SynopsisOne hundred psychiatric emergencies presenting to an inner London teaching hospital had formal assessments of psychopathology, personality disorder (using both ICD-10 and the Personality Assessment Schedule), social networks and social functioning before being randomly assigned to a multidisciplinary community-based team (Early Intervention Service (N = 48) or conventional hospital-based psychiatric services (N = 52) and treated for a period of 12 weeks. The ICD-10 classification yielded a higher proportion (50%) of personality disordered patients than the Personality Assessment Schedule (34%) and those from ethnic minorities (mainly Afro-Caribbean) and upper social classes had a lower incidence of personality disorder. Social networks were smaller in personality disordered patients and there were fewer attachment figures. Improvement in social function, and to a lesser extent with depressive symptomatology, was better in patients with no personality disorder referred to the community service compared with the hospital service. No differences were found between the numbers and duration of social contacts in the two services and it is concluded that the better outcome in the community-treated patients was independent of changes in social networks.

A pilot study of the effects of early intervention by a multidisciplinary community team in patients presenting as psychiatric emergencies to an inner London teaching hospital (St. Mary's) is described. Of 22 consecutive emergency referrals, 20 agreed to be formally assessed, 18 were successfully engaged and 16 had assessments of symptoms and social functioning assessed over a four week period. There was significant improvement in clinical symptoms but no important changes in social functioning. The discordance between improvement in social function and symptoms was unexpected and may reflect the unusual social circumstances of life in metropolitan London or be a consequence of the high proportion (55%) of patients seen with personality disorders.

Group training in anxiety management was compared with individual general practitioner appointments as a means of reducing dosage of benzodiazepines. Subjects were encouraged to withdraw at their own pace. Drug reduction, self-ratings of anxiety, depression and Health Locus of Control comprised the outcome measures. Process changes were recorded using visual analogue scales, and daily diary records. At the end of a six-week intervention period, the average dosage had significantly reduced in both groups. This reduction was maintined at a 15-week follow-up. There were no other differences in outcome between conditions. Although subjects reported a variety of symptoms during drug withdrawal, their occurrence was not significantly associated with the degree of dosage reduction.