Emotional and behavioural outcomes later in childhood and adolescence for children with specific language impairments: meta-analyses of controlled prospective studies

Background: Prospective evidence on psychological outcomes for children with specific language impairments (SLI) is accumulating. To date, there has been no attempt to summarise what this evidence says about the strength of link between SLI and later child and adolescent emotional and behavioural (EB) outcomes. Methods: We undertook a systematic review and meta-analysis (following PRISMA guidelines and involving a literature search to June 2012 of seven databases, including MEDLINE and PsychAPA) of prospective, cohort studies of children with SLI and typical language development (TLD) reporting on the incidence and severity of EB problems later in childhood or adolescence. Results: Nineteen follow-up reports of eight cohorts with 553 SLI children and 1533 TLD controls were identified. Initial assessment was at 38.8years of age and follow-up duration from 2 to 12years. Pooled across comparable studies, SLI children were about two times more likely to show disorder levels of overall internalising problems, overall externalising and ADHD problems than TLD children. Compared with the average TLD child (50 percentile), at follow-up, the symptom severity of the average SLI child was at the 72 percentile (95% CI 6579 percentile) on internalising symptoms, the 69 percentile (95% CI 6374 percentile) on externalising symptoms and the 60 percentile (95% CI 5268 percentile) on AHDH severity. The findings about risk to specific mental disorders and the severity of specific problems were inconclusive. Conclusions: Relative to typical children, SLI children experience clinically important increases in the severity of diverse emotional, behavioural and ADHD symptoms and more frequently show a clinical level of these problems. The small number of studies included in pooled analysis and methodological heterogeneity reduce the precision and generalisability of the findings. Most studies do not account for initial levels of EB problems.