Purpose: Our purpose was to evaluate county level sociodemographic predictors of county level mammography screening percentages in the United States.
Materials and Methods: Mammography screening percentages for each county were obtained from the most recent data from the National Cancer Institute’s Small Area Estimates (SAE) in 2010. SAE combines data from nationally representative surveys (National Health Interview Survey, Behavioral Risk Factor Surveillance System) with auxiliary variables obtained from relevant sources (Census, American Community Survey). Mammography screening percentage was defined as having mammography within the last two years before the time of the interview for women greater than 40 years old. County level sociodemographic predictors were derived from the County Health Rankings, derived from national data sources, including the National Center for Health Statistics, Behavioral Risk Factor Surveillance System, the American Community Survey, and several others. Univariable and bivariable linear regression analyses were conducted to evaluate the association between county level mammography screening percentages and sociodemographic variables.
Results: A total of 3,114 U.S. counties were included. Mammography percentages ranged from 34% in Starr County, TX to 95% in Manassas Park City, VA (Figure 1). Among counties for which the majority of residents were Black, mammography percentages ranged from 53 to 79%. Higher percentages of Black residents were associated with higher mammography screening percentages (Correlation Coefficient 0.163, p < 0.001) (Figure 2). Higher percentages of Hispanic residents (Correlation Coefficient -0.284, p < 0.001), higher percentages of uninsured residents (Correlation Coefficient -0.507, p < 0.001), and higher percentages of rural residents (Correlation Coefficient -0.410, p < 0.001) were associated with lower county level mammography screening percentages (Figures 2 and 3). Higher county level median incomes were associated with higher county level screening percentages (Correlation Coefficient 0.586, p < 0.001) (Figure 3).
Conclusion: Mammography screening percentages vary widely at the county level in the United States. Lower income levels and higher percentages of uninsured residents were associated with lower county level mammography screening percentages.
Clinical Relevance Statement: Improving breast cancer screening percentages will require targeted outreach efforts, particularly towards counties with higher proportions of low-income uninsured residents.
