Data concerning appropriate treatment in panic disorder following an initial response to acute therapy are limited.

To assess the safety and efficacy of continued fluoxetine treatment following successful acute therapy of panic disorder.

Patients who responded to acute fluoxetine treatment were randomised to 24 weeks of continued fluoxetine or placebo.

Fluoxetine responders randomised to continue on their acute-phase fluoxetine dose experienced statistically significant improvement in panic attack frequency and phobia rating scale score over 24 weeks of therapy, while those switched to placebo experienced statistically significant worsening in Hamilton Anxiety (HAM–A), Hamilton Depression (HAM–D) and SCL–90–R rating scores.

Fluoxetine was associated with improved clinical outcomes compared with placebo during continuation therapy.

Family interventions are effective in reducing relapse in patients with schizophrenia, but there is little work demonstrating the effectiveness of the interventions in routine service settings.

To test the effectiveness of a needs-based family intervention service for patients recruited as out-patients and their carers, including those of low expressed emotion status.

Carers of out-patient schizophrenia sufferers selected only on illness history factors were randomly allocated to receive either family support alone or in combination with systematic psychosocial interventions based on an assessment of need. Delivery of family interventions attempted to involve the clinical team.

Relapse outcomes were superior for family-treated patients at six-month follow-up, although most of the clinical and symptom patient variables assessed remained stable, as did measures of carer burden.

The study demonstrated the effectiveness of family interventions in routine service settings. Problems with staff, patient and carer engagement and participation were identified.

The EURODEP collaboration was formed to take advantage of existing studies of random community samples of older people in Europe, using GMS–AGECAT for case identification and diagnosis. Later, other centres joined, and the EURO–D scale was developed to harmonise the different methods used with the GMS. Previous studies had revealed different levels of depression in Europe but had been confounded by the use of unreconcilable methods. These studies attempt to overcome this problem.

To introduce the first set of publications from the EURODEP collaboration.

Presented in five accompanying papers (pp. 307–345, this issue)

Developmental impairments have been identified as a risk factor for early-onset schizophrenia. Affective symptoms are more common in children and adolescents with disordered neurodevelopment than in healthy controls.

To test the hypothesis that severe early-onset mood disorders are associated with developmental antecedents.

We retrospectively identified 38 adolescent cases (15 female, 23 male; mean age 14.4 years, range 11–18) who met ICD–10 Research Diagnostic Criteria for a manic episode, bipolar affective disorder or psychotic depression, and 41 controls (25 female, 16 male, mean age 14.2 years, range 11–18) with depression but without psychotic features.

Cases were significantly more likely to have experienced delayed language, social or motor development (OR 5.5, 95% CI=1.4–21.6, P=0.01). in particular those who develop psychotic symptoms (OR 7.2, 95% CI=1.8–28.6, P=0.003).

Compared to early-onset unipolar depression, neurodevelopmental antecedents are over-represented in early-onset bipolar disorder. The validity of this finding was supported by contemporaneous IQ scores that are not subject to the same potential biases as case-note ratings.

The HoNOS has been developed as a routine measure of outcomes in mental health.

To explore the validity and interrater reliability of HoNOS in a first-onset psychosis follow-up study.

Between 1992 and 1994 we ascertained a cohort of all persons with first-onset psychosis. We re-assessed these people at 3 years (n=166) with several outcome scales, including HoNOS. Patients' keyworkers also completed the HoNOS. We estimated concurrent validity by calculating correlations between HoNOS and other scales, and interrater reliability.

Researcher HoNOS correlated highly with other scales (0.46 < p < 0.86; P < 0.001). Keyworker HoNOS correlations were lower (0.41 < p < 0.51; P < 0.05), but still significant for all scores except the HoNOS-social subscale (0.12 < p < 0.28). Agreements between researcher and keyworker HoNOS were modest (0.47 < ICC < 0.85).

In this research cohort HoNOS correlates well with established outcome scales. Keyworker ratings show similar, but weaker, relationships; its use in routine settings may require further training for calibration of severity.

