An Assessment Of The Feasibility

An assessment of the feasibility, coverage and cost of fortifying maize meal and sugar with Vitamin A in South Africa M K Hendricks*, R Saitowitz*, J L Fiedler+, T Sanghvi++, I le Roux**, B Makan&, G Hussey*, H Maglagang&&, O Dary+& *Child Health Unit, Department of Paediatrics and Child Health, University of Cape Town +Social Sectors Development Strategies, Sturgeon Bay, Wis., USA ++Basics Support for Institutionalising Child Survival (BASICS) Project, Arlington, Va., USA **Nutrition Subdirectorate, Provincial Administration of the Western Cape, Cape Town &Health Economics Unit, University of Cape Town &&Resident Advisor to the MOST Project, Manila, Philippines +&Institute of Nutrition of Central America and Panama (INCAP), Gautemala S A J Clin Nutr 2001 May Vol 14 No 2 pp 46-52 Abstract Aim To investigate the feasibility, coverage and cost of a national vitamin A fortification programme using maize meal and sugar. Method Key components of a national fortification programme using maize meal and sugar were identified. Only added sugar and not sugar earmarked for industrial use was considered for fortification. The proportion of households consuming maize meal and sugar was determined from the Household Expenditure Survey (1994) of the South African Labour Development Research Unit (SALDRU). Consumption patterns of children aged under 6 years were obtained from two previous surveys and a food consumption survey done in the Northern Province. Different levels of fortification were selected to simulate the impact of vitamin A fortification using the raw data from these three surveys. Maximum safe intake of vitamin A was determined from data on the consumption of maize meal and sugar in the Risk Factor Study (BRISK) conducted among black peri-urban women. Key industry representatives were interviewed to determine the technological requirements needed in a fortification programme. Cost estimates were determined for the various inputs of the programme. Results The proportion of households consuming maize meal and sugar were 78% (range 57 93%) and 91% (range 78 - 97%), respectively. The maximum safe intake of vitamin A was 4 µg retinol equivalents per gram (RE/g) and 44 µg RE/g for fortified dry maize meal and added sugar, respectively. A higher level of vitamin A adequacy was achieved in rural children when fortified maize meal compared with fortified sugar was consumed (84% compared with 48%). Conversely, a higher level of vitamin A adequacy was achieved in peri-urban children with consumption of fortified sugar compared with fortified maize meal (79% compared with 56%). There are seven sugar-refining mills: six

in Kwazulu-Natal, where 90% of the sugar is produced, and one in Mpumalanga, where 10% of the sugar is produced. Six main maize millers process 70% of the maize meal while 100 - 150 small mills process at least 20% of the maize meal. Based on 100% of the RDA, the cost of fortifying sugar would be greater than that of maize meal (R23.2 million compared with R59.1 million). Conclusions This study shows that fortifying maize meal or sugar with vitamin A would be technically feasible and could achieve wide coverage. The costs would be lower for fortifying maize meal and could result in higher levels of vitamin A adequacy in rural children, who are at greatest risk of vitamin A deficiency. However, the impact would be less in peri-urban children, and in reaching the latter group another food vehicle such as wheat flour, in addition to maize meal, would probably need to be fortified. Fortifying sugar will result in higher levels of vitamin A adequacy in peri-urban children but may have less impact on rural children. The centralised nature of the sugar industry compared with the maize meal industry will facilitate monitoring of the fortified product. The final choice of food vehicles for fortification, however, would need to be guided by the results of the national food consumption survey that was conducted by the National Food Consumption Survey Group. Vitamin A deficiency (serum retinol < 20 µg/dl) affects one-third of South African children under the age of 6 years, and has been identified as a significant public health problem, according to a national survey conducted by the South African Vitamin A Consultative Group (SAVACG).1 SAVACG recommended a range of interventions aimed at addressing the problem sequentially. Vitamin A supplementation was suggested as an interim strategy, while food fortification and dietary diversification were recommended as medium to long-term interventions.1 There is a wealth of experience in fortifying foods in both developed and developing countries. Fortification has been important in reducing vitamin A deficiencies in countries such as Costa Rica, El Salvador, Guatemala, Honduras and Panama where sugar is fortified with vitamin A. There has also been relatively successful fortification of margarine with vitamin A in the Philippines, Brazil, Chile, Colombia, El Salvador and Mexico.2 The Department of Health (DOH) is committed to the prevention of vitamin A deficiency nationally, through the implementation of a vitamin A food fortification programme. An understanding of the design and expected costs of such a programme are critical to its gaining acceptance and an appropriate design is essential to the programme achieving its objectives. Knowledge of the appropriate food vehicle and the maximum safe intake of vitamin A, particularly by women of reproductive age who are at risk of potential side- effects, is essential in effectively implementing a vitamin A food fortification programme. The food vehicle chosen must be affordable, accessible, widely consumed among the at-risk groups, and unaffected organoleptically by the fortification process.3 At the time of this study there were no national representative data on the food consumption patterns of the at-risk population (children under 6 years and women of reproductive age). A recent

