Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Differential infant and child mortality rates in Bangladesh

M. Kabir and M. Mohiuddin Ahamed
Jahangirnagar University, Dhaka, Bangladesh

M. Moslehuddin
University of Dhaka, Bangladesh

Infant mortality has traditionally been viewed as an indicator of the social and economic well-being of a society. It reflects not only the magnitude of those health problems which are directly responsible for the death of infants but the effect of a multitude of other factors, including prenatal and post-natal care of mother and infant, and the environmental conditions to which the infant is exposed.

The purpose of this study is to estimate the general level of infant and child mortality in Bangladesh and to show how this varies between socio-economic characteristics and the geographic areas of the country.


All the data used in this study were derived from the 1983 CPS report [3]; for details of the methodology and data collection see that publication The estimates of infant and child mortality presented have been obtained by the Trussell technique [6].


Demographic surveys are an invaluable source of information on infant and child mortality in developing countries. Table 1 shows the proportion of children who die, according to the age of their mothers at the time of birth. This proportion is by itself an indicator of childhood mortality, though it is affected by factors other than mortality (such as the age pattern of fertility) and cannot, therefore, be used directly as a mortality measure. Estimates of infant and child mortality derived from the Trussell technique are also shown in table 1. The overall infant mortality as implied by the q2 in the West model life table was about 150 per thousand births. The infant mortality indicated by the average of q2, q3, and qs was in the range of 145 to 155 per thousand births.


Differentials in mortality levels documented between urban and rural regions and populations are associated with variations in socio-economic status and such family characteristics as education, literacy, income, and housing conditions.

Many of the determinants of child mortality are household properties, including levels of household income, adult literacy, health practices among members, sanitary facilities, and so on. Two central issues in mortality analysis can be effectively addressed: the role of paternal and, particularly, maternal education in reducing infant and child mortality.

TABLE 1. Estimation of infant and childhood mortality by Trussell technique from 1983 CPS.

Age of women i Proportion of children dead (Di) Average parity (Pi) Multiplying factors (Ki)a Probability of dying (qX=KiDi) Implied mortality level in the West model life table Implied infant mortality in the West model life table
15-19 1 0.196 0.580 0.842 0.165 10.6 166
20-24 2 0.197 2.125 0.958 0.189 11.6 150
25-29 3 0.207 3.703 0.963 0.199 11.9 145
30-34 4 0.230 5.269 0.991 0.228 11.3 155
35-39 5 0.249 6.350 1.015 0.253 10.3 156
40-44 6 0.282 6.946 1.005 0.283 10.3 171
45-49 7 0.268 7.343 0.996 0.267 12.0 144

a. Estimated on the basis of regression equation (West model life table), Ki = Ai + Bi(P1/P2) + Ci(P2/P3).

Caldwell [2] and Preston [4] have argued that advances in female education may represent a potent and cost effective means of reducing child mortality. Cochrance [5] confirms this view by reviewing a large number of studies. Table 2 presents cumulative probabilities of dying according to the mother's years of schooling. The evidence is that the higher the level of education, the lower the estimated child mortality; that is, child mortality invariably declines as maternal education increases. For all ages, the ratio of cumulative childhood mortality for those whose mothers have no schooling to those whose mothers have ten years of schooling is almost 3:1. On the other hand, the ratio of implied infant mortality for those whose mothers have no schooling to those whose mothers have 10 years and more is more than 2:1 (table 2). Both infant and child mortality declined as the mother's education increased. Table 2 also shows the differential effects of the mother's and father's education on infant and child mortality. It is striking that improvement in infant and child mortality associated with the increased years of parents' schooling is much greater for the mother's than for the father's. Since the father's education would be more closely related than the mother's to household income, the result suggests that education is working principally through other ways related to child-care and health practices.

As expected, children whose parents owned land have lower infant and child mortality than those whose parents owned none. Similarly, differentials by two religious groups suggest that both infant and child mortality were considerably higher for the children of Muslim mothers than for non-Muslims.

Perhaps the most interesting result presented in table 2 is that the urban-rural differential of infant and child mortality is not large. For instance, both infant and child mortality in the rural area is 16 per cent higher than in the urban areas. The lower mortality in the urban area can be explained by the availability and accessibility of health care services. This also suggests that environmental hazards in urban areas are greater than in the rural areas, and that only the most privileged group can overcome them.

Contrary to our expectations, children of working women had higher infant and child mortality than those of the non-working women. One explanation may be that children of working women are deprived of adequate child care and that women's work is an indicator of the low economic condition of the household. Among the working women who are employed for wages, the children's mortality is lower than that of children of the women who are employed without cash wages. This may be due to the fact that those who are employed without cash are generally from a low socio-economic stratum with little or no education. It should be noted that the effect of a mother's work on infant and child mortality depends on the structure of the family, particularly on the availability of other family members to take care of the children in their mother's absence.

