Trends in breast cancer mortality among Swedish women 1953-92: analyses by age, period and birth cohort.

Trends in breast cancer mortality among Swedish women were explored on the basis of all 51,048 deaths in women 30-89 years of age in Sweden during the period 1953-92. The age-standardised mortality rates were virtually unchanged during the observation period (with a mean of 32 deaths per 100,000 females and year), as were age-specific rates. In age-period-cohort analyses, age alone explained almost all of the variation in the rates. The effects of period and cohort were statistically significant, but very modest. Cohort effects seemed to explain more than period effects, and a weak downward trend starting with women born in 1883-92 was noted. A change in 1981 in the policy to classify the causes of death from the death certificates seemed to entail an artificial lowering of the mortality rates in women older than 75 years. It is concluded that breast cancer mortality in Sweden during the last 40 years has been remarkably stable, in spite of a substantial and constant increase in the incidence. This divergence between mortality and incidence reflects improved survival, which could in part be explained by earlier detection and more efficient treatment, or by an increasing occurrence of less aggressive tumours.