Using small area analysis to describe county-level disparities in mammography Free

B7

Prior studies have used national health survey data to analyze the associations of age, race, and ethnicity with cancer screening on the national and state levels; yet, research is needed on subpopulations defined by geographic units smaller than the state, such as the county. The purpose of this study was to evaluate the extent to which geographic disparities exist in mammography utilization across counties in the continental United States. County-level prevalence of recent mammography (past two years) for women 40-79 years old was estimated using synthetic-regression, a small area estimation method. The 2000 Behavioral Risk Factor Surveillance System (BRFSS), 2000 Decennial Census, and Area Resource File were merged by BRFSS respondent’s county of residence. Separate analyses were conducted to produce county-level prevalence estimates for each race and age group. Geographic variation existed in mammography utilization for women 40-79 years old (range, 69.6-84.0%). The location and magnitude of geographic disparities differed by race and age category. Non-white women showed the lowest prevalence of mammography and widest range in county-level estimates (range for Black, 65.0-92.0%; range for other non-White race, 36.0-83.0%). Women aged 40-49 had less variability in prevalence of mammography and overall lower prevalence than other age groups (range, 56.7-79.8%). Women aged 65-79 showed the greatest variation in county-level estimates of mammography (range, 53.0-88.0%). Small area estimation using BRFSS data is advantageous for surveillance of mammography utilization on the county-level. This method allows documentation of geographic disparities and improves our understanding of the spatial distribution of mammography prevalence. By considering this county-level geographic variation in future interventions, public health practitioners will be able to more efficiently and appropriately target women in the neediest of counties.