Dr Holloway's (Reference Holloway2001) stimulating, if ever so slightly mischievous, commentary on our paper (Reference Simmonds, Coid and JosephSimmonds et al, 2001) adds substance to the debate on this subject but leaves the reader with the unfair impression that community mental health teams are now out of date and have been replaced by ‘more exotic fruit’. Indeed, our labours have borne much more fruit than we expected, as Dr Holloway identifies our study as a mélange from a variety of species. We accept that the studies in our review showed great heterogeneity of service provision but all possessed the key central feature in the experimental group, a team-based community service. The fact that we were able to identify only five studies that satisfied the criteria for such a comparison, despite the widespread use of such teams, illustrates the consequences of deciding on policy in the absence of evidence. Once this is done, the subject cannot be researched through adequate randomised studies since policy makes the interventions statutory. Dr Holloway is right in concluding that community mental health teams have become the focus of mental health care in the UK and, although they are now universal here, it is still possible to carry out further randomised controlled trials elsewhere. We are in the process of developing similar studies in Eastern Europe, which should help to provide a stronger evidence base for our conclusions if they replicate the findings in the five studies we reported.

What would be most unfortunate at this stage of development of a community mental health team would be to move on to a new model based on the North Birmingham approach (Reference PeckPeck, 1999) without further evidence. The North Birmingham model has not been tested by any form of controlled comparison and there is now a strong body of evidence, to which Dr Holloway himself is a major contributor (Reference Holloway and CarsonHolloway & Carson, 1998; Reference BurnsBurns, 2000; Reference TyrerTyrer, 2000), which shows the standard community mental health team to be a robust and effective service model that is at least as effective as the new specialist approaches.

To return to the fruit metaphor, our review, and the work of others, seems to have established firmly that apples, grapes and oranges are good for your health when compared with other non-fruit diets. Recently, mangos, paw-paws and persimmons, have also been introduced and have attracted considerable numbers of devotees. To date, these exotic fruits have not proved in any way to be superior in their health-giving properties than the older fruits; until they do so we should not change our fruit policy. So we should stick with the community mental health team. James Lind, the originator of the first ever controlled trial — of citrus fruit juice for scurvy — would not have expected anything less.