Demographics (1950-1987) of breast cancer in birth cohorts of older women.

The effects of screening on breast cancer mortality, incidence, and prevalence were investigated using a general forecasting and simulation model. First, a biologically motivated model of disease incidence and mortality was fit to the breast cancer mortality experience of 15 U.S. White female birth cohorts followed for a 38-year period. The model assumed that breast cancer was the result of two different diseases. The first, or "premenopausal" disease, tends to have strong associations with the family history of disease and to be histologically more aggressive. The second, "postmenopausal" disease, occurs at more advanced ages, is apparently less strongly linked to family history, and is less aggressive with different histological characteristics (e.g., positive estrogen receptor status). Those results were used to forecast the effects of screening on the stage at diagnosis to simulate a screening program which reduced late-stage diagnoses by 50%. This produced large reductions in breast cancer mortality--an impact larger for disease associated with late age of onset.

Duke Scholars

Author Patrick J. Eric Stallard Social Science Research Institute

Author Kenneth G. Manton Demographic Studies