III - NUTRITION AND MORTALITY OUTCOMES

Growth Pattern of Children
Nutritional Status Of Preschool Children
Clinical nutritional deficiency signs
Birth Weight
Nutritional Status of Adults
Micronutrient Deficiencies
Infant and Child Mortality
Maternal Mortality

Growth Pattern of Children

Appropriateness of standards
Rural children
Secular trends

Anthropometric measurements like height, weight, arm circumference and fat-fold thickness are widely used for the assessment of nutritional status of individuals and communities. These measurements are influenced by age, sex, socio-economic and environmental factors. The National Nutrition Monitoring Bureau (NNMB) of the National Institute of Nutrition (NIN) is the major source of data on heights and weights of the rural population, while special studies conducted in different metropolitan cities provide information on growth and nutritional status of the urban population.

Appropriateness of standards

Assessment of growth involves the use of reference standards. The NCHS standards are commonly used for this purpose, though there is some debate as to whether these are appropriate for the Indian population. ICMR (1968) values of heights and weights have been found to be lower than NCHS standards, although these are based on data obtained mostly from poor rural communities, with constraints of malnutrition and infection, and hence do not reflect optimum growth. Special studies have been conducted among so-called ‘well-to-do’ school children to assess the normal growth pattern. The study conducted by NIN included a large number of children (Hanumantha Rao and Sastry 1979) aged 5-16 years from 14 public schools. The average weights and heights of these children were comparable to their American counterparts from 5-14 years in boys, but only up to 12 years in girls. Another study was conducted in a high income group, covering older boys (17-22 years) from the Institutes of Technology of four cities. Heights and weights of these boys were close to their American counterparts up to 14 years, but widened considerably during the period 14-16 years and remained more or less constant till 21 years. The differences in weight were more than the differences in height; the ‘well-to-do’ Indian man is about 5 cms shorter and 9 kg lighter than his American counterpart. The height of an Indian ‘well-to-do’ woman is also 5 cms less while the difference in weight is 6 kg. It is not clear whether these differences are due to genetic or environmental factors.

More recently, studies were conducted by the Nutrition Foundation of India (NFI) in affluent children under five years in seven different cities (Agarwal et al. 1991). Heights and weights of children in Ludhiana and Delhi corresponded to the NCHS standards, while the measurements were slightly lower in children of other cities. Although there is some regional variation, the use of NCHS standards has been recommended for all Indian children.

Rural children

Several studies (e.g. Hanumantha Rao et al. 1976) including NNMB surveys show that the heights and weights of rural children are lower than ‘well-to-do’ Indian children or NCHS standards. Longitudinal data are also available from special studies of linear growth in Indian children. In a study conducted by NIN, Hyderabad (Satyanarayana et al. 1986), a cohort of 700 rural children were followed from the age of 5 years to 20 years. They were classified into different nutritional grades, based on their height at the age of 5 years, and their growth rate was compared with that of ‘well-to-do’ children of Delhi. Peak height velocity was similar in all groups, while the peak weight velocity was lower in undernourished Indian children. Another interesting feature is the age at which the growth spurt occurs. The well-nourished Indian children had a growth spurt around 14 years, similar to that observed in Western children, while among undernourished boys the growth spurt is delayed by about 2 years. The total height gain between 10-20 years is around 40 cm in both Western and Indian boys. However, the rural children are shorter than the ‘well-to-do’ group at all ages, with the deficit in height being noted before puberty and then carried through to the adult stage. Unlike height gain, the increase in weight from 10-20 years was found to be lower in rural boys. If the rural adults are compared with the well-to-do group, the differences in heights and weights are about 7 cm and 10 kg respectively.

The heights and weights of rural girls were also found to be lower compared to the well-to-do group at all ages. The growth process in rural girls continued for a longer period and in fact, the increments in this group were much greater between 14-18 years. This is because the menarche, and consequently the adolescent growth spurt is delayed by 1-2 years in the case of poor rural girls. The mean age of menarche is around 12 years in well-to-do girls and 13-14 years in rural girls. If the rural women are compared to the well-to-do group the differences in heights and weights are about 6 cm and 7 kg respectively.

