The criticism of the PRiSM Psychosis Study (Reference Marshall, Bond and SteinMarshall et al, 1999; Reference Sashidharan, Smyth and OwenSashidharan et al, 1999) betrays several misconceptions about the nature and philosophy of community mental health teams in the UK. Unlike in the USA, where assertive community treatment (ACT) teams were set up in a desert of community care, any similar teams in the UK have to adjust to working in collaboration with other teams in the area and never aspire to providing a service for an entire catchment area, as Thornicroft et al (Reference Thornicroft, Becker and Holloway1999) have emphasised. Both Marshall et al and Sashidharan et al have failed to note that standard community care has improved enormously in the past 20 years and therefore can compete successfully with formal assertive approaches, including both ACT and intensive case management. Unlike drug/placebo comparisons, in which the effects of placebo are roughly similar whatever the year, complex psychosocial interventions such as those in a mental health service are changing constantly. I can predict with some confidence that the Cochrane review showing such excellent findings with regard to superiority of ACT in randomised controlled trials (Reference Marshall, Lock wood, Gray, Adams, Duggan and de Jesus MariMarshall et al, 1998) will show steadily decreasing benefits of ACT in future revisions. This is not because ACT has suddenly lost its effectiveness; standard treatments have caught up immensely in the past few years and have done so often by using different approaches to those of the original ACT programmes. The statement of Sashidharan et al (Reference Sashidharan, Smyth and Owen1999) that contemporary psychiatric care “continues to be dominated by thinking and practices which have their origin in the last century” is a travesty of the current position and a slur on the reputation and performance of many dedicated community mental health teams across the country. Such teams have cause for congratulation. Even though they are deprived of the resources that are allotted to ACT, particularly the requirement of a case-load of only 8-12 clients per worker, they are undoubtedly effective and may even have a positive effect on reducing suicide and other causes of undetermined death (Reference Tyrer, Coid, Simmonds, Adams, Duggan and de Jesus MariTyrer et al, 1999). But there is a limit to these benefits and some of those treated assertively may be better cared for in hospital. Sashidharan et al find it hard to conceive that intensive case management might increase violence in community settings. Unfortunately, antisocial behaviour in all its forms has been shown to be more prevalent in those with some personality disorders in this type of service than in one in which hospital treatment is given more readily (Reference Gandhi, Tyrer and EvansGandhi et al, 2000) and this could undermine progress towards better community care unless it is acknowledged as a problem.

It is time for the programme of assertive community treatment (PACT) model of community treatment to be judged in contemporary settings where, in most developed countries, there is reasonable statefunded community care. Len Stein, Mary Test and their colleagues in Wisconsin have made a great contribution to community care by the introduction of PACT but they should not adopt the reflex argument that all studies that show ACT or intensive care programmes to be less effective in other settings must be failing because they do not apply the PACT model properly. The fact is that PACT is primarily a care philosophy backed by a secondary compendium of interventions that have rarely been tested individually. One of its core features, the case-load of only 8-12 per worker, has recently been shown in a much larger trial than any others to be unimportant in influencing outcome (UK700 Group, 1999). There may be many other elements of PACT that are also unimportant. Stein & Santos (Reference Stein and Santos1998) quote our own work (Reference Merson, Tyrer and OnyettMerson et al, 1992) as indicating that ACT works outside the USA. Our service had case-loads of 20-25 per worker, did not have 24-hour cover and had psychiatrists working full-time in the team, all of which invalidates its description as an ACT model.

The other issue that must be taken account of by ACT enthusiasts is the need for ACT teams to have much closer liaison with existing teams when there is already well-established community care. This was never a problem at the time ACT was originally introduced as there was no competition. Now that there is a backbone of community care present in the UK and many other countries, it is really inappropriate for a new assertive team to come along and indulge in its autonomous activities with a small number of clients without establishing formal links with other teams, both in-patient and out-patient, in the relevant areas. I have drawn attention (Reference TyrerTyrer, 1999) to the similarity between ACT and plant succession in alien habitats; ACT is like a specialised plant that does extremely well in conditions that are alien to community care, but as it improves the circumstances for good care it gradually becomes redundant and can be replaced. This does not mean the principles of ACT are abandoned; it is just that the philosophy of seamless transfer between hospital and community is rarely supported by small teams based in the community with no responsibility for services beyond their immediate clients. Such teams and their evangelists should really be more humble before they advise others on how to run a comprehensive mental health service.