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

Criterion validation of nutrient profiling systems: a systematic review and meta-analysis

BACKGROUND

Nutrient profiling systems (NPSs) use algorithms to evaluate the nutritional quality of foods and beverages. Criterion validation, which assesses the relationship between consuming foods rated as healthier by the NPS and objective measures of health, is essential to ensure the accuracy of NPSs.

OBJECTIVE

We examined and compared NPSs that have undergone criterion validity testing in relation to diet-related disease risk and risk markers.

METHODS

Academic databases were searched for prospective cohort and cross-sectional studies published before November, 2022. NPSs were eligible if they incorporated multiple nutrients or food components using an algorithm to determine an overall summary indicator (e.g., a score or rank) for individual foods. Studies were included if they assessed the criterion validity of an eligible NPS. Validation evidence was first summarized in narrative form by NPS, with random effects meta-analysis where ≥2 prospective cohort studies assessed the same NPS and outcomes.

RESULTS

Of 4519 publications identified, 29 describing 9 NPSs were included in the review. The Nutri-Score NPS was assessed as having substantial criterion validation evidence. Highest compared with lowest diet quality as defined by the Nutri-Score was associated with significantly lower risk of cardiovascular disease (hazard ratio [HR]: 0.74; 95% confidence interval [CI]: 0.59, 0.93; n = 6), cancer (HR: 0.75; 95% CI: 0.59, 0.94; n = 5), all-cause mortality (HR: 0.74; 95% CI; 0.59, 0.91; n = 4) and change in body mass index (HR: 0.68; 95% CI: 0.50, 0.92; n = 3). The Food Standards Agency NPS, Health Star Rating, Nutrient Profiling Scoring Criterion, Food Compass, Overall Nutrition Quality Index, and the Nutrient-Rich Food Index were determined as having intermediate criterion validation evidence. Two other NPSs were determined as having limited criterion validation evidence.

CONCLUSIONS

We found limited criterion validation studies compared with the number of NPSs estimated to exist. Greater emphasis on conducting and reporting on criterion validation studies across varied contexts may improve the confidence in existing NPSs.

Defining Energy-Dense, Nutrient-Poor Food and Drinks and Estimating the Amount of Discretionary Energy

Overconsumption of energy provided by energy-dense, nutrient-poor (EDNP) food and drinks increases the risk of unhealthy weight gain and of obesity-related health outcomes. The aim of this study was to develop a nutrient profiling model for classifying EDNP food and drinks and to estimate the amount of discretionary energy for EDNP food and drinks in a recommended diet. A stepwise approach was used first to develop a nutrient profiling model for classifying EDNP food and drinks and then to estimate the amount of discretionary energy in a recommended diet using diet modeling. The nutrition profiling model comprised 24 macro- and micronutrients and energy density. The model classified 67% of 1482 foods and 73% of 161 drinks correctly as EDNP food and drinks compared with an expert-adjusted model. Sweets, chocolate, cake, cookies and biscuits, sweet and salty snacks, sugar-sweetened and artificially sweetened drinks, and alcoholic drinks were classified as EDNP food and drinks. The estimated amount of discretionary energy for EDNP food and drinks was 4–6% of the energy requirements for 4–75-year-old Danes. It seems prudent to have special attention on EDNP food and drinks in dietary guidelines and future public health initiatives to avoid overconsumption of energy.

Nutrition and health or nutrients and health?

Diet is an important contributor to human health and public health bodies are issuing guidelines aimed at favouring healthy food choices. The aim of our paper is to discuss the aspects underlying the concept of nutrient profiles, that is, defining levels of energy, some macronutrients, or salt which should not be exceeded in individual foods, according to the available evidence, to help in understanding to what extent such approach may actually be useful for improving nutrition and quality of life of European consumers. We list several pitfalls and oversimplifications of the current approaches to nutrient profiling and of the dichotomic classification of foods into "healthy" and "unhealthy" products. In view of the current "Facilitating healthier food choices - establishing nutrient profiles" EU initiative, we believe that further debate among all stakeholders is warranted and must consider all the limitations outlined in this paper.

