Measuring children's HRQoL via self-reports and parent proxy-reports in the French Enabee 2022 study: an evaluation of the psychometric properties of both versions of the French KINDL-R questionnaire.

Abstract

Introduction: Child well-being policies require quality data that takes children's perspectives into account, recorded through appropriate, accurate and reliable measurement instruments. This study evaluated the psychometric properties of two versions of the KINDL-R questionnaire used in the 2022 French National Study on Children's Well-being (Enabee) to assess Health-Related Quality of Life (HRQoL) in children: a self-report version for elementary school children aged 6-11 (n = 15,114), and a parent-report version for preschool and kindergarten children aged 3-6 (n = 3,785).

Methods: We examined the distributions of responses and the correlations between items. We performed confirmatory factor analysis (CFA) and exploratory structural equation modelling (ESEM) to assess the fit of first- and second-order factor models. Internal consistency was measured using McDonald's omega. We tested measurement invariance across child's gender, grade, type of school and whether or not they had a chronic illness. We also conducted a known-groups analysis via analysis of variance (ANOVA) based on child's chronic illness, mental health and parental educational attainment.

Results: The parent-report version demonstrated satisfactory CFA model fit (CFI = 0.915, TLI = 0.901, RMSEA = 0.072, SRMR = 0.073), supporting dimensional scoring. The acceptable fit of the hierarchical model (CFI = 0.916, TLI = 0.905, RMSEA = 0.073, SRMR = 0.075) provides preliminary support for reporting an overall score. Invariance was confirmed for the child's gender, grade, type of school and chronic illness. This version effectively distinguished between children with chronic illnesses or at least one probable mental health difficulty and healthy peers. However, the self-report version lacked factorial validity and showed poor CFA model fit (CFI = 0.757, TLI = 0.717). The ESEM model provided a detailed assessment of the fitting issues, which were primarily related to the emotional well-being dimension.

Conclusion: The parent-report version of the Kiddy-KINDL-R can provide a valid measure of HRQoL for French children aged 3-6 with comparable mean scores across groups. However, the proposed factor structure of the self-report version of the Kid-KINDL-R was not supported, and most subscales showed insufficient internal consistency in elementary school children. Further refinement is thus required. This was the first study to evaluate the psychometric properties of the KINDL-R measure in the French general population. It provided critical, reliable, and unprecedented information about measuring children's HRQoL.