F&N Bulletin Vol 20 No 2 - United Nations University · 2019. 4. 16. · 180 H. I. Tawfeek et al....

Contents

Human nutrition

Maternal dietary intake and pregnancy outcomes in Baghdad, Iraq—H. I. Tawfeek, J. N. Abdulla, and A. H. Rasheed ........................................................................................... 179

Interventions to improve intake of complementary foods by infants 6 to 12 months of age indeveloping countries: Impact on growth and on the prevalence of malnutrition and potentialcontribution to child survival —L. E. Caulfield, S. L. Huffman, and E. G. Piwoz .................................... 183

Salt iodization

Successful start of salt iodization in Laos —M. Coppens, A. Phanlavong,I. Keomoungkhoune, N. T. Dung, R. Gutekunst, M. G. Venkatesh Mannar, and C. Thilly ........................... 201

Field tests for iodate in salt —L. L. Diosady, J. O. Alberti, S. FitzGerald, and M. G. Venkatesh Mannar ..... 208Field tests for iodide in salt —L. L. Diosady, J. O. Alberti, S. FitzGerald, and M. G. Venkatesh Mannar ..... 215

Food science

Effect of tempe and sodium metabisulphite on the microbiological quality, development of rancidity,and sensory quality of Nile perch (Lates niloticus) sausages —V. O. Owino and N. M. Muroki ............. 221

Papers from the Institute of Nutrition at Mahidol University, Salaya, Thailand

The changing climate of health and nutrition in Thailand: A report from the Institute ofNutrition at Mahidol University —K. Tontisirin ........................................................................................ 228

Human nutrition

Nutrifit programme to improve health-related fitness among young Thai schoolchildren—K. Kijboonchoo, W. Thasanasuwan, and U. Yamborisut ......................................................................... 231

Recommending vitamin A–rich foods in southern Thailand—U. Chittchang, S. Jittinandana, P. Sungpuag, V. Chavasit, and E. Wasantwisut ...................................... 238

Sustaining behavioural change to enhance micronutrient status through community- andwomen-based interventions in north-east Thailand: Vitamin A —S. Smitasiri, K.Sa-ngobwarchar, P. Kongpunya, C. Subsuwan, O. Banjong, C. Chitchumroonechokchai,W. Rusami-Sopaporn, S. Veeravong, and S. Dhanamitta ................................................................................. 243

Food science

Induction of mutation in Drosophila melanogaster fed a hexane extract of vegetables grown insoil contaminated with particulates from diesel engine exhaust—K. Kangsadalampai, P. Laohavechvanich, and J. Saksitpitak ........................................................................ 252

Improvement of dietary density by the use of germinated cereals and legumes—C. Singhavanich, S. Jittinandana, W. Kriengsinyos, and S. Dhanamitta ..................................................... 261

Letter to the editor ................................................................................................................................................. 267

Books received ....................................................................................................................................................... 269

In memoriam .......................................................................................................................................................... 271

Food and Nutrition Bulletin, vol. 20, no. 2

© The United Nations University, 1999

United Nations University PressThe United Nations University53-70 Jingumae 5-chome, Shibuya-ku, Tokyo 150-8925, JapanTel.: (03) 3499-2811 Fax: (03) 3406-7345E-mail: [email protected]

ISSN 0379-5721

Design and Production by Desktop Publishing & Design Co., Newton, MA USAPrinted by Rivkind Associates, Stoughton, MA USA

Editor: Dr. Nevin S. ScrimshawAssistant Editor: Ms. Edwina B. MurrayManuscripts Editor: Mr. Jonathan HarringtonSenior Associate Editor—Clinical and Human Nutrition:Dr. Cutberto Garza, Director and Professor, Division of Nutritional Sciences,

Cornell University, Ithaca, N.Y., USASenior Associate Editor—Food Science and Technology:Dr. Ricardo Bressani, Institute de Investigaciones, Universidad del Valle de

Guatemala, Guatemala City, GuatemalaAssociate Editors:Dr. Abraham Besrat, Senior Academic Officer, United Nations University,

Tokyo, JapanDr. Hernán Delgado, Director, Institute of Nutrition of Central America and

Panama (INCAP), Guatemala City, GuatemalaDr. Joseph Hautvast, Secretary General, IUNS, Department of Human

Nutrition, Agricultural University, Wageningen, NetherlandsDr. Peter Pellett, Professor, Department of Food Science and Nutrition,

University of Massachusetts, Amherst, Mass., USADr. Zewdie Wolde-Gabreil, Director, Ethiopian Nutrition Institute, Addis

Ababa, EthiopiaDr. Aree Valyasevi, Professor and Institute Consultant, Mahidol University,

Bangkok, Thailand

Food and Nutrition Bulletin

Food and Nutrition Bulletin, vol. 20, no. 2 © 1999, The United Nations University. 179

Abstract

A survey was conducted among 103 pregnant womenattending a maternal and child health centre in Rasaf,Baghdad, to study the relationships between maternaldietary intake and pregnancy outcome. Dietary intakevalues were not significantly correlated with birthweightand birth length except for caloric intake, which was sig-nificantly correlated with pregnancy weight gain, deliv-ery weight, pre-pregnancy weight, birthweight, and birthlength (r = +.33, r = +.30, r = +.26, r = +.38, and r = +.32,respectively). The results showed that the pregnant womenwho delivered small-for-gestational-age infants had dietsthat provided less energy than those who delivered full-term infants (p < .05).

Introduction

Since pregnancy is a time in the life cycle when nutri-tion is of special importance, and maternal nutritionhas an important influence on the course and outcomeof conception [1–3], there is ample evidence from manycountries that there is a good correlation between ma-ternal nutritional deprivation and the birthweight ofthe infant [4]. Since 1990, the nutritional status of Iraqipregnant women has deteriorated, largely as a resultof the long economic sanctions. Many recent investi-gations have shown that the incidence of low birthweight(LBW), defined as less than 2,500 g, among Iraqinewborns is as high as 13.3% [5].

We studied the relationship between dietary intakeand pregnancy outcomes among Baghdadi pregnantwomen under conditions of economic hardship.

Maternal dietary intake and pregnancy outcomesin Baghdad, Iraq

H. I. Tawfeek, Jwameer Nihad Abdulla, and Amna Hameed Rasheed

H. I. Tawfeek is affiliated with the Medical Institute in Bab-Al-Mudam in Baghdad, Iraq. Jwameer Nihad Abdulla andAmna Hameed Rasheed are affiliated with the Departmentof Food Science in the University of Baghdad.

Materials and methods

Between November 1991 and December 1994, 103 preg-nant women were chosen for the study, regardless oftheir age, parity, and residence. All attended the ma-ternal and child health centres in Baghdad during thefirst trimester of pregnancy.

Socio-demographic data were collected by trainedinterviewers. One 24-hour dietary intake value wasobtained each month by the same interviewer. Portionswere estimated by standardized food models along withvisual aids, such as posters with photographs of food.The diets were scored for energy, protein, and othernutrients using a food composition table for the Mid-dle East [6]. Nutrient intakes were evaluated againstthe 1989 recommended dietary allowances (RDA) [7].

The anthropometric examination included measure-ment of height and body weight during each visit. Thepre-pregnancy weight and height were obtained fromself-reports. Pregnancy outcome data, including birth-weight, length, and gestational age, were obtained duringthe first 48 hours after delivery by trained medical staffat the Teaching Hospital in Baghdad. Trained homevisitors obtained the pregnancy outcome data from thosewomen whose infants were delivered at home.

The gestational age of the neonates was assessed fromthe date of the last menstrual period. LBW infants weredefined as infants with birthweight less than 2,500 g.Premature infants were classified as those with birth-weight less than 2,500 g and gestational age less than37 weeks. Infants were defined as small-for-gestational-age (SGA) if they had a birthweight less than 2,500 gwith a gestational age of 37 weeks or greater. Student’st values and correlation coefficients were calculatedas appropriate.

Results

Descriptions of the 103 pregnant women are given intable 1. The mean age was 28.6 years. The overwhelm-ing majority of the subjects (96.1%) were married, and

180 H. I. Tawfeek et al.

few were separated or divorced (0.9% and 2.9%, re-spectively). The mean pre-pregnancy weight was 52.6± 16.3 kg, and the mean pregnancy weight gain was13.0 ± 3.5 kg. Sixty-six percent had a high school edu-cation or lower, and 33.8% had college or advanceddegrees. Household incomes were generally low.

Table 2 summarizes the dietary intakes for the 103pregnant women. The mean energy consumption duringpregnancy was 1,990 ± 340 kcal/day. Protein intake av-eraged 57 ± 11.2 g/day and was below the 1989 RDA.During the entire pregnancy, the intakes of calcium,iron, thiamine, vitamin B6, and niacin were below the1989 RDA. The intakes of vitamins A and C were ad-equate.

On the other hand, these women had lower intakesbefore pregnancy than during pregnancy, but the dif-ference was not significant. The correlations betweendietary intake and maternal anthropometric measure-ments are shown in table 3. Energy intake was signifi-cantly correlated with pregnancy weight gain and de-livery weight (p < .001) and was weakly, but significantly,correlated with pre-pregnancy weight (p < .05). On theother hand, protein intake was not significantly corre-lated with pre-pregnancy weight, pregnancy weight gain,or delivery weight.

Women who had low dietary intakes generally hadLBW infants. Information on the outcome of pregnancyis given in table 4. All pregnancies resulted in singlelive births. The mean birthweight and height were 3,146± 44 g and 49.5 ± 0.2 cm, respectively; 14% of the in-fants weighed less than 2,500 g. The mean gestationalage at delivery was 39.2 weeks (range, 30 to 43 weeks).

