Front-of-Package Nutrition Rating Systems and Symbols: Phase I Report (2010)

As discussed in previous chapters, front-of-package (FOP) nutrition rating systems have been developed for different purposes. The systems have also been based on different approaches to setting nutrient criteria. Among the existing FOP nutrition rating systems and symbols, no two have been developed for exactly the same purposes; similarly, no two have the same underlying nutrient criteria. Concerns over the limitations of the nutrient criteria used in developing existing systems as well as concerns over the potential strength of FOP symbols in encouraging purchases have fueled much of the current debate over the use of FOP nutrition rating systems, and they were a motivating factor in the creation of this study.

Given the number of FOP systems in the market today and the potential for future systems to have a variety of attributes, it was not possible for the committee to undertake an exhaustive evaluation of each system. Instead, the committee took a more general approach to discussing the strengths, limitations, and challenges of developing FOP systems. In the following pages, existing systems are sometimes discussed in order to provide specific examples that illustrate a point, but it is important to keep in mind that these are only examples and are not intended to offer a comprehensive list of all systems that exhibit a certain attribute. The first section of this chapter describes various issues associated with developing the nutrient criteria of FOP rating systems, while the second section identifies the general strengths and limitations related to the committee-defined categories of FOP rating systems presented in Chapter 5.

The development of FOP nutrition rating systems requires a number of steps: making decisions related to the overall purpose (see Chapter 5); developing nutrient criteria, which includes selecting which nutrients and other components to include and choosing the basis for setting the criteria; determining the role of fortification; deciding whether to use the same criteria across all food categories or to use category-specific criteria; monitoring compliance; updating system criteria; and choosing between placing symbols on food packaging versus on shelf tags. This section addresses issues and challenges associated with each of these decision points.

As shown in Table 6-1, the nutrients and, in some cases, other food components included in the existing systems vary tremendously. Most systems focus solely or primarily on which nutrients to limit, some include

certain nutrients or food groups that are to be encouraged, still others pay attention to limiting some nutrients and to encouraging other nutrients and food groups, and a few also include some combination of additional food components, universal adjustors, and weighting coefficients. Most commonly, the nutrients to limit include some combination of calories, total fat, saturated fat, trans fats, cholesterol, sodium, and total or added sugars. Nutrients to encourage have usually been based on those required to be declared in the Nutrition Facts panel or the concerns identified by the Dietary Guidelines for Americans (HHS/USDA, 2005b), or both, and typically include one or more of fiber, calcium, potassium, magnesium, iron, vitamin A, vitamin C, and vitamin E. Systems that focus solely on nutrients that should be limited appear to be primarily concerned with reducing the risk for diet-related chronic diseases. Systems that attempt to assess the overall nutritional value of foods generally include nutrients to limit, nutrients or food groups to encourage, and sometimes other factors.

In determining which nutrients to include in FOP systems, existing systems have used dietary guidance recommendations from domestic and international governments and other authoritative bodies, and the opinions of scientific advisory panels assembled by rating system administrators. Determining which recommendations will serve as the basis both for selecting the nutrients to include and for setting their qualifying levels is a difficult process.

Of particular concern in determining the basis for which nutrients to include are the strength of the scientific

basis for setting the criteria and the changing state of the emerging science. Some FOP systems include only nutrients recognized as being of importance by consensus documents such as the Dietary Guidelines for Americans, while other systems include nutrients or other food components or attributes that have not been recognized by such groups. The inclusion of nutrients and other food components that have not been recognized by consensus bodies is concerning because it is likely that the reason consensus bodies have not recognized these as nutrients to encourage or discourage is that there is insufficient scientific evidence from which to draw a conclusion. Consensus bodies generally review guidelines and recommendations on a regular basis and modify criteria on the basis of the most recent data. Another factor related to selecting nutrients and qualifying criteria is whether to use international dietary recommendations. Some criteria for existing FOP systems are based on dietary guidance recommendations from other countries—usually because the criteria themselves were developed abroad. While many nations are experiencing the same diet and health concerns as the United States, it is important to keep in mind that there are various population and food supply differences among countries and thus dietary recommendations from another country may reflect public health concerns in that country that may not be of concern in the United States.

In summary, decisions about nutrients to include in front-of-package rating systems and the underlying nutrient criteria would be most properly grounded in current nutrition science if based on current consensus documents on the dietary needs of the U.S. population.

