The rate of mortality has declined so much in recent years and so many diseases have been nearly eliminated as causes of death that medical attention is now concentrated on a relatively small number. Of these diseases, cancer is of particular interest to industry.
Laboratory tests help us to predict whether or not a substance is carcinogenic but the development of knowledge is not yet advanced enough to allow us to rely on them and this ignorance prevents their use in predicting human risk quantitatively. Epidemiological observations are needed for three main reasons. The first is to detect unsuspected causes of cancer. This is easy if the cancer is normally rare and the hazard in industry is increased many times; it is difficult when the cancer is common and the relative increase is small. In the latter circumstances care must be taken to exclude bias, confounding (in particular with other local and social factors), and chance. Secondly, they are needed to demonstrate the size of a risk associated with a particular level of exposure and therefore help to determine the level that is socially acceptable. The risk may be shown to be, at the most, so small that its absence can effectively be regarded as having been demonstrated — although proof of a negative is not theoretically possible. For this purpose, all available evidence needs to be accumulated and examined and it may not be wise to lay down criteria for negative evidence. Thirdly, and most importantly, epidemiological observations help to maintain a sense of perspective. They suggest, for example, that the risk for all cancers combined, other than lung cancer, is relatively stable in the United States and that the main causes of current cancers must have been present in society for many years. The evidence suggests that some 30% of cancer deaths must be attributed to cigarette smoking and some 4% or 5% to occupational hazards. Dietary factors may be very important and there is the prospect that anti-promoting factors in the diet (possibly vitamin A or its precursor) may be used to reduce risks that have been incurred by previous exposure to initiators.
Epidemiology may not be the method of choice for the discovery of preventive measures, as it requires some people to have been affected before it can be employed; but at present its use is essential if efforts are to be concentrated in areas where they will do most good. One such area is the rigorous study of industrial cohorts to determine the degree to which they may be exposed.
©1981 The American College of Occupational and Environmental Medicine