Subtyping CD using CU traits identifies children with more severe and persistent psychopathology. Children with high CU traits but no CD diagnosis require further investigation.

Inpatient treatment is associated with substantive sustained health gain across a range of diagnoses. Lack of intensive outpatient-treatment alternatives limits any unqualified inference about causal effects, but the rigour of measurement here gives the strongest indication to date of the positive impact of admission for complex mental health problems in young people.

Previous twin and family studies have indicated that there are strong genetic influences in the etiology of autism, and provide support for the notion of a broader phenotype in first-degree relatives. The present study explored this phenotype in terms of one current cognitive theory of autism. Parents and brothers of boys with autism, boys with dyslexia, and normal boys were given tests of "central coherence", on which children with autism perform unusually well due to an information-processing bias favouring part/detail processing over processing of wholes/meaning. Results indicated that fathers of boys with autism, as a group, showed piecemeal processing across four tests of central coherence. This was not true for any other group. These findings raise the possibility that the broader autism phenotype may include a "cognitive style" (weak central coherence) that can confer information-processing advantages.

Adolescents' representational models of attachment are related to but distinct from current parenting quality and provide unique insight into the understanding of behavioural adjustment. The findings support a distinct conceptual role of attachment representations in adolescence. Clinical assessment and treatment models should include attachment patterns in this age group.