Existing systems vary greatly in the approaches they use for setting criteria once the nutrients and other components to include have been selected. As shown in Table 6-2, some criteria are based on U.S. Food and Drug Administration (FDA) label claims, some on dietary guidance recommendations from domestic or international governments or other authoritative bodies, and still others on the opinions of scientific advisory panels assembled by rating system administrators. Each approach has advantages and limitations. The sections that follow describe the strengths and limitations of setting nutrient criteria based on Daily Values and on existing nutrient content claim criteria versus setting criteria based on other forms of dietary guidance or nutrition expertise.

Nutrient-specific systems and some summary indicator systems in the United States use as their basis FDA and USDA criteria related to nutrition labeling, nutrient content claims, and health claims. The simplest approach is to provide the amount of nutrient per serving or the amount of nutrient as a percent of the Daily Value, or both. As described in Chapter 2, consumer education efforts have generally characterized 5 percent or less of the Daily Value as a “low” amount and 20 percent or more of the Daily Value as a “high” amount of a nutrient.

Some issues with the use of Daily Values deserve consideration. For example, not all nutrients of primary interest to public health—such as total calories, trans fat, and added sugars—have a Daily Value. The lack of a Daily Value means not only that there is no basis for developing criteria for a nutrient content claim but also that there is no way to inform consumers whether the amount of a nutrient is “high” or “low.” In the absence of a defined Daily Value for calories, some systems have improvised a reference total daily intake of 2,000 calories, which is consistent with the basis upon which the Daily Values for total fat and saturated fat were derived. Even for nutrients with already established Daily Values and claim criteria for “low” and “high,” there are no regulatory definitions for “medium” amounts of any nutrient, which makes it difficult to design criteria for systems that characterize nutrient contents in this way. Furthermore, many Daily Values based on dietary recommendations made 20 to 30 or more years ago would benefit from a reexamination to better reflect current science.

FOP nutrition rating systems that use symbols (or text) to indicate that a product meets the criteria for a nutrient content or a health claim have additional limitations. For instance, a product that claims to be a “good source of fiber” or an “excellent source of calcium” may not be low in the nutrients that should be limited, e.g., saturated fat and sodium. Even though FDA-regulated product labels must include a nutrient disclosure statement immediately adjacent to a claim when certain levels of fat, saturated fat, cholesterol, or sodium are

exceeded,¹ consumers may disregard these statements. Another issue is that the criteria for “low” may be too strict for some products that might otherwise be consistent with a healthful diet, such as fatty fish, tree nuts, peanut butter, and most vegetable oils. Or products may qualify for one or more nutrient content claims but not make the claim on the package label. Because consumers most likely do not know the nutrient amounts that qualify products for a nutrient content claim, it may not be easy for them to make comparisons and decisions among products with and without an FOP nutrient-specific symbol.

Overall, despite some limitations, using nutrient criteria based on regularly updated Daily Values or nutrient amount per Reference Amount Customarily Consumed (RACC) holds promise as a method of setting criteria for FOP systems. Using this structure would maintain consistency with other nutrition labeling requirements that are likely to remain in place in the future.

When no Daily Value exists or when the criteria for “low” cannot be met by products that system developers want to qualify, alternatives to the Daily Value must be found. In many cases in the past, system developers have looked to dietary guidance recommendations from domestic or international governments or other authoritative bodies as well as to the opinions of scientific advisory panels assembled by rating system administrators to set criteria.

One commonly used approach has been to apply dietary recommendations intended for the total diet to individual products. For example, because there is no Daily Value for total sugars or added sugars, criteria for individual products have been based on the WHO recommendation to limit free sugars intake to less than 10 percent of total energy intake, a recommendation that is based on data related to a low incidence of dental caries (WHO, 2003), or else on the IOM macronutrient report suggestion that added sugars should comprise no more than 25 percent of

total calories consumed, which was based on data related to decreased intake of some micronutrients of American subpopulations that exceeded this level (IOM, 2002/2005).

Another example is how criteria for fat and saturated fat are sometimes set. When the criteria for “low fat” and “low saturated fat” are difficult to meet, the dietary recommendations to keep total fat to no more than 35 percent of calories and saturated fat to less than 10 percent of calories (HHS/USDA, 2005b) are often applied to individual foods. The appropriateness of applying a total diet recommendation to an individual food has not been established, even though this approach was used in the development of criteria for “low fat” and “low saturated fat” claims for main dishes and meals, items which make a significant contribution to total dietary intake.²

Another approach used when criteria for “low” cannot be met is to apply nutrient disclosure amounts.³ These amounts are part of U.S. nutrition labeling regulations concerning the use of claims. If a food qualifies for a claim for one nutrient but exceeds certain prescribed levels for another nutrient, disclosure statements are placed adjacent to claims on food packages to alert consumers that some nutrients in the food may increase the risk of a diet-related disease or health condition.⁴ In all cases, disclosure amounts are considerably higher than the amounts required to meet “low” criteria (by regulation, 20 percent or more of the Daily Value). A main concern about such approaches is whether the criteria they adopt are too lenient.

