How does a simplified recipe collection procedure in dietary assessment tools affect the food group and nutrient intake distributions of the population

Abstract

AbstractTechnology advancements have driven the use of self-administered dietary assessment methods in large-scale dietary surveys. Interviewer-assisted methods generally have a complicated recipe recording procedure enabling the adjustment from a standard recipe. In order to decide if this functionality can be omitted for self-administered dietary assessment, this study aimed to assess the extent of standard recipe modifications in the Dutch National Food Consumption Survey and measure the impact on the food group and nutrient intake distributions of the population when the modifications were disregarded. A two-scenario simulation analysis was conducted. Firstly, the individual recipe scenario omitted the full modifications to the standard recipes made by people who knew their recipes. Secondly, the modified recipe scenario omitted the modifications made by those who partially modified the standard recipe due to their limited knowledge. The weighted percentage differences for the nutrient and food group intake distributions between the scenarios and the original data set were calculated. The highest percentage of energy consumed through mixed dishes was 10 % for females aged 19–79 years. Comparing the combined scenario and the original data set, the average of the absolute percentage difference for the population mean intakes was 1·6 % across all food groups and 0·6 % for nutrients. The soup group (−6·6 %) and DHA (−2·3 %) showed the largest percentage difference. The recipe simplification caused a slight underestimation of the consumed amount of both foods (−0·2 %) and nutrients (−0·4 %). These results are promising for developing self-administered 24-hour recalls or food diary applications without complex recipe function.