Development and psychometric validation of the Japanese version of the pediatric eating assessment tool (PediEAT-J)

This study aimed to develop and validate the Japanese version of the Pediatric Eating Assessment Tool (PediEAT-J). The original PediEAT underwent forward–backward translation and cultural adaptation. Caregivers of 697 children aged 6 months–7 years completed the PediEAT-J. A subset of 70 caregivers repeated the questionnaire after two weeks to assess test–retest reliability. Internal consistency was evaluated with Cronbach’s alpha; reliability with the intraclass correlation coefficient (ICC). Construct validity was examined via confirmatory factor analysis (CFA) of the original four-factor model; discriminant validity was examined by comparing scores of children with caregiver-reported feeding concerns to those without. Internal consistency was high across subscales and the total score (α = 0.806–0.913). Test–retest reliability was excellent (ICC = 0.902). Model-fit indices were acceptable: comparative fit index = 0.889; Tucker–Lewis index = 0.867; root mean square error of approximation = 0.039 (90% CI: 0.036–0.041); standardized root mean square residual = 0.071. Children whose caregivers expressed feeding concerns scored significantly higher than those without concerns on total and subscale scores (all p < 0.01), confirming discriminant validity. The PediEAT-J is a reliable and valid Patient-Reported Outcome Measure (PROM) that enables comprehensive assessment of feeding difficulties in Japanese children aged 6 months–7 years. It supports clinical assessment and intervention by providing structured caregiver-reported data on a child's feeding skills.