Comparison of nutrient profiling schemes for restricting the marketing of food and drink to children

Authors


Address for correspondence: Ms H Brinsden, International Association for the Study of Obesity, 12 Roger Street, London WC1N 6BQ, UK. E-mail: hbrinsden@iaso.org

Summary

What is already known about this subject

  • Many companies have signed voluntary agreements to restrict the marketing of foods high in fat, sugar and salt directly to children.
  • Several nutrient profiling schemes have been proposed for defining products which can be advertised to children.

What this study adds

  • Provides evidence that industry-led nutrient profiling schemes are less effective in restricting the advertising of energy-dense foods compared to government-led models.
  • Provides evidence that industry-led nutrient profiling schemes particularly favour the continued advertising of foods high in sugar.

Background

The food and beverage industry have made voluntary pledges to reduce children's exposure to the marketing of energy-dense foods and beverages, and in 2012 announced the replacement of company-specific nutrient profiling schemes with uniform sets of criteria from 2013 (in the USA) and 2014 (in the European Union [EU]).

Objective

To compare the proposed USA and EU nutrient profiling schemes and three government-led schemes, paying particular attention to the differences in sugar criteria.

Method

Food and beverage products permitted to be advertised in the USA under pre-2013 criteria were examined using five nutrient profiling schemes: the forthcoming USA and EU schemes and three government-approved schemes: the US Interagency Working Group (IWG) proposals, the United Kingdom Office of Communications (OfCom) regulations and the Danish Forum co-regulatory Code.

Results

Under the new USA and EU nutrient profiling schemes, 88 (49%) and 73 (41%) of a total of 178 products would be permitted to be advertised, respectively. The US IWG permitted 25 (14%) products; the Ofcom regulations permitted 65 (37%) and the Danish Code permitted 13 (7%).

Conclusion

Government-led schemes are significantly more restrictive than industry-led schemes, primarily due to their tougher sugar criteria. The Danish Forum (93%) and USA IWG scheme (86%) are the most restrictive of the five examined. Further harmonization of nutrient profiling schemes is needed to reduce children's exposure to the promotion of energy-dense foods.

Introduction

With the rapid rise in childhood obesity prevalence around the world, there is increasing public debate concerning the best course of action to improve children's food environments. One issue that has received attention among policymakers is children's exposure to the commercial promotion of foods and beverages. In May 2010, the 63rd World Health Assembly (WHA) endorsed a set of recommendations to limit children's exposure to the marketing of foods and non-alcoholic beverages (resolution WHA63.14) [1]. This followed earlier calls from the World Health Organization in 2004 for the private sector to address their marketing practices [2] and a 2005 report by the US Institute of Medicine which made a number of proposals for food and beverage industries, advertising agencies and government departments to reduce children's long term health risk associated with marketing practices for children [3]. Subsequent reviews have strengthened the evidence base for linking children's exposure to advertising of foods and beverages to their dietary choices and food intake [4, 5].

Aware of the concern regarding marketing to children, a number of voluntary and self-regulatory codes and pledges have been introduced by several leading food and beverage companies. Between 2005 and 2009, 13 voluntary pledges regarding food marketing practices were developed involving 52 companies [6]. A search in 2012 indicated a further 30 companies signed up to at least one pledge globally and that the number of pledges had increased to 21 [7]. These vary in nature, from collaborations at regional/national level (for instance in the USA and European Union [EU]) at industry level (e.g. quick service restaurants in Australia) and at company-specific level (e.g. Unilever, McDonald's). Typically, the collaborative pledges are characterized by a generic pledge which companies sign up to voluntarily, plus more detailed company-specific commitments which may, for instance, define a nutrient profiling scheme, the age of a child, or the promotional techniques and media channels to be covered if this is not included in the generic pledge. The fact that criteria are often set by each company independently has been criticized for tending to favour the product portfolio of each company concerned and for allowing too much inconsistency in what is and is not permitted to be advertised [8].

In the USA, the main voluntary scheme is coordinated by the Children's Food and Beverage Advertising Initiative (CFBAI) supported by the Better Business Bureau, a business interest organization [9]. Sixteen multinational companies have signed up to the CFBAI pledge with each company making its own commitments on media type, child's age and nutrient profiling criteria. In 2011, the CFBAI announced that it would update its pledge and include a uniform nutrient profiling scheme which all signatories will have to meet by June 2013 [10]. While the pledge is still voluntary, the new uniform criteria are designed to improve consistency among the signatories of the CFBAI pledge about the types of foods considered unsuitable for advertising to children. Flexibility at the discretion of each participating company still remains for the age groups and media channels covered. The age groups and media channels remain at the discretion of each participating company.

