Research Updates | Other
Total sugars are currently used in the HSR algorithm, despite the fact these include natural intrinsic sugars in dairy, fruits and vegetables. These Australian researchers wanted to explore whether the HSR would better discriminate between core and discretionary foods (based on the Australian Dietary Guidelines) by using added sugars data instead.
The modelling was based on nutrition data from the Australian FoodSwitch database and added sugar data from food composition databases. The ability of each of the nutrients in the HSR algorithm (as well as added sugar) to discriminate between core and discretionary foods was estimated using area under the curve (AUC). 15,9654 core and 18,350 discretionary foods were included. Fifty two percent (52%) of core foods and 87% of discretionary foods contained added sugar. The median HSR for core foods was 4.0 and 2.0 for discretionary foods. Median added sugar content was 3.3g/100g for core foods and 14.6g/100g for discretionary foods.
Of all the nutrients in the HSR algorithm, total sugar (not added) had the greatest capacity to discriminate between core and discretionary foods (AUC 0.692). A model using all nutrients in the current HSR has an AUC of 0.817 which increased to 0.871 (a 7% improvement) with inclusion of added sugar. The authors conclude the current HSR discriminates well between core and discretionary foods, but it would improve with addition of added sugar.
Peters SAE, Dunford E, Jones A et al. Incorporating added sugar improves the performance of the front-of-pack labelling system in Australia. Nutrients 2017; 9(7)