Sex differences in rates of depressive disorder have not been convincingly explained, and this reflects a more general failure of research to provide a comprehensive aetiological account of depression. The difference can be used as a probe for evaluating the research base of integrative models of depressive disorder (e.g. Akiskal & McKinney, 1975). It is particularly likely to be illuminating if the causes of the sex difference do not overlap completely the causes of depression itself. While there have been many reviews in the area (Weissman & Klerman, 1977; Kessler & McRae, 1981; Wolk & Weissman, 1995; Bebbington, 1996), this point has not been adequately expressed.

Several lines of investigation are necessary for assessing the relative importance of social, psychological and biological influences: the epidemiological study of macrosocial variables and of age effects; temperament, personality, and attributional and coping styles; the experience of psychosocial adversity; and the possibility of increased susceptibility to some forms of stress in women. Both the tendency to affiliation and the requirement for social support may differ by sex. The particular strains of the roles available to women may increase their risk of depression. Possible genetic explanations of the sex difference are of special relevance because of their implications for biological differences. The latter can also be studied directly: hormonal theories in particular must be evaluated.

Background. Women are consistently reported to have a greater prevalence of depressive disorders than men. The reason for this is unclear, and is as likely to be social as biological. There is some evidence that the excess of depression is greater during women's reproductive lives. Data from the National Survey of Psychiatric Morbidity were used to test the hypothesis that the excess disappeared in the post-menopausal years and that obvious social explanations for this were inadequate.

Method. Subjects (N=9792) from a random sample of the British population provided data for the analysis. Psychiatric assessment was carried out by lay interviewers using the CIS-R. Subjects with ICD-10 depressive episode or mixed anxiety/depression were compared with the remainder. Social variables that were likely to contribute to a post-menopausal decline in depressive disorders were controlled in logistic regression analyses.

Results. There was a clear reversal of the sex difference in prevalence of depression in those over age 55. This could not be explained in terms of differential effects of marital status, child care, or employment status.

Conclusions. This large and representative survey adds considerably to the increasingly held view that the sex difference in prevalence of depression is less apparent in later middle age. This may be linked to the menopause, and our attempts to explain it in terms of obvious conditions among social variables were not successful. More specific studies are required to clarify the finding.

Background. In the present paper, sex differences in the association between adult depression and childhood experiences were examined.

Methods. The study series of the Finnish UKKI Study consisted of a population sample of 501 men and 499 women. Information concerning childhood experiences was gathered retrospectively in a baseline survey carried out in 1969–72. After the initial phase, the mental health of the subjects was evaluated by interviews, questionnaires and register data at the 5-year follow-up (1974–6) and at the 16-year follow-up (1985–7).

Results. Twelve per cent of men and 21% of women had suffered from depression during the 16-year follow-up period. A disturbed mother–child relationship and neurotic symptoms in childhood were associated with depression in women but not in men in the logistic model that included gender interaction. In separate analyses by gender several childhood factors showed statistically significant associations with depression in women but only a few in men.

Conclusions. The finding suggests that childhood experiences are more highly predisposing factors to depression in women than in men.

Background. A previous paper (Abas & Broadhead, 1997) reported that among 172 women randomly selected from a Zimbabwean township 30·8% had a depressive or anxiety disorder during the previous year. Compared with London, the higher annual prevalence of disorders in Harare could mostly be accounted for by an excess of onset cases in the study year (annual incidence of depression 18%). This paper reports on the role of life events and difficulties in the aetiology of depression among these women.

Method. Randomly selected women (N=172) from a township in Harare were interviewed with a Zimbabwean modification of the Bedford College Life Events and Difficulties Schedule (LEDS).

Results. Events and difficulties proved critical in provoking the onset of depression in Harare. Far more events occurring in Harare were severe or disruptive. Furthermore, a proportion of the Harare severe events were more threatening than have been described in London. As in London, certain types of severe event were particularly depressogenic, i.e. those involving the woman's humiliation, her entrapment in an ongoing difficult situation, or bereavement. However, more severe events in Harare involved these specific dimensions.

Conclusions. Results indicate a common mechanism for the development of depression, as defined by international criteria, between Zimbabwe and London. The high frequency of severe events, and their especially adverse qualities, offer an explanation for the high incidence of depression in Harare.

Background. A prospective study, covering just over a 1-year period, sought to confirm an earlier finding that around 40% of women who experience a severe life event in the presence of two ongoing psychosocial vulnerability factors, (negative close relationships and low self-esteem) will develop a major depressive episode. Distal risk factors were examined to see if they improved prediction.