Secular trends

Aggregate data (pooled for all the States) on heights, weights and BMI over the period show a definite trend towards improvement. The average values of measurements in general, for almost all the age groups in both sexes show an increase; height increments tended to be more in children, with weight increments more visible in adults and adolescents. State-wise data clearly indicate that the heights of children and adolescents, and weights of adults and adolescents in the State, of Kerala and to some extent in Maharashtra and Gujarat were distinctly better compared to the 1970s.

Nutritional Status Of Preschool Children

Weight-for-age (underweight)
Stunting and wasting

Preschool children (defined as aged 1-5 years by NNMB) constitute the most vulnerable segment of the population and their nutritional status is considered to be a sensitive indicator of community health and nutrition. The prevalence of underweight, wasting and stunting among children is determined by anthropometric measurements and clinical assessment of deficiency signs. Weights and heights of children are compared with the reference standards (NCHS) and the degree of anthropometric deficit is usually expressed as the percentage of children below a specified cut-off such as 90% of the standard or below minus 2 standard deviations from the median value.

Weight-for-age (underweight)

In the Gomez classification, weight-for-age is used to classify children into different nutrition grades: normal: >90% of the standard, mild: 75-90%, moderate: 60-74% and severe <60% of the standard (NCHS). NNMB surveys between 1975-79 showed that about 6% children were ‘normal’ and about 15% were severely underweight, while a great majority were in the ‘moderate’ or ‘mild’ categories. Surveys repeated in the same areas during 1988-89 showed a decline in the prevalence of the severely underweight from 15% to 8.7% with a corresponding increase in the proportion of normal children (see Table 3). - There was no gender difference in underweight prevalences. These trends were observed in all states. The decline of severe grades was highest in Maharashtra, while in the states of Gujarat and Madhya Pradesh it was minimal (see Figure 3).

Table 3: Percentage distribution of children (1-5 years) according to Gomez classification and NCHS weight-for-age standards

Year	n	Normal	Mild	Moderate	Severe
1975-79*	6428	5.9	31.6	47.5	15.0
1983-84**	2244	5.4	34.7	46.3	13.6
1988-90*	13432	9.9	37.6	43.8	8.7

* repeat surveys, ** linked surveys

Source: NNMB Reports (1975-91)

Figure 3. Underweight Prevalences - 1-5 year olds, by state

Source: NNMB 1991

Stunting and wasting

The NNMB repeat surveys also classify children with respect to their height-for-age (stunting) and their weight-for-height (wasting). Children falling between median minus 2 standard deviations (- 2SD) and median - 3SD are considered as moderately wasted or stunted, while those below mean - 3 SD are classified as being severely so. When the data obtained from repeat surveys of NNMB were compared with those of earlier surveys, there was a substantial reduction in the percentage of children suffering from severe forms of underweight and stunting, with a corresponding increase in milder forms from 1975-79 to 1988-90 - a trend similar to that of Gomez classification (see Figures 4a to c). Effectively, this can be seen to be a manifestation of the rightwards shift in the distribution over this period, for these indicators. There was less of a shift in the distribution of proportions of children suffering wasting.

Clinical nutritional deficiency signs

The major nutritional deficiency signs encountered among preschool children are those of protein-energy malnutrition and vitamin A deficiency. The NNMB surveys conducted during 1988-90 showed that the prevalence of deficiency signs was lower compared to the earlier surveys between 1975-79 (NNMB 1991). The prevalence of marasmus decreased from 1.3 to 0.6% and kwashiorkor from 0.4 to 0.1% at the overall level. There were considerable variations between states, with Gujarat showed the highest prevalence of kwashiorkor (1.1%) and marasmus (4.9%). The prevalence of vitamin A deficiency (Bitot spots) declined from 1.8% to 0.7%. These observations thus also indicate an improvement in the nutritional status of children in the last 10-15 years.

Figure 4A. Weight For Age - Z score distribution

Figure 4B. Height For Age - Z score distribution

Figure 4C. Weight For Height - Z score distribution

Source: NNMB 1991

Birth Weight

Birth weight, an important determinant of child survival, is influenced by the nutritional status of the mother. National survey data indicate that the mean birth weights in India range from 2.49 to 2.97 kg with about 30% new-borns being less than 2.50 kg. There has been little change in the past three decades (Srikantia 1989). A gender difference has been noted in mean birth weights, female infants tending to be lighter than male counterparts. Hospital-based studies have shown differences ranging from 50-100 gm, while in the well-to-do group the weight difference was more pronounced, ranging from 100-300 gm.

