Você está na página 1de 9

Factors Associated With Underweight and Stunting Among Children in Rural Terai of Eastern Nepal

G. C. Pramod Singh, MA, MPH, Manju Nair, MBBS, MPH, Ruth B. Grubesic, DrPH, and Frederick A. Connell, MD, MPH

Asia-Pacific Journal of Public Health Volume 21 Number 2 April 2009 144-152 2009 APJPH 10.1177/1010539509332063 http://aph.sagepub.com hosted at http://online.sagepub.com

Malnutrition continues to affect a large proportion of children in the developing world. The authors undertook this study to identify biologic, socioeconomic, and health care factors associated with underweight and stunting in young children in an the eastern Tarai (plains) district of Nepal. Data were collected via questionnaires from mothers of 443 children aged 6 to 36 months in Sunsari district. Multistage cluster sampling was used to select villages and children. Anthropometric measurements were made on both children and their mothers. Logistic regression was used to measure the independent (adjusted) effect of risk and protective factors on the odds of underweight or stunting. More than half (53.3%) of the children were found to be underweight (<2 standard deviations weight for age below reference median) and more than one third (36.6%) had stunting (<2 standard deviations height for age below reference median). Low maternal body mass index, childs age, higher birth order, and lower standard of living score were strong predictors of underweight, whereas mothers education >5 years and participation in vitamin A and nutritional programs were protective. Infant age, low maternal body mass index, and low standard of living score were significant risk factors for stunting, whereas mothers education >5 years was strongly protective. These results suggest that underweight and stunting are the result of a nexus of biological, socioeconomic, and health care factors. Keywords: malnutrition; children less than 3 years; underweight; stunting; Nepal

he World Health Organization estimates reveal that malnutrition is associated with about half of the 10.7 million child deaths among children less than 5 years occurring each year in the developing world.1 According to the 2001 Nepal Demographic and Health Survey, the prevalence of underweight (defined as weight for age <2 standard deviations [SDs] below the mean) among children less than 5 years in Nepal is 48.3% underweight, of which 50% of the children are stunted (defined as height for age <2 SDs below the mean).2 This is higher than in most other countries; in 2000, it was estimated that 27%
From the School of Public Health, B. P. Koirala Institute of Health Sciences, Dharan, Sunsari, Nepal (GCS); Achutha Menon Center for Health Science Studies, Sree Chitra Tirunal Institute for Medical Science and Technology, Thiruvananthapuram, Kerala, India (MN); Texas Womens University, Houston, Texas (RBG); University of Washington School of Public Health and Community Medicine, Seattle, Washington (FAC). Address correspondence to: G. C. Pramod Singh, MA, MPH, School of Public Health, B. P. Koirala Institute of Health Sciences, Dharan, Sunsari, Nepal; e-mail: gharti_pr@yahoo.com.

144

Underweight and Stunting Among Children in Nepal / Singh et al 145

of the worlds children less than 5 years of age were underweight and 33% were stunted.3 The majority of underweight and stunted children live in Asia, especially Southern Asia, and the risk of being underweight is about 1.5 times higher in Asia than in Africa.4 Malnutrition is responsible not only for mortality among children but also seriously affects the health of survivors predisposing them to infections and other illnesses. The effects of childhood malnutrition lead to physical and psychological sequelae continuing through adulthood, cause intergenerational effects, and loss of human potential, leading to loss of social productivity.5 Malnutrition is deeply rooted in poverty and underprivileged social environments in addition to being caused by biomedical reasons. During the past 2 decades, global trends have shown progress, with the prevalence rates of underweight children falling from around 27% in the 1990s to around 22% in 2000.6 The United Nations Millennium Development Goal for child mortality aims to reduce by two thirds the mortality rate in children less than 5 years between 2000 and 2015, and in developing countries, tackling malnutrition is the biggest challenge in achieving this goal.7 This cross-sectional study, conducted in eastern Nepal, attempted to identify risk factors associated for underweight and stunting in children. Specifically, the objectives of this study were the following:
1. To measure the prevalence of underweight and stunting in a representative sample of rural Nepali children under 3 years of age living in Sunsari district 2. Measure the association of biological factors, socioeconomic and demographic factors, and nutritional interventions on underweight and stunting 3. Assess whether use of 2 nutritional programs was associated with a reduction in underweight and stunting

