Contents - Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
The findings from this evaluation highlight the poor nutrition status and diets of rural mothers in Bangladesh. The adequacy of women's diets did not improve after receiving education. Whereas nutrition education may have prevented mothers from suffering the sharp decline in dietary adequacy noted in the controls during the monsoon season, severe flooding occurred in both areas between the first and second post-education follow-ups, almost certainly influencing the final results. No significant differences were seen between the energy intakes of the treatment and control mothers at Post1, which was immediately after the education, when a difference would have been expected.
TABLE 6. Energy adequacy of breast-fed children's diets with and without breast-milk estimates.
(n = 21)
(n = 28)
|Complementary food (% energy req.)|
|Breast milk and food (% energy req.)|
Ideal energy requirement based on international reference median weight for child's age and sex. Breast-milk quantities at each age estimated from research on a comparable Bangladeshi sample [8, 9]. No significant differences were found in energy intakes between treatment and control children.
By contrast, an intervention to improve the diets of weaning-age children conducted the same year in the same site resulted in a significant increase in energy intakes and greater weight gains in the treatment children than in controls [29, 31]. As figure 4 (see FIG. 4. Energy requirements, intakes, and energy gaps of lactating Bangladeshi women compared with weaning-age children. Reported figures are averages for 78 women over 8 months and 61 control children age 4 to 17 months over 5 months during 1987 [29-31]. Data for children age 18 to 24 months were obtained from breast-fed children in control site in 1986. Energy requirements for children are based on growth at 1.5 SD below the median weight of NCHS/WHO reference for child's age and sex. One hundred per cent of children were breast-feeding, and food consumed includes estimates of breast-milk intake based on best country data available [8, 9].) illustrates, the weaning-age children's energy requirement and absolute energy gap are considerably less than those of lactating mothers; thus nutrition education for children confronts fewer economic barriers to compliance [30, 31]. As the children get older, the absolute differences between women's and children's energy gaps narrow somewhat. However, the widening energy gap stabilizes after age 18 to 24 months when the children begin consuming greater quantities of the adult diet. Thus, the additional food required for women is still considerably greater. The limited measurable impact of carefully developed education messages on women's dietary practices, combined with the concerns raised during the implementation phase, suggest the need to evaluate carefully the significant socio-economic and cultural barriers that preclude better nutrition for women.
Economic and cultural constraints
"When we have the food, we eat it; when we don't, we don't!" Despite the encouragement of low-cost foods in the messages, this was a common response of the mothers when asked if they were able to eat the amounts of foods suggested in the messages. With over 80% of the diet energy from grain, the women's diets were already very low cost [30, 31], with little animal food, and little room for substitution of other lower-cost foods to increase energy intake. To fill the lactating mother's energy gap would cost an additional 21% of the local daily wage, or almost double the cost of the food she was consuming [30, 31]. In addition, women knew that to increase their own share of food, cultural norms required that they would also have to increase the amount given to others in the family. Directing education to husbands, mothers-in-law, and other women in the household increased family support for the lactating women, but could not overcome the financial barriers to improving their diets. In contrast, weaning age children's energy gap could theoretically be closed for one-third of the cost of improving the mother's diets, or 8% of the daily wage. Further details on the financial implications of the messages for families are presented elsewhere [30, 31].
Most of the husbands in this study were very thin themselves, suggesting they were also receiving less than their required level of energy and other nutrients. Of a sampling of households interviewed, weekly food expenditures were stated to be greater than weekly incomes. When asked how this was possible, most families revealed they were constantly taking out loans from neighbours and, if possible, from banks [15, 31]. Husbands could agree to give more food to their wives only when they had it.
Some deep-rooted cultural practices of intrahousehold food allocation may be amenable to change in the long term through education. For example, although snacking was generally infrequent among the women in this study and others in Bangladesh , snacks did not carry the same taboos as meals. Therefore, messages to encourage more snacks, rather than more food at meals, may confront fewer cultural barriers to increasing women's food intake in Bangladesh where women are taught to be self-sacrificing and eat last at meals. Nevertheless, the sociocultural window of opportunity potentially offered by a snack-approach to education is likely to be closed by the overriding economic constraints to snacks, which are the same as for meal foods. The severe food insecurity of many households may prevent even a small reallocation of food to women in Bangladesh, and other settings where families are similarly destitute.
Decline in energy adequacy
The decline in energy intake after education was probably a seasonal effect. The baseline measure was taken during a more plentiful, cooler season. The first post-education measurement, in May-July, spanned the hot, rainy period that is often accompanied by swelling rivers and flooding, and lower food intakes [50, 51]. The final measure in October-November characterized a post-monsoon, pre-harvest season when the nutrition status of women in other studies was lowest [6, 52]. The few reliable longitudinal studies measuring women's dietary intake alluded to similar seasonal trends, yet also suggested year-to-year monthly variations [50, 51]. Adding to the difficulties, the final measurement was taken two to three months after the severe floods of August and September that affected food production, distribution, and consumption. Families were still struggling to repair damages to their homes and fields.