Companies operating within the EU announced in 2007 the launch of an EU Pledge to restrict marketing to children under 12 years of age [11]. The EU Pledge initially had 11 signatories, and this has since expanded to 19. Signatories adopt a set of core commitments covering media type and age definition of child, but set their own company-specific nutrient profiling criteria to define foods permitted to be advertised. In late 2012, the EU Pledge members announced a uniform nutrient profiling scheme which would apply to all of their members, from the end of 2014 [12]. This scheme is not identical to the CFBAI uniform scheme. Flexibility at the discretion of each participating company still remains for the age groups and media channels covered. The age groups and media channels remain at the discretion of each participating company.

In addition to these moves, proposals have been adopted by an industry-led organization, the Danish Forum for responsible Food Marketing Communication (Danish Forum) specifying criteria for restricting marketing to children, in order to comply with Denmark's Marketing Practices Act and specifically the Danish Consumer Ombudsman's 2006 Guideline to the Act [13], which states ‘Unhealthy foodstuffs should not be advertised in media targeted to children and young people’ (paragraph 5.8). The Danish Forum published a Code in 2008 and revised in 2010 [14] which includes a nutrient profiling scheme that differs significantly from the CFBAI and EU Pledge schemes. Although having no legal status, it provides a standard for compliance with the Marketing Practices Act, and as such may be considered co-regulatory. The Code applies to a wide range of broadcast, digital and print media ‘directed towards children’ aged under 13 years.

In 2007, a mandatory nutrient profiling model was introduced by the United Kingdom (UK) Office of Communications (Ofcom) and Department of Health to regulate food and beverages advertised on television during programmes seen by a significant proportion of children [12, 13] aged under 16 years. The nutrient profiling scheme formed part of a regulatory framework which restricted television advertising in specified scheduling, for products specified in the nutrient profiling scheme. This model has subsequently been adopted with slight modification by the Broadcasting Authority of Ireland [15] for similar purposes, by Food Standards Australia New Zealand [16] for the purposes of defining foods permitted to carry health claims and by McDonalds to define what they permit themselves to advertise to children.

In 2011, an interagency working group (IWG) drawn from four US agencies (the US Department of Agriculture (USDA), the US Food and Drug Administration (FDA), the Center for Disease Control and the Federal Trades Commission) published draft proposals for voluntary restrictions on advertising to children [17]. The proposals include nutrition criteria for defining which foods should and should not be promoted to children. The original proposals applied to children and adolescents up to age 17 years which was subsequently amended to 12 years; in addition, the original proposals included all media that targeted children but subsequently exclusions were made for marketing at charitable events, sporting events and to families and communities [18].

In 2012, a report by the International Association for the Study of Obesity (IASO) found that the nutrient profiling schemes used by companies signed up to the EU Pledge were generally weaker in comparison to the UK Ofcom, IWG and Danish Forum schemes and therefore failing to capture some of children's exposure to energy-dense food products [8]. The report found more than 30 examples of products which were permitted by EU Pledge members’ schemes but not permitted by other schemes, including chocolate, breakfast cereals and sweetened cereal bars.

The debate about the suitability of different nutrient profile scheme for protecting children remains unresolved. A review by Rayner et al. [19] suggested that nutrient profiling has much potential but that there are several obstacles to overcome before an ideal model for regulating the marketing of foods can be agreed, while an overview of the research opportunities related to marketing to children emphasized the need for research which can directly support policymaking and which can evaluate the effectiveness of controls [20].

The purpose of the present paper is to compare the forthcoming industry-led nutrient profiling schemes proposed by the CFBAI and the EU Pledge, with the Danish Forum scheme, the USA IWG scheme and the UK Ofcom scheme, in terms of the types of foods which each scheme permits to be promoted to children.

The paper forms part of the Pediatric Obesity issue focussing on sugar, and attention will be paid to the sugar criteria used by the different schemes and the foods which the schemes allow. Analysis by the US National Cancer Institute of National Health and Nutrition Examination Survey data shows sodas and fruit beverages to be responsible for over 40% of sugar intake in children, followed by cookies, cereals and confectionery [21]. In the UK, the leading sources of sugar in the children's diets are non-alcoholic beverages, cereals and confectionery (30%, 29% and 15% respectively and of all extrinsic sugars consumed by children aged 4–10 y, respectively) [22]. In the USA, research by Powell et al. [23] found the promotion of high-sugar products to be a feature of advertising to children, with some 48% of total calories among the products advertised coming from sugar. That study also found that within the breakfast cereal category, over 97% of advertisements seen by children aged 2–11 years were for high-sugar varieties.

