Examining the construct validity of food addiction severity specifiers

Purpose Food addiction (FA) is related to greater body mass index (BMI), eating-disorder psychopathology, food craving, and psychosocial impairment. Less is known regarding the utility of the FA severity specifiers, as measured by the number of symptoms endorsed on the Yale Food Addiction Scale (YFAS 2.0). Methods Participants (N = 1854) were recruited from Amazon Mechanical Turk to complete an online survey on eating behaviors. Participants completed self-report measures assessing FA, eating-disorder psychopathology (Eating Disorder Examination Questionnaire), and food craving (Food Craving Inventory). Based on the YFAS 2.0 specifiers, participants were classified into four FA groups: No FA (n = 1643), mild (n = 40), moderate (n = 55), and severe (n = 116). Results There were significant differences found in age, sex, BMI, and frequency of objective binge-eating episodes (OBEs) among the FA groups. Using ANCOVA, adjusted for multiple comparisons and covariates (e.g., BMI, sex, OBEs), the No FA group reported significantly lower levels of shape concern (eta(2) = 0.05;p < 0.001), weight concern (eta(2) = 0.04;p < 0.001), eating concern (eta(2) = 0.15;p < 0.001), and global eating-disorder psychopathology (eta(2) = 0.06;p < 0.001) than mild, moderate, or severe FA groups. The No FA group reported significantly lower levels of dietary restraint (eta(2) = 0.01;p < 0.01) than mild and severe FA groups. The severe FA group reported higher food craving scores (eta(2) = 0.02;p < 0.001) compared to the No FA group. Conclusion Our findings parallel the severity specifiers literature for eating and substance use disorders by also indicating the limited utility of severity specifiers based on symptom count. Future research should investigate alternative targets for discriminating among levels of FA.