The validity was assessed of the “Dietary Record by Cooked Dishes” (DR
cd) method, a newly developed and simplified self-reported dietary record, together with a smaller sample size for constructing the dish-based component table.
Scale-based one-day dietary records were compiled by 224 subjects from 2000 to 2002, the amounts of nutrients in each dish were calculated by using the 5th standard food composition table, and the results were used to construct the dish-based component table (DR
cd224). Databases for DR
cd125 and DR
cd75, which respectively used 125 and 75 samples from the 224 subjects randomly selected 5 times from each, were next constructed in the same way. A one-day dietary survey using the DR
cd databases was finally conducted on 106 subjects in 2003 and 2004 to assess the validity of the method.
The median values of Pearson's correlation coefficients for the intake of energy, of 29 nutrients, and of 21 food groups between DR and DR
cd224 were 0.808 and 0.921, respectively. The correlation coefficients of the minimum and maximum values of the 5-times calculation for DR
cd125 were 0.701 and 0.754 (nutrients), and 0.685 and 0.796 (food groups), and for DR
cd75 were 0.686 and 0.724, and 0.662 and 0.726, respectively. The ratio of the energy intake assessed by DR
cd224 to that by DR was 92.2%, and the ratios of the median values for nutrients assessed by DR
cd125 and DR
cd22275 to DR
cd224 were 97.5-102.8% and 99.2-104.8%, respectively.
Although the correlation coefficients between DR and DR
cd224 were higher than those for DR
cd125 and DR
cd75, there was little effect from this difference in estimating the dietary intake. A sample size of 75 would therefore be practical for constructing DR
cd in field research.
View full abstract