The incidence of low birth weight is highest in low-income groups. A number of factors have been identified as risk factors: maternal age, weight, height, parity, literacy, income, infections and pregnancy-related complications. Even within low-income groups, a gradual increase in birth weight is evident with rising income, the differences between the poorest and the less poor being 100-150 gm (NIN 1983). Differences of a similar magnitude have been recorded between infants born to illiterate and educated mothers. With increasing maternal height from below 145 cm to 160 cm, differences in birth weights are of the order of 200-400 gm. Similarly, with increasing maternal weights from around 35 to 55 kg, differences in birth weights are of the magnitude of 100-250 gm (NIN 1985-86).

A recent study showed a good correlation between birth weights and BMI of mother (Naidu et al. 1991). The mean birth weight in women with BMI less than 16 (severe CED) was 2.50 kg and it showed a progressive increase with increase in BMI status of mothers. The mean birth weight was around 2.80 kg in women with BMI 18.5-25. The incidence of low birth weights was highest (53%) in severe CED and gradually declined to about 15% in those with normal BMI.

Anaemia is common among pregnant women, the prevalence rates ranging from 40-50% in urban areas and 50-70% in rural areas (Gillespie et al. 1991). A significant fall in birth weight has been observed with decrease in haemoglobin levels. In one study (Ramachandran 1989), the incidence of low birth-weight babies was 40% in anaemic women with haemoglobin less than 8 gm as against 23% in women with normal haemoglobin levels.

One of the goals of the National Health Policy is to reduce the prevalence of low birth weight to 10% by 2000 AD. There is an urgent need to strengthen health care services so as to reach all pregnant women in need and to improve their nutritional status to achieve this goal. In the current programmes, women receive attention only after the onset of pregnancy. The preceding years of adolescence are crucial and special efforts are needed to reach girls at this stage.

Nutritional Status of Adults

Relationship between adult BMI and various factors
Maternal nutritional status

Anthropometric data available on rural adults from the NNMB surveys, as well as other studies, show that their heights and weights are lower than Western counterparts. Comparison of the recent surveys of NNMB (1988-90) with the earlier surveys (1975-79) did not show much difference between the two periods. Body Mass Index (BMI) expressed as weight (kg)/height² (m²) is considered to be a better indicator of chronic energy deficiency (CED) in adults. For classifying individuals according to BMI status - a proxy for CED - the cut-off levels suggested are 16, 17 and 18.5 (see Table 4).

Table 4: Classification of adult chronic energy deficiency

BMI class	Presumptive diagnosis
< 16.0	CED Grade III (Severe)
16.0 - 17.0	CED Grade II (Moderate)
17.0 - 18.5	CED Grade I (Mild)
18.5 - 20.0	Low Weight (Normal)
20.0 - 30.0	Normal
>30.0	Obese Grade II

Source: Ferro-Luzzi et al. (1988)

Recently the NNMB data of adults was analysed to see the distribution of BMI in the Indian population (Naidu et al. 1991). The average values of BMI were similar for both males and females and the same BMI classification was used for both. About 50% of the population had BMI less than 18.5 and only a negligible proportion (0.2 - 0.5) were found to be obese/overweight. The results thus show that the problem of CED is of primary importance in India, not obesity. Significant inter-state variations were seen in the distribution, as illustrated in Table 5. When the data of 1975-79 were compared with those of 1988-90, a clear shift to right is seen in the distribution of BMI suggesting an improvement in nutritional status of adult population over the period.

Figure 5. Male Mortality Rates - by BMI Grade

Source: Satyanarayana et al. (1991)

	A.P.	Gujarat	Karnataka	Kerala	Maharasthra	M.P.	Orissa	Tamil Nadu	All-India
Male	20.5	31.3	22.1	20.5	22.9	13.2	16.6	21.1	21.3
Female	26.2	30.5	26.8	18.6	29.6	17.5	23.1	22.9	24.2

Source: Naidu et al. (1991)

Relationship between adult BMI and various factors

Socio-economic status. When the BMI values were examined in relation to land ownership, it was observed that adult men and women from households with more than 5 acres of land had better BMI status compared to the landless group. Similarly, cultivators and artisans had a better BMI status compared to landless agricultural labourers (see Figure 6a and b). A strong association was found between per capita income and BMI status; about 60% of adult men and 50% of adult women had BMI values less than 18.5 when the income was less than Rs. 60/month, while the corresponding figures were 36% and 33% respectively in groups having per capita income of greater than Rs. 150 per month (Naidu et al. 1991).