Method

The study was conducted in September–November 2012. Information regarding industry pledges was extracted from two previous studies on marketing practices [6, 24], from the Yale Rudd Centre database on food marketing and pledge websites [7] and from the coordinating agencies: the CFBAI [9] and the EU Pledge [25]. Details of the other nutrient profile schemes were collected from the appropriate websites of the US IWG [17], UK Ofcom [26, 27] and the Danish Forum [13, 14].

In order to compare nutrient profiling schemes, a set of products has to be compiled which can be filtered by each scheme and the results analysed. For this paper, the authors chose a list of products that the CFBAI signatory companies themselves have stated as being suitable to be advertised to children in the period prior to the introduction of the forthcoming uniform nutrient profiling criteria. This product list was extracted from the CFBAI website (September 2012) [28] (see Supporting Information Table 1). The nutrition information for these products was also collected from the CFBAI website and company-owned websites. To maintain consistency, the product formulations and nutrient content data from the US sources were used in all analyses reported in the present paper. Products in the CFBAI list which were not listed on company websites in September 2012 were excluded from the final list used. The remaining list of products was then used as a means of comparing each of the five nutrient profiling schemes (see flow chart in Fig. 1).

Figure 1.

Methodology.

Table 1. Summary of criteria used under each nutrient profiling scheme
PledgeNumber of food categories (N + N subcategories)Product weightNutrients to limit (NL)Nutrients to promote (NP)
Per 100gPer portion as statedDefined portionRACC Portionenergyfatsat fatsugarsodiumRequired criteria to passBalanced against NL score
  1. CFBAI, Children's Food and Beverage Advertising Initiative; EU, European Union; RACC, reference amounts customarily consumed; US IWG, United States Interagency Working Group.
Industry pledgesCFBAI 201310 + 6x√ (for limited categories)x√ (kcal)x√ (as % and g)√ (g)√ (mg)x
EU Pledge 20149 + 18√ (for meals)xx√ (kcal)x√ (as % and g)√ (g)√ (mg)x
Government PledgesUK Ofcom2xxx√ (KJ)x√ (g)√ (g)√ (mg)x
Danish Forum Code10xxxx√ (g)X√ (g)√ (salt, g)x
US IWG2xx√ (kcal)x√ (g)√ (g)√ (mg)x

Products were categorized into appropriate categories by the two authors independently, using definitions within each nutrient profiling scheme. For products where there was disagreement or inadequate information, the following categorization rules were applied by the authors

  • Flavoured milk based drinks, e.g. chocolate milk, were classed as beverages not dairy (when not stated otherwise).
  • Popsicles (e.g. frozen sugared water) were classed as beverages where appropriate categories were not present.
  • Any product from a quick service restaurant chain was classed as ‘fast food’ when such a category existed.
  • Where multiple flavours of a product were available, the information repeated the declarations on the CFBAI list. For instance, if ‘all flavours’ was stated (rather than each flavour individually) then just one example was included. If information for every variant was given, then each variant was included in the present study.
  • Products classed as exempt from a model were categorized as passing except in the case of soft drinks under the EU Pledge, which states that they are banned outright.
  • A meal (e.g. drink, main dish and additional dish) was excluded under the Ofcom scheme if any one component failed the scheme (e.g. a burger).
  • USDA food composition tables [29] were used if required to get a better understanding of the characteristics of a product to help with classification, but were used as guidance only in informing our decision.

Nutritional information was calculated in the formats required for each of the nutrient profiling schemes. For instance, 100 g data was required for the Danish Code and Ofcom schemes, reference amounts customarily consumed (RACC), as defined by the FDA [30], were required for each category under the IWG scheme and company-recommended portion sizes were used for the majority of categories under the 2013 CFBAI and 2014 EU Pledge schemes (with some exceptions referring to RACC or per 100 g). Three schemes (the Danish Forum scheme and the 2014 EU Pledge and 2013 CFBAI schemes) used a series of food categories to determine nutrient cut-offs for different types of food and beverage products, while the Ofcom scheme uses different cut off values for food and drink. Products were also divided into eight categories chosen by the authors to allow cross scheme comparison (definitions of categories and summary of results shown in Table 2). Some ambiguity existed in categorizing products, particularly when differentiating between meal types under the CFBAI scheme. The authors discussed each of these ‘problematic’ products individually, based on product name, description, ingredients list and/or USDA food composition database, in order to reach consensus agreement. A summary of each nutrient profiling scheme is shown in Table 1.

