Author: Diana Khanna
Khanna, Diana, 2025 Investigating agreement in Proxy and Self-Reported Health-Related Quality of Life in children for informing Economic Evaluation and Quality assessment, Flinders University, College of Nursing and Health Sciences
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
Background: The application of child-specific preference-based measures to assess Health-Related Quality of Life (HRQoL) enables the derivation of utilities (preference weights) for incorporating into cost utility analysis of health and social care technologies or quality assessments targeted for paediatric populations. Challenges in paediatric self-assessment of HRQoL due to developmental and cognitive constraints necessitate proxy-reports alongside or in place of child self-reports. This thesis examines the inter-rater agreement between child-self and proxy assessments of child HRQoL using generic preference-based measures. The objective is to contribute to the existing evidence by focusing on reported measures of agreement in child and proxy assessments, particularly in highlighting dimension level discrepancies in addition to overall HRQoL.
Methods: The research methodology commenced with a systematic review, examining the existing evidence of agreement between child-self and proxy-reported HRQoL across both overall and at the dimension level using generic preference-based measures, the EQ-5D-Y-3L and the CHU9D. Subsequent empirical investigations involved two distinct studies. The first study was conducted through face-to-face interviews with 85 child-parent dyads, encompassing children aged 6-12 years. This study utilised a mixed-methods approach, combining quantitative and qualitative analyses to gain a comprehensive understanding of the HRQoL assessments. The second study was a larger quantitative analysis involving 845 online child-proxy dyads, where participants completed the EQ-5D-Y-3L for assessing child HRQoL.
Results: The systematic review revealed a generally low inter-rater agreement between child-self and proxy ratings of HRQoL, particularly in dimensions with less observable attributes such as emotional and social well-being. The first empirical study confirmed these findings, showing a consistent but low level of overall agreement across child-proxy assessments, with the lowest concordance observed in psychosocial health-related dimensions. However, the study also noted that adopting a proxy-child perspective resulted in higher agreement for the “feeling worried, sad, or unhappy” dimension in the EQ-5D-Y-3L. Additionally, with the CHU9D, children aged 6-7 years had comparable dimension level agreement to older age groups (8-10 and 11-12 years), whereas the PedsQLTM showed lower agreement for older age groups in similar psychosocial health dimensions. The mixed-methods approach further highlighted that the majority of younger children, especially those aged 6-10 years, tended to face significant challenges in reporting HRQoL dimensions as intended, leading to disparities with parental proxy reports. These discrepancies were not always reflective of actual health impairments but rather indicative of understanding and interpretation challenges. The second study’s findings, based on a larger sample, suggested that the Australian adult value set for the EQ-5D-Y-3L may provide agreement levels comparable to child-specific value sets in assessing HRQoL.
Conclusions: This thesis is one of the first to comprehensively examine the evidence that relates to the level of inter-rater agreement between child-self and proxy-reported child HRQoL, using generic preference-based measures. The findings significantly contribute to the methodological considerations in measuring child HRQoL, highlighting the complexities in using proxy reports as substitutes, particularly for subjective HRQoL, and in the intended interpretations of the dimensions by younger child-self reporters. The research underscores the need for enhancing current age-appropriate HRQoL measures and calls for ongoing research to improve the validity of both self and proxy reports. This research is vital for health economists, policymakers, and practitioners in paediatric health technology assessment, aiming to ensure that health technologies are evaluated accurately, reflecting the true needs and HRQoL experiences of the paediatric population.
Keywords: HRQoL, QoL, child, preference-based measures, generic, proxy, inter-rater agreement
Subject: Health Sciences thesis
Thesis type: Doctor of Philosophy
Completed: 2025
School: College of Nursing and Health Sciences
Supervisor: Prof Julie Ratcliffe