3. Food security information systems using combined methods

An international Nutrition Index: concept and analyses of food insecurity and undernutrition at country levels

Doris Wiesmann
Center for Development Research (ZEF)
Bonn, Germany

A review of international monitoring processes shows that food security and nutrition do not figure very prominently on the development agenda, despite the recognized importance and reconfirmed commitment to end hunger and malnutrition. The lack of a commonly accepted measure of food security and nutrition has impeded political advocacy and prevented beneficial synergies that might arise from integrated approaches. An international Nutrition Index (NI) has been developed in accordance with international comparison theory to demonstrate how this gap can be filled.

The theory of international comparisons formulated by Szilágyi (2000) states that there are three different types of socio-economic categories: (1) those that can be quantified by a single indicator (e.g. industrial production, dietary energy intake); (2) those that cannot be expressed in a scalar term according to pure theory, but for which conventions regarding a generally accepted indicator exist (e.g. GDP); and (3) socio-economic categories to which no a priori target variable can be assigned, for example infrastructure, technical level and health. Obviously, there is no single variable capturing the different aspects of food security, i.e. food safety, sufficient dietary energy and nutrient content of accessible food, as well as the subjective element of food preferences (see definition in FAO, 1996) and nutrition security, which includes the biological utilization of food (see Frankenberger et al., 1997). Food and nutrition security can be classified as not directly observable, abstract variables belonging to the third category. Composite indices attempt to capture such multifaceted phenomena, and international comparison theory describes rules for their construction concerning the selection of empirical indicators, operations carried out with them and the final aggregation to an index.

For the purpose of index calculation, the concepts of food and nutrition security are narrowed down to an operational definition of a “good nutrition situation”, including the dimensions of food adequacy, adequate nutritional status and survival. With regard to the selection of empirical variables, data availability at the international level is limiting. Therefore, only three indicators are chosen, namely one for each dimension. These are: (1) the percentage of undernourished in the population estimated by the FAO method (PUN), expressing energy deficiency; (2) the prevalence of underweight (percent of children with a weight-for-age score below -2 standard deviations) in children under five (CUW), indicating anthropometric shortfalls; and (3) the under-five mortality rate (deaths per 100 live births) (CM) as a proxy for premature death related to malnutrition. No non-linear transformation was applied to these components, and standardization was omitted since the high redundancy of standardized and nonstandardized versions could be demonstrated. Weights for aggregation were derived from a principal-component analysis, resulting in the following formula for the NI:

NI = 100 - (0.33 PUN + 0.32 CUW + 0.35 CM).

Regarding possible applications, the main purpose of the NI is policy advocacy. It is less useful for policy evaluation and not suitable for in-depth causal analysis. The NI has been used for an international ranking of 97 developing countries and nine countries in transition. Going beyond a purely descriptive comparison of NI trends and levels, the relation of the NI to other international indicators was analysed. Rank correlations of the NI and indicators of micronutrient deficiencies, which could not be included into the index owing to scanty data availability, were determined. Compared with its components, the NI performed quite well in terms of correlations with the goitre rate (indicating iodine deficiency), anaemia (reflecting iron deficiency), and subclinical and clinical vitamin A deficiency (Table 1).

The sensitivity of the NI to distribution and income poverty was likewise examined by means of rank correlations. Corresponding to similar results for the Human Development Index (HDI) in other studies, the analysis demonstrated that the NI bears an “equi-distributional bias”, which means that it is significantly more sensitive to inequality than GNP per capita. The NI also showed significantly higher correlations with four common measures of absolute poverty (proportion of the population below $1 a day, poverty gap at $1 a day and the respective two indicators for the poverty line at $2 a day) than GNP per capita and the HDI did (Table 2). The fact that the NI is more sensitive to income poverty than currently prevailing development indicators and even the single indicators of which it is composed (see Table 2) can be considered a noteworthy comparative advantage of the newly created index. Taking the NI seriously instead of aiming solely at maximization of GNP or HDI would compel decision-makers to address poverty more effectively.

