The modeling of breast cancer deaths in Vermont

Methodologic issues in modeling breast cancer deaths were investigated through modeling the number of breast cancer deaths in Vermont for the period 1975 to 1988. Age-specific incidence rates, case fatality rates, and secular trends of these rates were necessary to represent the observed trend in breast cancer deaths over this time period. Additional information, such as mortality from other causes, stage distribution, or screening history, was not necessary. Incidence and survival information were obtained from a statewide population-based breast cancer registry in Vermont and the Surveillance, Epidemiology and End Results (SEER) program. SEER incidence rates overestimated the actual number of deaths for long-term survival rates. Case fatality rates and secular trends in incidence and survival reported by SEER were appropriate. These results can be applied to the planning of community intervention studies by providing the essential expected numbers of deaths in control communities.