The Nutrition Facts panel provides nutrient information as an amount (in grams or milligrams) or as a percent of the Daily Value per serving, or both. Criteria for nutrient content and health claims are based on the Reference Amount Customarily Consumed (RACC) and sometimes also per labeled serving size and/or per 50 g or 100 g. As such, FOP systems based on nutrient-specific information are grounded in regulations for nutrition labeling and claims. FOP systems based on nutrient thresholds commonly express nutrient criteria per labeled serving size. The labeled serving size is the appropriate household measure of food or beverage that most closely approximates the RACC.⁵ Depending on the product and how it is packaged, individual items can have a labeled serving size as low as 51 percent or as high as 200 percent or more of the RACC.⁶ Consequently, some items can be manufactured in such a way as to reduce a labeled serving size to meet a FOP’s nutrient criteria.

Algorithm-based FOP rating systems generally evaluate nutrient content per 100 calories in order to take into account the importance of obtaining valuable nutrients within a limited number of calories.⁷ One example is the Nutrient Rich Foods Index (NRFI) (Fulgoni et al., 2009). Proponents of this approach emphasize that positive nutritional aspects of foods are given similar weight in the overall score, and scores for individual foods are considered in the context of the overall dietary pattern. One potential negative aspect of this approach is that scores for some lower- and reduced-calorie foods could be biased and difficult to interpret. For example, nutrient values per 100 calories would be higher for lower-calorie versions of some products, which would lead to a bias against the products when scored on the basis of nutrients to limit (e.g., reduced-calorie salad dressing might receive a poorer score than the regular version if the two products had similar levels of sodium) and cause a bias in favor of the products when scored on the basis of nutrients to encourage (e.g., a calcium-fortified, reduced-calorie beverage might have a more favorable score than the full-calorie version if both products contained similar calcium concentrations). Algorithms also tend to result in relatively low scores for foods, such as lean chicken, that are generally considered to be components of an adequate diet but that contain few of the targeted nutrients to encourage.

Another issue specific to FOP systems based on algorithms that calculate a product score using a combination of nutrients to encourage in the numerator and nutrients to discourage in the denominator is the relationship among

the ratios used to account for the risk of chronic disease. The merits of including constituents that should be encouraged as well as those to avoid, and the correct relative weighting of the factors in the numerator and denominator, cannot be known with certainty. Some systems allow beneficial nutrients to offset nutrients that should be limited in American diets, a practice that some view as questionable. There are also questions about the wisdom of giving food products good overall scores on the basis of providing vitamins and minerals that are actually not lacking in the U.S. food supply or for which there is some doubt about the level of public health concern.

Some FOP systems include criteria for a minimum amount of a food group or ingredient that should be encouraged in the diet—such as fruits, vegetables, low- or non fat dairy products, and whole grains. Other systems require quantification of the amount of a food group that is contained in a product. Threshold-based systems with food group criteria also include criteria for nutrients to limit, but systems based on food group information do not.

Many threshold-based systems include criteria for product categories like snack foods, sweets, and desserts despite general agreement that consumption of these items should be decreased because of their contributions to intakes of calories, saturated fat, trans fat, added sugars, and sodium. While these systems generally contain criteria for nutrients to limit as well as nutrients or food groups to encourage, the qualifying criteria may be viewed by some as too lenient. Furthermore, consumers may perceive foods in these categories that qualify for a symbol as being relatively healthy and not pay attention to the labeled serving size—and thus the amount eaten—when consuming them.

At the same time, replacing regular versions of products with more nutritious versions that meet FOP criteria may still improve total diet quality. Because of resource and time constraints, the committee was unable to carry out the necessary modeling to determine the effect of including products that are generally considered foods to limit in FOP systems. The committee recognized, however, that it would be useful to conduct such studies.

Fortification is another issue of concern in setting FOP system criteria. Some systems include nutrients to encourage as part of the system criteria. However, this raises questions about how foods will be rated that do not naturally contain nutrients to encourage and whether this situation might give food manufacturers incentives to alter their product formulations.

