Measuring quality of life in autistic children and young people: comparing the performance of common generic health related quality of life instruments

PURPOSE: To assess the comparative measurement performance of commonly used paediatric health related quality of life (HRQoL) instruments in autistic children. METHODS: The sample was 510 autistic children aged 5-18 years from the Australian Paediatric Multi-Instrument Comparison study. The psychometric performance of the PedsQL, EQ-5D-Y-5L, EQ-5D-Y-3L, CHU9D and KIDSCREEN-27 instruments was compared. Psychometric attributes assessed included acceptability (ease/difficulty completing), feasibility (time to complete), test-retest reliability, floor/ ceiling effects, convergent validity and known-groups validity (measured through schooling, special health care needs (SHCN), and the strengths and difficulties questionnaire (SDQ)). Sub-group analysis by reporter type (child vs. caregiver), child age (5-12 vs. 13-18 years) and child gender (male vs. female) was conducted. RESULTS: All measures demonstrated similarly good acceptability, feasibility, and convergent validity. Only the PedsQL, EQ-5D-Y-5L and EQ-5D-Y-3L demonstrated test-retest reliability (intraclass correlation coefficient 0.56-0.60). No floor or ceiling effects were found in the total sample however, ceiling effects for the EQ-5D-Y-3L were noted in 13-18 year olds (19.1%), males (15.6%), and children who self-reported (15.0%). For known-groups validity (Cohen's d > 0.5), only the EQ-5D-Y-5L differentiated between children in mainstream and specialist schools, only the CHU9D and PedsQL differentiated between children with and without SHCNs, and all measures differentiated between children with and without behavioural difficulties as per the SDQ. CONCLUSION: The PedsQL and EQ-5D-Y-5L were the strongest performing measures in autistic children. As performance varied by reporter type, child age, and child gender, those selecting a measure should assess evidence in their intended population of use.