Abstract
In recent years a decline in the rate of increase in male lung cancer mortality has led some observers to predict that this disease epidemic will soon peak and then ebb. This prediction is based on the extrapolation of curves the lengths of which are limited by many inherent deficiencies in the data. Perhaps the most important of these is the short period of observation available for the youngest cohorts who will provide the cases of the future. Manipulations of data for Philadelphia white men from 1955 to 1970 illustrate some of the analytic problems. Accurate predictions of lung cancer mortality cannot be made.