Reproducibility of food consumption frequencies derived from the Children's Eating Habits Questionnaire used in the IDEFICS study

Objective: To investigate the reproducibility of food consumption frequencies derived from the food frequency section of the Children's Eating Habits Questionnaire (CEHQ-FFQ) that was developed and used in the IDEFICS (Identification and prevention of dietary-and lifestyle-induced health effects in children and infants) project to assess food habits in 2- to 9-year-old European children. Design and methods: From a subsample of 258 children who participated in the IDEFICS baseline examination, parental questionnaires of the CEHQ were collected twice to assess reproducibility of questionnaire results from 0 to 354 days after the first examination. Weighted Cohen's kappa coefficients (kappa) and Spearman's correlation coefficients (r) were calculated to assess agreement between the first and second questionnaires for each food item of the CEHQ-FFQ. Stratification was performed for sex, age group, geographical region and length of period between the first and second administrations. Fisher's Z transformation was applied to test correlation coefficients for significant differences between strata. Results: For all food items analysed, weighted Cohen's kappa coefficients (kappa) and Spearman's correlation coefficients (r) were significant and positive (P < 0.001). Reproducibility was lowest for diet soft drinks (kappa = 0.23, r = 0.32) and highest for sweetened milk (kappa = 0.68, r = 0.76). Correlation coefficients were comparable to those of previous studies on FFQ reproducibility in children and adults. Stratification did not reveal systematic differences in reproducibility by sex and age group. Spearman's correlation coefficients differed significantly between northern and southern European countries for 10 food items. In nine of them, the lower respective coefficient was still high enough to conclude acceptable reproducibility. As expected, longer time (> 128 days) between the first and second administrations resulted in a generally lower, yet still acceptable
Group Author(s): IDEFICS Consortium