QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 139/1/135    most recent jn.108.092270v1 Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Santika, O. Articles by Ferguson, E. L. Search for Related Content PubMed PubMed Citation Articles by Santika, O. Articles by Ferguson, E. L.

---

Community and International Nutrition

Development of Food-Based Complementary Feeding Recommendations for 9- to 11-Month-Old Peri-Urban Indonesian Infants Using Linear Programming^1,2

³ Southeast Asian Ministers of Education Organization Tropical Medicine (SEAMEO-TROPMED) Regional Centre for Community Nutrition, University of Indonesia, 10430 Jakarta, Indonesia; ⁴ Department of Human Nutrition, University of Otago, 9015 Dunedin, New Zealand; and ⁵ London School of Hygiene and Tropical Medicine, WC1E 7HT London, United Kingdom

^* To whom correspondence should be addressed. E-mail: ufahmida{at}seameo-rccn.org.

	ABSTRACT

TOP ABSTRACT Introduction Subjects and Methods Results Discussion LITERATURE CITED

 
Effective population-specific, food-based complementary feeding recommendations (CFR) are required to combat micronutrient deficiencies. To facilitate their formulation, a modeling approach was recently developed. However, it has not yet been used in practice. This study therefore aimed to use this approach to develop CFR for 9- to 11-mo-old Indonesian infants and to identify nutrients that will likely remain low in their diets. The CFR were developed using a 4-phase approach based on linear and goal programming. Model parameters were defined using dietary data collected in a cross-sectional survey of 9- to 11-mo-old infants (n = 100) living in the Bogor District, West-Java, Indonesia and a market survey of 3 local markets. Results showed theoretical iron requirements could not be achieved using local food sources (highest level achievable, 63% of recommendations) and adequate levels of iron, niacin, zinc, and calcium were difficult to achieve. Fortified foods, meatballs, chicken liver, eggs, tempe-tofu, banana, and spinach were the best local food sources to improve dietary quality. The final CFR were: breast-feed on demand, provide 3 meals/d, of which 1 is a fortified infant cereal; 5 servings/wk of tempe/tofu; 3 servings/wk of animal-source foods, of which 2 servings/wk are chicken liver; vegetables, daily; snacks, 2 times/d, including 2 servings/wk of banana; and 4 servings/wk of fortified-biscuits. Results showed that the approach can be used to objectively formulate population-specific CFR and identify key problem nutrients to strengthen nutrition program planning and policy decisions. Before recommending these CFR, their long-term acceptability, affordability, and effectiveness should be assessed.

	Introduction

TOP ABSTRACT Introduction Subjects and Methods Results Discussion LITERATURE CITED

During the complementary feeding period, breast milk provides <50% of an infant's high nutrient needs for iron, zinc, calcium, thiamin, and riboflavin (3). At this time, complementary foods with a high nutrient density should be provided as snacks and meals (3), because deficits in calcium, iron, and zinc are common in home-based complementary diets fed to young children in developing countries (4,5). Indonesia is no exception. Complementary feeding diets often provide <20% of a child's estimated recommended dietary allowances for iron and zinc (6), with supportive evidence of biochemical iron and zinc deficiency in early childhood (7,8).

To improve complementary feeding practices, the Indonesian Ministry of Health released a brochure outlining 10 practical guidelines for complementary feeding for 4- to 24-mo-old infants. This brochure contains recommendations about breast-feeding, the time of introduction of complementary foods, the viscosity of complementary foods, ways to actively encourage a child to eat, feeding during illness, and hygiene practices. In addition, examples of nutrient-dense complementary feeding diets are provided for infants aged 4–6, 6–12, and 12–24 mo of age (9); however, the nutritional adequacy of complementary feeding diets based on them have not been evaluated. Further, for disadvantaged populations, these nutrient-dense diets may not be affordable and they do not take into account the cultural diversity and differences in food availability that exist throughout Indonesia (10). Affordable, locally contextual complementary feeding recommendations (CFR)⁶ are more likely to result in long-term improvements in complementary feeding practices than general recommendations for all Indonesian children (10–12).

Food-based dietary guidelines are generally formulated via expert consultation, where multiple factors, such as the key nutritional problems, food availability, acceptability, affordability, and food patterns, are simultaneously taken into account (12). This process is complex and time consuming. To help simplify the process, an approach based on linear and goal programming (LP approach) was recently developed (11). This approach formulates CFR that are locally contextual while ensuring that diets based on them come as close as possible to simultaneously meeting the theoretical nutrient requirements (10,11). Nutrients that are likely to remain low in diets based only on local foods are also identified, which is additionally useful information for nutrition program planning and advocacy.

