We audited 184 psychiatric wards against clinical practice guidelines for the management of violence. Staff and service users completed anonymous questionnaires. Environmental inspections were performed by two teams.

There were 4460 questionnaires returned. Nurses (78%) were significantly more likely to report the experience of violence than service users (37%). Drugs were reported by 72% of nurses and alcohol by 61% as causing problems. Other standards frequently not met included staffing levels, training, provision of activities, ward design and ambience.

Specific issues are identified that must be addressed by national and local action. A baseline is set against which the impact of this action can be judged. Priorities must include tackling drug and alcohol use in psychiatric wards.

A substantial minority of patients attending emergency departments do so with a primary mental health problem. A number will require admission to an acute psychiatric unit. There is little guidance about the best practice in terms of risk assessment and escorting of these patients. We conducted a telephone questionnaire survey of acute trusts, to identify current policies for the transfer of patients with mental illness to acute psychiatric units.

Eighty-two trusts were contacted. We could not verify the presence of any current policies. There was considerable variation in the assessment and management of risk in these clinical circumstances.

Current local service structure and transfer distance may influence the way in which risk is assessed. However, there is a need for trusts to ensure that basic standards of risk assessment are maintained for the transfer of these high-risk patients. We recommend that all emergency departments should have such policies and that liaison psychiatry is well placed to aid their development.

Trainee psychiatrists often perform emergency mental health assessments. Traditionally, it has been considered that having access to past psychiatric records will reduce the likelihood of a patient being admitted. We examined whether the availability of records had an influence on admission by recording all contacts to the duty junior psychiatrist in two district general hospitals over a 6-month period.

For those with chronic or enduring mental illnesses there is a 27% increase in the likelihood of admission if past records are available. For all other patients the increase is 10%.

Contrary to our expectations, the availability of records increases the likelihood of admission to mental health admission units.

We sought to identify changes in the quality of information in referrals to an old age psychiatry service before and after the introduction of the single assessment process. Referrals were compared in terms of length, legibility, information and clinical utility.

Compared with letters before the introduction of the single assessment process, referrals made on the new forms took longer to read (mean 96 v. 124 s, P=0.001), had more illegible sections (P=0.011), contained less information (P=0.026) and were judged to be less clinically useful (P=0.001).

The introduction of the single assessment process has impaired clinical communication between general practitioners and psychiatrists, and might be prejudicial to patient care.