Systems rank or rate foods in nutritional quality by using nutrient criteria that are either
- across the board, the same for all foods, or
- food category-specific, different for food categories (Table 2-1).
Most systems are category specific and are meant to compare the nutritional quality of a food to other foods in the same category. The number of food categories used in category-specific criteria varies; some systems use only a few categories and others use over 80 categories (e.g., Heart Check). The category-specific systems may reflect dietary guidelines, such as the recommendation to consume low-fat or nonfat milk instead of full-fat products.
An across-the-board approach allows for the comparison of foods across categories, although it also has the advantage of allowing within-food category comparisons. Another advantage of across-the-board systems is that they are simple because they have only one set of nutrient criteria.
Food category-specific systems incorporate intrinsic properties of foods that may be specific to a food group. For example, vitamin B12 is only in animal-based foods, so the nutrient criteria for milk may include vitamin B12, but B12 would not be included in the criteria for fruits. The consensus of participants in an International Life Sciences Institute (ILSI) Europe workshop on nutrient profiling was that a category-specific approach would be preferred, because this approach addresses the intrinsic differences in foods (Tetens, Oberdorfer, Madsen, & de Vries, 2007). A disadvantage of the approach is that it can be complex, because it may be difficult to categorize some foods, and different nutrient criteria thresholds must be developed for each category.
To assess whether a nutrient profiling system should be generic or category specific, Scarborough et al. (2010) compared the diets of participants in a British dietary survey with average scores using the algorithm from the UK Ofcom Nutrient Profiling Model, referred to as a WXYfm score. It was determined that to improve their diets similar to the level of the healthy-diet consumers, consumers of unhealthy diets would have to both consume healthier versions of foods within food categories and consume greater amounts of healthy foods. Individuals with a healthy diet consumed healthier versions of foods in the following categories: meat, dairy (except cheeses), bread (with minimal fiber), cereal products (except breakfast cereals), ready meals/soups/sandwiches, and breakfast cereals. These results suggest that a category-specific system may be preferable for nutrient profiling models designed to promote healthy diets. The authors suggest that these results could be used as a basis for deciding which food categories should be included in a system, that is, only those foods for which evidence exists that individuals with a healthy diet consume healthier versions of foods within those categories than those with an unhealthy diet. The results of this study also suggest that a limited number of food categories are probably necessary. As the authors state, "it would seem unlikely that people who achieve a healthy diet do so by consuming healthier versions of ice cream than those who consume an unhealthy diet" (p. 6).
Food category-specific systems created by a single manufacturer may be of limited use in general because they are likely designed to fit their product line and may not allow for comparison to other brands in that category. Other brand products in a category may have another profiling system using different criteria or may not have a profiling system at all.