Not all foods, even those considered to be important in health-promoting diets, are sources of nutrients to encourage, such as vitamin D, calcium, potassium, and dietary fiber. If nutrients to encourage are included in the criteria for FOP rating systems, foods that may otherwise be choices to encourage in the diet could receive less favorable ratings than foods that do contain these components. Alternatively, manufacturers might choose to fortify products in order to improve product ratings. The U.S. government has recognized that fortification “can be an effective way of maintaining and improving the overall nutritional quality of the food supply.”⁸ However, the government also recognizes that fortification could “result in over- or underfortification in consumer diets and create imbalances in the food supply … [and] it could also result in deceptive or misleading claims for certain foods.”⁹

In addition, including these dietary components in nutrition rating systems may encourage the addition of these nutrients to food systems in which the nutrient is unstable (because of their chemical compositions or storage conditions) or not biologically available, which would contradict FDA fortification policy. In the case of dietary fiber, fortification may also encourage consumers to eat foods that have had fiber added rather than increasing their consumption of naturally occurring, plant-based foods that are high in dietary fiber, as recommended by the 2010 Dietary Guidelines Advisory Committee (DGAC, 2010).

In summary, including nutrients to encourage (e.g., fiber and certain vitamins and minerals) in front-of-package systems may encourage overfortification or the addition of these nutrients to food systems in which the nutrient is unstable or not biologically available, which would contradict FDA fortification policy.

In developing FOP nutrition rating systems it will be important to consider whether to apply one set of nutrient criteria across all or most product categories or to develop criteria that are specific to individual food product categories.

When one set of criteria is used for all or most foods, federal regulations for nutrition labeling and nutrient content claims often form the basis for the criteria. As mentioned above, claim criteria are based on the RACC¹⁰ as well as on labeled serving sizes and Daily Values for nutrients. Claim criteria are defined so that any claim—such as “free,” “low,” “good source,” or “excellent source”—is the same across all categories of foods and beverages, with modifications for meat, fish, and poultry and for main dishes and meals. Developing FOP systems with the same criteria for all foods creates consistency in how individual products are evaluated, makes it possible to compare foods across all food categories, and makes it easier for consumers to understand the meaning of a claim.

Summary indicator systems typically develop nutrient criteria that are specific to food product categories and to their relative contribution to total intake. For example, because fiber content may be more relevant for fruits, vegetables, and grain products, while calcium is more relevant for dairy products, different nutrient criteria may be set for these different food categories. Developing category-specific criteria requires decisions about which and how many categories to include. Among the threshold-based systems reviewed, Smart Spot (PepsiCo) has nutrient criteria for 3 product categories, Smart Choices for 19 categories, and Healthy Ideas (Giant Food stores) for about 105 categories. Guiding Stars has algorithms for 3 food categories (with qualifying scores ranging from 1 to 3 stars); NuVal has one algorithm with “universal adjustors,” “weighting coefficients,” and other adjustors that are category-specific (with scores ranging from 1 to 100); and the Nutrient Rich Foods Index has one algorithm with no category-specific factors (most foods’ raw scores range from –150 to 300; theoretical raw scores range from –300 to 900) and raw scores are divided into quintiles and assigned a score of 1–5 for better comprehension. Decisions must also be made about which nutrients to consider for each category and what scientific basis to use. The summary indicator systems reviewed by the committee differ widely in how these decisions have been made and so influence the final evaluation of products. While tailoring nutrient criteria to specific food categories can be seen as beneficial in certain ways, it limits individual products to being compared only within product categories that have the same nutrient criteria and not across product categories that have different criteria. An example of product variability among summary indicator systems is shown on Table 6-3 and discussed in greater detail in Box 6-1.

Another issue that should be considered in developing FOP rating systems is compliance. Analytical detection methods are needed if one is to ensure that the products being evaluated actually contain the levels of nutrients or other components needed to meet FOP system criteria. Analytical methods are available to monitor compliance for those systems whose criteria are based on nutrients declared in the Nutrition Facts panel. However, there are no simple analytical tests available to ensure compliance for food components like fruit and vegetable content or added sugars.

Analytical methods are also lacking for monitoring compliance of foods that contain a mix of different food groups (e.g., pizza). However, for products that are not a mixture of different food groups, such as canned tomatoes, compliance can be monitored by comparing the declared serving size with the recommended food group servings, and if the product is 100 percent whole grains, compliance can be monitored by reviewing the ingredient list.

Distinguishing between added and naturally occurring sugars in food products has traditionally posed an analytical challenge, especially when food products contain multiple sources of sugars. According to the 2005 Dietary Guidelines for Americans, added sugars are sugars and syrups that are added to foods during processing or preparation or at the table (HHS/USDA, 2005a). They include the following: various types of beet and cane sugars (white sugar, brown sugar, and raw sugar), corn syrup, corn-syrup solids, high-fructose corn syrup, malt syrup, maple syrup, pancake syrup, fructose sweetener, liquid fructose, fruit juice concentrate (in some, but not all classifications), honey, molasses, anhydrous dextrose, and crystal dextrose. The most common sources of added sugars are refined beet or cane sugar (sucrose) and high-fructose corn syrups (Haley and Ali, 2007). Sometimes the terms intrinsic and extrinsic sugars are used as synonyms for naturally occurring and added sugars (HHS/USDA, 2005b). Intrinsic sugars are those sugars that occur naturally within a food, such as fructose and sucrose in fruits or lactose in milk, and extrinsic sugars are those that are added to foods.

