Does Quitting Smoking Save Lives?

Abstract

Claims like those of the U.S. Surgeon General (1982) that giving up smoking lowers lung cancer and other types of mortality are, of course, vital for any consideration of possible relations between smoking, cancer, and coronary heart disease (CHD). There are two ways of studying such effects. The first takes a sample of subjects, usually relatively homogeneous (e.g., British physicians), follows them up, and examines the death rates from various causes of those who give up smoking and those who continue to smoke (Doll & Peto, 1976). The alternative method is to form two groups, matched on as many relevant variables as possible, and to consider one of these as a control, receiving no instruction, while the other is a therapy group receiving instruction to give up smoking and possibly also receiving advice relating to other risk factors, such as poor diet, high blood pressure, and so on. Unfortunately, practically all the positive results of giving up smoking have been reported from studies using the first of these methods, that is, self-selection, and it is widely recognized that no relevant conclusions regarding causality can be drawn from studies of self-selected populations only. It is now quite clear that this is a crucial factor, making it impossible to compare ex-smokers and continuing smokers with the aim of establishing the causal link between smoking and disease.