Discriminative validity of the general behavior inventory using youth report

The present study investigated the ability of the General Behavior Inventory (GBI) to discriminate between diagnostic groups using youth self-report. One hundred and ninety-seven youths ages 10–17 years presenting at a midwestern urban outpatient clinic specializing in mood disorders completed the GBI as part of the intake process. Diagnoses were determined by a structured clinical interview (K-SADS) administered by either a child and adolescent psychiatrist or a research assistant trained to a high level of interrater reliability ( > .85). Games–Howell post hoc tests showed that the diagnostic groups significantly differed on the GBI's 2 subscales, Depression and Hypomanic–Biphasic. Logistic regression demonstrated that the scales discriminated between bipolar and disruptive behavior disorders, unipolar and bipolar depression, and mood and disruptive behavior disorders or no diagnosis. Receiver Operating Characteristic (ROC) curves further indicated the good diagnostic efficiency of the scales. Results indicate that the GBI's subscales might aid in making traditionally difficult differential diagnoses, such as between bipolar disorder and Attention Deficit Hyperactivity Disorder (ADHD) and between unipolar and bipolar depression.