Determinants of Nutritional Status among Children under 5 years in India: A Cross-sectional Analysis of the National Family Health Survey 2019–2021

Abstract

Objective

This study aims at understanding the important correlates of child undernutrition in India using a quantile regression approach.

Design

Cross-sectional study.

Data Source

The study utilizes the secondary data from the latest round of the National Family Health Survey (2019–2021) to assess the individual, maternal and household-level determinants of under-nutrition among Indian children aged 0–5 years.

Participants

The unit of analysis is children under the age of 5 years and the sample size is 224, 218.

Outcome Variable

The outcome measures of this study are stunting, wasting and underweight.

Results

Results from our study show that birth order and size of the child at birth have a statistically significant association with stunting, wasting and underweight. While currently breastfeeding is negatively associated with HAZ scores (below 50th quantile), maternal height is positively associated with HAZ, WAZ and WHZ scores in all quantiles. Mother’s age, body mass index and education level have significantly stronger association with HAZ and WHZ scores. Furthermore, children from richer and richest households are likely to have higher HAZ, WAZ, and WHZ scores compared to the children from poorest households. For instance, the WHZ scores increased from 0.15 in the 10th quantile among the poorest to 0.43 among the richest at the same quantile. Moreover, at the 90th percentile, this coefficient increased from 0.11 among the poorest to 0.46 among the richest.

Conclusion

It is crucial to acknowledge that to have healthy babies, we must have healthy mothers. Thus, interventions to address child nutrition in India should also focus on maternal health, education, increasing the women’s agency and women empowerment. The findings of the study suggest that concerted efforts addressing the manifold determinants of children’s nutritional status from the government, non-governmental organisations and civil societies will ensure bolstered nutritional outcomes for children in India.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The study uses nationally representative data, which provides a true picture of a child’s nutritional status.

• The use of quantile regression enables examination of covariate associations across the full distribution of nutritional z-scores, rather than at the mean alone.

• This is useful for identifying risk factors that operate differentially among the most severely malnourished children.

• This is a cross-sectional study; thus, no causal association should be drawn.

Introduction

Despite being treatable and preventable, child undernutrition is one of the biggest threats to public health globally.¹ It is the leading cause of about one-third of child deaths globally and has deleterious socioeconomic and health-based impacts throughout their life² course as well as across generations.³ However, despite impressive economic growth in the last two decades, the reduction in child undernutrition has been only about 14 percentage points or 26%.⁴ Even in places where there has been a notable decline in child undernutrition, the presence of large and graded inequalities looms as a threat.⁵ Recent evidence suggests that despite all the progress made over the years, the burden of child undernutrition is still unacceptably high in India, and its abetment needs a detailed understanding of all the risk factors.⁶ As per studies, the consequences of child stunting, wasting and underweight are both immediate and long-term, including increased risk of morbidity and mortality, cognitive impairment, increased risk of infections, higher risk of developing diabetes, hypertension, dyslipidaemia and adverse maternal and reproductive health outcomes in adulthood.¹ Other factors that have a significant bearing on child undernutrition in low- and middle-income countries are lack of household sanitation,⁷ low body mass index (BMI) among mothers,⁸ less parental education⁹ and lack of bargaining power of women within households.¹⁰ Recent studies suggest that, along with the factors mentioned earlier, household ownership of assets, household size, age at marriage and antenatal care are important predictors of the nutritional status of children and explain significant spatial variation in child undernutrition in India.¹¹

In previous literature, researchers have dichotomised continuous response variables for analysis, but this method ignores the within-category information. There have been extensive discussions in literature on the disadvantages of dichotomising predictor and response variables.12,14 The most common drawbacks include loss of information, leading to a decline in the statistical power of measuring the relationship between a variable and the outcome measure.¹² Additionally, dichotomisation increases the odds of a positive outcome being a false positive. Third, it may also cause underestimation of between-group variation.¹⁴ Linear regression does not provide optimum results in cases where there are extreme outliers in the dataset.¹⁵ Finally, the assumption that many ‘established’ risk factors linearly exert their effect is being challenged by emerging evidence of non-linear effects.¹⁶ Thus, to overcome such issues, researchers have suggested the use of quantile regression, which does not need the dichotomisation of response variables and paints a more reliable picture of the association between explanatory variables and the response variable. Furthermore, it is also useful when the distribution of the response variable is not normal.¹⁷

