![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
3 RTI International, Research Triangle Park, NC 27709 and 4 University of North Carolina School of Public Health, Chapel Hill, NC 27599
* To whom correspondence should be addressed. E-mail: bernadette_marriott{at}abtassoc.com.
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionMethodsResultsDiscussionLITERATURE CITED |
|---|
|
Introduction |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionLITERATURE CITED |
|---|
In this article, we describe the intake of infants in 20 developing countries using secondary analysis of data from the Demographic and Health Surveys (DHS). The purpose of this article is to describe the study population and initial results of reported feeding practices of infants. The foci of our analysis were the responses of mothers when asked whether they were currently breast-feeding and whether they had fed other liquids and solid foods to their infants in the previous 24 h.
|
Methods |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionLITERATURE CITED |
|---|
Under contract with the U.S. Agency for International Development since 1989, Macro International, Opinion Research Corporation (ORC Macro) updates the DHS basic model instrument approximately every 5 y and provides assistance to in-country agencies and institutions. The DHS core questionnaire is designed to collect information on anthropometry, child feeding, and child morbidity. Specific questions and answers are modified to be country or culturally specific. Countries may select to run surveys multiple times or once. Since the inception of the DHS in 1984, 76 countries globally have participated in surveys.
ORC Macro has implemented the DHS to provide standardization from country to country in terms of quality of data collection and data analysis. Because of the use of uniform DHS instruments, cross-country comparisons and data pooling are possible to study child feeding and other factors that have a bearing on health outcomes in the developing countries that participate in the DHS (9–11).
Samples. For the purpose of this study, we accessed MEASURE DHS+ data collected from developing countries between 1999 and 2003, using the Model B survey instrument. Only 24-h and 7-d intake questions were included in the DHS survey instruments between 1999 and 2003. During this period, 38 surveys were conducted in 5 global regions. The number of countries with surveys during the time period examined are listed with the total number of countries where surveys have ever been conducted in these global regions as follows (number during time period/total number): Sub-Saharan Africa (20/37), North Africa/West Asia/Europe (4/9), South Asia/Southeast Asia (7/11), Central Asia (2/4), and Latin America/Caribbean (5/15). We selected for the study only those countries that provided child-level focused datasets, had included 24-h and 7-d intake questions, and had data that had been made available by the country, including an English-language survey instrument. For our preliminary analysis we excluded the Latin America/Caribbean region. Our data sources therefore included (number of countries of number during time period): 9 of 20 in Sub-Saharan Africa [Ethiopia (2000), Ghana (2003), Kenya (2003), Malawi (2000), Namibia (2000), Nigeria (2003), Uganda (2001), Zambia (2001), Zimbabwe (1999)]; 3 of 4 in North Africa/West Asia/Europe [Armenia (2000), Egypt (2000), Jordan (2002)]; 7 of 7 in South Asia/Southeast Asia [Bangladesh (1999), Cambodia (2000), India (1999), Indonesia (2002), Nepal (2001), Philippines (2003), Vietnam (2002)]; 1 of 2 in Central Asia [Kazakhstan (1999)]; and 0 of 5 in Latin America/Caribbean. The Eritrea (2002) and Turkmenistan (2000) data that met our inclusion criteria had not been released by the countries when we began our study.
We verified that the data we downloaded were the correct dataset for our analyses. We also calculated the country-specific sample numbers and compared these numbers to reports on the ORC Macro website to verify that full datasets were included for each country. For our study, we selected the youngest child less than 1 y of age in each of the families, if the youngest child was living. We excluded deceased children and all children for whom age information was missing. Multiple children from 1 mother were included in the analysis only if the children were born at the same time (i.e., twins or triplets). Analyses were conducted separately for infants <6 mo old (0–6 mo) and infants 6 mo or older but <12 mo old (6–12 mo). For the 20 countries, we used data from 35,648 infants (18,174 aged 0–6 mo and 17,474 aged 6–12 mo). The number of infants sampled in each country ranged from as few as 99 (0–6 mo) and 115 (6–12 mo) in Kazakhstan to 5533 (0–6 mo) and 5083 (6–12 mo) in India. For the 2 age groupings, the mean age (SD) was 2.7 (0.1) and 8.5 (0.1) months for 0- to 6-mo-old and 6- to 12-mo-old infants, respectively.