Considering the pros and cons of the potential political impact of the NI or a similar composite measure, lessons can be learned from the application of the HDI. The HDI and related indices have proved to be powerful tools for policy advocacy because they are more “eye-catching” than a group of single indicators. In fact, the potential of the NI for advocacy should be its main virtue. Like the HDI, an international NI might promote a sense of competition and thereby enhance national efforts to combat hunger and malnutrition. The index would eventually encourage more comprehensive data collection, but it might also create incentives for data manipulation (see Wiesmann, in press for a detailed discussion and further information on an international Nutrition Index).

References

FAO. 1996. Rome declaration on world food security and world food summit plan of action. Rome.

Frankenberger, T.R., Oshaug, A. & Smith, L.C. 1997. A definition of nutrition security. Atlanta, GA, CARE (mimeo).

Szilágyi, G. 2000. International comparison theory - human development. Paper presented at the IAOS 2000 conference on “Statistics, Human Rights and Development”, Montreux, 4-8 September 2000.

Wiesmann, D. in press. An international Nutrition Index - concept and analyses of food insecurity and undernutrition at country levels. Frankfurt/Main, Peter Lang Academic Publishers.

Nutrition surveillance to monitor nutrition and food security: indicators, interpretations and action

Martin W. Bloem, Lynnda Kiess,
Regina Moench-Pfanner, S. de Pee,
H. Torlesse, M. Sari and S. Kosen
Helen Keller International
Jakarta, Indonesia

Experiences from Indonesia and Bangladesh show that surveillance systems with a combination of nutrition, diet and economic indicators are an effective way to measure and monitor nutrition, health and food security. In Indonesia, Helen Keller International (HKI) has followed measurements of household expenditure, dietary intake, anthropometry and micronutrient status through the ongoing HKI/Government of Indonesia Health and Nutrition Surveillance System to monitor the impact of, and recovery from, the economic crisis. Analyses using both epidemio-logical and economic techniques provide a clear picture of how the Asian economic crisis impacted on nutrition and health. Indicators of nutritional status were identified that would be the most responsive to food accessibility, diet quality and nutritional and micronutrient status. Analyses also show that micronutrient status and dietary quality are responsive in the postcrisis recovery period, particularly among the urban poor. The prevalence of child anaemia in urban Jakarta has declined during the postcrisis period, and this decline is paralleled by an increased micronutrient intake, estimated by household food expenditure, household consumption and individual dietary intake. An additional economic analysis is being conducted to identify the groups who were more vulnerable to the crisis and to identify coping mechanisms that were employed by households to minimize the negative impact of the crisis. Results from the surveillance were used to direct social and economic programmes to the urban poor, to revitalize the vitamin A capsule pro-gramme, to introduce fortified complementary foods and to re-examine food policies. Similarly, in Bangladesh, rice price, household food and nonfood expenditure, dietary intake and child anthropometry from ten years of surveillance data have been used to explore the relationships between food policy and nutrition. Lower rice prices were associated with a greater expenditure on, and intake of, micronutrient-rich foods and this in turn with a decline in the prevalence of child malnutrition. These innovative experiences in Bangladesh and Indonesia demonstrate the value and feasibility of incorporating a combination of nutrition, economic, health and dietary indicators into nutrition surveys and surveillance systems. This comprehensive approach to tracking nutrition and health helps to translate research findings into programme and policy actions.

The use of household budget survey data for assessing food disparities within and between populations: case studies of 13 European countries

Antonia Trichopoulou,
on behalf of the DAFNE team
University of Athens
Athens, Greece

Disparities in dietary patterns are observed both across and within European countries and stem from the multitude of cultural, economic, geographic and psychosocial factors that can influence food behaviour. Hence, there is a need in Europe for a dietary assessment tool that encompasses built-in mechanisms of continuity and extensive coverage, providing at the same time a comparable flow of information. Sound nutrition surveillance systems are prerequisites for detecting the main eating patterns and their time trends in the general population. Furthermore, there is a need to identify which segments of the population are inappropriately nourished (e.g. by geographical and socio-economic characteristics), when (e.g. in particular seasons) and why (e.g. purchasing capacity, ecological and seasonal constraints) they are inappropriately nourished and how these factors evolve with time.

Comparable information on food availability across countries can be provided by data collected in the household budget surveys (HBS). The HBS are periodically undertaken using nationally representative samples of households and allow cross-country comparisons of the dietary choices of the population and of population subgroups. The use of the HBS-derived data for nutritional purposes has been evaluated in the context of the European Union funded DAFNE (Data Food Networking) initiative, which has demonstrated that comparisons at the international level using food and socio-economic data from the national HBS are feasible (Trichopoulou and the DAFNE contributors, 2001). The methods and results of this evaluation are described below.

