Inequality in child undernutrition among urban population in India: a decomposition analysis

Background With increasing urbanization in India, child growth among urban poor has emerged as a paramount public health concern amidst the continuously growing slum population and deteriorating quality of life. This study analyses child undernutrition among urban poor and non-poor and decomposes the contribution of various factors influencing socio-economic inequality. This paper uses data from two recent rounds of National Family Health Survey (NFHS-3&4) conducted during 2005–06 and 2015–16. Methods The concentration index (CI) and the concentration curve (CC) measure socio-economic inequality in child growth in terms of stunting, wasting, and underweight. Wagstaff decomposition further analyses key contributors in CI by segregating significant covariates into five groups-mother’s factor, health-seeking factors, environmental factors, child factors, and socio-economic factors. Results The prevalence of child undernutrition was more pronounced among children from poor socio-economic strata. The concentration index decreased for stunting (− 0.186 to − 0.156), underweight (− 0.213 to − 0.162) and wasting (− 0.116 to − 0.045) from 2005 to 06 to 2015–16 respectively. The steepness in growth was more among urban poor than among urban non-poor in every age interval. Maternal education contributed about 19%, 29%, and 33% to the inequality in stunting, underweight and wasting, respectively during 2005–06. During 2005–06 as well as 2015–16, maternal factors (specifically mother’s education) were the highest contributory factors in explaining rich-poor inequality in stunting as well as underweight. More than 85% of the economic inequality in stunting, underweight, and wasting among urban children were explained by maternal factors, environmental factors, and health-seeking factors. Conclusion All the nutrition-specific and nutrition-sensitive interventions in urban areas should be prioritized, focusing on urban poor, who are often clustered in low-income slums. Rich-poor inequality in child growth calls out for integration and convergence of nutrition interventions with policy interventions aimed at poverty reduction. There is also a need to expand the scope of the Integrated Child Development Services (ICDS) program to provide mass education regarding nutrition and health by making provisions of home visits of workers primarily focusing on pregnant and lactating mothers. Supplementary Information The online version contains supplementary material available at 10.1186/s12889-020-09864-2.


Background
The rapid increase in the urban population due to urbanization in the last one decade has contributed to the urban poverty in India [47]. It is manifested in the form of inadequate provision of housing and shelter, water, sanitation, health, education, social security, livelihoods, and food security along with the special needs of vulnerable groups like women, and children [13]. The health of urban children as measured with malnutrition in the existing context of urbanization, has become an important issue in India [26]. Malnutrition among children under age five is a paramount public health concern in many developing countries, and India is no exception [33].
Child under-nutrition is expressed in different forms, including stunting (− > 2 SD length/weight for age), wasting (− > 2 SD weight for length/height), and underweight (− > 2 SD weight for age) [43]. In developing countries, an estimated 32% of children under five were stunted, 3.5% were severely wasted, and 20.2% were underweight [4]. The rate of stunting in India has declined from 48% to 38.4%, wasting has increased slightly from 19.8% to 21%, and underweight has decreased from 42.5% to 35.8% between the two time-periods of NFHS-3 and NFHS-4 [17]. Malnutrition in India is clustered in central, northern, and eastern regions [46]. In a study conducted in Indian settings, it was found that the spatial clustering of malnutrition was visible in those geographical pockets where poverty was high, women's education was low, and BMI level among women was below average [22].
Malnutrition is primarily caused due to the immediate causes (inadequate dietary intake, lack of care and onset of disease), and the underlying causes (an unhealthy environment and insufficient education) [9]. Studies in countries other than India had also raised the issue of child growth significantly. In a study, it was found that there was a substantial poor/rich gap in the nutritional status in the urban residence of Kenya [11]. Moreover, the same study argued that poor nutritional status in urban areas is attributed to poor environment and housing, poor access to drinking water, food, and health services. One of the studies found that the children belonging to poor households were undernourished not only because of poverty, but also due to limited use of maternal health care services and poor care resulting from the lower educational status of parents and poor health of mothers [5].