Methods. A population sample of 105 mainly working-class mothers with vulnerability and without depression were interviewed three times over a 14-month period to date the occurrence of severe life events and onset of major depression. Degree of vulnerability was assessed at first contact together with distal risk in terms of childhood neglect/abuse and any earlier episodes of depression.

Results. Thirty-seven per cent of these vulnerable women became depressed in the study period. The majority experienced a severe life event, and of these, 48% had onsets. Contrary to expectation, risk was only a little less among those with just one of the two vulnerability factors. Two-thirds of women with an onset had been depressed in previous years. Although this was associated with increased risk, the effect was greatest for those who had experienced an episode before age 20. A relationship between childhood neglect/abuse and onset was entirely accounted for by such early depression.

Conclusion. An aetiological model of depression outlined in earlier research was confined with a new factor of teenage depression shown to increase risk of onset.

Background. Previous work has indicated that the 2[ratio ]1 female[ratio ]male sex ratio in unipolar depressive disorders does not emerge until some time between ages 10 and 15.

Methods. Data from four annual waves of data collection from the Great Smoky Mountains Study (GSMS) involving children aged nine to 16 were employed.

Results. Pubertal status better predicted the emergence of the expected sex ratio than did age. Only after the transition to mid-puberty (Tanner Stage III and above) were girls more likely than boys to be depressed. The timing of this transition had no effect on depression rates. Before Tanner Stage III, boys had higher rates of depression than girls, and the prevalence of depression appeared to fall in boys at an earlier pubertal stage than that at which it began to rise in girls. In addition, recent transition to Tanner Stage III or higher had a transient effect in reducing the prevalence of depression in boys.

Conclusions. The period of emergence of increased risk for depression in adolescent girls appears to be a relatively sharply demarcated developmental transition occurring in mid-puberty. Previously reported effects of the timing of puberty (which have tended to be transient) appeared less important in increase of risk for depression than pubertal status.

Background. A prospective epidemiology study evaluated the role of specific social and psychological variables in the prediction of depressive symptomatology and disorders following childbirth in a community sample. Measures of social support used previously in clinically depressed populations facilitated further comparison.

Methods. Nulliparous pregnant women (N=507) were interviewed during pregnancy with the Interview Measure of Social Relationships (IMSR) and a contextual assessment of pregnancy-related support and adversity and 427 were followed up at 3 months postpartum with the 30-item GHQ, including six depression items. To establish the clinical representativeness of the GHQ, high GHQ scorers and a random subsample of low scorers were interviewed using the SCAN. Regression models were developed using the GHQ Depression scale (GHQ-D), the IMSR and other risk factor data.

Results. GHQ-D after childbirth was predicted by lack of perceived support from members of the woman's primary group and lack of support in relation to the event becoming pregnant; this held even after controlling for antenatal depression, neuroticism, family and personal psychiatric history and adversity. Informant-rated deficits in provision of social support also predicted later depression. The size of the primary social network group previously found to be related to depression in women, did not predict depressive symptom development.

Conclusion. Predictors of depressive symptom development differ from predictors of recovery from clinical depression in women. Interventions should be designed to reduce specific deficits in social support observed in particular study populations.

Background. Very little is known about the pharmacokinetics of neuroleptic drugs in breast-feeding mothers and their infants or about possible adverse effects in the infants.

Method. Twelve mothers who breast-fed their infants were prescribed haloperidol, chlorpromazine or trifluoperazine. Two methods, enzyme immunoassay (EIA) and high performance liquid chromatography (HPLC) were used to assay these drugs in samples from mothers, but infants' samples were assayed only by the more sensitive EIA. Repeated clinical and developmental assessments of the breast-fed infants were carried out up to 30 months of age. The control subjects were 18 bottle-fed infants whose mothers were also prescribed neuroleptic or mood-stabilizing drugs.

Results. The total concentrations of neuroleptic drugs and their principal metabolites in maternal plasma were correlated with concentrations in fore-milk. Infants were ingesting up to 3% of the maternal daily dose per kg body weight and small amounts of the drugs were detected in infants' plasma and urine. Concentrations of haloperidol in the adult range were found in plasma from 2 of 5 infants assayed by EIA but there was no evidence of any acute or delayed adverse effects. Three other breast-fed infants whose mothers were prescribed both haloperidol and chlorpromazine showed a decline in their developmental scores from the first to the second assessment at 12–18 months.

Conclusion. More extensive longitudinal studies are needed but, in the meantime, there appears to be grounds for caution if breast-feeding mothers are prescribed doses of single or two neuroleptic drugs at the upper end of their recommended ranges.

