Research Uses of the HEI
The Healthy Eating Index–2015 (HEI–2015) is the latest version of the HEI. The HEI–2015 and its predecessors, the HEI-2010 and the HEI–2005, are versatile metrics that can be used in many types of studies to answer a range of questions at multiple levels of the "food stream".
- What is the Food Stream?
- Types of Studies at Various Levels
- Future Opportunities for Research Using the HEI
What is the Food Stream?
Individuals do not make food choices in isolation. Rather, their eating behaviors are influenced by a myriad of contextual factors, including what types of food are available to them where they live, work, and shop. The food stream refers to the flow of foods from agricultural production, through processing and distribution channels, to the food that ends up on our plates.
Increasingly, nutrition researchers are realizing that if we can characterize all the points (sometimes referred to as "levels") along the food stream, we can build a better understanding of influences on consumer behavior. For example, examining the healthfulness of the US food supply, the output from major producers, the menu of offerings in a school system, sales in a local grocery store, or individual-level diets could provide insights into the extent to which individuals have the capacity to make food choices that are consistent with dietary guidelines.
The versions of the HEI are especially valuable tools in this regard because each of them can be used to evaluate any mix of foods. Specifically, the standards used to assign HEI scores are not based on any specific requirements or recommendations. Rather, each of these indices relies on a universal set of standards that apply equally well to any set of foods along the food stream, including diets of individuals. This is a feasible approach because the standards are density-based and are set using age- and sex-specific recommendations that are similar per 1,000 calories.
Types of Studies at Various Levels
Any given set of foods at each level of the food stream can serve as the unit of analysis when employing recent versions of the HEI. At any single level, the index can be used to describe and evaluate a set of foods or to examine relationships between the quality of a set of foods and some other factor. Alternatively, it can be used to examine relationships among multiple levels of the food stream, such as the influence of changes in the quality of manufacturer's output on the quality of individual dietary intake. The interpretability and comparability of studies of this nature are simplified because the HEI can be applied in these various ways at multiple levels.
Additional details on the types of studies that can be conducted at the various levels of the food stream are provided below. The way the HEI is implemented and analyzed varies according to the type of study and level(s) being examined. For more information about this, see HEI Tools for Researchers.
National Food Supply
The amount and types of food available in the food supply are the result of domestic production and imports, after accounting for exports, nonfood uses, inventories, and farm uses. The food supply might be considered the "headwaters" of the food stream, as it represents the source of all agricultural commodities that flow downstream to manufacturers, food outlets, and markets, on their way to individuals. It makes sense that if all individuals are to have access to a healthy diet, this mix of commodities must be appropriately balanced.
Researchers have used an earlier version of the index, the HEI–2005, to examine the healthfulness of the US food supply for a single year (see Reedy et al.) and over several decades (see Krebs-Smith et al.). Results indicate that substantial shifts are necessary for the US food supply to align with current dietary guidance. In particular, the food supply needs to provide substantially more fruit, vegetables, whole grains, and fat-free milk and less salt and fewer empty calories. A similar analysis using the HEI–2010 and more recent food supply data is available.
The next stop along the food stream for many foods in the United States is the manufacturer level, where agricultural commodities are processed into food products. This is the one level of the food stream to which the HEI has not yet been applied, but studies characterizing the output of major manufacturers could be valuable, considering their early position in the food stream and potential influence on levels further downstream.
For example, the HEI could be useful in examining the effects of the Healthy Weight Commitment. That effort involves about 150 food companies who have collectively pledged to remove 1.5 trillion calories from the food supply by 2015. Although the HEI is not necessary to evaluate whether or not they meet that goal, it could be used to examine whether their collective output has a higher diet quality after this effort is realized than before.
Obstacles to employing the HEI at this level relate to the lack of available data. For example, companies do not typically release information quantifying their total output of products. Even if such amounts could be obtained or reasonably estimated, data also are lacking on the composition of processed unprepared foods, such as macaroni and cheese or cake mix, which represent many of the foods associated with this level.
To calculate an HEI score, compositional data are needed in terms of both nutrients and food groups. For all packaged foods, the requisite nutrient composition data are available on the Nutrition Facts Panel, but data on the quantities of fruits, vegetables, whole grains, added sugars, and other components that also are required to calculate the HEI are not. Estimation of these components requires a specialized database that can disaggregate the product into its ingredients and tally the quantities of those ingredients with those of similar items. As described below, that type of database is available for ready-to-eat foods from Community Food Environment outlets and the Individual Food Intake level, but it is lacking for less-than-fully-prepared foods at the Food Processor level.
Community Food Environment
The Community Food Environment represents all of the places where individuals acquire food. These places can be broadly divided into markets, where consumers purchase food to prepare or serve at home, and outlets and other settings where consumers purchase or are served ready-to-eat food. At any of these locations, the HEI can be applied to either the set of foods available or the set of foods sold or served. The set of foods available -- using, for example, the menu at a fast food restaurant -- corresponds to the choices offered to consumers by the market or outlet. The set of foods served -- using, for example, sales data for the same restaurant in a given neighborhood -- represents the impact the location is having on eating habits there.
