A theoretical approach to the evaluation of the mortality effects of a 'low tar' cigarette

The two parameter Gompertz hazard function is shown to fit Hammond's 1960 American Cancer Society life tables which depict survivorship of men aged 25-100 yr for five levels of smoking frequency. Using data on tar levels of cigarettes on the market during the period 1960-1965, various models are proposed to estimate the covariation of the Gompertz parameters by smoking frequency and tar content. For each model, estimated life expectancy and age specific death rates are presented for various smoking frequency-tar groups and the mortality consequences of a 'low tar' and 'low frequency' public health policy are discussed. Multiple regression analysis is used to summarize the mortality findings and to provide a quantitative assessment of the relative importance of tar content and smoking frequency in determining life expectancy and death rate. Evidence from the models favors a 'low tar' policy over a 'low frequency' policy but considerable benefits can be obtained from a combined policy.