Variables. From the Model B Women's Questionnaire for 1999–2003, we used the following variables for this descriptive analysis: maternal last birth, next-to-last birth, child living or deceased, child twin, child age, breast-feeding still, times breast-fed in last 24 h, other liquids in last 24 h, solid foods in last 24 h, liquids in the last 7 d, and solids in last 7 d. For this cross-country comparison, we included those countries where an English language questionnaire was provided so we could carefully review each question set for skip patterns and country-specific variations in question wording or reply options. For each country, we reviewed coding of variables to confirm comparability among the 20 countries. Several countries introduced small wording changes into the survey questions; more frequently, cross-country variability was introduced by the addition or substitution of response options. Additional cross-country variation occurred because some countries chose to omit questions in the model survey. Thus, the numbers of countries providing responses are fewer for some variables than other variables. Among the variables presented in this article, Namibia did not inquire whether infants were fed tinned, powdered, or other milks; Ethiopia, Namibia, India, and Nepal did not inquire if infants were fed commercial formula.
One inclusion criterion for this analysis was that the model questionnaire used by the country included both 24-h and 7-d intake questions for liquid and food. Some of the countries that used these 2 questions, however, did not include both questions for all liquid or food types. In 4 countries (Jordan, Indonesia, Nepal, and the Philippines) a negative response to a food or liquid type on the 7-d intake question resulted in a missing data code in the parallel 24-h question. The other 16 countries instead recorded a zero value. We recoded all 24-h liquid and food intake variables so that if the 7-d variables for an item had a value of zero, then the 24-h variable also had a zero value.
For this analysis, food and liquid variables other than water and infant formulas were collapsed into broader food and liquid categories to permit cross-country comparisons. Specifically, we collapsed tinned, powdered, and any other specified animal milks; all other liquids such as fruit juices, herbals teas, and sugar water (excluding unique country-specific liquids) into other liquids, and the 10 types of solid food groups into an any-solid-foods category.
Analytical methods. In the model questionnaire, mothers were asked if they were still breast-feeding and how many times their infant was fed specific liquids or solids in the last 24 h. To compute the percentage of infants who consumed items, we recoded the variables so that any response indicating consumption was set to 1 (consumed item), and any response indicating no consumption or didn't know whether there was consumption was set to 0 to indicate no consumption. Missing values were treated as missing. Mean values are calculated across the 20 countries.
We used sample weights for all analyses. DHS+ data are weighted by country to account for variability of the country-based sampling models overall. In countries where only married women were sampled, we adjusted the rates to reflect all women in the country, and we standardized the weights for individual countries to represent the number of respondents on a variable basis.
We also pooled data from the 20 countries to provide a large sample size, increase statistical power, and present a more global perspective. According to Hatt and Waters (10), pooling can reveal broader results that are "often obscured by the noise of individual data sets." To calculate the pooled values an additional adjustment is needed to account for the variability in the number of individuals sampled in each country. This is accomplished using the weighting factor 1/(A*nc/nt), where A is the number of countries asked a particular question, nc is the number of respondents for the country c, and nt is the total number of respondents over all countries asked the question (11).
|
Results |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionLITERATURE CITED |
|---|
|
|
|
The pooled percentage of mothers who reported feeding their babies water in the last 24 h was twice as high for infants at 6–12 mo as for 0- to 6-mo-old infants. This difference reflects consistently higher reported water feeding by mothers for 6- to 12-mo-old infants for all countries. Reported water feeding of the younger infants varied widely among the 20 countries, ranging from 15.1% in Uganda to 87.9% in Cambodia. Water feeding of 6- to 12-mo-old infants ranged from 65.9% in Uganda to 97.5% and 99.3% in Cambodia and Jordan, respectively.
The use of tinned, powdered, or other milks was generally low for 0- to 6-mo-old infants but was higher (as was all reported fluid feeding) in 6- to 12-mo-old infants. The variability of response was high across countries where this question was asked. In Indonesia, Malawi, and Zimbabwe, <2% of mothers responded that they used these other types of milks for infants aged 0–6 mo, whereas over one-third of mothers in Kenya reported its use for the same age group (35.1%). All countries reported a higher use of these products with 6- to 12-mo-old infants with the country-specific data showing a 1.6- to 4-fold higher use among 6- to 12-mo-olds than 0- to 6-mo infants in each country. This question was omitted in the Namibian survey instrument, so no data are available.
