Abstract
The diagnosis of bipolar disorder (BD) in young children has been a topic of debate, in part owing to varied interpretation of manic-like symptoms. We examined how expert academic clinicians participating in the pediatric bipolar biobank varied in their interpretation and application of Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria and diagnoses. Study co-investigators reviewed 12 standardized narratives and for each marked a visual analog scale with their confidence in the presence of manic episodes and criteria. We analyzed raters’ confidence and inter-rater agreement using interclass correlation (ICC). Symptoms with good ICC ranging from 0.60 to 0.74 included inflated self-esteem/grandiosity and decreased need for sleep. Diagnoses and episodes with poor ICCs (< 0.4) included Hypomania and Bipolar Not Elsewhere Classified/Not Otherwise Specified. Despite efforts made to refine BD criteria with DSM-5, there was substantive variation in diagnostic interpretation among investigators working with children presenting with manic-like symptoms.