Tobacco Documents Online

/~ZPi ~1Z108t~L~ SCHOOL OF PUBLIC HEALTH DEPARTMENT OF BIOSTAT{ST1CS 109 SOUTH OBSERVATORY ANN ARBOR, MICHIGAN 48109 (313) 764-5450 The effects of smoking on health has been a subject of debate since long before the original Surgeon General's report in 1964. The debate has - intensified since that report, with increased attention being given to the methods of study, statistical procedures, and to scientific reasoning in general. "Ihe interpretation of statistical association in terms of cause- effect relationships has been at the center of this discussion. Statistical a$sociations, like the one between smoking and health, can reflect causal relationships, but it is not true that such an association proves causality. A statistical association, no matter how strong, can still result from confounding with other factors. This is particularly true in the case of association with chronic diseases or diseases with a long latent period. However, if association still persists after all non-causal alternative theories have been exhausted, one has little alternative than to act as if the association is in fact causal. In recent years a great deal of research on the relationship of smok- ing and health has been carried out. This research has investigated the role of a large number of possible risk factors in explaining the observed association between smoking and disease. Studies include animal trials, retrospective studies, prospective studies, as well as investigations into the basic physiology of smoking. These studies have consistently found a strong association between smoking and disease, particularly between smoking and lung cancer and smoking and cardiovascular disease. The wide range of populations studied in the U.S. and other countries, as well as the wide range of study types and analytical techniques lend credence to the possiblity that smoking is a contributing cause of lung cancer and of cardiovascular disease. The observation of a dose-response relationship enhances this possibility. (incidence of lung cancer and cardiovascular disease is-lowest in non-smokers, and shows a nonotone increase with the amount smoked). Further, groups of persons who discontinued smoking have shown reduced risk compared to those who continued to smoke. The strength and consistency of the observed relation- ship, together with the fact that an alternative non-causal explanation requires some factor or factors to be very strongly associated with both smoking behavior and with the diseases, makes it unlikely that the associa- tion between smoking and disease is entirely non-causal. I I I I I I I I t I ~ L -178-

( r I F I I I I F In sumnary the overwhelming evidence of association between smoking and disease, though not proving causality, does lead one to act as if the relation- ship is causal. However, it should be recognized that for public health, proof of causation, specification of the details of causation, or elucidation of the mechanism of causation is not the main concern. The weight of evidence strongly supports the conclusion that the risk of lung cancer and heart disease is lower for those who never smoked than for those who smoke, and the risk is lower for persons who-stop smoking than for persons who continue to smoke. Thus, programs to discourage people from starting to smoke, to encourage smokers to discontinue smoking, or to protect non-smokers from unwanted exposure to cigarette smoke is expected to result in fewer premature deaths from lung cancer and heart disease and better general health of the public. This is all that is needed to warrent the initiation of effective public health programs to reduce smoking, along with evaluation plans to ensure that these programs are, in fact, effective. Richard G. Cornell, Ph.D. Professor and Chairman J'rus D.- Flora, Jr., Ph. . Associate Professor William J. B er, Ph.D. Assistant Professor I- I- I- I- I L L Department of Biostatistics School of Public Health University of Michigan Ann Arbor, Michigan 48109 Testimony presented at the Public Hearing, Citizens' Panel on Smoking and Health, October 8, 1980. -179-