The Mission of the Manhattan Institute is
to develop and disseminate new ideas that
foster greater economic choice and
individual responsibility.

The Wall Street Journal.

Medical Mistakes Are No Cause for Alarm
July 18, 2002

by Theodore Dalrymple

The history of medicine is a record not only of brilliant success and stunning progress: It is also a litany of mistaken ideas and discarded treatments, some of which came to appear absurd or downright dangerous after having once been hailed as unprecedented advances.

Last week, thanks to research published in the New England Journal of Medicine and the Journal of the American Medical Association, two fashionable treatments will soon go the way of blood-letting and moxibustion: the arthroscopic washout of arthritic knee joints, and hormone replacement therapy to prevent disease in postmenopausal women. Both seem to do more harm, at least to the patients, than good; though it must be said that, at $5,000 a time, the knee operations -- 650,000 of them in America per year -- have benefited orthopedic surgeons enormously. I am quite sure that many of them feel a lot better after having performed one.

In the paper published in the New England Journal about knee surgery, real operations were compared with sham, or placebo, ones. It turned out that the real ones were not superior to the sham in relieving the pain of arthritis. If anything, their results were worse.

It was also a bad week for hormone replacement therapy. Over a third of healthy postmenopausal American women receive it, in the hope that it will prevent such conditions as osteoporosis and heart disease. The clinical trial published in JAMA established pretty clearly that hormone therapy does more harm than good. The women who received it were more likely to get heart attacks or strokes, blood clots on the lung, or cancer of the breast than those who took placebo. On the other hand, the therapy preserved them from cancer of the colon and osteoporosis.

The differences revealed by the trial were relatively small. Overall, there were no fatalities attributable to the therapy. And the trial indicates that if 10,000 women were given hormones for a year, only 19 of them would suffer from a serious medical condition that they would not otherwise have had. However, hormone replacement is a long-term treatment, and over ten years the figure would presumably rise to 190, or nearly 2%. And we should always keep it in mind that it is for proponents of hormone therapy (or any other treatment) to prove that it does good, not for opponents to show that it does harm.

So millions of Americans, it seems, have been expensively operated upon or have received therapy that not only has failed to do them good, but that might have caused them harm. Isn't this a scandal?

Well no, actually. Perhaps the trials could have been mounted earlier, but such trials are fiendishly difficult and expensive to organize, and in the meantime the physicians acted in good faith. There were theoretical reasons for believing that the knee operations would relieve arthritic suffering, and many patients reported precisely such relief afterwards. It subsequently turned out that this was only a placebo effect, but it is worth remembering that the placebo effect is what it says it is, namely an effect. Moreover, doctors always have to act in the light of imperfect knowledge.

The hormone trial, however, brings home a very important point about an increasing proportion of modern medical therapy: Very large numbers of people are being treated or screened in the hope of benefiting a very small number. Furthermore, the ones who have actually benefited from it will remain forever unknown. This is very different from taking antibiotics for pneumonia or insulin for diabetes.

Let us suppose for a minute that the hormone trial had demonstrated -- as it was possible that it might have done -- that the therapy did as much good as it now appears to do harm. Could a doctor then wholeheartedly recommend that 100 postmenopausal women take a treatment for ten years so that two of them could avoid a serious but not fatal medical condition? The individual women would have to decide for themselves, and there would be no indubitably correct decision.

Treatment for high blood pressure is given on the same logical basis. The chances that treatment will do those with moderately raised blood pressure harm rather than good are very significant, since side-effects are common and benefits rare. On the other hand, the side-effects are not very serious, but the benefits (the avoidance of stroke) are great. However long a patient takes his treatment, though, he will never be able to say that it has helped him. The harms are palpable, the benefits impalpable.

Screening is another example. Few people (I suspect) realize how shaky is the ground on which much modern medical screening is based. Most people think that screening for early disease can do no harm, but this is not so. No test is perfectly accurate, and so it gives both false negative results (leading some patients to live in a fool's paradise) and false positive ones (leading some, or indeed many, to undergo unnecessary further investigations and in the meantime to suffer from needless anxiety), all for a very small chance of actual benefit.

What is new in medicine is not error, for error springs eternal and is the price of progress, but the speed and thoroughness with which treatments are debunked. Blood-letting lasted two millennia before it occurred to anyone (a Frenchman called Pierre-Charles-Alexandre Louis) to test systematically the claims made on its behalf and find them wanting. Now, thanks to the method of the double-blind trial, disillusion with even the most lucrative methods of treatment sets in after only a relatively few years of useless or harmful activity. That is progress.

Theodore Dalrymple is the pen name of Anthony Daniels, a British physician and author of "
Life at the Bottom" (Ivan R. Dee, 2001).

©2002 The Wall Street Journal



Home | About MI | Scholars | Publications | Books | Links | Contact MI
City Journal | CAU | CCI | CEPE | CLP | CMP | CRD | ECNY
Thank you for visiting us.
To receive a General Information Packet, please email
and include your name and address in your e-mail message.
Copyright © 2009 Manhattan Institute for Policy Research, Inc. All rights reserved.
52 Vanderbilt Avenue, New York, N.Y. 10017
phone (212) 599-7000 / fax (212) 599-3494