AbstractThere is a great deal of indirect, nonexperimental evidence that a pattern of earlier‐stage disease at diagnosis has a better outcome. Increased early detection activities can change, these stage patterns while various biases and the question of generalizability need to be kept in mind in their interpretation. The indirect evidences of possible benefit from early detection activities includes an increase in the number of cases detected, a pattern of more early‐ and less advanced‐stage cases, an increase in the overall site‐specific survival rate, and a decrease in the case fatality rate. Unless these intermediate markers are favorable, it is unlikely that early detection will reduce mortality. In addition, one should also differentiate a reduced incidence or a change in treatment as a cause for reduced mortality. © 1992 Wiley