This study is an attempt to understand the association between a set of predictor variables on child nutrition indicators—stunting using the height-for-age z-scores (HAZ), underweight using weight-for-age z-scores (WAZ) and wasting using weight-for-height z-scores (WHZ). This identification of critical risk factors of child undernutrition will help formulate strategies and plan programmes to achieve the set of Sustainable Development Goals (SDGs). Furthermore, the study uses a quantile regression model as an in-depth statistical model to identify the risk factors for chronic and acute child undernutrition and aid in policy formulation targeted at their reduction, as used in previous studies.^{18 19}

Data and Methodology

Data Source

This study uses data from the fifth round of the Indian Demographic and Health Survey, commonly known as the National Family and Health Survey (NFHS-5). It is a nationally representative survey conducted during the year 2019–2021 under the stewardship of the Ministry of Health and Family Welfare, Government of India. International Institute for Population Sciences, Mumbai, has been the nodal agency for surveying all the rounds of NFHS. This survey provides crucial information on reproductive and child health and family welfare-related indicators aligned with various SDGs covering all 28 states and 8 union territories. The survey is an integrated multi-level survey of households, men and women in their reproductive ages and biomarkers. NFHS-5 adopted a uniform stratified sampling design representative at national, state/UT and district levels. Districts were stratified into urban and rural areas. In rural strata, villages were selected as primary sampling units (PSUs), and in urban strata, census enumeration blocks were selected as PSUs. PSUs were sorted based on literacy rates or scheduled castes (SC)/scheduled tribes (ST) population before selection. In the second stage, 22 households were systematically selected from each PSU following household listing. Overall 30,456 PSUs were selected across 707 districts, with fieldwork completed in 30,198 PSUs. A detailed description of the sampling design and data collection methods is provided elsewhere.⁴

In the interviewed households,747,176 eligible women aged 15–49 years were identified to participate in the survey, of which 724,115 were successfully interviewed with a response rate of 97%. All ever-married women who were usual members of the selected households and those who spent the night before the survey in the selected households were eligible to be interviewed in the survey. NFHS-5 data collection was conducted in two phases (2019–2020 and 2020–2021), with interruption due to the COVID-19 pandemic. Although anthropometric measurements followed standardised protocols, disruptions in health services and household access during the pandemic may have influenced child health outcomes. Information on certain covariates, particularly birth weight and maternal anaemia status, contained missing observations primarily due to recall limitations or non-availability of recorded measurements. Children with missing anthropometric measurements are excluded from the regression analyses. Online supplemental appendix table S1 presents the number and proportion of missing observations for all study variables. To evaluate the potential impact of missing data primarily observed in covariates such as birth weight and maternal anaemia, we conducted an empirical assessment of the missingness mechanism. A multivariable logistic regression is performed to predict the probability of an observation being excluded from the final analytical sample due to missingness. The analysis revealed that the missingness was significantly associated with fully observed sociodemographic characteristics, specifically the child’s age, maternal educational attainment and household wealth quintile (p<0.01). Because these structural and demographic predictors, along with the complex survey design elements, are comprehensively adjusted for in all primary quantile regression models, our analysis effectively conditions on the primary factors driving the missing data. Consequently, under the missing at random assumption, this conditioning minimises potential bias, ensuring the robustness of our complete case analysis. Nutritional status was assessed for children under the age of 5 years, that is, births preceding 5 years of the survey (n=224,218), of which 73,072 out of 206,025 were stunted; 37,553 out of 201,687 were wasted and 65,043 out of 210,524 were underweight.