To our knowledge, the LP approach has not been applied for nutrition programs. Before it can be recommended for use within an integrated system of nutrition program planning, its theoretical application requires testing under field conditions. Thus, the objective of this study was to use the LP approach to develop population-specific, food-based CFR for 9- to 11-mo-old peri-urban Indonesian infants and to identify nutrients that would likely remain low in complementary feeding diets based on local foods. The 9- to 11-mo age group was selected because of their high risk of growth faltering (13), the diversity of their diets exceeds that of younger infants, allowing for more substantive dietary change (3,14); and their requirements for key problem nutrients, such as iron, are higher than older infants (15).

	Subjects and Methods

TOP ABSTRACT Introduction Subjects and Methods Results Discussion LITERATURE CITED

 
    Location. The study was conducted in the Bogor Selatan subdistrict of Bogor District, West Java, Indonesia. This subdistrict was purposefully selected because of the high prevalence of malnutrition (categorized as a nutritionally vulnerable area by Bogor District Health Office), the large population of preschool children living in the subdistrict, and the poor quality of local complementary feeding diets (6,16). Within Bogor-Selatan subdistrict, 3 of the 16 peri-urban villages were randomly selected to participate in this study: Pakuan, Cikaret, and Genteng villages.

    Study design. CFR were formulated and evaluated using a 4-phase approach based on linear and goal programming. To define the model parameters, dietary data were collected in a cross-sectional survey of 9- to 11–mo-old infants (n = 100) from November 2005 to April 2006 and food costs were determined via a market survey of 3 local markets. In addition, anthropometric measurements were made (weight and length) by 2 trained anthropometrists and general sociodemographic data were collected using an interviewer-administered questionnaire.

    Ethical approval. Ethical approval was obtained from the Ethics Committee of the University of Otago, Dunedin, New Zealand. Informed, signed consent to participate in the study was obtained from all participating caregivers.

    Sample size and sampling procedures of the survey. One hundred infants were selected to estimate population mean food serving sizes for commonly consumed foods to within ±10% (95% CI), assuming a SD of 50% of the mean serving sizes in this age group and allowing for a 4% rate of attrition. The inclusion criteria for participation were an infant aged between 9 and 11 mo of age inclusively who was not suffering from an illness that changed their food intake on the day of the survey.

    Dietary assessment methods. Quantitative dietary data were collected using a 12-h weighed diet record (WDR) for all foods and beverages consumed between 0600 and 1800 (or until the infant's last meal on that particular day) and a 12-h recall for all foods and beverages consumed from 1800 on that day to 0600 the next morning. For the WDR, 6 trained research assistants who were fluent in the local language weighed (Salter Vista electronic scale model 3010, Salter Houseware; precision ± 0.1 g) and recorded all foods and beverages consumed by the infants. Serving sizes for the 12-h recall were estimated using standard household measures.

In addition to the WDR, a 24-h recall was performed to obtain information on the foods and beverages consumed 1 d prior to WDR day. Finally, qualitative dietary data were collected with a 5-d food tally using a self-administered questionnaire that was checked by 1 trained research assistant every 2 d. The 24-h recall, WDR, and 5-d food tally were analyzed to characterize each child's 7-d food group patterns, create a list of foods consumed by infants in Bogor-Selatan, and measure for each food its average serving size (g/d) and the percentage of infants consuming it.

    Development of CFR. The CFR were developed using the LP approach based on linear and goal programming, which has been described in detail elsewhere (11). In brief, in phase I, a draft set of CFR were formulated using goal programming. The objective function in this model aimed to select a diet that achieved or exceeded a desired nutrient content (15), respected habitual dietary patterns, and remained below a given cost (model no. 1; Table 1). The food group patterns of phase I's optimized diet formed the basis of the draft set of CFR. In phase II, the success of these draft of CFR for ensuring a nutritionally adequate diet was assessed via a series of linear programming models that sequentially minimized (worst-case scenario) and maximized (best-case scenario) selected nutrient levels and diet cost (series of model no. 2; Table 1). In this phase, the key problem nutrients were identified, i.e. nutrients that in the best-case scenario analysis did not achieve 100% of the desired nutrient levels (15), as well as potential problem nutrients, i.e. nutrients that in the worst-case scenario analysis did not achieve 70% of the desired nutrient levels (15). In phase III, optimal food sources of these key and potential problem nutrients were identified (model no. 3; Table 1). The phase III model was modified from the original model described (11) to control the number of foods selected to meet desired nutrient levels. Specifically, the model was transformed from a linear programming model, which minimized the number of servings of food selected, into a goal programming model, which aimed to select foods that came as close as possible to achieving the desired nutrient levels. A constraint limited the number of servings of foods selected. All other nongoal model constraints were identical to those used in the original phase III model (11). In phase IV, combinations of these foods were incorporated into the original draft set of CFR to produce alternative sets of CFR, which were compared (subset of model no. 2) on the basis of their cost, CFR flexibility/acceptability, and worst-case scenario nutrient levels to select a final set of CFR (10,11).

