Global Food Security 2 (2013) 5663

Contents lists available at SciVerse ScienceDirect

Global Food Security

journal homepage: www.elsevier.com/locate/gfs

Towards overcoming the food consumption information gap: Strengthening household consumption and expenditures surveys for food and nutrition policymaking
John L. Fiedler n,1
HarvestPlus, International Food Policy Research Institute (IFPRI), 2033 K Street NW, Washington DC 20006, USA

a r t i c l e i n f o
Article history: Received 28 January 2012 Accepted 14 September 2012 Keywords: Nutrition policy Evidence-based policy Household Consumption and Expenditure Surveys HCES Nutrition Micronutrients

a b s t r a c t
The dearth of nationally representative dietary assessment studies continues to severely constrain the nutrition evidence base and throttle the pace of global progress in improving nutrition. Despite their shortcomings, household consumption and expenditures surveys (HCESs) are increasingly being used to address the food and nutrition information gap because they contain a great deal of information about food acquisition and consumption; are done once every 35 years in more than 125 countries; have large samples (  8500 households); are statistically representative at subnational levels; and are much less costly than other dietary assessment data sources. To date, the nutrition communitys role has been that of a passive user of HCES that have already been conducted. Many HCES shortcomings, however, stem from design and implementation issues. If the nutrition community, with its unique skills and experiences were to get more proactively involved in the design, implementation and analyses of HCES, they could be strengthened substantially as a tool for evidence-based food and nutrition policy. This article describes the evolution in the use of HCES in addressing food and nutrition issues, identies HCES shortcomings and distills a shared agenda and a strategy for the nutrition community to work on, together with already existing HCES stakeholders, to strengthen HCES. A two-tiered approach and process for implementing this work is described. The rst tier of the approach consists of addressing a common set of activities at the global level, while the second tier is more country-level work that builds on a combination of the global-level workincluding the adoption and implementation of some of outcome of the rst tier activitiesbut may also include more idiosyncratic, country-specic work. The common global-level activities consist of addressing common, cross country, technical issues of questionnaire and survey design, implementation and data processing activities at the global level. A 115-country assessment of these aspects of HCES is already being conducted jointly by the World BankFAOInternational Household Survey Network. This work aims to distill better practices and lessons, recommend alternative ways to address common HCES shortcomings, and establish a global research agenda for improving understanding and identifying tradeoffs involving critical issues. The second tier of the approach consists of recognizes that HCES design and methodology has to be adapted to each countrys policy needs and strategies, while reecting each countrys technical and nancial constraints and building on its own experiences. Second tier activities are country level activities, and they are where the real work of strengthening HCES has to be done. That work should consist of the merging of the two tiers of the approach to create a partnership for implementing rigorous, experimental studies of the major, unsettled measurement issues confronting HCES, while providing a more sound foundation of evidence for nutrition policy. & 2012 Elsevier B.V. All rights reserved.

1. Introduction Since the mid-1940s, the Food and Agriculture Organizations (FAO) Food Balance Sheets (FBS) have been the principle data source for monitoring global food security, as well as an advocacy tool for focusing attention on hunger and malnutrition. FBS, however, contain only national level data about food availability. They do not provide information on: (1) access to available food, (2) how available food is distributed within a country, (3) how

Tel.: 1 703 966 9333. E-mail addresses: j.edler@cgiar.org, edler.jack@gmail.com 1 John (Jack) Fiedler is afliated with HarvestPlus, the Biofortication Project implemented by the International Center for Tropical Agriculture (CIAT) and the International Food Policy Research Institute (IFPRI), Washington DC, USA. 2211-9124/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.gfs.2012.09.002

J.L. Fiedler / Global Food Security 2 (2013) 5663

57

much of the available food is consumed, or (4) by whom. Despite these major shortcomings, FBS continue to be one key data source used in designing food and nutrition policiesin particular, for fortication programsdue largely to the perception that no other data are available. Few countries have food consumption data from what nutritionists generally regard as the preferred food consumption methodologiesviz., observed-weighed food records (OWFR) or 24-h recall (24-HR) surveysbecause they are expensive and difcult to conduct (Gibson, 2005; Neufeld and Tolentino, in press; Fiedler et al., 2011). Over the past 25 years, there has been a growing body of work that has demonstrated that data on food data collected in a family of multi-purpose surveyscollectively referred to here as household consumption and expenditure surveys (HCESs)can contribute to addressing the food consumption information gap and making nutrition policy more evidence-based (Trichopoulou, and Lagiou, 1997; Lagiou and Trichopoulou, 2001; Smith et al., 2006; Smith and Subandoro, 2007; Stein et al., 2007, 2008a, 2008b; Fiedler et al., 2008, 2012a, 2012b; Fiedler, 2009; Fiedler and Macdonald, 2009; WHO, 2009; Fiedler and Helleranta, 2010; n, 2008; Fiedler and Adra, 2010; Klemm et al., 2010; Sibria n et al., 2008; DAFNE-ANEMOS, 2011a, 2011b; Dop, 2012; Sibria Coates et al., 2012a). How well these surveys measure food consumption, however, what their key limitations are, and the extent to which their shortcomings can be ameliorated or eliminated, are critically important, as yet, largely unaddressed, issues. The nutrition community has become the newest HCES stakeholder, and it brings with it a new set of concerns and issues about these data, as well as new methodologies and applications for using them, new criteria for assessing their quality and relevance, and a growing, still evolving, set of suggestions for improving them. This article is intended to promote the dialog both within the large and diverse nutrition community, and between the nutrition community and the greater community of HCES stakeholdersby articulating how some members of the nutrition community would like to see HCES instruments improved, without creating false expectations. The development of a common understanding of the nutrition communitys needs is the critical pre-condition for devising a strategy for improving HCES, as judged from a nutrition perspective. The organization of the article is as follows: the next section discusses the appeal of HCES for food and nutrition analysis, and is followed by a discussion of the growth and evolution in the use of HCES. The fourth section discusses the diversity of HCES and their relative strengths and weaknesses. The nal section discusses some of the priority reform agenda items, and proposes a general, two-tiered process for rening the agenda and implementing it.