study showed,4 using available consumption as a proxy for actual consumption, that sugar, maize meal and wheat flour are among the main foodstuffs consumed per capita by South Africans. Similarly, the 1994 Household Expenditure Survey5 conducted by the South African Labour Development Research Unit (SALDRU), which included a national representative sample, confirmed high per capita household consumption of maize meal and sugar. Maize meal, sugar and wheat flour could serve as potential food vehicles for fortification. The aim of this study was to investigate the feasibility, coverage and estimated costs of a national vitamin A food fortification programme using maize meal or sugar. The rationale for the study was to contribute to informing the national vitamin A food fortification policy. Methods Design of a vitamin A fortification programme Key components were identified for a food fortification programme with maize meal and sugar (each considered individually) as the food vehicles. Consumption data on maize meal and sugar In order to develop estimates of the consumption of maize meal and added sugar, the data from previous surveys (secondary data sources) were analysed.5-7 SALDRU's 1994 Household Expenditure Survey,5 a statistically nationally representative sample of households, provided information on the proportion of households consuming maize meal and added sugar and the quantities of each consumed. As the SALDRU data gave no information on the consumption of maize meal or added sugar for children under 6 years old, data from two consumption surveys6,7 of children under 6 years were analysed. Both studies included dietary surveys using a 24-hour recall method. The first study6 included a survey of 118 rural Pedi children aged 3 - 5 years in the Northern Province. The second study7 included a survey of 163 peri-urban children aged 3 - 6 years in the Khayelitsha area of the Western Cape. A food consumption survey (primary data source) was conducted in two rural magisterial districts in the Northern Province.8 The study area consisted of 43 villages. Using a cluster sampling method, 366 children (under the age of 6 years) were selected from 21 villages. Dietary data were collected from each child's mother or caregiver by means of a 24-hour dietary recall method to determine food consumption and nutrient intakes of these children. The dietary data collected from the questionnaires were coded and quantified using the Medical Research Council's Food Composition Tables9 and Food Quantities Manual.10 The data were analysed using a mainframe nutrient programme. Two- thirds (67%) of the RDA was used as a cut-off point for nutritional adequacy.11,12 Level of the vitamin A fortificant Using raw data from the surveys by Steyn et al.,6 Bourne et al.,7 and Saitowitz,8 different levels of fortification with vitamin A (at 50% and 100% of the RDA) in maize meal porridge and added sugar were selected to simulate the impact of vitamin A

