Role of different nutrient profiling models in identifying targeted foods for front-of-package food labelling in Brazil

Abstract

Objective: To compare the degree of strictness and agreement of different nutrient profiling models (NPM) used to identify which foods would be required to show front-of-package (FOP) warning labels. Design: Using data of 11 434 packaged foods found in the five largest food retailers in Brazil, we used two published NPM: the Pan American Health Organization (PAHO) model and the NPM used in the Chilean nutritional FOP labelling policy, and compared them with a NPM proposed by the Brazilian National Health Surveillance Agency (Anvisa). The proportion of foods that would be required to show FOP warning labels was calculated overall and by food category. We also tested whether a modified version of the PAHO NPM would behave similarly to the original version. Setting: Brazil. Results: Two-thirds of the packaged products (62 %) would receive FOP warning labels under the PAHO NPM, as compared with 45 % of products using the proposed Anvisa NPM and 41 % if the Chilean NPM was applied. The PAHO NPM identified more foods high in critical nutrients such as sweetened dairy and non-dairy beverages, canned vegetables and convenience foods. Overall agreement between models was considered good with kappa coefficient ranging from 0.57 to 0.92 but was lower for some food categories. Conclusions: We found variations in the degree of strictness and agreement between assessed NPM. The PAHO NPM identified more foods and beverages high in sugar which are among the top contributors to sugar and energy intake in Brazil.