Despite high levels of stunting, wasting, and underweight among under-five children in India, studies related to measuring growth faltering as measured with z score are very limited [14]. Studies have found that growth in weight appeared to start right from birth, and there appeared a fall in WAZ score for the first 3 months, and after that, the rate of fall appeared to slow down [25]. The study was done to compare NFHS-1 and NFHS-3 data to provide a comparison for all India level growth among children aged 0-3 years [25]. Growth in infancy was associated with poorer educational outcomes in later childhood. However, these outcomes, when adjusted for confounding factors, suggested that childhood educational attainment was more strongly influenced by other socio-demographic factors, especially maternal education and the quality of the home environment [15,42]. Similarly, it was found that early childhood stunting is associated with deficits in cognition and educational achievement in late adolescence [42].
India ranks 102 in the Global Hunger Index [12], shows the severity of under-nutrition existence in the country. Moreover, other studies too carved out the fact that the poverty nutrition trap does exist in India [19]. There exist lack of literature focusing on examining the socio-economic gap for malnutrition among children in urban India. Therefore, this paper aims to measure undernutrition among urban poor and non-poor children below age five in India. It also analyses the changes in socio-economic inequality in undernutrition among children in urban India from 2005 to 06 to 2015-16.

Data
The data used in this paper have been taken from two rounds of National Family Health Survey namely NFHS-3 (2005-06) and NFHS-4 (2015-16). The primary objective of the NFHS was to provide essential data on health and family welfare, as well as data on emerging issues including the nutritional status of children under age five at national, state, and district levels. Therefore, data from NFHS-4 (2015-16) was considered as a benchmark to examine the progress made by the country in various health indicators over time. Decisions about the overall sample size required for National Family Health Survey (2005-06 and 2015-16) were guided by several considerations, paramount among which was the need to produce indicators at the national levels, as well as separate estimates for urban and rural areas. The data is a nationally representative data which provide essential data on health and family welfare, as well as data on emerging issues in these areas. Additionally, National Family Health Survey is a part of Demographic and Health Surveys (DHS) which are nationally-representative household surveys that provide data for a wide range of monitoring and impact evaluation indicators in the areas of population, health, and nutrition. The findings of this study are based on a total of 15,241 urban children under 5 years in NFHS-3 and 53,483 urban children under five years in NFHS-4. The information related to child's anthropometric measures covered measurements of height, weight and hemoglobin levels for children; Details about the NFHS-4 designs, tools, and protocols are presented in the national report of NFHS-4 [17], and all relevant information is available in the public domain on http://rchiips. org/ NFHS.

Categorization of variables Outcome variable
The three anthropometric indicators were used for children under five years: stunting (height for age < − 2 standard deviation (SD)), underweight (weight for age < − 2 SD), and wasting (height for age < − 2 SD). The measure used was recoded as per WHO standards [45].

Predictor variable
The study divided the predictors into five groups. 1) Maternal Factors 2) Socio-economic factors 3) Child-related factors 4) Environmental factors and 5) Health seeking factors. Maternal factors include age (15-24, 25-24 and 35+ years), Education (Illiterate and literate), age at first birth (< 18, 18-30, and > 30 years), Parity (1, 2, 3, 4+) and mother's anaemia status (no and yes). Socioeconomic factors include caste (SC/ST and non-SC/ST), religion (Hindu, Muslim, Christian, and Others) and Wealth status (Poorest, Poorer, Middle, Richer, and Richest). Child factors consist age (0-23, 24-47 months and 47-59 months), sex of the child (male and female), anemia status (no and yes), diarrhea, birth weight (normal and low birth weight (<2500grms), and colostrum feeding (yes and no). The environmental factors include the source of drinking water (protected and not protected), open defecation (opened and improved), stool disposal (yes and no), and geographical regions (North, Central, East, North-east, West, and South). Health seeking factors included the use of Skill Birth Attendants, Place of delivery (Home or any institution i.e., private/ public hospital), full immunization includes children who received BCG (for tuberculosis), measles, and three doses each of DPT (diphtheria, whooping cough, and tetanus), and Polio. Full ANC is defined as women who had four or more ANC visits, had at least two tetanus injection, and consumed 100 IFA tablets/syrup for the most recent birth. Further, Colostrum feeding, defined as a child received breastfeeding within 1 h after delivery.