Figure 6A. Male BMI Distribution - By occupational group

Figure 6B. Female BMI Distribution - By occupational group

Source: Naidu et al. (1991)

Child nutritional status. The ‘average household BMI’ values were derived from BMI of adult males and females in the household (Naidu et al. 1991).’ Children were classified as stunted, wasted and underweight based on their weights and heights. Bivariate distributions were derived between ‘household BMI’ values and child nutritional status. The results showed a close positive association between child nutritional status and ‘household BMI’, particularly with respect to stunting in children and less so with wasting, as would be expected.

Maternal nutritional status

Women of child bearing age constitute the vulnerable segment of the population because of their special reproductive needs. According to the 1981 census there were 143.5 million women of child bearing age, constituting about 21% of the total population. An average Indian woman has 6-7 pregnancies resulting in 5-6 live births, of which 4-5 survive. She is estimated to spend a greater part of her reproductive years in pregnancy and lactation. Chronic maternal undernutrition and overwork among the low income group pose a serious threat to the welfare of the mother and the infant.

Studies in Indian women of a high-income group (Ramachandran 1989) have shown that their dietary intakes range between 2000-2500 kcals per day during pregnancy. In this group, women generally do not perform hard physical labour and there is a reduction in activity during pregnancy (NIN 1983). The average weight of this group ranges between 45-55 kg and the mean weight gain during pregnancy is about 11 kg - observations similar to those reported in women from developed countries.

Studies in urban women of a low-income group have shown that their dietary intakes range from 1200-1600 kcals per day. The average pre-pregnancy weight of these women is around 43 kg and they gain 6 kg during pregnancy. In rural India, dietary intakes of women are slightly higher (1600-1900 kcal). However, rural women have to spend more energy in daily household chores. For instance, they have to fetch drinking water from sources which may be 1-2 kilometers away from home, and gather and bring firewood from forests miles away. Most rural women from low-income groups are heavily engaged in agricultural activities. It is therefore not surprising to find that these women weigh less than the urban women. The NNMB surveys showed that the average weight and height of rural women are 42 kg and 152 cm. About 33% of 18 year old women have body weights below 40 kg and 15 percent have heights less than 145 cms (NNMB 1991). These women fall into high risk category as they are likely to suffer from obstetric complications and give birth to small babies.

Micronutrient Deficiencies

Vitamin A deficiency
Iron deficiency anaemia
Iodine deficiency disorders

Apart from protein-energy malnutrition, deficiencies of specific micronutrients such as vitamin A, iron and iodine are common in India, affecting large segments of the population.

Vitamin A deficiency

Studies conducted by ICMR during 1965-69 showed that 7% of preschool children had ocular signs of vitamin A deficiency (ICMR 1977); 4.2% had Bitot spots. The NNMB surveys between 1975-79 (NNMB 1991) showed that the prevalence of Bitot spots in preschool children was around 2%. More recent surveys (1988-90) repeated in the same areas showed a decline in the prevalence (to 0.7%). There is a wide variation in the prevalence rate, ranging from 0.5% in Kerala to 1.0% in Andhra Pradesh (NNMB 1991).

Most of the community surveys are based on clinical signs of mild xerophthalmia while severe deficiency resulting in corneal xerophthalmia and blindness is relatively rare; the risk of mortality is also high (60%) in such cases. A nationwide survey conducted by the ICMR during 1971-74 (ICMR 1991) showed that, of the 9 million blind people in the country, 2% cases were attributable to corneal disease caused by vitamin A deficiency. In the more recent survey (1985) 0.04% of the total blindness has been attributed to vitamin A deficiency. This is clearly an underestimate of the problem as corneal scars have been excluded although vitamin A deficiency is an important cause of corneal blindness. Mortality is also high in such cases (Menon and Vijayaraghavan 1980). Assessment of nutritional blindness by cross-sectional surveys does not reveal the true picture since only the survivors can be examined in these surveys. The reported incidence of corneal xerophthalmia among preschool children is 0.05% The estimates based on these figures suggest about 30,000 cases of corneal xerophthalmia per year, of which nearly half would result in permanent blindness.