While it is possible to estimate how much added sugar a food might generally contain (e.g., for the purposes of creating databases for dietary surveys such as NHANES), it is not currently possible to determine the exact amount of added sugars in a product. There is no difference between the molecular structure of sugar molecules that occur naturally in the food and the structure of those added to the food (HHS/USDA, 2005b). Analytical methods approved by AOAC International are available for the qualitative and quantitative analysis of mono- and disaccharides in foods (BeMiller, 2003), but the structural equivalence between added and naturally occurring sugars makes it impossible to distinguish between the two types of sugars. A laboratory analysis of a breakfast cereal that contains both raisins and high-fructose corn syrup would, for example, be unable to distinguish the naturally occurring fructose in the raisins from the fructose in high-fructose corn syrup. Without an approved analytical method to make such distinction between types of sugars, it is essentially impossible to independently verify the amount of sugars added to a food product. Thus, the FDA has stated that it would be unable to enforce compliance with the disclosure of added sugars on nutrition labels since analysis only generates the level of total sugars, and historically the agency has maintained the position that it will not promulgate regulations it cannot enforce.¹¹ Concerns like this one clearly have implications for the development of FOP system criteria.

Another concern related to monitoring compliance is the fact that algorithms for some summary indicator systems are not publically available. This precludes scientific review and understanding of the algorithm components and how the nutrients and other factors included in the algorithm are evaluated. Although not necessarily unique to algorithm systems, there is also a concern that even if an algorithm is publicly available, it may include nutrients, food components, or weighting factors that were not analyzed specifically for the product being evaluated, but rather were imputed from the scientific literature or food composition databases.

In summary, to ensure that products actually meet FOP nutrient criteria, it is important that nutrient criteria be publically available and that analytical detection methods be available for the nutrients included in the criteria.

The currency of the Daily Value and of serving and portion sizes is important for FOP systems. One of the committee’s guiding principles was that information highlighted in FOP systems be consistent with the Nutrition Facts panel. Nutrient information presented in the Nutrition Facts panel is based on the amount per labeled serving or a percent of the Daily Value or both. The Nutrition Labeling and Education Act (NLEA) of 1990 also requires that nutrients be presented in the context of the daily diet; it also specifies that serving sizes should represent “an amount customarily consumed and which is expressed in a common household measure that is appropriate to the food.”¹²

The Daily Values comprise Reference Daily Intakes (RDIs) and Daily Reference Values (DRVs). RDIs were created during the implementation of NLEA when FDA changed the name of the U.S. RDAs to Reference Daily Intakes (RDIs) in order to reduce confusion with the RDAs developed by the National Research Council (NRC) of the NAS. FDA at the same time maintained the values based on the 1968 RDAs (NRC, 1968), rather than the newer 1989 RDAs (NRC, 1989b), as explained in Chapter 2. In addition, FDA established Daily Reference Values (DRVs) for total fat, saturated fatty acids, cholesterol, total carbohydrate, dietary fiber, sodium, and potassium, based largely on recommendations from The Surgeon General’s Report on Nutrition and Health (HHS, 1988), the NRC’s report Diet and Health: Implications for Reducing Chronic Disease Risk (NRC, 1989a), and the National Cholesterol Education Program’s Report of the Expert Panel on Population Strategies for Blood Cholesterol Reduction (NIH, 1990). The RDIs and DRVs form the basis for the Daily Values.

When NLEA was implemented in 1993, the scientific basis for the RDIs was already outdated. The Institute of Medicine, the health arm of the NAS, has since issued new Dietary Reference Intakes, but the RDIs and

DRVs—and thus the Daily Values—have not been updated in a timely manner to reflect current nutrition science and to be more relevant to public health.

To determine amounts customarily consumed, FDA used dietary intake data from the 1977–1978 Nationwide Food Consumption Survey (NFCS) and, to a lesser extent, the 1987–1988 NFCS augmented by other sources of information where available.¹³ However, since 1993, it is generally recognized that portion sizes as customarily consumed for many foods have increased, contributing to concerns about excess calorie intake and obesity (e.g., Young and Nestle, 2003). Consequently the RACCs used for determining labeled serving sizes may not reflect the larger portions of food actually being consumed today. While the NLEA mandate would suggest that serving sizes may need to be adjusted to reflect the new amounts customarily consumed, concerns about increased portion sizes and obesity have led to questions about whether labeled serving sizes might more appropriately be based on smaller serving sizes rather than the new, larger amounts generally consumed today.