The variable of wealth status was created using the information given in the survey. Households were given scores based on the number and kinds of consumer goods they own, ranging from a television to a bicycle or car, and housing characteristics such as source of drinking water, toilet facilities, and flooring materials. These scores are derived using principal component analysis. National wealth quintiles are compiled by assigning the household score to each usual (de jure) household member, ranking each person in the household population by their score, and then dividing the distribution into five equal categories, each with 20% of the population. Further, wealth status of the household was recoded as poor and non-poor. To depict the poor and non-poor differentials, we clubbed first two wealth quintiles as poor (poorest and poorer) and next three quintiles as non-poor (middle, richer and richest).

Statistical analysis Concentration index (CI)
The concentration curve is obtained by plotting the cumulative proportion of outcome variables (either stunting, underweight, or wasting) on y-axis against the increasing percentage of the population ranked by the socio-economic indicator (wealth index) on x-axis. The curves shows that whether the socio-economic status related inequality in the outcome variable (on x-axis) prevails or not. If the curve is above the line of equality (45 degree line) that means the index value is negative hence it shows that the outcome variable is disproportionally concentrated among the poor and vice-versa. Income-related inequality in stunting, underweight, and wasting was measured by the concentration index (CI) and the concentration curve (CC), using the wealth score as the socio-economic indicator and binary outcome as stunting, underweight, and wasting. The concentration index is defined as twice the area between the concentration curve and the line of equality. Concentration index measures the inequality of one variable (say stunting, underweight or wasting) over the distribution of another variable (wealth index). The index ranges from − 1 to + 1, where the index value of 0 (zero) shows no socio-economic inequality. However, positive value of index shows pro-rich inequality and vice-versa. Additional on the either scales higher the value, higher the extent of socio-economic inequality. The study used Wagstaff decomposition analysis to decompose the concentration index. Wagstaff's decomposition demonstrated that the concentration index could be decomposed into the contributions of each factor to the income-related inequalities [41]. For any linear regression model on health outcome (y) (say stunting, underweight or wasting), such as.
The concentration index for y, C, can be written as follows, Where μ is the mean of y, x k is the mean of x k , C k is the concentration index for x k (defined analogously to C), and GC ε is the generalized concentration index for the error term (ε). Equation (2) shows that C is equal to a weighted sum of the concentration indices of the k regressors, where the weight for x k is the elasticity of y with respect to x k ðη k ¼ β k x k μ Þ. The residual component captured by the last term reflects the socio-economic status related inequality in health that is not explained by systematic variation in the regressors by income, which should approach zero for a well-specified model. Each contribution is the product of elasticity with the degree of economic inequality. Moreover, the percentage contribution is obtained by dividing each absolute contribution by total absolute contribution multiplied by 100 to obtain the estimates. Table 1 shows the distribution of children under age five by selected background characteristics among urban poor and non-poor in India. It was found that significant health problems like anaemia and diarrhoea among under-five children were more prevalent among those from urban poor as compared to those from urban nonpoor. On contrary, health-seeking behaviour was better practised by urban residents as compared to non-poor population. Figure 1 portrays the prevalence of child growth as measured through stunting, wasting, and underweight among urban poor and non-poor in India, during 2005-06 and 2015-16. It is evident from the results that three key indicators of growth among children under age five, namely stunting, underweight, and wasting were higher among urban poor children than among urban non-poor children for both the time-periods. Further, results show that the prevalence of stunting and underweight has declined among urban poor as well as non-poor children during the last one decade, i.e., from 2005 to 06 to 2015-16. However, the prevalence of wasting has increased, from 2005 to 06 to 2015-16, among urban poor as well as urban non-poor children, which may be primarily due to seasonal variation in the month of surveys. By and large, a similar result has been found in case of growth for poor and non-poor children in urban India during 2005-06 and 2015-16. The result found that the average growth is higher among urban poor than in urban non-poor for each classification of child anthropometry measures (Figure-S1; Supplementary file). For standard error, please refer supplementary Table S1 and S2. Table 2 presents the average growth among children as measured through z scores for HAZ, WHZ, and WAZ by segregating them into different age groups. Mean  in 2015-16 for urban poor and non-poor respectively. Significant differences in z score in poor as well as