Methodology
Multistage cluster sampling method was used to select subjects of this study. Four village development committees (VDCs) were selected randomly from the 49 VDCs in this district, and 4 wards out of 9 wards were selected from each VDC using a random sampling method. The selected ward was considered as a cluster. Thus, there were a total of 16 clusters in the sample. A minimum sample size of 400 children was based on an expected prevalence of 48%1 of underweight children in this district. Within each cluster, children less than 3 years of age were selected from every alternate household, and all eligible children in each family were included. If there was more than one family having a child of eligible age in a household, one of the families was selected randomly on the spot for inclusion in the study. The selected household was excluded if there was no eligible child. The children in a household were excluded from the sample if the mother was not present. A total of 443 children were included in the study. Each mother was interviewed, from August to September 2005, using a pretested questionnaire that included items about demographic characteristics of the mother, father, and child; obstetric history (eg, number of previous pregnancies, birth order of each child); antenatal care; diet, drug use, alcohol use, and smoking during pregnancy; feeding practices (eg, length of exclusive breast-feeding); birth spacing; childs use of health services (eg, immunizations, vitamin A, nutritional programs); and family religion and caste. Religion was classified as either Hindu or other. Caste is an important source of identity in Nepal. Castes were classified either as Dalits, who inhabit the lowest rung in the Hindu caste hierarchy, or All Others, including non-Hindus.

146 Asia-Pacific Journal of Public Health / Vol. 21, No. 2, April 2009

In addition, the questionnaire obtained information to standard of living using a standard of living index. This index is computed based on inventory of 30 household assets, such as toilet facilities, ownership of various durable goods, housing type. Each of the 39 assets are scored, with scores of 0 to 14 indicating a low, 15 to 24 indicating a medium, and 25 to 67 indicating a high standard of living.8 Anthropometry was conducted at the same time as the interviews, using the UNICEF electronic scale (SECA 890) to measure the weight of children. An infantometer was used to measure the height of the children up to age 24 months. A stature meter (a wooden rod with centimeter markings) was used for children more than 24 months of age. Three female enumerators who were at least high school graduates were hired to conduct the interview with mothers and to measure the height and weight of children. They were trained theoretically and practically on how to interview mothers and how to measure weights and heights. Underweight was defined as a weight for age less than 2 SDs from the National Centre for Health Statistics/World Health Organization reference median value9; similarly, stunting was defined as a height for age less than 2 SDs of the reference median value. In addition, the mothers weight and height was measured, and body mass index (BMI) was computed using the usual formula: BMI = weight (kg)/height2 (m). We conducted bivariate analyses to examine the association of potential risk or protective factors with the 2 outcomes (underweight and stunting) using 2 tests to determine the statistical significance of these associations. Multiple logistic regression was then used to assess the strength of risk/protective factors that were found to be important in the bivariate analyses, controlling for potential confounding. SPSS, version 15, was used for the statistical analyses. Ethical Considerations The consent form was read to the mother, and verbal consent was obtained for participation in the study. The ethics committee of Sree Chitra Tirunal Institute for Medical Sciences and Technology, Thiruvananthapuram, India, reviewed the protocol and gave approval for the study.

