The present study reflects on the role of the middle manager in the implementation of what has become known as evidence-based health care. This movement advocates that clinical practice is continually informed by the results of robust research and evidence. In our work exploring the complexity of ensuring that practice is informed by evidence we have found that general managers have relatively little influence when compared with clinicians especially doctors. We argue that local professional groups work together in communities of practice, which are frequently uniprofessional. These boundaries affect the motivations for seeking improvement and upgrading and the way evidence and knowledge is perceived and interpreted. We argue that if the quality of health care is to be improved, we need to understand the complex historically and contextually informed interactions between different professional groups and to design diffusion strategies that acknowledge this complexity.