Measuring and interpreting outcome is more difficult in mental health services than in some other areas of health care, for at least five reasons. First, the effect of the treatment may be to slow decline or to maintain the current level, so the score on the outcome measure itself may not improve (or may even get worse) despite best quality clinical care. Secondly, the best available evidence in the United Kingdom indicates that clinical and social variables predict no more than 30% of the variance in an individual's quality of life (UK700 Group, 1999). Thirdly, different types of outcome are desynchronous (e.g. Drury et al. 1996), changing at different rates during an intervention. Fourthly, there may not be agreement regarding what is a positive change in outcome – the patient who has fewer episodes of mania as a result of treatment may see this as a negative outcome. Finally, three levels of mental health service can be differentiated: treatment (specific interventions); programme (combination of different treatment components); and system (all programmes for a defined target group in a given area) (Burns & Priebe, 1996). The outcome data needed to evaluate each level will be very different.

Background. Heavy users of psychiatric services, often defined as the population that uses the most beds, consume a large part of the resources used by the whole service, despite being relatively small in number. Any intervention that reduces heavy use is therefore likely to lead to significant savings, and enhancement of standard care using a form of intensive case management akin to assertive community treatment was thought to be a pragmatic strategy for testing in this group.

Methods. The effectiveness of enhanced community management (ECM) was compared with standard care alone in heavy users, who represented the 10% of patients with the highest number of hospital admissions and occupied bed days over the previous 6·5 years in an outer London borough. One hundred and ninety-three patients were randomly assigned to ECM or standard care and their use of services was determined after 1 and 2 years, with assessments of costs, clinical symptoms, needs, and social function made before entry into the study and after 1 and 2 years.

Results. Despite a 2·4 fold increase in community contacts in the study group, there were no significant differences between the two groups in any of the main outcome measures. Small savings on in-patient and day-hospital service costs were counterbalanced by the increased costs of outpatient and community care for the subjects assigned to ECM. Clinical outcome data derived from interviews in two-thirds of the subjects were similar in both groups.

Conclusions. Providing additional intensive community – focused care to a group of heavy users of psychiatric in-patient services in an outer London borough does not lead to any important clinical gains or reduced costs of psychiatric care.

Background. A 10-question screening scale of psychological distress and a six-question short-form scale embedded within the 10-question scale were developed for the redesigned US National Health Interview Survey (NHIS).

Methods. Initial pilot questions were administered in a US national mail survey (N = 1401). A reduced set of questions was subsequently administered in a US national telephone survey (N = 1574). The 10-question and six-question scales, which we refer to as the K10 and K6, were constructed from the reduced set of questions based on Item Response Theory models. The scales were subsequently validated in a two-stage clinical reappraisal survey (N = 1000 telephone screening interviews in the first stage followed by N = 153 face-to-face clinical interviews in the second stage that oversampled first-stage respondents who screened positive for emotional problems) in a local convenience sample. The second-stage sample was administered the screening scales along with the Structured Clinical Interview for DSM-IV (SCID). The K6 was subsequently included in the 1997 (N = 36116) and 1998 (N = 32440) US National Health Interview Survey, while the K10 was included in the 1997 (N = 10641) Australian National Survey of Mental Health and Well-Being.

Results. Both the K10 and K6 have good precision in the 90th–99th percentile range of the population distribution (standard errors of standardized scores in the range 0·20–0·25) as well as consistent psychometric properties across major sociodemographic subsamples. The scales strongly discriminate between community cases and non-cases of DSM-IV/SCID disorders, with areas under the Receiver Operating Characteristic (ROC) curve of 0·87–0·88 for disorders having Global Assessment of Functioning (GAF) scores of 0–70 and 0·95–0·96 for disorders having GAF scores of 0–50.

Conclusions. The brevity, strong psychometric properties, and ability to discriminate DSM-IV cases from non-cases make the K10 and K6 attractive for use in general-purpose health surveys. The scales are already being used in annual government health surveys in the US and Canada as well as in the WHO World Mental Health Surveys. Routine inclusion of either the K10 or K6 in clinical studies would create an important, and heretofore missing, crosswalk between community and clinical epidemiology.

Background. We describe for the first time the epidemiology of homicide–suicide incidents for England and Wales. Previous descriptions have been of incidents in London (1946–62) and Yorkshire and Humberside (1975–1992).