Results
Table 1 describes the demographic characteristics of the study group. Fifty-six percent of the study children were male. Approximately one fourth was 6 to 12 months of age at the time of the survey. The vast majority of families were Hindu (94%), and one third was of Dalit caste. Only 21% of families had a high standard of living score, whereas 44% had scores in the low range. Although almost half of the mothers married before the age of 18, 95% were 18 years or more at the time of the birth of the index child. Almost half of the mothers had less than 6 years of schooling, compared with 40% of the fathers. Only 5% of the mothers worked outside the home. Fathers were employed predominantly (86%) in agriculture or as laborers. Mothers BMI values ranged from 13.3 to 29.6; the median BMI of mothers was 19.1. Antenatal care was begun in the first trimester for approximately one third of pregnancies; however, almost 15% of pregnancies had no prenatal care visits (Table 2). More than a third of subjects (38%) were firstborn. Only 9% of the subjects were born within 24 months of the birth of the previous child. Among the mothers of subjects more than 24 months of age, 15% had another birth within 24 months of the birth of the index child. Approximately one third of the mothers reported eating more than usual during the index

Underweight and Stunting Among Children in Nepal / Singh et al 147

Table 1. Underweight and Stunting in Relation to Demographic Characteristics of Infants, Mothers and Fathers
n Percentage of Study Group Percentage Underweight 2 P Value Percentage Stunted 2 P Value NA NS

Total 443 100% 53.3 NA 36.6 Childs sex Male 249 56.2 52.3 NS 38.2 Female 194 43.8 54.6 34.5 Childs age 6-12 Months 105 23.7 35.2 <.001 22.9 13-24 Months 165 37.2 57.6 43.0 25-36 Months 173 39.1 60.1 38.7 Family ethnicity Dalits 145 32.7 62.8 .006 40.1 All others 298 67.3 48.7 32.6 Family religion Hindu 416 93.9 54.3 NS 36.5 Other 27 6.1 37.0 37.0 Standard of living score Low 194 43.8 63.4 <.001 45.4 Medium 155 35.0 51.0 31.6 High 94 21.2 36.2 26.6 Mothers age at delivery <18 Years 18 4.1 61.1 NS 33.3 18-29 Years 387 87.4 51.7 36.7 30+ Years 38 8.6 65.8 36.8 Mothers age at marriage <18 Years 209 47.2 58.4 <.05 38.3 18+ Years 234 52.8 48.7 35.5 Mothers education 0-5 Years 224 50.6 62.9 <.001 45.5 6+ Years 219 49.4 43.4 27.4 Mothers occupation Work at home 419 94.6 53.2 NS 35.8 Work outside home 24 5.4 54.2 50.0 Fathers age 15-24 Years 152 34.3 51.3 NS 38.2 25-34 Years 256 57.8 54.3 36.7 35+ Years 35 7.9 54.3 28.6 Fathers education 0-5 Years 176 39.7 63.1 .001 43.8 6+ Years 267 60.3 46.8 31.8 Fathers occupation Laborer or agriculture 380 85.8 55.3 .042 38.7 Clerical or nonmanual 63 14.2 41.3 23.8 Mothers BMI Below 18.5 (underweight) 169 38.1 62.7 <.001 40.8 18.5 and more 274 61.9 47.4 33.9 NOTES: NA = not applicable; NS = not specified; BMI = body mass index.

.003

.015

NS

.002

NS

NS <.001

NS

NS

.012

.024

NS

pregnancy, and 21% said that the index child appeared small at birth. More than 80% of the children had received 3 or more doses of polio and DPT vaccines. Among children more than 1 year of age, 56% had received 3 or more vitamin A treatments. Exclusive breastfeeding for less than 6 months occurred in 14% of the children more than the age of 1 year, and 31% were exclusively breast-fed for 9 months or more.

