Nutritional status and concomitant factors of stunting among pre-school children in Malda, India: A micro-level study using a multilevel approach

The latest edition of the Global Nutrition Report 2018 addressed that India tops in the number of stunted children (46.6 million), followed by Nigeria (13.9 million) and Pakistan (10.7 million). These three countries are home to almost half of the world’s stunted (47.2%) children [1]. A recent study published in Lancet found that malnutrition was the main cause of death among children below 5 years in every Indian state in 2017, accounting for 68.2% of the total deaths among children below 5 years of age [2]. The latest round of the Indian family and health survey shows that about 38, 21, and 36% of children are stunted, wasted and underweight, respectively, at the national level in 2015–16 [3]. While in Malda district, about 38, 23, and 37% of children aged under 5 years are stunted, wasted and underweight, respectively [3].

The Government of India has tried to combat the perseverance of malnutrition through several large-scale programmes and schemes which could not reach the level of expectations mostly. Integrated Child Development Services (ICDS) is one of the most inclusive programmes for addressing child malnutrition and child well-being and development, implemented in 1975 by the Ministry of Women and Child Development, Government of India and supported by the UNICEF [4]. ICDS is operating through Anganwadi Centres (AWCs). An AWC is operated by an Anganwadi Worker (AWW) and an Anganwadi Helper (AWH). Brief detail on the ICDS is provided in the supplementary file. Recently, another flagship programme is known as National Nutrition Mission (NNM), also known as POSHAN (Prime Minister’s Overarching Scheme for Holistic Nutrition) Abhiyaan has been implemented by the Indian Government in 2017. It targets to reduce stunting by two percentages annually, under-nutrition by two percentages annually, anaemia by two percentages annually among young children, adolescent girls, and women, and reduction of low birth weight by two percentages annually. Further, it sets the target ‘Mission 25 by 2022’ for stunting to reduce it from 38.4 to 25% by 2022 in India [5]. On the other hand, sustainable development goals (SDG) recommended the target to reduce stunting is a reduction by 50% by 2030 [6].

The prevalence of malnutrition varies substantially across the states as well as by the districts of the country [3]. There are several studies on the nutritional status of children in India, especially at the national and state levels. However, the in-depth research on nutritional status, particularly among pre-school children at the micro-level, is little known. Besides these, India is a vast country, having a great diversity in its geography, religion, caste, culture, food habits, and climate. Therefore, inferences based on a particular micro-level study setting may not be applicable to another particular micro-level area for policy interventions. In this regard, a review study based in India has recommended that the risk factors of malnutrition among children should be analysed by controlling diverse measures in a given set-up [7]. In addition to that, a recent study by Menon et al. (2018) demonstrated that inter-district variations in stunting among children under 5 years old in India are largely described by a large number of health, hygiene, economic and demographic factors [8]. Based on these observations, the present study attempts to know the level of nutritional status and to study this by various disaggregate levels, as well as to examine the risk factors of stunting among pre-school children aged 36–59 months in Malda which is a district of West Bengal in India. The findings of this study can be helpful for the local governing body to adopt proper interventions for reducing the prevalence of stunting and to reach the target ‘Mission 25 by 2022’ and also the target for stunting as embodied in SDGs.

