I find the placebo effect fascinating. In the London Review of Books‘ Scrivener’s Palsy, Carl Elliott reviews Constructing RSI: Belief and Desire by Yolande Lucire and Meaning, Medicine and the ‘Placebo Effect’ by Daniel Moerman and describes the importance of the doctor‘s belief in the placebo:
Whatever the placebo effect is, it isn’t reliably predictable on the basis of the characteristics of individual patients.A much better predictor is the characteristics and qualities of individual doctors. The more convinced a doctor is that a drug or a placebo will work, the more likely that it really will. The pioneering study here was conducted almost forty years ago. Physicians were treating anxious outpatients with meprobamate, or Miltown, a popular tranquilliser of the 1950s and 1960s. In two clinics, patients given Miltown became no less anxious than patients given placebos. But in the third, patients on Miltown became much less anxious. The difference was that doctors in this clinic had self-consciously adopted an enthusiastic, confident attitude towards the drug’s effectiveness. When they switched to a more neutral, experimental attitude, Miltown was no better than a placebo.
A more elegant study to the same effect was published by Richard Gracely and his colleagues in 1985. Gracely purported to be studying a pain reliever called fentanyl in patients recovering after having their wisdom teeth removed. But he was more interested in whether a clinician’s attitude relieved pain. So he recruited unknowing clinicians to administer the treatment. Gracely and his colleagues told half the doctors that they would be administering one of three possible treatments: fentanyl (a pain reliever), naloxone (a drug that blocked opiate receptors and could make the pain even worse) or a placebo. But he told the other half of the clinicians that there had been an administrative problem, and none of their patients would be getting fentanyl.
The results were striking. All the patients got placebos, and all were told the same thing about their chances of getting placebos. But their pain response differed tremendously. How it differed depended on what their clinicians thought they were administering. If a doctor thought there was a one-in-three possibility that he was giving the patient fentanyl, the patient was likely to feel a lot better after an hour. But if a doctor thought there was no chance he was giving the patient fentanyl – that he was giving either placebo or naloxone – then the patient’s pain was likely to get worse. Somehow, the clinicians were unknowingly transmitting their attitudes towards the medication to their patients.