These issues are of concern for nutrition labeling in general, but they affect FOP systems as well, especially if these systems are designed to maintain consistency with other nutrition labeling regulations. FOP systems have an added challenge in that additional dietary guidance recommendations such as the Dietary Guidelines for Americans may factor into criteria. These additional recommendations have the potential to change, creating another possible way in which FOP system criteria may become outdated.

In summary, it will be important to consider developing a formalized process that will trigger an automatic reassessment of FOP system nutrient criteria if changes are made in the dietary recommendations or nutrition labeling regulations on which the system is based.

A final issue associated with developing FOP rating systems that the committee identified was whether to use FOP symbols on shelf tags or on product packaging. Shelf tag symbols offer some advantages in that they provide an option for providing FOP symbols on unpackaged foods, such as fresh produce, and they may be more effective in getting consumer attention than the symbols on food packaging in retail stores. Because shelf tags can be used to label all foods in a store, they can potentially provide information on the whole diet and not just packaged foods. However, a limitation of shelf tags is that consumers may have difficulty determining which symbol goes with what product, especially if products get moved around on the shelf. An advantage of using symbols on packaging is that, unlike shelf tags, symbols on packaging stay with the food item once it has been purchased, while shelf tag symbols do not. Having the symbol remain with the product when it is brought into the home may help to reinforce the nutritional quality of the product with consumers and other members of the household besides the shopper. This potential benefit needs to be balanced with the inability to provide symbols on foods, such as fruits and vegetables, whose consumption should be encouraged but that are not traditionally sold in packages.

The previous section discussed issues associated with developing FOP nutrition rating systems. This section focuses on the strengths and limitations generally associated with each type of FOP system. Summaries of the strengths and limitations identified by the committee are provided in Tables 6-4 and 6-5.

Nutrient-specific systems provide information about nutrients and food components to limit or encourage and typically display some combination of (1) calories per serving, (2) targeted nutrients expressed as amount per serving,

%DVs; or “high,” “medium,” or “low” indicators; and (3) symbols or icons based on FDA or USDA nutrient content or health claim criteria. Current systems that provide nutrient amounts per serving include General Mills Nutrition Highlights, Kellogg’s Nutrition at a Glance, and the U.K. Traffic Light. Examples of symbols or icons based on FDA or USDA nutrient content or health claim criteria include Harris Teeter and Wegmans Wellness Keys.

Nutrient-specific systems all use one approach across all or most categories of foods and beverages: a declaration, description, or evaluation of calories or a nutrient amount per labeled serving. Many U.S. versions highlight nutrients shown on the Nutrition Facts panel that are considered of particular concern for the health of the American public. Highlighting nutrient amounts on the front of the package, putting the amounts into the context of a daily diet as a percent of the Daily Value, and characterizing the amounts as “high,” “medium,” or “low” can help individuals who want to comply with public health guidance or dietary recommendations from a healthcare provider. The approach is consistent with current regulations for declaring nutrient amounts and criteria for nutrient content claims. An additional strength is that, with the exception of added sugars, analytical methods and procedures for monitoring compliance are defined in regulations.¹⁴

Label space is limited, especially for small packages, which restricts the amount and type of information that can be presented. Including too many nutrients or icons may result in label clutter and interfere with consumers’ ability to use the information. Alternatively, consumers may read only the FOP symbol and reduce their use of the Nutrition Facts panel.

One limitation of nutrient-specific systems that characterize the amount of nutrient present is the lack of a Daily Value for some nutrients, and without a Daily Value there is no basis upon which to develop criteria for characterizing the amount of the nutrient. Even for nutrients with Daily Values and already established claim criteria for “low” and “high,” no regulatory definition exists for “medium” amounts.

Nutrient-specific systems based on FDA/USDA claim criteria that use symbols or text to indicate that a product meets the criteria for a nutrient content or health claim have additional limitations. Some products that qualify for a nutrient content claim may not have “low” amounts of the nutrients that should be limited in healthful diets. Even though regulations require that information concerning nutrients to limit be disclosed when certain claims are made, consumers may disregard the information. Furthermore, disclosure amounts may be too lenient for some product categories. At the same time, the criteria for “low” may be too strict for some products, especially those that might be consistent with a healthful diet. Finally, some products may qualify for one or more nutrient content claim but not actually make the claim on the package label. Because not all eligible products make nutrient content claims and because not all food items carry FOP nutrient-specific text or symbols, it may not be easy for consumers to compare and make decisions among products.

