In-person interview versus online self-assessment of the eating disorder examination: an empirical validation of data equivalence

PURPOSE: Online self-assessments in mental health are becoming increasingly common, and the COVID-19 pandemic has further accelerated the need for reliable digital tools. However, the degree of equivalence between online self-assessments and in-person interviews in the context of eating disorders remains under-researched. This study investigated the equivalence of adolescent patient responses on the Eating Disorder Examination (EDE) collected across these two modalities. METHODS: The EDE was administered in a cross-over test-retest design, both as an online self-assessment questionnaire and as an in-person interview, to 49 adolescents (47 females, 15.9 ± 1.9 years) attending specialized outpatient clinic services. The Eating Disorder Inventory-2 (EDI-2) served as a reference measure. Equivalence of EDE scores was tested with a two one-sided t test procedure (TOST) for paired data with bootstrapping using a non-inferiority margin of ± 0.5 raw points. Cramer's V assessed agreement between the two EDE-derived diagnoses. RESULTS: The analyses confirmed equivalence between the two EDE versions. However, when comparing diagnoses based solely on EDE responses with clinician-derived diagnoses, a non-negligible effect of assessment modality emerged (Cramer's V = 0.56; proportion of shared variability = 0.31), with the online self-assessment version being more accurate. When objective weight data were included, diagnostic accuracy increased to nearly 90% for both modalities. CONCLUSIONS: These results support the equivalence of the online EDE self-assessment questionnaire with the in-person EDE interview and suggest that the online format may promote trustworthy responses through its perceived anonymity. Objective weight remains an important criterion when translating diagnostic data across different assessment modalities in eating disorders.