In medicine, especially in clinical trials conducted for medical research, a placebo is a substance used as a control in a double-blind study. Half of a group of test subjects are given a medicinal substance being investigated, while the other half is administered an inert material, like a sugar pill, made to look indistinguishable from the medicine. In the optimal double-blind test, neither the research staff nor the test patients are allowed to know which is which until the study has been completed. By this process, psychological effects of the placebo are hopefully kept separate from the biological effects of the chemically active agent being tested.
The non-medical definition of the word placebo indicates the general phenomenon called the placebo effect. Any action, such as gift-giving, which is intended to soothe an agitated person without directly solving any problem is referred to as a placebo. As far back as the sixteenth century, the writer Montaigne commented that a patient's faith in a doctor had more bearing on the successful outcome of a therapy than any other factor.
The initial and often ongoing symptom being treated by a physician is pain, whether or not the cause of this pain is curable or even treatable. Sometimes treatment for an illness such as cancer leads to painful side effects, which must be tended. Only recent studies have begun to unlock the secrets of endorphins, analgesic or pain-reducing chemical agents produced by the human brain. They serve the same purpose as morphine, a narcotic first extracted from the poppy in the 1800s, and long used as an analgesic and anesthetic. There are still many questions as to what triggers an increase of endorphins in the body, how this contributes to the placebo effect, and how much endorphin production may be consciously controlled by a patient.
Other causes of pain are psychosomatic: stress-related, neurotic or phobic reactions with no detectable organic origin. Chronic discomforts describable as psychosomatic include allergies, ulcers, and hypertension. These conditions not only respond positively to placebos, they can also arise in a patient after taking a placebo, as negative aftereffects. Attempts to isolate a typical "placebo personality" have yet to succeed in predicting if any one person might be more susceptible to the placebo effect than another.
Even surgery can be used as a placebo, by cutting open a patient under anesthesia without actually operating. Control groups among angina sufferers have reported a decrease in chest pains after such "dummy" surgery, which indicates that angina may be at least partially psychosomatic. The problem with extreme placebos is the ethical issue of leaving any one patient untreated for the sake of being a control. The Tuskeegee syphilis experiment conducted in Alabama during the late 1930s is one example of an extreme clinical trial, during which penicillin was deliberately withheld from certain patients without their knowledge. While a few of these untreated patients survived, others died painful and preventable deaths.