programs. They have been used to estimate the coverage and impact of existing programs, as well as to conduct feasibility and cost-benet analyses. HCES have a number of appealing characteristics for food and nutrition analysts and policymakers. Most fundamentally, they contain a wealth of information about household food acquisition and consumption behaviors. They collect data on how food was acquired, differentiating whether it was purchased, homeproduced or received free-of-charge (e.g., from friends or relatives or a social program, or as payment in-kind). In addition, they are generally based on large samples of households and are statistically representative at the national level and almost always at a subnational (regional or state) level, as well. HCESs are also appealing because of their enormous coverage. Over the last two decades there has been a dramatic increase in the number, quality and availability of HCES in developing countries. The World Banks 1990 World Development Report presented original cross country analyses of HCES from 22 countries, with a single survey for each country. Today, there are more than 700 surveys for 116 countriesan average of six per countryand the collective sample from the latest surveys of each of the 116 countries totals 1.2 million households comprised of more than ve million persons (Ravallion, 2011). The cost of using HCES to analyze food and nutrition issues is another of its attractions. A recent study based on analysis of the costs of 24HR surveys in nine countries, estimated that it would cost $2.3 million to develop (from scratch) a clean, readyto-use, nutrition analytic le for 8500 households. The costs would cover questionnaire development, sample development and selection, the household interview survey eldwork, data entry and data cleaning, and (using this data along with food composition tables), constructing individual-specic variables of macro- and micro-nutrient intake (Fiedler et al., 2011). In sharp contrast, to develop a nutrition analytic le from an extant, already-processed HCES would cost about $40,000, roughly two percent of the 24HR survey costs.2 The fundamental, unanswered question, of course, is: What is the precision-cost tradeoff between of the individual-specic 24HR and the household level HCES measures? A nal attraction of HCES is that they are conducted routinely, and updated periodically, generally once every 3 years. In stark contrast, the only country in the world that routinely conducts an individual-based, nationally representative, 24HR survey is the Philippines, which has conducted them once every 5 years since (ca.) 1970.

3. The growth and evolution of HCES-based analyses of food and nutrition issues The history of using HCES in food and nutrition analyses goes back a quarter of a century and has involved ve distinct sets of landmarksmost of which involve projects, as distinct from more regularly funded and more permanent programs. The University of Athens DAta Food NEtworking (DAFNE) Project pioneered the use of HCES in 1987, to monitor trends in food habits and food availability (Trichopoulou and Lagiou, 1997; Lagiou and Trichopoulou, 2001; DAFNE-ANEMOS, 2011a, 2011b). Since then, DAFNE has become DAFNE-ANEMOS. The Project has harmonized
2 The key activity in processing the HCES would be the matching of each item in the HCES food item list with a food composition table entry to construct the household-level variables of macro- and micro-nutrient availability and then adjusting them using the FAO adult male consumption equivalent to take into account intra-household distribution to obtain estimates of individual nutrient intake.

2. The relevance and allure of HCES HCESs are relevant to food and nutrition analysts and policymakers because they contain a great deal of useful information about food that has already proven useful in addressing the food and nutrition information gap. By contributing to a stronger empirical basis for evidence-based nutrition policymaking, HCESs have provided the wherewithal to enhance the ability of nutrition program designers and policymakers to make more and betterinformed decisions. HCESs have been used to construct a number of nutrition measures including: the number, percentage and location of households that acquire specic types of foods or that purchase fortied or fortiable foods. They have also been used to assess variations in dietary patterns, to measure nutrient intakes, to identify the most common food sources of specic nutrients and to model the impact of fortication and biofortication

58

J.L. Fiedler / Global Food Security 2 (2013) 5663

HCES data in its 28 European country network, and developed software to facilitate its regularly conducting comparative analyses of food trends. The DAFNE-ANEMOS approach improved upon FBS by going beyond national measures of food availability to include the concept of access (FAO, 2011), and provided information on the sum of a households acquisition and consumption of food during a xed time period; referred to as apparent consumption (discussed in detail in section 6.2). It analyzed apparent consumption at the household level, and took into account the households total energy requirements by adjusting for household size and composition (FAO, 2003). More specically, DAFNEANEMOS used food composition table data to quantify the caloric content of the foods apparently consumed by the household, while adjusting for the age- and gender-specic energy requirements of each household member using the FAOs adult male consumption equivalent (AMCE). The second major landmark in the use of HCES for food and nutrition analyses was the International Food Policy Research Institutes Assessing Food Insecurity (AFINS) Project (Smith et al., 2006; Smith and Subandoro, 2007). From 2000 to 2006 AFINS analyzed the HCES of 20 developing countries of sub-Saharan Africa and Asia, and developed guidelines for processing, cleaning and using HCES databases to more precisely measure food insecurity. Like DAFNE, it developed household level indicators of caloric intake and nutrient intake inadequacy, using the AMCE to adjust for household size and composition. AFINS demonstrated the usefulness of HCES in better understanding and measuring household food security, in particular, and provided important insights about the distribution of food within a countrygeographically, as well as by a variety of household characteristics (e.g., income quintile, ruralurban residence, region/state). The AFINS legacy includes an excellent set of practical guidelines, Measuring Food Security Using Household Expenditure Surveys (Smith and Subandoro, 2007). But AFINS was a one-time assessment of the value of getting beyond FBS. It was not part of a longer term plan to improve food security indicators, and it did not have an institutional home that would facilitate more routine and wide use of the methods it had developed. When the project ended, concerted work on the application of the new methods it developed ended. A third landmark is the FAOs use of HCES since 2008. The FAO approach to measuring food insecurity has long been based on FBS data used in combination with HCES data. The FAO approach made only limited use of HCES data (within the framework a twoparameter log-normal distribution model) to provide information about the distribution of calorie consumption and the adequacy of caloric intake. Motivated by the need to assess food insecurity at a subnational level, FAO initiated a new approach in how it employed HCES, in 2008 it began routinely analyzing these databases to estimate household level food access, food security and inadequate nutrient intake and disaggregated the national analysis to a subnational level. That same year, FAO published a set of guidelines for its new approach, Measuring Hunger at Subnational Levels from Household Surveys Using the FAO Approach n, 2008). In the 4 years since the publication of the manual, (Sibria FAO has analyzed the HCES of 46 countries,3 but it has done little to publicize its use of this new routine approach, publishing the n, 2008; Sibria n et al., analyses of only 12 of the countries (Sibria 2008; Tanzania Bureau of Statistics, 2007). For the FAO, (at least

for the time being) the use of HCES is not intended to supplant, but rather to complement, FBSs.4 The fourth and fth landmarks in the evolution of the use of HCES have been its use by HarvestPlus and A2Z, projects of the Consultative Group in International Agricultural Research (CGIAR) and United States Agency for International Development (USAID), respectively. Both projects further elaborated the basic DAFNEAFINS approach, by: (1) making inferences about individual household members consumption of foods (usually based on the assumption that food is distributed within the household in direct proportion to household members shares of the households total AMCE) and (2) analyzing household micronutrient intakes, which in some of the studies included assessing individual household members adequacy of nutrient intake, a proxy measure of nutrient deciency (Stein et al., 2007, 2008a, 2008b; Fiedler et al., 2008; Fiedler and Macdonald, 2009; Fiedler, 2009; Fiedler and Helleranta, 2010; Fiedler and Adra, 2010, Fiedler et al., 2012a,b; Dop, 2012). Despite this 25-yr long history of using HCES for food and nutrition analyses, there remain many questions and unaddressed issues about its relevance and reliability. The uncertainties stem largely from heterogeneity in the design and implementation of these surveys and the general lack of analytic juxtapositions of directly comparable, individual level data using HCES-like data on apparent consumption with the nutrition gold standards of food consumption measurement methodologies, OWFR and 24HR.

