Performance of the EQ-5D-Y Interviewer Administered Version in young children

Amien, Razia

Performance of the EQ-5D-Y Interviewer Administered Version in young children

Master Thesis

2022

Abstract

Introduction The interest in Health-Related Quality of Life (HRQoL) in the paediatric population has grown over the last decade as it allows for a more holistic approach which has the potential to positively influence treatment outcomes (1–4). With an increase in interest, the need for alternative modes of administration of HRQoL instruments has become more important to allow for self-report in younger age-groups. Despite their age and/or literacy levels, their ability to understand the concept of HRQoL would allow for accurate self-report if the correct instrument is used (5). Proxy-report is used often as a default in these younger age-groups as only two interviewer-administered instruments are currently available (6,7), neither of which have been validated for African populations. The newly developed EQ-5D-Y-3L Interviewer Administered (IA) instrument would allow for self-report in younger paediatric populations, therefore limiting the reliance on proxy-report which does not account for the subjectivity of HRQoL or allow for the inclusion of the child's view (2,8–10). Aim The first aim of this study was to determine the performance and preference of the EQ-5D-Y-3L-IA and self-complete (SC), in children aged 8-10-years. The second aim was to determine the psychometric performance of the EQ-5D-Y-3L-IA version in children aged 5-7-years compared to those aged 8-10- years. Methods A cross-sectional, descriptive observational, analytical design was used. Children were recruited in two age-groups, 5-7-years (n=177, 46%) and 8-10-years (n=211, 54%). Participants were drawn from the General Population (GenPop) attending a Mainstream School (n=109, 28%), Special Schools for learners with special educational needs (n=55, 14%) and healthcare facilities caring for children with orthopaedic conditions (n=161, 41%) or chronic respiratory illnesses (n=63, 16%). All children completed the EQ-5D-Y-3L-IA, Faces Pain Scale-Revised (FPS-R), Moods and Feelings Questionnaire (MFQ). The researcher completed the observational Functional Independence Measure (WeeFIM). In addition, children in the 8-10-year group completed the EQ-5D-Y-3L-SC. Dimension responses of the EQ-5D-Y-3L-IA and SC were analysed for floor and ceiling effects, inconsistent responses, missing responses and differences in health states between age-groups and versions. Differences in reporting were determined by chi-square statistic (x2 ). Known-group validity across age (years), sex and health conditions were analysed using Spearman's rank order coefficients (rs) in addition to the median utility and Visual Analogue Scale (VAS) scores using Kruskal Wallis and Mann-Whitney U-test. Pearson's correlation was used to assess concurrent validity by comparing the utility and VAS scores between versions. Spearman's Rank Correlation was computed to assess the convergent validity of the EQ-5DY-3L-IA and SC compared to the FPS-R, MFQ and WeeFIM. Responses from structured cognitive debriefing interviews were grouped and coded by the researcher according to similar responses provided by participants. Cognitive debriefing was used to determine the acceptability, comprehensibility and where applicable, participants' preference between versions and the reasons for their preference. The researcher was aware of reflexivity and did not allow personal opinions to impact on participants' responses, nor the grouping and coding of responses. The EQ-5D-Y-3L-IA was retested 48 hours later only in children with a stable health condition, recruited from schools and analysed using weighted Cohen's kappa statistic (k) for dimension scores and the Intraclass Correlation Coefficient (ICC) for utility and VAS scores. Results There were no concerning differences in EQ-5D-Y-3L dimension responses, known-group validity, concurrent validity or correlation of VAS and utility scores between the IA and SC versions. The IA version had the advantage of no missing values and was preferred over the SC version by 8-10-yearolds (60%). When comparing the IA version between age-groups, the performance was similar. However, children aged 5-7-years reported significantly more problems with the Looking After Myself dimension (x2 =31.021; p<. 0001) by which cognitive debriefing revealed developmental difficulty with advanced dressing tasks such as laces and buttons. Conclusion Validity and test-retest reliability of the EQ-5D-Y-3L-IA version was successfully assessed in children aged 5-10-years. As the results were comparable to the SC version in children aged 8-10-years, it therefore indicates that versions can be used interchangeably. In settings with low literacy levels, such as South Africa, the IA version is recommended for young children, most notably those 8-years of age. The performance of the IA version across age-groups showed that younger children can reliably report on their HRQoL therefore also proved useful in younger age-groups, however, adaptations to the dimension of Looking after myself is suggested for improved developmental appropriateness. Therefore, it is recommended that EQ-5D-Y-3L-IA be included in children from 5-years in routine clinical practice and clinical trials.

Keywords

Physiotherapy

Reference:

Collections

Masters

Full item page