A new reference method for the validation of the nutrient profiling schemes using dietary surveys.

Nutrient profiles of foods are increasingly used as the scientific basis of nutritional labeling, health claims, or nutritional education. Nutrient profiling schemes are based on sets of rules, scores, or thresholds applied to the nutritional composition of foods. However, there is a lack of scientific validation of nutritional profiling schemes. To develop a reference method using existing dietary surveys, to define a set of indicator foods that are positively or negatively associated with a "healthy diet." Such indicator foods can be used both for establishing relevant nutrient profiles and for the validation of existing or future nutrient profiling schemes. The proposed validation method is based on food and nutrient intakes of adults participating in national dietary surveys in five EU countries: Belgium (n = 2,507), Denmark (n = 3,151), France (n = 1,474), Ireland (n = 1,379), and Italy (n = 1,513). The characterization of indicator foods is divided in two steps. First, "healthy diets" of individuals are identified in the five national dietary surveys by comparison to the Eurodiet reference intakes. Second, indicator foods associated positively or negatively to the "healthy diets" are determined. With a P-value of 10(-3) for the test of comparison of food intakes between the "most healthy eaters" and the "less healthy eaters," it was possible to identify 294 indicator foods out of 1,669 foods tested in the five countries. In all the countries except Italy, there were more indicator foods positively associated than indicator foods negatively associated with the "healthy diet." The food categories of these indicator foods were in good agreement with Food Based Dietary Guidelines like the USDA dietary guideline for Americans. A new reference method for the validation of profiling schemes was developed based on dietary intake data from using dietary surveys in five European countries. Only a minority of foods consumed in these dietary surveys could be used as indicator foods of healthy or unhealthy diets in order to subsequently test nutritional profiling schemes. Further work is needed to build a list of indicator foods that could be considered as a "gold standard."

Comparison of different nutrient profiling schemes to a new reference method using dietary surveys.

A new EU regulation on nutrition and health claims made on foods has entered into force in January 2007. The regulation provides for the use of nutrient profiles to determine which foods may bear claims but does not specify what the profiles should be or how they should be developed. Several nutrient profiling schemes have already been established. Therefore, it is necessary to develop approaches to test if the existing profiling schemes could fulfil the new regulation needs. The aim of the present study is to investigate how reference "indicator foods" derived from national dietary surveys in five different countries, are classified according to three existing nutrient profiling schemes: The UK Food Standards Agency (FSA) model, The Dutch Tripartite classification model and the US FDA model used for regulating health claims. "Indicator foods" that have been shown to be positively or negatively associated with healthy diets in adults in five EU countries were classified according to each of the three profiling schemes. The performance and effectiveness of each profiling scheme in correctly classifying the "indicator foods" were assessed using sensitivity and specificity ratios. The sensitivity and the specificity ratios of the three profiling schemes tested were relatively good. There were only small differences of performance between the three systems. A significant negative correlation between sensitivity and specificity was observed. The level of concordance between the classification of the "indicator foods" that have been selected because of being positively or negatively associated with a healthy diet and the classification by each of the three profiling methods tested was quite good. However, further improvement of the "indicator foods" approach is needed if it is to serve as a "gold standard".