    Data and statistical analysis. The Z-scores for weight-for-age, length-for-age, and weight-for-height of infants surveyed were analyzed using the National Center for Health Statistics, NCHS/WHO growth reference data 1977 in NutriSurvey for Smart, SEAMEO-TROPMED, University of Indonesia, 2005. Underweight, stunting, and wasting were defined as a Z-score –2 ± SD.

The average serving size of each food was defined as its median gram weight intake per day calculated from the WDR data for infants who had consumed that food, i.e. from the consumers' distribution. The weekly frequency of consumption for each food was defined by the 90th percentile (consumer's distribution) for the number of times it was reported in the WDR, the 24-h recall, and the 5-d food tally data. The maximum g/wk that could be selected for one food was 1.5 times its average serving size times its weekly frequency of consumption. A factor of 1.5 was arbitrarily chosen to represent a large serving size. For most foods, it corresponded closely with the 75th percentile of observed intakes except for porridge (close to the 90th percentile).

A fixed quantity of breast milk (554 g/d) was included in all phase I, II, and IV optimized diets (Table 1). The model constraint on breast milk intake of 554 g/d was derived from the difference between the estimated average energy requirements (i.e. 723 kcal/d)⁷ (15) and the observed mean energy intakes from complementary foods among nonmalnourished subjects (–2.00 weight-for-age Z-score +2.00) in this study (i.e. 306 kcal/d) and corresponded well with reported average breast milk intakes of 551 g/d for 9- to 11-mo-old breast-fed infants living in developing countries (5). For each food modeled, its cost/100 g of edible gram weight was the mean cost from the 3 local markets surveyed. The upper cost constraint of 2500 Indonesian rupiahs (Rp)/d was the mean daily expenditure on foods for infants reported in an unpublished study conducted concurrently with the current study in Bogor Selatan (3 group interviews with 15 mothers of 9- to 11–mo-old infants).

The linear and goal programming models were conducted using the solver function in Microsoft Excel 2003. The population-descriptive statistics were analyzed using SPSS for Windows, version 15.0.

	Results

TOP ABSTRACT Introduction Subjects and Methods Results Discussion LITERATURE CITED

 
Background characteristics

    Characteristics of respondents. The average infant age was 10.2 ± 0.9 mo and 53% were boys. Only 1% of the infants were wasted, 12% were underweight, and 15% were stunted. Fifty percent of the mothers had a primary school-level education, 41% had a junior or senior high school-level education, and 9% had a college or university-level education; the majority were housewives (88%). The most common occupation among the fathers was laborer (32%); 22% worked in the informal sector, 22% as private employees, 4% as government employees/pensioners, and only 3% were unemployed.

    Breast-feeding and complementary feeding practices. The majority of infants were currently being breast-fed (85%). Close to one-half of the mothers initiated breast-feeding within 1 h of delivery, although 95% gave prelacteal feeds, particularly of honey (45%) and formula milk (39%). Only 72% of mothers fed their infants colostrum. The median age of introduction of complementary foods was 3.2 mo.

Food pattern data that defined linear and goal programming parameters

    Food frequency, food serving sizes, and food patterns. The infants consumed 51 foods, of which 6 were staples, 4 were plant protein foods, 7 were animal protein foods, 5 were fruits, 9 were vegetables, and 20 were miscellaneous foods (Table 2). The most common foods consumed (>20% of children) were rice, infant cereals, carrots, biscuits, crackers, and soup broth. Despite the availability of cheap plant protein source foods (e.g. tempe and tofu), <20% of the infants consumed them. Similarly, most vegetables and fruits were consumed by <20% of the infants. Fish was not consumed despite its availability, because mothers thought fish caused worm infestations.

