This paper investigates the effect of health insurance on health and the use of health services by exploiting a change in insurance status that occurs for most Americans at age 65; that is, eligibility for the U.S. Medicare program. A regression discontinuity design is employed to identify discontinuities at age 65 in the relationship between age and access to care and health status, especially for groups more likely to be uninsured prior to age 65, such as those with less than a high school education or blacks and Hispanics. The paper focuses on the use of health services and health outcome related to breast cancer, a common cause of death among women, and one for which good access to early detection services is thought to significantly improve survival. Results show that the use of health services including mammography increases discontinuously at age 65, especially for women without a high school degree and for black and Hispanic women. A modest decrease in the probability of late-stage breast cancer diagnosis at age 65 is also found for white and Hispanic women.