Table 2. Summary of products permitted in each category
CategoryCategory criteriaTotal in categoryCFBAI permittedEU Pledge permittedOfcom permittedDanish Forum Code permittedUS IWG permitted
nn (%)n (%)n (%)n (%)n (%)
  1. CFBAI, Children's Food and Beverage Advertising Initiative; EU, European Union; Ofcom, Office of Communication; US IWG, United States Interagency Working Group.
Savoury bakeryAll savoury bakery grain based products such as crackers and bread124 (33%)2 (17%)3 (25%)0 (0%)2 (17%)
Sweet bakeryAll sweet bakery grain based products such as biscuits and cookies2113 (62%)8 (38%)0 (0%)0 (0%)0 (0%)
Breakfast cerealsAll breakfast cereals, including porridge oats2621 (81%)6 (23%)4 (15%)2 (8%)6 (23%)
DairyAll dairy products such as unflavoured milk, yogurt and cheese139 (69%)7 (54%)9 (69%)4 (31%)7 (54%)
Composite dishesAny product made up of multiple ingredients such as ready meals, canned spaghetti, soup,267 (27%)15 (58%)20 (77%)0 (0%)0 (0%)
MealsAny product made up of multiple components, e.g. main, side, drink2616 (62%)0 (0%)13 (50%)3 (12%)2 (8%)
SnacksAny product classified as a snack such as popsicles and sweet products not categorised as a bakery item314 (13%)17 (55%)4 (13%)0 (0%)0 (0%)
OtherAny product which did not fall into another category, includes juice, nut spreads, chewing gum, flavoured milk2314 (61%)18 (78%)12 (52%)4 (17%)8 (35%)
Totals17888 (49%)73 (41%)65 (37%)13 (7%)25 (14%)

Products were first analysed based on their content of ‘nutrients to limit’ (saturated fat, total fat, sugars, salt, energy) using the nutrient profiling criteria from each scheme. Where the sugar criteria was based on ‘added sugars’, an assessment was made based on the ingredients list and/or known natural sugar content of ingredients, e.g. milk from food composition databases. When a product ‘passed’ based on ‘nutrients to limit’, a further analysis was conducted based on the content ‘nutrients to promote’, as appropriate under each scheme. The methodology for doing this was done on a product by product basis. In many cases, the nutrients to promote guidelines were based on a specific nutrient content and therefore it was possible to collect this information using the nutrition information provided on company websites and/or in food composition databases. In certain cases, particularly for the application of fruit and vegetable scores under the Ofcom scheme, ‘nutrients delivered through ingredients’ as part of the EU Pledge scheme and principle A under the IWG scheme, determining the products attributes was more complex. In these cases, an assessment of the products components was made by each of the reviewers using the ingredients lists, the product name and product description.

Any product which was listed as being exempt from needing to meet the nutrient profiling criteria of a specific scheme (i.e. not covered by the scheme) was classed as meeting the guidelines, unless there were statements to the contrary (e.g. outright commitment not to advertise the particular category of food).

In the case of the Danish Forum Code, all foods are subject to ‘further consideration’ which asks if the food contains over 0.5 g salt per 100 g. The Code is ambiguous about how and when this criterion should be applied. We interpreted the Code to apply the salt criterion to all products, on the basis that this appears to have been the interpretation of most advertisers, given the dramatic fall in products advertised during Danish children's commercial TV [31]. We understand that salt criteria for each product category in the Code are under discussion.

All data and classification was cross checked by the two authors.

Results

Of the 16 companies that were part of the CFBAI pledge at the time of conducting this study, four companies stated that they do not advertise to children (Coca-Cola, Mars, Ferrero and Hillshire). In September 2012, CFBAI listed 194 products that were allowed to be advertised in accordance with the CFBAI member companies’ commitments. Of these, 16 products were excluded because no nutrition details were available and/or because products no longer appeared on company websites, bringing the total number of included products to 178 [28]. Included products are listed in the Supporting Information Table S1.

A summary of each scheme is shown in Table 1 and the number of products which would be permitted under each scheme is shown in Table 2. However, the absolute number does not reveal some of the important differences between schemes, and these are described in more detail here.

Children's food and beverage advertising initiative (USA)

The CFBAI uniform nutrient profiling scheme will come into force for voluntary signatories from June 2013 [10]. The scheme includes 10 categories of food: juice; dairy (with subcategories milk, yogurt, dairy desserts and cheese); grains, fruit and vegetables and others; soups and meal sauces; seeds, nuts and nut butters; meat and fish; mixed dishes; main dishes; small meals; meals. For the majority of categories, the nutrient criteria are based on the serving size stated on the product labels, with the exception of milk (8 fl. oz.), yogurt (6 oz.) and seeds/nuts (1 oz. or 2 tbsp). The model provides limits for calories (kcal), saturated fat (g), total sugar (g), sodium (mg) as well as having criteria for ‘nutrient components to encourage’ (NCTE), which include category-dependent requirements for fruit, vegetables, dairy and/or wholegrain and in some cases essential nutrients, such as vitamin D, C or A, B vitamins, iron and calcium.

