Over the last ten years a new approach to psychiatric knowledge has developed under the influence of social anthropology. Its origins, assumptions, methods, achievements, and limitations are reviewed.

Using data from the first community-based, epidemiological survey of Puerto Rico, this paper examines the current prevalence of use of spiritist healers by Puerto Ricans, the role of spiritism in the provision of mental health services, and the association between spiritism and psychiatric disorders and symptoms. Those who visit spiritists were found to be more likely to work outside the home, to have a low family income, to have sought help for emotional problems from mental health professionals, and to have mild symptoms of depression.

The present study is based on the so-called 1957 Lundby cohort, a geographically defined normal Swedish population of 2612 individuals who were evaluated for mental disorders in 1957 and 1972. The annual age-standardised first incidence of depression, with or without other psychiatric symptoms, all degrees of impairment included, was found to be 4.3 per 1000 person years in men and 7.6 per 1000 person years in women. Up until 70 years of age, the cumulative probability of suffering a first episode of depression was 27% in men and 45% in women.

Adults in a village in Lesotho, Africa, were interviewed to determine the community prevalence of major depression, panic disorder, and generalised anxiety disorder. The prevalence data were compared with data from a large epidemiological study in the United States utilising the same research instrument. There was a significantly higher prevalence of all three diagnoses in Lesotho as compared with the United States. As in the United States, women were at an increased risk for these disorders, although statistical significance was not demonstrated for depression. The majority of people (77%) who had experienced panic attacks said they had sought help for their symptoms, with the majority attending Western-trained doctors. The relationship between explanatory models and help-seeking behaviour was explored in people who had had panic attacks. Less than 40% of those with generalised anxiety disorder said they sought help.

A two-year follow-up was conducted of a subsample of the Chandigarh cohort of first-contact schizophrenic patients from the WHO Determinants of Outcome project. The patients were those living with family members who had been interviewed initially to determine their levels of expressed emotion (EE). The interview was repeated for 74% of the relatives at one-year follow-up. A dramatic reduction had occurred in each of the EE components and in the global index. No rural relative was rated as high EE at follow-up. Of the patients included in the one-year follow-up, 86% were followed for two years. In contrast to the one-year findings, the global EE index at initial interview did not predict relapse of schizophrenia over the subsequent two years. However, there was a significant association between initial hostility and subsequent relapse. The better outcome of this cohort of schizophrenic patients compared with samples from the West is partly attributable to tolerance and acceptance by family members.

A study of 36 French-speaking families in Geneva identified 66% of high-EE household. At nine-month follow-up, the relapse rate for the high-EE group (33%) was significantly higher than for the low-EE group (0%). These results seem to confirm the relevance of EE as predictor of relapse. However, in our sample, it appears that one component of EE, critical comments, was principally responsible for the prognostic validity.

Hospital-based annual incidence rates for schizophrenia in Croatia over 1965–84 did not change significantly. Rates ranged from 0.21 to 0.22 per 1000 population (0.26–0.29 per 1000 population aged over 15). Factors that could influence these rates were analysed, but it appeared that the rates were a true reflection of the incidence rate of schizophrenia in the Croatian population.

Based on information from a case register, patient age and diagnosis at first admission are analysed in a Croatian cohort of schizophrenics first admitted in 1972 and followed up through the register for 12 years. Diagnosis was analysed on the same basis and over the same period. Although the male and female differences in incidence rates for schizophrenia were not large, hospital incidence rates in younger age groups were higher in males. Males were also more commonly diagnosed as schizophrenic at first admission, females more frequently receiving diagnoses of affective psychosis and other organic psychosis, except for alcohol-induced psychosis.

Age at onset was determined in a sample of 360 patients representative of the 8069 schizophrenics hospitalised in SR Croatia. The 95% confidence interval for mean age at onset was 22.9–26.7 years. The difference between males and females was not significant, unlike the age difference between the sexes reported for age at first admission for schizophrenia. Results may be influenced by attrition of the original population of patients.

Between October 1983 and December 1985, 268 patients were brought by police to a psychiatric hospital in south London under section 136 of the Mental Health Act 1983. Comparisons were made between ‘blacks' and ‘whites' on several clinical and demographic variables. The vast majority of admissions received a psychiatric diagnosis. An excess of black admissions was recorded. Black men were younger, were more likely than whites to be given neuroleptics, to be put on compulsory orders, and to be given an out-patient appointment when discharged from hospital. More black men were given a case-note diagnosis of schizophrenia or drug-induced psychosis. The differences in clinical management between ethnic groups could be at least partly accounted for by these differences in diagnosis. Treatment did not appear to be independent of diagnosis among the black admissions.