Sixteen of the 20 countries sampled included a specific question on commercial formula use in the last 24 h. No data are available on infant formula use for Ethiopia, Namibia, India, and Nepal. Commercial infant formula was least reported of all fluids by mothers with infants in both age categories in the 16 other countries. For the 0- to 6-mo age category, 11 of the 16 countries reported <10% formula use. The exceptions were Indonesia, Jordan, and the Philippines, in which mothers respectively reported 24.2%, 25.1%, and 30.0% use of infant formula. Mothers in these 3 countries also reported the highest use of commercial infant formula in the 6- to 12-mo age category (Jordan 28.8%, Indonesia 38.8%, and the Philippines 47.1%). Generally, reported use of formula was greater for older infants. The 2 exceptions were Zimbabwe and Armenia, where fewer mothers reported feeding their older infants formula than did mothers with infants in the younger age category (Zimbabwe: 0–6 mo, 3.2%; 6–12 mo, 2.4%; and Armenia: 0–6 mo, 13.1%; 6–12 mo, 11.6%).
As described above, use of a wide variety of other liquids was reported. A portion of these other liquids were either country- or culture-specific and had infrequently reported use. In contrast, the use of fruit juice, herbal tea, and sugar water was queried in all 20 countries. Combined, these 3 types of sweetened liquids were in widespread use. More mothers reported the use of other liquids than reported use of other milks or formulas. Use of other liquids ranged from 2.8% in Nepal to 48.1% in Armenia for 0- to 6-mo-old infants. Higher use was reported in all countries for the older infants, ranging from 15.8% in Nepal to 91.8% in Kazakhstan.
These pooled multicountry data indicate that over 20% of mothers are providing some type of solid food to infants aged 0–6 mo, but the practice varies greatly among countries. Ethiopian and Indian mothers reported the lowest percentage of solid food use for their 0- to 6-mo-old infants (5.1% and 6.5%, respectively), whereas >40% of mothers in Malawi, Indonesia, Zimbabwe, and Kenya reported feeding solid foods to their 0- to 6-mo-old infants. Reported use of solid foods more than tripled to 80.1% (pooled value) for 6- to 12-mo-old infants, with reported percentages ranging from a low of 44.1% by Indian mothers to over 90% by mothers in Zambia, Zimbabwe, Malawi, and Jordan.
|
Discussion |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionLITERATURE CITED |
|---|
These pooled values indicate that, overall, 
97% of mothers report that they currently breast-feed their 0- to 6-mo-old infants at least some portion of the day, with 88% of mothers saying they currently breast-feed their 6- to 12-mo-old infants. These data do not speak to the exclusivity of this practice. The DHS do not include an exclusive breast-feeding question but rather recommend using a combination of intake questions to derive an estimate of exclusive breast-feeding in a population (8). Nonetheless, the percentage of mothers of infants of both age categories who state that they are breast-feeding is encouraging in light of many recent concerns about changes in breast-feeding practices (12). In fact, the pooled mean for any-current-breast-feeding through 12 mo (nearly 88%) was substantially higher than that reported for the sample of mothers selected for study in the construction of WHO growth curves, 68% (13). This suggests that other aspects of feeding recommendations, such as the type and timing of other foods introduced, should be the focus of feeding interventions.
Pooled values indicated that mothers in these countries reported using commercial formula less than the 3 other categories of liquids to feed their babies and most often reported using plain water or other liquids with all infants 0–12 mo old. Large differences were seen among countries in mothers' use of these 4 categories of liquids. However, we are concerned by the high percentage of mothers in each country and overall who reported feeding their infants the "other liquids" category because this category consists of nearly all sweetened beverages. We expect that many of these differences will be related to other variables such as family economic status, maternal education, and household sanitation when we complete our ongoing modeling analysis (14,15).
Solid food feeding, based on pooled analysis, was higher than expected for both age categories with mothers reporting feeding solid food to 21.9% of infants 0–6 mo of age and 80.1% of infants aged 6–12 mo. These results were variable across the 20 countries. WHO recommends introduction of solids at 6 mo (16); our data show that 20% of infants in the 20 developing countries in our sample are not compliant with this recommendation. As mentioned earlier, the DHS includes questions on specific food types that we have combined into the category of any-solid-food for the purpose of this analysis. Our future analyses will evaluate this solid food intake by country and by food type including the relation to socioeconomic variables. Based on the research reported by Dr. Krebs in this supplement, we are particularly interested in evaluating the use of meat and fish in 6- to 12-mo-old infants in this 20-country sample as well as other complementary food types and their relation to infant health (17).
