Psychometric evaluation and derivation of the Kannada version of the modified (29-item model) Indian Vision Function Assessment Questionnaire

Résumé fourni par la source

BACKGROUND/AIMS: India's linguistic and cultural diversity necessitates a region-specific validated Visual Functioning Questionnaire. The objective of this study was to translate the Indian Vision Function Questionnaire-33 (IND-VFQ-33) into the Kannada language and test its psychometric properties, underlying factor structure, and model fit. METHODS: A cross-sectional study was conducted among 330 participants, and basic psychometric properties (reliability, convergent, discriminant, construct validity, responsiveness, etc.) were assessed. The underlying factor structure was evaluated by exploratory factor analysis (EFA) and model fit by confirmatory factor analysis (CFA). RESULTS: The Kannada version of IND-VFQ-33 exhibited robust psychometric properties [excellent content validity, reliability (Cronbach α = 0.91), test-retest reliability (Rho S = 0.99; P < 0.001), convergent (Rho S = 0.58; P < 0.001), and discriminant validity (Rho S = 0.06; P = 0.274)]. On EFA, items in the general functioning subscale fell into two categories, namely "Mobility" (10 items: Q1, Q2, Q4-Q9, Q15, and Q16) and "Daily Activities" (8 items: Q3, Q10, Q11, and Q17-Q21), and two other subscales "psycho-social impact" subscale (Q22-Q26) and the "visual symptoms" subscale (Q27-Q32, excluding Q33). Thus, a new Kannada-translated VFQ-29 questionnaire was derived with excellent psychometric properties. All observed variables showed significant factor loadings on their respective latent constructs (>0.50), establishing convergent validity on CFA by using the structural equation model. The model-fit to the data assessed by various indices [Root Mean Squared Error of Approximation = 0.08 (90% CI: 0.081-0.091); Comparative Fit Index = 0.877; Tucker-Lewis Index = 0.866; Akaike's Information Criterion = 23447.133; Bayesian Information Criterion = 23800.45, and Coefficient of Determination = 0.998] demonstrated acceptable fit. CONCLUSION: The refined 29-item Kannada-translated IND-VFQ model not only offers an improved factor structure across four subscales but also consistently demonstrates excellent reliability, convergent and discriminant validity, and responsiveness.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.