The Government of India launched the national vitamin A prophylaxis programme in 1971. Under this programme, sponsored by the Ministry of Health & Family Welfare, children between 1-5 years are given a massive dose of 200,000 IU every six months. The programme is now in operation in almost all the states in the country covering about 30 million children. An evaluation study (Vijayaraghavan and Prahlad Rao 1982) has shown that in areas where the vitamin A programme was implemented well, there was a significant reduction in the prevalence of xerophthalmia, while in other areas the coverage was unsatisfactory. Reasons for poor coverage include inadequate supplies of vitamin A, irregular distribution of the dose, poor coordination between the various health functionaries, non-involvement of village level workers and absence of supporting nutrition education. Based on the recommendations of a review committee, the programme is now being modified to improve the outreach of the target population. The new approach involves integrating the vitamin A distribution programme with other child care services and involving village level workers (anganwadi workers) in the distribution of the dose.

Iron deficiency anaemia

Anaemia is a problem of serious public health significance, given its impact on physical work capacity, mental performance, maternal morbidity and mortality. The most vulnerable groups are pregnant women and preschool children. A number of sample surveys carried out during 1960s (Ramachandran 1989) showed that more than 50% of the pregnant women have haemoglobin levels below 11 gm/1 and are thus classified as ‘anaemic’. More recent surveys indicate that anaemia is common even in other segments of the population. In rural areas around Hyderabad and Delhi, the prevalence of anaemia ranged from 40-70%, while in villages near Calcutta where hookworm infestation was common, more than 90% of the population were anaemic (AJCN 1982). In all the areas, women of child-bearing age had the highest prevalence of anaemia, followed by preschool children, school children and adult men. These data indicate that anaemia in India is much more widespread than hitherto believed and suggests the need to cover the entire population of pregnant and lactating women in rural areas in any intervention programme designed to control anaemia (see Gillespie et al. 1991).

The Government of India started a national anaemia prophylaxis programme in 1970 for reducing the prevalence of anaemia in vulnerable groups. The target population comprise pregnant and lactating women, women acceptors of family planning, and children between 1-12 years. Adult beneficiaries are given iron folate tablets containing 60 mg elemental iron and 500 ug of folic acid, while children are given smaller tablets containing 20 mg iron and 100 ug folic acid. The programme is implemented through all the institutions providing MCH services e.g. primary health centres, health sub-centres and maternity clinics. An evaluation study conducted by the Ministry of Health and Family Welfare (1989) revealed a poor performance in all the states. The coverage of pregnant women ranged from 3-26%. The number of children covered was also negligible. The reasons for poor coverage included inadequate supplies of iron folate tablets, poor supervision, and poor compliance of women due to lack of knowledge and ignorance. More than 80% of the pregnant women had haemoglobin less than 11 gm. Thus there was no impact on the prevalence of anaemia. There is an urgent need to strengthen the national anaemia prophylaxis programme with better control over actual delivery of tablets. Adequate supply of iron folate tablets, involvement of village-level workers in the distribution of tablets and a greater motivation of health workers and the community will be important for the successful implementation of the programme.

