Use of mammography screening in the general population has been credited with improvements in the diagnosis of breast cancer. Yet, the proportion of women with invasive breast cancers who are diagnosed at later, distant stages has not been declining. Two statewide data sources were linked to compare mammography screening compliance of female Medicare beneficiaries and stage of breast cancer at diagnosis at a population level. Breast cancer diagnoses were ascertained from the Kansas Cancer Registry. Mammogram use was identified using Medicare claims files and were classified as annual, biennial, less than biennial, or never use. The study included women aged ≥69 years at the time of their primary invasive breast cancer diagnosis between 2011 and 2012 who were continuously enrolled in Medicare parts A and B for at least 4 years prior to diagnosis. This approach enabled demographic comparisons to be made at the population level, providing novel and important insights about how sociocharacteristics of residency affect mammography compliance in an invasive breast cancer population. Of the 915 female Kansans who met the eligibility criteria, 73% (672) had received at least 1 mammography screening during the 4 years prior to their primary invasive breast cancer. The percentages of these women who received annual, biennial, less than biennial, and no screening were 25.8%, 27.5%, 20.1%, and 26.6%, respectively. Women aged ≥74 years, those living in counties with a median income below $42,000, and those living in counties with a higher percentage of individuals without a college degree at the time of diagnosis had higher odds of not having had an annual screening. Linkage of data from the statewide cancer registry with Medicare claims data on mammogram screening may help to identify subgroups of older women to be targeted for promoting breast cancer screening.
Keywords: cancer; mammography; population surveillance; record linkage; screening compliance.