fortification. Individual dietary intake data and age- related RDAs were used in undertaking the simulations. Due to no available information on vitamin A stability in these food items during storage and food preparation, the simulations consider only the levels of vitamin A at the point of consumption. In order to determine the maximum safe intake of vitamin A in women of reproductive age, the median, minimum and maximum levels of consumption of maize meal and added sugar were determined in women, using the data of the Risk Factor Study (BRISK Study) conducted among the peri-urban black population of Cape Town.13,14 Based on intake, vitamin A gap and vitamin A losses, different scenarios suggested by the three surveys were then simulated and the maximum safe level at which maize meal or added sugar could be fortified was determined, assuming a maximum intake of 3 000 µg retinol equivalents (REs) per day for women of reproductive age. Identifying other vitamin A fortification programme requirements Food industry-related information was gathered from a review of the literature on the technical requirements of vitamin A food fortification programmes, based specifically on maize meal and sugar. Key representatives of the sugar and maize milling industries were interviewed to determine the potential role of industry in the training, monitoring and evaluation of a national vitamin A food fortification programme. Estimated costs of a vitamin A fortification programme using maize meal or sugar Cost estimates were developed for fortification programmes based on maize meal, or, alternatively, added sugar. The cost estimates included an assessment of: (i) capital costs (e.g. equipment and buildings); (ii) recurrent costs (the fortificant, personnel, pre-mix and fortified food production, monitoring and evaluation). Results Design of a vitamin A fortification programme Consumption data on maize meal and sugar Analysis of the 1994 SALDRU household survey,5 which included 9 128 473 households, resulted in sugar and maize meal being identified as the two most promising potential food vehicles for a vitamin A fortification programme (Table I, [not shown]). According to SALDRU data, 91% and 78% of households nationwide consumed sugar and maize meal, respectively.5 The proportion of households consuming maize meal varied from 53% to 90%. The range of sugar consumption varied from 78% to 97%. The mean amount of sugar and maize meal consumed based on the amount purchased was 52 g/person/day and 104 g/person/day for sugar and maize meal, respectively (see Table I for additional details). Table II [not shown] provides information on the consumption of maize meal porridge and added sugar by children under 6 years, based on the three surveys.6-8 The mean consumption of maize meal porridge in the studies by Steyn et al.6 and Saitowitz8 was 570 (standard deviation (SD) 245) g/day and 590 (SD 261) g/day. This is compared with the lower consumption rates of maize meal porridge in peri-urban children in the study by Bourne et al.,7 namely 184 (SD 162) g/day. The mean consumption of sugar in the two former studies was 15 (SD 16) g/day and 15 (SD 15)

g/day; this is compared with the latter study where consumption rates were higher at 36 (SD 22) g/day. Level of the vitamin A fortificant Based on the mean levels of consumption of maize meal and sugar, 100% of the RDA was equivalent to 2 000 µg RE/100 g for sugar and 270 µg RE/100 g for maize meal porridge in peri-urban children;7 100% of the RDA was equivalent to 2 500 µg RE/100 g for sugar and 80 µg RE/100 g for maize meal porridge in rural children.6,8 Table III [not shown] includes simulations of the impact of fortifying maize meal and sugar, using the data from the three consumption surveys.6-8 Fortification of maize meal porridge with vitamin A at 100% of the RDA level would result in vitamin A adequacy (i.e. an intake of 67% of the RDA) in 94%, 84% and 56% of children in the studies by Steyn et al.,6 Saitowitz8 and Bourne et al.,7 respectively. Fortification of sugar at 100% of the RDA level would result in vitamin A adequacy in 83%, 48% and 79% in the three respective studies. Vitamin A fortification programme requirements The sugar industry A review of industry data and the interviews with industry representatives revealed that the sugar industry in South Africa is highly concentrated, both geographically and economically. Ninety per cent of the sugar produced nationally comes from KwazuluNatal. The remaining 10% is produced in Mpumalanga. There are seven sugar-refining mills in South Africa, with six located in Kwazulu-Natal and one in Mpumalanga. According to the latest available industry data, 1 278 000 metric tons of sugar are consumed annually in South Africa. Seventy- one per cent of this total, namely 915 520 metric tons of added sugar, was sold directly to consumers, while the remainder (362 480 metric tons) was sold for industrial use (C Browne, South African Sugar Association personal communication, 1998). South African sugar industry representatives expressed concern about the illegal importation of sugar. These illegal imports avoid value-added tax as well as the 10% customs duty, which places them at a price advantage relative to South African sugar, and in the event that a South African vitamin A fortification programme is established, it would allow unfortified sugar into the country. An industry association spokesperson also noted that the market for these illegal imports overlaps substantially with the areas of the country where the incidence of vitamin A deficiency is the highest. While favourably disposed to participating in a vitamin A fortification programme, the industry would like the government customs office to improve its enforcement of trade controls, stemming illegal imports of sugar in order to: (i) eliminate what it perceives to be potential unfair competition and (ii) to improve coverage and effectiveness of the proposed vitamin A fortification programme in reducing vitamin A deficiency.