Of the 178 products included in this paper, 103 products failed the CFBAI nutrition criteria due to exceeding the threshold for calories, fat, sugar and/or salt. A further 17 products failed for not containing sufficient NCTE (defined as: at least one-half serving of fruit, vegetables, whole grain or dairy or not containing ≥10% of any essential nutrients), including eight popsicles, seven grain products and two juice drinks. Of the 88 (49%) products that will continue to be permitted, four (one fruit, three chewing gum) ‘passed’ by default as they were excluded from the models assessment process and the remaining 84 products were deemed to meet the minimum nutrition criteria outlined in the model. The new scheme will permit products across a number of categories including a selection of ready meals, tinned pasta, soup, sweet biscuits and snacks, crackers, breakfast cereals, juices, yogurt and milk-based drinks.

EU pledge (Europe Union)

The nutrient profiling scheme for the EU Pledge was proposed in November 2012 [11] and is expected to be introduced in 2014. The scheme includes nine food categories (with 20 subcategories): vegetable and animal based oils; fruit, vegetable and seed based products; meat based products; fish products; dairy; cereal based products; soups, mains, composite dishes, sandwiches; children's meals; edible ices. The criteria include energy (kcal) per portion and sodium (mg), saturated fat (g) and total sugar (g) per 100g. There is a further category specific ‘components to encourage’ such as fruit and vegetables, presence of polyunsaturated fatty acids, fibre, protein, vitamin D/Calcium, B vitamins and wholegrain which must also be met.

Of the 178 products examined here, the EU Pledge 2014 proposals will permit 73 (41%) of products. The main types of products permitted include ready meal/composite dishes, meals, tinned pasta, yogurt, peanut butter, oat-based cereals (e.g. porridge) and iced popsicles. The restricted products were restricted on the basis of having high content of fat, sugar or sodium and/or high calories. No product failed on the sole basis of containing insufficient ‘components to encourage’. Fourteen products were exempt from the scheme, including nine sugary drinks (including sugar-sweetened milk) but were classed as ‘failing’ due to not being permitted by the wider text of the Pledge. One fresh fruit product, one fruit juice and three chewing gum products were classed as passing the scheme. One in four (26%) of the failed products, did so due to their sugar content.

Code of the forum of responsible food marketing (Denmark)

The Danish Code scheme uses 10 product categories: dairy, cheese, meat, bakery, cereals, fruit and vegetables, sauces, beverages, desserts, snacks and candy, and ready meals and convenience foods) [13, 14]. It is based on per 100 g data and specifies levels of fat and/or sugar (g) depending on the category. A limit of 0.5 g salt (equivalent to 0.2 g sodium) per 100 g applies across all categories [31]. 1 For ‘meals and convenience foods’ which pass the fat, sugar and salt threshold criteria, further discretionary consideration is given based on whether the product contains the main food groups and would be approved by nutrition experts as appropriate for children.

Of the 178 products examined here, the Danish Code scheme permits 10 (6%) of products based on meeting nutrition thresholds, primarily in the dairy categories, and a further three products (chewing gum) which are exempt from the scheme. The majority of restricted products were restricted on the basis of their sugar content (100 products) and and/or salt content (112 products). Ten products which passed fat, sugar and salt were failed for not being deemed suitable for child consumption (fast food), as noted as part of the discretionary criteria.

Interagency working group (USA)

The proposals include a nutrient profiling scheme based on two principles: Principle A is based on ‘positive nutrition’ stipulating that individual foods should contain a significant proportion by weight of key food groups, defined as at least at least 50% of what could be considered as meaningful for an entire day by weight of the following: fruit; vegetable; whole grain; fat-free or low-fat (1%) milk products; fish; extra lean meat or poultry; eggs; nuts and seeds; or beans. Main dishes should contain at least 50% by weight from a combination of at least two of these groups and meals should contain at least 50% by weight from a combination of at least three of these groups. Principle B is nutrient based, and proposes limits to the content of energy (kcal) sodium (mg), saturated fat (g), trans fat (g) and added sugars (g). The scheme uses RACC portion sizes [16] (or per 50 g if the portion weight is less) [17].