Daily individual food availability was estimated using data collected in the 1990s through the HBS of 13 European countries: Belgium, France, Germany, Greece, Hungary, Ireland, Italy, Luxembourg, Norway, Poland, Portugal, Spain and the UK. The harmonization of the food and related HBS data of the 13 countries was undertaken in the context of the DAFNE project. The process included iterative cross-coding and the establishment of operational criteria for the classification of food and socio-economic variables (Lagiou and Trichopoulou, 2001).

Individual food availability for 57 analytical and 15 main food groups was estimated from harmonized cross-country data. Individual availability was estimated without making allowances for edible proportion and under the assumption of equal distribution of food within the household and during the survey period.

The dietary information available in the DAFNE data bank supports previous findings on the variation in the food habits of Europeans. The availability of meat and meat products, for example, exceeds 180 g/person/day in Hungary, Poland and Luxembourg, while in Norway and the UK, it is around 130 g/person/day. Variation also exists in the type of meat consumed: Greeks seem to prefer veal meat (52 g/person/day), Spaniards show a preference towards poultry (58 g/person/day), Hungarians preferably consume pork, and Germans and Poles prefer processed meat products.

The total added lipid availability also varies across countries - from 84 g/person/day in Greece to 26 g/person/day in the UK. In the Mediterranean countries, vegetable oils, olive oil in particular, provide practically all of the added lipid availability. Vegetable fat is the prevailing lipid in central and northern Europe, whereas in eastern European countries, butter and animal fat also figure prominently.

With respect to fresh and processed vegetables, Greece is leading with a total availability of 271 g/person/day, whereas Norway is trailing with 109 g/person/day. Spain is leading for fresh and processed fruits with a total availability of 308 g/person/day, whereas Ireland and Poland are at the lower end.

The DAFNE data bank provides evidence of the presence of food inequalities among segments of the populations, even in the affluent regions of Europe. Educational attainment has been reported as the strongest and most consistent indicator in assessing socio-economic differentials (Liberatos, Link and Kelsey, 1988). Based on data retrieved from the DAFNE data bank, a low educational level and, by extension, a low socio-economic status are generally associated with a lower consumption of fresh fruit, low-fat milk, animal lipids and soft drinks, and increased consumption of sugar products, potatoes, cereals (including bread), meat and meat products. In southern European countries, a lower educational level is associated with a higher availability of fresh vegetables. The observation that fresh vegetable availability decreases with the educational level of the household may reflect the adherence of less well-educated southern European households to traditional dietary choices. In northern European countries, more educated households residing in urban areas are observed to follow healthier dietary practices. Norwegians, for example, who live in urban areas tend to lower the overall consumption of lipids while increasing their vegetable oil consumption. Easier access to health information and current nutrition advice may be partly responsible for such observations.

Another indicator commonly used as a proxy to socio-economic status is the food-purchasing capacity of the household, often expressed as the ratio of the household’s food expenses over the total household’s expenses. High values of this ratio have been shown to indicate a low socio-economic class or low income. Using data from the Greek HBS undertaken in 1998-99, households of a lower socio-economic status, as expressed by their food expenditure ratio, report a greater availability of cereals, vegetable lipids, potatoes, fresh vegetables, legumes, fish, meat and meat products.

Barriers responsible for the suboptimal efficiency of current nutrition campaigns have been identified in several studies (Cox et al., 1998; Williams, 1995). Strategies should be tailored according to people’s attitudes towards eating. The DAFNE data bank may contribute to the formulation and monitoring of a nutrition policy and could prove useful to health professionals to ensure that health benefits are conferred upon the population subgroups who need them the most.

Acknowledgements

The DAFNE initiative was supported by the European Commission in the context of the “Cooperation in Science and Technology with Central and Eastern European Countries”, the “Agricultural and Agro-Industry, including Fisheries” (AIR), the “Agriculture and Fisheries - FAIR”, the “COST 99 - Food Consumption and Composition Data” programmes and the “Health Monitoring Programme” of DG-SANCO.