Background. Women with anorexia nervosa have a reduced fertility but they may have borne children before the onset of their illness or after partial recovery. Little is known on how to identify the anorexic mothers who underfeed their children and how to manage them. This article aims to remedy these gaps.

Methods. The clinical scientific method is the only means of identifying the children of anorexic mothers who are at risk. Eight such mothers were identified as a result of obtaining serial measurements of the children's weights and heights over time. Tanner–Whitehouse charts were used to plot weight for age and height for age. A simple rating scale was devised to measure the acceptance of treatment involving both mother and child.

Results. Nine children (eight boys and one girl) were found to have suffered food deprivation: with severe reduction in weight-for-age in six and in height-for-age in eight. Five siblings were not affected. Catch-up growth was correlated with the degree of engagement in treatment of both mother and child. Long-term treatment of one mother, combining family therapy with admissions to hospital, resulted in catch-up growth in her two sons.

Conclusions. The mechanisms underlying the privation of the children stem from the anorexic mother's abnormal concerns with body size extending to her children. The children may become unduly accepting of the underfeeding. It is essential to obtain the confidence of anorexic mothers suspected of underfeeding their children and to adopt a whole family approach to treatment.

Background. As part of a longitudinal study, prevalence findings of DSM-IV disorders are presented for a random sample of 3021 respondents aged 14 to 24, with response rate 71%.

Method. Assessment included various subtypes of disorders, subthreshold conditions and disorders that have only rarely been studied in other epidemiological surveys. The computer-assisted Munich-Composite International Diagnostic Interview (M-CIDI) was used to derive DSM-IV diagnoses.

Results. Substance disorders were the most frequent (lifetime 17·7%; 12-month 11·4%), with abuse being considerably more frequent than dependence. Other mental disorders had a lifetime prevalence of 27·5% (12-month, 17·5%) with depressive disorders (16·8%) being more frequent than anxiety disorders (14·4%). Eating disorders (3·0%) and threshold somatoform disorders (1·2%) were rare disorders. Subthreshold anxiety and somatoform disorders, however, were more frequent than threshold disorders. Prevalence of disorders was equally high for males and females, although specific disorder prevalence varied significantly by gender. The co-occurrence of disorders (co-morbidity) was substantial and was significantly related to greater reductions in work productivity and increased rates of professional helpseeking behaviour.

Conclusions. Findings underline that mental disorders in young adults are frequent and impairing, limiting work and education ability and social interaction. Given the fact that adolescents and young adults are in a key phase of socialization in terms of professional career and interpersonal relationships, our findings indicate a considerable risk potential for an accumulation of complicating factors and future chronicity. This paper is the first report of this ongoing longitudinal study about early developmental conditions of mental disorders.

Background. To assess the prevalence of delusional ideas in primary-care patients.

Method. A survey was carried out with the Aquitaine Sentinel Network of general practitioners (GPs). Consecutive practice attenders were invited to complete the Peters et al. Delusional Inventory (PDI-21) self-report questionnaire, designed to measure delusional ideation in the normal population. GPs, blind to the questionnaire results, provided information on patients' psychiatric history.

Results. Of the 1053 attenders included in the survey, 348 (35%) had a lifetime history of psychiatric disorder, of whom 20 (2%) had a history of broadly defined psychotic disorder. The self-report questionnaire was completed by 790 patients. The range of individual PDI-21 item endorsement in subjects with no psychiatric history varied between 5 and 70%, suggesting that delusional ideation is a dimensional phenomenon lying on a continuum with normality. The main discriminative items between psychotic and non-psychotic patients were those exploring persecutory (OR=15·2, 95% CI 4·3–53·7), mystic (OR=6·4, 95% CI 1·9–22·4) and guilt (OR=5·8, 95% CI 1·5–23·2) ideas.

Conclusions. This survey demonstrates that questions that explore delusions and hallucinations are well-accepted by most primary-care patients. More research is needed on psychotic disorders in primary-care settings to improving early identification of these disorders.

Background. Primary care attenders with a common mental disorder (CMD) frequently present with somatic symptoms. This study aimed to examine somatic and psychological models of CMD in primary care attenders in India.

Methods. Cross-sectional survey of attenders at two primary care clinics. Psychiatric caseness was determined on three criteria: standardized psychiatric interview (biomedical criterion), patients' self-assessment of emotional disorder (emic criterion) and health care provider diagnosis. The GHQ-12 and the PPQ, which emphasize psychological and somatic symptoms respectively, were used as screening instruments.

