This issue features substantial groups of papers on anxiety disorders and on depression. A review article (pp. 589–599) provides a meta-analysis of antidepressant treatment trials in people with schizophrenia. The authors find suggestive evidence of benefit but this depends on small studies. No evidence is found of psychotic worsening.

This issue features groups of papers on recurrent brief depression, psychological treatment of schizophrenia, psychological aspects of depression, attentional processes in ADHD and recovering alcohol abuse, and anxiety disorders. In addition an editorial reviews the role of GABA in mood disorders (pp. 387–393) and a systematic review of psychological autopsy studies of suicide finds mental disorder to be the strongest associated factor, with high population attributable fraction of 47–74% (pp. 395–405).

The issue of whether an association exists between cannabis and psychosis exists, and why, has received considerable attention in recent years (Hall, 1998; Mueser et al. 1998; Blanchard et al. 2000; Degenhardt & Hall, 2002). There are probably a number of reasons for the sustained interest. First, psychotic disorders such as schizophrenia are often chronic or recurring (Mason et al. 1996), are associated with significant disability (Keith et al. 1991) and they place a considerable burden upon the community at large (Hall et al. 1985; Knapp, 1997). Secondly, over the past few decades, high rates of cannabis use and use disorders have been observed among persons with schizophrenia and other psychoses (Barbee et al. 1989; Fowler et al. 1998), suggesting the possibility that cannabis use may be causally related to psychotic disorders. Thirdly, given what is known about the psychotomimetic effects of cannabis (Hall et al. 2001), it is plausible that high doses of cannabis may produce psychotic symptoms. Finally, clinical research with persons with psychotic disorders has found that problematical substance use is correlated with a range of negative outcomes including relapse, rehospitalization, poor medication compliance, poorer social functioning and increased treatment costs (Salyers & Mueser, 2001), suggesting drug use may be related to worsened clinical outcomes.

This issue features groups of papers on genetics, long-term outcome of depression, fatigue syndromes and somatization, and the neuropsychology of schizophrenia, autism and obsessive–compulsive disorder.

Psychological treatment in primary care. This issue features four papers on psychological treatment in primary care. An editorial by Scott & Sensky (pp. 191–196) also discusses the methodology of such studies. Bower et al. (pp. 203–215) report a meta-analysis of controlled trials of counselling. There are benefits but they are modest. Consistent with this, Simpson et al. (pp. 229–239) report a controlled trial of psychodynamic counselling in chronic depression, with little benefit over routine GP treatment. Proudfoot et al. (pp. 217–227) report benefit from an innovative computerized interactive cognitive-behavioural program for depression and anxiety. Gabbay et al. (pp. 241–251) report on agreement on baseline problems between patient, GP and therapist from a controlled comparison of cognitive-behavioural therapy and non-directive counselling.

A number of studies have recently used modern neuropsychological techniques to investigate abnormalities in impulsive personality disorders. In the lead editorial in this issue, Rogers (pp. 1335–1340) reviews this growing field.

This issue features groups of papers on life stress, neuropsychology of psychiatric disorders, substance use disorders.

This issue features groups of papers on deliberate self-harm, neuroimaging, cognitive impairment, treatment of depression in primary care, neuroticism and attachment as aetiological factors. In an additional editorial Tavares et al. (pp. 959–967) review recent studies of cognition in mania and depression.

Deliberate self-harm (DSH) is the most important factor associated with risk of suicide. It also poses a considerable burden on health and social services. Repetition of DSH after hospital presentation is common, with most studies reporting between 12 and 25% of patients repeating within a year and again presenting to hospital (Owens et al. 2002) and many more being involved in repeat episodes which do not bring them to clinical attention (Guthrie et al. 2001). Repetition itself increases the risk of a fatal outcome (Sakinofsky, 2000). The results of both psychosocial and pharmacological interventions to prevent repetition have been fairly disappointing (Hawton et al. 1998, 2000), although the findings of treatment trials have not all been negative. Dialectical behaviour therapy had a substantial initial impact on repetition of self-harm, although this did not persist in a treatment study of female patients with borderline personality disorders (Linehan et al. 1991) and depot flupenthixol reduced repetition in frequent self-harmers (Montgomery, 1987). A recent trial of brief psychotherapy was more effective than treatment as usual (Guthrie et al. 2001) and a meta-analysis of trials of problem-solving therapy suggested beneficial effects for depression, hopelessness and specific problems (Townsend et al. 2001).

