RESUMO
INTRODUCTION: Obesity is a major health problem with an increasing risk of mortality, associated with comorbidities and high rates of dropout. Research demonstrated that pathological eating behaviors could help to phenotype obese patients thus tailoring clinical interventions. Therefore, our aim was to develop (study 1), validate (study 2), and test in a clinical setting (study 3) the Eating Behaviors Assessment for Obesity (EBA-O). METHOD: Study 1 included the exploratory factor analysis (EFA) and McDonald's ω in a general population sample (N = 471). Study 2 foresaw the confirmatory factor analysis (CFA) and convergent validity in 169 participants with obesity. Study 3 tested the capability of the EBA-O to characterize eating behaviors in a clinical sample of 74 patients with obesity. RESULTS: Study 1. EFA identified five factors (i.e., food addiction, night eating, binge eating, sweet eating, and prandial hyperphagia), explaining 68.3% of the variance. The final EBA-O consisted of 18 items. McDonald's ω ranged between 0.80 (hyperphagia) and 0.92 (binge eating), indicating very good reliability. STUDY 2: A second-order five-factor model, through CFA, showed adequate fit: relative chi-square (χ2/df) = 1.95, CFI = 0.93, TLI = 0.92, RMSEA = 0.075, and SRMR = 0.06, thus suggesting the appropriateness of the EBA-O model. Significant correlations with psychopathological questionnaires demonstrated the convergent validity. Study 3. Significant associations between EBA-O factors and emotional-related eating behaviors emerged. CONCLUSION: The EBA-O demonstrated to be a reliable and easy-to-use clinical tool to identify pathological eating behaviors in obesity, particularly useful for non-experts in eating disorders. LEVEL OF EVIDENCE: Level V, descriptive research.