The sugar industry is also concerned about the costs of fortification. In particular, it is concerned about who will bear the costs of the machinery, training, and monitoring and evaluation required to implement the vitamin A fortification programme. The maize industry South Africa's maize milling industry is dominated by six major companies, most of which mill both maize and wheat. These six mills process 70% of the maize meal. A handful of intermediate-sized facilities account for about 10% of milled maise, and the remaining 20% is produced by an estimated 100 and 150 gristing mills. About 2 400 000 metric tons of maize meal was produced for consumption in 1996/97, the latest year for which data were available at the time of this study. Almost all the maize meal produced in South Africa is for domestic consumption (H Zunkel, South African Maize Millers Association - personal communication, 1998). The main kinds of maize meal produced (from the most highly to the least processed) are: (i) 'super', with a low extraction rate and high price; (ii) 'special', with an intermediate extraction rate and intermediate price; and (iii) 'sifted' maize, with a very high extraction rate and low price. A major concern raised by the milling industry is that the small-scale maize millers have operations that will be logistically difficult to identify and monitor. According to industry, the small gristing mills produce sifted maize which is the product generally bought by those most likely to be vitamin A- deficient. Therefore, industry spokespersons point out, inadequate monitoring of these small-scale operations will have a substantially deleterious effect on the coverage and effectiveness of the fortification effort, while simultaneously undermining the big mills' competitive position, thereby encouraging non- compliance/non-participation in the proposed programme. Another industry concern is who will bear the cost of fortification. Food industry activities required for fortification The food industry's role goes beyond 'just' fortifying the chosen product. The food industry also needs to undertake training, monitoring and evaluation (of the quality of the pre-mix and the vitamin A content of the fortified food). Fortification-related training is a one-time, start-up activity. If sugar is to be fortified, the personnel responsible for preparing the pre-mix (2 - 3 persons) would need to be trained, as would the quality assurance personnel at each of the seven refineries in the country (two or three persons per refinery for a maximum of about 20 persons). At least four laboratories in South Africa were identified, based on infrastructure and expertise, as being capable of monitoring the vitamin A content of the pre-mix. Through a tender process, one of these (or another capable institution) could be contracted (on a multiple-year contract) to provide this service. The frequency and specific procedures for conducting these quality assurance tests still need to be identified. The same basic procedures identified for monitoring the sugar pre-mix quality could be used to monitor the vitamin A content of the fortified food (be it sugar or maize flour) at a sample of retail points. A framework will have to be devised for determining how retail

outlets will be identified and sampled. The frequency and procedures for conducting these quality analysis tests will need to be determined. Estimated costs of a vitamin A fortification programme using maize or sugar Different fortificant levels developed for dry maize meal (per 100 g) were 90, 120, 240 and 400 µg RE; for sugar (per 100 g) levels were 1 000, 1 500, 2 000 and 2 500 µg RE (Table IV, [not shown]). The fortificant levels are based on the assumption that no more than 400 µg RE/100 g of vitamin should be added to dry maize meal to remain within safe limits of intake for women. For sugar fortification, there does not appear to be a danger of excess intakes by women within the ranges tested for meeting children's dietary needs; safe levels ranged from 8 to 44 µg RE/g of sugar. Table IV presents the capital and recurrent costs of the fortification programme for maize and sugar assuming the different levels of fortification with vitamin A. In developing cost estimates of the maize-based fortification programme, it is assumed that 36 sites will fortify maize meal (6 large, 10 intermediate and 20 small plants, ideally those serving areas of high vitamin A deficiency. The 6 large millers already have fortification equipment such as dosifiers and feeders used for riboflavin and niacin fortification. The remaining 30 will need to be equipped with dosifiers, mixers and scales. Pre-mix preparation (for sugar-based fortification only), mixing in bulk, and qualitycontrol procedures will follow international guidelines. Estimates assume the costs of monitoring and evaluation. These costs do not take into account the losses of vitamin A during storage and cooking. Based on a vitamin A intake at 100% of the RDA, the cost of fortifying dry maize meal (at 240 µg RE/100 g) will be R23.2 million compared with R59.1 million required to fortify sugar (at 2 000 µg RE/100 g). Discussion This study identified the programme components and inputs regarded as essential in a national vitamin A food fortification programme. Identification of the appropriate food vehicle is of crucial importance in designing a food fortification programme.15 A major constraint on the preliminary design of such a programme has been the lack of information on food consumption patterns of children under 6 years and women of reproductive age. The former is likely to be addressed by the findings of the national food consumption survey, which was undertaken by the National Food Consumption Survey Group Based on our findings, there are a number of reasons favouring the selection of sugar as the potential food vehicle: (i) the findings of the Household Expenditure Survey (SALDRU)5 show that nationwide households more commonly consume sugar than maize meal; (ii) there is little inter-provincial variability in the percentage of households consuming sugar compared with maize meal; and (iii) the centralised nature of the sugar industry would reduce total required capital costs (Table IV) and would facilitate monitoring of the vitamin A content of the fortified product. While nationwide more households consume sugar than maize meal, this is probably a reflection of white households that consume maize meal less frequently than black

