Reproducibility, validity, and responsiveness to change of a short questionnaire for measuring fruit and vegetable intake

The reproducibility, relative validity, and responsiveness to change of an eight-item. food frequency questionnaire designed to measure fruit and vegetable consumption was assessed among 157 women (mean age=41 years) in the Netherlands from spring 2001 to spring 2002. Plasma concentrations of vitamin C and total and specific carotenoids served as biomarkers against which validity was assessed. The questionnaire was completed and biomarker concentrations were determined three times: immediately preceding and following a controlled intervention of 1 month aimed at increasing fruit and vegetable consumption and 1 year after the start of the intervention. The 1-month and 1-year reproducibility of total fruit and vegetable consumption assessed in the control group was 0.80 and 0.79 (Spearman's r). Correlations between consumption and plasma carotenoids and vitamin C at baseline were 0.39 and 0.37, respectively, for fruits and 0.24 and 0.26, respectively, for vegetables. Correlations between changes in consumption and plasma carotenoids and vitamin C were 0.32 and 0.33, respectively, for fruits and 0.28 and 0.30, respectively, for vegetables. On the basis of similar correlations reported in the literature, the authors conclude that the questionnaire appears to be suitable for ranking individuals according to their consumption of fruits and vegetables and according to changes in their consumption. However, the validity of the questionnaire remains to be established in males, other age groups, and populations of lower educational levels.