I find myself in agreement with many of the observations of Manchanda and colleagues. There are many examples in medicine (for example surgery) where the need to advance clinical care, to keep up with new approaches and to develop research, has led to greater specialisation within a discipline. Dr Pelosi's charge of elitism seems to me a professional one; in early psychosis, in line with the predictions of Manchanda and colleagues, I think this focus in the UK is directly responsible for the increase in public recognition of the underinvestment in these services (Reference RethinkRethink, 2002) and for the development of the political will for reform. The longitudinal studies have shown clearly that long-term disability and course trajectories are in place within 3 years, yet all resources are downstream (assertive community treatment, rehabilitation); thus, this new investment has been warmly welcomed by consumer groups (Reference RethinkRethink, 2002).

This service structure now provides an unparalleled opportunity for further research and service innovation. Important research questions now come into focus. What kind of intervention will bring the early cycle of relapse under control, and will young people find it acceptable? What strategies are effective in encouraging help-seeking to reduce duration of untreated psychosis, and what is its impact? I think it is important to emphasise that early intervention services can only provide vehicles for intervention and are not an intervention in themselves; the litmus test of a service is its ability to engage (a major problem in early psychosis) and to fix existing service problems. For example, as Dr Manchanda illustrates, the early intervention focus enables us to think creatively about how to improve continuity of care between child and adolescent mental health services and adult services and to infuse the concepts so familiar to child and adolescent services into the adult arena and vice versa (Reference BirchwoodBirchwood, 2003). I agree with Dr Manchanda that continuity can work forward in time, too; however, there is a risk that early intervention, like existing services, could trap people unnecessarily in long-term services. Preparing for exit and developing community support strategies and identifying cases of ‘prolonged recovery’ are also important.

Dr Manchanda comments about Dr Pelosi's concerns about the ethics of ‘prodromal intervention’. I too share these, but this continues to be a research issue and does not form part of the vision for early intervention services. However, the cases thrown up by the ‘ultra high risk’ or prodromal research involve people suffering from distressing psychotic experience that has not reached the ICD threshold; these people are all seeking help and the majority are already receiving care from secondary services. All clinicians will be familiar with such individuals, who present a therapeutic challenge where equipoise is acknowledged. One benefit of this research, therefore, is its potential to inform a non-pharmacological protocol of treatment, capitalising on the efficacy of cognitive-behavioural therapy in psychosis and emotional disorders.