Browsing by Author "Juntunen, Iiris"

Alternative research methods are needed to complement traditional dietary assessment methods in risk assessment, because traditional methods have bias related to misreporting and a short period of monitoring. The aim was to study whether food purchase data could be used in risk assessment to complement traditional food consumption data to estimate food intake. The study compared statistics from two different Finnish studies. This substudy of LoCard is based on the original grocery purchase data, a subsample provided by consenting retailer loyalty card holders (n=15 756) from 2017. FinDiet 2017 is a national dietary survey with individual level data based on two 24-h-dietary recalls (n=1 655). Four food groups were selected from both data sets: almonds and nuts, bananas, apples, and eggs. Post-stratification weights to correct selection bias for LoCard data were used and the degree of loyalty was taken into account. FinDiet data showed higher consumption of almonds and nuts, and apples than LoCard data. In LoCard data, banana consumption was relatively high compared to FinDiet data. When using post-stratification weights, the magnitude of differences did not change significantly. In conclusion, grocery purchase data cannot be used in risk assessment to estimate food intake as such, because bodyweight information is missing. Thus, consumption estimates per bodyweight cannot be executed. However, when combined with other food consumption data, purchase data have the potential to remediate some limitations of self-reported food consumption, such as misreporting and a short period of monitoring. Food purchase data also add timely information on population consumption and special dietary products. Future studies are needed to evaluate how food purchase data could practically be used in risk assessment process.