This paper examines how the education gradient in the take-up of screening mammography is affected by institutional features of the health care system. It does this by comparing the gradient in the Netherlands, which has an organized screening program, with that in the United States, which does not have such a program. For the Netherlands, we have found no educational disparities in screening use among women invited by the screening program. For the United States, educational differences are explained by income, insurance and recommendations. Being recommended to undergo screening has the largest impact on use. Insurance also emerges as a key variable which is closely related to receipt of recommendations and screening itself.