Introduction: To analyze the trends in cancer mortality between 1968 and 2000, one needs to consider both the increase in the size of the population and the ageing of this population, since both phenomena lead to a mechanical increase in the number of deaths.
Methods: We describe the analysis of the variation in number of deaths to distinguish the demographic effects from the variation in risk. This variation in risk must be studied, controlling for both the size and the age of the population.
Results: The number of cancer deaths in France increased by 15.5% between 1980 and 2000. Based on the demographic changes, one would have expected an increase of 29.1%: 10.3% because the population has grown and 18.8% because it has aged. The observed increase is less than expected because the demographic effects have been canceled out in large measure by a decrease in the risk of cancer, which fell by 13.6% (29.1% minus 15.5%) between 1980 and 2000.
Discussion: It is a mistake to describe the increase in the number of cancer deaths as worrisome when it is due only to demographic changes. The risk of dying from cancer is not rising in France: it is clearly falling.