Although stroke is a major public health issue little is known about the need for stroke services or the most cost-effective methods of treating stroke patients. This paper discusses how stroke registry data can be used to inform and promote health service changes to meet these needs. Stroke registers were established in three districts of southern England. The initial study findings were that incidence rates reflected mortality rates in these districts, with the higher rates being observed in inner city districts. Hospitalization rates were high with significant use of health service resources but with only 5% being spent on rehabilitation. As a result of these findings a stroke steering group was established in the district with the highest incidence. Standards for care were developed and are currently being audited. A multidisciplinary rehabilitation team was set up in collaboration with the stroke registry. The local purchasers of health care have commissioned an evaluation of a community rehabilitation team to address the problem of low rehabilitation rates in the community. Although it is not possible to control for external influences within a health service this paper illustrates how research findings can influence local health service provision in a tangible way. Epidemiological research can only hope to effect such change if the original studies are undertaken rigorously and the researchers are proactive in the dissemination process, creating opportunities for further research to resolve questions raised by the studies.