The infants generally consumed 2–3 staples/d and 1–3 snacks/d (Table 3) from the fruits, legumes, roots, and miscellaneous food groups. The miscellaneous snacks were consumed more frequently than fruit, vegetable, root, and legume snacks. Protein source foods were consumed 4 times/wk, with animal protein source foods being consumed more frequently than plant protein source foods. Vegetables were consumed 5 times/wk.

The draft CFR formulated in phase I recommended 3 servings/d of staples; 4 servings/wk of all protein source foods, including 3 servings/wk of animal protein foods; 5 servings/wk of vegetables; and 16 servings/wk of all snacks, including 2 servings/wk of fruits. The phase I optimized diet showed that 100% of the 2002 FAO/WHO recommended nutrient intakes (RNI) for most nutrients could be achieved using local foods at a cost of Rp 2,320/d. The notable exceptions were iron (50% of RNI), zinc (97% of RNI), and niacin (76% of RNI), assuming moderate bioavailability for iron and zinc (15) (Table 4, phase I, optimal case). Niacin was not adjusted for the contribution from tryptophan.

In phase III, the nutrient-dense foods selected to improve dietary adequacy were fortified infant cereals, meatballs, chicken liver, eggs, fortified biscuits, tempe-tofu, bananas, and spinach. These foods were then incorporated individually or in combination into the phase I draft CFR to produce 12 alternatives sets of CFR to test in phase IV. The phase IV worst-case scenario results showed that calcium, iron, zinc, and niacin levels remained <70% across all alternative sets of CFR tested, whereas thiamin, riboflavin, vitamin B-6, and folate exceeded 70% in some (Table 4, phase IV, alternative worst case). Of the alternative sets of CFR tested, combination 2 was selected as the final CFR, because it ensured >70% of recommended intakes for all nutrients except iron, calcium, and zinc; its cost was lower than combination 1; its worst-case scenario for vitamin B-6 was higher than combination 3; and it was cheaper than 13 servings/wk of fortified infant cereal (Table 4). The final CFR includes general recommendations about breast-feeding; the number of times meals, snacks, and foods from different food groups should be consumed; and specific recommendations to consume at least a stated number of servings of fortified infant cereal, tempe-tofu, chicken liver, bananas, and fortified biscuits and their serving sizes (Table 5).

	Discussion

TOP ABSTRACT Introduction Subjects and Methods Results Discussion LITERATURE CITED

Our approach, unlike traditional approaches, objectively identifies nutrients that are likely to remain low in local complementary diets, which in this study were iron, zinc, calcium, thiamin, and niacin. These potential problem nutrients were consistent with those that were low in complementary feeding diets in different areas of the world, including rural Indonesia (4,6,11), and with studies showing that biochemical iron and zinc deficiency are common (7,8).

For iron, it also indicated that complementary feeding diets based on local foods will not provide theoretical requirements. Even in the best-case scenario, when iron-dense foods were preferentially selected, the optimized diet provided only 63% of iron requirements (Table 4, best-case scenario for iron) and in the worst-case scenario analysis, the iron content of a diet conforming to the CFR could be as low as 26% of requirements (Table 4, worst-case scenario for iron), thereby indicating alternative strategies such as fortification or supplementation are required to ensure iron requirements are met.

The linear programming approach is particularly useful for countries such as Indonesia where there is district level autonomy in health/nutrition program planning and implementation and geographic and ethnic diversity in food availability and dietary practices. For example, based on our data, the current Ministry of Health of Indonesia guidelines (i.e. feeding 9- to 12-mo-old infants minced mixed porridge of rice, meat, red pumpkin, and minced tofu 3 times/d; snacks, such as biscuit or cake 2 times/d, and the use of oil when cooking infant foods) (9) are not consistent with current dietary practices in Bogor Selatan, where meat was not consumed daily and some recommended foods were not locally available (i.e. red pumpkin). The need for district-specific, food-based recommendations is not unique to Indonesia. To accommodate different dietary patterns, district-specific nutrition education programs were needed in, for example, rural Kenya (17).

Our approach will help nutritionists rapidly formulate CFR that respect local food patterns and identify key problem nutrients and the food sources of them. Animal protein foods were identified as a key food group source of problem nutrients in our study, which is consistent with an earlier study showing them to be the first limiting food group in diets modeled to minimize diet energy content while satisfying the theoretical nutrient requirements of 3- to 6–y-old rural Malawian children (18). Together, these results highlight the need to ensure young children in this community are frequently fed animal source foods, as well as other micronutrient-rich foods, although barriers to their consumption, including food availability, affordability, or local preferences, may need to be addressed.