households, leading to a 'dilution effect' for maize meal consumption. On the other hand, maize meal fortification could allow for a more focussed programme compared with sugar fortification, as more of the recipients will probably have low vitamin A reserves. Looking specifically at the Northern Province, which has the highest prevalence of vitamin A deficiency, a higher proportion of the children achieved vitamin A adequacy when maize meal as opposed to sugar was fortified with vitamin A. In the two studies by Steyn et al.6 and Saitowitz8 the fortification of maize meal at 100% of the RDA resulted in vitamin A adequacy in 94% and 84% of the children, respectively. This is compared with 83% and 48% adequacy achieved in the two respective studies when sugar was fortified at the same vitamin A level. The latter level of vitamin adequacy obtained in the study by Saitowitz8 is probably a more realistic estimate as consumption patterns for the children were determined by means of a household survey, compared with the study by Steyn et al.6 where dietary patterns were determined for children visiting a central point. The children in the study by Steyn et al.6 may therefore have represented a more select group with higher consumption rates of sugar compared with the children in the study by Saitowitz.8 The higher proportion of children reaching vitamin adequacy following fortification of maize meal is probably due to larger numbers of children consuming maize meal than sugar. This is an important factor to be considered in choosing the appropriate food vehicle for fortification as these children represent the group at greatest risk of vitamin A deficiency, as was found in the SAVACG study.1 Conversely, in the study by Bourne et al.,7 vitamin A adequacy was achieved in a higher proportion of the children consuming sugar than maize meal (79% versus 56%). This is evidence of the transitional dietary patterns of peri-urban communities with their higher intakes of refined carbohydrates. The fortification of maize meal will achieve less impact in children from these communities. These studies, however, were not based on a national probability sample and must, therefore, be regarded as providing only a starting point for estimating the food consumption-related programme parameters, rather than serving as the basis for a definitive programme design. These findings show that fortification of maize meal with vitamin A may achieve a high degree of adequacy in rural children, the group at greatest risk of vitamin A deficiency. This may need to be linked to fortification of another food vehicle, e.g. wheat flour, or to targeted vitamin A supplementation in order to achieve comparable adequacy in at-risk children living in peri-urban areas, where maize meal is less commonly consumed. Using different vitamin A fortificant levels, the cost of fortifying sugar appears to be more than twice that of maize meal. The main reason for this difference in cost is due to the higher fortificant levels needed in sugar because of the small quantities of sugar consumed by the vitamin A-deficient population relative to the quantities of maize meal consumed (Table II). The fortification simulations were based on vitamin A intake at the point of consumption. No account was taken of losses of vitamin A during cooking or storage of the fortified product, which could be higher for maize meal than sugar. However, even with relatively greater losses of vitamin A in fortified maize meal, the

costs of fortification will be far less as considerably lower levels of the fortificant are required. Besides determination of the most appropriate food vehicle for fortification and the level of the fortificant, other important pre-conditions need to be met before the implementation of a food fortification programme. These include: political commitment to bearing the costs of the fortification process; political determination of the incidence of those costs (including, if necessary, the development of legislation that favours and/or protects collaborating producers), or, alternatively, the political acceptability of the market's distribution of the incidence of those costs; and a policy to ensure sustained financial support for the programme. The government, industry, consumer agencies and international donor agencies must all play an active role in addressing some of these issues.16 Food fortification must be driven by the creation of a demand for the fortified foods, or, alternatively, by legally mandated changes, when costs are not high and demand is not sensitive to associated price changes. The health and education sectors, agricultural initiatives and consumers have a key role to play in this regard. As the demand is generated, supply must be guaranteed through effective programme implementation and management. In conclusion, this study shows that it would be feasible to implement a national fortification programme based on industry's technical infrastructure and expertise. The selection of either maize meal or sugar as the food vehicles in a food fortification programme would achieve wide coverage. Fortification of maize meal would incur less cost compared with sugar and result in higher levels of vitamin A adequacy in rural children, who are at greatest risk of vitamin A deficiency. However, it may have less impact in reducing vitamin A deficiency in peri-urban children. In order to reach the latter group, another food vehicle such as wheat flour, in addition to maize meal, would probably need to be fortified. The decentralised nature of the maize milling industry, however, makes it difficult to monitor the fortification process. Fortifying sugar will result in higher levels of adequacy in peri-urban children but may have less impact on rural children. The centralised nature of the sugar industry will facilitate monitoring of the fortified product. The final choice of food vehicles for fortification would need to be guided by the results of the national food consumption survey. We would like to thank Drs L Bourne and N Steyn for making their data available to us, Mr H Zunkel (National Association of Maize Millers) and Ms C Browne (SA Sugar Association) for providing industry-related information and the United States Agency for International Development (USAID) for funding the study. References

