The Best Medicine

In fact, we know a good deal about what placebos can do for people with depression, because over the past 40 years thousands of depressed patients have been given placebos in double-blind clinical trials—research studies in which a new drug is compared to a placebo, and both the patient and doctor are "blind" to which is being administered. In the most severe depression—the kind that makes it impossible to carry out the simplest tasks and requires hospitalization—placebo treatment is less effective than the standard treatments, antidepressant drugs and electroconvulsive therapy (ECT), or "shock treatment." About 20 to 30 percent of severely depressed people recover with placebo treatment, whereas 60 to 70 percent recover with antidepressants and 80 to 90 percent with ECT.

But in the most common type of depression—the milder version that affects almost 20 percent of people at some time in their lives but that doesn't require hospitalization—placebo pills are almost as effective as antidepressants. About 60 to 70 percent of such depressed patients recover with antidepressant medication, 30 to 70 percent with placebos. The less severe the depression, the more likely it is to improve with placebo treatment.

Placebos also stack up well next to another standard treatment for less severe depression, the various types of psychotherapy—interpersonal therapy, behavior therapy and, most prominently, cognitive therapy. Comparisons between these treatments and placebo pills have consistently shown that placebos are as effective as any of the psychotherapies, at least in the short term.

Why Do They Work?

Many diseases, such as the common cold, improve on their own. So some patients who get better with placebo treatment—and some who improve with standard treatment—would have gotten better if left to themselves. But placebo treatment is not simply the same as no treatment; it offers more than the passage of time. For example, although up to 70 percent of depressed patients improve when treated for a few weeks with a placebo, depressed patients awaiting treatment for a few weeks rarely improve. And when people experiencing pain are treated with a placebo they get maximum relief not at various times after treatment, as would be expected if "pain relief" was a result of spontaneous fluctuations, but one to two hours after treatment, just as they do with "real" pain medicine. Yet another example: A recent study in the European Journal of Psychiatry shows that people with schizophrenia who take placebo pills are less likely to relapse than those who receive no treatment.

So a placebo offers something therapeutic. What are its active ingredients? Jerome Frank, M.D., Ph.D., a psychotherapy scholar and author of Persuasion and Healing, pointed out that placebo treatment has the four features shared by all psychotherapies, features he believes are the curative elements: a person in distress; an expert; an explanation for the condition; and a healing ritual. According to Frank, these features reverse demoralization and promote the expectation of recovery.

Expectation undoubtedly makes an important contribution to the healing power of placebos. Across a wide range of illnesses, patients who expect to improve are more likely to improve. A recent British study, for example, examined the effects of both reassurance and expectation on patients with physical complaints but no identifiable disease. The doctor told some of these patients that no serious disease had been found and that they would soon be well. He told others only that the cause of their complaint was unclear. Two weeks later, 64 percent of the first group but only 39 percent of the second had recovered.

Expectation operates more specifically as well. For example, when people were given an alcohol-free drink but told that it contained alcohol, they often felt intoxicated, acted intoxicated, and showed some of the physical signs of intoxication. In another study, when people with asthma were given an inhaler containing only saltwater and told that they would be inhaling an irritant or allergen, they reported increased airway obstruction and had more difficulty breathing. But when told that the same inhaler had a medicine to help asthma, their airways opened up and they breathed more easily.

Yet another study reveals how expectations enhance the effects of "real" medicine. Patients with asthma were given a bronchodilator, a substance that opens airways and makes breathing easier. Its effect was twice as great when people were told that it would open up their airways and make it easier to breathe than when they were told it would tighten their airways and make breathing difficult.

Placebos that harm, as in the airway obstruction example above, are sometimes called "nocebos." Here, the expectation that one will get ill can actually produce illness, as with symptoms that are spread by being witnessed. Similarly, placebos can produce side effects; in a recent study of 303 men with enlarged prostates, more than 80 percent reported adverse effects while taking "inactive" pills.