Methods. Death certificates were obtained for all who died in homicide–suicide incidents in England and Wales (1988–1992) that were reported by the police to the Home Office. Incidents were included in the analysis if the interval between death or fatal injury of victim and suspect was 3 or fewer days.

Results. Three hundred and twenty-seven people died in 144 incidents (180 victims and 147 suspects). Eighty per cent of incidents had one victim and one suspect. Three incidents were also suicide pacts between two suspects killing their children. Eighty-eight per cent of incidents exclusively involved members of the same family, 9% acquaintances or strangers, and 3% both family and acquaintances or strangers. Seventy-five per cent of victims were female, 85% of suspects male. The victims of male suspects were predominantly their womenfolk, past and present, and their children, and of female suspects their young children. Car exhaust and firearms accounted for 40% of victim and 50% of suspect deaths. Of all homicides during 1988–1992, 3% of male, 11% of female and 19% of child deaths occurred in homicide–suicide incidents. Similarly, of all suicides, 0·8% of male and 0·4% of female deaths occurred in homicide–suicide incidents.

Conclusions. Homicide–suicide in England and Wales is mostly ‘a family matter’, men of predominantly lower social class killing their kin, and pre-menopausal mothers their young children, before they kill themselves. A few men kill strangers during a crime and then themselves.

Background. Somatic and psychiatric morbidity may cluster because of reciprocal effects between them but also as a result of common underlying factors.

Methods. The data come from the 1946 MRC birth cohort (N = 5362). Clustering of 20 chronic medical conditions at the participants' 43rd year was analysed using multi-morbidity coefficients. Associations of somatic and psychiatric morbidity at 43 years, with temperament and behaviour, assessed from adolescence to early adulthood, and background variables like paternal social class, were examined using structural equation modelling. The focus was on whether links applied to both morbidity types or one only, and were direct or were indirect, mediated by prior morbidity.

Results. Reported chronic medical conditions clustered strongly. Somatic multi-morbidity and psychiatric ill-health at 43 years were associated with temperamental and behavioural features assessed between the subjects' 13th and 26th birthday, like neuroticism and aggression, as well as with external variables like parental death before the participants' 16th birthday. However, only neuroticism holds direct links with somatic as well as with psychiatric ill-health, 28% of the former and 52% of the latter association being independent of the simultaneous presence of the other morbidity type and of participants' health status 7 years earlier.

Conclusions. Personality traits like neuroticism not only raise the risk of psychiatric disorder but also, irrespective of whether manifest psychiatric disorders have developed, of a broad spectrum of chronic somatic diseases. This suggests that clinicians and researchers should focus not only on the psychiatric disorders associated with such personality traits but also on their medical consequences.

Background. Previous epidemiological studies indicate large cross-national differences in prevalence of depression.

Methods. At 15 centres in 14 countries, 25916 primary care patients were screened for common mental disorders. A stratified random sample of 5447 primary care patients completed a baseline diagnostic assessment and 3197 completed a 12-month follow-up assessment. Psychiatric symptoms and diagnoses were assessed using the Composite International Diagnostic Interview (CIDI). Interviewer-rated disability was assessed using the Social Disability Schedule (SDS).

Results. Prevalence of current major depression varied 15-fold across centres. When centres were divided into three groups according to prevalence rates, the symptom pattern or latent structure of depressive illness was generally similar at low-, medium-, and high-prevalence centres. Depression was universally associated with disability, but this association varied significantly (t = 3·51, P = 0·0005) across centres. At higher-prevalence centres, depression was associated with lower levels of impairment. At 1 year follow-up, higher prevalence centres had both significantly higher rates of depression onset (t = 3·11, P = 0·002) and higher rates of persistence among those depressed at baseline (t = 2·49, P = 0·013).

Conclusions. Large cross-national variations in depression prevalence cannot be attributed to ‘category fallacy’ (cross-national differences in the nature or validity of depressive disorder). Use of identical measures and diagnostic criteria may actually identify different levels of depression severity in different countries or cultures. Cross-national differences in the onset and outcome of depression may reflect either true prevalence differences or differences in diagnostic threshold.

Background. Previous studies have shown that the presence of conduct disorder may contribute to the persistence of attention deficit-hyperactivity disorder (ADHD) symptomatology into adolescence; however, the aetiological relationship between the two phenotypes remains undetermined. Furthermore, studies utilizing multiple informants have indicated that teacher ratings of these phenotypes are more valid than maternal reports.