non-poor children were visible from the first 5 months itself. Of the total growth differential as measured with z score for height-for-age (HAZ) at 59 months among urban poor and non-poor in 2005-06 (− 2.51 and − 1.59), 38% (− 0.95) and 22% (− 0.36) of growth took place at first 5 months after birth; whereas, in 2015-16, the distribution increased to 41% (− 0.86) and 25% (− 0.33) respectively. Similarly, of the average growth in weight-for-age (WAZ) at 59 months among urban poor and non-poor in 2005-06 (− 2.18 and − 1.54), 71% (− 1.54) and 68% (− 1.05) of growth took place at first 5 months after birth; whereas, in 2015-16, the distribution was 41% (− 0.86) and 25% (− 0.33) respectively. For weight-for-height (WHZ) at 59 months among urban poor and non-poor in 2005-06 (− 0.99 and − 0.88), 112% (− 1.11) and 113% (− 1.00) of growth took place at first 5 months after birth; whereas in 2015-16, the distribution was 99% (− 1.09) and 131% (− 1.25) respectively. Growth for HAZ among children was found to be highest at age 21-25 months among urban poor, and at age 46-50 months among urban non-poor during 2005-06, whereas, growth in stunting was highest at age 21-25 months among urban poor and non-poor during 2015-16. It leads to the premise that the children were more susceptible to be stunted during the age of 21-25 months. Z score for WAZ among children was found to be highest at age 31-35 months among urban poor and at age 56-59 months among urban non-poor during 2005-06, whereas, z score for underweight was highest at age 21-25 months among urban poor and at age 56-59 months among urban non-poor during 2015-16. It means that urban non-poor children aged 56-59 months were more prone to be underweight. The z scores for WHZ among children was found to be highest at age 6-10 months among urban poor and at age 0-5 months among urban non-poor during 2005-06, whereas, z score for wasting was highest at age 6-10 months among urban poor and at age 0-5 months among urban non-poor during 2015-16. It means that children in their early months were more prone to be wasted.

Discussion
Despite recent achievements in economic progress, India failed to ensure better outcomes in child growth. Results further suggested that there exists socio-economic clustering that is further widening the gap in securing development among children. Through this study, we have tried to explain the factors contributing to economic inequality in stunting, underweight, and wasting among children below five years in urban India.
There is little data to understand the timing of growth faltering in under-five age children in India, and whatever limited study conducted to understand the growth faltering among children measured growth faltering  through reference growth standards of NCHS (National Centre for Health Statistics) methodology [25]. This study has the edge over previous studies in measuring growth faltering as this study used WHO standards to measure the growth faltering among children over NCHS methodology [34]. The result confirms that wasting among urban poor and non-poor children was more prone during the first year of birth, and stunting among urban poor and non-poor children is more prone during the second year of birth. Growth flattering in underweight is more during the second year of birth among urban poor children and the fifth year of birth among urban non-poor children. Critically, children aged 21-25 months were more prone to growth faltering than the children of any other age group. Studies found that children in low-middle-income countries experience rapid growth faltering during the first two years of life [40]. Maternal factors remained a prime contributor to inequality. Stunting in simple terms is defined as -> 2 SD height-for-age [38], and measures the past or chronic child undernutrition. The mother's related factors explained a little more than half of the inequality for stunting among children during 2015-16. Various studies in Indian set up are in coordination with this study, concluded that maternal factors such as mother's age mother's education, age at first birth, parity, and anaemia were highly associated with stunting [28,36]. Of all the mother's factors, mother's education is the most significant contributor to the economic inequality in defining stunting among children. Not only in India [10], but in developed [44] and developing countries [2,20] also, mother's education was found significantly associated with stunting among children. Among environmental factors, defecation and stool disposal explained the highest economic inequality in stunting among children. A study conducted in Rwanda for 10 years, using decomposition analysis reported that defecation has no association with stunting [30]. Another study conducted in rural Zimbabwe employing cluster-randomized trials found that WASH (Water, Sanitation, and Hygiene) practices implemented in rural areas are unlikely to reduce stunting [16]. However, few studies found that sanitation and defecation significantly affect stunting among children [24]. A study conducted in Indian settings utilizing NFHS data concluded that the practice of open defecation influences the inequality in stunting [35]. Health seeking factors explained about 30% of the inequality in stunting in 2005-06, and in 2015-16, it could only explain about 10% of the inequality. From various works of literature, it can be summed up that full Antenatal Care and Skill attendants at birth had increased over time in India [1].