The South African Vitamin A Consultative Group (SAVACG). Children aged 6 to 71 months in South Africa, 1994: their anthropometric, vitamin A, iron and immunisation coverage status. Isando, Gauteng: SAVACG, 1995. Darnton-Hill I. Overview: Rationale and elements of a successful food- fortification programme. Food and Nutrition Bulletin 1998; 19: 92-100. Administrative Committee on Coordination-Subcommittee on Nutrition (ACC/SCN) state-of-the-art series nutrition policy discussion paper No. 14. Controlling vitamin A deficiency. A report based on the ACC/SCN Consultative Group Meeting on Strategies for the Control of Vitamin A Deficiency, 28 - 30 July 1993, Ottawa, Canada, January 1994. Steyn NP, Robertson HL, Mekuria M, Labadarios D. Household food security - what health professionals should know. S Afr Med J 1998; 88: 75-79. South African Labour Development Research Unit (SALORM), Project for Statistics on Living Standards and Development. South Africans Rich and Poor: Baseline Household Statistics. Cape Town: SALDRU, 1994. Steyn NP, Badenhorst CJ, Nel JH, Ladzani R. Breastfeeding and weaning practices of Pedi mothers and the dietary intakes of their preschool children. South African Journal of Food Science and Nutrition 1993; 5(1): 10-13. Bourne LT, Langenhoven ML, Steyn K, Jooste PL, Laubscher JA, Bourne DE. Nutritional status of 3 - 6 year old African children in the Cape Peninsula. East Afr Med J 1994; 71: 695-702. Saitowitz R. A baseline survey of the nutritional status of children under 6 years living in two rural districts of the Northern Province. Cape Town: Child Health Policy Institute of the Child Health Unit, 1998. Langenhoven ML, Kruger M, Gouws E, Faber M. Medical Research Council Food Composition Tables. 2nd ed. Parowvallei: Medical Research Council, 1991. Langenhoven ML, Conradie PJ, Wolmarans P, Faber M. Medical Research Council Food Quantities Manual. 2nd ed. Parowvallei: Medical Research Council, 1991. Food and Nutrition Board, United States National Research Council. Recommended Dietary Allowances. 10th ed. Washington DC: National Academy Press, 1990. National Research Council. Nutrient Adequacy: Assessment using food consumption surveys. Subcommittee on criteria for dietary evaluation, co- ordinating committee on evaluation of food consumption surveys. Food and Nutrient Board, Commission on Life Sciences. Washington, DC: National Academy Press, 1986. Bourne LT, Langenhoven ML, Steyn K, Jooste PL, Laubscher JA, Bourne DE. Nutrient intake in the urban African population of the Cape Peninsula, South Africa. The Brisk Study. Cent Afr J Med 1993; 39: 238-247. Bourne LT, Langenhoven ML, Steyn K, Jooste PL, Nesamvuni AE, Laubscher JA. Food and meal pattern in urban African population of the Cape Peninsula, South Africa. The Brisk Study. Cent Afr J Med 1994; 40: 140-148. Arroyave G, Dary O. Manual for Sugar Fortification with Vitamin A. Part 1. Guidelines for the Development, Implementation, Monitoring and Evaluation of Vitamin A Sugar Fortification Program. 2nd ed. Washington DC: United States Agency for International Development, 1996.

World Bank. Enriching Lives: Overcoming Vitamin and Mineral Malnutrition in Developing Countries. Washington, DC: International Bank for Reconstruction and Development/World Bank, 1994.