4. Heterogeneity in the design and implementation of HCES HCESs are a collection of surveys that countries use to measure, among other things, food acquisition and/or consumption, and include: (1) Household Budget Surveys (HBSs); (2) Household Income and Expenditure Surveys (HIESs); (3) Living Standards Measurement Study (LSMS); (4) Integrated Household Surveys (IHSs); (5) Core Welfare Indicator Questionnaire (CWIQ); (6) Comprehensive Food Security and Vulnerability Assessment (CFSVA) surveys; and (7) Welfare Monitoring Survey (WMS). These surveys have diverse objectives, including the provision of input into consumer price indices and national accounts, understanding poverty and other dimensions of welfare (including food and nutrition security), and the monitoring of select socio-economic indicators. While there are international guidelines and recommendations for the design and implementation of each of these surveys (International Labor Organization, 2003; Grosh and Glewwe, 2000; United Nations, 1984), they are specic to each type of survey and generally leave considerable exibility to the implementing countries. HCESs reect the diversity of their objectives, and vary in terms of: data capture methods; recall periods; questionnaire design; the length and composition of their food lists; whether they collect only information about food purchases or also about food consumed; what food sources they collected data onfood consumed from own production and/or about food received as payment in-kind and/or food received free-of-charge; and whether they only collect information about the value of food purchases, or also about the quantities of food purchased. Table 1
4 In fact, food consumption data remain an indispensable part for the estimation of at least one of the parameters of the distribution underlying FAOs parametric approach and may serve a key role in validating the one parameter derived from FBSs; viz., the mean of the distribution. Thusabstracting for a moment from issues of ideological disagreements on the use of a parametric approach and the feasibility of improving the data underlying the FBSimproving the quality and availability of food consumption data from HCES may be seen as being consistent with, and not in conict, with the FBS-based approach.

3 Personal communication, May 4, 2011, Seevalingum Ramasawmy, Food Security & Social Statistics, Economic and Social Development Department, Food and Agriculture Organization.

J.L. Fiedler / Global Food Security 2 (2013) 5663

59

Table 1 Key characteristics of a subset of 74 HCES surveys aspects of heterogeneity. Source: Development Data Group, World Bank. The table includes surveys processed by the World Bank under its Knowledge for Change Programfunded research project Modeling Household Consumption Patterns. 1. Sample size (no. of households) Mean Median 2. Data capture method Interview/recall Diary 3. Recall period Mode Range 4. Number of food items Range Mean Median 5. Food measures Quantity and value Only value 6. Food sources/acquisition method Purchases Consumed from own production Received in-kind, free
n

11,958 9555 54, 73% 20, 27% 7 days (20, 37%n) 3 days12 months 213536 212 146 58, 78% 16, 22% 73, 99% 52, 70% 34, 46%

Naska et al., 2007; Nelson et al., 1985 Sekula et al., 2005), (6) mixed results in testing the statistical signicance of differences between the estimated median intakes of a specic macroand micronutrient intakes (Sekula et al., 2005; RambelosonJariseta et al., 2012). While the results are generally encouraging, their external validity is questionable. More comparative analyses are warranted in order to better understand the sources and nature of these differences. To date, the entire body of work using HCES to analyze food and nutrition has consisted exclusivelywith one important exception (Beegle et al., 2010)of retrospective studies. In this work, the role of the analystsand by extension the nutrition communityhas been one of a passive user of secondary data. How might HCES be modied to improve their reliability and precision in measuring food acquisition and consumption?

6. HCES strengths and weaknesses from a nutrition community perspective Strengthening HCES as a tool for analyzing food and nutrition issues requires the more active participation of the nutrition community. The nutrition community will have to be strategic in terms of the processes by which it engages with current HCES stakeholders as well as the substantive issues it identies as being in need of change. In most countries, HCES have been implemented for more than a decade, and there are already many existing HCES stakeholders. The window of opportunity for introducing changes, therefore, will be relatively small, especially in the short term. The nutrition community must recognize that the HCES will need to retain its ability to fulll its primary objectives, and, should expect that changes that entail breaks in HCES data seriesthereby precluding or vitiating the ability to perform trend analysiswill be resisted. How then to proceed? The rst order of business in distilling a prioritized nutrition community agenda for addressing HCES shortcomings is to recognize that the nutrition community is itself a diverse group with diverse information needs. Members of this community will have different views about which are the most important variables that we must get right; they range from measures of nutrition status (including micronutrient intakes and deciencies), to food security, fortication (including biofortication and point-of-use or home fortication), to supplementation and behavior change (including food storage and preparation methods, and dietary habits). Given the breadth of potential nutrition topics that are potentially of interest to the nutrition community how the nutrition community goes about identifying its priorities is not likely to be as simple or straightforward as it might initially appear. Bearing this in mind, the following discussion should be regarded as a rst take and one that is offered by an economist, not a nutritionist. This is but a rst step in a process that needs to involve more, and more diverse, members of the nutrition community. 6.1. The shared HCES agenda Most of the purposes for which HCES food data are collected including poverty analysis, national accounts and the construction of price indiceswould benet from improved measurement of household food consumption. Most HCES data producers and survey sponsors would also welcome the increased use of HCES data and increased visibility of discussions about it by the nutrition community: they demonstrate the relevance of the data, thereby helping to justify HCES budgets and may provide a case for additional investments to make the data more even useful. These considerations suggest that the participation of

Percent of interview-based surveys.

summarizes key survey design characteristics of a sample of 74 HCESs, and testies to their heterogeneity. Given their heterogeneity, it is difcult to make general statements about HCES strengths and shortcomings. Judgments regarding what is a strength or what is a shortcoming of an HCES, depends upon the specic type of HCES survey (e.g., HBS, LSMS, HIES, etc.), country-specic variations in how the specic survey was designed, implemented and processed, and the specic food and nutrition issues or applications of interest. This high level of variability is a major factor explaining why even after 25 years of using HCES for food and nutrition analyses there remain fundamental issues of relevance and reliability.