The correlation coefficients between maternal dietaryintakes and infant birthweight and length are shownin table 5. The positive correlation between caloric in-take and birthweight was highly significant (p < .001).

Discussion

The average intakes of calories, protein, calcium, iron,thiamine, and riboflavin were below the RDA. Therewas no significant relationship between the intakes ofthese nutrients and pregnancy outcome, except for therelation between caloric intake and birthweight. Thenegative correlations that have been reported between

TABLE 1. Socio-demographic characteristics of the 103 mothers

Characteristic Value

Age (yr)—mean ± SD 28.6 ± 1.3

Age (yr)—%16–19 15.520–24 24.225–29 33.930–34 16.535–39 4.8≥ 40 4.8

Marital status—%Married 96.1Divorced 2.9Separated 0.9

Maternal anthropometric measurements(kg)—mean ± SDSelf-reported pre-pregnancy weight 52.6 ± 16.3Weight gain in pregnancy 13.0 ± 3.5

Education—%High school or less 66.2College degree 24.2Advanced degree 9.6

Annual household income(Iraqi dinars)—% < 2,500 15.52,500–4,999 42.95,000–6,999 21.37,000–9,999 6.5≥10,000 13.8

TABLE 2. Nutrient intakes of the 103 mothers (means ± SD)

During pregnancy (average of all measurements)

Mothers of Mothers ofNutrient Before pregnancy All mothers SGA infants premature infants

Energy (kcal) 1,984 ± 280 1,990 ± 340 1,630 ± 356 1,774 ± 780Protein (g) 56 ± 18.3 57 ± 11.2 41 ± 31 49 ± 23Calcium (mg) 694 ± 110.2 700 ± 80.3 716 ± 81.5 694 ± 113.4Iron (mg) 11.2 ± 2.1 11.2 ± 3.1 11.6 ± 4.1 10.3 ± 2.8Vitamin A (IU) 4,593 ± 807 4,586 ± 902 4,008 ± 607 4,579 ± 1,030Thiamine (mg) 1.12 ± 0.1 1.21 ± 0.2 1.04 ± 0.3 1.19 ± 0.4Riboflavin (mg) 1.43 ± 0.1 1.53 ± 0.2 1.43 ± 0.2 1.51 ± 0.3Vitamin C (mg) 85 ± 16.4 87 ± 17.3 81 ± 20.3 86 ± 20.4

181

maternal intakes of calories, protein, and calcium andinfant birthweight and length [2] were not observedin this study.

Although there was a general agreement that the in-cidence of prematurity increased when the dietary in-take was poor, there was a disagreement about the re-lationship between dietary intake and the size of theinfant at birth [2].

The results from the population survey showed thatthe pregnant women who delivered SGA or prematureinfants had lower dietary intakes than those womenwho delivered full-term infants. These results are inkeeping with observations by others who studied riskfactors for SGA births in India [8].

A good relationship of maternal status to risk factorsfor SGA was found. Thomson reported that the influ-ence of current diet on birthweight was very small [9].These conclusions have been confirmed by others [10].

In a study in Cairo, slight but significant differenceswere observed in birthweight, height, and skull cir-cumference between infants of well-nourished and mod-erately nourished mothers [11]. Since the differenceswere slight, it was concluded that moderate malnutri-tion did not seriously affect the health of the infant.

The relationships between dietary intake and preg-nancy outcome have been reported with conflictingfindings and inconsistent results. Worthington-Robertsand Klerman reported that the inconsistencies may bedue in part to methodological problems in evaluating

TABLE 4. Pregnancy outcomes

Outcome Value

Full-term birth—% 74.7Premature birth—% 3.9SGA infant—% 7.8LBW infant—% 13.6Birthweight (g)—mean ± SD 3,146 ± 44 Length (cm)—mean ± SD 49.5 ± 0.2Gestational age (wk)—mean ± SD 39.2 ± 0.3

TABLE 5. Correlation coefficients (r) between nutrientintakes and pregnancy outcomes

Outcome

Nutrient Birthweight Birth length

Calories .38** .32*Protein .17 .04Calcium .21 .20Iron .21 .22Vitamin A .16 .08Thiamine –.18 .09Riboflavin .12 .14Vitamin C .07 .08

* p < .05; ** p < .001.

dietary intake accurately and in part to variations indaily nutrient intake among the women in manypopulations studied [12].

The interaction between nutrition and other envi-ronmental and biological factors frequently interfereswith the proper interpretation of the results. The inter-relation between the mother and her growing foetusand the extent to which the foetus depends on the ma-ternal diet are problems that need more investigation[3]. Clearly there is a need for more such studies inthe future.

References

1. Alberman E. The influence of social environment onthe nutrition of mother and child. In: Turner MR, ed.Preventive nutrition and society. San Diego, Calif, USA:Academic Press, 1981:85–94.

2. Beal VA. Nutritional studies during pregnancy. II. Di-etary intake, maternal weight gain, and size of infant. JAm Diet Assoc 1971;58:321–6.

3. Gabr M. Malnutrition during pregnancy and lactation.World Rev Nutr Diet 1981;36:90–9.

4. Antonov AV. Children born during the siege of Lenin-grad in 1942. J Pediatr 1947;30:250–6.

5. Al-Hadi AH, Al-Ani MM. Factors affecting birth weightof the new born babies in two hospitals in Baghdad. In:Proceedings of the Second Annual Scientific Congress.Baghdad: Sadam University, 1994:6.

6. Food composition tables for use in the Middle East.Publication No. 20. Beirut, Lebanon: Division of FoodTechnology and Nutrition, Faculty of Agricultural Sci-ences, American University of Beirut, 1963.

7. National Research Council. Recommended dietaryallowance. 10th ed. Washington, DC: National AcademyPress, 1989.

TABLE 3. Correlation coefficients (r) between anthropometricmeasurements and nutrient intakes

Pre-pregnancy Pregnancy Weight atNutrient weight weight gain delivery

Energy .26* .33** .30**Protein .09 .18 .13

* p < .05; ** p < .001.

Maternal dietary intake

182

8. Mavalankor DV, Gray RH, Trivedi CR, Parikh VC. Riskfactors for small gestational age births in Ahmedabad,India. J Trop Pediatr 1994;40:285–90.

9. Thomson AM. Diet in pregnancy. 3. Diet in relation tothe course and outcome of pregnancy. Br J Nutr1959;13:509–25.

10. Metcoff J. Association of fetal growth with maternalnutrition. In: Fallmer F, Tanner JM, eds. Human growth.London: Ballière Tindall, 1979:415–60.

11. Kamal I. Standard length, weight and cephalid diam-eters in new-born Egyptian babies. Effect of race, sex,and parity on various measurements of body and head.J Egypt Paediatr Assoc 1962;10:1–35.

12. Worthington-Roberts BS, Klerman LV. Maternal nutri-tion. In: Merkatz IR, Thompson JE, Mullen PD,Goldenberg RL, eds. New perspectives on prenatal care.New York: Elsevier Science Publishing 1990:235–71.

In the article “New issues in developing effectiveapproaches for the prevention and control of vitaminA deficiency,” by Martin W. Bloem, Saskia de Pee, andIan Darnton-Hill (Food Nutr Bull 1998;19:137-48), onp. 144, column 2, in the sentence that begins on line

Erratum

10, the authors intended to say, “A recent study in Nepalshowed an impact of both vitamin A and caroteneseparately on maternal mortality [5].” The authorsregret the misleading phrasing of the sentence aspublished.

H. I. Tawfeek et al.

Food and Nutrition Bulletin, vol. 20, no. 2 © 1999, The United Nations University. 183

Abstract

To evaluate programmatic efforts to improve dietaryintake and growth in 6- to 12-month-old infants indeveloping countries, we reviewed the results of 5 effi-cacy trials and 16 programmes conducted in 14 coun-tries. Efficacy trials were able to improve infant dietaryintakes by 65 to 302 kcal/day and infant growth by 0.04to 0.46 SD. Programmes reported large improvements inmaternal knowledge and practices concerning infantfeeding. Four programmes that provided informationreported improvements in dietary intakes of 71 to 164kcal/day and changes in growth of –0.08 to 0.87 SD.Despite variability in the results, the majority of researchand programmatic efforts improved growth rates by 0.10to 0.50 SD. In absolute terms, this range of improvementin growth would reduce prevalences of malnutrition(< –2 SD) at 12 months of age by 1% to 19% and couldreduce deaths due to malnutrition by 2% to 13%, de-pending on the underlying prevalence of malnutritionin the community.

Introduction

Childhood malnutrition is a major public health prob-lem throughout the developing world and is one ofthe principal underlying causes of death for many ofthe world’s children [1, 2]. Research conducted overthe past 20 years throughout the world in a variety of

Interventions to improve intake of complementaryfoods by infants 6 to 12 months of age in developingcountries: Impact on growth and on the prevalenceof malnutrition and potential contribution tochild survival

Laura E. Caulfield, Sandra L. Huffman, and Ellen G. Piwoz

Laura Caulfield is affiliated with the Center for HumanNutrition, Department of International Health, Johns HopkinsSchool of Hygiene and Public Health, Baltimore, Maryland,USA. Sandra Huffman and Ellen Piwoz are affiliated withthe Academy for Educational Development (AED) inWashington, DC.

Funding for this activity was provided by the United StatesAgency for International Development (USAID) underCooperative Agreement HRN-A-0097-0007-00 and by fieldsupport from USAID/Bolivia. The contents of this documentdo not necessarily reflect the views or policies of USAID.

settings has demonstrated that post-natal growth fal-tering begins at around six months of age, just as in-fants begin to receive foods to complement their breast-milk intake [3].

