The Reference Diet Deprivation (ReDD) index: A new diet quality measure for more effective nutrition-sensitive policies

Poor diet quality is a major cause of various forms of malnutrition and noncommunicable diseases. Tackling this public health problem is an important development policy priority; and doing so successfully requires access to reliable measures of diet quality that can be used for policy planning and evaluation purposes. Recent methodological developments have mostly focused on the costs and affordability of healthy diets. While diet costing methods are useful for assessing whether nutritionally desirable diets are achievable, they provide little information on how far consumers are from that diet, what dietary shifts are required to get closer to the diet, and how food or related policies might encourage such behavioral change.

In a recently published open-access article in Food Policy, we address this knowledge gap by proposing a new, quantifiable measure of diet quality in populations with useful applications in development policy analysis. Our Reference Diet Deprivation (ReDD) index is based on a comparison of household food consumption as reported in commonly available household consumption and expenditure surveys across a distinct number of food groups against optimal consumption amounts defined by any selected reference diet. Drawing on the Alkire-Foster method for measuring multidimensional poverty, the ReDD index can be decomposed into three sub-indicators:

1. the incidence of diet deprivation, i.e., the percentage of people with an inadequate diet relative to the reference diet;
2. the breadth of diet deprivation, i.e., the average proportion of food group deprivations that diet deprived people face; and
3. the depth of diet deprivation, i.e., the average consumption shortfall in each food group relative to the reference consumption threshold.

The composite ReDD index is the product of the three sub-indicators and ranges from 0 to 1, with higher values indicating a greater degree of diet deprivation and hence a lower diet quality. The index allows users to compare the quality of people’s diets across population subgroups or across countries, or to track dietary changes over time, using Living Standard Measurement Study (LSMS) household survey data, for example. While informative as a standalone measure, the ReDD index can also be integrated into economic simulation models to assess the dietary implications of different policy scenarios or external shocks.

In the article, we demonstrate the indicator’s usefulness in development policy analysis by linking it to the Rural Investment and Policy Analysis (RIAPA) model, an economywide modeling framework developed by the International Food Policy Research Institute (IFPRI). The presented simulation analysis suggests that, out of more than twenty agricultural value chains, productivity growth in the dairy, pulse and nut, fruit, and red meat value chains have the greatest potential to reduce diet deprivation in Nigeria. These findings, we argue, can provide important evidence for nutrition-sensitive agricultural policy formulation and prioritization of actions.

The ReDD index has several useful features, many of which distinguish it from existing diet quality measures or those under development:

First, the ReDD index is derived from estimates of food consumption gaps across multiple food groups relative to a reference diet representing a nutritionally desirable diet. The approach is motivated by the policy-relevant question of how to move people closer to an ideal diet, rather than only counting the number of people that cannot afford the overall diet or lie below a reference calorie threshold. The emphasis on the composition of the diet further allows analysts to measure the contribution of underconsumption of different food groups to overall diet deprivation.

Second, while the index uses the EAT-Lancet healthy reference diet as default, it can easily be modified to accommodate proposed derivations of the EAT-Lancet diet, quantitative national food-based dietary guidelines, or other constructed global reference diets. What sets the ReDD index apart from other measures of diet quality and affordability is its emphasis on whether food-group specific reference calorie amounts are being attained.

Third, the ReDD index can be computed in two ways depending on data availability and user preferences. Under an expenditure-based approach, the index, denoted by ReDD-X, is derived from food expenditure shortfalls relative to the estimated minimum costs of food groups in a reference diet. Under a calorie-based approach, the index, denoted by ReDD-C, is derived from calorie consumption shortfalls relative to food group-specific reference calorie intakes given by the reference diet. ReDD-X is widely applicable as most countries collect household-level food expenditure data. Cost thresholds for the index are computed from the International Comparison Program (ICP) data, as is common in the diet costing literature. ReDD-C, on the other hand, requires detailed food consumption quantity data and reliable calorie conversion factors. While adequate food consumption data are less widely available, ReDD-C avoids having to cost the diet—a potential source of bias in ReDD-X values.

Lastly, although it is likely that the ReDD index will most often be computed using household-level food consumption and expenditure data, it can be applied to individual-level data such as those from 24-hour dietary recalls and food diaries, provided food-item-specific expenditures or consumption quantities are reported.

In summary, the new ReDD index is easily computed from commonly available household consumption and expenditure survey data, and yields rich information about the incidence, breadth, and depth of diet deprivation within populations as well as the contribution of underconsumption of different food groups to overall diet deprivation. When the ReDD index is integrated into IFPRI’s RIAPA model, it is possible to simulate the effects of policies or shocks on diet quality alongside other outcomes such as economic growth, job creation, or poverty reduction. This simulation framework is therefore useful not only for designing more effective nutrition-sensitive policies, but also to assess the benefits and trade-offs of those policies across different development outcomes.

Referenced article:

Pauw, Karl, Olivier Ecker, James Thurlow, and Andrew R. Comstock. 2023. "Measuring changes in diet deprivation: New indicators and methods." Food Policy 117: 102471. https://doi.org/10.1016/j.foodpol.2023.102471

This is a modified version of the post that first appeared on the International Food Policy Research Institute (IFPRI) Blog.