5. Comparing HCES with the gold standard food consumption measurement methodologies5 A small, but growing, number of studies have juxtaposed 24 h recall and HCES data and have found relatively high levels of consistency between the surveys in quantifying: (1) the proportion of households reported to be consuming most food items (Lambe et al., 1998; Friel et al., 2001; Naska et al., 2001; Lividini et al., 2012; Rambeloson-Jariseta et al., 2012), (2) the proportion of households purchasing most food items (Lambe et al., 1998; Friel et al., 2001; Naska et al., 2001; Lividini et al., 2012; Rambeloson-Jariseta et al., 2012), (3) total caloric intake (Sekula et al., 2005; Naska et al., 2007; Rambeloson-Jariseta et al., 2012), (4) the nutrient density6 of most food items consumed (Rambeloson-Jariseta et al., 2012; Naska et al., 2007; Nelson et al., 1985; Sekula et al., 2005) and (5) the nutrient density of most food items purchased (Rambeloson-Jariseta et al., 2012;
5 As Gibson (2005) makes clear in what is commonly regarded as the bible of nutritionists dietary assessment guidelines, the gold standard methodology itself is subject to a wide variety of pitfalls and shortcomings. See Coates et al. (2012a, 2012b) for direct comparisons of HCES to 24HR and observed-weighed food records using several criteria. 6 Nutrient density is a measure of the nutrient content (grams) per 2000 kcal of the edible portion of a food.

60

J.L. Fiedler / Global Food Security 2 (2013) 5663

the nutrition community is likely to be welcomed by current HCES stakeholders. The nutrition community should target its initial efforts so as to reinforce perceptions of its being a source of value added for the wider HCES community. It can do so by targeting its initial efforts on the identication and implementation of a shared agenda, rather than focusing on considerations that are only or primarily of interest to only the nutrition community. The common ground of all HCES stakeholders is the need to address general HCES shortcomings in the information collected about food acquisition and consumption, so as to improve the accuracy of food- and nutrition-related indicators based on them. The experiences and skills of nutritionists in general, and of dietary assessment experts (a subset of the nutrition community) in particular, make them uniquely qualied to make signicant contributions to improving HCESs by helping to improve three areas that are commonly problematic: the standardization of units of measurement, optimizing the recall period7 (Grosh and Glewwe, 2000; Deaton and Grosh, 2000; Gibson, 2005), better capturing food away from home (Gewa et al., 2007; Ma et al., 2006; Smith, 2011) and improving the food item list (Bermudez et al., 2012; Rambeloson-Jariseta et al., 2012). Most HCESs collect data on the volume or weight of food by using a substantial number of different reporting units. Many of these units are standardized and universal, such as grams, ounces or liters. Others, however, may be regional or local measures (e.g., Ugandas ngorogoro or Kenyas pakaacha or debe), and still others may have common names but may not be standardized in terms of how they are measured (such as a piece or heap, bunch, or stack). Dietary intake specialists knowledge, skills and tools can help to improve methods and contribute to more accurate reporting of the quantities of foods consumed and also with translating to standardized units of measurement. Dietary assessment specialists are also likely to have special insights about how to improve the HCES food list. The length and composition of the food item list has been shown to be an important determinant of the accuracy of food consumption reporting (Grosh and Glewwe, 2000; Deaton and Grosh, 2000; Gibson, 2005). While short, overly aggregated food lists have been shown to contribute to the systematic under-reporting of food consumption, highly detailed, lengthy lists are not the answer either: they may contribute to over-reporting of food consumption, add to interviewer and interviewee fatigue, increase the possibility that the household will refuse to participate in the survey or that it will prematurely terminate the interview (Deaton and Grosh, 2000; Gibson, 2005). A balance must be struck, and the terms of the tradeoff between accuracy and interview time (and costs) need to be elucidated. Dietary assessment specialists can help with striking this balance by making sure that a food list has sufcient detail to enable undertaking nutrient analysis (by distinguishing between forms of foods with differing edible portions and nutrient contents), but are not unnecessarily populated with detailed food types which add little to the substantive nutrient intake estimates. Other common, quantity-related issues that the nutrition community could help to address include: (1) the tradeoffs involved in using interview/recall as opposed to diaries to collect data, and, if using a diary method, whether individual household members should each keep a diary or if one member should keep a diary for the entire household (Sudman and Ferber, 1971;
7 Optimizing the recall period refers to striking the best balance between too short a recall period (which risks inadequately capturing the variability of food items purchased and/or consumed) and too long a recall periodwhich risks memory lapse).

