New policies have rained down on the NHS over the last ten years, from internal markets to risk management, care planning to quality assurance. Yet the National Service Framework (NSF) for Mental Health deserves particular attention, because it will play such a central role - or good or ill - in shaping mental health provision over the next decade.

And there is much to be welcomed in the NSF. Few readers will want to disagree with autonomy and respect, user and carer involvement, accessibility, effectiveness, safety and accountability as the NSF's basic principles.

 The problem with the NSF as I see it does not lie in its genuinely good intentions and stated aims. My concerns arise from the questions that the NSF does not ask; the assumptions that are not challenged, or even discussed. Chief among these assumptions is the underlying model of mental distress that informs the document, which is quite clearly a biomedical one.

A few extracts illustrate this: 'One adult in six suffers from one or other form of mental illness... One person in 250 will have a psychotic illness such as schizophrenia or bipolar affective disorder... Many of these patients have not been getting the treatment and care they need... [The NSF] covers health promotion, assessment and diagnosis, treatment, rehabilitation and care...'

The vocabulary (illness, patient, treatment, diagnosis, and the uncritical use of terms such as schizophrenia, bipolar disorder and so on) clearly indicates an acceptance of the orthodox biomedical view of mental distress. Why is this so important? The reason is that the way in which we understand and conceptualise mental distress, which is both reflected in and reinforced by the language we use, determines just about everything else we do. It determines the interventions we think are most appropriate, the professionals who are most centrally involved, the settings they use, the training they get, the perceived role of service users, the kind of research we carry out, and so on.

If, for example, we see mental distress as primarily the result of social and relationship factors, we will end up with a very different set of aims for our ideal mental health service than if we see distress as deriving from genetic and biological factors.

As I have argued elsewhere, the biomedical model of mental distress is unproven, but it nevertheless remains the basis of most current psychiatric practice.' The danger of having a policy framework based on unexamined (and arguably mistaken) assumptions is that, however much money and effort you put into it, however many extra professionals are recruited and new teams set up, you will essentially be delivering the same old service under a different name, or at a different site, or by a different set of professionals. The apparently sweeping and desperately-needed changes will not actually amount to anything very new at all. All the faults of the old system will be carried through into the new one.

Pat Bracken and Phil Thomas, writing in this magazine, have made exactly this point in relation to community care: 'While the move...to community-based approaches opens up the possibility of a radically different mental health service, it does not in itself guarantee it. Some home treatment services work within a conventional medical model. We see dangers in this. For example, a central role of traditional psychiatric units has been surveillance, a concern with containment and monitoring. Users of such a service find this highly oppressive. Without a critical analysis of the role of psychiatry, home treatment can simply mean that. surveillance is transferred into the community. This results in an even greater intrusion of psychiatry into the lives of patients, the medicalisation of distress and an increasing dependence on professionals."

For an illustration of how this might happen, we can look at the NSF's enthusiasm for assertive outreach (AO) teams for what it describes as 'severely mentally ill people in the community'. In the most positive version, AO involves teams of professionals who work intensively with a small number of people in their own homes, offering 24-hour support and help with finances, housing and other practical matters.

For the most alarming version, we can look to the United States. There, much the same population is now being served by their version of AO, called Programme for Assertive Community Treatment, or PACT. This is wrapped up in quite an attractive package- round-the-clock support, low staff/client ratio and so on -but the central feature is the daily 'dose drop', where people are visited at home by professionals who supervise them taking their medication, backed up by court orders. (The court orders can mean immediate detention in a locked unit if people refuse.)

To advocate something called 'assertive outreach' without a detailed discussion of the very different philosophies that can be grouped under this heading is to open the door to some extremely undesirable developments.

Because these basic principles have not been thought through, it is actually very difficult to work out what some of the NSF's targets are about, and what they will achieve. Take for example, the NSF goals of drawing up agreed sets of treatment procedures for depression. What you consider good practice will depend entirely on your understanding of depression. Do you want, for example, to see everyone with a diagnosis of depression offered ten sessions of counselling. Or should they - as per the temporary guidelines adopted by the NSF -be making more use of medication?

The same kind of confusion can be seen in the selection of services which have been given a Beacon award (recognised as examples of good practice to be emulated). On the one hand there is commendation for the Royal College of Psychiatrists' current national anti-stigma campaign 'Changing minds', which issues information leaflets framed in strictly medical terms ('schizophrenia is the name given to a set of symptoms...it is something to do with the connections between nerve pathways in the brain' and so on.)

On the other hand, the NSF also praises the Bradford Home Treatment Service, which works within an explicitly social framework of mental distress and which avoids where possible using psychiatric diagnoses at all. It is impossible for both these projects to be examples of best practice, because they start from models that are contradictory.

Interestingly, though, there are hints in the NSF of an alternative model of mental distress running alongside the biomedical one. It notes that: 'Mental health problems can result from the range of adverse factors associated with social exclusion For example, unemployed people are twice as likely to have depression as people in work, children in the poorest households are three times as likely to have mental health problems than children in well off households ', and so on.

This is clearly a social model of mental distress. This may be a risky viewpoint for any government to put forward -because the solutions are likely to be so much more complex and expensive than immediate medical ones. But it is also actually incompatible with the biomedical one that seems to be dominant in government thinking.

If people with psychiatric labels are really suffering from the long-term effects of poverty, unemployment, violence and poor housing, then how is medication going to help them? The more you see people's distress as an understandable response to their circumstances, the less sense it makes to describe it as the 'symptom' of an 'illness'.

Administering drugs then starts to look less like treating an illness and more like suppressing and sedating their entirely appropriate despair. But no government can afford to explore this contradiction very far. If pressed, they would probably resort to an intellectual trick that is commonly seen in psychiatry -the social factors are described as merely the triggers of an underlying biologically-based illness. This is not how service users themselves see it: 'It was not these factors which led to their problem; these were the problem Respondents saw their difficulties as meaningful in the context of their life experiences."

This kind of debate raises the whole awkward question of what, or who, psychiatry is for. The biomedical model gives a simple answer: psychiatry is for diagnosing and treating mental illnesses. Questions such as what are 'mental illnesses'? What does it mean to 'diagnose' someone? What are we really 'treating'? What role does psychiatry play in society? threaten to unravel this naive view.

The failure to have such a debate, or even to realise that it is necessary, will inevitably result in policies and practices that are as confused and contradictory as the thinking behind them. The National Service Framework, for all its good intentions, may turn out to be the biggest missed opportunity of the decade.

1. Johnstone, L. (2000) Users and Abusers of Psychiatry: A Critical Look at Psychiatric Practice 2nd edition (Routledge). £15.99+ 10%p&p from Mind Mail Order, 15-19 Broadway, London E15 4BQ, tel. 0208221 9666, e.mail publications@mind.org.uk

2. Bracken, P and Thomas, P(1999) 'Post-psychiatry', Openmind 95, p.17

3. Rogers, A., Pilgrim, D. and Lacy, R. (1993) Experiencing Psychiatry: Users' Views of Services, (Macmillan/Mind)

Illustration by Gary Kempston