How should babies be fed during their first year oflife? For the first four to six months of life, it is recom-mended that infants receive no food or liquid otherthan breastmilk (not even water). From the age of aboutsix months onward, infants should continue to receivebreastmilk and, in addition, should be fed safe and ad-equate amounts of local foods frequently throughoutthe day. A recent review [3] suggests that on the as-sumption of an average intake of energy from breast-milk, infants 6 to 8 months of age should receive 270kcal/day and those 9 to 11 months of age should re-ceive 450 kcal/day from complementary foods. The vari-ability in energy requirements is wide at any age; thus,some infants will need considerably more, whereas oth-ers will need considerably less than these average val-ues. The total length of time that a child receives breast-milk is left open for the mother and child to decide,but it is recommended that mothers continue breast-feeding throughout the second year of life and for aslong as possible.

The provision of safe and adequate amounts of localfoods appropriate for 6- to 12-month-old infants whoare just learning to eat is not as simple as it seems. Com-plementary foods fed to infants in the second six monthsof life (and beyond) are often inadequate in energy den-sity, protein, and micronutrient concentration or qualityand are often prepared, stored, or fed to children inways that increase their risk of illness. Because of thecomplexity of the nutritional and behavioural issuesinvolved, careful, detailed, interdisciplinary work mustbe undertaken to define the nutritional problems of olderinfants and identify appropriate and effective interven-tions to improve complementary feeding and infantgrowth and development in country settings.

Many nutrition programmes have been implementedover the past 20 years to improve complementary feedingfor young children throughout the world, and there isrenewed interest in strengthening this component of

184 L. E. Caulfield et al.

child survival programmes. Therefore, it is a timely pointto review the scientific and programmatic literature toidentify what has been accomplished and what can belearned from earlier programmes and applied in fu-ture efforts to improve complementary feeding.

The purpose of this review is to evaluate the extent towhich research and programmatic efforts over the last20 years have been able to improve the dietary intakesof 6- to 12-month-old infants in developing countries.First, we review the results of efficacy trials conductedto improve the growth of these older infants, and thenthe information available on programmes designed toincrease dietary intakes and reduce malnutrition indeveloping countries. The following questions areaddressed: What improvements in dietary intakes andgrowth of infants have research projects (with efficacydesigns) been able to achieve? What improvements indietary intakes and growth of infants have develop-ing-country programmes been able to achieve? Whatare the features of successful programmes to improvecomplementary feeding of infants in developing coun-tries? What is the likely impact of improvements ingrowth rates (brought about by such programmes) onmalnutrition rates? On deaths due to malnutrition?

Methods

A search was undertaken to identify scientific papersand programme project documents with data perti-nent to this analytic review. The search was not restrictedto work on older infants; rather, we sought to identifyresearch or programmatic efforts conducted with pre-school children and then to determine whether perti-nent data were available on the 6- to 12-month agegroup. To identify efficacy studies, we conducted a com-puter-based search for the years 1970 to 1997 and ex-amined all citations in the identified studies. We alsoexamined the references in reviews on dietary intakesand growth of infants and children in developing coun-tries. The final list is probably complete with respectto published reports of controlled trials to improve di-etary intakes and growth of non-hospitalized 6- to 12-month-old infants in developing countries. To iden-tify relevant programme projects, we reviewed thecitations of published reviews as well as other docu-ments (published by ACC/SCN or the World Bank)detailing programmatic efforts to reduce child malnu-trition. We also enlisted the help of public health pro-fessionals working in the area of child feeding andgrowth in developing countries to identify relevantdocuments. In all, 5 efficacy trials were identified forreview, as well as 16 programmes conducted in 14 coun-tries. Complete information (design, coverage, andimpact on outcomes of interest) was not always avail-able for all 16 programmes; however, because we feltthe information available would be useful to research-

ers and programme planners, we have made maximaluse of the information available.

One important goal of the review was to describethe impact of the interventions on the growth rates ofchildren. To do so, we converted (when necessary) thereported nutritional status data into standard devia-tion (SD) units. This was accomplished in a variety ofways, depending on how the published data were re-ported. For example, the pre-supplementation nutri-tional status of each treatment group was convertedto Z scores or SD units by subtracting the mean weightor length at six months of age in each group from thegender-averaged median of weight or length from theinternational growth reference [4], divided by the gen-der-averaged reference SD of weight or length at thatsame age. The differences in the post-supplementationaverage Z scores of weight or length describe the im-pact of the intervention on growth rates in SD units.The treatment effect on growth in SD units could alsobe calculated by taking the difference in the simple post-supplementation means (in kilograms or centimetres)and dividing by the gender-averaged reference SD forchildren of that age.

Calculating nutritional impact in terms of shifts inthe distribution of nutritional status in SD units is useful,because the results allow us to project reductions inthe prevalence of malnutrition (% < –2 Z score) asso-ciated with various levels of programme impact. Thiscan be done if we assume that the post-intervention Zscores are normally distributed and have a given SD(in our case, we chose 1.0), and we utilized informa-tion on the area under a normal curve at selected cutpoints [5]. As a simple example, suppose that the av-erage weight-for-age or length-for-age of children in acommunity is 3.0 ± 1.0. In this case, about 83% of chil-dren would be considered underweight or stunted, re-spectively, because –2 Z score falls at +1 SD on the com-munity Z-score distribution, and the proportion of thearea under the normal curve [5] to the left of +1 SD is0.83 (or 0.50 + 0.33). Similarly, if the average growthrates of infants in the community were improved(greatly) by +1.0 SD with the variance of growth un-affected, we would expect the post-intervention Z scoresin the community to be –2.0 ± 1.0 and the prevalenceof malnutrition to be reduced to 50%.

To examine the potential impact of improvementsin growth on child survival, we used the method devel-oped by Pelletier et al. [6] for calculating the propor-tion of deaths attributable to malnutrition in childrenin developing countries. The population attributable risk(PAR), or the number of deaths due to malnutritionas a proportion of all deaths, is calculated as follows:

PAR =× −( )[ ]

+ × −( )[ ]∑∑

P RR

P RR

i i

i i

1

1 1

185

where Pi is the prevalence of malnutrition of a givendegree of severity and RRi is the relative risk of mor-tality associated with that degree of malnutrition [6].On the basis of the analysis of data from eight studies,the relative risk of dying for children of a given weight-for-age (as a percent reference median) as comparedwith children with weight-for-age of 90% can be cal-culated as RR = 10–0.0264 (WA – 90), where WA is weight-for-age [6]. This information can also be used to cal-culate the expected reduction in deaths due tomalnutrition given improved growth rates of children.Pelletier et al. [6] determined that the method is validacross the age range of 6 to 36 months and thus is ap-propriate for the reduced age range of 6 to 12 monthsthat is of interest here. Further, although the methodis based on nutritional status calculated as percent ref-erence median, their analyses support the applicationof their method to nutritional data calculated as Z scores.For our purposes, the distributions of weight or lengthZ scores were converted to units of percent referencemedian at 12 months of age and then used to calcu-late the PAR depending on the theoretical prevalencesof varying degrees of malnutrition and associated RR.By calculating the PAR for different theoretical Z-scoredistributions, the impact of improving growth rates ofchildren on reducing malnutrition and therefore thePAR could be quantified.

Results

Efficacy studies to improve dietary intakes and growth

It is widely held that the largest improvements in theintakes of complementary foods of older infants willbe achieved in scientific research trials because of tech-nical, financial, and other factors that differentiate re-search from programmes. Because of this, it is helpfulto review the achievements of small-scale randomizedintervention trials in this area to provide us with anidea of what optimally might be achieved. From thescientific literature, five studies [7–13] were identifiedfor review. The studies, conducted in Indonesia, Bang-ladesh, Colombia, Guatemala, and Jamaica, are sum-marized in table 1.

The interventions varied with respect to design andprecise objectives. However, most of them wererandomized controlled trials with the common goal ofimproving the growth of children through improvementsin their dietary intakes. None of the trials focused ex-clusively on 6- to 12-month-old infants. For this rea-son, most of the information on programme impactwas derived from the results reported for the subsampleof study participants of the appropriate age. In Colom-bia, however, it was not possible to separate out theimpact of the programme on 6- to 12-month-old infants,and therefore we report the impact on infants 3 to 12

months of age. Further, in Bangladesh the interventioninvolved 8-month-old infants who were followed upuntil they were 13 months of age. The study is includedhere under the assumption that the results obtainedare appropriate to what would be observed in 6- to12-month-old infants. Most of the trials gave food assupplements to the usual dietary intakes of the infants,but the studies varied with respect to the foods offeredand the delivery mechanism. The exception to this wasthe study conducted in Bangladesh [13], in which spe-cific feeding advice and cooking demonstrations wereprovided to caretakers in their homes. Although ourfocus is not on the statistical significance of the find-ings, it is important to note that each of the five stud-ies found that the nutritional interventions significantlyimproved the growth rates of participating children.

With respect to our objectives, there are four con-clusions that can be drawn from these published stud-ies. Each of these points is described below.