In the previous studies, it has been found that individual child characteristics, e.g., child’s age, sex, birth interval, birth order, and birth weight, are the associated factors of stunting [9‐14]. Many studies found that children born with low birth weight are more prone to become stunted in their childhood [15‐21]. A multi-country analysis based in 137 developing countries found that preterm birth or short gestational age is the leading cause of stunting among children of 24–35 months old [22]. Investigations revealed that birth interval is also another leading risk factor of stunting among children in developing countries, where along with the increase in the birth interval, the probability of stunting decreases [23‐26]. In the Indian context, researchers have also demonstrated that birth interval is an important factor of stunting among children [27‐31]. A cohort study based in Kenya exhibited that a longer duration of breastfeeding is positively associated with child growth [32]. Further, it has been found that mothers’ characteristics, e.g., maternal age at birth, mother’s education, maternal nutritional and occupational status affect stunting among children. A study based in India showed that maternal age is positively associated with stunting among children [28], while some studies showed that stunting is inversely related to maternal age [33, 34]. Many studies have documented that with the increase of maternal education, the chances of stunting decrease among children [14, 24, 35‐38]. Studies also reported that mothers’ employment is related to poor nutritional status among children [39]. On the contrary, a study found that the employment status of mothers does not matter for stunting, but the type of remuneration does matter for the same, in which children of unpaid workers are significantly more stunted than the paid workers [40]. It has also been observed that household characteristics, e.g. religion, caste or social groups [14, 35], wealth index, and the place of residence [41, 42], are also associated factors of stunting among children. Besides these, studies that evaluated the effects of ICDS at the national level in India show that there are no significant effects on reducing stunting among children who are accessing the services from ICDS [43‐45].

Table 1 presents the univariate descriptive statistics of 731 sampled children. About more than half of the study children are male. Around one-fifth of all children were born with < 2.5 kg weight at the time of birth, defined as low birth weight. About 61% of children breastfed the WHO recommended duration of 24 months and above [56]. Among all mothers or caregivers, homemaker or housewife is the dominant category of occupation. Nearly two-thirds of children are attending AWCs or receiving supplementary food from ICDS schemes. More than half of the children belong to Muslim households. Among caste or social groups, OBC is the most prevalent caste in the study area, followed by the general category, SC and ST. About 86% of children belong to rural areas.

Figure 1 depicts a comparison between NFHS-4 and primary survey estimation of nutritional status among children of 36–59 months old. Estimation from the primary field survey data shows that about 40, 19, and 38% of children aged 36–59 months are stunted, wasted and underweight, respectively, in Malda. While the estimation from the NFHS-4 demonstrates that the prevalence of stunting, wasting, and underweight are about six percentage points higher, six percentage points lower and two percentage points higher, respectively, in Malda than the estimation from the primary data of Malda. However, the estimation from NFHS-4 for India shows that the prevalence of stunting, wasting, and underweight are almost equal to the estimation from primary survey data for Malda district.

The bivariate estimations of childhood stunting by background characteristics of children in the Malda district are also presented in Table 1. Female children are almost six percentages points more stunted than male children. By and large, the prevalence of stunting increases from the first birth order to the higher order of birth. Children who born with < 36 months of the preceding birth interval are about 15% points more stunted than those born with ≥36 months preceding birth spacing. Similarly, children who born with < 2.5 kg weight are 16% points more stunted than those born with 2.5+ kg weight. The prevalence of stunting decreases with the increasing duration of breastfeeding. With the increase in mothers’ educational level, the proportion of stunted children decreases. The prevalence of stunting among children of mothers of agricultural workers or manual labourers is the higher one in occupation factor. Unexpectedly, children who are accessing the AWCs almost 18% points more stunted as compared to those who do not access the same. The prevalence of childhood stunting is highest among the ST population in comparison to other groups. Children belonging to Hindu households are about 15% points less stunted than children of Muslim or other households. Children from rural areas are almost 25% points more stunted as compared to the children living in urban areas.

Table 2 presents the results of multivariate multilevel logistic regression showing the effects of fixed-effects parameters as well as random-effects parameters, e.g., individual child and community-level effects on stunting among children in the collected sample. A lower Wald chi-square statistic (< 0.001) for all the conditional models implies that childhood stunting varied significantly by the individual, mother, household, and community-level factors. In the random-effect part, the p-values of the log-likelihood ratio test (LR) versus logistic regression vary from 0.001 to 0.024, which also indicate that the variance in child stunting differs significantly by an individual child or household level and AWC or community level.

