Diet, nutrition intake, and metabolism in populations at high and low risk for colon cancer. Concordance of nutrient information obtained by different methods.

Data obtained from two vegetarian and two nonvegetarian populations by 24-h diet recall, 3-day weighed diet diary, analysis of 3-day composite food samples, and measurement of certain biochemical parameters were compared. Correlations for group means (both sexes combined) for values obtained by 24-h recall versus 3-day diet diary were protein, r = 0.96 and total calories r = 0.86. For 3-day diet diary with food analysis, protein r = 0.89, fat r = 0.90 and for reported protein intake with serum prealbumin r = 0.74, with urinary urea nitrogen r = 0.87. Correlations were low for comparisons of means of male groups for most parameters. Individual correlations for reported protein intake and biochemical parameters were lower for males than for females. This suggests that male subjects may require special attention to insure accuracy of reporting and specimen collection. These findings indicate comparability of data from 24-h recall and 3-day diary, and agreement of dietary records with certain biochemical parameters. Protein and fat intake as determined from chemical analysis of a composite food sample correlated well (r = 0.89; r = 0.90) with that calculated from food tables.