Tardive dyskinesia is important in the side-effect profile of antipsychotic medication.

The development of tardive dyskinesia was evaluated in patients treated with double-blind, randomly assigned olanzapine or haloperidol for up to 2.6 years.

Tardive dyskinesia was assessed by the Abnormal Involuntary Movement Scale (AIMS) and Research Diagnostic Criteria for Tardive Dyskinesia (RD-TD); it was defined as meeting RD-TD criteria at two consecutive assessments. The risk of tardive dyskinesia, the relative risk, incidence rate, and incidence rate ratio were estimated.

The relative risk of tardive dyskinesia for the overall follow-up period for haloperidol (n=522) v. olanzapine (n=1192) was 2.66 (95% CI=1.50–4.70). Based on data following the initial six weeks of observation (during which patients underwent medication change and AIMS assessments as frequently as every three days), the one-year risk was 0.52% with olanzapine (n=513) and 7.45% with haloperidol (n=114). The relative risk throughout this follow-up period was 11.37 (95% Cl=2.21–58.60).

Our results indicated a significantly lower risk of tardive dyskinesia with olanzapine than with haloperidol.

Two recently developed measures intended for clinical use are HoNOS (Health of the Nation Outcome Scales), measuring social functioning, and CANSAS (Camberwell Assessment of Need Short Appraisal Schedule), for measuring need.

We investigated the association between CANSAS and HoNOS.

For some domains there were substantial associations, with high HoNOS rating correlated with CANSAS rating of the presence of a need. For other areas the agreement was less than might be expected. Seven factors were identified within the 22 CANSAS domains, and the presence of needs in the Drug/alcohol and Activities of daily living factors was associated with high scores in the related HoNOS domains.

CANSAS and HoNOS assessments differ. HoNOS can track changes in social functioning over time, but may be less useful for treatment planning, and should not be used to infer the level of morbidity in a case-load. CANSAS indicates when treatment should be commenced or continued. It can also be used as a case-load measure, but may be insufficiently sensitive to be used as an outcome measure at the individual level.

The burdens experienced by relatives of mentally ill persons are substantial.

To study the relationship between family burden and sociocultural context.

A comparative study of Euro-Americans and Latinos ascertained whether dimensions of family response are (a) non-specific to diagnostic groups; and/or (b) variable across cultural settings.

Regardless of diagnosis or ethnicity, patient misery was found most burdensome and distressing. However, considerable difference in shades of meaning and nuance across groups appears in relation to what is classed similarly as ‘misery’. Only gender was significantly associated with social performance (males reported to have greater deficits). A complex cultural – ecological effect was observed among the Latino–schizophrenia group.

Findings suggest similarities and differences in levels of family burden in relation to socio-cultural factors across cultural and diagnostic groups. The specificity of results by objective and subjective measures, types of burden, gender, ethnicity, diagnosis, and living situation confirm the importance of context and heterogeneity in understanding family burden and distress.

Despite considerable interest, there is no consensus regarding the prevalence of depression in later life.

To assess the prevalence of late-life depression in the community.

A systematic review of community-based studies of the prevalence of depression in later life (55+). Literature was analysed by level of caseness at which depression was defined and measured.

Thirty-four studies eligible for inclusion were found. The reported prevalence rates vary enormously (0.4–35%). Arranged according to level of caseness, major depression is relatively rare among the elderly (weighted average prevalence 1.8%), minor depression is more common (weighted average prevalence 9.8%), while all depressive syndromes deemed clinically relevant yield an average prevalence of 13.5%. There is consistent evidence for higher prevalence rates for women and among older people living under adverse socio-economic circumstances.

Depression is common in later life. Methodological differences between studies preclude firm conclusions about cross-cultural and geographical variation. Improving the comparability of epidemiological research constitutes an important step forward.

There have been many studies reporting reduced volume of the hippocampus or other limbic structures in patients with schizophrenia, but the literature is inconsistent.

To compare patients with either first-episode or chronic schizophrenia with controls using high-resolution volumetric magnetic resonance imaging (MRI) scans.