As mentioned in the section on Food Processing, compositional data are needed in terms of both nutrients and food groups in order to calculate HEI scores. Studies of markets, such as grocery or convenience stores, have been hampered because compositional data on many of the HEI components are lacking for many items -- namely, the unprepared foods. These types of stores also sell many ready-to-eat foods, for which the necessary data are available, but because unprepared foods make up a large portion of foods available and sold in markets, such studies are painstaking. Nonetheless, Volpe and Okrent assessed the healthfulness of consumers' grocery purchases (market baskets) using the HEI–2005 and found a similar pattern of discordance with recommendations as found at other levels (e.g., the food supply) -- too few fruits, vegetables, whole grains and too many refined grains, fats, and sugars/sweets. They also determined that the healthfulness of purchases varies across geographic regions and markets. Another group, Jahns et al., examined grocery store circulars using the HEI-2010 and found the circulars had lower scores than the U.S. population scores.
A database of all types of foods available in markets, including both ready-to-eat foods and unprepared foods, linked to compositional data on nutrients and dietary guidance-based food components would facilitate the expansion of research at this level.
Food Outlets & Other Settings Where People Purchase or Are Served Ready-To-Eat Food
Studies of food outlets and other settings where people purchase or are served ready-to-eat food -- such as cafés, restaurants, and schools -- do not have the same limitations regarding compositional databases as those of food markets, because the foods served are ready-to-eat, which enables the comprehensive databases that have been developed for the individual level generally to suffice. Reedy et al. described the methods for conducting an analysis of a food outlet with the HEI–2005, and Kirkpatrick et al. [PDF] used that approach to compare the healthfulness of menus at five major fast food restaurants.
When sufficient compositional data can be obtained to calculate HEI scores, studies employing the HEI at this level could:
- examine the healthfulness of a set of foods marketed by major grocery chains (say, by tracking foods featured in weekly newspaper ads) (see Jahns et al.);
- determine how well foods offered in the café at a local health care facility conform to recommendations;
- assess the healthfulness of packages provided by food banks and food pantries (see Nanney et al.); or
- compare the diet quality of in-flight offerings across major airline carriers.
Individual Food Intake
When interest is in describing the diet quality of an individual, data must be collected at the individual level. Most of the time, researchers are interested in describing or making inferences about groups of individuals, or populations. A dietitian might also be interested in the HEI for clinical use to assess diet quality for an individual client. In this section, we describe individual level dietary data, and the uses of such data pertaining to the HEI.
Data collected at the individual level commonly use food frequency questionnaires (FFQ), 24 hour dietary recalls, or food records. Before diet quality can be assessed using the HEI, the intake reported by these methods must be summarized into nutrients and food groups. Data collected at the individual level through 24-hour recall methodology are often coded using either the USDA's Food and Nutrient Database for Dietary Studies (FNDDS) or the Nutrition Coordinating Center's (NCC's) Nutrition Data System for Research (NDSR). Both of these databases provide compositional information for a full array of nutrients. However, the HEI requires food serving equivalents in addition to nutrients. The FNDDS links to the Food Patterns Equivalents Database (FPED, formerly the MyPyramid Equivalents Database, MPED), which characterizes the foods reported according to components needed to calculate the HEI. The NDSR has some food component information as well, and has developed a guide to help users create the variables needed to calculate scores using data from NDSR output files. Data collected through an FFQ that links to FNDDS can similarly be linked to the FPED and be used to estimate HEI scores. Data from FFQs that have not been linked with FNDDS would have to be appropriately coded to capture and summarize each of the dietary components required to calculate the HEI. Researchers interested in how to apply HEI scoring should see Tools for Researchers for details regarding scoring and analysis.
Population of Group Level Intake
Studies at the population or group level represent the most frequent type of HEI application. In these studies, individual persons are the units of data collection; however, inference about the HEI is made at the population level. Researchers interested in employing the HEI with data analyzed at the group level can use it to: describe the diet quality of the population overall or for subgroups defined by income, race/ethnicity, and other characteristics; examine relationships between overall diet quality and outcomes, such as mortality or incidence of some chronic disease; or describe the impact of dietary interventions.
Many studies have been conducted using the HEI–2005 and the HEI–2010. The US Department of Agriculture (USDA) has published a brief report on HEI–2010 population scores for 2007 - 2008 and 2001 - 2002 [PDF]. Similar to findings at other levels, the overall US population scores are suboptimal compared to federal dietary guidance. Americans consume far too few fruits, vegetables, whole grains, and milk products or fortified soy beverages, while over-consuming refined grains, solid fat, and added sugars.
Individual Food Intake
Individual-level food intake data can also be used to calculate HEI scores for a given person, for example, within a clinical setting. However, there are caveats in interpretation of the scores at the individual level. The application of the HEI does not involve the comparison of a person's intake to his or her individual requirements, but rather to standards based on national recommendations based on the minimal criterion. Furthermore, the recommendations upon which the HEI is based are meant to be met over time, not every day. An individual's HEI score based on a given day's intake would not necessarily reflect the score based on his/her usual or habitual intake, particularly for components that are episodically consumed such as whole grains.
Researchers interested in employing the HEI at the individual level can use it to:
- describe the diet quality of the population or subgroups defined by income, race/ethnicity, and other characteristics; or
- examine relationships between overall diet quality and outcomes, such as mortality or incidence of some chronic disease.
Future Opportunities for Research Using the HEI
Each new version of the HEI provides an opportunity for further research in these areas, especially if the requisite supporting databases are made available. Collectively, such studies could add to our understanding of the food stream, the influence of different levels on food available to consumers, and the potential impact of environmental and policy changes at each level.