Model 6 is a final and comprehensive model. After controlling for all the explanatory variables in this model, it is found that birth interval, low birth weight, duration of breastfeeding, mother’s age at birth, mother’s education, and occupation are the significant risk factors of stunting. Among them, low birth weight (OR 2.22, 95% CI: 1.44–3.41) and bidi worker (OR 1.92, 95% CI: 1.18–3.12) are the most influencing factors of stunting, which show a consistent effect throughout the models. Besides these, the mother’s age of > 25 years at birth (OR 1.98, 95% CI: 1.04–3.78) is also another important risk factor of stunting. Nevertheless, the proportion of stunted children among the mothers who are agricultural workers or manual labourers and children belonging to the ST community are extremely high, as shown by bivariate estimations, but these are not significant risk factors of stunting in the full regression model. On the other hand, the value of the community level ICC is 0.14, which indicates the correlation among the children within the same community is 0.14 in the case of child stunting. In other words, about 14% variation in stunting lies at AWC or community level, and the larger part of the variation in stunting lies at the child or household level, as shown in the random part of the model. Further, the highest degree of changes in the random part of the model is observed after including the mother’s characteristics in the third model.

The prevalence of stunting in the study area is 40% among pre-school children aged 36–59 months, which is a very high prevalence as per the WHO’s cut-off values (≥40%) for public health significance [49]. If the WHO’s cut-off value (≥40%) is considered as the very high level of prevalence, several categories of background characteristics of children have shown a very high prevalence of stunting in Malda district, e.g., female child, 3+ birth order, < 36 birth intervals, low birth weight, < 25 months of duration of breastfeeding, > 25 years of mother’s age at birth, illiterate or poorly educated mothers, mothers’ occupation as agricultural workers or manual labourers and bidi workers, children attending AWCs, children belong to Muslim or other minor religious groups, children belong to OBC and ST categories, children living in poor as well as middle-class family, and children living in rural areas. However, the highest proportion of stunted children is found among the most deprived ST community, almost seven children out of 10 children are stunted in this community; followed by illiterate mothers, mothers’ occupation as agricultural workers or manual labourers, low birth weight, and poor households, where more than half of children are stunted. Here, it should be noted that after having several flagship programmes and schemes implemented by the Indian Government to eradicate malnutrition among children, still, this region, as well as India, have a very high level of prevalence of stunted children.

According to Arjan de Wagt, is the nutrition chief, UNICEF India, and his colleagues, 2019, huge potential interventions are still in the operations under the nutrition and health schemes and programmes, but the services of these programmes and schemes are not reaching the children and women at a satisfactory level. Therefore, implementation constrictions need to be appropriately addressed by that high-impact nutritional interventions can reach the maximum level, and that should be followed with the Coverage, Continuity, Intensity, and Quality (C2IQ) principle [57] in the study area as well as in India. Unexpectedly, this study found that the proportion of stunting is higher (47.3%) among the children who access the services from the AWCs or ICDS (one of the most inclusive programmes for addressing child under-nutrition) than those who are not accessing the same. Although, it might be justified by saying that children of socio-economically disadvantaged families are mostly the beneficiaries of AWCs or ICDS schemes.

The results of the multilevel analysis revealed that preceding birth interval, low birth weight, duration of breastfeeding, mother’s age at birth, mother’s education, and occupation are the only associated risk factors of stunting in this study. Low birth weight is the strongest predictor of stunting, followed by mothers’ occupation and mothers’ age at birth. Association between low birth weight and stunting has been shown by several studies [15, 18, 19, 21]. As per medical concerns, intra-uterine growth retardation (IUGR) is the main cause of low birth weight in developing countries [58]. Children who suffer from IUGR during mothers’ pregnancy are effectually born malnourished. Maternal malnutrition, low maternal weight, low maternal stature during conception, and insufficient maternal weight gain during pregnancy all together are attributed to almost half of the cases of IUGR in developing countries [20, 58]. Further, anaemia and iron deficit also accompany IUGR and subsequently results in low birth weight [59, 60]. Children of mothers engaged in bidi making are at higher risk of child stunting that requires further investigation. Usually, most of the bidi workers live in poverty, mostly below the poverty line that might be the underlying cause of the higher risk of child stunting among them. In addition to that, normally bidi making women are less careful towards their children while they sit and make the bidis; perhaps this is another reason for the higher risk of child stunting among them.

