Evaluation of breast cancer incidence: is the increase due entirely to mammographic screening?

OBJECTIVES: To examine the trends in the incidence rates of breast cancer in a population with mammographic screening and in the unscreened women within that population.
METHODS: Data consisted of incident cases of breast cancer notified to the Victorian Cancer Registry in Victoria, Australia, between 1988 and 1996 and cases detected in the population-based BreastScreen Program. These data were grouped by age (25-39, 40-49, 50-59, 60-69 and > or = 70 years of age) and size of tumor (< or =10 mm, > 10-< or =15 mm, and > 15 mm). Poisson regression modeling was used to examine trends by age, tumor size, calendar year and availability of screening.
RESULTS: The incidence rate of breast cancer in the total population increased between 1988 and 1996. The greatest increase was seen after 1993 when population-based screening became available. In unscreened women, modeling demonstrated a statistically significant (p < 0.01) 1.5% annual increase in the incidence rate. The annual increase in this rate differed by size of tumor and was approximately 8% (p < 0.01) for small tumors (< or = 10 mm) but not significant for tumors > 10 mm. The greatest increase was in small tumors for women > or = 50 years of age.
CONCLUSION: The incidence of breast cancer has increased since population-based mammographic screening was introduced in 1994. The rate in unscreened women also showed a significant increase. This was greatest in small tumors for women > or = 50 years of age. Whether this will translate into an increase in mortality is uncertain and long-term monitoring is required to determine if cohort and period effects impact on the underlying incidence of breast cancer in Victoria.