Professor Singh is the expert on early intervention services and provides a characteristically scholarly and elegant reappraisal. ^{Reference Singh1} Although I am surprised that the Lambeth Early Onset (LEO) ^{Reference Craig, Garety, Power, Rahaman, Colbert and Fornells-Ambrojo2} and OPUS ^{Reference Petersen, Jeppesen, Thorup, Abel, Øhlenschlaeger and Christensen3} trials are interpreted as quite so definitive (when most of us read them as very promising but far from conclusive), I am reluctant to disagree with an admired colleague in his area of expertise. However, I must take issue with one conclusion in his ‘future directions’. Singh argues that generic community mental health teams (CMHTs) have no evidence for them and that ‘The logical next step in the move from institutions to community is from generic community teams to specialist teams’. In this I believe he is mistaken.

Community mental health teams suffer from having evolved before the era of intensive mental health services research. Nobody ‘owns’ them, so few have actively researched them; they have most often been the comparators in randomised controlled trials of other innovative specialist teams. Despite this, research-based conclusions can be drawn about their comparative effectiveness. The body of assertive outreach research is overwhelmingly greater than for any other specialised team. What a series of over 60 assertive outreach team trials shows is that reductions in in-patient care are more highly dependent on the nature of the comparator services than the experimental services. ^{Reference Burns, Catty, Dash, Roberts, Lockwood and Marshall4} Where these comparator services are poor and fragmented there is a substantial reduction; where they are not, then there is little or no reduction. Often this has been where the comparator is a generic community team. ^{Reference Killaspy, Bebbington, Blizard, Johnson, Nolan and Pilling5}

We have rather myopically interpreted these findings as a failure to demonstrate superiority of the specialist team over CMHTs. However, ‘As health services enter a period of economic austerity’, ^{Reference Singh1} we need to recognise that the findings tell us much more than that. What they demonstrate is that generic CMHTs have routinely matched the specialist teams in major outcomes yet for a significantly lower cost. ^{Reference Burns6} They are, in short, more cost-effective and therefore currently our best buy.

Experimentation and innovation in specialised teams must continue if we are to progress. However, if we conduct research we must pay attention to its findings, no matter how unwelcome. The current evidence supports the superiority of CMHTs, no matter how much that they may grate.