Stroke is an enormous public health problem, the magnitude of which can be reduced mainly by effective stroke prevention and less so by effective treatment of acute stroke. The greatest effect is likely to be achieved by a mass approach to prevention, which consists of modification of lifestyle behaviors (eg, less smoking and less intake of salt, alcohol, and fat) among the general population through public education and, more importantly, government legislation. The appropriate identification and treatment of high-risk individuals by neurologists is likely to have a smaller but complimentary impact on the population burden of stroke and a substantial impact on the burden of stroke among individuals. The most cost-effective interventions for patients with transient ischemic attack and ischemic stroke are organized multidisciplinary acute care and rehabilitation in a stroke unit and early secondary prevention with aspirin, blood pressure control, smoking cessation, and, in the appropriate patient, oral anticoagulant therapy and carotid endarterectomy. The cost-effectiveness of carotid endarterectomy for asymptomatic carotid stenosis is highly questionable until data from ongoing trials (eg, Asymptomatic Carotid Surgery Trial) become available. Screening for asymptomatic carotid stenosis is more likely to be harmful than helpful, except perhaps among populations with a very high prevalence (pretest probability) of severe carotid stenosis. It is essential that the impact of these strategies on the incidence, outcome, and cost of stroke is measured and monitored. Currently, this is done simply, but unreliably, by examining changes in statistics that are already being measured, such as mortality (eg, among those younger than 70 years old, for greater accuracy). A growing priority in many countries is the development and implementation of valid, reliable, practical, and inexpensive methods of routinely collecting and evaluating data on stroke incidence, outcome, and cost.