Race, gender and age effects on the assessment of bipolar disorder in youth

Abstract

Background: Little is known about racial, age or gender disparities in the assessment and diagnosis of pediatric bipolar disorder in children and adolescents. Objectives: To assess the psychometric properties of five commonly-used screening instruments in detecting bipolar disorder in youth. Design: A secondary analysis will be used to determine if a test item bias (or differential item functioning) exists within each of these instruments. Data from Dr. Eric Youngstrom's existing database of 758 African American and White families, with 303 female and 456 male subjects) will be used. Results: Measurement variance was found in the instruments between different racial and age groups and between genders. As seen in the results section, DIF was detected on specific items, indicating that the probability of a parent endorsing that item differs between equal groups due to group membership, rather than the construct (mania), which it is intended to measure. However, most differences were small and unlikely to be clinically-meaningful or influence diagnostic or treatment decisions in research or clinical practice. Conclusions: Further research using advanced methodology is needed to better understand how these instruments function between different racial and age groups and between genders and to better understand how the differences will affect research and clinical practice. Keywords: pediatric bipolar disorder, assessment, diagnosis, race, age, gender, screening instruments, rating scales and disparities.