Characterizing Interpersonal Perception in Adolescent Eating Disorders

Eating disorders (EDs) are associated with significant morbidity and mortality, and recovery is a challenging and often protracted process. Improved understanding of EDs, including identification and refinement of treatment and recovery targets, is necessary to improve intervention and treatment outcomes. Cognitive and interpersonal models of EDs propose interpersonal function as a risk and maintenance factor, and the adult ED literature documents challenges in self-concept and interpersonal function during illness that improve with recovery. Adolescence is a significant period for social development as well as ED onset and early intervention. Despite this, little research has prospectively examined interpersonal perception as a potential target in adolescents. This observational study examined baseline differences in self-referential thinking in adolescents with EDs (n = 29) compared to healthy control adolescents (n = 31) using the Internal, Personal, and Situational Attributions Questionnaire self-report and a Social Identity behavioral task. Self-attribution bias was also examined prospectively as it related to both time and clinical symptoms in 25 ED and 30 HC. When controlling for depression severity, the ED group had a more negative self-attribution bias than the HC group (p = .006). Additionally, the ED group exhibited less positive self (p < .001), social (p = .015), and socially reflected appraisals (p = .011) but no group differences for friend and friend reflected appraisals during the Social Identity task. After including depression as a covariate, however, the model was no longer significant, suggesting group differences in social appraisals may be mediated by depression. Additionally, there was no significant effect of clinical symptoms on self-attribution bias, although more severe clinical symptoms were significantly associated with lower valence indices for self, social, and socially reflected appraisals. Although significant group differences remained for self-attribution bias at follow up, there was no effect for time or group × time, nor was there a relationship between self-attribution bias change and clinical symptom change. Findings support valence in self-perception as a potential target in adolescent EDs. Future research should replicate results in larger samples and include longer follow-up duration to clarify the relationship between self-attribution bias and clinical status in adolescents.