148 Asia-Pacific Journal of Public Health / Vol. 21, No. 2, April 2009

Table 2. Underweight and Stunting in Relation to Obstetric and Other Factors


n Percentage of Study Group Percentage Underweight 2 P Value Percentage Stunted 2 P Value NA .022

Total 443 100% 53.3 NA 36.6 When antenatal care began First trimester 142 32.1 45.8 .048 30.3 Second trimester 214 48.3 54.7 36.0 Third trimester or none 87 19.6 62.1 48.3 Number of antenatal care visits None 66 14.9 60.6 .026 51.5 1-3 178 40.2 58.4 37.1 4 or more 199 44.9 39.0 31.2 Birth order of index child First child 166 37.5 47.0 .016 30.7 Second or third child 214 48.3 53.7 36.9 Fourth or more 63 14.2 68.3 50.8 Age of previous sibling at birth of index child None or unknown 169 38.1 46.2 NS 30.8 Less than 24 months 39 8.8 53.8 46.2 24 months or older 235 53.0 58.3 39.1 Subsequent child born (if subject >24 months old) Within 24 months 25 14.5 64.0 NS 48.0 After 24 months 148 85.5 59.5 37.2 Diet during pregnancy Did not eat more than usual 292 65.9 57.5 .012 38.7 Ate more than usual 151 34.1 45.0 32.5 Babys size at birth Appeared small 94 21.2 68.1 .001 50.0 Did not appear small 349 78.8 49.3 33.0 Exclusive breast-feeding (children 12 months and more) 0-5 Months 48 14.2 58.3 .038 37.5 6-8 Months 187 55.3 53.5 39.0 9 Months or more 103 30.5 68.9 45.6 DPT/polio <3 Doses 75 16.9 62.7 NS 46.7 3 Doses or more 368 83.1 51.4 34.5 Vitamin A (subjects 12+ months) 0-2 Doses 147 43.6 66.7 .009 44.9 3 Doses or more 190 56.4 52.6 37.4 Nutritional program Yes 23 5.2 30.4 .024 26.1 No 420 94.8 54.5 37.1 NOTE: NS = not specified

.012

.019

NS

NS

NS

.002

NS

.046

NS

NS

More than 53% of the children were underweight and 36.6% were stunted, defined as less then 2 SDs between the reference median weight or height for age. Severe underweight or stunting (<3 SDs below the median for age) was found in 16.9% and 12.4% of the children, respectively. Several demographic characteristics were found to be significantly associated with being underweight: age greater than 12 months, being in the Dalit caste, low family standard of living score, maternal or paternal education of less than 6 years, mothers being married

Underweight and Stunting Among Children in Nepal / Singh et al 149

Table 3. Logistic Regression Model for Underweight


Childs age (continuous as ln[months]) Mothers BMI (continuous) Birth order (continuous) Poverty score (continuous) Mothers education (6 years or more) Vitamin A treatments (3+ vs <3) Participation in nutrition program (yes vs no) OR 110.97 0.94 1.48 1.06 0.96 0.65 0.92 95% CI for OR Lower 31.15 0.86 1.25 1.03 0.62 0.37 0.33 Upper 8.74 0.78 1.06 1.01 0.40 0.21 0.12 Significance <.001 .001 .010 .018 .032 .001 .033

NOTES: OR = odds ratio; CI = confidence interval; BMI = body mass index.