Description of the Variables

Dependent Variable

To assess nutritional status, stunting, wasting and underweight were employed as the dependent variables in this study. The anthropometric data on height, weight and age collected in the survey enabled us to measure the nutritional status of young children. Weight was measured with an electronic SECA 874 flat scale, while height was measured with a SECA 213 stadiometer. Children younger than age 24 months were measured lying down (recumbent length using a Seca 417 infantometer), while standing height was measured for the older children. The z scores for the calculation of stunting, wasting and underweight were defined using WHO growth standards. Children whose HAZ score was below −2 SD from the median of the reference population were considered short for their age (stunted). Children whose WHZ score was below −2 SD from the median of the reference population were considered wasted. Children whose WAZ score was below −2 SD from the median of the reference population were classified as underweight. The variables stunting, wasting and underweight were dichotomised to understand the prevalence of stunting, wasting and underweight in the sample. Further, the corresponding HAZ, WHZ and WAZ scores (continuous) for stunted,wasted and underweight were used for analysis.

Independent Variables

The study investigates whether the following socioeconomic and demographic characteristics of mothers and children are associated with the risk of stunting, wasting and being underweight. Independent variables were selected based on the literature, the conceptual framework used in this study and data availability within this dataset (NFHS-5). These variables have been widely used in earlier studies to capture key sociodemographic, household and health-related factors influencing the nutritional status of children.

The child characteristics included sex of the child (male or female), age of the child (less than 6 months, 6–11 months, 12–23 months and 2–5 years), birth order (1, 2, 3 and 4 and above), size of the child at birth (large, average, small), weight of the child at birth (low birth weight ‘less than 2500 g’, normal birth weight ‘2500 g and more’ and not weighted), Birth Interval (1st birth,<24 months, 24 months or higher), anaemia (not anaemic and anaemic), had a recent illness such as fever, cough or diarrhoea in past 2 weeks, Breastfeeding (never, ever and current). Birth weight was recorded from either a written record or the mother’s report. The size of the child at birth was self-reported based on the mother’s estimate of the baby’s size being very large, larger than average, average, smaller than average or very small. Baby size was categorised to be ‘large’ if the mother reported the baby’s size as very large/larger than average, ‘average’ and ‘small’ if the reported size was smaller than average/very small.

The mother’s characteristics included mother’s age at birth (=<19 years, 20–24, 25–29 and 35+), educational attainment (no education, primary, secondary and higher), body mass index (underweight, normal, overweight and obese),²⁰ mother’s height (<145 cm, 145–149 cm, 150–154 cm and ≥155 cm),²¹ anaemia status, delivery by c-section (yes and no), place of delivery (home, public facility and private facility).

The household characteristics included place of residence (rural and urban), drinking water source (improved, unimproved), wealth index (categorised as poorest, poorer, middle, richer, richest), self-reported caste based on ethnicity (SC, ST, other backward castes (OBC) and others), religion (Hindu, Muslim, others) and region (north, central, east, north-east, west and south) based on their geographic location. The wealth index was calculated using principal component analysis 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 and then the distribution was divided into five equal categories, each with 20% of the population.

Statistical Analysis

Descriptive statistics were used to understand the sample distribution and to find the preliminary results. Bivariate analysis was used to investigate the prevalence of children’s nutritional status across a number of sociodemographic variables. Chi-square (χ²) tests were used to assess the significance of association between categorical sociodemographic variables and dichotomised nutritional outcomes. Independent variables found to be significantly associated with the dependent variable were considered for further analysis. Dichotomised indicators were used to estimate the prevalence of undernutrition, while continuous z-scores (HAZ, WAZ, WHZ) were employed in quantile regression analyses to preserve distributional information and assess covariates effects across different quantiles. Further, quantile regression ws used to identify the factors affecting stunting, wasting and underweight by dividing the population into various quantiles (10th, 25th, 50th, 75th, 90th) instead of a population average score.²² Since the majority of the undernutrition indicators lie in the tail ends of the distribution, this regression technique allows the explanatory variable to vary across the different distributions of these undernutrition indicators.

