Abstract
Purpose (a) To compare the agreement between adolescent assessments of their quality of life (QoL) and that of their mothers; (b) to explore how the comparison is influenced by the method of analysis. Methods Forty-nine adolescents aged 12–18 years who received liver transplants, and their mothers completed the Child Health Questionnaire self (CF87) and parent (PF50) report. Results There was wide variation in agreement between adolescent and parent responses depending on the method of analysis used. Analysis with t test showed no differences in physical function (t = 1.42, P = 0.16), role/social-physical (t = 0.07, P = 0.94), mental health (t = 0.55, P = 0.59) and family activities (t = −0.40, P = 0.69). Using Pearson correlation coefficients, there were significant correlations in every domain; however, there were no intraclass correlation or concordance correlation coefficients ≥0.80 suggesting less than strong agreement. Finally, the Bland–Altman comparison indicated wide variation in the 95% limits of agreement ranging from −46 to 58.5. Conclusions There was considerable inconsistency in agreement according to the methods of analysis. The wide variation in scores between adolescent and parent assessment of QoL suggests self rather than proxy report should be used as the primary outcome where possible.