The Methodology of Epidemiological Studies of Smoking
What are the major arguments in favor of a close connection between smoking and disease, possibly interpretable in causal terms? The first argument, already considered to some extent, deals with the mortality ratios for different diseases, that is, figures purporting to demonstrate that smokers of a given age and sex die more frequently of a given disease than nonsmokers. The difference can be expressed as the “mortality ratio” (age-standardized mortality rate, or SMR), indicating the proportion of smokers to nonsmokers who are certified as having died of a particular disease. These ratios are usually in excess of 1.0 for a great variety of diseases, with the highest mortality ratio being that of cancer of the lung, followed by bronchitis and emphysema and cancer of the larynx and the oral cavity. Some mortality ratios, however, are below 1: for example, for cancer of the rectum,.90 (Kahn, 1966); colorectal cancer in women (.78 and.66 for women who smoked heavily, Hammond, 1966); and Parkinson’s disease (.26, Kahn, 1966;.81 for an older group, Hammond, 1966). Eysenck (1987) gives a survey of the literature. He stresses that these figures should not be taken to indicate a beneficial effect of smoking, just as the positive mortality ratios for lung cancer and so on do not necessarily indicate the nefarious effect of smoking; the problem of inferring causation from correlation is, of course, too complex for such easy inferences.
KeywordsLung Cancer Death Certificate Surgeon General Environmental Tobacco Smoke Passive Smoking
Unable to display preview. Download preview PDF.