Response

Extract

The suggestion that secular changes in risk factors were ignored in our evaluation of trends in breast cancer mortality rates is puzzling. In the second paragraph of our paper ( 1 ) we indicated, as we had in an earlier paper ( 2 ), the importance of taking such changes into account when assessing the impact of treatment or mammography on trends in breast cancer mortality rates. Accordingly, we used piecewise regression to analyze trends in age-specific mortality rates to detect recent changes in existing trends ( 1 ). We demonstrated significant declines of a similar magnitude in the slope of the mortality curves after 1989 in every decade of age from 40 through 79 years. Trends for these age groups prior to 1989 were largely consistent with trends in age at first birth ( 1–4 ). However, the simultaneous decline in slopes across multiple age groups after 1989 is not consistent with a change in risk factors but rather provides evidence of improved medical interventions ( 5 ). The calendar period effects from an age-period-cohort model fit to breast cancer mortality rates for U.S. white women through 1994 ( Fig. 1 , A) clearly evidence this improvement in breast cancer mortality after 1989.