Predicting change in psychopathology in youth referred to mental health services in childhood or adolescence.

BACKGROUND Little evidence is available on factors associated with persistence and change of psychopathology, and little is known about the predictive value of factors regarding change once problem behaviours exist. This study aims to evaluate change in level of scores of empirically derived problem patterns and to study factors that influence this change for children and adolescents referred to mental health services. METHOD A referred sample (N = 1,652), aged 4 to 18 years at initial assessment, was followed up after a mean interval of 6.2 years. We used standardised information from parents, teachers and subjects, including the CBCL, YSR and TRF at both assessments. RESULTS Subjects at follow-up scored significantly above the expected mean norm scores, although for most scores improvement was found. The strongest predicting factor for time 2 psychopathology was the corresponding time 1 score, odds ratios ranging from 1.6 to 21.7. Males and children older at intake improved more than females and younger children, respectively. CONCLUSIONS Few child, family and treatment-related factors had additional predictive value over and above earlier psychopathology, and their contribution to the prediction of outcome was small. Findings indicate continuity of behavioural and emotional problems in clinically referred children and adolescents, and these problems should be viewed as chronic conditions. Girls referred for behavioural and emotional problems may form a group especially at risk for poor outcome.