Guiding Framework Home Illustrative Questions

Quality: What is a quality diet and is it the same everywhere?

General Question:

  • What is diet quality and what are some of the ways to measure it?

Importance: Diet quality is a key component of the definition of food security. With rapidly changing diets and increasing concern for all forms of malnutrition, measuring diet quality, instead of focusing on energy sufficiency or on single nutrients, is of growing relevance. Tracking individual nutrients is insufficient for understanding the causes and consequences of poor health and nutrition outcomes and thus there is increasing interest in looking at dietary patterns.

Despite widespread use of the term “diet quality”, it is poorly defined with limited agreement on how to measure it. One of the reasons it is hard to define is that the exact configuration of a quality diet varies with dietary customs, cultural context, locally available foods, and individual needs (e.g. age, sex, physical activity level). A general way of thinking about diet quality is grouping foods into healthy and unhealthy components whereby one should consume adequate consumption of healthy foods and nutrients (e.g. fruits, vegetables, whole grains, fiber etc.) and moderate (or very limited) consumption of unhealthy foods and nutrients (e.g. saturated fat, sugar, sodium etc.) (WHO Healthy Diet Factsheet, Guenther et al. 2013). Diversity can be considered either a component of diet quality, or a proxy for diet quality based on the idea that a diverse diet consists of both sufficient nutrients (adequacy) and will necessarily limit other nutrients (moderation).

Diet quality can be measured in different ways using either a priori or a posteriori analytical approaches (Hu 2002). Using previous knowledge of a healthy diet, composite indices have been created that operationalize the concept of diet quality (a priori approach). Indicators that capture a key facet of diet quality (e.g. dietary diversity) are also used. Alternatively statistical modeling, such as factor analysis and cluster analysis, can be applied to dietary data in order to ascertain the types of dietary patterns (a posteriori approach) (Hu 2002).

Several examples of a priori indices exist such as the Healthy Eating Index 2010 (HEI 2010), which has two components consisting of adequacy and moderation (Guenther et al. 2013), while the Diet Quality Index-International (DQI-I) is proposed to have four components (Kim et al. 2003), including:

  • Variety / Diversity: within and across food groups
  • Adequacy: sufficiency of nutrients compared to requirements or quantity of consumption of specific food groups
  • Moderation: specific nutrients or foods that should be consumed with restraint
  • Overall Balance: composition of macronutrient intake

There is no single, validated index that has been used to measure diet quality across low- and middle-income countries, in part because determining a single definition of diet quality that can be applied across different contexts and cultures is difficult. Given the lack of a globally relevant diet quality index, individual indicators such as dietary diversity scales, are frequently used as proxy measures for overall diet quality.

Potential data sources and ways of measuring diet quality: Food Balance Sheets (FBS) can be useful as a way of measuring the dietary patterns overtime. While FBS do not provide information on consumption, they can illustrate how the food supply and consumption patterns have changed over time (see for example the analyses in the Global Panel on Agriculture and Food Systems for Nutrition 2016). Such analyses can provide a macro picture of shifting diets but with no insight into the distribution of foods within the country. FBS can also be used to calculate indicators such as quantity of available fruits and vegetables in a given country, the share of energy from non-staple foods, and the share of energy from animal source protein (FAOSTAT). These types of indicators are important in that they provide information about availability of diversity foods in a country but are not able to provide specific insights to the other dimensions of diet quality. FBS are freely accessible and can provide limited information on trends over time and shifting dietary patterns.

Household Consumption and Expenditure Surveys (HCES) can be used to calculate similar indicators as those discussed for FBS in the paragraph above, in addition to other indicators such as an adapted household dietary diversity score (ADePT, 2016). HCES collect data at the household level and have the added advantage of being nationally representative so indicators can be constructed to assess dietary patterns in urban and rural areas (for example see Hirvonen et al. 2015), such indicators can provide a measure of dietary quality and insight on the nutrient density of the overall diet.

HCES data could be a good option for calculating various indices related to dietary quality, particularly because much of the existing data can be used for analyses at a sub-national levels. However, one possible shortcoming in many countries is that HCES collect limited information (and sometimes no information) on food away from home. In some countries that are urbanizing with rapidly changing dietary patterns and increasing quantities of food that are consumed outside the home the ability to accurately assess overall diet quality without information on food consumed away from home may be limited.

