Dr MacMillan ( , this issue) is quite right to point out the importance of appropriate access to in-patient beds as a critical component of mental health services. That mental health systems should provide a balance of in-patient beds and community services tailored to the mental health needs and resources of the local community being served is something all mental health practitioners across the world can probably agree on. My commentary did not suggest that increased specialisation means we should do away with in-patient services, it simply stated the fact that investment in specialist community mental health teams (particularly crisis teams) through the National Service Framework for Mental Health was associated with a reduced need for in-patient admissions. Where I believe Dr Lodge and I also agree is on the need for continued investment in mental health rehabilitation services to prevent the inappropriate use of out-of-area placements for the small number of people with particularly complex and long-term psychoses. ^{Reference Killaspy and Meier1,2}

Professor Burns’ response (see letter above) states: ‘It is ungenerous and unjustified for Helen Killaspy to accuse George Lodge of nostalgia and wearing rose-tinted spectacles just because she disagrees with him. Newer is not automatically better.’ This accusation is not only unjust and ungenerous to those who have been working without feeling conflicted in both specialist and generalist services for many years, but it is without basis in fact. My commentary made clear, evidence-based justification for my view. I included reference to the lack of evidence for the effectiveness of assertive community treatment in the UK context that probably influenced subsequent disinvestment in this model. However, our research group, while contributing to such findings, simultaneously participated in a multicentred international study which suggested that assertive community treatment in the UK may have not performed as effectively as in Australia owing to lack of implementation of critical components that Professor Burns’ own team identified through meta-analyses. ^{Reference Harvey, Killaspy, Martino, White, Priebe and Wright3,Reference Burns, Catty, Dash, Roberts, Lockwood and Marshall4} His further accusation that I was ‘disingenuous’ is a little ironic given his lack of reference to the robust international evidence on which investment in the new specialist teams was made, not to mention the expanding evidence base for early intervention services.

Dr Dodwell's response ( , this issue) accuses me of dismissing evidence on therapeutic alliance, yet I did not mention it. It is a truism to say that the therapeutic alliance is important. Who would argue against the importance of being treated with humanity and respect in the therapeutic encounter? However, therapeutic alliance is not the same as continuity of care, which was, after all, the focus of Dr Lodge's piece.

The fundamental issue that seems to have prompted such vociferous response is whether psychiatrists can go on being Jacks and Jills of all (psychiatric) trades. My view is that our increased specialisation is a sign of the maturation of our profession and allows us to deliver better treatment, tailored to our patients’ needs. This does not equate to support for some kind of anarchic service redesign with the aim of promoting turf wars and passing patients from pillar to post. We need to get on with the business of incorporating the evidence we have available from our research to design systems of care that are appropriate, effective and cost-efficient, and accept that the process is iterative and subject to socioeconomic and political vagaries. Perhaps we are more likely to succeed in this if we start with a focus on the areas where we have consensus.