Nutrition Quality of Packaged Foods in Bogotá, Colombia: A Comparison of Two Nutrient Profile Models

BACKGROUND

Front-of-package warning labels are an increasingly common policy and have been implemented to inform consumers of the nutritional quality of ultra-processed foods. This study examined the proportion of Colombian products that could be subjected to such regulations.

METHODS

Two nutrition profile models were compared: the Pan American Health Organization (PAHO) model, and the nutrient profile established under the Chilean food labeling and advertising law (Chilean model). Products (n = 6708) exceeding nutrient criteria based on each model were identified as subject to regulation.

RESULTS

A total of 80.2% (PAHO model) to 66.4% (Chilean model) of foods met the criteria for regulation. The categories with the highest proportion of regulated products were meats (97.3% PAHO model; 87.5% Chilean model), sweets (95.6% PAHO) and snacks (Chilean model). The category with the lowest proportions of regulated products were cereals (47.3% PAHO model) and miscellaneous foods and fish/seafood (39.0% and 39.5%, respectively, Chilean model).

CONCLUSIONS

Under both the PAHO and Chilean nutrient profile models, the majority of packaged foods available in Bogotá would be eligible to receive front-of-package warning labels. These results suggest a warning label law could have a major impact on the Colombian food supply, especially in the context of the growing prevalence of diet-related chronic diseases in Colombia.

Voluntary policies on checkout foods and healthfulness of foods displayed at, or near, supermarket checkout areas: a cross-sectional survey

OBJECTIVE

To determine if voluntary policies on supermarket checkout foods are associated with a difference in the healthfulness of foods displayed at, or near, supermarket checkout areas.

DESIGN

Cross-sectional survey of foods at, or near, supermarket checkouts categorised as less healthy or not according to the Food Standards Agency's Nutrient Profiling Model.

SETTING

One city in Eastern England (population about 125 000).

SUBJECTS

All stores in nine supermarket groups open for business in June-July 2017 in the study city. Supermarket checkout food policies were categorised as clear and consistent, vague or inconsistent, or none.

RESULTS

In thirty-three stores, 11 434 checkout food exposures were recorded, of which 8010 (70·1 %) were less healthy; and 2558 foods in areas near checkouts, of which 1769 (69·2 %) were less healthy. After adjusting for a marker of store size, the odds of a checkout food exposure being 'less healthy' was lower in stores with vague or inconsistent checkout policies (OR=0·63; 95 % CI 0·49, 0·80) and in stores with clear and consistent checkout policies (OR=0·33; 95 % CI 0·24, 0·45), compared with no policy. There was no difference in the odds of foods near, but not at, checkouts being less healthy according to checkout food policy.

CONCLUSIONS

Supermarket checkout food policies were associated with lower odds of checkout foods but not foods near, but not at, checkouts being less healthy. Further research is required to explore impacts on purchasing and consumption.

The nature of UK supermarkets' policies on checkout food and associations with healthfulness and type of food displayed: cross-sectional study

Background

Food choices are often determined by stimuli from our immediate surroundings, including strategic placement in shops to encourage impulse purchases. One example of this is food in shop checkout areas. Recently a number of UK supermarkets have voluntarily committed to providing healthier checkout foods. The aim of this study was to document the nature of current UK supermarket checkout food policies; determine whether there are any differences in the healthfulness and type of food displayed at checkouts in supermarkets according to the presence or nature of policies; and determine whether supermarkets are adhering to their checkout food policies.

