There is enormous geographic variation (up to eight, nine, ten to one) in the frequency of certain medical procedures across US areas that cannot be justified by variations in epidemiology, and higher procedure rates do not necessarily produce better health outcomes (e.g., coronary bypass and angioplasty).