The linear programming approach alone will not improve local complementary feeding practices. It must be integrated into a comprehensive nutrition program planning system using ethnographic/formative research methods such as the Trial of Improved Practices or Integrated Management of Maternal and Child Illnesses (19,20) to help illuminate the complex barriers and motivating factors of behavior change and produce rigorously tested materials, messages, and approaches with which to encourage CFR adoption. Within an integrated system of nutrition program strategy development, the linear programming approach can serve as a practical tool for program planning that allows rapid, objective formulation and refinement of CFR, identification of key problem nutrients (phase II, best-case and worst-case scenario analysis), and a rapid analysis of the nutritional and cost implications of alternative CFR (phase IV analysis) to inform CFR choice. Further, any future changes in local food prices or availability can be easily remodeled using our approach to produce revised CFR that better reflect the changing situation. As such, policy makers can continually produce practical, dynamic, and flexible CFR that reflect existing situations, a characteristic in agreement with the FAO/WHO report on the preparation and use of Food-Based Dietary Guidelines (9,12).

The soundness of the CFR-formulated linear programming will depend on the accuracy of data defining model parameters, especially those related to important food sources of the key problem nutrients. Further, food composition data and assumptions related to nutrient bioavailability need to be correct. The complexity of the modeling process is another limitation; however, this can be overcome once a user-friendly interface is developed. With a user-friendly interface, nutritionists working for governments (down to the regional office level), universities, or nongovernmental organizations could use our approach to formulate and evaluate their local-specific CFR to help ensure optimal health, growth, and development of young children.

In conclusion, the 4-phase approach based on linear and goal programming is a powerful tool to formulate and evaluate culturally appropriate, population-specific CFR. Its use is recommended for policy makers and nutrition program planners interested in improving current dietary practices, because it objectively formulates population-specific CFR and identifies problem nutrients to justify nutrition intervention efforts that focus on them (10,11,21). For nutrition intervention efforts in Bogor Selatan, the current study provides a set of population-specific CFR that can be further refined via community-based trials based on Trial of Improved Practices/Integrated Management of Maternal and Child Illnesses methodology to assess the community's response to them and their long-term feasibility. It also provides substantive evidence that a food-based approach alone will not ensure dietary iron adequacy, pointing to a need for alternative iron intervention strategies such as iron fortification or the use of complementary food supplements. To effect successful dietary behavior change, the linear programming approach alone is not sufficient. Rigorous testing of messages, materials, and promotional strategies as well as commitment from the public health sectors are required. Its integration with these other program planning methods is recommended (12,20,22,23).

	ACKNOWLEDGMENTS

 
We thank Maria Widjaja for her useful comments on the tool development and Rachael Stevenson for her team support in the field. Eka Rosiyati and Ina Amelia are greatly appreciated for their assistance during data collection.

	FOOTNOTES

 
¹ Supported by University of Otago Research Grant, University of Otago, Dunedin, New Zealand and a grant from SEAMEO-TROPMED Regional Centre for Community Nutrition, University of Indonesia.

² Author disclosures: O. Santika, U. Fahmida, and E. L. Ferguson, no conflicts of interest.

⁶ Abbreviations used: CFR, complementary feeding recommendation; LP, linear programming; RNI, recommended nutrient intakes; Rp, Indonesian rupiah; WDR, weighed food record.

⁷ 1 kcal = 4.184 kJ.

Manuscript received 27 May 2008. Initial review completed 8 July 2008. Revision accepted 12 October 2008.

	LITERATURE CITED

TOP ABSTRACT Introduction Subjects and Methods Results Discussion LITERATURE CITED

 

1. UNICEF and Government of Indonesia (civil society partner). Challenge for new a generation: the situation of children and women in Indonesia. Jakarta: UNICEF; 2000.

2. WHO/UNICEF. Global strategy for infant and young child feeding. Geneva: WHO; 2003.

3. WHO. Complementary feeding of young children in developing countries: a review of current scientific knowledge. WHO/NUT/98.1. Geneva: WHO; 1998.