Summary indicator systems include both systems based on nutrient thresholds and systems based on algorithms. The two types both attempt to assess the overall healthfulness of a food product, in one case by setting nutrient or food component thresholds and in the other by integrating information about various nutrients to limit, nutrients to encourage, and other factors. Examples of current systems based on nutrient thresholds include Choices (EU), Sensible Solutions (Kraft), and Smart Spot (PepsiCo); examples of those based on algorithms include Guiding Stars and NuVal.

Summary indicator systems typically develop criteria specific to food categories and their relative contribution to total intake. Both threshold- and algorithm-based systems tend to include nutrients thought to be important to public health concern and also to consider nutrient density. Analytical methods are available for some, but not all, nutrients that make it possible to monitor compliance for those systems whose nutrient thresholds or algorithms are publically available and that are based on nutrients listed in the Nutrition Facts panel.

Similar to the case with systems based on nutrient-specific information, a lack of Daily Values can present challenges for setting nutrient criteria for summary indicator systems if the Daily Values are used as the basis for the criteria. For threshold-based systems, challenges for setting the nutrient criteria include a lack of definition for “low,” “medium,” and “high” for some nutrients, criteria for “low” that are too low, and lenient disclosure amount criteria for some nutrients. One approach to dealing with these issues has been to apply dietary recommendations intended for the total diet to individual products, but the appropriateness of this approach has not been established.

Generally, nutrient criteria for summary indicator systems are publicly available on the sponsoring organization’s website or in peer-reviewed journals (see Table 6-2), but two of the algorithm systems reviewed by the committee are not publicly available. A lack of transparency makes it impossible to have independent assessment of the scientific basis underlying the algorithm or to monitor compliance. In addition, the analytical methods necessary for compliance monitoring are not available for some nutrients (e.g., added sugars and bioflavanoids) (BeMiller, 2003; Robbins et al., 2006; Kwik-Uribe and Bektash, 2008).

Another limitation for summary indicator systems is that consumers do not know at the point of purchase how individual nutrients in a product contributed to the product’s evaluation. With threshold systems, consumers can assume that nutrient amounts meet specific targets, but with algorithm systems consumers cannot know what led to the final rating or what might comparatively be considered the best and worse ratings within a product category.

For both threshold- and algorithm-based systems, decisions need to be made about how many and which product categories to include, which nutrients to include, and the basis upon which to evaluate each nutrient. While tailoring nutrient criteria to specific food categories can be seen as a positive, the systems reviewed differ widely in how these decisions have been made and how they influenced the final evaluation of products, which in some cases may have led to the approval of foods that are generally identified as items to limit in the diet. Finally, as discussed previously in this chapter, the way that summary indicator systems are developed may encourage discretionary fortification in order to meet threshold criteria or improve algorithm scores unless rules are put in place to prevent it. Fortification is a particular concern for summary indicator systems; in these systems, more than in other system types, fortification may be likely to improve the rating of a product that contains nutrients of concern.

This type of FOP rating system provides information about the contribution a product makes to the recommended intake of food groups or ingredients. The food groups to encourage are usually fruits, vegetables, whole grains, and fat free or low fat dairy products. Start Making Choices (ConAgra) is an example of a system based on food groups and the Whole Grain Stamp is an example of a system based on an ingredient.

The FOP systems based on food group information that were reviewed by the committee appear to apply a single, consistent approach across all product categories. For example, to carry a Start Making Choices (ConAgra) logo, products must have at least 10 percent of the daily recommended amount of a food group, and the percent

daily amounts must be displayed in increments of 5 percent.¹⁵ Likewise, the Whole Grains Council has specific instructions that apply to all products for how to determine the whole grain contents.¹⁶

A limitation of systems based on food groups is the lack of analytical methods for monitoring compliance of foods that contain a mixture of different food groups (e.g., pizza). FOP systems based on food groups and ingredients may have an additional limitation if there are no criteria for evaluating the amounts of nutrients to limit in the diet. Foods that qualify may not, for instance, be low in nutrients to limit such as sodium and saturated fat.