Methods

Survey of checkout food policies. Cross-sectional observations in 69 supermarkets (covering 14 store formats) in the East of England in Feb-May 2017 of the number and type of checkout foods on each ‘checkout journey’ (each possible route through the checkout area). Checkout foods were categorised as less healthy or healthier, using the UK Food Standard’s Agency’s Nutrient Profile Model, and into food groups. Checkout food policies were categorised as clear and consistent, vague or inconsistent, or absent.

Results

Checkout food policies differed between store formats in some supermarket groups. Across the 14 store formats included, two had no checkout food policy, six had ‘clear and consistent’ policies, and six ‘vague or inconsistent’ policies. In supermarkets with clear and consistent policies there were a median of 13 products per checkout journey, of which 35% were less healthy. Comparable figures for supermarkets with vague or inconsistent, and absent policies were 15 (57%) and 39 (90%) respectively (ps for trend < 0.001).

Whilst most supermarkets with a clear and consistent checkout food policy were fully adherent to their policy, those with vague or inconsistent policies were not.

Conclusions

Most UK supermarkets have checkout food policies, but not all are clear and consistent. Supermarkets with clear and consistent policies display fewer checkout foods and a lower proportion of these are less healthy than in other supermarkets. Supermarkets with clear and consistent policies adhere well to these. More stores should be encouraged to develop a clear and consistent checkout food policy. This may require non-voluntary intervention.

Electronic supplementary material

The online version of this article (10.1186/s12966-018-0684-2) contains supplementary material, which is available to authorized users.

Nutrient profiling and adherence to components of the UK national dietary guidelines association with metabolic risk factors for CVD and diabetes: Airwave Health Monitoring Study

CVD is the leading cause of death worldwide. Diet is a key modifiable component in the development of CVD. No official UK diet quality index exists for use in UK nutritional epidemiological studies. The aims of this study are to: (i) develop a diet quality index based on components of UK dietary reference values (DRV) and (ii) determine the association between the index, the existing UK nutrient profile (NP) model and a comprehensive range of cardiometabolic risk markers among a British adult population. A cross-sectional analysis was conducted using data from the Airwave Health Monitoring Study (n 5848). Dietary intake was measured by 7-d food diary and metabolic risk using waist circumference, BMI, blood lipid profile, glycated Hb (HbA1c) and blood pressure measurements. Diet quality was assessed using the novel DRV index and NP model. Associations between diet and cardiometabolic risk were analysed via multivariate linear models and logistic regression. A two-point increase in NP score was associated with total cholesterol (β -0·33 mmol/l, P<0·0001) and HbA1c (β -0·01 %, P<0·0001). A two-point increase in DRV score was associated with waist circumference (β -0·56 cm, P<0·0001), BMI (β -0·15 kg/m2, P<0·0001), total cholesterol (β -0·06 mmol/l, P<0·0001) and HbA1c (β -0·02 %, P=0·002). A one-point increase in DRV score was associated with type 2 diabetes (T2D) (OR 0·94, P=0·01) and obesity (OR 0·95, P<0·0001). The DRV index is associated with overall diet quality and risk factors for CVD and T2D, supporting its application in nutritional epidemiological studies investigating CVD risk in a UK population.

Testing Two Nutrient Profiling Models of Labelled Foods and Beverages Marketed in Turkey

AIM

The objective of this study was to evaluate the nutrient profile of labelled foods and also understand the application of two international nutrient profiling models of labelled foods and beverages.

METHODS

WXYfm and NRF 9.3 nutrient profiling models were used to evaluate 3,171 labelled foods and beverages of 38 food categories and 500 different brands.

RESULTS

According to the WXYfm model, pasta, grains and legumes and frozen foods had the best scores whereas oils had the worst scores. According to the NRF 9.3 model per 100 kcal, the best scores were obtained for frozen foods, grains and legumes and milk products whereas the confectionery foods had the worst scores. According to NRF 9.3 per serving size, grains and legumes had the best scores and flavoured milks had the worst scores. A comparison of WXYfm and NRF 9.3 nutrient profiling models ranked scores showed a high positive correlation (p=0.01).

