This study investigates the effects of assigned diagnostic labels, with their accompanying predictions of therapeutic response, and prescribed psychotropic drugs. It was thought that such ‘labelling’ effects might be important in ambiguous situations, such as neurotic anxiety–depressive states, where diagnoses of ‘reactive depression’ or ‘anxiety state’ might justifiably be made and treatment with either diazepam or nortripyline legitimately given. The depression label and its concomitant two-weekly prediction of improvement with antidepressants produced a set towards slower response with higher self-report depression. Nortriptyline produced significantly more improvement in self-report depression than diazepam in the first 2 weeks of treatment. No significant interaction effects were detected between diagnosis or drug. One month after the initial diagnostic evaluation there were no significant effects from either diagnostic label or drug.

In a prospective controlled investigation chronic patients, the majority with neurotic conditions, were randomly allocated either to community psychiatric nurses (CPN) as the main therapist or to routine out-patient psychiatrist follow-up. There was no statistically significant difference in the mean public expenditure for the two modes of care. The direct costs of psychiatric care comprised a small proportion of total public expenditure and were initially greater in the nursing group. Subsequently, and over the whole 18-month study period, nursing was the cheaper option in terms of these direct costs. Clinical and social outcomes were comparable in both follow-up groups. Consumer satisfaction was significantly greater among CPN patients. On balance, these results confirm the benefit of community psychiatric nursing for this patient group.

Life-event data for a 3-month period were collected for 1058 adults. Six ways of categorizing life events and difficulties were analysed: namely, loss (L), threat (T), anti-social act (A), hopeless situation (H), uncertainty of outcome (U), and choice of action (C). Reliabilities were moderate to good. The number and the pattern of characteristics within a single event or difficulty were found to be the most important elements in predicting a depression score criterion. Two patterns were particularly predictive: CH present in an event or difficulty, and LH present with U absent. Approximately 21·9% of the depression score variance maybe explained using life events together with the respondent's social class, sex and social support. This represents a significant improvement over prediction based solely on the scoring methods of Brown & Harris (1978).

A comparison was made between the characteristics of patients admitted to the only psychiatric hospital in Newfoundland, Canada, and those admitted to general hospital psychiatric units within the same area. Patients were compared on demographic characteristics, diagnosis, incidence of violence and suicidal attempts, treatment, length of stay and previous admissions. The results suggest that patients admitted to the mental hospital are more likely to be single, male, of lower socio-economic status and to suffer from more chronic conditions. There is also a higher incidence of violence among patients admitted to the mental hospital. These findings have implications for the types of changes that will be required in general hospital units if the role of mental hospitals continues to decline.

A cohort of 181 users of community psychiatric services in Camberwell with at least one year's history of contact was traced for two years following a census day. Four of the eight patients who had been given a primary diagnosis of an organic condition dicd during this period. For the others the population-standardized mortality risk was 2·4, similar to rates previously found for in-patients with acute disorders. The causes of death were those that are common in the population. Only one suicide was recorded.

In their recent paper on smoking and personality, McManus & Weeks (1982) administered the Eysenck Personality Questionnaire (EPQ) and questionnaires regarding smoking to 100 subjects, and calculated correlations between the personality traits measured (psychoticism, extraversion, neuroticism, and lie scale scores) and smoking. Contrary to expectation, there was no correlation with extraversion (E), only slight correlation with neuroticism, but a significant correlation with psychoticism (P). They argue that the lack of correlation with E is due to the changed contents of the scale, which has been changed from earlier versions, such as the Eysenck Personality Inventory (EPI), by the omission of some impulsiveness items. They quote a paper by Rocklin & Revelle (1981) which argues in a similar vein.