Community Factors and County-Level Cancer Screening, Prevalence, and Mortality.

Importance: Cancer screening, prevalence, and mortality vary across US counties; the broad set of community factors associated with these differences are not well understood. Objective: To assess the relative importance of diverse community measures for explaining county-level variance in cancer screening, prevalence, and mortality rates for breast, colorectal, lung, and prostate cancers. Design, Setting, and Participants: This geospatial cross-sectional analysis across all US counties used random forest algorithms to estimate the relative importance of 24 community measures-including health behaviors and lifestyle, socioeconomic, and environment factors-as explanatory factors for cancer screening, prevalence, and mortality. For each cancer, county-level cancer screening and prevalence rates were based on a nationally representative 5% sample of Medicare fee-for-service beneficiaries (2020). The 5-year (2016-2020) mean mortality rates were from the National Cancer Institute's State Cancer Profiles. Neiman Cancer disparity maps geospatially display significant community-factor to cancer-outcome associations via county-level choropleth maps. Data were analyzed from September 2024 through February 2025. Exposure: Residence in the 50 states, Puerto Rico, and the District of Columbia. Main Outcomes and Measures: Outcomes were county-level rates of cancer screening, prevalence, and mortality for breast, colorectal, lung, and prostate cancers. Separate random forest algorithms estimated the relative importance of 24 community measures for each outcome-cancer type combination, and maps display county-level associations. Results: Among a nationally representative 5% of 2020 Medicare Fee-For-Service beneficiaries, 87% were aged 65 years or older. The top ranking explanatory community factors (100% relative importance) for mortality were smoking rate for both lung and colorectal cancers, and the non-Hispanic Black population share for both breast and prostate cancers. For prevalence and screening rates, the top factors were unique for each cancer type. Uninsured rate, unemployment, limited access to health foods, and poor physical health were the top ranked factors for breast, colorectal, lung, and prostate cancer prevalence, respectively. Hispanic population, poverty, air pollution, and Air Toxics Cancer Risk were ranked highest for screening rates, respectively, for each cancer. Two of only 3 environmental factors, Environmental Justice Index and Air Toxics Cancer Risk, had the most top-5 associations for mortality and screening, respectively. Uninsured rates for prevalence and poverty rates for screening were also important across cancer types. Conclusions and Relevance: In this cross-sectional study of the relative explanatory importance of community measures on cancer screening, prevalence, and mortality rate we found that the ranking of factors varied for each specific cancer and outcome, which can inform targeted population health efforts. Factors that ranked high across cancer types for a given outcome, such as uninsured rates for cancer prevalence, represent opportune targets for future study and broader policy change.