Validation of a semiquantitative food frequency questionnaire for estimation of intakes of energy, fats and cholesterol among Singaporeans

Publisher Website
Google Scholar
Abstract
The aim of this study was to assess the relative validity of a 159‐item semiquantitative food frequency questionnaire (FFQ) for use among adult Singaporeans. This FFQ should be able to classify intakes of energy, total fat, saturated fat, polyunsaturated fat, monounsaturated fat and cholesterol into quintiles of intakes for purposes of epidemiological studies. A total of 126 subjects (84 women and 42 men) took part in the study which included an interview using the newly developed FFQ (assess past month’s intake) and three 24‐h (24 h) recalls (reference method, collected over a period of 1 month). Subjects also collected two 24 h urinary samples for urea from which total nitrogen excretion was assessed to validate protein intake. When compared to the reference method, the FFQ slighty overestimated the intakes of energy, total fat and types of fat as reflected by the difference in means and the ratio of FFQ to 24 h intakes. The overestimation ranged from 1 to 11% of the reference method. Dietary cholesterol was underestimated by 17% by the FFQ. These differences were however, not statistically significant. Pearson’s correlation coefficients (95% confidence intervals (CI)) between intakes assessed by FFQ and reference method varied from 0.58 (0.45, 0.69) for total fat to 0.39 (0.23, 0.53) for polyunsaturated fat. Cross‐classification into quintiles resulted in correct classification into the same or adjacent quintiles in 70% of subjects, with only one or two subjects being grossly misclassified. Nitrogen (N) intake from the 24 h recalls did not differ significantly from that estimated from the urinary nitrogen excretion. The mean (± SD) difference was 0.0 ± 0.4 g and the Pearson correlation coefficient (95% CI) was 0.55 (0.31, 0.72). It is concluded that the newly developed FFQ is an adequate tool for classifying individuals’ intakes into quintiles for epidemiological studies among Singaporean adults.