A Community Informed Geospatial Assessment of Breast Cancer Screening and Outreach Priorities in Kansas

Project BRA (Breast Cancer Risk Assessment), conducted in partnership with the Breast Cancer Health Equity Task Force, used a community-informed geospatial approach to evaluate inequities in breast cancer (BC) screening and incidence across the University of Kansas Cancer Center (KUCC) catchment area, with the goal of informing risk-stratified, targeted outreach.

County-level data were obtained from the OPTIK data warehouse, integrating Kansas Cancer Registry records, BRFSS indicators, and socioeconomic measures including poverty, median income, food desert prevalence, rurality, and insurance coverage. Outcomes included BC screening prevalence (2018-2022) and age-adjusted BC incidence rates (2014-2018). Multivariable logistic regression estimated adjusted odds of screening, with Geographically Weighted Binomial Logistic Regression (GWBLR) used to assess spatial heterogeneity. Incidence analyses used multivariable linear regression and Geographically Weighted Regression. Seven Kansas City-area counties were compared with the remaining catchment using χ² and Welch t-tests. This analysis was informed by input from a community advisory group (BCHETF) to guide geographic priorities and contextual interpretation.

Screening proportions were highest in northeastern and north-central Kansas and lower in southern and southwestern counties. Rural counties had a slightly lower pooled screening proportion than urban counties (66.79% vs. 67.79%; 95% CI for difference: -1.27 to -0.73; p < 0.0001). In the multivariable model, lower screening odds were associated with rurality, food desert prevalence, smoking prevalence, and higher proportions of Black and Hispanic or Latino residents, while higher screening odds were associated with below-poverty percentage and uninsured percentage. The GWBLR showed spatial heterogeneity in screening-related associations. Among the seven focal counties, screening rates ranged from 62.7% to 72.5%, and all were below the 2021 national benchmark of 75.9%. Age-adjusted breast cancer incidence was lower in rural than urban counties (119.93 vs. 135.19 per 100,000 women; p = 0.0073), and five of the seven focal counties exceeded the national incidence benchmark of 130.8 per 100,000 women.

BC screening disparities at the county level were more strongly associated with structural and contextual factors than with racial composition after adjustment; however, these findings should be interpreted as area-level associations and do not imply individual-level causation. Community-engaged geospatial analytics can identify high-need communities and guide equitable prevention and outreach strategies.