Self-report versus clinical interview: Discordance among measures of binge eating in a weight-loss seeking sample

Abstract

Purpose Obese, behavioral weight-loss (BWL) seeking individuals may be prone to over-reporting binge-eating (BE). However, many studies rely on self-reported measures of BE in this population, which may be inaccurate. As such, this is the first-ever study to examine the concordance rates among one self-reported and one clinician- administered measure of BE in a BWL-seeking sample with overweight/obesity. Methods At baseline of a BWL trial, participants (N = 94) completed two measures of BE: The Eating Disorders Examination Questionnaire (EDE-Q) and the interview-based Eating Disorder Examination (EDE, Overeating section). Results Cohen’s kappa detected poor agreement between measures (κ < 0). A paired samples t-test detected large, significant differences in OBE frequency across the EDE-Q and EDE, p < 0.001. The self-reported EDE-Q detected a significantly greater frequency of OBEs compared to the EDE (M_EDE-Q = 0.73, SD = 1.29 vs. M_EDE = 0.06, SD = 0.34). The EDE-Q detected that approximately 50% of participants have experienced OBEs, while the EDE detected that only 5% of participants have experienced OBEs. The frequency of OBEs detected by the EDE-Q was statistically greater than the frequency of OBEs detected by the EDE, p < 0.001. Discussion Results suggest poor agreement between one self-reported measure and the “gold-standard,” clinician-administered measure of BE in a BWL-seeking sample with overweight/obesity. The EDE-Q exhibited high sensitivity but low-to-moderate specificity of OBEs, with the number of false positives (41) outweighing that of true positives (4). Studies measuring BE in this population should consider relying solely on assessor-administered measures, as this sample may require clinical guidance or clarification on the definition and features of BE. Level of evidence Level V, cross-sectional, descriptive study.