The quantile function is given by

$${\displaystyle {\displaystyle Q_Y(\tau /X)=F_Y^{-1}(\tau /X)=\mathrm{i}\mathrm{n}\mathrm{f}\left\{\mathrm{y}:F_Y(\mathrm{y}/\mathrm{X})\ge \tau \right\}};\mathrm{w}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{e}\tau \in [0,1]}$$

And y is a real-valued random variable with a cumulative distribution function $F_Y(\mathrm{y}/\mathrm{X})=P(Y\le \mathrm{y}/\mathrm{X})$.

The parametric quantile regression model can be described as $Q_Y(\tau /X)={\beta }_{0\tau }+{\beta }_{1\tau }{\mathrm{x}}_{1\mathrm{i}}+\dots +{\beta }_{\mathrm{k}\tau }{\mathrm{x}}_{\mathrm{k}\mathrm{i}}\mathrm{\forall }\mathrm{i}\in \{1,2,\dots ,\mathrm{n}\}$ where ${\beta }_{0\tau },{\beta }_{1\tau },{\beta }_{\mathrm{k}\tau }$ are the quantile coefficients that may depend on τ.

The outputs obtained from the quantile regression are presented in coefficient terms. The distribution of the z-scores for the three anthropometric indices has been shown using boxplots. All statistical analyses were conducted using Stata V.16 and R software. Appropriate sampling weights have been applied to adjust for differences in probability of selection and interview between cases in a sample to produce nationally representative estimates. Sampling weights are provided in the NFHS data to adjust for selection probabilities to be used for further analysis. Both descriptive and regression analyses used weighted data. Survey design elements such as sampling weights, PSUs and stratification variables were incorporated using survey (svy) commands. The standard sample weight (v005) was applied after normalising it by dividing by 1 000 000. In Stata, the complex survey design was specified using clusters (v021) as PSUs and sample design strata (v023). The pre- and post-weighted demographic table has been included as online supplemental appendix table S2. SEs are adjusted for clustering at the PSU level. Multicollinearity among independent variables was assessed using variance inflation factors, where no collinearity was observed.

Given the exploratory nature of this study and the large number of statistical tests performed across multiple quantiles and outcome variables, results with p<0.001 should be interpreted with greater confidence, while borderline significant findings (p<0.05 and p<0.10) should be interpreted with caution as potentially subject to type I error. No formal multiple-testing correction was applied, consistent with the descriptive and hypothesis-generating intent of the analysis.

Patient and Public Involvement

As the study used the nationally representative NFHS 2019–2021 data, no patients or the public were involved in this research. The data are freely accessible through the web link (https://dhsprogram.com/data/dataset/India_Standard-DHS_2020.cfm?flag=1).

Results

Characteristics of the Study Population

The distribution of the z-scores for the three anthropometric indices is positively skewed, as shown by the coefficient of skewness of the z-scores and the boxplot in figure 1.

Figure 1. Boxplot of the z-scores of three anthropometric indexes. HAZ, height-for-age z-scores; WHZ, weight-for-height z-scores; WAZ, weight-for-age z-scores; .

Table 1 displays the percentage distribution for the study variables of interest. Overall, around one-third (36%) of the study sample is stunted, 19% are wasted and 32% are underweight. One-half of the study sample are of second and third birth order and 39% are first order birth . Around 10% of the children have ‘small’ size at birth and almost three-fifths (75%) have normal birth weight. The majority of the births that took place in the last 5 years occurred to women aged 20–24 years (44%) and around 12% of total births preceding 5 years of the survey occurred to teenage mothers. Approximately one-fifth of the births occurred by c-section, two-thirds of mothers had a height above 149 cm and the majority delivered at a health facility (89%).

Table 1. Background Characteristics of the Study Population, National Family and Health Survey (NFHS-5) (2019–2021), India.

