Abstract
BACKGROUND: Governments globally aim to reduce the intake of unhealthy foods. Many policies exist that aim to address foods high in saturated fat, salt and sugar (HFSS) but the identification of ultra-processed foods (UPF) have presented a greater challenge due to the lack of an appropriate policy definition. To support policymakers, we provide approaches that can support governments to identify both HFSS foods and UPFs. METHODS: Four approaches combining elements of UPF definitions (i.e., presence of additives) and HFSS definitions were compared attempting to simplify and standardize the identification of less healthy products. Nationally representative food purchase data from NielsenIQ linked with nutrition facts label data were used to examine the mean proportion of product volume purchased by US households to be targeted. Differences between approaches were examined using Student t test; Bonferroni adjusted P value < 0.0001 was considered significant. FINDINGS: In 2020, 50% of 33,054,687 products purchased by US households were considered UPFs (65% of foods and 38% of beverages) and 43% HFSS (65% of foods and 26% of beverages), however there was not 100% agreement between the two definitions (P < 0.0001). By starting with HFSS criteria and adding elements of UPF (colors and flavors), we were able to provide a method with 100% agreement between the identification of UPFs and HFSS products. INTERPRETATION: Results demonstrated how combining HFSS criteria with UPF criteria can be used to identify less healthy foods and ensure policymakers have both a simple and accurate method to target products for policy intervention. FUNDING: Bloomberg Philanthropies and the Global Food Research Program of UNC-Chapel Hill provided funds.