An analysis of the 178 products examined for this paper indicates that the IWG scheme restricts 130 (71%) products based on principle B, with a further 23 restricted for failing principle A (totalling 86% restricted). The remaining products allowed under the IWG proposals include fruit juice, milk, bread and some ready meals.

Ofcom model (UK)

The Ofcom scheme is based on per 100g data for energy (KJ or kcal), saturated fat (g), sodium (mg), sugar (g), protein (g), fibre (g) and fruit and vegetable content (g) [26, 27]. The scheme is based on a ‘balancing’ approach whereby an overall score is given based on both ‘positive’ and ‘negative’ criteria. A product is given a score of 1–10 for each ‘negative’ nutrient and a score of 1–5 for each positive nutrient (protein, fibre and fruit and vegetables) and will fail if it receives an overall score of 4 or more, while a beverage will fail if it receives a score of 1 or more.

Of the 178 products analysed, the Ofcom scheme permits 65 products (37%). It permits a selection of dairy products, ready meals, tinned pasta, soup, milk based drinks and juice. The scheme excluded 113 out of 178 products such as savoury and sweet crackers, breakfast cereals, fast food and other meals and iced popsicles.

Products permitted under all schemes

Of the 178 products examined, only five (3%) would be permitted under all schemes. The five products were low-fat or zero-fat non-flavoured yogurts (three products) and non-flavoured oat-based cereals (two products).

Sugar: profiling the sugar content of products

Of the products that failed under each scheme, more than half did so because of the sugar content (60% Danish, 61% CFBAI, 52% EU, 45% IWG). In these four schemes, no other single nutrient caused such a high proportion of failures. (Note that the Ofcom scheme does not set a threshold but scores the sugar, fat and salt, and allows offsetting scores from other components.)

In particular, the schemes differed considerably in their treatment of sweet biscuits/cookies and sweetened breakfast cereals (Table 3). The new CFBAI scheme excluded 19% of the sample's breakfast cereals from advertising, while the US IWG and EU scheme excluded 77% and Ofcom and the Danish scheme excluded more than 80%. In all cases, at least 75% of the breakfast cereals that failed, failed due to sugar content. The CFBAI scheme includes a Nutrition.

Table 3. Treatment of breakfast cereals and biscuits/cookies under each scheme
SchemeCFBAIEU PledgeOfcomDanish Forum CodeUSA IWG
  1. CFBAI, Children's Food and Beverage Advertising Initiative; EU, European Union; Ofcom, Office of Communication; RACC, reference amounts customarily consumed; US IWG, United States Interagency Working Group.
Breakfast cereals     
Category for breakfast cerealsGrain, fruit and vegetable products, and items not included in other categoriesBreakfast cerealsn/aCerealsn/a
Sugar criteria used for breakfast cereals10 g/portion (if <150 kcal/portion), or 12 g/portion (if 150–200 kcal/portion) typically equivalent to 25–35 g/100 g<30 g/100 gn/a<15 g/100 g13 g/RACC typically equivalent to 25–35 g/100 g
n (%) of breakfast cereals that fail (Total n = 26)5 (19%)20 (77%)22 (85%)24 (92%)20 (77%)
- of which number failed due to high sugar levels515n/a2220
Sweet biscuits/cookies     
Category for biscuits/cookies/cakesGrain, fruit and vegetable products, and items not in other categoriesSweet biscuitsn/aCerealsn/a
Sugar criteria used for biscuits/cookies/cakes10 g/portion (if <150 kcal/portion), or 12 g/portion (if 150–200 kcal/portion)<35 g/100 gn/a<10 g/100 g13 g/RACC
n (%) of biscuits/cookies that fail (Total n = 21)8 (38%)13 (62%)21 (100%)21 (100%)21 (100%)
- of which number failed due to high sugar levels53n/a1913

In the case of sweet biscuits/cookies, three schemes – Danish Forum, Ofcom and US IWG – restricted all sweet biscuits/cookies in the product sample used. However, the CFBAI scheme restricted just 38% while the EU Pledge restricted 62%.

The CFBAI and IWG schemes use per portion criteria, while the other schemes use per 100g data. The CFBAI, EU, Ofcom and Danish Forum scheme all work on the basis of ‘total sugars’ (unless otherwise stated), while the IWG is based only on ‘added sugars’.