when less than 18 years of age, father working in agriculture or as a laborer, and low maternal BMI (Table 1). With the exception of mothers age at marriage and BMI, these same characteristics were also significantly associated with stunting. Earlier and more frequent antenatal care were associated with a lower likelihood of both underweight and stunting, as was being firstborn. Children whose mothers reported being small at birth were significantly more likely to be underweight and stunted, whereas mothers who reported eating more than usual during pregnancy were less likely to have an underweight child. Infants who were exclusively breast-fed either less than 6 months or more than 8 months were more likely to be underweight. Both vitamin A supplementation and participation in a nutritional assistance program were associated with a lower likelihood of being underweight. Birth intervals, either before or after the birth of the index child, were not related to either underweight or stunting in this study group. Because it was likely that many of the factors found to be associated with either underweight or stunting were interrelated, we used multivariate logistic regression to simultaneously adjust for their potentially confounding effects. In creating the logistic regression modes, we excluded (a) variables that were not associated with the outcomes in bivariate analyses (eg, infant sex), (b) variables that were not associated with the outcomes after adjusting for standard of living (eg, ethnicity/caste, fathers education, and others), or (c) variables that were not logically causal (eg, immunization). We also did not include the variables describing mothers assessment of infants size at birth or her recollection the amount of food eaten during pregnancy, as these were likely to be inaccurate due to recall bias. Therefore, 7 variables were included in the multiple logistic regression analyses: childs age (entered as the natural log of the age in months), mothers education (0-5 vs 6+ years), vitamin A treatment (3+ vs <3), nutritional program participation (yes vs no), birth order (entered as a continuous variable), mothers BMI (entered as a continuous variable), and poverty index (standard of living score with order reversed so that high scores indicate greater poverty). The resulting models are shown in Table 3. In addition to age of child, all 6 variables in the logistic model were significantly related to a child being underweight. Poverty (odds ratio [OR] = 1.03 for each point higher in the poverty index) and increasing birth order (OR = 1.25 for each increment in birth order) were highly significant risk factors for underweight, whereas mothers BMI (OR = 0.86 for each unit of BMI), mothers education (OR = 0.62 for 6+ years of schooling), vitamin A treatments (OR = 0.37 for 3 or more treatments), and participation in a nutritional program (OR = 0.34) were strong protective factors. In the logistic model for stunting, the childs age was a strong risk factor, whereas mothers BMI (OR = 0.90) and mothers education (OR = 0.57) were significantly protective (Table 4).

150 Asia-Pacific Journal of Public Health / Vol. 21, No. 2, April 2009

Table 4. Logistic Regression Model for Stunting


Childs age (continuous as ln[months]) Mothers BMI (continuous) Birth order (continuous) Poverty score (continuous) Mothers education (6 years or more) Vitamin A treatments (3+ vs <3) Participation in nutrition program (yes vs no) OR 8.80 0.898 1.11 1.03 0.57 0.637 0.67 95% CI for OR Lower 2.62 0.82 0.95 1.00 0.37 0.38 0.123 Upper 29.52 0.98 1.30 1.05 0.89 1.08 1.82 Significance <.001 .019 .176 .048 .013 .091 .428

NOTES: OR = odds ratio; CI = confidence interval; BMI = body mass index.

Discussion
This study found very high rates of both underweight and stunting; these rates were similar to, but higher than, those that have been measured for Nepal as a whole.2 We found that biologic, socioeconomic, and health care factors were all associated with underweight or stunting. Biologic Factors Similar to earlier reports,2,8,10-12 children more than 12 months of age were more likely to be underweight and stunted than younger children. It is likely that nursing during early life is protective and that undernutrition becomes more likely as the child becomes more dependent for caloric intake on foods that have to be grown or bought. Interestingly, although either short (<6 months) or long (>9 months) duration of exclusive breast-feeding was found to be weakly associated with underweight in bivariate analyses, we found that duration of breast-feeding was not significantly related to underweight after adjusting for other factors. Birth order of the baby had an effect on underweight, as shown in previous studies.2,8,13,14 The higher the birth order, the higher the prevalence of underweight and stunting. Repeated pregnancies drain the mothers health and also further impoverish the family. A higher birth order and more surviving children reflect repeated pregnancies and possibly less care for the individual children. The BMI of the mother was strongly related to underweight in the bivariate analyses and to both underweight and stunting in the multiple logistic regression analyses, suggesting perhaps that the nutritional status of the mother, especially when she is pregnant, may affect the subsequent growth of her children. It is also possible, however, that unmeasured factors (in addition to the familys standard of living, which was controlled for in the multivariate analyses) may be causally related to both maternal and child nutritional status. We found no relationship, however, between various measures of birth spacing and either underweight or stunting. Nor was the nutritional status of children in this study significantly different according to gender. Other studies in Nepal and India2,8,10 show that the prevalence of malnutrition was similar in both the genders in this age group. Socioeconomic Factors As found in many previous reports,15-17 this study demonstrates the significant relationship between socioeconomic status and child malnutrition. The standard of living score was

