One of my first medical publications was on the nocebo effect, the unpleasant symptoms patients may suffer as a result of being made aware of potential side effects of a treatment they are about to receive or a procedure they are to undergo. Thus patients who were having a lumbar puncture were either told or not told they might suffer a headache afterwards; and lo and behold, those who were told that they might get headaches duly got headaches while those who were not told didn’t.
On the whole, as an article in a recent edition of the Journal of the American Medical Association points out, doctors are well aware of the placebo effect, that is to say the good that their treatment may do patients by means of mere suggestion, but have little awareness of the opposite nocebo effect, the harm that their treatment may do their patients by mere suggestion.
The nocebo effect poses an ethical dilemma for doctors, say authors of the article. On the one hand, doctors are supposed to do their patients no harm; on the other, they are supposed to be open and honest with their patients about the potential harms of drugs and other treatments. The dilemma is this: foreknowledge of those harms can harm some patients. Should the need for honesty trump the ethical injunction to do no harm?
It is not as if the nocebo effect is only occasional or of no practical effect. In one trial of an anti-depressant, nearly half the patients who were given placebo developed unpleasant side effects of the kind that they were told in advance that they might develop with the active drug, and five percent of them withdrew from the trial because these supposed side effects were so unpleasant.
The authors suggest ways of reducing the nocebo effect, though short of total dishonesty it cannot be eliminated altogether. There is evidence, indeed, that, as general public knowledge of side effects gains currency so does the nocebo effect increase in frequency and severity.
At the moment, patient information leaflets are obliged to list all side effects, even those that are suspected rather than proved, reports of some of which may be themselves the product of the nocebo effect, for news of side effects spreads fast in these days of instant communication. But the patient-information leaflets give little indication of the frequency of the listed side effects, or the degree of certainty about the causal connection between the side effects and the drug, so that that the patient is left with no real knowledge of the statistical likelihood of being afflicted by them.
The authors suggest that, to counteract the nocebo effect, doctors always append information about the potential harmful effects of their treatment with information about the good it will do. The example they give of this is to tell the patient, on prescribing medication that will lower his blood pressure, that it “will reduce his chances of a stroke by 20 percent.” In other words, accentuate the positive.
This is an interesting example because, as given, it is so grossly incomplete as to be almost dishonest. A reduction of 20 percent within what period? A month, a year, a decade, a lifetime? Impossible to say from the wording.
Moreover, to make a proper decision, the patient needs to know the absolute risk and not just the relative one. A 20 percent reduction may be worth pursuing or it may not, depending on the size of the initial risk. Only the patient can decide whether the benefits outweigh the risk for him: and even then, in my experience, he may not be able to.