The significant association of the average energy adequacy of the mothers' diets with the growth (WAZ and MUAC) of their breast-fed infants suggests that if women's diets can be improved, the growth of their breast-fed infants may also improve. The education campaign may also have been successful at preventing severe malnutrition.
Stable arm circumference
The women's stable arm circumference, in the light of the declining energy intake, suggests their reduced activity during the hot, rainy season, and/or that they produced less breast milk. The reduction in breast-milk production may have been slightly greater for the controls if this explanation is true, although the decline was not detectable in the growth (change in WAZ or MUAC) of their breast-feeding children.
Alternatively, changes in arm circumference may not be a sensitive indicator in these women to changes in dietary adequacy and its functional effects on nutrition status. The average maternal MUAC (both treatment and control) was near the fifth percentile of the reference population , highlighting the lean state of the lactating mothers. Estimates of body mass index from a study the previous year also imply that most of these women had very low fat reserves (average BMI 18 kg/m2), with 13% of women meeting the classification used in developing countries of anorexic, which is defined here as BMI less than 16 kg/m2 .
Although it would have been desirable to have more measures of mothers' nutrition status, for example, weights, heights, and triceps skinfold thickness, the intervention was designed with one simple assessment measure. The MUAC was selected since measurement tapes were readily available in the country at very little cost and could be easily carried with the workers. Our find that women's MUAC did not decline despite a decline in energy intake suggests more research should be done on the use of this indicator for rapid assessment and monitoring in lean women. In addition, despite care taken in standardizing measurements, errors cannot be ruled out.
Other studies in Bangladesh have suggested that the nutrition status of women, measured by weight or MUAC, does not decline over the lactation period [10, 52, 53]. Women who breastfeed their children from 1.5 to 3 years appear to maintain body weight , suggesting that this may be due to metabolic adaptation caused by lactation, or the probable low quantities of milk produced. In the first year postpartum, women's arm circumference did not decline; however, their weights did, suggesting either differences in change of body composition or that MUAC may not be as sensitive as weight to changes in nutrition status . In addition, seasonal differences were seen, with the leanest period being August-October, and the postpartum deterioration in nutrition status was thought probably to be due to seasonal factors.
Very low energy intakes
The low energy intakes of the women studied longitudinally the previous year in the same site  were consistent with this study, averaging 1,259 kcal/day (5,422 kJ) and 1,354 kcal/day (5,669 kJ), respectively. Other published studies on impoverished women, including lactating and pregnant women, in Asia, Africa, and Latin America recorded similarly low energy intakes [55-57]. For the purposes of this study, the relative differences in energy intake between controls and treatments are more important than the actual measured values. However, this large deficiency in energy intake warrants further comment on the appropriateness of the requirement and the accuracy of the diet recalls.
First, it is possible that the additional energy requirement for lactation overestimated actual needs because of the women's low milk production. It is apparent from the deterioration in nutrition status of their breast-fed infants that the mothers were not producing sufficient amounts of milk to support normal child growth, and thus may have actually required less than the lactation adjustment of 500 kcal (2,094 kJ) to maintain their own nutrition status.
Second, at such a deficient level of energy intake, the mothers were probably not as active as they would otherwise be, and as was estimated in deducing the factor for energy expenditure. Activity reduction during lactation has been observed in healthy women .
Third, the FAO/WHO/UNU predictive equations used in this study to derive BMR may slightly overestimate requirements for people in tropical environments [59-62], owing to a combination of factors such as climate, diet, ethnic background, and body configuration [59, 63-65]. However, this discrepancy would account for only about 60 to 100 kcal, putting the average energy adequacy of the women's diets at about 61% of recommended levels. The WHO metabolic constants may also overestimate actual energy expenditure in small individuals . Indian women were able to perform 15 work tasks at a lower cost than men primarily because they are smaller , although the researchers' hypothesized social processes also enabled these seemingly undernourished women to maintain a high level of productivity. One questions whether these women with consistently low energy intakes may represent a metabolically more efficient population, similar to healthy women described as "small eaters" who have similar levels of activity and fat-free mass as "big eaters" .
Fourth, underestimation of actual food consumption by the 24-hour dietary recall cannot be ruled out. Indeed, measuring dietary intake accurately under field conditions is extremely difficult, and both under- and over-estimations are potential errors . Careful measures were taken to obtain accurate estimates of intakes. Field workers were experienced and well trained in probing for additional foods and quantities. As natives of the study village, they had close rapport with the women. The dietary recall instrument had been pre-tested and validated by observation in rural Bangladesh [36, 37], finding only small average discrepancies (both over- and under-reporting) of less than 50 kcal (209 kJ) between observed and reported intakes. This may be due, in part, to the monotony and simplicity of the traditional Bangladesh diet.
If the recall underestimated actual energy intakes, or if the ideal requirement overestimated actual needs, we assume the measurement error is consistent for both treatment and controls and thus would probably not affect interpretation of the study conclusions.