CONCLUSIONS

The two nutrient models evaluated yielded similar results. Further studies are needed to test other category specific nutrient profiling models in order to understand how different models behave.

Construct and criterion-related validation of nutrient profiling models: A systematic review of the literature

Nutrient profiling (NP) is defined as the science of ranking foods according to their nutritional composition for the purpose of preventing disease or promoting health. The application of NP is ultimately to assist consumers to make healthier food choices, and thus provide a cost effective public health strategy to reduce the incidence of diet-related chronic disease. To our knowledge, no review has assessed the evidence to confirm the validity of NP models. We conducted a systematic review to investigate the construct and criterion-related validity of NP models in ranking food according to their nutritional composition for the purpose of preventing disease and promoting health. We searched peer-reviewed research published to 30 June 2015 and used PUBMED, Global Health (CABI), and SCOPUS databases. Within study bias was assessed using an adapted version of the QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies -2) tool for all diagnostic studies and the Cochrane Collaboration's Risk of Bias tool for all non-diagnostic studies. The GRADE (Grades of Recommendation, Assessment, Development, and Evaluation) approach was used to guide our judgement of the quality of the body of evidence for each outcome measure. From a total of 83 studies, 69 confirmed the construct validity of NP models; however most of these studies contained methodological weaknesses. Six studies used objective external measures to confirm the criterion-related validity of NP models; which inherently improved quality. The overall quality of evidence on the accuracy of NP models was judged to be very low to moderate using the GRADE approach. Many carefully designed studies to establish both construct and criterion-related validity are necessary to authenticate the application of NP models and provide the evidence to support the current definition of NP.

Food at checkouts in non-food stores: a cross-sectional study of a large indoor shopping mall

OBJECTIVE

To investigate the display of food at non-food store checkouts; and to classify foods by type and nutrient content, presence of price promotions and whether food was at child height.

DESIGN

Cross-sectional survey of checkout displays at non-food stores. Foods were classified as 'less healthy' or healthier using the UK Food Standards Agency's Nutrient Profile Model. Written price promotions were recorded. Child height was defined as the sight line of an 11-year-old approximated from UK growth charts.

SETTING

A large indoor shopping mall, Gateshead, UK, February-March 2014.

SUBJECTS

Two hundred and five out of 219 non-food stores in the shopping mall directory which were open for trading.

RESULTS

Thirty-two (15·6%) of 205 non-food stores displayed food at the checkout. All displayed less healthy foods, and fourteen (43·8%) had healthier foods. Overall, 5911 checkout foods were identified. Of these, 4763 (80·6%) were 'less healthy'. No fruits, vegetables, nuts or seeds were found. Of 4763 less healthy foods displayed, 195 (4·1%) were subject to price promotions, compared with twelve of 1148 (1·0%) healthier foods (χ 2(df=1)=25·4, P<0·0001). There was no difference in the proportion of less healthy (95·1%) and healthier (96·2%) foods displayed at child height.

CONCLUSIONS

Almost one-sixth of non-food stores displayed checkout food, the majority of which was 'less healthy' and displayed at child height. Less healthy food was more likely to be subject to a written price promotion than healthier food. Further research into the drivers and consequences of checkout food in non-food stores is needed. Public health regulation may be warranted.

The proportion of unhealthy foodstuffs children are exposed to at the checkout of convenience supermarkets

OBJECTIVE

To investigate the proportion of foods that are unhealthy to which children are exposed at the checkout of convenience supermarkets.

DESIGN

We performed a cross-sectional survey of foodstuffs displayed at the checkout. Products displayed at or below children's eye-level were designated as healthy, unhealthy or unclassifiable using the Food Standards Agency's scoring criteria.

SETTING

Thirteen convenience supermarkets from the three leading UK supermarket chains were selected on the basis of proximity to the town hall in Sheffield, England.