Methods. The genetic structure underlying the persistence of ADHD and oppositional-defiant disorder/conduct disorder (ODD/CD) symptomatologies as rated by mothers and teachers at two occasions of measurement was investigated on a sample of 494 male and 603 female same sex adolescent twin pairs participating in the Virginia Twin Study of Adolescent Behavioral Development (VTSABD).

Results. Using structural modelling techniques, one common genetic factor was shown to govern the covariation between the phenotypes across informants and occasion of measurement with additional genetic factors specific to ODD/CD symptomatology and persistence of symptomatology at reassessment. Genetic structures underlying the phenotypes were, to some extent, informant dependent.

Conclusions. The findings indicate that it is unlikely that the co-morbidity between ADHD and ODD/CD is due to environmental influences that are independent of ADHD. Rather it is likely to be due to a shared genetic liability either operating directly, or indirectly through gene–environment correlations or interactions. The covariation between phenotypes across informants and time is governed by a common set of genes, but it seems that ODD/CD is also influenced by additional genetic factors. Developmentally, different forms of genetic liability control ADHD in males and inattention in females.

Background. Routine administration of Health Related Quality of Life (HRQoL) and needs assessment instruments has been advocated as part of clinical care to aid the recognition of psychosocial problems, to inform clinical decision making, to monitor therapeutic response and to facilitate patient–doctor communication. However, their adoption is not without cost and the benefit of their use is unclear.

Method. A systematic review was conducted. We sought experimental studies that examined the addition of routinely administered measures of HRQoL to care in both psychiatric and non-psychiatric settings. We searched the following databases: MEDLINE, EMBASE, CINAHL, PsycLIT and Cochrane Controlled Trials Register (to 2000). Data were extracted independently and a narrative synthesis of results was presented.

Results. Nine randomized and quasi-randomized studies conducted in non-psychiatric settings were found. All the instruments used included an assessment of mental well-being, with specific questions relating to depression and anxiety. The routine feedback of these instruments had little impact on the recognition of mental disorders or on longer term psychosocial functioning. While clinicians welcomed the information these instruments imparted, their results were rarely incorporated into routine clinical decision making. No studies were found that examined the value of routine assessment and feedback of HRQoL or patient needs in specialist psychiatric care settings.

Conclusions. Routine HRQoL measurement is a costly exercise and there is no robust evidence to suggest that it is of benefit in improving psychosocial outcomes of patients managed in non-psychiatric settings. Major policy initiatives to increase the routine collection and use of outcome measures in psychiatric settings are unevaluated.

Background. In the closing decades of the twentieth century, changes in population sociodemographics took place that might be thought to have an adverse influence on the nation's psychological distress. Here, we examine the stability of social and gender inequalities in psychological distress throughout the 1980s and 1990s.

Methods. The study uses data from the 1958 National Child Development Study and the 1970 British Cohort Study collected when the cohort members were aged between 23 and 42. Multilevel logistic regression models were used to examine the effects of social class, gender, age, period and cohort on psychological distress as measured by the Malaise Inventory.

Results. We identify clear social inequalities in psychological distress during 1981–2000 that reduced in magnitude over this period. Non-linear age effects were observed: psychological distress improves in early adulthood but declines again on approaching mid-life. The 1970 cohort had poorer psychological distress than the 1958 cohort. Although women had higher rates of psychological distress than men, gender differences reduced in magnitude. Declining rates of women's psychological distress over time have not been matched in men. A reduction in social inequalities over time was also observed. Improvements in the psychological health of those in manual occupations were not equalled among those in non-manual occupations.

Conclusions. Both social and gender inequalities have narrowed in the last two decades of the twentieth century.

Background. Psychiatric disorders are purported to play a role in the aetiology of violent crime, but evidence for their role in sexual offending is less clear. The authors investigated the prevalence of psychiatric morbidity and personality disorders in elderly incarcerated sex offenders compared with elderly non-sex offenders.

Method. One hundred and one sex offenders and 102 non-sex offenders aged over 59 years were interviewed using standardized semi-structured interviews for psychiatric illness (the Geriatric Mental State) and the personality disorder (Structured Clinical Interview for DSM-IV personality disorders). Data on demographic, offence and victim characteristics were collected.