The studies were largely successful in identifying foods thatwere liked and accepted by children in this age group andtheir caretakers in each cultural setting

The foods offered to the children to enhance their en-ergy intakes from non-breastmilk foods varied acrossthe settings. The foods were either offered daily as ready-to-eat foods [7–9] at feeding centres [8, 9] or day-caresettings [7], or provided weekly as raw ingredients tothe family to be prepared and fed to the study chil-dren [10–12]. Of the foods offered, many provided oneor two nutritious foods to be added to the diet. How-ever, in Indonesia [7] local foods were used to develop20 distinct snack foods to augment the dietary intakesof children while they attended a day-care centre. Inseveral settings, drinks as opposed to semi-solids orsolids were offered. In Guatemala atole, a high-energy,moderate-protein, micronutrient-fortified drink, wasprovided [8, 9]. Whole powdered milk was distributedto the families of the study children in Colombia [10,11], and a milk-based formula and skim milk powderwere distributed in Jamaica [12]. Again, only in the studyin Indonesia [7] were combinations of semi-solids andsolids offered to increase the energy intakes of 6- to12-month-old infants.

In general, the foods were well liked and consumedby infants in this age group. In three sites [7–9, 13],part of this success can be attributed to the formativeresearch conducted before the intervention to identifyinfant-feeding issues, foods appropriate for young chil-dren, and the level of energy required from comple-mentary foods in the population. In Guatemala [8, 9]infants were offered either of two different drinks, atoleand fresco, a low-energy, no-protein, micronutrient-fortified drink. Consumption of atole was good, be-cause mothers thought it was a good food for infants,whereas consumption of fresco was very low, becausemothers did not consider it an acceptable food.

Interventions to improve intake of complementary foods

186

Regardless of the acceptability of the supplementalfoods, an important drawback in some of the studieswas related to the texture (liquid) and uniformity of thesupplements chosen [8–12]. Offering a variety of texturesand flavours to 6- to 12-month-old infants is important,because two of the developmental goals of complemen-tary feeding are to introduce the infant to varying foodtextures in preparation for mastication [14, 15] and tobegin the transition of the infant from a single-source(liquid) diet to one characterized by diversity and va-riety. Many psychologists believe that delays or failuresto introduce textures to older infants may lead to later

feeding problems [14, 15]. Whether this happened inany of the published studies is not known. However,the implication for programmes is that the additionof single-source or liquid supplements to the diets ofolder infants may not be the best approach to improv-ing the dietary intakes and well-being of children liv-ing in underprivileged environments in the long term.

Energy intakes from complementary foods were improved bythe interventions

It is important to note that these studies did not com-pare the outcomes of infants who did or did not re-

TABLE 1. Efficacy trials to improve dietary intakes and growth of infants 6–12 months of age in developing countries

Baseline Change inNet increase WAZ and WAZ and

Site Subjects Duration Food/nutrient profile (kcal/d)a HAZ HAZ

Abbreviations: HAZ, height-for-age Z score; WAZ, weight-for-age Z score.a. Net increase refers to the difference in energy intakes between treatment and control infants net of home and breastmilk energy intakes.

In Guatemala, the intake estimate is also net of the intake of fresco.

Guatemala[8, 9]

150–200pergroup orvillage, 4villages

From birth,60–80%participation

Atole: 40–60% d

Fresco: 10–30% d

Two drinks: atole (high energy,moderate protein) and fresco(low energy, no protein), eachwith added vitamins and minerals.Offered in unlimited amounts

+83

Colombia[10, 11]

70 pergroup

From about3 mo to3 yr

Whole powdered milk + commer-cial high-protein vegetable mix.Provided 428 kcal/d. Also gave10 mg/d iron and 1,552 IU/dvitamin A

At 18 mo:+178

At 6 mo:–0.21 to

–0.50WAZ

–0.16 to–0.36HAZ

3–12 mo:+0.40

WAZ+0.35 HAZ

Jamaica[12]

128stuntedchildren9–24 moold

12 mo 1 kg/mo milk-based formula and1 kg/mo skim milk powderand cornmeal for the family,delivered weekly to home. 750kcal/d with 20 g/d protein

+106 –2.5 WAZ–3.0 HAZ

+0.44WAZ

+0.35 HAZ

Indonesia[7]

112children6–20 moold

90 d 20 snacks of local foods containingbread, rice, wheat flour, sweetpotato, coconut milk, cassava,potatoes, sugar, and oil. 400 kcal/don average with 5 g/d protein in aday-care setting

+ 317(9–23mo old)children

–1.6 WAZ–2.4 HAZ

+0.29WAZ

+0.04 HAZ

Bangladesh[13]

120 infants8 mo old

5 mo In home demonstrations to enrichmeals with energy and protein,using oil, molasses, fish, lentilflour, fruits, and vegetables

Frequent and persistent feeding ofchop-chop: wheat flour, oil, andbrown sugar. Feed more snacks(not meals). No information onkcal/d or nutrient profile

+65 –1.94 to–2.11WAZ

+0.46WAZ

+0.25WAZ

+0.17HAZ

L. E. Caulfield et al.

187

ceive complementary foods, but rather the outcomesof infants who had higher versus lower intakes of com-plementary foods in addition to breastmilk—differencescreated by the intervention itself. As shown in table 1,the study infants were offered at least 400 kcal/day inaddition to their usual energy intake from both breast-milk and non-breastmilk foods given them at home.

Humans, including infants, self-regulate their energyintake. At all sites, study infants did not consume allof the food offered. Although the calculation is diffi-cult to make for all studies, it appears that the studyinfants consumed 15% to 75% of the extra energyoffered from non-breastmilk foods. Thus, the inter-vention strategies were able to increase energy intakesfrom non-breastmilk foods by 90 to 500 kcal/day, withmost efficacy studies reporting intakes in the range of150 to 300 kcal/day.

Replacement of energy from breastmilk occurred in many ofthe studies, and thus the net improvements in total energyintake were somewhat smaller but still important

As stated by Brown et al. [3], optimal complementaryfeeding in this age range involves adding energy andmicronutrients while maintaining high energy intake frombreastmilk. One disturbing finding reported in some ofthese studies is that additional intake of non-breastmilkfoods replaced some of the energy previously providedby breastmilk. Thus, the total energy intake (breast-milk energy plus complementary food energy) was lowerthan expected. Information is not available for all studies,but the decrease in the amount of energy obtained frombreastmilk was as much as 200 kcal/day in infants offeredadditional complementary foods. Across the studies,the net increase in total energy intake due to the inter-ventions ranged from 65 to 300 kcal/day.

The growth of infants was improved by the interventions

The nutritional status of the infants at enrollment inthe studies was generally poor, with average Z scoresof –2.5 to –0.21 for weight-for-age and –3.0 to –0.5for height-for-age.

To assess the impact of increasing total energy in-take by 65 to 300 kcal/day, most of the studies com-pared rates of growth between infants in the interven-tion and control groups. All five studies reportedstatistically significant positive impacts of the interven-tions on somatic or linear growth, or both. Each ofthe studies provided evidence that the intervention andcontrol groups of infants had similar nutritional statusat baseline. At the end of the interventions, which lastedfrom five to nine months, differences in nutritional statusdue to the intervention in the five studies rangedfrom –0.25 to +0.46 SD units for weight-for-age andfrom –0.04 to +0.35 SD units for height-for-age.

In summary, the results suggest that improvementsin complementary food intakes of 6- to 12-month-oldinfants in developing countries can be made, and that

such improvements in dietary intakes will translate intoimproved rates of growth. However, these were researchprojects in which the delivery of and compliance tothe intervention were tightly controlled. Further, theprojects were small in scale, involving only several hun-dred infants. The next question is, “How effective haveprogrammes been for improving complementary foodintakes of older infants?”

Programmes to improve complementary feeding

This section summarizes the experiences of 16 pro-grammes in 14 countries to improve the complemen-tary food intakes of 6- to 12-month-old infants. Theapproach taken in the majority of programmes has beencomprehensive, involving formative research for thedevelopment of specific complementary foods as wellas complementary feeding education and counsellingand some form of mass-media communication strat-egy. During implementation, most programmes con-ducted some form of monitoring or evaluation in or-der to modify or enhance the impact of the programme.Many of these programmes were implemented underthe umbrella of growth-monitoring programmes, andmost of them were large in scale, reaching populationsof 1,000 or more potential beneficiaries. It should alsobe noted that these programmes were not focused solelyon the feeding of 6- to 12-month-old infants; rather,they were designed to improve feeding practices in chil-dren up to 3 years of age, including appropriate feed-ing during illness. The programmes were conductedin Peru [16–18], the Dominican Republic [19], Indo-nesia [20, 21], the Philippines [22], Mali [23, 24], BurkinaFaso [25, 26], Nigeria [27, 28], the Gambia [29],Cameroon [29], Tanzania [29], Swaziland [29], Ghana[29], Niger [29], and Senegal [29].

Most of the programmes used a combination of quali-tative and quantitative research techniques in five stagesin order to formulate the behavioural change strate-gies [29]. The stages were the following:1. Review of pre-existing information on feeding prac-

tices and diet, inter alia ;2. Ethnographic study of health and nutrition beliefs

and practices of the health providers, communitymembers, and mothers, and their sources of infor-mation;

3. A nutritional assessment of existing diets and prac-tices, their potential for enrichment or improvement,and possible resistances or obstacles to improvingdiet quality and feeding practices;

4. Individual and group trials of new feeding practices,foods, and recipes to determine the most feasiblealternatives for improving dietary intake and peo-ple’s reactions to new products and behaviours;

5. Development of an overall strategy for improvingchild feeding in the population, based on the find-ings from steps 1 to 4.


188

Because of the nature and complexity of the endeav-ours undertaken, it is important to consider the achieve-ments of these programmes at two stages. The first stageconsiders the results of step 4 above, in which the fea-sibility of new feeding practices and recipes was testedon a relatively small scale with intensive motivationand counselling. Following that, we will consider theoverall impact of these programmes on feeding behav-iours, dietary intakes, and infant growth.

