Abstract
OBJECTIVE: Although several assessments have been developed to diagnose or measure avoidant/restrictive food intake disorder (ARFID) symptoms, few studies have validated these tools in nonclinical and adult samples. This study explored the validity of two self- and parent/guardian-report ARFID screening measures in identifying adults and children who may have ARFID within a large community sample. METHOD: Fifty participants (divided into two groups: 25 adults and 25 parents/guardians of children) were selected from the ARFID Genes and Environment study, which enrolled over 3000 adults and parents/guardians of children who screened positive for ARFID on either the Pica, ARFID, and Rumination Disorder Interview-ARFID Questionnaire (PARDI-AR-Q) or the Nine Item ARFID Screen (NIAS) self- and parent/guardian-report measures. Participants then completed the ARFID portion of the Pica, ARFID, and Rumination Disorder Interview (PARDI) to determine ARFID diagnosis. RESULTS: Correlations between the PARDI-AR-Q and PARDI (r = 0.31-0.67) were weaker than the correlations between the NIAS and PARDI (r = 0.53-0.64) in both groups. The diagnostic positive predictive value for the PARDI-AR-Q was numerically higher (adults = 55.0%; parents/guardians = 76.0%) than the NIAS (adults = 45.8%; parents/guardians = 64.0%). Most PARDI-AR-Q dimensions and all NIAS dimensions were significant predictors of their corresponding PARDI dimensions in both groups. DISCUSSION: The PARDI-AR-Q more accurately identified adults and children with ARFID, whereas the NIAS was a better estimator of ARFID symptoms. These findings provide partial support for using these self- and parent/guardian-report screeners. Results highlight the need to better understand diagnostic presentations of ARFID within community samples, particularly in adults, and to refine these tools within those populations.