The Pediatric Quality of Life Inventory 3.2 Diabetes Module: Validity, Reliability and Factor Structure of the Croatian Version

Background/Objectives: Type 1 diabetes mellitus (T1DM) diagnosed in primary school children presents unique challenges due to developmental dependence on adults, limited self-care abilities, and the need for continuous medical supervision. These challenges may affect health-related quality of life (HRQoL), particularly within early educational environments. The aim of this study was to test validity, reliability and factor structure of the Croatian version of the Pediatric Quality of Life Inventory (PedsQL) 3.2 Diabetes Module. Methods: The sample included 70 children aged 7-14 years and their parents or caregivers. HRQoL was measured using the existing Croatian version of the PedsQL 4.0 Generic Core and the Croatian version of the PedsQL 3.2 Diabetes Module which we validate in this study. Croatian version of the PedsQL 3.2 Diabetes Module was developed using forward-backward translation. Reliability was assessed through Cronbach’s α and test-retest analysis. Spearman’s correlation examined the relationship between the PedsQL 3.2 Diabetes Module total scale and its scales. Construct validity was evaluated with exploratory factor analysis. Results: Children with T1DM reported higher overall HRQoL comparable than their parents, except in Diabetes symptoms scale of the PedsQL 3.2 Diabetes Module and School functioning scale of the PedsQL 4.0 Generic Core. All scales of the Croatian version of the PedsQL 3.2 Diabetes Module demonstrated satisfactory internal consistency (Cronbach’s α = 0.71-0.85) and favourable pattern of Spearman’s correlations with total scores (ρ = 0.61-0.92). The test-retest reliability of all PedsQL 3.2 Diabetes Module scales and total scores were excellent (ICC: 0.982-0.996). Bartlett’s test of sphericity indicated a high and significant correlation for child self-report (χ2 =1398.57, p<0.001) and parent proxy-report (χ2 =1302.74, p<0.001). The Kaiser-Meyer-Olkin value measured was sufficient, 0.65 for child’s and 0.68 for parents. The factors extracted for child self-report accounted for 66.30% of the total variance, with factor loadings ranging from 0.41-0.89. The factors extracted for parent proxy-report accounted for 61.80% of the total variance, with factor loadings ranging from 0.41-0.85. Conclusions: The Croatian version of the PedsQL 3.2 Diabetes Module is reliable, valid and feasible instrument for assessing HRQoL in Croatian primary school-aged children diagnosed with T1DM.