Trials of new feeding practices, foods, and recipes

Studies of the complementary foods currently fed toolder infants in developing countries have highlightedthe problems of low energy density and poor nutri-tional quality [30, 31]. To address these problems, manyprogrammes have attempted to develop new comple-mentary food recipes, building largely on what moth-ers were already “doing right.” An example we havealready seen is from the study in Bangladesh [13] inwhich a complementary food called chop-chop was pro-moted. The food—a mixture of wheat flour, oil, andbrown sugar—was already being fed to children whenthey had diarrhoea; the project sought to promote its usefor healthy children as well. To determine the natureand feasibility of a new (or improved) food product,household trials were conducted with small numbersof individual mothers or groups of mothers. In thesetrials, programme planners and mothers worked to-gether to develop and evaluate new food products andfeeding behaviours.

Shown in table 2 are the results obtained from forma-tive research involving recipe development and test-ing with the purpose of designing nutritionally improvedcomplementary foods for young children. Three pointscan be drawn from these results.

Nutritionally improved complementary foods could bedeveloped in diverse cultural settings

Eleven of the programmes cited above involved thedevelopment of new or improved complementary foods.As shown in table 2, most of the foods developed weregrain-based porridges to which protein-, energy-, andmicronutrient-rich ingredients were added. It is inter-esting to note the similarity of the approaches takenin very diverse settings. Throughout sub-Saharan Af-rica, a twofold strategy was followed. First, protein andmicronutrients were added to a staple porridge usingpeanuts, peanut butter, or cowpea flour. Second, in somesettings, fermented flour was added to reduce the vis-cosity of the porridge, thus enhancing energy densityas well as the acceptability to children. In Indonesia[20, 21] and the Philippines [22], fish flakes, vegeta-bles, and oil were added to rice.

It should not be concluded from these results thatformative research is no longer a necessary step in thedevelopment of projects to improve complementary

feeding. Rather, experience thus far suggests that somecommonality exists in the nature of the problems in com-plementary feeding as well as in the means of address-ing them. This commonality will probably streamlinethe formative research necessary for the developmentof new projects.

From a nutritional standpoint, the foods developedin these recipe trials represented clear improvements overthe traditional foods offered to children in each setting.However, little information is available to quantify theextent to which regular consumption of these foods(as well as others offered to children) would translateinto overall improvements in the energy, protein, andmicronutrient intakes of older infants and children.

Poor mothers are willing to prepare new foods and theirchildren are willing to eat them

Are poor mothers in developing countries willing totry new ways of preparing complementary foods fortheir infants? As shown above, the overriding conclu-sion to be drawn from the recipe trials is “yes!” Themajority of women participating in the trials were willingto prepare new foods and feed them to their childrenand were active participants in their development.However, the trial results suggest that although moth-ers were generally willing to try new foods, they wereless able to feed the foods regularly and to incorporate(adopt) them into their usual feeding repertoire be-cause of time and resource constraints. Children likedthe foods and were willing to eat them, and this was amajor motivation for mothers to change and continuenew foods and practices.

There are common barriers to using new foods and foodpreparations

Despite the willingness of mothers to try somethingnew, several common barriers to using the new foodswere identified. In a variety of settings, mothers stateda willingness to try the new foods and to continue tofeed them as long as their children liked them. Sec-ond, the foods for enrichment had to be affordable; akey strength of several projects was the developmentof options for mothers at varying levels of poverty. Forexample, in Senegal [29] mothers who could not affordto add peanuts were willing to add milk or butter totheir infants’ porridge. Time is an important constraintto many women; to be successful, enriched foods mustinvolve minimal changes in maternal time spent inpreparation and feeding. Finally, it was reported in sev-eral sites that mothers would discontinue giving thefood if their infants became sick or had other negativereactions that the mothers attributed to the food. Thisis a difficult barrier to overcome, because infants indeveloping countries will invariably get sick, and somemay happen to get sick after consuming the enrichedcomplementary food.

Although these results appear optimistic, it is impor-


189

TABLE 2. Household and recipe programme trials to improve complementary food intakes of infants 6–12 months of age indeveloping countries

Site Enriched weaning food Nutrient profile Time period Acceptance

Peru [16] Sanquito: wheat flour,toasted pea flour,brown sugar, oil,carrots, water

237 kcal/100 g,recommended tofeed 1 cup/d,providing 550 kcal,11.9 g protein, and206 µg RE

Small groups ofwomen in 10rural highlandcommunities

88% (on 1st try,accepted 4 mouthfuls);compared withpudding, childrenconsumed 6 timesmore kcal/kg/12-h day

Nigeria[27, 28]

Eko ilera, a fortifiedogi, containing maizeor sorghum ogipaste, cowpea flour,red palm oil, sugar,water, sorghummalt flour

85 kcal/100 g (+50kcal/100 g fromtraditional eko);designed for 12%increase in netenergy intake/d

Good acceptance, butno estimates available

Cameroon[29]

Bouillie enrichie, thetraditional papenriched with milk,egg, or peanut butter


The Gambia[29]

Millet pap ogi,enriched with peanutpaste; alternativeadditional ingre-dients were beanflour, butter, milk,and dried fish

31 women visited 4times at home over2-wk period

68% liked the peanuts,an additional 13%added alternativeingredients

Tanzania [29] Maize peanut gruel ugiwith germinatedsorghum flour(“power flour”)or kimea

40 children aged 5–60mo in 3-mo trial

28% prepared gruelwith kimea regularly;85% used gruel about25% of the time

Ghana [29] Traditional cerealporridge, thickenedand enriched withlegumes, fish powder,or oil; frequency offeeding also increased

About 50–70 mothersover 4–7 d

79% willing to thicken;82% willing to enrich;82% willing toincrease feedingfrequency

Swaziland[29]

Maize porridgeliphalishi with germi-nated sorghum malt,enriched with relish,oil, or peanut butter

28 families over 1 wk 86% were able toenrich; 90% addedthe malted grains

Niger [29] Millet flour-basedporridge with sugarand peanut solids(kulikuli), sourskimmed milk, or afried bean, millet,or wheat cake

83 kcal/100 g enrichedwith cakes; 51 kcal/100 g enriched withpeanut solids or milk

116 mothers ofmostly ill children4–24 mo old in a1-wk trial

73% used food at leastonce a week; 11% usedit for 3 or more days;4% never tried it; 26%modified the recipeslightly; 91% wouldcontinue to give thefood to their child

continued on next page


190

tant to keep in mind that several key features inherentin the design of these recipe trials influence the typeof results obtained. First, the approach involves smallnumbers of women or other target audiences. Second,the approach is participatory; women are asked to pro-vide suggestions regarding the ingredients, methods ofpreparation, and ways to overcome barriers to their use.Thus, the dynamic created during these sessions is a one-on-one partnership between the programme plannersand the target beneficiaries. Third, the nature of theprocess is to begin with a list of possible options (in thiscase, recipe variations), with the goal of the researchbeing to refine the list. Thus, the results presented nec-essarily reflect the most successful approaches, and detailson everything that was tried but not found to be suc-cessful are often not described. It is therefore importantto consider whether the new foods or promoted feedingpractices will be adopted when the intervention is con-ducted on a larger scale with less intensive interactionand participation among all mothers in the population.

Programme messages and communication strategies

How did these programmes try to improve comple-mentary feeding? Presented in table 3 are brief descrip-tions of the content and communication strategiesemployed in nine of the programmes. Several featuresof these programmes are evident from the table.

Improving complementary feeding practices requires acomprehensive approach

The majority of the programmes used a comprehen-

sive approach for improving complementary food in-takes of older infants. First, the design of the programmecontent indicates that optimal complementary feedingbegins at birth and that what constitutes optimal feedingchanges as the infant grows and develops. Thus, formothers to feed their infants optimally, they need simple,action-oriented information that is age-appropriate andthat changes as their infants grow and mature. Second,the content is built upon current local practices, beliefs,and concerns—just as the food products or recipes thatare developed are built upon local beliefs and prac-tices. Third, the messages used reflect the fact that moth-ers need advice and information not only on what tofeed, but also on how to feed their infants.

As described in table 3, the key features of the edu-cational strategies are the following:

Promote exclusive breastfeeding at birth as the criticalfirst step. Promoting exclusive breastfeeding until fourto six months of age delays the introduction of com-plementary foods until the appropriate time. Breast-milk is a critical component of an optimal feeding strat-egy for infants: the more optimal the breastmilk intake,the lower the amount of energy required from com-plementary foods. The programmes differed with respectto the recommended duration of exclusive breastfeed-ing because of changes in the international recom-mendations over time and because of variations amongnational policies. Besides promoting exclusive breast-feeding, several of the programmes had messages thatattacked the reasons given by women for not exclu-sively breastfeeding their infants for the first four tosix months of life. For example, in many countries water

TABLE 2. Household and recipe programme trials to improve complementary food intakes of infants 6–12 months of age indeveloping countries (continued)

Site Enriched weaning food Nutrient profile Time period Acceptance

Senegal [29] Porridge enrichedwith peanut butter,milk, butter, or oil

31 mothers 77% tried food; 68%would continue; milkand butter werepreferred; butter forinfants 6–11 mo old.Milk most likely to becontinued

Indonesia[20, 21]

Nasi tim bayi, riceenriched with fish,vegetables, and oil

22 mothers [21] 86% [21] made it asmilk; 55% continuedto make it with slightchanges; 45% werewilling to add fatsource to the recipe

Philippines[22]

Rice or lugao enrichedwith oil plus flakedfish plus a vegetablefrom the family pot



191

TABLE 3. Programmes to improve complementary feeding of infants 6–12 months of age in developing countries: Messagesand communication strategies