4. Gibson RS, Ferguson EL, Lehrfeld J. Complementary foods for infant feeding in developing countries: their nutrient adequacy and improvement. Eur J Clin Nutr. 1998;52:764–70.[Medline]

5. Dewey KG, Brown KH. Update on technical issues concerning complementary feeding of young children in developing countries and implications for intervention programs. Food Nutr Bull. 2003;24:5–28.[Medline]

6. Latief D, Fallah TS, Sunawang. Program ASI Ekslusif dan Makanan Pendamping Air Susu Ibu (MP-ASI) (Exclusive Breast feeding and complementary food). Kumpulan makalah diskusi pakar bidang gizi tentang ASI, MP ASI, Antropometri dan BBLR. Bogor: Persagi, LIPI dan Unicef; 2000.

7. Dijkhuizen MA, Wieringa FT, West CE, Muherdiyantiningsih, Muhilal. Concurrent micronutrient deficiencies in lactating mothers and their infants in Indonesia. Am J Clin Nutr. 2001;73:786–91.[Abstract/Free Full Text]

8. Fahmida U, Rumawas JSP, Utomo B, Patmonodewo S, Schultink W. Zinc-iron, but not zinc alone supplementation, increased linear growth of stunted infants with low haemoglobin. Asia Pac J Clin Nutr. 2007;16:301–9.[Medline]

9. Departemen Kesehatan RI. 10 Petunjuk praktis penyiapan makanan untuk usia 4–24 bulan. Brochure. Departemen Kesehatan, Dirjen Pembinaan Kesehatan Masyarakat, Direktorat Bina Gizi Masyarakat. Jakarta: Depkes; 1999.

10. Ferguson EL, Darmon N, Briend A, Premachandra IM. Food-based dietary guidelines can be developed and tested using linear programming analysis. J Nutr. 2004;134:951–7.[Abstract/Free Full Text]

11. Ferguson EL, Darmon N, Fahmida U, Fitriyanti S, Harper TB, Premachandra IM. Design of optimal food-based complementary feeding recommendations and identification of key "problem nutrients" using goal programming. J Nutr. 2006;136:2399–404.[Abstract/Free Full Text]

12. FAO/WHO. Preparation and use of food-based dietary guidelines. Geneva: WHO; 1998.

13. Dewey KG. The challenges of promoting optimal infant growth. J Nutr. 2001;131:1879–80.[Free Full Text]

14. WHO. Complementary feeding: family food for breastfed children. WHO/NHD/00.1. Geneva: WHO; 2000.

15. FAO/WHO. Human vitamin and mineral requirements. Report of a joint FAO/WHO expert consultation, Bangkok, Thailand. Rome: Food and Nutrition Division, FAO; 2002.

16. Dinas Kesehatan Bogor (Bogor Health Office). Profil Kesehatan Kota Bogor. Bogor: Dinkes Bogor; 2003.

17. Mwadime RK, Muita JW, Omwega AM, Havemann K. Dietary considerations in designing nutrition intervention programmes in rural areas of Kenya. East Afr Med J. 1995;72:442–8.[Medline]

18. Darmon N, Ferguson E, Briend A. Linear and non-linear programming to optimize the nutrient density of a population's diet: an example based on rural Malawian preschool diets. Am J Clin Nutr. 2002;75:245–53.[Abstract/Free Full Text]

19. WHO. Integrated management of childhood illness IMCI adaptation guide. Part 3. The study protocols. Geneva: WHO, Department of Child and Adolescent Health and Development; 2002.

20. Dickens K, Griffiths M. Designing by dialogue. A program planner's guide to consultative research for improving young child feeding. Washington, DC: Academy for Educational Development; 1997.

21. Briend A, Darmon N. Determining limiting nutrients by linear programming: a new approach to predict insufficient intakes from complementary foods. Pediatrics. 2000; 106 Supplement 5:1288–9.[Free Full Text]

22. Hulshof KFAM, Valsta LM, Welten DC, LoÈwik MRH. Analytical approaches to food-based dietary guidelines in the European setting. Public Health Nutr. 2001;4: 2B:667–72.

23. Anderson GH, Zlotkin SH. Developing and implementing food-based dietary guidance for fat in the diets of children. Am J Clin Nutr. 2000;72 Suppl:S1004–409.

This Article Abstract Full Text (PDF) All Versions of this Article: 139/1/135    most recent jn.108.092270v1 Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Santika, O. Articles by Ferguson, E. L. Search for Related Content PubMed PubMed Citation Articles by Santika, O. Articles by Ferguson, E. L.