a1 Department of Nutrition and
a2 Department of Medical Statistics, Institute of Basic Medical Sciences, University of Oslo, PB 1046 Blindern, N-0316, Oslo, Norway
a3 Directorate for Health and Social Affairs, Department for Nutrition, Oslo, Norway
The validity of fruit and vegetable intake estimated by 14 d weighed records, a twenty-seven-item food frequency questionnaire (FFQ) and a 180-item FFQ was investigated using serum carotenoids as the validity criterion. In addition, the method of triads was used to assess the validity of fruit and vegetable intake estimated from the FFQ and serum carotenoids. One hundred Norwegian men completed 14 d weighed records and the 180-item FFQ. Eighty-six of them also completed the twenty-seven item FFQ. The partial correlation coefficients between serum carotenoids and fruit and vegetable intake were slightly higher for the 14 d weighed records than for the two FFQ, but no difference was observed between the 180- and the twenty-seven item FFQ. The strongest correlations were observed between vegetable intake and serum α-carotene. The highest validity coefficients (VC) were observed for vegetable intake estimated from weighed records, the 180-item FFQ, the twenty-seven item FFQ and by the biomarker serum α-carotene, with VC of 0·77, 0·58, 0·51 and 0·67, respectively. In conclusion, the short FFQ gave as valid estimates for fruit and vegetable intake as the long FFQ. Both the estimated partial correlation coefficients and VC suggest that serum α-carotene is the best biomarker for intake of vegetable alone and total intake of fruit and vegetables in this population of Norwegian men, but the biomarkers did not perform any better than the FFQ.
(Received February 23 2004)
(Revised November 11 2004)
(Accepted November 19 2004)