Comparing the validity of clinician-generated diagnosis of conduct disorder to the diagnostic interview schedule for children.

Clinician diagnoses of conduct disorder (CD) were compared to the diagnoses of CD generated by a structured interview against an observed criterion. Participants were 534 youth from a large residential program in the Midwest for delinquent youth. Rates of in-program CD behaviors were gathered from staff observations of the youth over a 9-month time period. Youth diagnosed with CD by the Diagnostic Interview Schedule for Children (DISC) displayed significantly more CD behaviors in the first 6 months of treatment compared to both youth without an externalizing disorder and youth diagnosed with CD by a clinician. Youth diagnosed with CD by a clinician had rates of CD identical to youth without an externalizing disorder. Clinicians may have weighted contextual information more heavily, as this group was significantly more likely to have an arrest record. Results support the use of structured interviews and provide evidence that typical clinician diagnoses may lack adequate validity.