Comparison of the CFBAI 2013 and EU pledge 2014 schemes

A number of similarities and differences were found between the CFBAI and EU Pledge nutrient profiling schemes. The schemes share 11 supporting signatory companies: Burger King, Coca-cola, Danone, Ferrero, General Mills, Kellogg's, Kraft, Mars, McDonalds, Nestle and Unilever. While both schemes use a set of product categories and have criteria for energy, sugar, saturated fat, sugar and ‘nutrients to promote’, the schemes also show notable differences, including:

  1. the categories used: The EU Pledge scheme provides a much greater number of product categories (nine categories with 22 subcategories) compared to the CFBAI pledge (10 categories with only ‘dairy’ and ‘grains, fruit and vegetable and others’ categorized further). Most notably, the CFBAI scheme includes one broad category to cover ‘grain, fruit and vegetable products, and items not in other categories’ while the EU Pledge scheme offers a category for fruit and vegetable (further categorized into five subcategories) and one for cereal based products (categorized into four subcategories);
  2. the EU Pledge scheme is based on per 100g information for saturated fat, sugar and sodium, and per portion data for calories, while the CFBAI scheme is based on the labelled serving size (with the exception of the milk, yogurt, seeds/nut spreads categories which included a defined serving size).

Of the 88 products passing the new CFBAI criteria, 56 fail the EU pledge. Of the 73 products passing the EU Pledge, 43 fail the new CFBAI criteria. This represents a significant discrepancy between the two nutrient profiling schemes.

Products that passed CFBAI but failed EU Pledge criteria include a selection of bakery products (three bread, one savoury biscuit, nine sweet biscuits), cereal (15), dairy (2), meals/composite dishes (23), other (2) and snacks (2). Products that passed EU Pledge but failed CFBAI criteria include savoury biscuits (5), meal/composite (15), snacks (15, primarily iced popsicles), other (6). Fifty products fail both schemes: primarily sweet biscuits (5), savoury biscuits (6), cereal (5), dairy (4), main/composite (4), meals (10), other (3) and snack (12). Thirty-nine passed both schemes: sweet biscuits (5), cereal (6), dairy (7), meal/composite (6), meals (2), other (12), snacks (2).

Discussion

Our analysis suggests that the uniform nutrient profiling scheme introduced by the CFBAI during 2013 should restrict half (51%) of the products previously permitted under the separate company schemes. On the basis that very few of the products promoted to children are likely to promote healthier eating habits, this increase in the numbers of products restricted constitutes an improvement on the previous situation, assuming full compliance with the voluntary scheme. Equally, the EU Pledge 2014 scheme will restrict 59% of products examined, again implying an improvement dependent on compliance. However, both sets of uniform nutrient profiling schemes are less restrictive than any of the other three schemes examined here. The Ofcom scheme would restrict 63% of the products examined, the Danish Forum Code would restrict 93% of the products, and the US IWG scheme (when both principle A and principle B are applied) would restrict 86% of the products examined.

Comparison between CFBAI and EU pledge

A number of similarities and differences were found between the CFBAI 2013 and EU Pledge 2014 nutrient profiling schemes. While they both use a series of product categories and consider energy, sugar, saturated fat, sugar and ‘nutrients to promote’ content of products, they also have a number of differences in the definitions of the categories and therefore the nutrient criteria applied to each product. This will impact on the ‘strictness’ of thresholds applied to products under the different schemes and goes someway to explaining the differences found.

The two pledges were found to share 11 signatories, representing over half of those signed up to each pledge. Such overlap between the signatories raises questions over the rationale behind the production of two different nutrient profiling models, and the justification for the criteria used within each category.

A number of differences can be seen when comparing the EU Pledge 2014 scheme to the CFBAI 2013 scheme: The EU Pledge scheme restricts a high proportion of the breakfast cereal products looked at (on the basis of sugar content) and bakery products (on the basis of sugar and sodium content). At the same time, the CFBAI restricts more ready meals and meals (on the basis of sugar and sodium content) compared to the EU scheme, as well as more ice popsicles (on the basis of sugar content).

Sugar thresholds

The sugar criteria are especially noteworthy. By way of example, when compared to the UK Ofcom scheme, the new CFBAI scheme can be seen to restrict a number of ready meals, meal combinations, soups and tinned pasta while permitting some sweetened cereals and sweet and savoury crackers/biscuits. The Ofcom scheme on the other hand permits many of the ready meals and savoury items but restricts most of the sweetened cereals and biscuits (Table 2).

Of particular note is the difference in the sugar criteria used for breakfast cereals and sweet biscuits/cookies/cakes (Table 3). The new CFBAI scheme permits 21 out of 26 breakfast cereals (81%) while the other schemes all permit fewer than 6 of the 26. Further, the EU Pledge scheme contains a subcategory (6c) for breakfast cereals, while CFBAI puts cereals in a broad category ‘3. Grain, fruit and vegetable products, and items not included in other categories’. The CFBAI scheme makes some attempt to take into consideration ‘nutrients to encourage’; however, most cereals are fortified with nutrients and therefore meet the criteria of having ≥10% Daily Values (DV) of any essential nutrient. This may go some way to explaining why the sugar cut-offs are not adequately restricting the advertising of high-sugar breakfast cereals which are excluded under the other schemes.