Iodine deficiency disorders

Goitre has been recognised as an endemic problem in the Himalayan and sub-Himalayan regions in the past half century. Surveys between 1945-53 in sub-Himalayan belt, stretching from Kashmir in the North-West to Nagaland in the East indicated a prevalence of goitre ranging from 26-90% (ICMR 1981). Surveys carried out most recently by the Directorate General of Health Services and other agencies indicated the presence of goitre even outside the conventional goitre belt of Himalayas. ICMR conducted a survey covering 14 districts of 9 states with diverse geological, metrological and geochemical characteristics (ICMR 1989). The study confirmed a high prevalence of goitre in all the areas. In some districts like Dibrugarh, which lies in the Himalayan belt, the prevalence was as high as 66%. In the extra-Himalayan regions the prevalence ranged from 24% in hilly districts to 19% in coastal districts and 12% in the plains, with prevalences higher among tribals than non-tribals. The prevalence rates in each of these districts were well above the level recognised for endemicity. More alarming is the prevalence of endemic cretinism in all the 14 districts surveyed, though in varying proportions, with Manipur having the highest prevalence of 6.1%. Apart from cretinism, children in endemic areas show varying degrees of thyroid deficiency and developmental defects. A study conducted by the All India Institute of Medical Sciences, Delhi revealed a high prevalence of neonatal chemical hypothyroidism (NCH), as defined by thyroxine <30 ug/dl and TSH >50 ug/ml. In one particular district (Gonda) of Uttar Pradesh, the prevalence rate of NCH was as high as 15%. Without treatment, these infants are likely to develop physical and mental defects. In the villages of Uttar Pradesh and Bihar where goitre prevalence was high, deaf-mutism, mental retardation and other clinically detectable problems of environmental iodine deficiency were found in 4% of the children. It is estimated that, today, more than 54 million people in India are suffering from endemic goitre and 8.8 million from different grades of mental/motor handicaps (Khatri 1990). The magnitude of the problem is thus far greater than hitherto believed.

Although the national goitre control programme has been in operation for the last three decades, it has gained momentum only recently. Initially, iodised salt production was managed by the Hindustan Salt Ltd., a public undertaking. The production was grossly inadequate and could not meet the demands of even the endemic areas. Transportation of salt from production centres also posed serious problems, as did poor supervision and the lack of quality control.

Recently, following the recommendations of the review committee, the Government of India has liberalised the production of iodised salt to include the private sector, in addition to the public sector. More important is the historic decision taken in November 1984 to iodate the entire edible salt in the country in a phased manner. The Government of India stands committed to universal iodisation of the entire edible salt supply by 1992. A financial outlay of Rs. 300 million has been provided during the 7th Five Year Plan, and goitre control cells and quality control laboratories are being set-up in all states.

Infant and Child Mortality

Relation between prevalences of severely underweight and mortality

The infant mortality rate (IMR) is a valuable indicator of health and development. It has shown a decline from 129 per 1,000 live births in 1971 to 91 in 1989. There is a wide variation between urban and rural areas (see Figure 7) and between states, ranging from 28 in Kerala to 123 in Uttar Pradesh in 1988 (see Figure 8). The data provided by the office of Registrar General of India shows the factors associated with infant deaths. These include maternal age, education, age at marriage, religion, social status and annual income of the household. The figures for 1984 show that the mean IMR for women married before 12 years of age was 135.3 as compared to 78.2 for women married between 21-23 years old. The IMR for Hindu women was highest of religious groups, being 107.6, as compared to 98.9, 54.4 and 40.1 for Muslims, Sikhs and Christians respectively. Scheduled castes had the highest IMR of 126.5, compared to non-scheduled castes and tribes, 99.2. IMR varied markedly with educational status of the mother, ranging from 119.9 for illiterates to 21.0 for women educated to matriculation or above. Household income also had a substantive effect as expected, with IMRs of 124.2 for annual household income below Rs. 5,000, compared to 79.7 for household incomes above Rs. 10,000. There were no gender differences in IMR. The under-five mortality rate has declined from 282 per 1000 live births in 1962 to 142 per 1000 live births in 1990 (UNICEF 1992).

Figure 7. Infant Mortality Rates - Rural and urban

Source: Registrar General at India

Figure 8. IMR by State - 1988

Source: Registrar General at India

Relation between prevalences of severely underweight and mortality

Data available on the prevalence of malnutrition and child mortality were examined to see the relationship between the two. The percentage of severely underweight children was taken from an ICDS baseline survey of 23 districts in 12 states during 1983-84. The district-wise child mortality rates for 1981 available from the office of the Registrar General, Ministry for Home Affairs were also examined. There was a significant positive correlation between the mean district-level severely underweight prevalences and the child mortality rate among children under 5 years (see Table 6).