Grootaret, 1986; Grosh and Glewwe, 2000; Beegle et al., 2010; Gibson and Kim, 2012); (2) the tradeoffs inherent in the choice of reference period (Scott and Amenuvegbe, 1991; Grosh and Glewwe, 2000; Gisbon and Kim, 2011; Beegle et al., 2010); (3) how to better deal with seasonality and the identication of usual consumption (Grosh and Glewwe, 2000; Gibson, 2005); (4) the signicance of stocks (Deaton, 1997; Boizot et al., 2001; Gibson and Kim, 2011), and (5) asking about consumption as well as/or in addition to acquisition (Martirosova, 2008; Pereira et al., 2008; Kaara and Ramasawmy, 2008). Some members of the nutrition community have already begun to systematically address some of these issues, by, for instance, incorporating into their 24-HR surveys HCES-like questions about food acquisition (by source) in order to improve understanding of the relative strengths and weaknesses of the two methodologies and to enable crossfertilization (Hotz et al., 2011; Moursi, 2011; Rambeloson-Jariseta et al., 2012, 31, 4244). This practicewhich has been found not to add markedly to the required interview timemust be encouraged (Moursi, 2011; Rambeloson-Jariseta, 2011). Table 2 identies some other specic HCES issues in which nutritionists could make important contributions to strengthening these surveys. 6.2. More nutrition community-specic agenda items Where the nutrition community will confront the greatest challenges in attempting to introduce changes in HCES will be in trying to address shortcomings that are of lesser or no concern to those who are not nutrition analysts. There are two key sets of issues in this agenda: (1) better sorting out the current mix of purchases and consumption that currently characterize most HCES so as to be able to better distinguish food acquisition and food consumption, and (2) to improve the empirical basis for transforming HCES household level data into individual level data. These sets of issues have separate, but overlapping agendas. While HCESs are comprised of a mixture of food consumption and acquisition data, many nutrition-related applications of HCES data use it as a proxy for food consumption. There are two reasons why the quantity of food consumed may vary from the quantity of food acquired during a given period of time: (1) HCESs generally do not collect information on food that is acquired but wasted, given to animals or given away to other people, and thus not consumed by household members, and (2) HCESs do not collect information on food that is acquired during the recall period, but not consumed during it.8 If HCES data on consumption and acquisition in (for instance) the past 7 days are used to estimate consumption in the past 7 days, and there is some wasting or giving away food, unadjusted HCES-based estimates will result in over-estimating consumption. Similarly, relying on unadjusted HCES-based estimates when food has been acquired but not
8 Much nutrition analysis requires having information not only about the quantities of food consumed, but about the time period during which those quantities are consumed; i.e., they require a rate of consumption. For instance, developing a standardized proxy measure of consumption (e.g., grams per day) and assessing food nutrient intake levels by comparing them to the Estimated Average Requirements Institute of Medicine (IOM) of the United States National Academy of Sciences, 2006), entails introducing a time dimension. The construction of such a measure (which is an approximation of usual intake) requires rst standardizing the units in which the food purchases are reported and then dividing by the number of days in the recall period. This approach implicitly assumes that the consumption of all food items occurs at a constant pace throughout the recall period. This is obviously, a gross over-simplication given what is known about the intra-individual variability of consumption of foods, and especially that of particular essential micronutrients, such as vitamin A (Gibson 2005; Institute of Medicine (IOM) of the United States National Academy of Sciences, 2006).

J.L. Fiedler / Global Food Security 2 (2013) 5663

61

Table 2 How nutritionists can help to improve HCES. A. Ways to improve the precision and policy relevance of HCES food data 1. Ensure that the HCES food list captures the most important foods: (a) the most popular foods (b) the most important sources of essential micronutrients (c) the most likely food fortication vehicles (including common consumed foods that contain the fortication vehicle (e.g.,bread, crackers, bakery) 2. Assure that for all important foods in the country, that the food list items are dened to ensure that they can be closely matched to specic FCT entries 3. Develop a more country-specic Food Composition Table that reects how the HCES food items are generally prepared and eaten 4. Analyze intra-country regional variations in HCES food data to prioritize geographic areas in which to conduct 24 HR surveys 5. Help improve understanding of how food is distributed within the the HH to enable improving the common assumption that it is distributed in direct proportion to energy requirements 6. Conduct comparative analysis of 24HR and HCES data to provide insights for improving the precision of HCES B. Where nutritionists can have value-added in helping to enhance the precision of HCES: informing fortication program design and monitoring 1. Include in the food item list popular foods that contain signicant quantities of potentially fortiable foods 2. Identify how the food was purchased (purchased, own production in-kind payment, in-kind from a social program, friend, etc.) 3. Identify where the food was purchased (e.g., open air market, market, kiosk, etc.) 4. Identify who within the household usually consumes potentially fortiable foods 5. Estimate the quantities of each potentially fortiable foods that is consumed outside the home

consumed during the recall period will over-estimate consumption.9 Oftentimes food that is not consumed and is wasted or given away is assumed (oftentimes implicitly) to be quantitatively insignicant, or, it is assumed that some xed proportion of the food (commonly 5%10%) is lost. Similarly, it is oftentimes assumed that the impact of food stocks averages out over the recall period; i.e., the food purchased prior to the recall period but eaten during the recall period is roughly equal to the food purchased during the recall period but eaten subsequentlyi.e., not during the recall period. Given these various (oftentimes only implicit) assumptions and shortcomings of HCES in serving as a proxy for consumption, HCES-derived estimates of consumption are most accurately labeled apparent consumption. The concept of apparent food consumption should be analyzed to better understand the limitations of using acquisition data to proxy food consumption. Among two priority areas to investigate are the signicance of food stocks and the tradeoffs involved in using alternative methods for distinguishing consumption and acquisition. Gibson (2002) has noted that the potential bias will be more problematic for those wishing to estimate individual nutrient intake from these surveys than for those only interested in measuring average expenditures. The nutrition communitys concern about seasonality has parallels: it too is an issue that can be adequately addressed from the perspective of most current HCES stakeholders by implementing the survey over the course of the entire year in order to be able to capture seasonal variations, but that is not an adequate solution for the nutrition analysts (Coates et al., 2012b). Armenia, Cape Verde and Kenya have attempted to collect complete information on both acquisition and consumption, but the effort to distinguish households acquisition and consumption has been of limited success (Martirosova, 2008; Pereira et al., 2008; Kaara and Ramasawmy, 2008). More work is needed on this issue; which is one of fundamental signicance for nutritionists. A more immediate alternative approach that can be used in retrospective analyses of HCES is to constrain estimates of daily consumption to plausible levels, or to adjust implausible daily consumption levels using data from 24-HR where available, or more simply (and prospectively), to introduce changes in questionnaires to capture infrequent bulk purchases of a limited number of staples foods in particular, in order to be better able to accurately characterize
9 In contrast, a countervailing shortcoming of HCESone that contributes to their under-estimating consumptionis that they commonly do not ask about or do not adequately capture food that is purchased for consumption outside of the home, which is discussed elsewhere in this article.

usual consumption or usual intake. Some progress can be made by including a few more questions in the interview that might be asked selectively about specic foods (e.g., common weaning foods) and/or about specic target populations (e.g., children under 5), to ensure that they do not become onerous. The second key issue in the more nutrition community-specic agenda is how to estimate individual nutrient intakes using household level data. While household level analysis is adequate for all or nearly all of the current HCES stakeholders, for nutrition analysts this is a major shortcoming. It is a shortcoming, however, that the nutrition community will likely need to work largely on its own, and that it will be able to make progress on primarily in only an incremental manner. The additional time and costs that would be required to go from collecting data at the household to collecting it on the individual level are likely enormous, and it may very well be a feasibility, rather than simply, a time and cost issue.10 6.3. Country-specic agenda items A third and nal category of HCES shortcomings are those that are country-specic in nature. There are two major sources of these shortcomings: (1) specic questions, and more generally questionnaire design, and (2) survey implementation (data collection and data processing). Given that there is relatively little documentation on the actual implementation of HCES (both eldwork and data processing), the promotion of guidelines and protocols based on state-of-the-art practices in select countries would be helpful.