Variable	Label	No.	%	Total
Child stunted	No	132 953	64.5	206 025
	Yes	73 072	35.5
Child wasted	No	164 134	80.8	201 687
	Yes	37 553	19.2
Child underweight	No	145 481	67.9	210 524
	Yes	65 043	32.1
Sex of child	Male	115 825	51.8	224 218
	Female	108 393	48.2
Child age	<6 months	23 093	10.2	224 218
	6–11 months	21 732	9.8
	12–23 months	43 436	19.5
	2–5 years	135 957	60.5
Birth order	1	85 828	39.1	224 218
	2	74 171	33.9
	3	35 082	15.2
	≥4	29 137	11.7
Size of child at birth	Large	40 533	19.0	221 288
	Average	158 816	70.6
	Small	21 939	10.4
Birth weight	Low	34 226	16.2	224 218
	Normal	168 869	75.2
	Not weighted	21 123	8.6
Birth interval	1st birth	86 456	39.5	224 218
	<24	35 704	16.3
	≥24	102 058	44.3
Child anaemia	Not anaemic	113 096	61.2	183 855
	Anaemic	70 759	38.8
Breastfeeding	Never	13 764	7.3	180 008
	Ever	59 225	34.0
	Current	107 019	58.7
Recent illness	No	175 948	77.5	224 218
	Yes	48 270	22.5
Mother age at birth	<19	25 182	12.4	224 218
	20–24	95 445	44.1
	25–29	67 002	29.3
	30+	36 589	14.2
BMI	Underweight	38 955	19.9	203 265
	Normal	128 032	61.0
	Overweight	28 334	14.7
	Obese	7944	4.5
Educational level	No education	48 527	21.0	224 218
	Primary	28 729	12.2
	Secondary	115 857	50.9
	Higher	31 105	15.9
Mother’s height	<145 cm	25 259	12.0	218 557
	145–149 cm	56 531	25.9
	150–154 cm	73 928	33.7
	≥155 cm	62 839	28.4
Mother anaemia	Not anaemic	144 360	66.7	215 992
	Anaemic	71 632	33.3
Place of delivery	Home	29 933	11.0	223 742
	Public health facility	145 047	62.1
	Private health facility	48 762	26.9
Delivery by c-section	No	180 726	78.3	224 218
	Yes	43 492	21.7
Place of residence	Urban	45 884	26.9	224 218
	Rural	178 334	73.1
Source of drinking water	Unimproved	14 718	4.2	213 592
	Improved	198 874	95.8
Wealth index	Poorest	60 222	24.2	224 218
	Poorer	52 184	21.6
	Middle	43 592	19.6
	Richer	37 968	18.6
	Richest	30 252	16.1
Caste	SC	45 723	23.1	224 218
	ST	45 308	10.0
	OBC	85 782	43.5
	Others	47 405	23.4
Religion	Hindu	164 319	79.3	224 218
	Muslim	32 397	16.3
	Others	27 502	4.4
Region	North	41 775	13.5	224 218
	Central	57 460	27.5
	East	43 271	26.0
	North-East	33 199	3.7
	West	19 951	12.6
	South	28 562	16.8

OBC, other backward class; SC, scheduled caste; ST, scheduled tribe.

Prevalence of Child Undernutrition according to Individual, Maternal and Household Characteristics

Table 2 displays the prevalence of stunting, wasting and underweight among children by sociodemographic factors. The prevalence of stunting, wasting and underweight is highest among children with following characteristics: children aged 12–23 months (for underweight: 2-5 years), 4+ birth order, ‘small’ size of child at birth, low birth weight and having never breastfed and currently breastfeeding. Findings indicate that the prevalence of stunting, wasting and undernutrition is marginally higher for males than females and among anaemic children. The prevalence of all three indicators of undernutrition showed an increasing pattern with higher birth orders.

Table 2. Prevalence of child malnutrition (stunting, wasting and underweight) by background characteristics, National Family and Health Survey (NFHS-5) (2019–2021), India.