Developing the nutritional criteria underlying FOP nutrition rating systems requires decisions about a variety of factors: the nutrients and other components to include and the basis for setting the criteria, the role of fortification, whether the same criteria should be used across all food categories or whether category-specific criteria should be developed, how best to monitor compliance, how to update system criteria, and the placement of symbols on food packaging versus on shelf tags. Decisions related to these issues will affect the outcome of product evaluations. It may be valuable to use a set of test foods to determine how products fare under evaluation by systems that are based on different types of product categories and nutrient criteria and to see whether ratings and rankings of the test foods are consistent with dietary guidance and useful for informing consumers about the usefulness of products in a health-promoting diet.

Decisions about the nutrients to include in FOP systems and about the underlying nutrient criteria will be most effectively grounded in current nutrition science if they are based on current, government-endorsed, consensus documents on the dietary needs of the U.S. population.

Including nutrients to encourage (e.g., fiber and certain vitamins and minerals) in FOP systems may encourage overfortification or the addition of these nutrients to food systems in which the nutrient is unstable or not biologically available, which would contradict FDA fortification policy.

To ensure that products actually meet FOP nutrient criteria, the criteria need to be publically available, analytical methods need to exist for detecting the nutrients and other components included in the criteria, and products need to be evaluated based on their specific nutrient content and not on values imputed from databases and the literature. In addition, it is important that system developers consider creating a formalized process that would trigger an automatic reassessment of nutrient criteria if changes are made in the dietary recommendations or the nutrition labeling regulations on which the system is based.

Based on the committee’s review, several options exist for setting criteria, but these require further testing of consumer use and understanding to assess their overall viability. These options will be further discussed in the next chapter.

BeMiller, J. N. 2003. Carbohydrate analysis. In Food Analysis, 3rd ed., edited by S. S. Nielsen. New York: Kluwer Academic. Pp. 143–174.

DGAC (Dietary Guidelines Advisory Committee). 2010. The report of the Dietary Guidelines Advisory Committee on Dietary Guidelines for Americans, 2010. Washington, DC: U.S. Department of Health and Human Services, U.S. Department of Agriculture.

Fulgoni, V. L., D. R. Keast, and A. Drewnowski. 2009. Development and validation of the Nutrient-Rich Foods Index: A tool to measure nutrition quality of foods. Journal of Nutrition 139:1549–1554.

Haley, S., and M. Ali. 2007. Sugar backgrounder. Outlook Report No. (SSS-249). Washington, DC: U.S. Department of Agriculture Economic Research Service.

HHS (U.S. Department of Health and Human Services). 1988. The Surgeon General’s report on nutrition and health. Washington, DC: U.S. Public Health Service, Office of the Surgeon General.

HHS/USDA (U.S. Department of Health and Human Services/U.S. Department of Agriculture). 2005a. Dietary Guidelines Advisory Committee Report, 2005. Washington, DC: Government Printing Office.

HHS/USDA. 2005b. Dietary Guidelines for Americans, 2005. Washington, DC: Government Printing Office.

IOM (Institute of Medicine). 2002/2005. Dietary Reference Intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and amino acids. Washington, DC: The National Academies Press.

Katz, D. L., V. Y. Njike, Z. Faridi, L. Q. Rhee, R. S. Reeves, D. J. A. Jenkins, and K. T. Ayoob. 2009. The stratification of foods on the basis of overall nutritional quality: The Overall Nutritional Quality Index. American Journal of Health Promotion 24(2):133–143.

Kwik-Uribe, C. and R. M. Bektash. 2008. Cocoa flavanols—measurement, bioavailability and bioactivity. Asia Pacific Journal of Clinical Nutrition 17(S1):280–283.

NIH (National Institutes of Health). 1990. National Cholesterol Education Program: Report of the Expert Panel on Population Strategies for Blood Cholesterol Reduction. Washington, DC: U.S. Government Printing Office.

NRC (National Research Council). 1968. Recommended Dietary Allowances. Seventh revised edition. Washington, DC: National Academy of Sciences.

NRC. 1989a. Diet and health: Implications for reducing chronic disease risk. Washington, DC: National Academy Press.

NRC. 1989b. Recommended Dietary Allowances. Washington, DC: National Academy Press.

Robbins, R. J., C. Kwik-Uribe, J. F. Hammerstone, and H. H. Schmitz. 2006. Analysis of flavanols in foods: What methods are required to enable meaningful health recommendations? Journal of Cardiovascular Pharmacology 47(S2):S110–S118.

WHO (World Health Organization). 2003. Diet, nutrition, and the prevention of chronic diseases; Joint WHO/FAO Technical Report Series no. 916. http://www.who.int/dietphysicalactivity/publications/trs916/download/en/index.html (accessed June 3, 2010).

Young, L. R., and M. Nestle. 2003. Expanding portion sizes in the U.S. marketplace: Implications for nutrition counseling. Journal of the American Dietetic Association 103(2):231–234.