scans to look for activity in certain areas of the brain after a person has received a placebo.
“A lot of the current enthusiasm for this idea of ‘placebo as medicine’ comes within the context of a belief that we may now be in a position to elucidate mechanisms underlying the placebo effect,” says Harrington.
Better understanding of the placebo effect raises questions about the design of traditional placebo trials. Ader suggests that because conditioning is part of the placebo effect, trials in which volunteers take either the placebo or a drug and then switch won’t provide a true estimate of the placebo effect.
He proposes that researchers design trials in which half the participants are told they are getting a placebo and half are told they’re receiving the active treatment. Yet each of the two halves of the study would be split again, so half of each group would get the active drug and half the placebo. Thus, some people would mistakenly believe they were receiving the placebo when they were actually taking the active drug, and vice versa.
Researchers are also considering the use of so-called active placebos, compounds that lack the effective ingredient in a drug but may cause some of its side effects. Not only would this make it more difficult for people to guess which treatment they were getting, it might actually increase the power of the placebo, says Kleinman.
Some types of placebos may be more effective than others, Kaptchuk says. For example, some experiments find sham surgery more effective at reducing pain than placebo pills are, he says. Kaptchuk is examining whether sham acupuncture is better than placebo pills for treating a chronic pain condition.
Better understanding of the placebo effect could influence physicians’ current clinical practices. Giving antibiotics to people with colds is a common example of what is, most physicians agree, application of the placebo effect.
In fact, “placebo use outside of clinical trials used to be rather common,” says Sissela Bok of the Harvard School of Public Health in Cambridge, Mass. As late as the early 20th century, doctors often prescribed tonics they knew were inactive and others with little proven medical value. No one knows how often doctors today knowingly prescribe placebos, says Bok.
Several researchers, such as Kirsch, have endorsed placebos for treating people with depression. In this disease, the scientists claim, placebos in clinical trials have been almost as effective as antidepressant drugs.
“The placebo is becoming fashionable” in popular culture, says Harrington. She notes increasing numbers of headlines such as this one on the cover of the Jan. 9, 2000 New York Times Magazine: “Astonishing Medical Fact: Placebos Work! So Why Not Use Them as Medicine?”
But, she cautions, “there are real questions about the need for deception and the role of deception in achieving benefits. If you’re not a bit troubled by the placebo effect, as well as intrigued by it, you haven’t been paying attention.”
Rising concerns about placebos
The gold standard of medical research is the double-blind, randomized, placebo-controlled trial: a test in which researchers randomly assign volunteers to one of at least two groups. People in one group get the drug or other treatment being tested, and people in another group get an inactive treatment—and during the trial, neither researcher nor volunteer knows which is which. Such procedures rule out inadvertent bias in assessing an experimental drug and can give clear-cut proof that a drug is effective.
In the United States, the Food and Drug Administration usually asks companies to perform such studies to prove the safety and efficacy of new drugs and some medical procedures. Placebo use is generally required when there’s no current treatment for a condition. Further, researchers often use placebos to test new treatments for high blood pressure, allergies, and depression and some other mental disorders—conditions for which the risk of serious harm is low if patients briefly discontinue currently available treatments. However, for ethical reasons, researchers must avoid using placebos when currently available drugs are known to save lives—such as in tests of drugs to prevent heart failure.
Last October, however, scientists assembled in Edinburgh under the aegis of the World Medical Association amended the Declaration of Helsinki, a document that has defined medical ethics internationally since World War II. This amendment calls for new drugs to be tested against “the best current” treatment—a statement that many researchers say implies that using placebos is unethical even when the risk of harm is low.
While the document has no legal authority in any country, it exerts considerable moral sway over national governments and ethical-review boards around the world, according to Kenneth J. Rothman of the Boston University School of Public Health.
“If there were no ethical concerns with placebo use, we would all endorse it” because of the scientific benefits of placebo-controlled trials, he says. However, even briefly taking people off drugs that lower blood pressure or prevent psychotic episodes may in fact be harmful, he says. In such cases, informed consent from a volunteer—saying he or she understands and accepts the risks of participating in a clinical trial—may be ethically meaningless, Rothman adds.
The Helsinki amendment suggests that “the rights of the individual patient [to the best available treatment] take precedence over the rights of science and society in general,” Rothman concludes.
Many researchers—and bioethicists—disagree. In July 2000, before the Helsinki declaration was amended, the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use released a guidance document explicitly stating that placebo use is, in general, ethically acceptable. This group of representatives from the drug industry and regulatory agencies, in contrast to the Helsinki group, concluded that placebo-controlled trials are justifiable when researchers can reasonably assume that patients will experience, discomfort but not serious harm.
“There’s a lot of anger about the Helsinki document,” says medical anthropologist Arthur Kleinman of Harvard Medical School in Boston. “These ethical issues go to the heart of biomedical research.”
As alternatives to using placebos, researchers can study different doses of the drug being investigated or compare an experimental drug with a currently approved drug. However, such trial designs have limitations, says Robert J. Temple, director of medical policy at the FDA’s Center for Drug Evaluation and Research in Rockville, Md.
If poor study design—such as not enrolling enough patients to show a significant difference—causes the study to show no differences between two treatments when there actually is one, a placebo-controlled trial would be regarded as showing no benefit of the experimental treatment but a test of a current drug versus an ineffective new treatment might be interpreted as proving that the treatments are equal in value. “You don’t have to be cynical to think that’s risky,” Temple says.