Background: Research has not established (1) if breast cancer screening varies by county-level proportion of uninsured or (2) whether county-level-proportion of uninsured correlates with county-level early-stage and late-stage breast cancer incidence. Methods: A multilevel study was conducted to determine if individual-level self-reported breast cancer screening data from the 2000 Behavioral Risk Factor Surveillance System (BRFSS) was associated with county-level-proportion-uninsured data from the 1999-2001 BRFSS. An ecologic study was conducted to determine if county-level proportion of uninsured correlated with incidence of early-stage and late-stage breast cancer using the 1999-2001 BRFSS data from the overlapping counties in the Surveillance, Epidemiology, and End Results (SEER) program. Data were analyzed in 2005. Results: Women were less likely to be screened (prevalence odds ratio: 0.95; 95% confidence interval=0.93-0.97) with every 5% increasing county-level proportion of uninsured. African-American and Hispanic women who resided in counties with a proportion of uninsured of 9%-19% had higher screening utilization than white non-Hispanic women. The county-level-proportion of uninsured had little effect on screening use among women with household incomes less than $25,000 or greater than $75,000. Screening prevalence decreased with increasing county-level proportion of uninsured among women with intermediate income. The rate of T1 (<2 cm diameter) tumors decreased with increasing county-level proportion of uninsured while controlling for poverty rate; Spearman correlation -0.294. Conclusions: High county-level proportions of uninsured may lead to lower early-stage breast-cancer incidence through lower screening use among women living in these less-well-insured counties.