SUBJECTS

Convenience supermarkets were defined as branches of supermarket chains that were identified as being other than superstores on their company's store locator website.

RESULTS

In almost all of the convenience supermarkets surveyed, the main healthy product on display was sugar-free chewing gum. On average, when chewing gum was not included as a foodstuff, 89% of the products on display at the checkouts of convenience supermarkets were unhealthy using the Food Standards Agency's criteria. One store was a notable outlier, providing only fruit and nuts at its checkout.

CONCLUSIONS

The overwhelming majority of products to which children are exposed at the convenience supermarket checkout are unhealthy. This is despite all the supermarket chains surveyed having signed up to the UK Government's 'responsibility deal'.

Future body mass index modelling based on macronutrient profiles and physical activity

Background

An accurate system of determining the relationship of macronutrient profiles of foods and beverages to the long-term weight impacts of foods is necessary for evidence-based, unbiased front-of-the-package food labels.

Methods

Data sets on diet, physical activity, and BMI came from the Food and Agriculture Organization (FAO), the World Health Organization (WHO), the Diabetes Control and Complications Trial (DCCT), and Epidemiology Diabetes Intervention and Complications (EDIC). To predict future BMI of individuals, multiple regression derived FAO/WHO and DCCT/EDIC formulas related macronutrient profiles and physical activity (independent variables) to BMI change/year (dependent variable). Similar formulas without physical activity related macronutrient profiles of individual foods and beverages to four-year weight impacts of those items and compared those forecasts to published food group profiling estimates from three large prospective studies by Harvard nutritional epidemiologists.

Results

FAO/WHO food and beverage formula: four-year weight impact (pounds)=(0.07710 alcohol g+11.95 (381.7+carbohydrates g per serving)*4/(2,613+kilocalories per serving)–304.9 (30.38+dietary fiber g per serving)/(2,613+kilocalories per serving)+19.73 (84.44+total fat g)*9/(2,613+kilocalories per serving)–68.57 (20.45+PUFA g per serving)*9/(2,613+kilocalories per serving))*2.941–12.78 (n=334, R²=0.29, P < 0.0001). DCCT/EDIC formula for four-year weight impact (pounds)=(0.898 (102.2+protein g per serving)*4/(2,297+kilocalories per serving)+1.063 (264.2+carbohydrates g per serving)*4/(2,297+ kilocalories per serving)–13.19 (24.29+dietary fiber g per serving)/ (2,297+kilocalories per serving)+ 0.973 (74.59+(total fat g per serving–PUFA g per serving)*9/(2,297+kilocalories per serving))*85.82–68.11 (n=1,055, R²=0.03, P < 0.0001). (FAO/WHO+ DCCT/EDIC formula forecasts averaged correlated strongly with published food group profiling findings except for potatoes and dairy foods (n=12, r=0.85, P = 0.0004). Formula predictions did not correlate with food group profiling findings for potatoes and dairy products (n=10, r= −0.33 P=0.36). A formula based diet and exercise analysis tool is available to researchers and individuals: http://thehealtheconomy.com/healthTool/.

Conclusions

Two multiple regression derived formulas from dissimilar databases produced markedly similar estimates of future BMI for 1,055 individuals with type 1 diabetes and female and male cohorts from 167 countries. These formulas predicted the long-term weight impacts of foods and beverages, closely corresponding with most food group profiling estimates from three other databases. If discrepancies with potatoes and dairy products can be resolved, these formulas present a potential basis for a front-of-the-package weight impact rating system.

Potential effects of nutrient profiles on nutrient intakes in the Netherlands, Greece, Spain, USA, Israel, China and South-Africa

INTRODUCTION

Nutrient profiling is defined as the science of categorising foods based on their nutrient composition. The Choices Programme is a nutrient profile system with criteria that determine whether foods are eligible to carry a "healthier option" stamp. The Daily Menu Method which has been developed to evaluate these criteria is described here. This method simulates the change in calculated nutrient intakes which would be the result of consumers changing their diets in favour of food products that comply with the criteria.