Site Messages and interventions Communication strategies

Burkina Faso[25, 26]

Breastfeed exclusively to 4–6 moGive enriched bouillie and mashed fruits at 4–6 mo,

continue breastfeeding, give 3 bowls/d plus fruitor other snack once child begins chewing food

Increase well-baby clinic attendance

Drama, story-telling, print materials,counselling, handouts, flip charts, songs andcassettes; in counselling, negotiate ratherthan educate

Cameroon [29] Prepare and offer bouillie enrichieAdvice on feeding frequency, meal composition

for healthy and sick children, 0–3, 4–9, 10–15,and 16–36 mo old

Animateurs led monthly discussion groups,held cooking demonstrations, performedgrowth monitoring, gave individualizednutrition counselling, made home visits forthose moderately malnourished

Nigeria [27,28] Exclusive breastfeeding to 4 moAt 4 mo, prepare eko ilera instead of traditional

eko to help your child be “active and powerful”Feed other solids as well, starting at 6 moContinue breastfeeding at least until child’s 1st

birthdayHygienic methods of food preparationEncourage spoon feeding (as opposed to hand

feeding)Strongly discourage bottle-feeding of any kind

In each village, 10 village-selected “teachingmothers” trained by community healthworker each taught 10 mothers via 2 cookingdemonstrations; mothers were given flyers totake home and 3-d supply of ingredients

Peru [16] Feed 1 cup/d of sanquito to children 6–18 mo oldwhen ill, 2 cups/d if > 18 mo old

Continue breastfeedingFeed other foods as wellDo this during and for 1–2 wk after episode of

diarrhoea

Radio messages, flip chart, calendar with recipeon it, training materials for healthprofessionals

Peru [17, 18] Booklet: “You can learn to produce enoughbreastmilk”

Breastfeeding helps keep an infant sanoBreastmilk only until 6 mo; if the baby is thirsty,

the mother should drink waterAt 6 mo feed thick, easy-to-prepare foods, feed

thick foods before soupGeneral messages on “How to feed your child

during the first year of life”

Posters, mobile loudspeakers, radio spots, 3- to8-min instructional videos in health clinics,leaflets, recipe booklet and infant-feedingguide, nurse-led lactation classes,nutritionist-led cooking demonstrations

DominicanRepublic [19]

Growth monitoringOnly breastmilk until 4 moAt 5–8 mo, also feed 3 times/d if child is growing

well; feed 4 times/d + fruit (banana, orange, ormango) if child is not growing well

Give thick foods (add little water), since “Waterfills but does not nourish”

Feed 9- to 24-mo-olds 3 times/d + 2 snacks offamily food, continue breastfeeding, if notgrowing well—give family foods (don’t addwater), feed 4 times/d + 2 snacks, continuebreastfeeding

Mass marketing and individual and groupeducation (community growth chart);also community development education,counselling cards, individualized growthmonitoring, and counselling in mother’shome



192

is given to young infants to curb their thirst, and sev-eral programmes developed specific messages to counterthis idea. In several sites, messages suggested that mothersshould drink more to quench their infants’ thirst [17,18, 27, 28]. In one of these programmes, the decisionwas made not to try to dissuade mothers from givinginfusions (teas), because the beliefs in the populationregarding this practice were so tightly held. Through-out the world, “insufficient milk” is the principal rea-son mothers begin introducing complementary foodsbefore the appropriate time [32]. Given that this is likelyto be true only rarely, the projects sought to overcomethis barrier by increasing mothers’ self-confidence withrespect to breastfeeding, discouraging the use of paci-fiers and bottles, and teaching women behavioural strat-egies for optimizing breastmilk production, such asincreasing frequency, breastfeeding at night, and feed-ing from both breasts each time. In Peru a booklettitled “Como tener bastante leche” (How to have enoughmilk) was produced and disseminated [17, 18].

Begin complementary feeding with small amounts ofmashed foods; texture is important. Although a five- orsix-month-old infant is beginning to need more energy

than that provided by breastmilk alone, the need de-velops slowly over a few months, depending on themother’s milk production and the infant’s individualgrowth pattern. Thus, the goal of complementary feedingat the beginning is to introduce new flavours and tex-tures to the infants, prepare them to begin chewing,and encourage their interest in food. Mothers through-out the world, however, are worried about giving in-fants semi-solids or solids before they can handle them.For this reason, many of the programmes recommendedthat mothers start by giving mashed or pureed foods.Two of the programmes also recognized maternal con-cerns regarding the time involved in preparing specialcomplementary foods; one programme provided recipesor food ideas that were “fast and easy,” and the otherpromoted the mashing of the foods fed to the rest ofthe family. By focusing the issue around texture andease of preparation, the mothers’ initial concerns arevalidated, and this is likely to be a key factor influenc-ing maternal acceptance of the recommendation.

Provide energy and nutrient-dense complementary foodson a daily basis. As described earlier, most of the foodproducts or recipes recommended in the programmes

Mali [23, 24] Growth monitoringOnly breastmilk until 4 mo, no water until after

4 moPromotion of vitamin A–rich foodsUse small feeding bowl, supervise feeding, feed

patiently and persistently

5-step counselling approach, live drama, radioprogrammes, flip charts, counselling cards,training in negotiation

TABLE 3. Programmes to improve complementary feeding of infants 6–12 months of age in developing countries: Messagesand communication strategies (continued)

Site Messages and interventions Communication strategies

Indonesia [21] Growth monitoringBreastmilk only until 4 mo, feed from both breastsIn addition, at 5–8 mo, feed enriched rice

porridge bubur campur 4 times/d, feed patientlyAt 9–24 mo feed adult food 4 times/d, including

tempe, tahu, or fish, and green vegetables; offersnacks

Continue breastfeeding as long as possible

Radio education spots, Kader (communityhealth worker) training, posters for familiesto have in house

Indonesia [20] Growth monitoringGive calorie- and nutrient-dense mashed family

foods at 4 mo; mix with fat for infants 6–9 moold (nasi tim bayi); add adult foods to providecomplete meal at 10 mo

Feed 6- to 9-mo-olds 4–6 times/d, plus breastmilk;feed 10- to 24-mo-olds 5 times/d (meal + snack)plus breastmilk; feed children ≥10 mo old largerportions and get more variety in diet

Children don’t know what is best for them:mothers do

Feed patientlyEnhance parental aspirations for child

Mass media (jingles, posters, cassettes, leaflets)Age-specific counselling cards with messages,

feeding schedule insert for growth card


193

were designed to provide affordable, energy- and nu-trient-dense foods suitable for daily consumption byolder infants. These foods, along with breastmilk, formthe base of the diet of older infants as they make thetransition to the solid foods consumed by the family.As described earlier, the energy- and nutrient-densefoods tended to be protein- and micronutrient-enrichedversions of local complementary foods, enrichedporridges with malted flour added to decrease the vis-cosity, adaptations of local foods not typically providedto older infants, or local enriched foods usually fed onlyduring illness that are promoted for everyday use.

Offer foods frequently throughout the day. As the in-fant’s energy needs increase, the need to consume morenon-breastmilk foods increases; however, the infant’sstomach cannot handle large amounts of food at onemeal. Therefore, it is recommended that smalleramounts of foods be offered several times throughoutthe day. How frequently should they be offered? Thatdepends on the energy density of the foods offered:the more energy-dense the foods, the less frequentlyinfants need to be fed in order to meet their energyrequirements [3]. The projects made decisions regardingfeeding frequency based on the energy density of thelocal complementary foods, with most programmesrecommending that infants between 6 and 12 monthsof age be offered three to five meals each day. As shownin the programme in the Dominican Republic [19],mothers of children identified as not growing well wereadvised to increase feeding frequency and variety, andwere reminded that foods should not be watered down.

Feed patiently and persistently. Feeding older infantscan be difficult and therefore requires patience andpersistence. Such messages about feeding style wereincorporated into several of the projects. Mothers werealso encouraged to take a more proactive feeding styleand to keep trying to feed the infants even when theydid not appear to be hungry or when they rejected foods.In Indonesia mothers were told that mothers and notinfants know what is best. In Mali [23, 24] the pro-gramme encouraged mothers to feed their children fromsmall bowls and to supervise them while they ate.

Increase variety in the infant’s diet. A well-acceptedprinciple in nutrition is that a diverse diet is a healthfuldiet. Children need to be offered a variety of foods sothat they can develop tastes and preferences and moveaway from a one-food liquid diet (breastmilk) to thevariety of foods that will characterize their diet for therest of their life. One programme in Indonesia [20]explicitly incorporated the concept of variety in theireducational messages to caretakers of infants 10 monthsof age and older. Other programmes stressed varietyimplicitly by encouraging the consumption of foodsfrom distinct food groups, particularly foods rich invitamin A.

Continue breastfeeding for as long as possible. Eventhough the infant is consuming other foods, breast-

milk is still a nutritious food and should continue tobe offered. Recognizing the multiple benefits of breast-feeding for both the mother and the child, the pro-grammes supported continued breastfeeding. It shouldbe noted, however, that none of the projects developedmessages about maintaining high levels of breastmilkintake during the first year of life as complementaryfoods are introduced and their consumption increases,or about how to do it.

