The prediction of disruptive behaviour disorders in an urban community sample: the contribution of person-centred analyses

Abstract

Background: Variable- and person-centred analyses were used to examine prediction of middle childhood behaviour problems from earlier child and family measures.

Method: A community sample of 164 families, initially recruited at antenatal clinics at two South London practices, was assessed for children's behaviour problems and cognitive ability, maternal mental health, and the family environment when the children were 4 years old. At age 11, children, mothers, and teachers reported the child's disruptive behaviour, and mothers and children were interviewed to identify cases of disruptive behaviour disorders (DBD).

Results: Neither social class nor ethnicity predicted the child's disruptive behaviour at age 11. Rather, path analyses and logistic regression analyses drew attention to early behavioural problems, maternal mental health and the child's cognitive ability at 4 as predictors of disruptive behaviour at age 11. Cluster analysis extended these findings by identifying two distinct pathways to disruptive symptoms and disorder. In one subgroup children who showed intellectual difficulties at 4 had become disruptive by 11. In a second subgroup mothers and children both showed psychological problems when the child was 4, and the children were disruptive at age 11. The person-centred approach also revealed a high-functioning group of cognitively able 4-year-olds in supportive environments, at especially low risk for DBD.

Conclusions: Combining variable- and person-centred analytic approaches can aid prediction of children's problems, draw attention to pertinent developmental pathways, and help integrate data from multiple informants.