In 25 Jewish and 26 non-Jewish depressives resident in Hackney, the PSE revealed that ‘hypochondriasis' and ‘tension’ were more common in the Jewish group, and special features of depression, containing symptoms connected with guilt, was less common. These differences could not be explained by variation in severity of illness as both groups had similar scores on the HRSD. The ‘social maladjustment’ scores were similar, very probably reflecting an elderly, deprived, inner-city population.

Asian and non-Asian referrals to the department of child psychiatry at the London Hospital in 1987 were analysed. Asian referrals were under-represented with respect to the local population but, contrary to our expectations, there were no significant differences between the groups with regard to demographic data, the nature of the presenting problem, and attrition rates.

Using data from the WHO International Pilot Study of Schizophrenia and from our own previously reported series of 120 patients receiving treatment for schizophrenia, we ascertained the degree of agreement between 13 different systems for diagnosing schizophrenia. We identified a triad of symptoms similar to that from the International Pilot Study: auditory hallucinations, passivity feelings, and disturbances of affect. This triad correlated very strongly with the diagnosis of schizophrenia as determined by the 13 diagnostic systems, as well as with response to fluphenazine in our series. The triad should serve as a core set of symptoms in the study of schizophrenic illness. It represents the leading phenomena in a group of patients having what is generally considered to be schizophrenia, although it does not provide a definitive diagnosis. Evidence is needed on its specificity.

Resolution of presenting problems at three-month follow-up was studied in 228 parasuicides. ‘Non-resolvers' presented with less severe problems, but prodromes were longer, prior episodes more frequent, and they experienced more powerlessness and internally directed hostility than the resolvers. At follow-up resolvers had improved more on measures of depression, externally and internally directed hostility, locus of control, powerlessness, self-esteem, sensitivity to criticism, and social adjustment. Baseline problem severity, powerlessness (negatively), and ‘normlessness' ‘predicted’ degree of problem resolution, while baseline problem severity, the number of previous parasuicide episodes, and baseline internal hostility (the latter two variables negatively) were associated with problem resolution as a dichotomous variable. Resolvers improved more in social functioning during the follow-up than non-resolvers, but a similar proportion (16%) in each group reported repeat episodes of parasuicide at follow-up.

To discover why parasuicides who resolved their difficulties repeated self-harm within three-month follow-up at the same rate as those who had not resolved problems, a group of 30 repeaters was compared with 156 non-repeaters. The repeaters had a history of more frequent episodes, beginning younger. Their problems were more severe, the acts of self-harm potentially less lethal. They experienced greater feelings of externally directed hostility, powerlessness, and ‘normlessness'. A constellation of nine variables correctly predicted 81.5% of the repeaters and 77.5% of the non-repeaters. At follow-up the non-repeaters had improved on several parameters but the repeaters were essentially unchanged. The resolvers among the subgroup of repeaters were more like repeaters in the overall sample than the resolvers, which may explain why some parasuicides repeat in spite of resolving their problems.

Thirty-one patients dependent on benzodiazepines were randomly assigned to either slow withdrawal (SW) or abrupt withdrawal under propranolol cover (PW). Of 16 patients in the SW group, 11 successfully withdrew from their drugs, while only 4 out of 15 in the PW group did so. Patients in the SW group had only mild withdrawal symptoms, while those in the PW group suffered more severe symptoms, which lasted around four weeks. In all, 81 % of the whole group suffered withdrawal symptoms of some kind. Patients in both groups were significantly less anxious at the end of the study than at baseline. Younger subjects and those who were more severely anxious at the start of the trial had more difficulty in withdrawing than older and less anxious patients.

The heads of centres collaborating with the WHO in biological research on mental disorders decided to develop a series of consensus statements on key issues in biological research and treatment of psychiatric problems. The statement on prophylactic use of anticholinergics in patients on neuroleptic treatment is the second of the proposed series.

We have reservations about accepting the conclusions made by Crow et al (1989), namely that there is evidence that susceptibility to schizophrenia may be transmitted by a pseudoautosomal locus. In our view the methodology of the study is unsound. Additionally the data presented by the authors, when analysed correctly, do not support their hypothesis.