Results. Six per cent of the elderly sex offenders had a psychotic illness, 7% a DSM-IV major depressive episode and 33% a personality disorder; and 1% had dementia. These prevalence figures were not different from the elderly non-sex offenders interviewed in this study. Differences emerged at the level of personality traits with sex offenders having more schizoid, obsessive–compulsive, and avoidant traits, and fewer antisocial traits compared with non-sex offenders.

Conclusions. Elderly sex offenders and non-sex-offenders have similar prevalence rates of mental illness. However, elderly sex offenders have increased schizoid, obsessive–compulsive, and avoidant personality traits, supporting the view that sex offending in the elderly is associated more with personality factors than mental illness or organic brain disease.

Background. The literature on executive function in antisocial populations is unclear due to variation in diagnostic criteria and variation in the inclusion of healthy control comparison groups. Some studies suggest prototypical psychopathy is associated with specific deficits in ventromedial prefrontal (VMPFC) function rather than dorsolateral prefrontal (DLPFC). Meta-analytical studies, however, suggest that antisocial personality disorder may be associated with a broader range of executive deficits. This study assessed DLPFC and VMPFC function in antisocial personality disorder subjects and controls using the Cambridge Neuropsychological Test Automated Battery (CANTAB) and a Go/NoGo task respectively.

Methods. All subjects were screened for Axis I pathology, substance misuse and prescribed medication. The performance of 29 subjects with antisocial personality disorder DSM-IV and 20 male right-handed controls (matched for age and IQ) on the neuropsychological test battery was compared.

Results. Subjects with antisocial personality disorder displayed impairments on DLPFC executive function tasks of planning ability and set shifting. Impairments were also seen in VMPFC Go/NoGo tasks and in visual memory tasks.

Conclusions. Antisocial personality disorder is associated with a broad range of deficits in DLPFC and VMPFC function. Future studies need to examine relationships between the interpersonal and behavioural components of antisocial personality disorder and neuropsychological function.

Background. Social support may reduce the risk of psychiatric illness. Though perceived as an environmental measure, genetic factors may influence levels of social support. A relationship between social roles and personality with social support suggests possible sex effects on the sources of individual differences in social support.

Method. We used the responses of MZ and DZ same and opposite sex twins to 16 questions regarding their social life. Six factors – friend support, relative support, friend problem, relative problem, confidants and social integration were used for structural equation modelling. Factor derived scales were analysed for genetic, shared and unique environmental influences. Quantitative and qualitative gender differences were analysed using the software package Mx.

Results. Except for relative support and confidants, no qualitative sex differences were seen. Genetic and individual specific environmental influences accounted for the variance for friend support, friend problems, relative problems and social integration and no quantitative gender differences were seen. For relative support genetic factors were detected in females but not males, while for confidants, the shared environment was important in females but not males.

Conclusions. Except for relative support in males, genetic factors influence variation in all dimensions of social support. Shared environmental factors influence relative support and relative problems in both sexes. Sex differences were detected for confidants and relative support.

Background. Despite a number of studies that have indicated impaired memory function in patients with schizophrenia, there have been few that have used a sensitive measure of right medial temporal lobe pathology. Given the reported findings of reduced hippocampal volume in schizophrenia, we used a theoretically sensitive test of the right medial temporal lobe to determine the nature of the visuospatial memory deficit in the disorder.

Methods. Seventy-six patients (37 with a first-episode schizophreniform psychosis, and 39 with established schizophrenia) were compared with 41 comparison subjects on a number of tests of visuospatial memory. These included spatial working memory, spatial and pattern recognition memory and a pattern-location associative learning test.

Results. Both patient groups displayed recognition memory deficits when compared to the comparison group. However, only those patients with established schizophrenia (of 9 years duration on average) were impaired on the associative learning test.

Conclusions. The results indicate either a progressive decline in visuospatial associative learning ability over the course of the disorder, or that poor visuospatial associative learning is a marker for poor prognosis. In addition, these results have implications for our understanding of the role of the right medial temporal lobe in the pathophysiology of schizophrenia.

Background. Exposure to an ill parent in childhood may be a risk factor for adult somatization. This study examines the hypothesis that somatizing adults are more likely to have been exposed to illness as a child and that in turn, their children are more likely to report ill health and to have more contact with medical services than children of other mothers.

Method. A cross-sectional comparative investigation of three groups of mothers and their children of 4–8 years of age: (i) 48 mothers suffering from chronic somatization; (ii) 51 mothers with chronic ‘organic’ illness; and (iii) 52 healthy mothers was carried out.

