OBJECTIVE: To test whether the model of 'diffusion of innovations' enriches understanding of the implementation of evidence-based thrombolysis services for stroke patients. METHODS: Four case studies of the implementation of evidence on thrombolysis in stroke services in England and Sweden. Semistructured interviews with 95 staff including doctors, nurses and managers working in stroke units, emergency medicine, radiology, the ambulance service, community rehabilitation services and commissioners. RESULTS: The implementation of thrombolysis in acute stroke management benefited from a critical mass of the factors featured in the model including: the support of national and local opinion leaders; a strong evidence base and financial incentives. However, while the model provided a starting point as an organizational framework for mapping the critical factors influencing implementation, to understand properly the process of implementation and the importance of the different factors identified, more detailed analyses of context and, in particular, of the human and social dimensions of change was needed. CONCLUSIONS: While recognising the usefulness of the model of diffusion of innovations in mapping the processes by which diffusion occurs, the use of methods that lend themselves to in-depth analysis, such as ethnography and the application of relevant bodies of social theory, are needed.