Validation of a quantitative food frequency questionnaire for Inuit population in Nunavut, Canada

Abstract

Validation of a quantitative food frequency questionnaire (QFFQ) developed specifically for Inuit is necessary to determine its usefulness in assessing dietary intake and adequacy and in identifying dietary risk factors for chronic disease in this population. Seventy-five randomly selected Inuit adults in Nunavut, Canada, were recruited. Mean daily intake of nutrients from one to three 24-h recalls was used as the reference to measure QFFQ validity. Crude and energy-adjusted Spearman rank correlations (ρ), cross classification and weighted kappa were computed as measures of concordance. Bland-Altman plotting was used for additional assessment. Excluding four participants with daily energy intake of >25.1 MJ, 71 participants were included in the analysis. For all nutrients, mean daily intake from the QFFQ was higher than the recall. ρ's for macronutrients were in the range 0.71 for carbohydrate to 0.25 for protein. The best ρ amongst micronutrients was observed for vitamin C (0.66). Overall correlation between the two dietary tools improved after correction for within-person variance (from 0.46 to 0.49), although adjusting for energy did not improve the overall coefficient. When nutrient intakes were categorised into quartiles, the QFFQ and 24-h recalls indicated relative agreement proportion (same or adjacent quartiles) of 83% for energy, 94% for total sugar, 83% for macronutrients and 77% for micronutrients. Bland-Altman plots showed a tendency for increased scatter of the differences in nutrients at higher intakes. The QFFQ developed is valid and can be used to assess usual dietary intake and dietary adequacy, determine the contribution of foods to specific nutrient intakes, and identify dietary risk factors for chronic disease amongst this population.