Abstract
Disagreement between parents and children about a child’s symptoms, termed parent/child informant discrepancies (PCIDs), is one of the most robust findings in the child clinical literature (e.g., De Los Reyes, 2011). Prior theoretical (e.g., Ferdinand, van der Ende, & Verhulst, 2004) and empirical (e.g., Kolko & Kazdin, 1993) work suggests greater discrepancies are associated with a host of factors which may impede treatment outcomes (e.g., family functioning, child insight, parental psychopathology). This suggests that PCIDs contain clinically meaningful information, and likely represent important and easily collected indicators for treatment prognosis. The current study examines how discrepant information gathered from parents and children, long considered a pervasive problem in the child psychology literature, can help identify youth at risk for treatment failure. Using data from the largest treatment trial of pediatric anxiety to date, the Child/Adolescent Anxiety Multimodal Study (Walkup et al., 2008; Ginsburg et al 2011), the current study applies the latest statistical advancements in the study of informant discrepancies (i.e., polynomial regression coefficients; Laird & De Los Reyes, 2013), as well as latent growth curve modeling, to examine how PCIDs prior to and throughout treatment predict treatment outcomes. Findings suggested that when parents reported more symptoms than youth, children were less likely to be free of their diagnoses post-treatment; however, this was only true if the child received CBT alone. Additionally, latent growth curve modeling suggested that significant changes in informant agreement occurred over the course of treatment, and increasing concordance between parents and youth predicted better youth outcomes. Findings suggest that how parents and youth “co-report” is an important indicator of treatment outcome and present an exciting new direction for these discrepancies - they can potentially be utilized to both predict and monitor treatment progress.