BRAIN STIMULATION IN THE TREATMENT OF MAJOR DEPRESSION

Repetitive transcranial magnetic stimulation (rTMS) is an experimental medical procedure that is currently under investigation for its potential therapeutic value in major depression and other psychiatric and neurological disorders (Wassermann & Lisanby, 2001). The idea of using brain stimulation to treat depression dates back to the origins of ECT, and includes more recently developed techniques such as deep brain stimulation and vagus nerve stimulation. The value of brain stimulation in psychiatry is still most clearly seen in the, as yet, unparalleled efficacy of ECT in treating severe depression (American Psychiatric Association, 2001). While ECT is the most effective and most rapidly acting treatment for depression, it also causes a variable degree of undesirable cognitive side effects that limit its clinical utility and prevent many patients who could benefit from receiving this often life-saving treatment (McElhiney et al. 1995; Lisanby et al. 2000b). The search for an effective somatic treatment for medication resistant depression with fewer cognitive side effects than ECT has motivated much of the work with rTMS in psychiatry.

Since the mid-1980s Ian Goodyer and his colleagues have been engaged in trying to understand the dynamics of the relationship between acute and chronic life stresses and depression in adolescents (Goodyer et al. 1985, 1986). One of the great strengths of their work has been their attention to an array of potential risk factors ranging from the apparently exogenous (such as life events, e.g. Goodyer et al. 1988; Goodyer & Altham, 1991; Goodyer, 1999), through those of uncertain origin (e.g. peer and family relationships and problems, Goodyer et al. 1990), to those that can be regarded as being endogenous to the child (e.g. measures of temperament or endocrine function, Goodyer et al. 1993, 1996, 1998; Herbert et al. 1996). They have then typically attempted to determine how subsets of these factors interact in the generation of disorder (see Goodyer, 2002 for a speculative presentation of his current thinking about the relationships between social adversity and depression). A second characteristic of their research lies in its frequent attention to outcomes over time, as opposed to cross-sectional relationships (e.g. Goodyer et al. 1997a,b). Thirdly, in line with growing evidence that the relationships between environmental stress (and perhaps neuro-anatomical correlates) and depression change as repeated episodes occur (Lewinsohn et al. 1999; Daley et al. 2000), they have distinguished among first, recurrent, remitting and chronic depressive episodes. The group's latest work, presented in this issue (Goodyer et al. 2003), nicely illustrates these characteristics, and their extension to consideration of potential dynamic relations among endocrine parameters in relation to psychopathology.

Classifications of depression based on secondary care samples may have little relevance in primary care settings, where many patients do not fulfil criteria for major depression; for example, through having depressive symptoms that are either too mild or too brief. And yet many such patients receive antidepressant treatments (antidepressants and various psychotherapies). At a time when the appropriate recognition and treatment of people with depressive symptoms is the subject of much discussion, and there is concern about the burgeoning number of new psychiatric diagnoses, the burden and hazards of depressive disorders need to be emphasized. The paper by Pezawas et al. (2003) (in this issue) reports the findings of an epidemiological study in adolescents and young adults, and provides a detailed description of the prevalence and features of one such disruptive disorder, namely recurrent brief depression.

Randomized controlled trials (RCTs) of psychological interventions are still much less frequently published than drug trials. To have four papers (Bower et al. 2003; Proudfoot et al. 2003; Simpson et al. 2003; Gabbay et al. 2003) within a single issue of a leading psychiatry journal is indeed unusual. Together these four major papers provide evidence of the benefits and limitations of psychological interventions for common mental disorders.

It is more than 40 years since a burst of studies in the late 1960s started what has become a substantial corpus of work, establishing the role of life events in psychiatric disorders (Brown & Birley, 1968; Paykel et al. 1969). Findings depended on advances in methodology. First, came development of a life events questionnaire and a scaling of their stress magnitude (Holmes & Rahe, 1967); then, replacement by more reliable and valid interview methods and better ways of distinguishing major and minor events (Brown & Harris, 1978; Paykel, 1996).

There has been revived interest within the last few years in neuropsychological investigations of personality disorders (PDs) associated with impulsive and/or anti-social behaviour. Neuropsychological studies of such PDs offer two potential benefits: (i) an improving characterization of the cognitive and emotional difficulties experienced by individuals diagnosed with these disorders and, in particular, borderline PD and anti-social PD; (ii) the development of more precise hypotheses about the role of brain dysfunction in the aetiology and behavioural problems associated with these conditions. The most recent neuropsychological studies of impulsive PD have been interpreted against a background of accumulating evidence for structural (Driessen et al. 2000; Raine et al. 2000; Dolan et al. 2001), neurochemical (Coccaro & Siever, 1995; Steinberg et al. 1997; Leyton et al. 2001) and functional brain abnormalities (Donegan et al. 2003). Future studies of impulsive PDs will be able to exploit advances in structural imaging (e.g. diffusion tensor imaging/tract mapping, e.g. Hoptman et al. 2002), as well as pharmacological and behavioural genetics (McGuffin & Thapar, 1992; Lesch et al. 1996; Skodol et al. 2002).

