Reproducibility and Validity of the KIDMED Index in a sample of Portuguese adolescents.

The KIDMED Index is frequently used to evaluate adherence to the Mediterranean Dietary Pattern among children and adolescents, through 16 questions with the associated total score ranging from -4 to 12. However, in the authors' best knowledge, the psychometric properties of this index had not yet been investigated in Portugal. Thus, the main purpose of the present study was to investigate the reproducibility and the validity of the KIDMED Index in a sample of 185 Portuguese adolescents. The reproducibility was tested by comparing the application of the KIDMED Index at two different times (two-week interval), using McNemar test and Kappa statistics. There was moderate agreement (κw = 0.591, 95% CI = 0.485-0.696) and no significant change (p-value = 0.201) in the KIDMED Index classification, between the two applications. The validity was explored by comparing the results obtained by the KIDMED Index and by the average of 3-days Dietary-Record (DR), using Spearman's correlation coefficient and Kappa statistics. There was weak correlation (ρ = 0.317; p-value < 0.001) and slight agreement (κw = 0.167, 95% CI = 0.071-0.262) between the KIDMED Index classification and the 3-days DR-derived KIDMED score; and moderate correlation (ρ = 0.423; p-value < 0.001) and fair agreement (κw = 0.344, 95% CI = 0.202-0.486) between the terciles of the KIDMED Index and the Mediterranean Adequacy Index scores. The results suggested an acceptable reproducibility and validity of the Portuguese version of the KIDMED Index, in alignment with the few studies investigating psychometric properties of this index in other countries.