Determining the patient's best interests is an underpinning principle of capacity legislation (Brindle & Branton, pp. 430–437, Box 2) and surely of all clinical practice. It is perhaps nowhere more controversial than in the use of medication to treat young people with mental health problems – described as a ‘blunt’ approach and ‘bad medicine’ by Des Reference SpenceSpence (2010), a practising GP and regular columnist in the BMJ. A more measured JAMA editorial (Reference Varley and McClellanVarley 2009) discusses some of the concerns: the substantial increase in the use of antipsychotics in young people (including children younger than 5 diagnosed with bipolar disorder); diagnostic problems, including the overlap of bipolar disorder with behaviour disorders and the possible discontinuity between child and adolescent and adult forms of bipolar disorder; the relative prioritising of medications and psychosocial interventions; the long-term health implications of prolonged treatment with second-generation antipsychotics (not least obesity); and the extent to which the evidence base is formed by industry-sponsored investigations. It is clear, then, that determining appropriate treatment and the best interests of a young person with an affective disorder is unlikely to be straightforward.

Dubicka et al (pp. 402–412) consider prescribing by examining and reflecting critically on the evidence without shying away from the controversial issues. Accurate diagnosis, necessary for both good prescribing (British Pharmacological Society 2010) and steering clear of causing avoidable harm, is problematic. So, too, is the evidence base, which is sparse and at times depends on extrapolating from the adult literature. Even randomised controlled trials, often referred to as the gold standard (e.g. Coia, pp. 474–475), are not necessarily so and are ‘particularly weak in relationship to generalisability and most especially in the assessment of harms’ (Reference RawlinsRawlins 2008). Importantly, Dubicka et al do not consider pharmacological treatment as a standalone intervention but always in addition to ‘specialised treatment as usual’ (STAU). Distinct from specialised psychological treatments such as cognitive–behavioural therapy and interpersonal therapy, STAU comprises no fewer than 16 elements (Box 1, p. 404) and describes the complex and multifaceted quality of ‘ordinary’ multidisciplinary and multi-agency clinical practice.