How do adult inpatients with ARFID or AN-R compare on self-report eating disorder assessments?

Abstract

The Pica, ARFID, and Rumination Disorder ARFID Questionnaire (PARDI-AR-Q) and the Eating Disorders in Youth Questionnaire (EDY-Q) were developed to assess avoidant/restrictive food intake disorder (ARFID), but research is ongoing regarding their discriminant validity. This study examined scores of adult inpatients with ARFID or anorexia nervosa, restricting type (AN-R) on the PARDI-AR-Q, EDY-Q, and Eating Disorder Examination Questionnaire (EDE-Q) at admission to a medical stabilization unit for severe eating disorders. Response differences between 78 adult inpatients with ARFID or AN-R were analyzed using a two-sample t-test or Fisher's exact test. Sensitivity and specificity of self-report measures were examined using receiver operating characteristic curves. Patients with AN-R had significantly higher severity of impact and lower concern about aversive consequences scores than patients with ARFID on the PARDI-AR-Q (p < .001), with no significant differences on sensory-based avoidance or lack of interest or EDY-Q total score. Patients with AN-R scored significantly higher on all domains of the EDE-Q (p's < .001) and the EDE-Q subscales had better discrimination than ARFID measures. This study provides further evidence of the need for assessments that can distinguish ARFID from AN, and may also highlight the need for future research on the nosological accuracy of differentiating these disorders.