Results. Somatizing mothers were more likely than other women to report exposure to childhood neglect and to physical illness in a parent (OR 2·9; 95% CI 1·4–6·1). The children of these somatizing mothers were more likely to have health problems than were the children of organically ill or healthy women and had more consultations with family doctors (average annual rates: somatizers 4·9 (S.D. 3·8), organic 3·0 (S.D. 3·5) and healthy 2·8 (S.D. 2·6)). Multivariate modelling of consultation rates among children found significant main effects for maternal somatization, maternal childhood adversity, the child's tendency to worry about health and a two-way interaction of maternal childhood adversity and her somatization status.

Conclusions. The hypotheses are broadly supported. However, it is important to emphasize the extent to which these findings are based on maternal reports.

Background. Many studies suggest that pregnancy and birth complications (PBCs) are environmental risk factors for child psychopathology. However, it is not known whether the effects of PBCs occur independently of genetic predisposition. The current study examined the possibility of gene–environment interaction in a twin design.

Method. The East Flanders Prospective Twin Survey prospectively records the births of all twin pairs born in East Flanders, Belgium. The current study included 760 twin pairs aged 6–17 years. Multilevel regression analysis was used to assess the effects of several PBCs collected around the time of birth. Using structural equation modelling, ACE models assuming additive genetic (A), shared environmental (C) and unique environmental (E) influences, were compared in order to examine whether the contribution of genetic factors to parent-rated child problem behaviour varied as a function of exposure to dichotomously and continuously defined PBCs.

Results. A main independent effect of lower birth weight, corrected for gestational age (small for gestational age – SGA), on child problem behaviour was found. In addition, there was an interaction between genetic influence and SGA, in that being smaller for gestational age resulted in less influence of additive genetic factors on individual differences in problem behaviour.

Conclusions. Results are suggestive of negative gene–birth weight interaction. Children who are SGA are less sensitive to the genetic effects, and those with high genetic vulnerability are less sensitive to the effects of being SGA in bringing about post-natal mental health effects.

Background. It is well documented that females have higher rates of internalizing disorders (anxiety, depression) than males. It is also well known that females have higher exposure to childhood sexual abuse and sexual assault. Recently, it has been proposed that the higher levels of internalizing disorders in females may be caused by their greater exposure to sexual violence.

Method. Data were gathered as part of the Christchurch Health and Development Study. In this study a cohort of 1265 children born in Christchurch, New Zealand, in 1977 have been studied from birth to age 21 years. The measures collected included: major depression and anxiety, childhood sexual abuse and adolescent sexual assault.

Results. Findings confirmed the established conclusion that internalizing disorders are over twice as common in females than males (ORs 2·2–2·7). In addition, it was found that females were exposed to higher rates of sexual violence than males (ORs 5·1–8·4). Statistical control for gender related differences in exposure to sexual violence reduced the associations between gender and anxiety and depression. Nonetheless, even after such control, gender was significantly (P<0·0001) related to both anxiety (OR = 1·8; 95% CI, 1·3–2·4) and depression (OR = 1·9; 95% CI, 1·4–2·3).

Conclusions. Greater female exposure to sexual violence may be a factor that contributes to greater female susceptibility to internalizing disorders. However, even after adjustment for gender differences in exposure to sexual violence it is clear that a substantial relationship between gender and internalizing disorder persists.

Background. As part of a collaborative linkage study, the authors obtained clinical and demographic data on 160 families in which more than one sibling was affected with a bipolar illness. The aim of the study was to identify clinical characteristics that had a high degree of familiality.

Method. Data on age at onset, gender, frequency of illness-episodes and proportion of manic to depressive episodes were examined to determine intra-pair correlations in affected sibling pairs. Dimension scales were developed measuring frequency and severity of lifetime mania, depression, psychosis and mood-incongruence of psychotic symptoms; degree of familial aggregation for scores on these dimensions was calculated.

Results. Sibling pairs correlated significantly for age at onset (ρ = 0·293, P<0·001); dimension scores for psychosis (ρ = 0·332, P < 0·001); and proportion of manic to depressive episodes (ρ = 0·184, P = 0·002). These findings remained significant when correcting for multiple testing. Of the other test variables; mania (ρ = 0·171, P = 0·019); incongruence dimensions (ρ = 0·242, P = 0·042); frequency of manic episodes (ρ = 0·152, P = 0·033); and frequency of depressive episodes (ρ = 0·155, P = 0·028) were associated with modest correlations but these were not significant after correction. Degree of familial aggregation was not significant for sex (κ = 0·084) or dimension scores for depression (ρ = 0·078, P = 0·300).