Underweight and Stunting Among Children in Nepal / Singh et al 151

strongly related to underweight and stunting in both the bivariate and the multiple regression analyses. Children from Dalit families were found to have significantly higher malnutrition rates when compared with other children. Sah,10 in his study in Dhanusha, did not find a significant relationship between ethnicity and child malnutrition, but the prevalence was higher for Dalits than non-Dalits. The National Family Health Survey, India,8 revealed that the children belonging to schedule castes, tribal groups, and other backward classes were found to have relatively higher prevalence of child malnutrition. It is important to point out, however, that the effect of caste was minimal and not statistically significant when standard of living was controlled for, suggesting that the major nutritional disadvantage of lower caste status is due to poverty. Like many previous studies,2,12,13,18-22 this study found that higher educational attainment of both mothers and fathers were strongly protective in relation to underweight and stunting. Fathers education, however, was not significant after controlling for standard of living. It is important to emphasize that mothers educational level was strongly protective for both underweight and stunting, even after controlling for other important variables in the multiple regression analyses. This effect may be due to the fact that an educated mother has more opportunities to be informed of, and be aware, of health care, better nutrition, and child development when compared with uneducated mothers. It is noteworthy that the use of vitamin A treatments (OR = 0.37) and participation in a nutritional program (OR = 0.33) was strongly protective for underweight, even after controlling for other important factors such as mothers education, standard of living, and childs age. This finding suggests that health care programs can and do make a difference in the nutritional status of poor, rural children. This study was restricted to a single district in eastern Nepal and may not be generalizable to other populations. Data on risk and protective factors were mostly obtained from questionnaires, and it is not known how much recall bias or other threats to validity may affect the accuracy of these data. For this reason, we did not use mothers assessment of the amount she ate during pregnancy or the size of the child at birth in the logistic regression models. Finally, the statistical analysis did not take the cluster sampling methodology into account when calculating P values and confidence intervals. If we had adjusted the analyses for clustering, it is likely that some of the weaker results would not be significant at the .05 level. These results suggest that underweight and stunting are the result of a nexus of biological, socioeconomic, and health care factors. In addition to programs that offer direct nutritional support or supplementation, efforts to improve the nutritional status of poor children should consider interventions that can improve the overall health of mothers; advance socioeconomic status, especially income and maternal education; and limit the number of pregnancies. Policies and strategies have to be long term and strategic at the individual, community, and country levels:
Improving the status of women through primary and secondary education, including nonformal skill training to increase economic independence. Providing health information packages aimed at popularizing better locally available nutritious food, supplementary feeding programs for children aged less than 5 years through a network of child care centers. Supporting behavior change interventions designed to address locally relevant child care, maternal care, and hygiene practices developed with an understanding of the local culture and practices.