Underweight is defined as -> 2 SD weight-for-age, and it depicts the past as well as present undernutrition. During both the time, mother's factors remained the highest contributory factors that explained the inequality in underweight among children aged 0-5 years. Employing the decomposition approach, a study concluded that maternal factors improve WAZ among children aged 0-5 years in India [27]. In other studies also, the role of mother's education in explaining the inequalities in underweight was highlighted in Indian settings [29,35]. Of the environmental factors, defecation remained the highest contributor to the inequality in underweight with nearly 17%, during both the time-periods. Open defecation is a problem among masses in India, and recently the Indian Government has done well to tackle the issue [6]. On the base of the decomposition approach, a study conducted in India confirmed that open defecation positively influences the inequality in underweight [35]. A study conducted in five Indian states; Uttar Pradesh, Karnataka, Madhya Pradesh, Rajasthan, and West Bengal, found that children from open defecation free (ODF) districts were less underweight than children in non-ODF districts [37], thus confirming the importance of open defecation in understanding the inequality in underweight. The result found that healthseeking factors explained nearly one-fourth (27.7%) of the inequality in underweight in 2005-06, but the same declined to nearly 14.4% in 2015-16. The decline in the contribution of the place of delivery in explaining the inequality signifies the equitable access of institutional deliveries in urban areas among poor and non-poor. A study found that in India, the utilization of institutional delivery care has increased from 43% in 2004 to 83% in 2014 [21]. This increase in institutional delivery had helped in shrinking the gap of inequality in underweight, as explained by place of delivery.
Wasting is defined as -> 2 SD weight-for-height among children. It is a measure of current nutrition status, also known as acute undernutrition. The inequality in wasting among children has decreased, but the prevalence of wasting has increased. Around the world, wasting had reduced at a plodding pace over the last 40 years, but in countries like India and Sri Lanka, wasting has increased [39]. Maternal factors (44.4%) were the highest contributory factors in explaining the inequality in wasting among children during 2005-06, but the contribution of maternal factors in inequality in wasting had declined significantly to 22.2% in 2015-16. In previous studies, maternal factors were deemed as the contributory factors in explaining the inequalities in wasting among children aged below five years [23]. Nearly half of the inequality in wasting among children had been explained by open defecation (46.9%) in 2015-16, citing the importance of sanitation. Since the weight-for-height is a measure of current nutritional status and is likely to be affected by small illnesses episodes and infectious diseases [23], the contribution of defecation in inequality in wasting among children is highly visible. Unavailability of toilet facilities within households promotes open defecation, and in turn, it facilitates water-borne diseases, which negatively affect the current health and nutrition status among children. Open defecation reinforces undernutrition (wasting) via increased susceptibility to infections and through reducing immunity [31].

Strengths and limitations
One of the strengths of the study is the considerable sample size for both the survey periods. This study explores the inequalities in stunting, underweight, and stunting among urban poor by employing survey data, which is 10 years apart, thus giving a clear picture of decadal change in economic inequality in stunting, underweight, and wasting among urban children. Decomposition allows our understanding of various factors  became the most significant contributory factors in explaining the rich-poor inequality in wasting among children during 2015-16. Third, more than 85% of the inequality in stunting, underweight, and wasting among urban children was explained by maternal factors, environmental factors, and health-seeking factors. Fourth, growth faltering among children reaches its peak by the end of 2 years of age. Rich-poor inequality in malnutrition calls out for the expedite policy interventions aimed at poverty reduction in urban areas. There is also a need to provide mass education regarding nutrition and health, along with focussing on the correlates that aim at improving mother's education. Environmental factors also promoted the inequality in stunting, wasting, and underweight in urban children; there is a need to focus the enigma of open defecation in India. After the launch of Swachh Bharat Mission in 2014, the government of India has focused a lot on building individual household latrines [6]. Despite having access to latrines, people prefer to defecate in open [3,8,18,32]. There is a need to promote the use of latrines among Indian masses. There is a need to bring a change in behavioural aspects and preferences of people to promote latrine use. Information, Education, and Communication (IEC) activities are encouraged to clamp down the economic inequality in stunting,