Conclusions. Significant but modest familial resemblance has been shown for some specific features of bipolar illness, particularly age at onset and degree of psychosis. Further research may establish the extent to which these findings are mediated by genetic and/or environmental factors.

Background. Studies that examined community samples have reported several risk factors for the development of depressive episodes. The few studies that have been performed on primary care samples were mostly cross-sectional. Most samples had originated from highly developed industrial countries. This is the first study that prospectively investigates the risk factors of depressive episodes in an international primary care sample.

Methods. A stratified primary care sample of initially non-depressed subjects (N = 2445) from 15 centres from all over the world was examined for the presence or absence of a depressive episode (ICD-10) at the 12 month follow-up assessment. The initial measures addressed sociodemographic variables, psychological/psychiatric problems and social disability. Logistic regression analysis was carried out to determine their relationship with the development of new depressive episodes.

Results. At the 12-month follow-up, 4·4% of primary care patients met ICD-10 criteria for a depressive episode. Logistic regression analysis revealed that the recognition by the general practitioner as a psychiatric case, repeated suicidal thoughts, previous depressive episodes, the number of chronic organic diseases, poor general health, and a full or subthreshold ICD-10 disorder were related to the development of new depressive episodes.

Conclusions. Psychological/psychiatric problems were found to play the most important role in the prediction of depressive episodes while sociodemographic variables were of lower importance. Differences compared with other studies might be due to our prospective design and possibly also to our culturally different sample. Applied stratification procedures, which resulted in a sample at high risk of developing depression, might be a limitation of our study.

Background. Complaints of poor memory and concentration are common in veterans of the 1991 Persian Gulf War as are other symptoms. Despite a large research effort, such symptoms remain largely unexplained.

Method. A comprehensive battery of neuropsychological tests and rating scales was administered to 341 UK servicemen who were returnees from the Gulf War and peace keeping duties in Bosnia, plus non-deployed military controls. All were drawn from a large randomized survey. Most were selected on the basis of impaired physical functioning defined operationally.

Results. Group comparisons revealed an association between physical functioning and symptoms of depression, post-traumatic stress reactions, increased anger and subjective cognitive failures. Poorer performance on some general cognitive measures, sequencing and attention was also seen in association with being ‘ill’ but virtually all differences disappeared after adjusting for depressed mood or multiple comparisons. Deployment was also associated with symptoms of post-traumatic stress and subjective cognitive failures, independently of health status, as well as minor general cognitive and constructional impairment. The latter remained significantly poorer in the Gulf group even after adjusting for depressed mood.

Conclusions. Disturbances of mood are more prominent than quantifiable cognitive deficits in Gulf War veterans and probably lead to subjective underestimation of ability. Task performance deficits can themselves be explained by depressed mood although the direction of causality cannot be inferred confidently. Reduced constructional ability cannot be explained in this way and could be an effect of Gulf-specific exposures.

Background. Mental disorders have an increased mortality risk. However, most data have been provided by few countries, some mental disorders have received little attention, long-term studies of large samples are scarce, and insufficient control for confounding variables has lead to artefactual inconsistencies across studies. The aims of this study were: to quantify the mortality risk in psychiatric patients 5 to 21 years after hospital admission and to investigate temporal trends in mortality risk and predictive factors associated with mortality.

Method. All patients admitted to an in-patient psychiatric unit in Italy between 1978 and 1994 were included and vital status and death causes were determined up to 21 years after admission. The observed number of deaths in the sample was compared with the expected number of deaths in the general population. Cox proportional hazard models were fitted to identify predictors of mortality.

Results. Mortality from natural and unnatural causes was higher than expected across all mental disorders. Standardized mortality risk was higher in males (SMR = 4·55; 95% CI 4·17–4·97) than in females (SMR = 3·43; 95% CI 3·07–3·83). Individuals aged less than 40 years were at higher risk in both sexes. The first several years following admission were characterized by a faster decline in survival. Several demographic and clinical factors were predictors of mortality.

Conclusions. Mortality is high in individuals with mental disorders. Prevention of unnatural death causes is an important goal though insufficient to abate excess mortality, since natural death causes account for it to a larger extent.