152 Asia-Pacific Journal of Public Health / Vol. 21, No. 2, April 2009

References
1. World Health Organization. World Health Report: shaping the future. http://www.who.int/ whr/2003/en/whr03_en.pdf. Published 2003. Accessed April 7, 2005. 2. Ministry of Health Nepal, New Era, and ORC Macro. Nepal Demographic and Health Survey 2001. Calverton, MD: Family Health Division, Ministry of Health, New Era, and OCR Macro. 3. World Health Organization. Fifty-Third World Health Assembly: infant and young child nutrition. ftp.who.int/gb/archive/pdf_files/WHA53/ea7.pdf. Published 2000. Accessed April 8, 2005. 4. de Onis M, Monteiro C, Akr J, Clugston G. The worldwide magnitude of protein-energy malnutrition: an overview from the WHO Global Database on Child Growth. Bull World Health Organ. 1995;71:703-712. 5. United Nations Administrative Committee on Coordination (ACC)/Sub-Committee on Nutrition (SCN). Fourth Report on the World Nutrition Situation. Geneva, Switzerland: ACC/SCN. 6. de Onis M, Blossner M, Borghi E, Morris R, Frongillo EA. Methodology for estimating regional and global trends of child malnutrition. Int J Epidemiol. 2004;33:1260-1270. 7. United Nations. Millennium Development Goals. http://www.un.org/millenniumgoals. Accessed April 7, 2005. 8. International Institute for Population Sciences (IIPS) and ORC Macro. National Family Health Survey (NFHS-2), 1998-99. Mumbai, India: IIPS; 2000. 9. World Health Organization. Physical status: the use and interpretation of anthropometry. Technical Report Series No 854. Geneva, Switzerland: World Health Organization; 1995. 10. Sah N. Determinants of child malnutrition in Nepal: a case analysis from Dhanusha, central Terai of Nepal. http://iussp2005.princeton.edu/download.aspx?submissionId=51628. Published 2005. Accessed September 25, 2005. 11. Yasoda Devi P, Geervani P. Determinants of nutrition status of rural preschool children in Andhra Pradesh. http://www.unu.edu/unupress/food/8F154e/8F154E0c.htm. Accessed March 23, 2005. 12. Girma W, Genebo T. Determinants of nutritional status of women and children in Ethiopia. http:// www.measuredhs.com/pubs/pdf/FA39/02-nutrition.pdf. Published November 2002. Accessed April 16, 2005. 13. Madise NJ, Mpoma M. Child malnutrition and feeding practices in Malawi. http://www.unu.edu/ unupress/food/V182e/ch13.htm. Accessed March 12, 2005. 14. Melville B, Williams M, Francis V, Lawrence O, Collins L. Determinants of child malnutrition in Jamaica. http://www.unu.edu/unupress/food/8F101e/8F101E08.htm. Published 1998. Accessed March 7, 2005. 15. Setboonsarng S. Child malnutrition as a poverty indicator: an evaluation in the context of different development interventions in Indonesia. http://www.adbi.org/files/2005.01.14.dp21.malnutrition.poverty.indonesia.pdf. Published January 2005. Accessed March 13, 2005. 16. Haddad L, Alderman H, Appleton S, Song L, Yohannes Y. Reducing child malnutrition: how far does income growth take us? World Bank Econ Rev. 2003;17:1107-1131. 17. Levinson FJ, Mehra S, Levinson D, et al. Morinda revisited: changes in nutritional well-being and gender differences after 30 years of rapid economic growth in rural Punjab, India. Food Nutr Bull. 2004;25:221-227. 18. Smith LC, Haddad L. Overcoming child malnutrition in developing countries: past achievements and future choices. Food, Agriculture and the Environment Discussion Paper 30. Washington, DC: International Food Policy Research Institute. http://www.ifpri.org/2020/dp/2020dp30.pdf. Published February 2000. Accessed March 13, 2005. 19. Borooah VK. The role of maternal literacy in reducing the risk of child malnutrition in India. University of Ulster and ICER. http://ideas.repec.org/p/icr/wpicer/31-2002.html. Published June 2002. Accessed March 14, 2005. 20. Waters H, Saadah F, Surbakti S, Heywood P. Weight-for-age malnutrition in Indonesian children, 1992-1999. Int J Epidemiol. 2004;33:589-595. 21. Kikafunda JK, Walker AF, Collett D, Tumwine JK. Risk factors for early childhood malnutrition in Uganda. Pediatrics. 1998;102:E45. 22. Chamarbagwala R, Ranger M, Waddington H, White H. The determinants of child health and nutrition: a meta-analysis. http://www.wam.umd.edu/~ranger/home/papers/meta-analysis.pdf. Accessed April 21, 2005.