Thirteen patients with first-episode schizophrenia, 27 with chronic schizophrenia and 31 controls had 1.5 mm coronal slices taken through the whole brain using a spoiled-grass MRI acquisition protocol.

The parahippocampal gyrus was reduced significantly on the left side in patients with chronic schizophrenia compared with controls for both male and female patients, whereas the hippocampus was reduced significantly on both sides only in female patients. There were no significant reductions in any structure between patients with first-episode schizophrenia and controls.

Volumetric reduction seen in patients with chronic schizophrenia may be due to an active degenerative process occurring after the onset of illness.

Earlier work on the protective role of social support in onset and course of depressive disorder suggested that its provision might be relevant for outcome.

To evaluate volunteer befriending as an intervention among women with chronic depression in inner London.

A randomised controlled trial, with awaiting list control design, with outcome measured as Present State Examination (PSE)-based full or partial remission after one year.

A statistically significant effect upon remission was found for befriending. Other treatments monitored naturalistically did not relate to remission, nor did initial duration of chronic episode or comorbidity, but there was an association with initial PSE score. The findings are discussed in relation to standard drug-trial results and to another befriending intervention with the elderly.

The results are not unpromising. Additional trials with less restricted intake conditions, and in more naturalistic general practice settings, might confirm volunteer befriending as a useful adjunct to current treatments.

The adequacy of pharmacotherapy received in practice by patients after an acute episode of depression has been little studied.

To describe and assess adequacy of drug continuation and maintenance in patients with depression.

Patients with depression were interviewed 18 months after discharge from hospital. Quantitative assessments of drug treatment doses and compliance were made monthly over this period, and qualitative ratings in continuation and maintenance phases.

About 20% of patients were prescribed low drug doses after discharge and 10% were prescribed no drugs at all. Reported compliance was around 70%. About 30% failed to receive adequate longer-term treatment, mostly due to the continuation phase being too short. Deficiencies of dosage and compliance were greater in patients who never achieved full recovery. Patient refusal was the most common reason for not using antidepressants. Further episodes of depression were not particularly associated with inadequate treatment.

There were deficiencies in drug treatment that did not appear to be the principal cause of further episodes but may be important in non-recovery. Patient fears require discussion.

Alzheimer's disease and other types of dementia are characterised by numerous psychiatric and behavioural changes. Little is known of their natural history.

To investigate the sequence and pattern of these changes throughout the course of dementia.

One hundred people, initially living at home with carers, entered a prospective, longitudinal study. At four-monthly intervals, behavioural and psychiatric symptoms were assessed using the Present Behavioural Examination and Mini-Mental State Examination. Follow-up continued for up to nine years (mean 3.3 years; s.d. 2.4). Patterns of onset and disappearance of these symptoms, their sequence and association with time of death and cognitive decline were analysed. Autopsy confirmed a diagnosis of pure Alzheimer's disease in 48 subjects. Data for this subgroup are presented.

Some changes tend to occur earlier than others but changes can occur at almost any time in the course of dementia.

The natural history of behaviour changes in Alzheimer's disease shows great individual variation although some changes tend to follow a recognisable sequence.

This is the first report of results from the EURODEP Programme.

To assess the prevalence of depression judged suitable for intervention in randomised samples of those aged ⩾65 in nine European centres.

The GMS-AGECAT package.

Differences in prevalence are apparent, 8.8% (Iceland) to 23.6% (Munich). When sub-cases and cases are added together, five high- and four low-scoring centres emerge. Women predominated over men. Proportions of sub-cases to cases revealed striking differences but did not explain prevalence. There was no constant association between prevalence and age. A meta-analysis (n=13 808) gave an overall prevalence of 12. 3%, 14.1% for women and 8.6% for men.

Considerable variation occurs in the levels of depression across Europe, the cause for which is not immediately obvious. Case and sub-case levels taken together show greater variability, suggesting that it is not a matter of case/sub-case selection criteria, which were standardised by computer. Substantial levels of depression are shown but 62–82% of persons had no depressive level. Opportunities for treatment exist.