Similarly, when looking at sweet biscuits/cookies, CFBAI guidelines permit 13 out of 21 products and the EU pledge permits 8 out of 21. The Danish Forum, Ofcom and US IWG however, all restrict 100% of the sweet biscuits/cookies in the product sample used. As with breakfast cereals, the sugar criteria used for sweet biscuits/cookies differs between the schemes. The EU Pledge is the only scheme to include a category specifically for sweet biscuits.

Complexity of the nutrient profiling models

There is concern that the multiple categories and subcategories used in both the EU Pledge and CFBAI schemes add considerable complexity to the system. There is scope for disagreement and misinterpretation when defining a food as belonging to a particular category, e.g. there are different nutrient thresholds according to whether a products is a ‘meal’, a ‘small meal’ and a ‘main dish’ or a ‘mixed dish’. The ambiguity in these criteria and category definitions reduces the clarity and transparency of the scheme, makes evaluation and comparison complicated, is likely to result in inconsistencies regarding the products that are permitted to be marketed to children, and will make compliance difficult to monitor. The use of criteria such as ‘would a nutrition specialist agree that the food is appropriate for children’ in the Danish code, and determining whether a product, by weight, contains at least 50% of a particular food group is ambiguous, and may result in a lack of transparency and consistency between products that are permitted and those that are not.

Limitations of the present study

The authors came across a number of points of ambiguity categorizing products, and when assessing the products in relation to ‘nutrients to promote’ criteria, as outlined in the methodology. This affected calculations for the Danish Code, the CFBAI 2013 and IWG proposals, especially in deciding categories (CFBAI 2013 and Danish Code), calculating portion sizes (CFBAI 2013 and IWG), calculating nutrition information for meal combinations and assessing the content of ‘nutrients to promote’. In the case of calculating nutrients to promote, some ambiguity exists where ingredients lists and nutrition panel information was not sufficiently detailed. Furthermore, when assessing products against the Danish Forum model, an assumption was made that the salt criteria (0.5 g per 100 g) should be applied to all products. It is however unclear and open to interpretation in the Code whether or not this is in fact the case.

It should also be noted that this paper does not compare schemes in respect of the media channels, children's age or proportion of viewing audience, or other aspects of the schemes which determine how well the schemes are able to protect children from exposure to the promotion of food and beverages. A broader critique has been made by the Obesity Policy Coalition in their report Exposing the Charade [32].

Conclusions

The new nutrient profiling schemes which form the basis for the CFBAI and EU Pledge may be welcomed as a simplification of the previous multiplicity of schemes, and should have some effect on reducing the exposure of children in the US and EU to the marketing of foods and beverages. However, the schemes fail to restrict all of the unhealthy products that other schemes restrict, particularly in the breakfast cereal category where high levels of sugar are permitted. This indicates that further steps need to be undertaken by the CFBAI and EU Pledge members to increase the protection of children's health.

The authors acknowledge some difficulty in comparing the products with the CFBAI and EU Pledge schemes, primarily due to the complexity of the schemes. Such complexity may lead to inconsistencies between companies’ interpretations of the categories and the nutrient criteria and thus reduce consistency in the application of these schemes, and increase the difficulty of independent monitoring of the implementation of the schemes.

In the present analysis, the schemes that appeared to have the greatest ability to protect children from the advertising of high fat, sugar or salt foods, was that devised by the Danish Forum, restricting 93% of products, closely followed by the US IWG model (86%), when both principle A (nutrients to promote) and principle B (nutrients to limit) were applied. While these schemes can be viewed as having a number of weaknesses, primarily the complex nature of calculating Principle A in the case of the US IWG and in determining salt applicability in the case of the Danish model, the nutrient profiling schemes they have adopted are the strongest in terms of restricting food marketed to children. They therefore deserve attention from policymakers when developing guidelines to restrict marketing of food to children.

Conflicts of interest statement

No conflict of interest was declared.

Authors contribution

TL conceived the study idea, HB collected the data, HB and TL analysed the data. Both authors were involved in writing the paper and had final approval of the submitted and published versions.

Footnotes

  1. 1

    The Danish Code definition is ambiguous on this point. We interpreted the rule to mean that a product had to meet the salt criteria, as well as the other criteria, as this appears to be the interpretation of the advertisers, as the number of advertisements on children's programming on commercial TV fell sharply in a wide range of categories after 2007.

Ancillary