Many will remember the launch of the Tory white paper, Working for Patients, in 1989. Softened up by the razzmatazz more usually associated with boxing matches, everyone in the NHS was herded together in halls across the nation to watch a video of Maggie. Big reforms hit us with the subtlety of a heavy-weight's right to the jaw.

What a difference this time. The New NHS was presented with the subtlety of an opening gambit in a complex game of chess. A hushed silence has followed while everyone contemplates their next moves.

Rather than interpret the white paper as an art form created for its own sake, it is more helpful to question whether the new policy direction is likely to produce better care on the ground. For mental healthcare, key objectives have to be strategic commissioning, effective service configurations and a well-trained workforce.

The scourge of fragmented commissioning has had a profound effect on mental health services. Trusts had to negotiate with a plethora of independent fundholding practices and some with several health authorities. Community teams had to cope with the consequences, applying different models of care street by street.

Few tears will be shed about the disappearance of fundholding, and the idea of local commissioning by primary care groups is exciting. Sensibly, each group can choose its preferred model on the basis of its interests and expertise. Considerable variation is likely, all for the better as we all believe in responsiveness to local choice. But how will HAs manage if groups in their area go for the full range of options, requiring HAs to commission comprehensive services for some patches, but to delegate complete budgets to others? And what if some groups want to merge with the community trust, but others reject the option?

Merging HAs into sub-regional units, more distant from the action and even shorter of resources, while relying on the individual capacity of approximately 500 new primary care groups, is risky. Much will depend on the ability of practice managers and GPs, and the control which future HAs will dare to impose through health improvement programmes. It's useful to consider that so far no total purchasing pilot or other primary care fundholding model has succeeded in commissioning a complete mental health service. Let's not be sanguine about the skill pool of staff, relatively low paid since pounds3 per head of population will not buy much expertise, even if it were available.

Whoever or whatever will be responsible for mental health, the partnership with social service commissioners will remain crucial if we want to eliminate bed-blocking and duplication of community services.

The last government's green paper (remember the four options?) was a big step in the right direction, but seems to have been binned. The need for joint planning runs through The New NHS, and a statutory duty of partnership is introduced, but no consistent structural solution is offered.

Unintentionally, the white paper makes a strong case for integrating mental health and social care when discussing the option of mental health becoming part of primary care trusts. Yet this is rejected because 'health and social care boundaries are not fixed and... joint work is particularly important'.

Rightly, the white paper observes that 'an integrated range of services from community to hospital care' is required in mental healthcare. Fuzzy boundaries may be an indication of an artificial split, and the merging of health and social care may be a solution, irrespective of participation in primary care trusts.

This leads to the second criterion - provider reconfiguration. Presently, most mental healthcare is provided in combined trusts, mainly with community services, and most function well. The creation of large mental health trusts may give rise to new monopolies not very interested in GPs' local ideas. Whether large specialist mental health trusts are the best mechanism to achieve joint working is open to question, especially if co-terminosity with HAs, primary care groups and social services is subject to continuing change.

The new boundaries with primary care trusts may re-open old wounds over the responsibility for people with minor mental illness, always a great frustration among community teams. Increasingly, the rejected option of including mental healthcare in primary care trusts, combining health and social care commissioning and providing, seems attractive. It may be an idea to reconsider when the dust settles over the community trust primary care experiment.

The bottom line of any reform has to be its impact on direct patient care. It never ceases to astonish me how distant most reforms are to service users and clinical staff. When I gave a lecture about policy to NHS workers recently, most had no idea about the roles of HAs or fundholders, nor were they interested. Why should they be?

Staff are struggling because of poor dissemination and implementation of evidence-based practice. Users are frustrated because they perceive that bad clinicians have little accountability. Proposals to introduce clinical governance and a National Institute for Clinical Excellence could make more impact than any of the proposed structural changes.

Yes, the white paper is exciting, with considerable scope for strategic evolution. It rightly addresses key structural flaws and introduces ideas to improve quality of care. It also proposes a performance framework based on clinical outcomes. This is most likely to be achieved if we prioritise best practice, retaining and developing high-quality practitioners alongside the massive energy that will be invested in reconfiguration.

The point of the latter is, after all, merely to achieve the former. Dr Matt Muijen is director of the Sainsbury Centre for Mental Health.

'The new boundaries with primary care trusts may re-open old wounds over the responsibility for those with minor mental illness... Increasingly, the rejected option of including mental healthcare in primary care trusts, combining health and social care commissioning, seems attractive'