7. A forward-looking agenda for strengthening HCES The content of the three sets of agendas that have been identied suggest that a two-tiered approach would be useful: one at the global level one and one at the individual country level. The common technical issues of questionnaire design, survey
10 This work could adopt the same approach as noted earlier in the discussion about improving estimates of usual intake: with adding a few simple, selective questions that ask about specic foods and/or specic target populations. The most modest goals are probably those that would improve HCES for analyzing food group diversity indicators. Probably the most ambitious goalwhich is not shared by the entire nutrition communityis to use HCES to conduct nutrient intake analysis. An intermediate set of ambitions would relate to strengthening HCES to improve their usefulness in designing and monitoring food fortication programs. In all of these instances, comparative analysis of HCES and 24-HR surveys are likely to provide insights about how to modify HCES.

62

J.L. Fiedler / Global Food Security 2 (2013) 5663

design, implementation, and data processing could be addressed at the global level. Recognizing that the survey design and methodology has to be adapted to each countrys policy needs, objectives, and strategies and be a reection of each countrys technical and nancial constraints, while building on its own experiences and good practices, the guiding principle in addressing these shortcomings should not be the provision prescriptive solutions, but rather the provision of a menu of alternatives that clearly articulates and, to the maximum extent possible, quanties the tradeoffs involved in selecting from the different options. The many country-specic issues and modications related to a countrys socio-economic, demographic, agro-ecological characteristics and its food culture and dietary patterns and dynamics need to be addressed by each individual country. The global level work has in fact recently begun. In December 2011, the World Bank, the International Household Survey Network (IHSN) and FAO signed a formal Memo of Understanding and embarked on a joint workplan. The initial activity in the workplan calls for conducting a retrospective assessment of the most recent HCES databases of 115 countries. Common criteria for assessing HCES have been identied, and are being used to identify key HCES shortcomings. The report due in the second half of 2012, will include a distillation of lessons and better practices, as well as recommendations of alternatives ways to address common HCES shortcomings, including a global research agenda. But where the real work of strengthening HCES has to be done is at the country level. While the global level work helps to develop focus and momentum by identifying key areas for strengthening the technical aspects of HCES, it is through implementation of the research agendawhich would mark the beginning of taking this work forward in a proactive modethat it could payoff in terms of improving HCES, in general, and, more specically, help provide a more sound foundation of evidence for nutrition policy. To realize this potential, it will be necessary for the World Bank, IHSN and FAO and others to work with countries and conduct rigorous, experimental studies of the major, unsettled measurement issues confronting HCES, including: the measurement of food taken away from home, the relative merits and tradeoffs involved in using the various data collection methods; in designing the optimal (length and composition) food item list; in identifying the optimal length of the recall period and in identifying the additional requirements of improving food consumption versus nutrient intake data. The work of Beegle et al. (2010) in Tanzania is a model for this work. But this proactive phase must move beyond research. It should start with research but also involve the implementation of pilot surveys, including technical support to implement or test and rene recommendations and produce guidelines of best-practices. The agenda must also address the issue of the external validity of any ndings. As Beegle and colleagues have noted, how well a particular survey approach works is often conditioned by the characteristics of a country. This suggests that it would be useful to identify typologies of countries categorized by such conditioning characteristics. These typologies could then be prioritized in terms of where there is the greatest potential for improving HCES and the greatest urgency for doing so. In conclusion, there is growing evidence that HCESs are useful in addressing many important nutrition issues and in understanding the root causes and dynamics of malnutrition and food insecurity. To make HCES a more precise and more powerful tool, it is imperative that the nutrition community becomes more actively involved in their design, implementation and analyses. Developing a more sound empirical basis for nutrition policymaking is essential to making nutrition policy more evidencebased, which, in turn, is essential to making nutrition programs more effective.

Acknowledgment The author gratefully acknowledges The Bill and Melinda Gates Foundations Nutrition and Economic Research Support to HarvestPlus for Grand Challenge #9 Projects Grant OPP52013.

