A22 22..22

Three papers in the Journal this month consider issues related to measuring, understanding and treating mental ill health in childhood. Baron-Cohen et al (pp. 500–509) undertook a study to establish the prevalence of autism-spectrum conditions in Cambridgeshire using two case ascertainment methods. First, a survey of individuals on the Special Educational Needs register and second, a screening of the mainstream primary school population. Estimated prevalence was found to be approximately 1% based on either method, in keeping with previous studies. On the basis of a formula devised to combine estimates, the authors report a prevalence of 157 per 10 000, including previously undiagnosed cases. In light of an increasingly emphasised need for routine outcome monitoring to be undertaken by clinical services and the problem of assessing the benefits of interventions in the absence of a control group, Ford et al (pp. 552–558) developed an algorithm for producing a Strength and Difficulties Questionnaire (SDQ) Added Value Score. Using data from a longitudinal community study, the authors derived the algorithm and then tested it using data from the Welsh Sure Start randomised controlled trial (RCT). The SDQ Added Value Score acted as predicted when applied to the RCT data and consequently the authors conclude that their work provides preliminary support for the use of such a tool in the assessment of the outcome of interventions provided by child mental health services. Utilising data from the ALSPAC cohort, Thomas et al (pp. 521– 526) found that fetal growth measures were associated with risk of psychosis-like symptoms at 12 years but measures of childhood growth were not. Controlling for IQ only partially accounted for the association between birthweight and psychosis-like symptoms.