If the objective is to measure all dimensions that encompass the definition of ‘diet quality’, using individual level quantitative data from a 24-hour dietary recall or weighed food records is required. A 24-hour dietary recall, for example, provides quantitative information on the individual level, which allows for not only quantification of dietary diversity, but also nutrient adequacy, moderation, and overall balance of the diet. These data also allow for the assessment of nutrient deficiencies and excesses in populations (their primary use). Food frequency questionnaires (FFQ) can also be can be used to determine some aspects of diet quality (e.g. fruit and vegetable consumption, animal source protein), particularly if they seek to quantify intake. FFQs can be developed to measure moderation (intake of nutrients or foods that should be consumed with restraint), and overall balance (macronutrient composition in the diet). The major drawback of using an FFQ is that they are context specific since diets vary from place to place and thus the food list needs to be updated and the FFQ validated in each context.

Another option is to focus solely on dietary diversity by relying on one of the existing dietary diversity modules. The benefit of using a standard dietary diversity measure, such as the Minimum Dietary Diversity for Women (MDD-W), is that they rely on pre-established, validated, short FFQs. Individual dietary diversity scores (e.g. MDD-W) have demonstrated that consuming foods from different food groups results in a greater probability of meeting micronutrient needs than women consuming foods from fewer food groups (FANTA), while the Household Dietary Diversity Score (HDDS) provides a proxy measure of energy adequacy (Hoddinott and Yohannes 2002 and Leroy et al. 2015). Dietary diversity modules can easily be inserted into longer surveys and are designed for rapid assessment, however they do not collect information on the quantity consumed so are unable to estimate nutrient adequacy.

Bottom line advice: Currently few indicators capture every dimension of diet quality and those that do require investments in collection of individual-level quantitative dietary data. If you are interested in measuring some dimension of diet quality but have limited time and resources, the best option may be to collect information on dietary diversity. The Minimum Dietary Diversity for Women (MDD-W) indicator has recently been validated and can provide useful information at baseline and endline, for example. However if you are interested in measuring diet quality in its full form, it is recommended to use quantitative individual-level dietary data (e.g. from a 24-hour recall) to ensure that you are able to measure all components of diet quality.

Program and Policy Questions:

  • Question 1: People in my country are becoming overweight, especially in urban areas, but many children are still stunted.  How can I figure out how changing dietary patterns may be contributing to this trend?

Many countries are increasingly facing the double burden of malnutrition in which high and stagnant rates of stunting and underweight are quickly being overshadowed by increasing rates of overweight and obesity (Popkin et al. 2012; Shrimpton and Rokx 2012). The nutrition transition and shifting dietary patterns has been caused in part by globalization, urbanization, and economic growth.

Given your interest in measuring changing dietary patterns, and how they may be linked with changes to body weight over time, it will be necessary to gather 24-hour recall data from a sample. If dietary guidelines exist you can adapt an existing diet quality index (e.g. the HEI 2010) to the dietary guidelines in use in your country. If, however, nationally relevant dietary guidelines do not exist (and in many cases they do not) then the best option would be to use study specific data ad apply appropriate statistical methods. Keding et al. (2011) take this approach in Tanzania to explore the relationship between dietary patterns and the nutritional health of women in Tanzania.

In addition to collecting repeated 24-hour recall data from your defined population over time (i.e. longitudinal data) it will also be important to collect data on body mass index (BMI), and possibly other factors such as attitudes toward obesity, and measures of socioeconomic status. Based on the 24-hour recall data the mean intakes of the foods can be calculated and used to identify the dietary patterns by applying the statistical technique of principle component analysis (see for example Keding et al. 2011). With this a posteriori approach, identified dietary patterns can be correlated with BMI and other health outcomes to understand the relationship between outcomes of interest (e.g. overweight) and dietary patterns. Dietary patterns that are identified can also be used to inform dietary guidelines.

  • Question 2: How can I figure out the prevalence of micronutrient deficiency for the population as a whole? What about for urban versus rural areas?