Mother’s age at birth is also another important risk factor of stunting, where children of higher maternal age are more likely to be stunted than their counterparts. A similar association is observed in another study based in India [28]. Duration of breastfeeding is also another important risk factor of stunting in this research. Children with a lower duration of breastfeeding are more likely to be stunted as compared to children with a higher duration of breastfeeding. This finding is also consistent with other studies [32, 61]. Children with lower preceding birth intervals also have higher chances of being stunted as compared to those of higher birth intervals. A similar observation is reported by other studies as well [23‐25]. A birth interval may affect under-nutrition among children by its relationship with premature birth and low weight at birth. If a woman becomes pregnant too soon after birth, the mother does not get adequate time to fill their depleted body fats and nutrients during the pregnancy and breastfeeding, which may lead to premature birth and low weight at birth. Thus, a higher birth interval allows the mothers to recover and become healthy for subsequent pregnancy. Mothers with a higher birth interval can provide their children with the required nutrition for growth and a resilient immune system that reduces the chances of childhood stunting [62]. Although the effect of the ICDS schemes operates through AWCs is not a significant factor of stunting after controlling for household’s characteristics and place of residence, it does have a positive association with stunting if children are accessing the services in the case of child’s, and mother’s characteristics are only taking into consideration in the regression analysis. Likewise, the same observation is noticed in the case of the mother’s occupation, especially for agricultural workers or manual labourers. Conceivably, the multicollinearity between mother’s occupation, children accessing the services from ICDS, household and community characteristics viz. religion, caste groups, wealth status, and place of residence, is the reason for altering the effect of mother’s occupation and ICDS schemes. Therefore, further studies are needed to investigate the risk factors of stunting within the community level as well. Besides these, a higher level of changes in the level variances of the multilevel model (two-levels) was reflected after adjusting for mother’s characteristics along with the child’s characteristics. Thus, it can be said that a mother’s characteristics also play a crucial role in her child’s physical growth. On the other hand, among total variation of the two-levels model in child stunting, about 14% variation was due to the heterogeneity between the communities or AWCs. So, it can be believed that a community or particular AWC itself also has an influence on stunting among children in the study area.

An attempt has been made in the present research to explore the level of nutritional status, to study this by disaggregate levels, and to examine the risk factors of stunting among pre-school children aged 36–59 months in Malda. Accordingly, a primary field survey was conducted to study the same in 2018. It was found that Malda district suffers from the burden of a high-level prevalence of child stunting. This district has already got priority from the policy-making body, and the Indian Government implemented a flagship program NNM (National Nutrition Mission) in the first phase (2017–18) to eradicate malnutrition in this district. Before that, another flagship program, ICDS was implemented by the Indian Government during the 1990s and still in operation. Nevertheless, the prevalence of child stunting is at an alarming level. All the public health machinery from the grassroots level to a higher level, e.g., the local governing body to the State Government and Central Government are requested to monitor the implementing process and effectiveness of the programmes if the target ‘Mission 25 by 2022’ for stunting, that is still to be achieved. Further, special attention needs to be placed on the modifiable risk factors of child stunting as this study has found that preceding birth interval, low birth weight, duration of breastfeeding, mother’s education, and occupation are the associated risk factors of stunting in this district. Policy interventions should direct community health workers, e.g., ASHA (Accredited Social Health Activist), AWW (Anganwadi Worker), and ANM (Auxiliary Nurse Midwife), to encourage women as well as their male partners to increase birth interval using various family planning practices, provide extra care for preterm birth or low birth weight baby following WHO’s recommendation 2015 [63], increase educational awareness, that can help to reduce childhood stunting and thereby result in healthy childhood development. Mothers who are bidi workers and their children, as well as particular communities (AWCs), should be targeted groups as they need special attention because their children are at higher risk of being stunted. Socially most deprived ST community also needs special attention as the proportion of stunted children are highest among social groups and all other factors in this study setting. Further investigation is required as to why the children of mothers working as bidi workers are at higher risk of childhood stunting. Also, studies are needed to investigate the risk factors of stunting within the community level as well.