Assessment of Adolescent Mental Health and Behavioral Problems in Residential Care: Discrepancies Between Staff-Reported CBCL Scores and Adolescent-Reported Ysr Scores

Purpose:  Throughout the world, more than 2 million children are estimated to reside in residential care. For these children, staff act as primary caregivers and often provide assessment of child psychopathology. The Youth Self-Report (YSR) and the Child Behavior Checklist (CBCL) are valid and reliable measures frequently used to ascertain emotional and behavioral problems experienced by youth.  However, minimal research exists on how and when residential staff are best positioned to report on mental health.  This research examined discrepancies between adolescent- and residential staff-reported psychopathology and whether variation between raters may be explained by demographic or relationship factors.

Methods: Interviews were conducted with 60 Jordanian adolescents (age 11-17) in residential care and their residential staff-caregivers. Two types of discrepancy scores were derived using data from the Youth Self-Report (YSR) and Child Behavior Checklist (CBCL) to demonstrate a fuller picture of how raters view both the severity and number of problems endorsed. First, problem severity difference scores were based on the summed raw scores for individual items in each scale.  Second, we recalculated the scales based on item endorsement to create a dichotomized score and then derived problem frequency difference scores. Discrepancies were examined using visual analysis, Pearson-correlations, and Kappa coefficients; associations with participant characteristics and relationship factors were examined using t-tests and ANOVAs.

Results: Small to modest correlations were found between YSR/CBCL scores (mean r=.18 across all scales). These finding show that, on average, adolescents report higher severity of problems than residential staff on all three problem severity difference score scales. Despite caregivers underreporting severity of adolescent problems, the problem frequency difference scores indicate that the residential caregivers recognize more overall problems on average than adolescents. Adolescent attachment style was associated with levels of agreement on problem frequency difference scores; specifically, staff-youth dyads with adolescents who endorsed secure attachment reported higher levels of agreement on number of internalizing problems (t(58)=-2.77, p=.007). Relationship duration was associated with problem frequency difference scores (t(58)=-2.68, p=.009); indicating when the staff knew the child for one year or more there was greater agreement on number of endorsed items for internalizing problems.  Adolescents endorsing high perceived closeness to staff showed less agreement on anxiety/depression symptoms (t(58)=1.98, p=.05). Residential staff reporting high perceived fit with the youth displayed better agreement on total problems and externalizing problems.

Implications: Worldwide, staff are positioned to evaluate youth mental health for children in residential settings. Adolescents endorsed higher standard scores indicating higher perceived severity in problems; dichotomized scores indicate that residential staff recognized more overall problems. Although residential staff recognize a higher number of problems than adolescents, staff appeared to perceive these problems as less severe. It is also possible that adolescents minimize minor problems until they become severe. Quality of the relationship appears to influence the quality of reporting; findings indicate residential staff with high perceived fit are better able to approximate adolescent-reported psychopathology. Results indicate that residential staff are able to identify child-endorsed psychopathology through the CBCL. Implications for social work practice will be discussed.