ABSTRACTObjective This study examines Child Behavior Checklist 1.5-5 (CBCL) validity for externalizing, internalizing, and total problems T-scores when screening for clinically significant behavioral problems in a Navajo (Diné) sample, that was not included in CBCL norming. Method Mothers were recruited during pregnancy through the Navajo Birth Cohort Study/Environmental influences on Child Health Outcomes (NBCS/ECHO). Families completed in-person, clinical assessments when children were 42 to 72 months-old. Caregivers completed the CBCL and provided histories. Best estimate clinical diagnoses were established by experienced clinician-researchers using records, along with comprehensive neuropsychological, psychiatric, and physical examinations. Results Caregivers completed CBCLs for 192 children (98 boys, 51.0%; mean age 4.0±0.61, range 3 to 6 years). 125 children (65.1%) received at least one diagnosis, with language disorder being the most common, followed by speech sound disorder. Every group with a Diagnostic and Statistical Manual of Mental Disorders-5 (DSM-5) diagnosis had T-scores within the normal range when they were derived according to the CBCL manual. T-scores obtained by normalization within our study population yielded higher scores for the diagnosed group; however, Area Under Curve (AUC) values still ranged from 0.6 to 0.7, displaying poor CBCL discriminative power. Externalizing problems T-scores for those with any DSM-5 diagnosis were most discriminative (AUC = 0.70). Conclusion The CBCL was an ineffective single screening tool in the Diné population compared to in-person clinician evaluations, possibly because the Diné were not included in the CBCL validation or cross-cultural adaptations. Our findings may be relevant to other minority populations not included in norming.
Cho et al. (Fri,) studied this question.