Shall We Count the Living or the Dead?

WHEN we wish to estimate the beneficial effect of a therapeutic agent, or the harmful effect of a factor such as exposure to radiation, we have the choice of a number of methods. If a new treatment is associated with lower mortality rates, we can estimate its beneficial effect from the absolute increase in the number of survivors, from the percentage increase in the living or from the percentage decrease in the dead. Often, these different methods provide estimates of quite different magnitude. Similarly, we may get quite different estimates of the harmful effect of radiation from the number of . . .