Patterns of Parent-Youth Reports of Internalizing Problems in a Psychiatric Inpatient Sample: Links to Clinical Outcomes

Low-to-moderate correspondence among informants (i.e., rs in the 0.20s to 0.30s) is common and particularly for youth internalizing problems (Achenbach et al., 1987; De Los Reyes et al., 2013). The task of integrating discrepant reports is complex, and the approach taken strongly affects estimates of prevalence rates of disorders and levels of risk, case conceptualization and treatment planning, and gauging treatment response (Hawley & Weisz, 2003; De Los Reyes & Kazdin, 2005). However, there is a dearth of research examining the extent to which multi-informant reports provide incremental validity in the prediction of important criterion variables (e.g., treatment response; De Los Reyes et al., 2015). Such information would be particularly useful in severe clinical samples, in which clinical decision-making is associated with high-risk outcomes (e.g., suicide attempts following discharge from an inpatient unit). To address this, we examined patterns of parent-child reports of internalizing problems in a sample of youth in a psychiatric inpatient sample. In addition, we seek to explore whether patterns of parent-child report predict treatment outcomes.

In an inpatient sample of 1344 youth (59.6% female; M_age=13.5; 51.6% African American), we examined patterns of parent-child reports on the Behavioral Assessment System for Children (BASC; Reynolds & Kampaus, 2004) internalizing subscales (i.e., anxiety, depression, somatization). We observed low levels of correspondence between parent and youth reports of anxiety (r=.15, p<.001), depression (r=.22, p<.001), and somatization (r=.08, p<.01). Latent class analyses (LCA) revealed four distinct patterns of parent-youth reports of internalizing problems: (a) Parent and Child Report Low (n = 795; 59.2%); (b) Parent Report Low and Child Report High (n = 78; 5.8%); (c) Parent Report High and Child Report Low (n = 203; 15.1%); (d) Parent and Child Report High (n = 268; 19.9%). We also plan to explore associations between patterns of parent-youth report and treatment variables (i.e., length of hospital stay, aftercare, use of seclusion and restraint, inpatient readmissions) to determine whether patterns of parent-child report predict important criterion variables.

The study’s findings suggest that parent-youth correspondence in reports of internalizing problems is particularly low in psychiatric inpatient settings. Despite this overall low correspondence, parent-child dyads displayed profound individual differences in reporting patterns, with some dyads displaying a relatively high level of convergence between reports. Findings from the present study can inform clinical decision-making by aiding our understanding of how multi-informant reports can be used to make predictions about treatment outcomes.