It is not controversial to say that we live in an era of unprecedented preoccupation with the notion of ‘risk’ in our public mental health services, not to mention other areas of public life. Notions of ‘risk’ and ‘safety’ have long been at the heart of current government policy on mental health as epitomized by the guidance document entitled Modernising Mental Health Services: Safe, Sound & Supportive (Department of Health [DoH] 1998). This is partly a consequence of the continuing switch of emphasis by both major political parties over the last 20 years or more away from hospital or in-patient care and towards community-based services. This is not the place, however, to focus on the political context. However, in clinical terms what really is the difference between a ‘safe’ or ‘secure’ mental health service and an ‘unsafe’ or ‘insecure’ one? The more I have been invited to think about this question, the more complex and unfamiliar the territory has revealed itself to be.