Development and Validation of a Nutrient Profiling Model for Shopping Baskets: The Grocery Basket Score Methodology.

BACKGROUND

Rating a person's diet as a whole, as opposed to rating individual food items, could help better inform consumers about the health value of their diet.

OBJECTIVES

Our goal was to develop an automated health rating of grocery shopping baskets, based on the nutrient composition of all the food items contained within the basket, and without knowing how many people consume the food items over what period of time. It is envisaged that the rating, or score, can be deployed by retailers on top of existing loyalty or digital shopping basket programs.

METHODS

We developed a novel model for calculating Grocery Basket Scores (GBS) that uses nutrient energy densities rather than absolute quantities. It was based on self-reported daily food intake from the National Health and Nutrition Examination Survey as well as mortality follow-up data. We conducted a validation against the Alternate Healthy Eating Index (AHEI) as well as the Nutri-Score.

RESULTS

The nutrient energy density (GBS) model penalized consuming calories from sugar and saturated fats, high salt intake, as well as consuming calories from beverages. Furthermore, a penalty was applied to foods that have a low ratio of protein, vitamin C, or iron relative to the total number of calories. Fiber consumption was rewarded. The model showed a high degree of correlation with the AHEI (Pearson correlation coefficient: 0.62 for the AHEI without the protective effect of moderate alcohol consumption and 0.60 with it) and the Nutri-Score (PCC: -0.60).

CONCLUSIONS

The proposed nutrient energy density model is aligned with the recommendations formulated in nutritional guidelines, as indicated by a high degree of correlation with the AHEI and Nutri-Score. Long-term use of a GBS from a single or multiple grocery stores could help consumers adhere to dietary guidelines.