It would be surprising if a public service was tolerated when it was feared by its customers, who are put at risk; unable to show evidence of its effectiveness; very expensive; and paying its staff uncompetitively. It would be astonishing if, nevertheless, such a service could not cope with demand.

This is a recognisable picture of acute hospital care in the NHS. The 14 000 NHS acute in-patient beds in England for adults aged 16-64 years cost about £750 million annually. Recent visits by the Commission for Health Improvement (CHI) confirmed earlier studies (Sainsbury Centre for Mental Health, 1998; Standing Nursing and Midwifery Advisory Committee, 1999) identifying acute care as particularly challenging because of high bed occupancy rates, poor hygiene, lack of space, boredom, lack of privacy, scarcity of therapeutic interventions, violence and bed blocking. Although not all these problems occurred on every ward, none was without its challenges and patients were highly critical. The consensus between a College Council Report on admission wards (Royal College of Psychiatrists, 1996) and a user report (Reference RoseRose, 2000) is striking, both highlighting similar factors as the CHI reports.

No recent studies, nor any in the distant past of which I am aware, have demonstrated the therapeutic effectiveness of hospital care as compared with other services. Indeed, hospital care is often used as a control group or outcome measure in health service research, allowing a range of alternative services to show their relative clinical and economic advantages. The evidence against hospital care as a therapeutic and cost-effective intervention is strong and consistent.

The image of hospital as a place of last resort, which is applied reluctantly, and the development of alternatives such as crisis homes and community teams has meant that the threshold for admission has risen over time. The increasing rates of involuntary admissions and the reduced rates of non-psychotic disorders are indicators of increasing severity of case mix. The development of prestigious community services such as assertive out-reach and crisis resolution is an irresistible draw to staff, especially since they offer easier access to higher grades and salaries. In combination, this means a spiral of case mix of increasing severity, growing fear and violence, recruitment problems, worsening care and a greater search for alternatives, leading to a yet more disturbed group of people on the wards receiving poorer care.

Despite all these critical arguments, few would argue for the abolition of hospital care. There is, and I postulate will be, a need for a secure and therapeutic place to care for people when they are a serious risk to themselves or others for psychiatric reasons, potentially complicated by medical, psychological or social factors. The question to consider is, what structure and process of care such places should offer to be therapeutic and acceptable. To put it differently, we need therapeutic and safe 24-hour care, but the concept of ‘hospital’ needs reforming.

Until recently, government policy ignored these questions and concentrated on the development of community services. The strategy was to reduce demand rather than review and adjust supply. Many pages of the National Service Framework (Department of Health, 1999) are committed to descriptions and examples of good community services, and a few to the failings of hospital care and how to reduce their use further. The NHS Plan (Department of Health, 2000) adds greater specificity to the structure of community care, ignoring hospital care altogether. The only policy-based developments have been the increase in numbers of a variety of secure beds and the establishment of single gender wards (NHS Executive, 2000).

The publication of the Mental Health Policy Implementation Guide, Adult Acute Inpatient Provision (Department of Health, 2002), is a major step forward. It places in-patient services in a system of care, makes recommendations and sets standards for issues such as care, leadership, communications, environment and human resources. Most, inevitably, are rather generalised, and the importance of this document is probably more a sign of a sensible readjustment of priorities rather than attempting to impose centralised solutions. The good practice examples are very helpful. Depending on one's perspective, the additional £25 million earmarked for 2002 to improve ward conditions is either a significant gesture or a drop in the ocean.

However, the Government strategy and the islands of innovative practice cannot hide a general poverty of vision. While there is agreement about the problems and failures, few visions or solutions have emerged from research or practice, and there is certainly no consensus on the way forward. The explanation for this is most likely to be a combination of the complexity of hospital care and the low status assigned to it within the care system. Some factors that cannot be ignored but are difficult to disentangle are estate, human resources, the balance between therapy and safety, diversity of needs, the pressure to implement community services and the tremendous cost implications. It is highly dubious whether a single magic bullet exists for the range of problems hospital care is expected to solve, involving a broad range of conditions affecting a wide variety of people.

The fundamental issue hospital care struggles to address is how to create an optimal therapeutic environment while maintaining safety. Most other questions flow from this. The current situation, where everyone is hoarded together — young men with schizophrenia and old ladies with depression — irrespective of age, gender, culture, risk, diagnosis and treatment, is irrational and unacceptable. The widely prevalent practice to group beds primarily on the basis of the responsible medical officer and secondarily on geography is supply driven and possibly pragmatic for wardrounds, but it is hardly in the interest of patients. Even if this is agreed, the alternative is not obvious. Assume that each of the above six categories is bimodal (e.g. age in two groups) and ignore the question of whether such a degree of specialisation is sensible, the answer to both obviously being negative, this would require 64 units. Unless we believe that large regional institutions are a good idea, such divisions are impractical. This does not mean that sensitive, small-scale restructuring offers no gains, but a centrally imposed solution is unlikely to be best suited to local priorities, such as the presence of specialist expertise or specific cultural needs.

The structural question is related to the case-mix issue. Some advocate centralised units that would allow constant medical cover and flexible staffing to deal with emergencies. In reality, this means a minimum of about 45 beds divided into three wards, covering at least five community mental health teams (CMHTs) (Royal College of Psychiatrists, 1998). Others prefer small, community resource centres providing 8-10 beds, linked to a single team and integrated into a locality, but obviously not offering any specialist care. Again, local context, including geography, may determine preferences, but there are obvious pay-offs. Either way, a major investment in new building programmes will be necessary to replace or transform existing units that are sometimes not more than 25 years old.

However well we design hospital wards, the quality of care will only be as good as staff are able to provide. The challenge is how hospitals are able to recruit competent staff in sufficient numbers. For recruitment, three layers of stigma have to be neutralised: the NHS as a poor employer, mental health and the wards themselves. A review is needed of training for practitioners, to include skills specific to hospital care. The development of competencies by the College is an important development. The question is whether hospital care is posing such unique challenges that it should become a speciality in its own right. We need to create incentives that will make working on the wards attractive.

Finally, we have to design hospital care around the wishes of the patient, if only because otherwise wards will neither be therapeutic nor safe. Long periods of boredom with a cheap radio booming in the background, interspersed with medication dispensing three times a day and a weekly alienating wardround is no longer acceptable. People expect access to staff and a choice of activities in addition to evidence-based therapies.

Assumptions about hospital care have to be challenged, new models tested and, if effective, disseminated. Such a project, ‘Acute Solutions’ is being developed in a partnership between user groups, the Department of Health, Royal College of Psychiatrists, Royal College of Nursing and Sainsbury Centre for Mental Health. Pilot sites have been selected in Colchester, Derby, Liverpool and Worthing. The tremendous interest of clinicians, managers and users testifies to the need for change.

The combination of a functional structure and personalised care, delivered by sufficient and competent staff, may allow hospital to be a therapeutic component in a comprehensive system of mental health care, respected by staff and patients. It has to be done, since the status quo is unacceptable and unsustainable. It will take energy, imagination and money. Whether it is a priority to design care for the most vulnerable people in society is the one question we can answer easily.

Declaration of interest

None.