Twenty years ago it was not unusual for psychiatrists in the UK to spend at least some of their time working in a primary care setting (Reference MitchellMitchell, 1985; Reference Pullen and YellowleesPullen & Yellowlees, 1988). A considerable literature developed on ways of working within, and across the interface between, primary (generalist) and specialist care (Reference Gask, Sibbald and CreedGask et al, 1997), with clarification of the evidence base for particular ways of working (Reference Bower and GilbodyBower & Gilbody, 2005). The intervening years, however, have seen a widening rather than a narrowing of the gap between primary care and mental health services. Recent data are difficult to track down but a common perception would be, at least in urban areas, that psychiatrists now rarely visit general practitioner (GP) clinics or meet their GP colleagues. The focus of mental health services on severe and enduring mental health problems over the past decade has seen a shift in GPs' perceptions of the accessibility of psychiatrists for consultation or advice, particularly about the assessment and management of people who do not have a psychotic illness, but who may nevertheless present with a complex mixture of mood disorder, social difficulties, risky behaviour and often comorbid physical health problems (Reference Chew-Graham, Slade and MontanaChew-Graham et al, 2007; Reference CohenCohen, 2008, this issue).

It is not always easy for generalists and specialists to understand each other's view of the world, and a variety of barriers stand in the way of better integration between primary and specialist mental healthcare (Reference Lester, Glasby and TyleeLester et al, 2004; Reference GaskGask, 2005a ). General practitioners currently work in a system in which they are the first point of access for all comers with every type of health problem; a high-volume, low-intensity provider system with centrally imposed targets for access times set by central government. Psychiatrists who work in crisis teams may also feel something akin to this constant flow of human need through their services, but many of our profession now work in highly specialised teams to which access is increasingly limited by protocols and referral criteria.

With the development of GP commissioning, GPs are being encouraged to play a greater role in determining the overall shape and functioning of mental health provision in the community. If we psychiatrists rarely meet our GP colleagues we never get to explore commonalities and differences in the views and ideas that we each hold about mental health policy, service provision, the world and each other. More important, however, is that patients and service users surely cannot be receiving optimal care in a system that does not encourage better inter-professional communication (Reference Miller, Charles-Jones and BarryMiller et al, 2005).

We think that it is time to question and re-examine mental health policy in the light of the evidence that has accrued over the past two decades about the interface between primary and secondary care. What are the benefits to be gained for the patient, the practitioner, the organisations and the wider health economy through primary care and psychiatry working more closely together? What are the optimal ways of achieving this? Where are the examples of good practice? What research still needs to be conducted to strengthen the evidence and inform the direction of policy?

To follow up these lines of inquiry, we need to work on three different fronts: policy and strategy, training and professional behaviour.