References
Beegle, K., DeWeerdt, J., Friedman, J., Gibson, J., 2010. Methods of household consumption measurement through surveys: experimental results from Tanzania. World Bank Policy Research Working Paper #5501. The World Bank, Washington DC. Bermudez, O., Smitz, M.F., Lividini, K., Fiedler, J.L., 2012. Estimating micronutrient intakes from HIES: an example using the Bangladesh 2005 household income and expenditure survey. Food & Nutrition Bulletin. (Supplement, September). Boizot, C., Robin, J., Visser, M., 2001. The demand for food products: an analysis of inter-purchase times and purchased quantities. Economic Journal 111, 391419. Coates, J., Colaiezzi, B., Fiedler, J., Wirth, J., Lividini, K., Rogers, B., 2012a. Applying food consumption measurement methods for decision-making in food fortication and other nutrition programs. Food & Nutrition Bulletin 33. (Supplement, September). Coates, J., Colaiezzi, B., Fiedler, J., Wirth, J., Lividini, K., Rogers, B., 2012b. Food consumption measurement methods for decision-making in food fortication and other nutrition programs: A Joint GAIN/HarvestPlus/Tufts Working Paper. Global Alliance for Improved Nutrition (GAIN), Geneva. DAFNE-ANEMOS, 2011a. National and Kapodistrian University of Athens, School of Medicine, Department of Hygiene, Epidemiology and Medical Statistics. DAFNE-DAta Food Networking, /http://www.nut.uoa.gr/dafneENG.htmlS (accessed 7.11.11). DAFNE-ANEMOS, 2011b. Software Package for Harmonizing National Household Budget Data from the 28 DAFNE-ANEMOS Countries, /http://www. hhf-greece/dafnesoftweb/S (accessed 7.11.11). Deaton, A., 1997. The Analysis of Household Surveys: A Micro-Econometric Approach to Development Policy. Johns Hopkins University Press, Baltimore. Deaton, A., Grosh, A., 2000. Consumption. In: Grosh, M., Glewwe, P. (Eds.), Designing Household Survey Questionnaires for Developing Countries: Lessons From 15 Years of the Living Standards Measurement Survey. World Bank, Washington D.C. Dop, M.C., 2012. Using HIES data to explore dietary and nutrient intake patterns: a case study of the nutrition transition of Cape Verde. Food & Nutrition Bulletin. (Supplement, September). Fiedler, J.L., Smitz, M.F., Dupriez, O., Friedman, J., 2008. Household income and expenditure surveys: a tool for accelerating the development of evidencebased fortication programs. Food & Nutrition Bulletin 29 (4), 319332. (December). Fiedler, J.L., 2009. Strengthening household income and expenditure surveys as a tool for designing and assessing food fortication programs. Working Paper no. 001 of the International Household Survey Network. Available from: /http://www.internationalsurveynetwork.org/home/index.php?q=activities/ working-papersS (accessed 7.11.11). Fiedler, J.L., Macdonald, B.A., 2009. A strategic approach to the unnished fortication agenda: feasibility, costs and cost-effectiveness analysis of fortication programs in 48 countries. Food & Nutrition Bulletin 30 (4). (December). Fiedler, J.L., Helleranta, M., 2010. Recommendations for improving Guatemalas food fortication program based on household income and expenditure survey (HIES) data. Food & Nutrition Bulletin 31 (2). (June). Fiedler, J.L., Adra, R., 2010. Vitamin A fortication in Uganda: comparing the feasibility, coverage and cost-effectiveness of vegetable oil and sugar. Food & Nutrition Bulletin 31 (2). (June). vel, Y., Mourisi, M., 2011. Comparing the Costs of 24 h Fiedler, J.L., Martin-Pre Surveys and Household Consumption & Expenditures Surveys (HCES): A Nutrition Analysis Perspective. Fiedler, J.L., Babu, S., Bermudez, O., Lividini, K., Smitz, M.F., 2011. Indian social safety net programs as platforms for introducing wheat our fortication: a case study of Gujarat, India. Food & Nutrition Bulletin 33 (1). (March 2012). Fiedler, J., Lividini, K., Bermudez, O., Smitz, M., 2012a. Household Consumption and Expenditures Surveys (HCES): a primer for food and nutrition analysts. Food & Nutrition Bulletin. (Supplement, September). Fiedler, J., Carletto, C., Dupriez, O., 2012b. Still waiting for Godot? Improving Household Consumption and Expenditures Surveys to enable more evidencebased nutrition policies. Food & Nutrition Bulletin. (Supplement, September). Food and Agriculture Organization (FAO), 2003. Food energy methods of analysis and conversion factors. Report of a Technical Working Group. Food and Nutrition Paper 77, 36 December 2002, Rome /http://www.fao.org/DOC REP/006/y5022e/y5022e00.htm#ContentsS (accessed 12.02.12). Food and Agriculture Organization (FAO), 2011. Revising FAOs methodology for measuring hunger. State of Food Insecurity in the World. UN FAO, Rome (2011). Friel, S., Nelson, M., McCormack, K., Kelleher, C., Thriskos, P., 2001. Methological issues using household survey expenditure data for individual food availability estimation: Irish experience in the DAFNE pan-European project. Public Health Nutrition 4 (5b), 11431147.

J.L. Fiedler / Global Food Security 2 (2013) 5663

63

Gewa, C.A., Murphy, S.P., Neumann, C.G., 2007. Out-of-home food intake is often omitted from mothers recalls of school childrens intake in rural Kenya. Journal of Nutrition 137 (9), 21542159. Gibson, J., 2002. Why does the Engel method work? Food demand, economies of size and household survey methods. Oxford Bulletin of Economic Statistics 64 (4), 341359. Gibson, J., 2005. Statistical tools and estimation methods for poverty measures based on cross-sectional households surveys. In: Morduch, J. (Ed.), Handbook on Poverty Statistics. United Nations Statistical Division, (still in draft form). Available from: /http://www.welfare.gov.sd/soruce/HPSCM.pdfS (accessed 13.02.12). Gibson, R.S., 2005. Principles of Nutritional Assessment, second ed. Oxford University Press, New York. Gibson, J., Kim, B., 2012. Testing the infrequent purchases model using direct measurement of hidden consumption from food stocks. American Journal of Agricultural Economics 94 (1), 257270. (January) http://dx.doi.org/10.1093/ ajae/aar135. Grootaret, C., 1986. The use of multiple diaries in a household expenditures survey in Hong Kong. Journal of the American Statistical Association 81 (396), 938944. Grosh, M., Glewwe, P. (Eds.), 2000. Designing Household Survey Questionnaires for Developing Countries: Lessons from 15 Years of the Living Standards Measurement Study, vol. 13. World Bank, Washington DC. (2000). Hotz, C., Loechl, C., de Brauw, A., Eozenou, P., Gilligan, D., Moursi, M., Munhaua, B., van Jaarsveld, P., Carriquiry, A., Meenakshi, J.V., 2011. A large-scale intervention to introduce orange sweet potato in rural Mozambique increases vitamin A intakes among children and women. British Journal of Nutrition (October 10), 114, http://dx.doi.org/10.1017/S0007114511005174. Institute of Medicine (IOM) of the United States National Academy of Sciences, Otten, J.J., Hellwig, J.P., Meyers, L.D. (Eds.), 2006. Dietary Reference Intakes: The Essential Guide to Nutrient Requirements. National Academy Press, Washington DC. International Labor Organization, 2003. Household income and expenditure statistics. Report II of the 17th International Conference of Labor Statisticians, Geneva. Kaara, J., Ramasawmy, S., 2008. Food data collected using acquisition and consumption approaches with a seven-day recall method in Kenyas KIHBS n, R. (Ed.), Deriving Food Security Information from 2005/2006. In: Sibria National Household Budget Surveys: Experiences, Achievements, Challenges. Food and Agriculture Organization of the United Nations, Rome. Klemm, R.D.W., West, K.P., Palmer, A.C., Johnson, Q., Randall, P., Ranum, P., Northrop-Clewes, C., 2010. Vitamin A fortication of wheat our: considerations and current recommendations. Food & Nutrition Bulletin 31 (1), S47S61. (Supplement). Lagiou, P., Trichopoulou, A., 2001. The DAFNE Contributors, 2001. The DAFNE initiative: the methodology for assessing dietary patterns across Europe using household budget survey data. Public Health Nutrition 4 (5B), 11351142. Lambe, J., Kearney, J., Becker, W., Hulshof, K., Dunne, A., Gibney, M.J., 1998. Predicting percentage of individuals consuming foods from percentage of households purchasing foods to improve the use of household budgets in estimating food chemical intakes. Public Health Nutrition 1 (4), 239247. Lividini, K., Fiedler, J.L., Bermudez, O., 2012. An empirical comparison of the HIES and an observed weighed food record conducted in Bangladesh. A Harvest Plus Working Paper. Consultative Group on International Agricultural Research, Harvest Plus/International Food Policy Research Institute and Centro Internacional de Agricultura Tropical (CIAT), Institute of Research for Devleopment (IRD).Washington DC, /http://www.HarvestPlus.orgS. Ma, H., Huang, J., Fuller, F., Rozelle, S., 2006. Getting rich and eating out: consumption of food away from home in urban China. Canadian Journal of Agricultural Economics 54 (1), 101119. Martirosova, D., 2008. Food data collected using acquisition and consumption approaches with daily diaries in Armenias ILCS 2004. In: Sibrian, R. (Ed.), Deriving Food Security Information from National Household Budget Surveys: Experiences, Achievements, Challenges. Food and Agriculture Organization of the United Nations, Rome. Moursi, M., 2011. Unpublished Results from the HarvestPlus 24 Hour Recall Household Survey in Burkina Faso. Naska, A., Vasdekis, V.G.S., Trichopoulou, A., 2001. A preliminary assessment of the use of household budget survey data for the prediction of individual food consumption. Public Health Nutrition 4 (5b), 11581165. (2001). Naska, A., Oikonomou, E., Trichopoulous, A., Wagner, K., Gedrich, K., 2007. Estimations of daily energy and nutrient availability based on nationally representative household budget survey data. The Data Food Networking (DAFNE) project. Public Health Nutrition 10 (12), 14221429.

