Psychometric Evaluation of Self- and Proxy-Reported EQ-5D-Y-3L and EQ-5D-Y-5L in Children With Attention-Deficit/Hyperactivity Disorder

OBJECTIVES: This study aimed to assess the psychometric properties of 2 versions of EuroQol 5-Dimension Youth version (EQ-5D-Y) (3 level [Y-3L] and 5 level [Y-5L]) in children with attention-deficit/hyperactivity disorder (ADHD) and to compare their performance across self- and proxy-reported responses. METHODS: Overall, 188 child-caregiver dyads (mean age 9 years with a clinical ADHD diagnosis) completed a cross-sectional survey, followed by a 2-week assessment to examine test-retest reliability. Outcome measures included the Y-3L and Y-5L, Child Health utility 9D, Child Behavior Checklist, and the Swanson, Nolan, and Pelham Rating Scale for ADHD severity. Ceiling, informativity, convergent and known-group validity, patient-proxy agreement, and test-retest reliability were assessed. RESULTS: Self-reported Y-3L (48%) and Y-5L (28%) had higher ceilings than proxy versions (32% and 25%), and the Y-5L showed higher absolute informativity for both self- and proxy-reports. Self-reports correlated more strongly with the child-completed Child Health utility 9D, whereas proxy reports were more aligned with Child Behavior Checklist. Self-report version showed good known-groups validity: the Y-3L discriminated 11 of 14 groups, and the Y-5L discriminated 13 of 14. The proxy versions showed slightly weaker discrimination. Item-level test-retest reliability and patient-proxy agreement were good (Gwet's agreement coefficient [AC] 0.71-0.98 and 0.70-0.97), but scale-level reliability and agreement were lower. CONCLUSION: In Chinese children with ADHD, the Y-3L appears more suitable for self-report, whereas the Y-5L may be better for proxy report, depending on the population and study aims. Further research is needed to identify the optimal version across age groups and clinical conditions.