Elevated Breast Cancer Mortality in the Northeastern U.S. is Secondary to Poorer Survival Rather Than Increased Incidence

摘要

Mortality rates for breast cancer are approximately 25% higher in the northeast U. S. than the south or west. Virtually all previous investigations into this phenomenon have assumed that the elevated mortality rates reflect elevated incidence, and this putative elevation in incidence has been attributed to a number of factors. We hypothesized that the geographic variation in breast cancer mortality was secondary to differences in survival with breast cancer as well as differences in incidence. We proposed to test this by identifying women aged >70 years newly diagnosed with breast cancer in 1991 in different regions of the country using Medicare charge data and assessing their survival. Initially we developed algorithms to determine incident breast cancer and date of diagnosis from the Medicare data and validating them using the SEER-Medicare linked data base. We found that an algorithm incorporating data from hospital inpatient, hospital outpatient, and physician services claims produced levels of sensitivity and specificity> 90%. However, the positive prediction value was low (67 to 70 percent), precluding use of Medicare data alone to examine survival from breast cancer. Our current work is using SEER-Medicare linked data to examine geographic variation in incidence, survival and mortality of women with breast cancer.