Pedi‐IKDC exhibits questionable measurement properties in a cohort of pediatric patients with ACL rupture

Abstract

BackgroundPedi‐IKDC is commonly used to evaluate anterior cruciate ligament (ACL) deficiency in children. However, its construct validity has not been thoroughly assessed. The aim was to examine the measurement properties of the Pediatric International Knee Documentation Committee (Pedi‐IKDC) by modern test theory (MTT) models, confirmatory factor analysis (CFA), and item response theory (IRT).MethodsThe cohort consisted of all children and adolescents in Denmark (n = 535, age 9–16) treated with physeal‐sparing ACL reconstruction 2011–2020. Patient‐reported outcome measure (PROM) data were collected before surgery and at 1 year follow‐up. Structural validity of Pedi‐IKDC was assessed with MTT models. Reliability was reported as McDonalds coefficient omega. Responsiveness was evaluated with standardized response means.ResultsSufficient PROM data were available for 372 patients. The original unidimensional construct did not fit CFA model expectations neither before surgery (χ2 = 462.0, df = 163, p < 0.0001; RMSEA: 0.109, CFI: 0.910, TFI: 0.895) nor at follow‐up. Neither did a two‐factor CFA model with “Symptoms” and “Sports activities” as individual subscales (χ2 = 455.6, df = 162, p < 0.0001) nor a bifactor model (χ2 = 338.9, df = 143, p < 0.0001), although fit indices improved with the latter (RMSEA: 0.094, CFI: 0.941, TFI: 0.922). The IRT models confirmed this pattern. The scale was responsive (SRM 1.66 (95% CI: 1.46–1.88)). Coefficient omega values were 0.866 before surgery and 0.919 at follow‐up.ConclusionsThe Pedi‐IKDC exhibited inadequate structural validity. Neither the original construct, a two‐factor model, nor bifactor models fitted data well. We advise that data obtained by Pedi‐IKDC are interpreted with caution.