OBJECTIVE: In paediatric research, investigators rely on proxy reports of outcome, such as the proxy-completed Childhood Health Assessment Questionnaire (P-CHAQ), to assess function in JIA. As children mature, they may self-complete the adult Health Assessment Questionnaire (HAQ) or the un-validated adolescent-specific CHAQ (A-CHAQ). It is unclear how these measures compare and whether they are directly interchangeable. METHODS: Adolescents aged 11 to 17 years participating in the Childhood Arthritis Prospective Study, a UK multicentre inception cohort, were included. In a CAPS sub-study, adolescents self-completed the A-CHAQ and HAQ and proxies simultaneously completed the P-CHAQ at baseline and one year. Correlation and agreement between scores were assessed at baseline. Agreement and ability to similarly classify clinically important changes over time were assessed at one year following initial presentation to rheumatology. RESULTS: 107 adolescents (A-CHAQ and HAQ) or their proxies (P-CHAQ) had completed all three measures at baseline. Median age at diagnosis was 13 years and 61% were female. Although the three scores demonstrated strong correlations (r>0.8), they were not completely interchangeable, with agreement ranging between 70 and 80%. There was similar agreement between the changes in scores between baseline and one year. Using P-CHAQ minimal clinically important cut-offs, the A-CHAQ and HAQ similarly classified 80 to 90% of adolescents as having improved or worsened. CONCLUSION: Whilst there is relatively high agreement and similar classification of change between HAQ and two CHAQ scores, these are not completely interchangeable. This impacts the comparison of function when measured in different ways over the lifespan. This article is protected by copyright. All rights reserved.