METHODS

Average intakes of energy, trans fatty acids (TFA), saturated fatty acids (SAFA), sodium, added sugar and fibre were derived from dietary intake studies and food consumption surveys of 7 countries: The Netherlands, Greece, Spain, the USA, Israel, China and South Africa. For each of the key nutrients, these average intakes were translated into three Typical Daily Menus per country. Average intakes based on these three menus were compared with average intakes from three Choices Daily Menus. To compose the Choices Menus, foods from the Typical Menus that did not comply with the Choices criteria were replaced with foods that did comply and are available on the market.

RESULTS

Comparison of intakes from the Choices Menus with the survey data showed that calculated intakes of energy, SAFA, TFA, sodium and added sugar were reduced. Fibre intakes were increased. The size of the effect differed per country.

CONCLUSION

The Daily Menu Method is a useful means to predict the potential effects of nutrient profiles such as the Choices criteria, on daily nutrient intakes. The method can be applied internationally and confirms that the criteria of the Choices Programme are in line with the aim of the programme: to improve nutrient intakes in the direction of the recommendations.

Quantifying the contribution of foods with unfavourable nutrient profiles to nutritionally adequate diets

That 'all foods can fit' into a healthy diet is a long-standing principle of dietetic practice. The present study quantified the relative contributions of foods to encourage and foods to limit, using new techniques of individual diet optimisation and nutrient profiling. Individual foods from every food group were assigned to four nutrient profile classes based on the French SAIN,LIM system. Foods with the most favourable nutrient profiles were in class 1, and foods with the least favourable nutrient profiles were in class 4. An optimised diet that met the recommendations for thirty-two nutrients and that respected the existing eating habits was designed for each adult in the nationally representative 'Enquête Individuelle et Nationale sur les Consommations Alimentaires 1' dietary survey (n 1171). The relative proportions of the four nutrient profiling classes were assessed before and after the optimisation process. The contribution of fruits and vegetables, whole grains, milk and fish was significantly increased, whereas the contribution of refined grains, meats, mixed dishes, sugars and fats was decreased. The optimised diets derived more energy (30 v. 21 % in the observed diets) from class 1 foods and less energy (41 v. 56 %) from class 4 foods. They also derived a higher amount of class 1 foods (61 v. 51 %) and a lower amount of class 4 foods (22 v. 32 %). Thus, nutrient adequacy was compatible with the consumption of foods with an unfavourable nutrient profile (one-fifth the basket weight), provided that the diet also contained almost two-thirds of foods with the most favourable profile. Translating these results into concrete and quantified advice may have very tangible public health implications.

Nutritional profiles in a public health perspective: a critical review

Nutritional profiling is defined as 'the science of categorizing foods according to their nutritional composition' and it is useful for food labelling and regulation of health claims. The evidence for the link between nutrients and health outcomes was reviewed. A reduced salt intake reduces blood pressure, but only a few randomized controlled trials have verified the effect of salt on overall and cardiovascular mortality. Evidence linking a reduced fat intake with cardiovascular mortality and obesity is generally non-significant. Studies that have examined the relationship between obesity and diet have produced contrasting results. A simulation exercise that demonstrated that the impact of a reduced salt and fat intake on overall mortality would be negligible in the European population was carried out. Consideration of the literature and the results of this simulation exercise suggest that the introduction of nutritional profiles in Europe would be expected to have a very limited impact on health outcomes.

Where is the science? What will it take to show that nutrient profiling systems work?

