(PS7-75) A Response Surface and Polynomial Regression Analysis for Predicting Youth Impairment at Discharge: What Can Caregiver-youth Reports on Anxiety Tell Us?

Abstract Text: Multi-informant reporting has long been recommended as one of several best practices for assessing children and adolescent psychopathology (Kraemer et al., 2003). Gathering information from multiple perspectives yields benefits through examining cross-informant reporting patterns that might affect assessment findings and treatment recommendations (De Los Reyes & Kazdin, 2008). However, it has been acknowledged that in both clinical and research settings, informants often provide diverging perspectives when reporting about youth emotional and behavioral concerns. Additionally, for youth anxiety problems, many studies support the idea that the best informant on children’s symptoms is dependent on their developmental stage (Navarro et al., 2020). Recent studies have begun to investigate the extent to which informant discrepancies predict youth outcomes; however, the field should move beyond outdated strategies rooted in examining difference scores and instead embrace more modern analytic approaches.
New approaches such as response surface analysis allow for a comprehensive overview of the association between a dependent variable and varying levels of agreement and disagreement with predictor variables. Using results from polynomial regression analysis, coefficients are graphed in a three-dimensional space visualizing the extent to which caregiver-youth dyad scores, characterized as having minimal discrepancy compared to large discrepancy, can predict criterion variables.
Participants included youth and caregivers receiving public mental health services between January 2014 to December 2018 through the Hawaii State Department of Health. A total of 366 youths (Mage = 15.03) and their caregivers were included, with majority of youth identified as multiethnic (39.9%) and female (52.2%). Focusing on their responses to the (a) Child and Adolescent Functional Assessment Scale (CAFAS) and (b) Ohio Scales Short Form, Problem Severity domain Anxiety Subscale, a response surface analysis was conducted to examine how varying levels of informant agreement and disagreement on the Anxiety Subscale related to youth discharge CAFAS scores. Results indicated that youth discharge CAFAS Total Scale scores were higher (i.e., increased impairment) when caregiver and youth Anxiety Subscale scores converged (i.e., interrater agreement) at higher levels. Functional impairment at treatment discharge could not be predicted by youth-caregiver disagreement on the Anxiety Subscale at treatment intake.
The present findings contribute to the current literature by providing some support for the association between initial caregiver-youth informant discrepancy scores and youth impairment at discharge. Furthermore, the present study builds upon work in this area by utilizing the latest statistical approach recommendations (Laird & De Los Reyes, 2013) and a large ethnically diverse public mental health sample of youths and their caregivers.