There is increasing evidence-based knowledge in the drug treatment of psychotic and behavioural symptoms in dementia, but drugs do not possess a formal licence for these indications. Drug companies, health authorities, NHS trusts and medical defence unions were asked for their advice on the medico-legal implications for the prescribing clinician.

Drug companies, health authorities, medical defence unions and NHS trusts are aware of out-of-licence prescribing and leave ultimate accountability with the clinician. A suggested best practice is that of obtaining the patient's consent.

Out-of-licence prescribing for psychotic and behavioural symptoms in dementia is widespread. This patient group may be unable to grant consent. The accountability of individual clinicians should be supported by more adequate medico-legal frameworks.

To examine service-level variables predicting treatment adherence in a specialist eating disorder unit. We analysed a sample of 157 individuals consecutively referred to the unit over an 18-month period. Associations were determined using odds ratios.

Individuals with a formal care programme at the point of referral were more likely to stay in treatment. Treatment adherence was not predicted by illness severity or waiting time. Follow-up by a dietician and acceptance of referral to a support group predicted better treatment outcomes.

Although the standard care programme approach may be relinquished in the UK, we recommend that this approach or its equivalent be used in specialist eating disorder services to improve treatment adherence.

In this paper, we summarise and critique a network meta-analysis (NMA) of antidepressant efficacy and tolerability for paediatric depression and an accompanying editorial. Although we agree that many of the extant studies are flawed, this meta-analysis showed clear efficacy of fluoxetine in the NMA, and for sertraline and escitalopram in pairwise analyses. Consequently, these papers underestimate the benefits of antidepressants for paediatric depression, and provide support for current practice guideline, which recommends the use of an antidepressant if the patient does not respond to psychotherapy. In these circumstances, fluoxetine should be the first choice, with escitalopram and sertraline as alternatives.

D.A.B. receives royalties from Guilford Press, has or will receive royalties from the electronic self-rated version of the C-SSRS from eResearch Technology, Inc., is on the editorial board of UpToDate, and is a reviewer for Healthwise. R.D.G. serves as an expert witness for the US Department of Justice, Pfizer, Wyeth and GSK; and is the founder of Adaptive Testing Technologies. P.W. receives personal fees from Lundbeck and Takeda. B.D. reports a licensing agreement with Lundbeck for a psychosocial treatment manual for depression. No other disclosures were reported.

We conducted the first systematic territory-wide survey of electroconvulsive therapy (ECT) practice and utilisation in Hong Kong. All public ECT services were evaluated with a postal questionnaire, a visit and observations of ECT administration. All private ECT practitioners were surveyed using a postal questionnaire.

Most practice areas at the public ECT services in Hong Kong complied with College guidelines. However, a few private practitioners were still using outdated ECT machines. The number of ECT courses completed in 1998 was 0.34 per 10 000 population, which was much lower than the rate of ECT use in Western countries.

We obtained basic information on ECT services in Hong Kong and on areas requiring immediate attention.

A cross-sectional survey was conducted to assess patient knowledge and information provision about opioid substitution treatment among individuals with opiate dependence receiving treatment at four treatment centres in South London.

In total 118 people were recruited to the study. Participants answered a mean of 14 out of 34 questions assessing a range of factors such as medication, blood-borne viruses and overdose correctly. Participants overestimated their performance on average by almost 40%. Individuals with a history of previous treatments scored significantly higher than those in their first treatment episode. The majority reported having been given written information on most of the topics assessed.

The results of this study highlight the need to improve education about opioid dependence and its treatment. Poorly informed patients are unlikely to make optimal treatment choices. Improving patients' knowledge and understanding about treatment may lead to better engagement, retention, treatment adherence and, ultimately, better health outcomes.

A cross-sectional survey of the use of the Mental Health (Scotland) Act 1984 in a defined urban area. Patients initially detained under civil sections (Sections 24, 25, 26 and 18) between 1 April 1997 and 31 March 1998 were identified using the hospital information system and a hand search of section papers.

There were 283 detentions involving 204 patients that lasted a median of 6 days. A total of 98% of patients were initially detained on a 72-hour ‘emergency section’. A total of 61% had non-organic psychotic disorders (172/283). Less than half of detentions were during the working week. Consent was usually provided by the mental health officer or relatives, but was not provided for 11% of detentions. Patients detained after admission were more likely to be detained for a longer period (29 v. 3 days) and to have their detention extended over 72 hours (64% v. 41%) compared with those detained in the community.

This paper provides information on some of the gaps identified by recent reviews of mental health legislation in Scotland and discusses the possible impact of the changes proposed by the Millan Committee. The workload of general adult consultant psychiatrists is likely to increase and the proportion of patients detained without consent could also increase. The study supports the differentiation of patients detained after admission from those detained in the community, as the patterns of detention are different.