Podium: Anthony Daniels: A little medical knowledge is a dangerous thing

From a lecture given by the medical practitioner and writer at a Lanesborough Lunch in London on the Risk of Freedom

THE MISUSE of science is a large subject. I don't intend to deal with the question of the correct or appropriate use of science, which might at first sight seem an essential preliminary to any discussion of the misuse of science. I am not convinced that it is, however; almost everyone can recognise the misuse of public funds but few can define their proper use.

The field with which I am most familiar is medicine, a field in which the interplay of ends and means is peculiarly complex, and in which the possibilities of the misuse of science (and here I include technology that is science based) are almost infinite.

Of course, if there is a supply of bad science, or of science that leads to wild speculation dressed up as science, the chances are that there is a demand for it also. Let me illustrate what I mean in the case of epidemiology. Properly used, its achievements are considerable. The identification of Aids, as well as the demonstration that it was most likely caused by a virus, was a recent triumph of epidemiology. But there is little doubt that there is much misuse of the science. It often seems to have forgotten the lessons taught it by one of the founders of modern medical statistics, Sir Austin Bradford Hill. Time and again we read of a statistical correlation between disease A and lifestyle or environmental factor B, and however much lip-service we pay to the principle that a statistical correlation does not imply a causative relationship, we end up thinking that A is caused by B, and that if only we avoid B we shall avoid A.

Bradford Hill lay down certain principles to establish whether a statistical correlation was likely to imply causation. In the vast majority of studies his principles are now forgotten entirely. But the fact is that if you take a disease, call it A, and examine enough factors in the lives of people who suffer from it, you must, for purely statistical reasons, come up with some that are correlated with it. And when you find these factors, you report them in medical journals, omitting to mention all the factors that you examined and found not to be correlated with disease A.

Your findings are spotted by the eagle-eyed media of mass communication, who disseminate them to a very wide audience. This audience is completely uncritical and does not remember that last week another environmental factor was correlated with the very same disease. And action soon results; when it was suggested not long ago that people with ischaemic heart disease had lower levels of selenium in their blood than non-sufferers, walnuts swiftly disappeared from supermarkets because they were said to contain a lot of selenium.

Of course, not all correlations are equal, and not all epidemiological evidence is treated in the same way. Much research is conclusion-driven rather than purely investigative.

Pasteur said that, in the field of observation, chance favoured only the prepared mind. The same might be said of epidemiological pseudo-information; it affects only those whose minds (and emotions) are prepared. And when it comes to health matters in these days of greatly expanded life expectancy, the public attitude may be described as one of unrealistic expectation tempered by baseless fears.

There is a public that demands - or allegedly demands - this kind of information. Of course, the more they get used to it, the more they will demand it. Perhaps people want to feel that their world is extraordinarily hazardous - danger is one way of investing your life with significance, after all - when in fact life is far safer than it has ever been in the history of the world. But we invent epidemiological hazards as children invent goblins.

What is clear is that in our current climate, the kind of misuse of science, or of scientific information, that I have described is bound to flourish. And the only antidote is not yet more information, but wisdom, which in this instance consists of a sense of perspective. For information without perspective is like a graph, the meaning of whose axes is unknown; information without perspective is a higher form of ignorance.