If maternal deprivation causes differentials in infant and child mortality, then the effect of the mother's absence should be greater where the mother is educated. Much of the effect of a mother's work on infant mortality in the first year following childbirth can be attributed to the very low income among the husbands of working women. There was also a significant direct effect of work through the shortening of breast-feeding time by working mothers.

Table 2 also shows the infant and child mortality by contraceptive use status. The influence of the mother's age on the survival of newborn babies is among the most frequently studied factors, since it is known that a late age of childbearing represents a grave risk factor for the newborn child. Birth interval also appears to have a role in the survival of newborn infants. For instance, for those mothers who had at some time used contraception, the infant and child mortality was more than 30 per cent lower than for the children of women who had never used contraception. The differentials in infant and child mortality are even higher when current users and non-users are compared. For women who breast-feed their children, an infant death tends to curtail the postpartum infecundity following a birth and thus, in the absence of contraception, reduces the interval between births. Short birth intervals cause high infant mortality.


There is relatively little information regarding mortality differentials by socio-economic characteristics in Bangladesh. Part of the problem results from the absence of data and part from failure to exploit data that already exist. The 1983 CPS has given the possibility of utilizing the data by background characteristics. Although infant and child mortality estimates derived from the survey data cannot be expected to be as precise as those obtained from the traditional vital registrations system, the 1983 CPS has yielded information determining levels and trends in infant and child mortality that seems to be reasonable.

The availability of more reliable information and the recent efforts to expand mortality investigation have made it apparent that infant and child mortality decline in Bangladesh has not progressed as expected. Differentials in infant and child mortality levels that have been found between urban and rural populations are associated not only with variation in the availability of and access to health services, but also with variation in socio-economic status and such family characteristics as education, literacy, income, and housing conditions.

TABLE 2. Infant and child mortality differentials by selected background characteristics


Number of deaths on which estimates are based

Estimated child mortalitya

Implied mortality level in the West model life table

Implied infant mortality in the West model life table

Education of mother
No schooling
q2 561 0.195 11.2 0.156
q3 945 0.216 11.1 0.156
q5 1,072 0.243 10.8 0.162
P1P2 = 0.378, P2/P3 = 0.604
1-5 years of schooling
q2 180 0.148 13.7 0.119
q3 305 0.183 12.4 0.138
q5 260 0.203 12.5 0.136
P1P2 = 0.383, P2/P3 = 0.577
6-9 years of schooling
q2 24 0.114 15.7 0.094
q3 33 0.127 15.5 0.096
q5 34 0.139 15.5 0.096
P1P2 = 0.494, P2/P3 = 0.546
10 years and above schooling
q2 9 0.066 19.0 0.057
q3 6 0.074 18.7 0.060
q5 3 0.081 18.7 0.060
P1P2= 0.358, P2/P3 =0.631
Education of husband
No schooling
q2 409 0.197 11.7 0.158
q3 706 0.223 10.8 0.162
q5 126 0.250 10.5 0.167
P1P2 = 0.386, P2/P3 = 0.584
1-5 years of schooling
q2 146 0.165 12.8 0.132
q3 265 0.210 11.4 0.153
q5 307 0.225 11.6 0.150
P1P2= 0.349, P2/P3= 0.641
6-9 years of schooling
q2 93 0.171 12.5 0.136
q3 149 0.188 12 4 0.138
q5 143 0.205 12.4 0.138
P1P2 = 0.382, P2/P3 = 0.542
10 years and above schooling
q2 76 0.120 15.4 0.098
q3 95 0.133 15.2 0.100
q5 89 0.146 15.1 0.101
P1P2 = 0.447, P2/P3 = 0.621
Place of residence
q2 718 0.192 11.4 0.153
q3 1,171 0.202 11.7 0.148
q5 1,259 0,230 11,4 0.153
P1P2 = 0.276, P2/P3 = 0.585
q2 170 0.163 12.9 0.131
q3 346 0.175 13.0 0.129
q5 319 0.202 12.5 0.136
P1P2 = 0.263, P2/P3 = 0.498
q2 705 0.177 12.2 0.141
q3 1,176 0.199 11.9 0.145
q5 1,216 0.235 11.1 0.158
P1P2 = 0.384, P2/P3 = 0.595
q2 72 0.163 12.9 0.131
q3 115 0.196 12.0 0.144
q5 149 0.207 12.3 0.139
P1P2 = 0.418, P2/P3 = 0.604
Landholding status
Owned land
q2 514 0.172 12.4 0.138
q3 782 0.185 12.5 0.136
q5 892 0.228 11.4 0.153
P1P2 = 0.369, P2/P3 = 0.607
Owned no land
q2 263 0.182 11.9 0.145
q3 508 0.226 10.6 0.166
q5 477 0.240 10.9 0.161
P1P2 = 0.427, P1/P3 = 0.571
Employment status
Employed with cash
q2 66 0.199 11.0 0.159
q3 137 0.219 10.9 0.161
q5 120 0.238 11.0 0.159
P1P2 = 0.386, P2/P3 = 0.590
Employed without cash
q2 29 0.218 10.1 0.174
q3 22 0.239 10.1 0.174
q5 36 0.259 10.2 0.172
P1P2 = 0.393. P2/P1 = 0.551
Not employed  
q2 682 0.180 12.0 0.144
q3 1,132 0.199 11.9 0.145
q5 1,211 0.217 11.9 0.145
P1P2 = 0.386, P2/P3 = 0.594
Contraceptive use status
Used at some time
q2 258 0.144 14.0 0.115
q3 455 0.154 13.8 0.118
q5 498 0.194 12.9 0.131
P1P2 = 0.457, P2/P3 = 0.639
Never used  
q2 519 0.193 11.7 0.148
q3 836 0.230 10.5 0.167
q5 873 0.263 10.0 0.175
P1P2 = 0.379, P2/P3 = 0.575
Currently using
q2 108 0.120 15.4 0.098
q3 227 0.150 14.2 0.113
q5 271 0.180 13.2 0.126
P1P2 = 0.488, P2/P3 = 0.650
Currently not useing
q2 661 0.187 11.7 0.148
q3 1,064 0.213 11.2 0.156
q5 1,100 0.248 10.6 0.166
P1P2 = 0.382, P2/P3 = 0.585