The DAFNE (DAta Food NEtworking) team: Greece (University of Athens, coordinating center): A. Trichopoulou, A. Naska, A. Antoniou, H. Economou Belgium (University of Ghent): AM. Remaut, AP. Cueto Eulert, France (AFSSA, Maisons-Alfort): JL. Volatier, J. Maffre, Germany (JL University Giessen): IU. Leonhauser, Hungary (National Institute of Food Hygiene and Nutrition): G. Zajkas, Ireland (University College, Galway): S. Friel, J. Dalton, Italy (Istituto Nazionale di ricerca per gli Alimenti e la Nutrizione): A. Turrini, S. Barcherini, S. Martinez, Luxembourg (Interdisciplinary Nutrition Policy Group): A. Schmitt, Norway (University of Oslo): K. Trygg, E. Mork, K. Lund-Iversen, Poland (National Food and Nutrition Institute): W. Sekula, Portugal (University of Porto): MD. Vaz de Almeida, S. Rodrigues, Spain (University of Madrid): O. Moreiras, C. Cuadrado, United Kingdom (King’s College, London): M. Nelson, D. Rimmer, S. Burr.

References

Cox, D.N., Anderson, A.S., Lean, M.E.J. & Mela, D.J. 1998. UK consumer attitudes, beliefs and barriers to increasing fruit and vegetable consumption. Public Health Nutr., 1(1): 61-68.

Lagiou, P. & Trichopoulou, A. 2001. The DAFNE initiative: the methodology for assessing dietary patterns across Europe using household budget survey data. Public Health Nutr., 4(5B): 1135-1141.

Liberatos, P., Link, B.G. & Kelsey, J.L. 1988. The measurement of social class in epidemiology. Epidemiol. Rev., 10: 87-121.

Trichopoulou, A. and the DAFNE contributors. 2001. The DAFNE databank as a simple tool for nutrition policy. Public Health Nutr., 4(5B): 1187-1190.

Williams, C. 1995. Healthy eating: clarifying advice about fruit and vegetables. Br. Med. J., 310: 1453-1455.

Using anthropometric data: case study of a nutrition management information system in Ethiopia

Jean Gladwin
London School of Hygiene and Tropical Medicine
London, UK

Introduction

The Ethiopian government’s Disaster Prevention and Preparedness Commission (DPPC) is responsible for improving short-term food security in emergency-affected areas, and the DPPC’s Early Warning Department (EWD) provides information and supports resource-targeted decision-making.

Ethiopia recently achieved stability after years of conflict; however, the democratic government was accused of favouring particular regions. Existing information systems (ISs) were proving inadequate, and so the DPPC was unable to plan and monitor supplementary food allocation. In 2000, the EWD established an Emergency Nutrition Coordination Unit (ENCU) to “facilitate the use of good quality nutrition and nutrition-related information to enable the rational use of food aid and other resources in emergency-affected areas”. Consequently, attempts to improve survey quality and develop a nutritional IS were initiated. This paper describes the early development of the nutrition management information system using an organizational change model and indicates the challenges and lessons learned.

Management information system design

Management information system (MIS) design and development should be undertaken within the context of the organization where the management decisions are made, to provide useful and beneficial information. Previous MIS development research indicates the usefulness of a dynamic equilibrium model identifying five forces: Structure; Management styles, tools and procedures; Individuals and roles; Policy; and Information management strategies or IS within organizations that need to be aligned (Gladwin, 1999). The initial design of the Ethiopian nutrition MIS was based on this model and implied that these forces needed to be examined and possibly changed in the DPPC to produce high-quality targeting decisions. Wider country and international contexts have to be accounted for. A participatory approach involving stakeholders should be used; this begins by reviewing how management decisions, e.g. planning and resource targeting, would benefit from nutrition and nutrition-related information.

Features

The features of the nutritional MIS are described below.

Institutional arrangements

Surveys undertaken by the EWD and partners feed into the national-level ENCU. Plans for local ENCU offices are under development.

Data collection and processing

Anthropometric data with short-term food security data from surveys conducted by agencies form the database. Surveys are undertaken in areas of food emergency. The preferred sampling frame is 30 clusters of 30 children each, especially recommended when the surveys are conducted by agencies with inexperienced staff or those without nutritionists or epidemiologists on their team. ENCU supports agencies lacking experience. More experienced nutritionists/epidemiologists will be able to design their own sampling frames taking into consideration local conditions.