Mass media and one-on-one counselling approaches were usedjointly to improve complementary feeding

Each of the programmes used a variety of communi-cation approaches to impart its messages to the popu-lation. First, mass-media techniques were used to reachthe target population of mothers with infants as wellas the larger population. Such approaches are neces-sary to change cultural norms regarding optimal com-plementary feeding. Second, one-on-one interactionsbetween community health workers and mothers wereused to provide individualized information and sup-port to mothers of infants. In several of the programmes,the counselling component was explicitly integrated withgrowth-monitoring programmes, and communityhealth workers were provided with training in growthmonitoring and promotion, complementary feeding,and specific counselling techniques. Importantly, theapproach to counselling was both process- and action-oriented. For example, in Mali [23, 24] a five-step coun-selling approach was followed that stressed negotiationwith mothers as opposed to education per se. Finally,each programme produced printed materials such ascounselling cards, recipe booklets, leaflets, and postersfor mothers to take home and use.

Impact of programmes on maternal knowledge andpractices

How effective were the educational strategies used inthese programmes? Presented in table 4 are the reportedimpacts of the interventions on maternal knowledgeand practices related to complementary feeding. How-ever, before examining programme impact, it is im-portant to consider the coverage achieved in theseprojects as well as the evaluation strategies.

Most of the projects achieved good coverage of thetarget population, but the coverage rates varied depend-ing on the communication strategy. As expected, mass-media strategies reached a greater percentage of thetarget population than did strategies involving indi-vidualized or one-on-one interaction. A word of cau-tion should be added here, because little documenta-tion was available that described how coverage wascalculated. Nevertheless, the data provided suggest thatthe interventions did in fact take place and that a size-able proportion of the potential beneficiaries wereexposed to one or more messages.


194

TABLE 4. Programmes to improve complementary feeding of infants 6–12 months of age in developing countries: Impacton maternal knowledge and feeding practices

ImprovedSite Evaluation design Exposure or coverage Appropriate knowledge feeding practices

Nigeria[27, 28]

1-wkintervention

Post-interventionsurvey of 295participants and 301non-participantsconducted 2–8 wkafter intervention

937 mothers from 11communities wereexposed

57% of participantscould accuratelydescribe eko ilera

2% dissemination tonon-participants

48% tried eko ilera17–24% adopted the

recipe and had thenecessary ingredientsat home

Peru [16]5-mo

intervention

30 cluster survey of648 representativefamilies conductedafter intervention

87% radio25% home clubs19% print material16% friends5% health personnel

82% knew of sanquito 16% tried food in class,12% tried it at home(19% and 15% ofthose who knew,respectively)

12% overall adoption(15% of those whoknew, 76% of thosewho ever tried, 98%of those who tried athome)

Peru [17, 18]1-yr

programmedivided into2 6-mophases,conductedin 9 commu-nities

Pre-post cross-sectional survey of~200 infants 0–12mo old from the 9communitiesconducted 1 yrapart. Interview,24-h recall, structuredobservations,weighed intakes, andanthropometry.Ethnography basedon maternal adoptionpost-programme

98% any exposure80% posters29–45% had booklets57% knew of courses29% attended courses61–88% heard

breastfeedingmessages

44–58% heardweaning foodmessages

73% “Every mothercan produce enoughmilk”

88% “Don’t give othermilks”

61% “Don’t give teas”86% “Begin

complementaryfoods at 6 mo”

44–55% “Purees andpuddings are morenutritious thansoup”

58% “Add 1 teaspoonoil to complementaryfoods”

8% ↑ breastfeeding6–12 mo

24% ↓ giving water< 4 mo

50% ↓ complementaryfoods at 3 mo

32% ↓ complementaryfoods at 5 mo

83% ↑ feeding thickfoods 2 times/d

No change in specificfoods

400% ↑ use of purees,but not puddings

No change in no. offoods per day

DominicanRepublic[19]

Cross-sectionalsurveys of 18randomly selectedprogrammecommunities and18 matched non-programmecommunities. Alsofollowed programmefamilies through the3-yr programme

70% overall contactwith programme

18% ↑ belief inexclusivebreastfeeding for< 4-mo-olds

30% ↓ introductionof complementaryfoods < 4 mo

70% ↓ use of cow’smilk and other foodsinstead of breastmilk0–4 mo

200% ↑ breastfeedingon demand; nodifference in no. offeeds or snacks for 4-to 12-mo-olds

200% ↑ 3 food groups(diversity)


It should also be noted that significant exposure toprogramme activities (contamination or crossover)occurred in the comparison communities in some sites.

In several programmes, the level of exposure wasminimal (2% to 3%). However, in Mali [23, 24] 18%of respondents in the comparison (non-programme)


195

communities had been exposed to one or more pro-gramme activities or messages. Exposure of non-pro-gramme communities to the programme complicatesthe evaluation process and leads to an underestimationof programme impact. It also speaks well of the inter-vention itself, in that it was diffused beyond the limitsof the target population.

A variety of approaches were taken to evaluate theimpact of these programmes. In a few of them, pre- andpost-intervention surveys were conducted in programmecommunities. The principal problem with this approachis that it may overestimate the impact of the programmebecause it does not separate the impact of the pro-gramme from secular improvements in knowledge and

TABLE 4. Programmes to improve complementary feeding of infants 6–12 months of age in developing countries: Impacton maternal knowledge and feeding practices (continued)


Burkina Faso[25, 26]

2.5-yrintervention

Baseline and finalsurveys ofprogrammecommunities

52% receivednutritional advicefrom communityhealth workers

60% recalled thecounselling card

21% radio dramas

Overall low retentionof messages

50% reported thatchildren should befed 3 times/d withvariety

20% ↑ correct timingof introduction ofsolids

Mali [23, 24]3-yr interven-

tion involv-ing multiplenon-govern-mentalorganizationsand privatevoluntaryorganizations

Baseline, mid-term,and final evaluationsurveys inintervention andcomparisoncommunities

45% recalled aprogramme activity

18% cross-treatment

50% ↓ giving water< 4–6 mo

Consumption at4–6 mo:

29% ↑ porridge140% ↑ fruit57% ↑ cow’s milk300% ↑ greens150% ↑ meat and liver

Cameroon[29]

Varyingdurations ofprogrammeactivities

Post-interventionsurvey in 23intervention(n = 463) andcomparison (n = 302)communities after6–12 mo ofprogrammeimplementation

41% ↑ “Feed moreafter illness”

8% ↑ “Continuefeeding duringdiarrhoea”

11% ↑ “Can encouragea sick child to eat”

13% ↑ bouillie enrichie

Indonesia[20]

1-yrprogramme

Baseline population-based survey andfinal surveyconducted 1 yrafter programmeimplementation in3 programme and 3matched-comparisonkecamatan. Alsofollowed asubsample ofprogramme childrenlongitudinally

50% recalled messagecorrectly

25% recalled messagebut not correctly

50% ↑ exclusivebreastfeeding at 3mo

At 6–9 mo childshould receive:

animal protein 15% ↑vegetable protein

22% ↑vegetables 21% ↑oil 51% ↑feed at least 3 times/d

16% ↑fish 15% ↑1 spoon/mo of age

9% ↑feed 4 times/d at 12

mo 51% ↑

6% ↓ complementaryfoods before 4 mo

21% ↑ nasi tim bayoverall

288% ↑ nasi tim bayprepared correctly

266% ↑ maternaladvice sharing



196

practices. Two projects compared outcomes amongparticipants with those among non-participants in thesame community. The potential problem with this ap-proach is one of selection bias; that is, an undocumentedprocess exists through which individuals become par-ticipants or remain as non-participants, and this processcan lead to a lack of comparability between these twogroups of individuals within the community. This lackof comparability may lead to either overestimates orunderestimates of programme impact. Other projectsexamined impact by comparing outcomes betweenprogramme and comparison communities. To improvecomparability, the comparison communities werematched to programme communities on the basis oftheir socio-economic and demographic characteristics,and in a few cases, both baseline and final evaluationswere conducted. The strongest design would have beento randomize individuals or communities to receivethe programme activities. However, to the best of ourknowledge, none of the projects used randomizationprocedures to identify programme and comparisoncommunities.

Maternal knowledge and practices were assessed usingknowledge, attitude, and practice (KAP) survey tech-niques. Mothers were asked to recall specific messagesof the programme, and the accuracy of their responseswas noted. They were also asked whether they agreedwith a series of belief statements regarding breastfeed-ing and complementary feeding. Most projects com-pared persons in programme and non-programme com-munities with respect to changes in their knowledgeof specific intervention messages over time. Only oneproject created knowledge composite scores based onthe percentage of correct responses [21]. In general,the results indicate that the interventions resulted inlarge shifts in maternal knowledge regarding optimalfeeding during the first year of life.

The impact of the programme on infant-feedingpractices was assessed via interviews or structured ob-servations and weighed intakes. Most of the projectsused survey techniques that provided information oninfant-feeding practices as reported by the mother.Mothers were asked how they were currently feedingtheir infants and were administered a 24-hour recalltest to describe what their infants had eaten on the daybefore the interview. The 24-hour recall also providedquantitative information on dietary intakes. The re-sults suggest large differences in the frequency withwhich recommended feeding practices were reportedin programme versus comparison communities. In thefollowing section, we consider whether the changes inmaternal knowledge and reporting practices translatedinto measurable improvements in the dietary intakesand nutritional status of programme participants.