Nutrient profiling is defined as the science of ranking or categorizing foods on the basis of their nutritional composition. Validity is a general term meaning accuracy. Nutrient profiling systems in the United States have not undergone any systematic validation effort to assess their accuracy against a comparison measure or group of measures. Different types of validation studies should be conducted: content, face, convergent, criterion, and predictive. This article provides a conceptual framework for establishing the validity of nutrient profiling systems with the desired objective of assisting US consumers with food selection to improve diet quality. For a profiling system to work successfully in the American marketplace, it must function well with consumers from most or all cultural groups, from all racial groups, and with low-literate as well as highly literate people. Emphasis should be placed on conducting different types of validation studies and multiple studies with different subpopulation groups. The use of consistent standards to assess the accuracy and usefulness of multiple profiling systems is imperative to successfully identify a nutrient profiling intervention that will have the potential to lead to improved diet quality and eventually to an improved health status in US consumers.

Nutrient profiles discriminate between foods according to their contribution to nutritionally adequate diets: a validation study using linear programming and the SAIN,LIM system

BACKGROUND

The nutrient profile concept implies that it is possible to discriminate between foods according to their contribution to a healthy diet on the basis of their nutrient contents only.

OBJECTIVE

The objective was to test the compatibility between nutrient profiling and nutrient-based recommendations by using diet modeling with linear programming.

DESIGN

Food consumption data from the French "Individuelle et Nationale sur les Consommations Alimentaires" dietary survey and its associated food-composition database were used as input data. Each food was allocated to 1 of 4 classes, according to the SAIN,LIM system -- a nutrient profiling system based on 2 independent scores, including a total of 8 basic plus 4 optional nutrients. The possibility to model diets fulfilling a set of 40 nutrient recommendations (healthy models) was tested by using foods from a given nutrient profile class only or from a combination of classes. The possibility to fulfill a set of nutrient constraints in contradiction with the recommendations (unhealthy models) was also tested. For each model, the feasible energy range was assessed by minimizing and maximizing total energy content.

RESULTS

With foods from the most favorable nutrient profile class, healthy diets could be modeled, but it was impossible to design unhealthy diets within a realistic range of energy intake with these foods. With foods from the least favorable class, unhealthy, but not healthy, diets could be designed. Both healthy and unhealthy diets could be designed with foods from intermediate classes.

CONCLUSION

On the basis of a few key nutrients, it is possible to predict the ability of a given food to facilitate -- or to impair -- the fulfillment of a large number of nutrient recommendations.

Nutrient profiling can help identify foods of good nutritional quality for their price: a validation study with linear programming

Nutrient profiling ranks foods based on their nutrient content. They may help identify foods with a good nutritional quality for their price. This hypothesis was tested using diet modeling with linear programming. Analyses were undertaken using food intake data from the nationally representative French INCA (enquête Individuelle et Nationale sur les Consommations Alimentaires) survey and its associated food composition and price database. For each food, a nutrient profile score was defined as the ratio between the previously published nutrient density score (NDS) and the limited nutrient score (LIM); a nutritional quality for price indicator was developed and calculated from the relationship between its NDS:LIM and energy cost (in euro/100 kcal). We developed linear programming models to design diets that fulfilled increasing levels of nutritional constraints at a minimal cost. The median NDS:LIM values of foods selected in modeled diets increased as the levels of nutritional constraints increased (P = 0.005). In addition, the proportion of foods with a good nutritional quality for price indicator was higher (P < 0.0001) among foods selected (81%) than among foods not selected (39%) in modeled diets. This agreement between the linear programming and the nutrient profiling approaches indicates that nutrient profiling can help identify foods of good nutritional quality for their price. Linear programming is a useful tool for testing nutrient profiling systems and validating the concept of nutrient profiling.

A simple visual model to compare existing nutrient profiling schemes

Nutrient profiling is a highly pressing issue. However, as there are currently various nutrient profiling schemes it may be difficult to maintain an overview. We therefore developed a simple visual model where the various choices that can be made are indicated. This allows for easy comparison of existing schemes. The model is available in PowerPoint format and attached as a separate file to this paper (see Supplementary files under Reading Tools online).