The two paradigms that now dominate the increasingly active field of psychiatric genetics – genetic epidemiology and gene-finding methods – are well illustrated by five papers in this issue of Psychological Medicine. Genetic epidemiology, which uses the classical ‘work-horse’ methods of family, twin and adoption studies, infers the action of genetic and environmental risk factors by observing the pattern of resemblance of traits or disorders in various classes of relatives. Methods that are utilized in genetic epidemiology can range from great simplicity, such as the odds ratio of a disorder in first-degree relatives of affected versus matched control probands, to substantial complexity, as seen in some advanced multivariate twin-family models (Truett et al. 1994). Of the quasi-experimental methods available to psychiatric geneticists that can tease apart the effects of genetic and shared-environmental factors (nature versus nurture), the twin method has become increasingly popular. This is probably due to the increasing availability of large population-based or volunteer twin registries (see the Oct 2002 issue of Twin Research on ‘Twin Registers as a Global Resource for Genetic Research’) that can be utilized to study psychiatric and drug abuse disorders. By contrast, due to changing social circumstances in most Western countries, adoption, even in the Scandinavian countries where most of the classic studies have been done, is becoming so rare that new adoption studies with younger age cohorts are hardly feasible. By contrast, gene-finding methods, which utilized variants of two different methods of linkage or association, have the goal of determining, on the human genome, the location and potential identity of susceptibility genes.

The inclusion of cognitive symptoms in the DSM-IV criteria for major depressive and manic episodes highlight the importance of cognition in both of these psychiatric disorders (American Psychiatric Association, 1994). For example, criteria for diagnosis of these conditions include a diminished ability to concentrate and indecisiveness. In addition, numerous studies have demonstrated wide-ranging cognitive deficits in depression (for example Elliott et al. 1996; Purcell et al. 1997; Murphy et al. 2003) and mania (Goldberg et al. 1993; Murphy et al. 1999, 2001; Sweeney et al. 2000). These include deficits in early information processing (Tsourtos et al. 2002), recollection memory (MacQueen et al. 2002) and planning (Elliott et al. 1996) as well as affective biases (Murphy et al. 1999) and abnormal responses to negative feedback (Elliott et al. 1996, 1997; Murphy et al. 2003). Some residual deficits are also evident in a proportion of remitted subjects, even when controlling for mood (for example, Clark et al. 2002).

Background. Childhood adversity significantly increases the risk of depression, but it is unclear whether this risk is most pronounced for depression occurring early in life. In the present study, we examine whether three aspects of childhood adversity – low socio-economic status (SES), family disruption, and residential instability – are related to increased risk of depression during specific stages of the life course. We also examine whether these aspects of childhood adversity are related to the severity of depression.

Method. A sample of 1089 of the 4140 births enrolled in the Providence, Rhode Island cohort of the National Collaborative Perinatal Project was interviewed between the ages of 18 and 39. Measures of parental SES, childhood family disruption and residential instability were obtained upon mother's enrolment and at age 7. Age at onset of major depressive episode, lifetime number of depressive episodes, and age at last episode were ascertained via structured diagnostic interviews. Survival analysis was used to identify risk factors for depression onset and remission and Poisson regression was used to model the recurrence rate of depressive episodes.

Results. Low parental SES, family disruption and a high level of residential instability, defined as three or more family moves, were related to elevated lifetime risks of depression; the effects of family disruption and residential instability were most pronounced on depression onset by age 14. Childhood adversity was also related to increased risk of recurrence and reduced likelihood of remission.

Conclusions. Childhood social disadvantage significantly influences risk of depression onset both in childhood and in adulthood. Early childhood adversity is also related to poor prognosis.

The problems of understanding what causes chronic fatigue, its relationship to psychiatric disorders, and how best to treat it continue to baffle and bemuse clinicians and patients alike (Report to the CMO, 2002; Stanley et al. 2002). This edition of the journal carries four original papers that provide much needed knowledge in a subject riddled by polemic and prejudice (White, 2002).

Background. Studies that assess the rates of psychotic patients who fail to adhere to treatment programmes have generated heterogeneous results, with estimates ranging from 24 to 90%. This paper presents findings from a systematic review on adherence to treatment by patients with psychosis. Its purpose is to provide an overall estimate of treatment non-adherence in community psychiatric services, and to analyse and study patient characteristics explaining the between-study heterogeneity in rates of non-adherence.

Method. A systematic review of published studies that report rates of non-adherence with medication and scheduled appointments by psychotic patients in community settings has been undertaken.

Results. A total of 103 studies were included in this systematic review. Eighty-six of these studies were suitable for data re-analysis. The overall weighted mean rate of non-adherence, calculated in a sample of 23796 patients, was 25·78%. A linear regression analysis of non-adherence rates on background characteristics showed that sample size was negatively associated with non-adherence rates, while first-contact cases and low-adherence cases, in comparison with ongoing cases, were associated with higher non-adherence rates. Factors associated with poor compliance included: lack of insight; positive symptoms; younger age; male gender; history of substance abuse; unemployment; and low social functioning.

Conclusions. Approximately one in four patients with psychosis fails to adhere with treatment programmes. Preventive evidence-based clinical interventions should be routinely implemented in community settings to reduce patient non-adherence.