Impact of programmes on complementary foodintake and nutritional status

Changes in the adequacy of the dietary intake of par-ticipating infants were reported for only three pro-grammes. This information was converted to intakein kilocalories per day for comparison with the resultsof efficacy trials. As shown in table 5, the results sug-gest that the projects were able to increase the energyintake of 6- to 12-month old infants by about 70 to165 kcal/day. This change is impressive, because it coversthe lower range of net improvements in dietary intakereported by the efficacy studies described earlier (65–300 kcal/day). It is important to note, however, thatseparate information on energy intakes from breast-milk and non-breastmilk foods was not provided, andthus we do not know whether increased consumptionof complementary foods reduced breastmilk intake. Onthe basis of the results of the efficacy studies, this cer-

TABLE 4. Programmes to improve complementary feeding of infants 6–12 months of age in developing countries: Impacton maternal knowledge and feeding practices (continued)


Indonesia[21]

3-yrprogramme

Baseline survey ofproject communities,plus final cross-sectional survey of600 project and 400matched-comparisonhouseholds withinfants < 24 moof age

88% got advice fromkadar (health worker)

78% attended growthmonitoring regularly

85% growth monitor-ing education

37% personalcounselling

79% received at least1 poster

44% heard radio spotsHigh rate of

contamination ofcomparison group

Knowledge scores (%correct): 75% vs45% overall. Formothers of infants0–4 mo: 70% vs48% correct; ofinfants 5–8 mo: 58%vs 10% correct; ofinfants 9–24 mo:90% vs 85% correct.

Infants 5–8 mo oldconsuming onprevious day:

greens 200% ↑coconut milk 253% ↑oil 54% ↑fish 200% ↑tempe 290% ↑negative effect on tahububur campur 350% ↑+ effects also seen at

9–24 mo


197

TABLE 5. Programmes to improve compleentary feeding of infants 6–12 months of age in developing countries: Impact oninfant dietary intakes and nutritional status

Site Change in dietary intakes Baseline WAZ, HAZ Change in WAZ, HAZ

Dominican Republic [19] –1.0 WAZ, 0–11 mo +0.24 WAZ, 0–11 mo–1.4 WAZ , 12–23 mo +0.36 WAZ, 12–23 mo

Peru [17, 18] 30% ↑ energy density: –0.23 WAZ, 6–12 mo –0.08 WAZ+19 kcal/100 g at 6–12 mo

Mali [23, 24] –1.18 WAZa, 12 mo + 0.41 WAZ–1.75 HAZ, 12 mo + 0.87 HAZ

Indonesia [20] No change at 4–5 mo –2.0 WAZa, 12 mo +0.50 WAZ↑ 71 kcal/d at 6–9 moa –1.6 HAZa, 12 mo +0.40 HAZ

Indonesia [21] ↑ 79 kcal/d, 5 moa –1.7 WAZa, 12 mo +0.35 WAZa↑ 141 kcal/d, 9 moa –1.8 HAZa, 12 mo +0.30 HAZa (corrected for↑ 164 kcal/d, 12 moa differences at baseline)

Abbreviations: HAZ, height-for-age Z score; WAZ, weight-for-age Z score.a. Extrapolated estimates based on data available in reference.

tainly seems likely. Thus, the actual improvements incomplementary food intakes may be larger.

Also presented in table 5 are estimates of the im-provements in nutritional status of infants achievedin the project area. This information was not presenteduniformly across all programme reports; therefore, theinformation was transformed to make it comparableto that presented in the efficacy studies. The methodsused were similar to those described earlier; exact de-tails on the transformations used are available from theauthors. With the exception of the programme in Peru[17], all of the projects with available data showed posi-tive programme impacts on nutritional status. The lackof impact of the project in Peru may be related to theevaluation strategy, which consisted of cross-sectionalnutritional status surveys in the communities beforeand after the intervention. When using this strategy,one cannot separate programme impact from seculartrends in nutritional status over the life of the project.The range of positive impacts was on the order of +0.24to +0.87 SD for both weight-for-age and height-for-age. For several reasons described above, however, it islikely that these estimated impacts are somewhat over-stated. Thus, the actual improvements in nutritionalstatus are probably smaller. However, even if the trueimpact was overestimated by 50% (i.e., infants of 0.10–0.40 SD), the impact would still probably be largeenough to translate into tangible reductions in mal-nutrition of older infants in developing countries.

Discussion

The results of this analysis provide evidence that in-

creasing complementary food intakes of 6- to 12-month-old infants is both an efficacious and an effective ap-proach to reducing early childhood malnutrition indeveloping countries. The first section reviewed fiveefficacy studies that provided an additional 65 to 300kcal/day, resulting in improvements in growth of 0.1to 0.5 SD. These studies used randomized designs, andconfounding issues were addressed. Overall, they werewell conducted and provide causal evidence that im-proving dietary intakes of infants living in impover-ished environments will improve their growth.

In the next section, we reviewed programme projectsthat attempted to reduce malnutrition by improvinginfant-feeding practices. The information is less com-prehensive, because several of the projects examined werenot formally evaluated for impact. The evaluation de-signs, when available, were not optimal, although mostat least conducted pre-post evaluations of maternalknowledge, infant-feeding practices, and infant nutri-tional status in both programme and comparison com-munities. One way to strengthen these types of designswould be to randomly assign individuals or commu-nities to programme or comparison groups. Programmeplanners and policy makers, however, are reluctant toadopt such strategies. If randomization is not done, dem-onstration of comparability of the groups and an analysisof congruence of the findings become more impor-tant. Such analyses were performed to a greater or lesserextent in these evaluations. Thus, the results provideplausible evidence that comprehensive, multifacetedintervention approaches involving breastfeeding pro-motion and improved complementary feeding can iden-tify affordable and acceptable means for caretakers toimprove infant feeding, improve caretaker knowledge


198

and beliefs regarding optimal infant feeding, improveinfant-feeding practices, increase total energy intake,and improve the nutritional status of older infants.

The evaluations provide evidence that improvementsin dietary energy intake of 70 to 165 kcal/day were as-sociated with changes in growth rates on the order of–0.04 to +0.87 SD. Thus, it would appear that the pro-gramme projects were able to produce equivalent orgreater improvements in nutritional status with aboutone-half the improvement in dietary energy intakes.How can this be? The answer probably lies in the factthat whereas the efficacy trials used one method toimprove infant feeding (direct provision of additionalfoods), the programmes intervened in multiple waysand at multiple moments, beginning at birth and con-tinuing throughout infancy and beyond. Thus, it is likelythat some of the improvements in nutritional statusare a direct result of improved breastfeeding practices(particularly improvements in the duration of exclu-sive breastfeeding) as well as improved means andmotivation for improved complementary feeding. It isalso likely that evaluations that provide plausible evi-dence of programme impact are more subject to smalllevels of positive confounding that lead to a slight over-estimation of programme impact.

Despite the variability in the impact of programmeson growth rates of infants, most efficacy studies as wellas programmatic efforts increased the growth rates ofchildren by 0.10 to 0.50 SD. What do these rates ofimprovement in growth mean for reducing malnutri-tion in a population? Assuming that the distributionsof Z scores post-intervention are normal, with an SDof 1.0, we can calculate the impact of an upward shiftin the average Z score of 0.10 to 0.50 SD and the ex-pected prevalence of malnutrition (% < –2 Z score).This is done in table 6A and B for various levels ofnutritional status in the population and interventionimpact. As shown, improvements in growth on the orderof 0.10 SD imply absolute reductions in the prevalenceof malnutrition of 2% to 4%, whereas improvementson the order of 0.50 SD imply reductions of 5% to19%, depending on the underlying prevalence of mal-nutrition in the population (assumed to be between7% and 69%). If we consider the average positive im-pacts on growth from both studies and programmesto be on the order of 0.2 to 0.3 SD, then we can expectabsolute reductions in the prevalence of malnutritionon the order of 3% to 12%, depending on the under-lying prevalence in the population.

It is well known that poor nutritional status con-tributes to more than half of child deaths in develop-ing countries. For this reason, interventions to improvenutritional status are integral components of child sur-vival efforts worldwide. Therefore, it is likely that im-

provements in nutritional status will also translate intoreductions in mortality rates for infants and young chil-dren. Recently, Pelletier et al. [6] provided a methodof quantifying the proportion of deaths among chil-dren attributable to malnutrition, based on the tradi-tional statistic, population attributable risk (PAR). Weused this technique to provide an indication of thepotential reduction in mortality attributable to mal-nutrition that would result if programmatic efforts wereable to improve the growth rates of older infants by0.10 to 0.50 SD. These results are shown in table 6C.The underlying PAR in the community is a functionof the underlying prevalence of malnutrition (charac-terized here by the average Z score), and the changein PAR depends on the improvement in nutritionalstatus. These extrapolations indicate that programmaticefforts that improve the nutritional status of childrenby 0.10 to 0.50 SD might reduce the proportion of deathsattributable to malnutrition by 2% to 13%, depend-ing on the underlying prevalence of malnutrition inthe community.

Several words of caution must be said regarding thereview and the analytic results extrapolated from thesedata. First, it is likely that positive publication bias ex-ists for both types of interventions, meaning that studiesnot able to improve the growth rates of older infantswere less likely to be published in the scientific litera-ture, and that few evaluations of ineffective programmesmay have been available for review. Thus, the overallfindings presented here may be overstated to an unknowndegree. Second, in the evaluation we have emphasizedthe validity and magnitude of the effect sizes reportedrather than the statistical significance (or lack thereof)of specific estimates. In fact, some of the results fromprogramme evaluations were not statistically signifi-cant at traditional alpha values of 0.05. Statistical sig-nificance is arguably less important in this type of analy-sis, given the relative dearth of information that wasavailable a priori for designing optimal evaluations.

Despite these limitations, nutritionists, health pro-fessionals, and policy makers should take note of thelikely magnitude of the positive contribution that com-plementary feeding inter

F&N Bulletin Vol 20 No 2 - United Nations University · 2019. 4. 16. · 180 H. I. Tawfeek et al....

Documents

Transcript of F&N Bulletin Vol 20 No 2 - United Nations University · 2019. 4. 16. · 180 H. I. Tawfeek et al....