Results. Although somatic symptoms were the presenting complaints for 97% of subjects, 51% of subjects with a biomedically defined CMD had a psychological illness attribution. Patients with psychological attributions were more likely to be women, to have a longer duration of illness, to have higher CISR scores and were more likely to be recognized by the primary health care (PHC) physician. The GHQ-12 was superior to the PPQ in identifying cases of CMD against the biomedical criterion for both psychologizers and somatizers; both instruments performed equally well against the emic and care provider criteria.

Conclusions. Psychological models may be acquired by patients as CMD becomes more chronic or severe, making them more likely to be detected by PHC physicians. Psychological symptoms are superior to somatic symptoms in detecting CMD. Shorter versions of the GHQ have comparable discriminating abilities to the 12-item version and offer the practical advantage of brevity, which may make them more acceptable to PHC physicians as a clinical screening tool.

Background. Factor analysis has been employed to identify latent variables that are unifying constructs and that parsimoniously describe correlations among a related group of variables. Confirmatory factor analysis is used to test hypothesized factor structures for a set of variables; it can also, as in this paper be used to model data from two or more groups simultaneously to determine whether they have the same factor structure.

Method. Non-psychotic psychiatric morbidity, elicited by the Revised Clinical Interview Schedule (CIS-R), from four culturally diverse populations was compared. Confirmatory factor analysis was employed to compare the factor structures of CIS-R data sets from Santiago, Harare, Rotherhithe and Ealing. These structures were compared with hypothetical one and two factor (depression–anxiety) models.

Results. The models fitted well with the different data sets. The depression–anxiety model was marginally superior to the one factor model as judged by various statistical measures of fit. The two factors in depression–anxiety model were, however, highly correlated.

Conclusions. The findings suggest that symptoms of emotional distress seem to have the same factor structure across cultures.

Background. The purpose of the present study was to examine the measurement properties of positive affect items among the Japanese population.

Methods. Responses to the Japanese version of the Center for Epidemiologic Studies Depression Scale and four additional negatively revised items of the original positive affect items were compared for 85 Japanese psychiatric out-patients with dysphoric-mood-related symptoms and 255 demographically matched controls.

Results. Responses to positive affect items were generally comparable between the two groups, whereas responses to negative symptom items were markedly different (P<0·002 for all comparisons). The group difference was most marked for symptom persistence. Responses to the four negatively revised items of positive affect revealed a similar picture to that of the negative symptom items. The internal consistency of the scale significantly improved when the original positive affect items were replaced by the negatively revised items (P<0·001 for both).

Conclusions. Positive affect items with positive wording cannot be used to assess depressive disorders in the Japanese population adequately, but this can be done with the corresponding negatively revised items.

Background. Increasingly, evaluations of mental health services include an economic component, although often only summary statistics such as the mean or median are reported. Measures of variation are often limited to the standard error or standard deviation, though costs are rarely normally distributed and vary substantially between patients. The aim of this study is to identify factors that can explain variations in the cost of mental health services for epidemiologically representative cases of psychosis.

Methods. Cases with ICD-10 diagnoses of functional psychosis were drawn from a sample that included all identified cases of psychosis in two geographically defined sectors in Camberwell, south London. Mental health service use was measured and costed. A predictive model was developed using multiple regression analyses, in which patient characteristics and previous service use indicators were used as predictor variables. Services were measured using the Client Service Receipt Interview.

Results. Among the 147 cases included, the amount of cost variation explained by the model was 31·5%. The most significant predictive factor was social functioning, which was highly negatively associated with cost.

Conclusions. Current mental health service use can be predicted to a reasonable extent by previous service use and patient characteristics, especially the degree of social disability. Identification of such factors can aid the effective allocation of scarce resources. In particular, subgroups of patients who use most resources can be identified and targeted by mental health services.

Background. The patients' ability to appraise their quality of life in schizophrenia was studied by examining the reliability and the validity of self-rated quality of life estimates.

Methods. Sixty-three symptomatically stable patients with schizophrenia (DSM-IV) receiving maintenance treatment were evaluated over a 4-week period. The subjects were asked to appraise their quality of life at weekly intervals on a single item global quality of life measure, as well as the self-administered sickness impact profile. The patients' quality of life was also rated by a clinician using the social performance schedule and the global assessment scale of functioning; and clinical aspects such as the severity of psychotic symptoms, neurocognitive deficits, dose of medications, and side effects were documented with standardized measures.

Results. The results indicated that the patients' self-reports were highly consistent over the 4 weeks, and the quality of life ratings correlated significantly with the clinician's estimates. The patients' quality of life was predictably influenced by the severity of their symptoms, side effects, cognitive deficits and the dose of their antipsychotic medication, but the reliability of their reports was not materially affected by these factors.