The extent to which maternal intake of specific nutrients influences milk volume is unclear . Inadequate consumption of specific nutrients affects maternal health, however, and the quality of the women's diet warrants comment.
The strikingly low levels of fat (5% of dietary energy) are consistent with other studies on lactating women in Bangladesh . In fact, in a sample of urban Bangladesh women consuming 1.7% to 7.8% of their dietary energy from fat, fat intakes reported by dietary recall were greater than the levels actually observed . This low intake may limit absorption of fat-soluble vitamins.
As noted, protein and iron intakes paralleled energy intakes, and thus comparisons between treatments and controls were not reported. In general, since the women's energy intake was extremely limited, it is difficult to interpret the functional implications of the low protein intake, because protein will be used to meet energy needs under such circumstances . However, even lactating women with energy-adequate diets who consume less than I g protein per kilogram body weight may be in negative nitrogen balance , and the women in our sample were consuming well under this (0.6 g/kg adjusted for low protein quality).
Our data suggest the women consumed sufficient dietary iron using the requirement of 15 ma. However, iron-deficiency anaemia is prevalent in Bangladeshi women of child-bearing age [6, 52]. Although this requirement takes the low biological availability of dietary iron into consideration , increased iron demands in these women from frequent pregnancies and infections (e.g., hookworm), combined with the poor bioavailability of dietary iron, may increase their requirement. Using a higher daily requirement of 26 to 56 mg recommended by some for (non-pregnant) women consuming a diet with low iron bioavailability , the women in our study consumed only 31% to 67% of their iron requirement (excluding iron from tube well water and other non-food sources).
Whereas the low intakes of vitamin A are consistent with findings of other studies [11, 7476], they should be interpreted with caution owing to the lack of complete food analysis data  and possible deficient intakes of other nutrients (fat, protein, alphatocopherol, iron, zinc) that may affect vitamin A transport, storage and use . With this limitation in mind, we found some improvements in the vitamin A content of treatment women's diets relative to controls immediately after education (Post!). This increase could be a programme effect, although differences between the groups cannot be ruled out, since more treatment families owned small plots of land on which they potentially could grow vegetables. However, in a study the previous year in the same site, agricultural wealth was not significantly associated with greater consumption of dark leafy greens . No significant differences between the groups in our study were found at the final measure; however, the floods in August and September destroyed many of the home gardens, reducing the supply of greens in October and November.
Women's formal education
The link between maternal education and child health was highlighted in a number of countries, including Bangladesh, and, to a lesser extent, the link with education and maternal health was described [27, 78, 79-88]. We found that a woman's exposure to formal education, even after controlling for wealth, was significantly associated with the energy adequacy of her diet. In another study in the same site, mother's education level was also predictive of the children's nutrition status and weight gain [29, 31]. Significant positive differences were seen consistently in child feeding and hygiene practices of Bangladeshi women with any education, compared with women with no education, although some negative practices such as breast-feeding less frequently and terminating feedings more often were noted . The specifics of the relationships between education and health have to be teased out , but the role of basic education as a pathway to better maternal and child health must be recognized.
The results of the lactation intervention suggest Bangladeshi mothers were not able to comply with nutrition messages designed by the community to improve their own diets. The study spanned the hot, dry and monsoon seasons, when hunger is greatest, in a period compounded by unusually severe flooding. Whereas families liked to receive information that might keep their children healthy, and the women generally wanted to eat more food, they felt they had little means to do so. Each breast-feeding mother knew that if she increased her own share of the food, she would have to increase the amounts given to all family members. Enlisting the support of her husband and others for the idea that she should eat more at meals did not make her believe that she could take more than she served to them.
The findings raise concerns about the effectiveness of nutrition education for improving women's diets in an impoverished area such as Bangladesh. The financial barriers to compliance lead us to conclude that nutrition education campaigns should not be conducted in isolation of supports that make these improvements affordable to women. Examples of such programmes might include women's saving or credit groups, income-generating projects, agricultural production, home gardens, and subsidized or donated food programmes.
The evaluation of this project was supported by the Social Sector Policy Analysis Project operated by the Academy for Educational Development, for the Bureau for Program and Policy Coordination, Office of Policy Development, and Program of Review of the US Agency for International Development. The field project was funded by the Asia and Near East Bureau of AID through the US Office of International Health under Public Health service contract.
The authors thank Mustaque Chowdhury, Fazle Abed, Azmat Ahmad, Suaib Ahmed, Sobita Datta, Halida Khondaker, the village workers and volunteers, and other BRAC staff without whom this project would not have been possible. The many individuals who provided technical, editorial, and administrative assistance and advocacy included Judy Brace, Melanie Marlett, Nancy Pielemeier, Helen Armstrong, Michael Levitan, Linda Vogel, Collette Chabbott, Mary Lee Mclntrye, and Karen Nurick.
Contents - Previous - Next