To date, we have initially examined only the 24-h intake data. Ongoing analyses include data from the 7-d recalls and feeding practices for the infant's first 3 d of life. For the preliminary analysis we combined solid food into 1 category, but we will be examining specific food types and the timing of their introduction by month across the first 12 mo of life. Maternal and infant demographic and socioeconomic variables such as maternal education, maternal employment, and household variables, such as urban or rural location, water sources, the availability of refrigeration, and cooking fuels will be examined in relation to outcome variables of infant health. We are using a series of modeling techniques to elucidate the relation among these distal, intermediate, and proximate variables on infant health outcomes in these 20 countries as well as how actual practices in developing countries reflect international guidelines for infant feeding.
Question and answer session
[Q1]: The term "current breast-feeding" means any breast-feeding and does not necessarily indicate exclusive breast-feeding for 0- to 6-mo-old infants, correct? So, this extraction of DHS data does not indicate if women are meeting guidelines for breast-feeding.
[Dr. Marriott]: Correct. The question posed by DHS is "Are you currently breast-feeding?" In the DHS there isn't an exclusive-breast-feeding question in the model questionnaire for 1999–2003. A variable can be derived from available data to provide an estimate of exclusive breast-feeding, which is something we intend to do.
[Q2]: I noticed that there are no countries included from Latin America. Is that by design or by chance?
[Dr. Marriott]: That's by design, because there were no countries in Latin America with a DHS questionnaire for the period 1999–2003 with both 24-h and 7-d intake questions that was also in English. When the country creates their country-level report and makes the data available, the questionnaire is attached. In order for us to pool the data, we needed to read each question and each variable response option and be able to verify the coding and the recoding for programming. One of our next phases is to look at the Spanish-speaking countries.
[Q3]: I come from Kenya and I've had some experience in data collection at household levels. How did you confirm that the data were obtained for/from the mothers and not for/from the primary caregiver? Given the high prevalence of HIV and AIDS in some of these developing countries, sometimes the child is in a different place than the mother.
[Dr. Marriott]: ORC Macro works very closely with the in-country survey collectors on the survey questions and methods for conducting the surveys. They also perform quality control measures before and after each survey is conducted. Survey data are not released unless they are sure they represent the design for the survey. This level of quality control is one of the reasons why these data are excellent to use, and another reason is that it allows for cross-country comparisons.
[Q4]: When you work with the data concerning diarrhea, how will you correct for instances where the family has reacted by reducing the child's food intake?
[Dr. Marriott]: There are actually DHS questions that address this specific topic, asking, "If your baby had diarrhea," and there are different time periods that this question is asked around: "Did you reduce the food intake for your baby?" "Did you provide different fluids?" And specific potential responses are elicited. So these data are available.
[Q5]: WHO recommends the method that DHS has used to derive and define "exclusive breast-feeding," in terms of excluding other liquids and foods to arrive at a derived variable. When mothers are directly asked whether they are exclusively breast-feeding, they can interpret this question differently. There has been the experience in some DHS countries where this question was asked and the findings were erroneous. One other comment, these data are cross-sectional, so there is no evidence of causality. And it's difficult to control for many aspects related to socioeconomic status that really affect both infant feeding practices and affect risk of different outcomes, be it morbidity or growth failure. I remember some early analyses of DHS by Shea or perhaps Rutstein showing that, if I remember right, breast-feeding was associated with greater outcome of diarrhea. So, this was indicative of the poorer mothers, who were breast-feeding more, under conditions in which their babies were more at risk of diarrhea. So, just a word of caution, as you look at your data, to be critical of the associations you find.
