Validation of a Food Frequency Questionnaire Based on Food Groups for Estimating Individual Nutrient Intake.

A food frequency questionnaire (FFQg) was developed, based on 29 food groups and 10 kinds of cookery, for estimating the energy and nutrient intakes of an individual subject during the previous one to two months. This questionnaire was evaluated by a comparison with weighed dietary records for 7 continuous days (7d records) of 66 subjects aged 19-60 years.The correlation coefficients between the FFQg and 7d records for energy, protein, fat, carbohydrate and calcium intakes were 0.47, 0.42, 0.39, 0.49 and 0.41, respectively. The intakes of 26 of the 31 nutrients were not significantly different by paired t-tests between the two methods (p≥0.05). The ratio of the value obtained by the FFQgmethod against that by the 7d records method ranged from 72%(vitamin B12) to 121%(vitamin C). The average value of this ratio was 104%.The correlation coefficients for the intakes of rice, bread, meat, fish, milk, dairy products, green-yellow vegetables, other vegetables, and fruits were 0.66, 0.76, 0.27, 0.27, 0.72, 0.58, 0.46, 0.53, and 0.64, respectively, between the FFQg and 7d records, and there was significant correlation between the two methods at the p<0.05 level for 22 of the 29 food groups.Apart from those food groups for which “Less than once a month or never” was selected by 50% or more of the respondents, 34% of respondents could estimate portion sizes in the FFQg with an error of less than 25%. The results show that the FFQg record is a useful instrument for estimating individual energy and nutrient intakes.