a. Estimated on the basis of the Trussell variant (West mortality pattern).

Even in the urban areas enjoying better health facilities, large differentials in mortality exist between the low-income, illiterate, inadequately housed majority and those who are better-off [8] . In many urban areas of Bangladesh, the conditions of sanitation and hygiene for the poor are worse, or at least no better, than in the rural areas.

The mother's attainment of primary schooling has a significant impact on child survival. The father's education also has a bearing on mortality during early childhood but the mother's education emerged as an important determinant of child survival in Bangladesh. In general, the effects of parental education may be greater than those of income and access to health facilities combined. It is quite reasonable to expect that women with at least some primary education know more about proper nutrition, the dangers of contamination from water, vaccination, and other preventive measures. At the same time, such knowledge alone is not the key factor where socio-economic conditions are at their worst, as is the case in Bangladesh.

It appears from the evidence that work outside the home had a negative impact on infant and child mortality. In the Bangladesh situation, mothers may be forced to work in order to supplement insufficient family income or even to provide its principal source. Their work may thus result in less care for their children, but this needs more investigation before a meaningful interpretation can be made.

The relationship between contraceptive use status and infant and child mortality indicates one often neglected area that may be susceptible to programme intervention, through family planning and maternal and child health services. A family planning programme should aim at improving the health of both mothers and children by encouraging and assisting women to space their pregnancies

If women can be successfully encouraged to extend the period between pregnancies, both fertility and infant and child mortality could be expected to decline.

In general, it is now recognized that in some situations investment in fields not directly associated with health care programmes, such as environmental sanitation, provision of potable water, eradication of illiteracy, and improvement of educational levels {including adult education}, may yield, even in the short run, more substantial health benefits than further expansion of the health care service.


1. W. Brass, Methods for Estimating Fertility and Mortality from Limited and Defective Data (Laboratory for Population, University of North Carolina at Chapel Hill, Chapel Hill, N.C., 1975).

2. J. C. Caldwell, "Education as a Factor in Mortality Decline: An Examination of Nigerian Data," Population Studies, 33: 3 ( 1979).

3. Mitra and Associates, Bangladesh Contraceptive Prevalence Survey 1983 Key Results (Dhaka, 1984).

4. Samuel H. Preston, "Mortality, Morbidity and Development," Population Bulletin of the United Nations Economic Commission for Western Asia (United Nations, New York, 1978).

5. Susan H. Cochrance, The Effects of Education on Health, Staff Working Paper, no. 405 (World Bank, Washington, D.C., 1 980) .

6. T. J. Trussell, "A Re-estimation of the Multiplying Factors for Determining Childhood Survival," Population Studies, 29: 1 11975)

7. UNO, Indirect Techniques for Demographic Estimation, Population Studies, no. 81 (United Nations, Beirut, 1983).

8. World Bank, Poverty and Development of Human Resources: Regional Perspectives, Staff Working Paper, no. 406 (World Bank, Washington, D.C., 1980).