[Dr. Marriott]: Yes. I'm aware of the limitations with these approaches. With the 1999 to 2003 DHS+ dataset, ORC Macro has had trouble with their "exclusive-breast-feeding" derivation and analysis. They applied the same algorithm that they had always used, but for some reason "exclusive breast-feeding" appeared to change dramatically in this time period. As a result, they are conducting a more in-depth analysis of the DHS+ data to see if changes in the order in which the questions were asked could have affected the results. The DHS+ 1999 to 2003 survey instruments were the first to include both 7-d and 24-h liquid and food intake questions. ORC Macro is considering whether the addition of the 7-d recall questions may have confounded their "exclusive-breast-feeding" derivation. So they have cautioned us against moving ahead with the "exclusive-breast-feeding" derivation until they reach a conclusion about their derivation process for this dataset.
|
FOOTNOTES |
|---|
2 Author disclosure: Supported by contract 0280500.901 with Wyeth Nutrition.
5 Present address: Nutrition & Health Research, Abt Associates, Inc., Durham, NC 27703.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTIntroductionMethodsResultsDiscussionLITERATURE CITED |
|---|
1. Black RE, Morris SS, Bryce J. Where and why are 10 million children dying every year? Lancet. 2003;361:2226–34.[Medline]
2. Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and trends in obesity among US adults, 1999–2000. JAMA. 2002;288:1723–7.
3. Millard AV. A causal model of high rates of child mortality. Soc Sci Med. 1994;38:253–68.[Medline]
4. Mosley WH, Chen LC. An analytical framework for the study of child survival in developing countries. 1984. Bull World Health Organ. 2003;81:140–5.[Medline]
5. Raisler J, Alexander C, O'Campo P. Breast-feeding and infant illness: a dose-response relationship? Am J Public Health. 1999;89:25–30.
6. Bahl R, Frost C, Kirkwood BR, Edmond K, Martines J, Bhandari N, Arthur P. Infant feeding patterns and risks of death and hospitalization in the first half of infancy: multicentre cohort study. Bull World Health Organ. 2005;83:418–26.[Medline]
7. Bhandari N, Bahl R, Mazumdar S, Martines J, Black RE, Bhan MK. Effect of community-based promotion of exclusive breastfeeding on diarrhoeal illness and growth: a cluster randomised controlled trial. Lancet. 2003;361:1418–23.[Medline]
8. Measure DHS. DHS history. Over 30 years of history. Measure DHS 2006 [cited April 17, 2006]; Available from: URL: http://www.measuredhs.com/aboutdhs/history.cfm
9. Ruel MT, Menon P. Child feeding practices are associated with child nutritional status in Latin America: innovative uses of the demographic and health surveys. J Nutr. 2002;132:1180–7.
10. Hatt LE, Waters HR. Determinants of child morbidity in Latin America: a pooled analysis of interactions between parental education and economic status. Soc Sci Med. 2006;62:375–86.[Medline]
11. Peng YK, Hight-Laukaran V, Peterson AE, Perez-Escamilla R. Maternal nutritional status is inversely associated with lactational amenorrhea in Sub-Saharan Africa: results from demographic and health surveys II and III. J Nutr. 1998;128:1672–80.
12. World Health Organization. Effect of breastfeeding on infant and child mortality due to infectious diseases in less developed countries: a pooled analysis. WHO Collaborative Study Team on the Role of Breastfeeding on the Prevention of Infant Mortality. Lancet. 2000;355:451–5.[Medline]
13. WHO Multicentre Growth Study Reference Group. Breastfeeding in the WHO Multicentre Growth Reference Study. Acta Paediatr Suppl. 2006;450:16–26.[Medline]
14. Checkley W, Gilman RH, Black RE, Epstein LD, Cabrera L, Sterling CR, Moulton LH. Effect of water and sanitation on childhood health in a poor Peruvian peri-urban community. Lancet. 2004;363:112–8.[Medline]
15. VanDerslice J, Popkin B, Briscoe J. Drinking-water quality, sanitation, and breast-feeding: their interactive effects on infant health. Bull World Health Organ. 1994;72:589–601.[Medline]
16. Dewey K, Pan American Health Organization. Guiding principles for complementary feeding of the breastfed child. 2002 [cited 2006 Apr 26];Available from: URL: http://www.paho.org/English/AD/FCH/NU/Guiding_Principles_CF.pdf.
17. Caulfield LE, Huffman SL, Piwoz EG. Interventions to improve intake of complementary food by infants 6 to 12 months of age in developing countries: Impact on growth and on the prevalence of malnutrition and potential contribution to child survival. Food Nutr Bull. 1999;20:183–200.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