Anthropometric indicators include: global acute malnutrition (the proportion of children with a weight-for-height Z-score of -2 or below or observed oedema). Other indicators include: mortality rates; household food security, e.g. receipt of general rations; crop outcomes; diverse livelihoods; climate and geographic public health indicators; unusual migration patterns; etc. Processing is undertaken by the collecting agency, and survey reports are forwarded to ENCU.

Presentation and information outputs

The national nutrition database set up in the ENCU utilizes a geographical information system (GIS). Updates of summarized data in the form of maps by wereda (a group of villages), highlighting areas of concern, are sent to operational partners and government officials on a regular basis.

Use of the information

The survey data are to be used locally to assess needs, plan, evaluate or advocate, and are also used by DPPC for national-level planning. Typical uses include annual fundraising calls, multiagency biannual food security assessments, annual selection of weredas to receive supplementary food, and allocation of a “Rapid Response Reserve” by the Multi-agency Food Security Coordination Group. Difficulties arose when the DPPC was reluctant to allocate Rapid Response Reserve food to areas where nutritional levels were not below cutoffs, even when other indicators identified need. This indicates an inability of guidelines to cater for every eventuality, a need to accept that nutritionists are necessary to interpret collected data and a need for management tools such as guidelines to be aligned to the IS.

The reluctance of the central government to relinquish control of resources to the regional and zonal levels became clear over discussions on use of the data, as the central EWD often conducted their own nutritional assessments rather than accept the local government assessment. Furthermore, there was insufficient decentralization of decision-making to subna-tional levels, and so there is clearly a need for skills to be developed at the subnational level.

Development and challenges

In practice, the development of a nutrition MIS in Ethiopia is still at an early stage, is complex and presents challenges. Many of the challenges contributing to the situation remain, such as dealing with diverse livelihoods, involving the community and linking with other IS.

Organizational structure

Identifying decision-makers was difficult. National decentralization policy devolved some food-aid decisions to regional, zonal and wereda levels, but senior managers were loath to lose control of food aid.

Policy and strategy

The organizational policies, strategies and conceptual models were incorporated into the MIS design. DPPC’s role is to coordinate the emergency response, and the ENCU aims to promote an integrated approach. Thus, morbidity, mortality, food security, public health, social and care environment data are all relevant. A consensus emerged that nutritional status was not an early indicator of food crisis, which has implications for identifying the stage of the food crisis and appropriate interventions.

Management tools and procedures

The people involved in this process lacked the knowledge and experience of conducting surveys. EWD guidelines focused on longitudinal surveys, and not short-term surveys. These guidelines were sometimes used inappropriately, which at times led to poor-quality surveys. New guidelines are being developed, but there is still a conflict between providing adequate methods for inexperienced persons and the flexibility needed by experienced nutritionists.

A coordinated response is essential. A multiagency group was set up at the national level and chaired by EWD, but its powers need expanding, and regional and zonal committees are needed.

Individuals and roles

Staff training focuses on nutrition concepts and the ways in which nutrition and related data are to be collected and processed. Poor staff retention compromises sustainability.

Information management and use

Time is needed to agree upon data collection and processing activities. It is difficult for non-technical people to accept technical interpretations of complex indicators. There is some degree of narrowing of acceptable indicators, so the GIS highlights anthropometric and mortality data, but not the risk indicators that are more difficult to present.

Wider issues

DPPC had a strong need to retain political control. Cash-flow problems owing to complex funding arose.

Conclusion

The Ethiopian MIS is still in the early stages of development, however:

The dynamic equilibrium organizational model was useful. If organizational forces are not aligned, change is necessary to ensure information use. IS development should continue using this at the national and subnational levels.

Stakeholder involvement, including communities, is essential.

National and subnational nutrition MIS should be linked.

Countries accustomed to central government control may have problems with an informational approach to management. Managers using this approach may need management training, training in how to interpret and use information, knowledge of nutrition concepts and the causes of malnutrition, and access to decision-making tools that may lead to the commissioning of nutrition surveys.

References

Gladwin, J. 1999. An informational approach to health management in low-income countries. Ph.D. thesis, Sheffi eld, UK, University of Sheffi eld.