ASEBA research

A group of Dutch researchers tested the predictive value of parent, teacher, and clinician ratings of child psychopathology over a 3-year period (Ferdinand et al., 2003). Parent and teacher ratings were obtained with the CBCL and TRF. Clinician ratings were obtained with the Semi-structured Clinical Interview for Children and Adolescents (SCICA). Ninety six children referred for mental health services were initially evaluated at ages 6 to 12 and were evaluated again after a mean interval of 3.2 years. The initial CBCL, TRF and SCICA scale scores were tested as predictors of the following measures of poor outcomes: school problems, police/judicial contacts, parent's wish for help, need for outpatient treatment, and need for inpatient treatment. The cross-informant correlations of SCICA scores with CBCL and TRF scores were low to moderate, as is typically found for assessment of psychopathology. All measures of poor child outcomes were significantly predicted by CBCL, TRF, and SCICA scores. On the whole, CBCL and TRF scores tended to be better predictors of poor outcomes than clinicians' ratings. Nevertheless, after all other scores were partialed out, the SCICA Attention Problems scale, based solely on clinicians' observations during the interview, significantly predicted need for inpatient treatment, which is an especially compelling sign of poor outcomes. This was evidently the first study to systematically compare the predictive value of parent, teacher, and clinician ratings of poor outcomes for children referred for mental health services. The findings suggest that the predictive power of clinical assessment can be enhanced by the integration of reports provided by multiple informants. Exclusive reliance on clinicians' ratings of child psychopathology may limit the predictive power of clinical assessment. Nevertheless, clinician ratings can add important predictive power to parent and teacher ratings.