Nelson, M., Dyson, P., Paul, A.A., 1985. Family food purchases and home food consumption: comparison of nutrient contents. British Journal of Nutrition 54, 373387. Neufeld, L.M., Tolentino, L. Nutritional surveillance: development countries. In: Caballero, B., Allen, L., Prentice, A. (Eds.), Encyclopedia of Human Nutrition, 3rd ed., Oxford, Elsevier Press, in press. Pereira, C., Troubat, N., Sibrian, R., 2008. Food data collected using acquisition and intake approaches in Cape Verdes IDRF 2001/2002. In: Sibrian, R. (Ed.), Deriving Food Security Information from National Household Budget Surveys: Experiences, Achievements, Challenges. Food and Agriculture Organization of the United Nations, Rome. Rambeloson-Jariseta, Z., 2011. Unpublished Results of the A2Z/GAIN 24 Hour Recall Household Survey in Uganda. Rambeloson-Jariseta, Z., Dary, O., Franklin, N., Fiedler, J.L., 2012. Assessing nutrient density using a Household Consumption and Expenditure Survey (HCES) among women of reproductive of age and children 2459 months in Uganda. Food & Nutrition Bulletin. (Supplement, September). Ravallion, M., 2011. Global poverty measurement: current practice and future challenges. PowerPoint Presentation: The National Academy of Sciences, A Sustainability Challenge: Food Security for All. Workshop 2: Measuring Food Insecurity and Assessing the Sustainability of Global Food Systems, February 16, 2011. Venable LLP Conference Center, Washington DC. Available from: /http://sites.nationalacademies.org/PGA/sustainability/foodsecurity/ PGA_060826S (accessed 17.06.11). Scott, C., Amenuvegbe, B., 1991. Recall loss and recall duration: an experimental study in Ghana. Inter-Stat 4 (1), 3155. Sekula, W., Nelson, M., Figurska, K., Oltarzewski, M., Weisell, R., Szponar, L., 2005. Comparison between household budget survey and 24-hour recall data in a nationally representative sample of Polish households. Public Health Nutrition 8 (4), 430439. n, R. (Ed.), 2008. Deriving Food Security Information From National Sibria Household Budget Surveys: Experiences, Achievements, Challenges. Statistics Division, FAO, Rome /http://foodsecuritylink.net/index.php?option=com_ remository&Itemid=27&func=leinfo&id=47S (accessed 20.01.12). n, R., Seevalingum, R., Mernies, J., 2008. Measuring hunger at sub-national Sibria levels from household surveys using the FAO approach. Statistics Division Working Paper Series, ESS 5, FAO, Rome. Smith, L.C., Alderman, H., Aduayom, D., 2006. Food insecurity in sub-Saharan Africa: new estimates from household expenditure surveys. IFPRI Research Report #146. International Food Policy Research Institute, Washington, D.C. Smith, L.C., Subandoro, A., 2007. Measuring food security using household expenditure surveys. Food Security in Practice Series. International Food Policy Research Institute, Washington, D.C. /http://www.ifpri.org/sites/default/les/ publications/sp3.pdfS (accessed 20.01.12). Smith, L.C., 2011. The great Indian calorie debate: An investigation of divergent trends in poverty and undernourishment during Indias rapid economic growth. Consultancy Report Submitted to the FAO. Technical Assistance to NGOs (TANGO) International, Tucson, Arizona. Stein, A.J., Nestel, P., Meenakshi, J.V., Qaim, M., Sachdev, H.P.S., Bhutta, Z.A., 2007. Plant breeding to control zinc deciency in India: how cost-effective is biofortication. Public Health Nutrition 10 (5), 492501. Stein, A.J., Menakshi, J.V., Qaim, M., Nestel, P., Sachdev, H.P.S., Bhutta, Z.A., 2008a. Potential impacts of iron biofortication in India. Social Science & Medicine 66, 17971808. Stein, A.J., Sachdev, H.P.S., Qaim, M., 2008b. Genetic engineering for the poor: golden rice and public health in India. World Development 36 (1), 144158. Sudman, S., Ferber, R., 1971. Experiments in obtaining consumer expenditures by diary methods. Journal of American Statistical Association 66, 725735. Tanzania Bureau of Statistics, Ministry of Finance and Economic Affairs, 2007. Tanzanian National and Nutrition Analysis of Tanzania Household Budget Surveys 2000/1 and 2007. Dar es Salaam. Trichopoulou, A., Lagiou, L. (Eds.), 1997. Methodology for the exploitation of HBS food data and results on food availability for ve European countries. Data Food Networking (DAFNE I), Directorate-General, Science, Research and Development, European Commission, Athens. Available from: /http://www. nut.uoa.gr/downloadsENG.htmlS (accessed 7.11.11). United Nations, 1984. Handbook of Household Surveys. In: Murdoch, J. (Ed.), United Nations, New York. World Health Organization (WHO), 2009. Recommendations on wheat and maize our fortication. Meeting Report: Interim Consensus Statement /http://www.who.int/nutrition/publications/micronutrients/wheat_maize_ fort.pdfS (accessed 22.02.12).