Conclusions. It is concluded that clinically compliant and stable patients with schizophrenia can evaluate and report their quality of life with a high degree of reliability and concurrent validity, implying that self-report measures are potentially useful tools in clinical trials and outcome studies.

Background. Analysing costs measures in conjunction with psychiatric case register (PCR) data can provide important epidemiologically-based information on resource utilization. Costing the service use patterns of first-ever patients can indicate the shape and likely resource consequences for mental health services operating within a community-based system of care.

Methods. Yearly costs were calculated for the 299 first-ever patients and 768 longer-term patients who contacted the South-Verona Psychiatric Case Register between 1 January 1992 and 31 December 1993. Bivariate and multivariate analyses were used to compare costs between these groups and to test the associations between costs and the sociodemographic and diagnostic data recorded on the PCR.

Results. For all diagnostic groups identified, first-ever patients were found to be less costly to support than longer-term patients, even after adjustment for various factors, including whether patients were single consulters. When multivariate analyses were employed, between 20% and 69% of the cost variation for first-ever patients could be explained by patient and other characteristics, and the effect of the contact (first or subsequent) variable was reduced.

Conclusion. This study considered only the costs to the specialist psychiatric services but the methodology allows the likely annual resource implications of supporting new patients to be predicted from data collected at first contact. Such information can help ensure that services are adequately funded and that the resources are deployed appropriately between client groups.

Background. We examined a group of subjects at familial risk of depression and explored the relationship between the perceptions of parents and a history of depression. We also investigated: (a) whether any difference in perceived parenting found between those with and without a past history of depression was an artefact of the depression; and (b) whether the relationship between parenting and depression was explained by neuroticism.

Method. We took a sample of first-degree relatives selected from a family study in depression and subdivided them by their history of mental illness on the SADS-L, into those: (a) without a history of mental illness (N=43); and (b) those who had fully recovered from an episode of RDC major depression (N=34). We compared the perceptions of parenting, as measured by the Parental Bonding Instrument (PBI), in these two groups having adjusted for the effect of neuroticism and subsyndromal depressive symptoms. We also had informants report on parenting of their siblings, the latter being subdivided into those with and without a past history of depression.

Results. Relatives with a past history of depression showed lower care scores for both mother and father combined compared with the never ill relatives. The presence of a history of depression was associated with a non-significant reduction in the self-report care scores compared to the siblings report. Vulnerable personality (as measured by high neuroticism) and low perceived care were both found to exert independent effects in discriminating between the scores of relatives with and without a history of depression and there was no interaction between them.

Conclusion. This study confirmed that low perceived parental care was associated with a past history of depression, that it was not entirely an artefact of having been depressed, and suggested that this association was partially independent of neuroticism.

Background. Empirical evidence indicates that manipulations of attention may facilitate changes in cognition and stress symptoms in emotional disorder.

Methods. The present study reports the effects of Attention Training (ATT) in a brief case series of three patients with primary hypochondriasis using an A–B–A design.

Results. ATT produced clinically significant improvements in self-reported measures of affect, and illness-related behaviour and cognition. Treatment gains were maintained at 6 months follow-up assessments. Measures of body-focused attention indicated that the ATT procedure acted on attentional processes as intended.

Conclusions. The present case series extends the effects of ATT to problems of hypochondriasis. A number of studies now suggest that ATT is associated with a reduction in anxiety and negative beliefs across disorders of panic, social phobia and hypochondriasis. Controlled clinical trials are now required to establish firmly the effects of ATT as a component of cognitive therapy.

Background. Various studies from Australia, Canada and the United States have shown significant rank correlations between the suicide rates of immigrants and those of their country of birth (COB). This study compares the rank ordering of age standardized suicide rates of immigrants in Australia for two periods, 1961–70 and 1979–90: (a) between each period; and (b) with their COB for each period.

Methods. Data were obtained from the World Health Organization Annual Statistics and from the Australian Bureau of Statistics. Comparisons were made for 11 countries from England and Wales, Ireland and Europe, for which there was a sufficiently large number of immigrant suicides to warrant statistical analysis.

Results. The data showed considerable heterogeneity in rates of immigrants from various countries, with increased rates in Australia compared with their COB. There were consistently significant Spearman rank correlations between the rates after immigration and those in their COB for each period, and between rates in the two periods for both immigrants and for their COB, despite increases in suicide rates, and considerable socio-economic demographic changes between the various countries over that time span.

Conclusions. The findings are used to argue two conclusions: (i) the important influence of pre-migrant social and cultural experiences in subsequent suicide rates in immigrants in their host country; and (ii) to support the case for the reliability of using international suicide data for comparative epidemiological research.