Oral rehydration therapy-the need for a proper perspective

C. Gopalan
Nutrition Foundation of India, New Delhi, India

Diarrhoeal diseases account for a considerable proportion of child deaths in developing countries and contribute significantly to the prevailing picture of malnutrition among children in these countries. There has, unfortunately, been a great deal of mismanagement and unnecessary medication in the treatment of diarrhoea in children. Antibiotics and other drugs, which are wholly uncalled for and ineffective, have been used indiscriminately. The promotion of oral rehydration therapy has contributed towards checking these undesirable trends and has served to underscore the importance of timely correction of dehydration as the central life-saving measure in the management of childhood diarrhoeas

Unfortunately, however, in recent years, "oral rehydration" has been vigorously promoted and propagated in a manner likely to give the impression to the unwary that it is the total answer to the problem of childhood diarrhoea. Much of the debate has centred on such issues as the composition of the oral electrolyte solution, on "adaptations" of the formula to suit home conditions, and on methods of ensuring supplies and deliveries of "oral electrolyte packets" to rural homes. While not minimizing the possible life-saving effect of the oral rehydration procedure in many cases of diarrhoea, this measure has to be seen in the proper perspective, a perspective mostly lacking at present.

The developed countries of Europe and North America did not achieve their success in prevention and control of diarrhoea through a reliance on a programme of distribution of oral electrolyte packets to their people; nor can the developing countries hope to achieve such success with this procedure. Oral rehydration is not a preventive procedure, nor even a curative one; it only addresses the problem of dehydration and nothing more, merely providing a partial therapy in the acute phase. By mitigating severe dehydration, it facilitates better child survival, but it can by no means eradicate the basic factors responsible for the problem of diarrhoea. It cannot reduce the incidence of diarrhoea, and it most certainly cannot solve the problem of undernutrition in the survivors, a problem further aggravated by the diarrhoea itself. In short, it is no more than a "holding operation" which can tide over the emergency posed by severe dehydration.

Heavy reliance on this procedure as the answer to the problem of diarrhoea at the public health level would imply failure on the part of health agencies to attack the root causes of the problem by: (a) providing safe water for their people and basic facilities for excrete disposal; (b) improving environmental sanitation; (c) mounting a programme of health education to promote personal hygiene and hygienic procedures for food handling and child feeding; and, most importantly, (d) improving the nutritional status of their people to levels which will endow them with a reasonably adequate immuno-competence. The current euphoria about "oral rehydration therapy" should not obscure the hard reality and mislead health agencies, planners, and policy-makers in developing countries from seeing the problem in its true perspective. The essential prerequisites for ensuring proper environmental sanitation and personal hygiene, and for promoting good nutrition, cannot be bypassed; there are no short-cuts.

This is not to suggest that promotion of oral rehydration must be given up, but merely that the current trend of overstating the case for oral rehydration alone, to the point of obscuring a balanced perspective of the diarrhoea problem, could prove counter-productive, leading to the relative neglect and de-emphasis of other essential programmes that alone can address the causes of the problem and not just its symptoms.

The importance of oral rehydration therapy in the management of the acute phase of diarrhoea cannot be minimized, but the prevention and treatment of diarrhoea involves more than the correction of dehydration. An overemphasis on the symptom of diarrhoea currently tends to neglect its nutritional consequences. These is an imperative need to ensure adequate caloric and protein intake even during the acute phase of the disease to compensate for the loss of nutrients. There is an even greater need for higher intake of calories, proteins, and other nutrients during convalescence in order to catch up growth and restore immuno-competence. Health agencies should educate health workers on dietary management of diarrhoea cases as much as stressing the use of oral rehydration procedures. This aspect is almost totally ignored at present among health agencies.

Under the circumstances currently prevailing in many developing countries, the provision of clean water and of facilities for excrete disposal will by themselves not solve the problem. More important are measures which promote personal hygiene and better hygienic practices in the matter of food handling and child feeding. Currently there is a great deal of intra-household contamination of foods through unhygienic practices of storage and handling of food, and as long as these persist heavy investment in programmes for providing safe water and excrete disposal facilities alone will not yield dividends.

As far as infants are concerned, one step which more than any other can help minimize diarrhoea is encouraging mothers to feed their infants exclusively on breast milk (ideally for the first half of their infancy, at least for four months). Given the present circumstances, it is best for the poor urban slum dweller to avoid commercial milk foods, the use of which calls for considerable precautions. Health workers could at least educate mothers about simple methods of avoiding food contamination through better storage of cooked foods and better handling. They should learn the more effective use of locally available inexpensive foods for feeding their children.

The prevention and control of diarrhoeal diseases in children is a major challenge to health agencies of developing countries. There are unfortunately no instant remedies, short cuts, or magic solutions. The problem will not be solved with crisis management strategies alone. It will be met only if the basic factors involved are addressed adequately.

Contents - Previous - Next