Table 6: Regional nutrition and mortality profile among children in ICDS project areas during 1984-85*

State	Type of project	% severely underweight		Child mortality rate**
		0-3	3-6	0-2	0-5
Andhra	Rural	8.8	11.3	99	142
Pradesh	Tribal	22.2	32.5	169	196
Karnataka	Rural	15.4	12.9	124	182
	Rural	17.2	10.8	109	149
Tamil Nadu	Urban	5.7	4.6	70	80
	Urban	5.7	0.3	74	101
Kerala	Rural	1.2	4.3	66	104
	Rural	0.7	4.0	74	116
Madhya	Rural	34.4	16.9	220	265
Pradesh	Tribal	26.0	17.3	138	174
Uttar Pradesh	Rural	14.6	13.0	163	211
	Rural	6.1	3.1	154	187
Rajasthan	Rural	12.4	NA	214	252
	Tribal	14.2	0.7	170	169
Maharashtra	Rural	22.3	25.1	177	227
	Urban	12.0	10.2	87	107
Gujarat	Tribal	4.6	3.3	91	96
	Tribal	7.3	5.6	115	129
Orissa	Rural	3.2	2.8	196	204
West Bengal	Rural	21.3	12.2	115	149
	Tribal	6.5	8.2	77	89
Haryana	Rural	6.7	4.2	120	146
	Rural	5.5	0.7	114	129

Source: Kakwani and Subbarao (1990b)

* The nutritional data are taken from ICDS project-wise progress report for 1984-85, for those projects which were sanctioned during 1983-84. As such, these figures reflect the nutrition situation at about the commencement of ICDS in these districts.

** District-wise child mortality rates for 1981 for the 0-2 and 0-5 year olds were made available for the first time in the Occasional Paper 5, 1988 of the Office of Registrar General, Ministry for Home Affairs.

Maternal Mortality

It is estimated that out of half a million maternal deaths in the world each year about 20% are in India. Maternal deaths are caused either by direct causes arising from complications of pregnancy, delivery or their management; or indirect causes due to aggravation, by pregnancy or child birth, of an existing abnormal condition.

India’s maternal mortality rate, estimated at 3.4 per 1000 live births, on average between 1980-87 (UNDP 1990 pp.148) compares with an average of 2.9 for ‘all developing countries’ and 0.24 for industrialized countries. There is no figure for the maternal mortality rate of the country which can be considered as reasonably conclusive, more so because levels as high as 13 have been noted in certain rural areas. The actual life-time risk of an Indian woman dying from a maternity-related cause is far greater than comparative rates between India and the industrialized countries would suggest owing to the higher total fertility rate (4.0 in 1990 according to the World Bank 1992). Maternal age and number of births have a strong effect on maternal mortality. A woman giving birth to children at 20-35 years of age faces a much lower risk than women below 20 and over 35 years. An estimated 8% of the 26-27 million annual births in India are to mothers below 19 years, whose growth and maturation may be retarded. Maternal illness and death rose significantly with the fourth pregnancy and reached a high level after the fifth. Some 35% of live births in rural areas and around 29% of live births in urban areas are of the fourth birth order and above (Registrar General of India 1984). In India, as in many developing countries, maternal mortality accounts for the-largest, or near-largest, proportion of deaths among women in their prime years (UNICEF 1990 p14). The trend in the percentage distribution of deaths by causes related to child birth and pregnancy in rural India is given in Table 7.

Table 7: Percentage distribution of deaths by cause related to child-birth pregnancy 1981-86

Specific causes	1981	1982	1983	1984	1985	1986
Abortion	13.7	10.1	10.7	10.8	11.5	8.0
Toxaemia	8.0	12.5	12.1	10.8	6.7	11.9
Anaemia	17.7	24.4	18.9	23.3	23.1	17.0
Bleeding of pregnancy	23.4	26.2	23.5	18.8	15.9	21.6
Obstructive labour	9.2	7.2	8.3	6.2	7.7	6.2
Puerperium sepsis	13.1	8.3	11.6	10.8	13.9	13.1
Non-classifiable	14.9	11.3	14.6	19.3	21.2	22.2
Total	100.0	100.0	100.0	100.0	100.0	100.0
Sample number of deaths	175	168	206	176	208	176
Percentage of total deaths	1.0	1.0	1.2	1.0	1.2	1.0

Source: Registrar General of India, Survey of Censuses of Deaths (Rural) 1984 and 1987