Concurrent Validity of the CBCL DSM-Oriented Scales: Evidence from the ABCD Study.

Abstract

ObjectivePrevious research has yielded mixed findings regarding the validity of the Child Behavior Checklist (CBCL) DSM-Oriented Scales (DOSs) in identifying DSM diagnoses. The present study evaluates the concurrent validity of the DOSs by comparing scale classifications with a gold-standard clinical interview.MethodsParticipants (N = 11,851) were children in the NIH Adolescent Brain Cognitive Development (ABCD) study whose caregivers completed the CBCL and the Kiddie Schedule for Affective Disorders and Schizophrenia (KSADS) semi-structured interview at baseline (9.92 years, SD = 0.63), two-year follow-up (12.03 years, SD = 0.67), or both.Results2,321 (19.59%) children at baseline and 1,708 (15.57%) participants at two-year follow-up met criteria for one or more KSADS disorders. Sensitivity of the CBCL relative to KSADS diagnoses was generally low (10.36% - 75.00%), with only two scales above 50%. Specificity was high (92.47% - 97.07%). Positive predictive values were low (0.71% - 48.20%), and negative predictive values were high (89.84% - 99.90%). Area under the curve (AUC) ranged from 0.55 to 0.84.DiscussionIn the present sample, the DOSs showed poor concurrent validity with a gold standard clinical interview. Caution should be exercised when using the CBCL due to the high rates of false negatives.