Variable	Child stunted	Child wasted	Child underweight
Sex of child
Male	36.2	20.0	32.9
Female	34.6	18.4	31.2
Child age
<6 months	24.3	26.9	28.5
6–11 months	24.7	23.3	26.4
12–23 months	39.8	19.6	31.1
2–5 years	37.6	17.3	34.0
Birth order
1	31.4	18.3	28.5
2	34.7	19.3	31.4
3	39.6	20.4	36.4
≥4	45.5	20.4	40.3
Size of child at birth
Large	34.2	18.6	30.3
Average	34.8	18.8	31.2
Small	41.5	22.8	40.7
Birth weight
Low	42.9	23.2	42.4
Normal	32.7	18.2	29.0
Not weighted	45.5	20.2	40.1
Birth interval
1st birth	31.5	18.4	28.6
<24	43.1	19.4	37.9
≥24	36.2	19.9	33.0
Child anaemia
Not anaemic	33.4	17.9	30.2
Anaemic	42.0	19.6	36.3
Breastfeeding
Never	36.5	18.1	31.8
Ever	32.9	16.8	29.7
Current	34.3	22.2	32.0
Recent illness	ns
No	35.4	19.0	31.8
Yes	35.6	20.1	33.1
Mother’s age at birth		ns
<19	38.7	19.4	35.2
20–24	36.1	19.1	32.2
25–29	33.5	19.2	30.7
30+	34.8	19.5	31.7
BMI
Underweight	43.1	24.2	42.9
Normal	34.8	19.7	31.6
Overweight	27.8	13.5	21.9
Obese	23.9	11.2	17.1
Educational level
No education	46.4	21.5	42.1
Primary	41.6	19.9	37.1
Secondary	33.3	18.9	30.2
Higher	22.9	16.8	20.8
Mother height (cm)
<145	53.5	21.0	46.5
145–149	42.4	20.7	37.7
150–154	32.9	19.2	30.1
≥155	24.6	17.1	23.1
Mother anaemia
Not anaemic	34.0	19.0	30.8
Anaemic	38.4	19.7	34.7
Place of delivery
Home	46.2	20.4	41.1
Public health facility	36.7	19.7	33.4
Private health facility	28.0	17.6	25.1
Delivery by c-section
No	37.5	19.7	34.1
Yes	28.1	17.3	24.9
Place of residence
Urban	30.2	18.5	27.3
Rural	37.4	19.5	33.8
Source of drinking water
Unimproved	36.3	22.3	35.8
Improved	35.6	19.1	32.0
Wealth index
Poorest	46.1	22.5	43.1
Poorer	39.7	19.9	35.7
Middle	34.4	18.3	30.3
Richer	28.2	17.6	25.4
Richest	22.9	16.1	20.0
Caste
SC	39.4	19.6	35.0
ST	40.2	23.1	39.4
OBC	35.0	19.0	31.5
Others	30.4	17.6	27.1
Religion
Hindu	35.5	19.2	32.2
Muslim	36.8	19.9	32.7
Others	30.9	16.3	26.8
Region
North	29.6	14.6	24.0
Central	38.3	17.7	32.2
East	38.7	21.6	37.1
North-East	35.2	19.2	29.5
West	36.6	25.3	37.2
South	29.7	17.2	27.2
Total	35.5	19.2	32.1

BMI, body mass index; OBC, other backward class; SC, scheduled caste; ST, scheduled tribe.

The prevalence of child undernutrition was higher for the following maternal characteristics:

Higher prevalence of stunting and underweight are observed for children born to mothers aged less than 19 years at the time of delivery, underweight mothers, without any formal schooling, short maternal height (less than 145 cm) and anaemic mothers. With the increase in the mother’s educational attainment, the prevalence of all these nutrition indicators decreased. For instance, 46%, 22% and 42% of children born to mothers without any formal education are stunted, wasted and underweight, whereas 23%, 17% and 21% of children born to women with higher and above levels of education are stunted, wasted and underweight, respectively.

The prevalence of undernutrition is higher for the children from households in rural residence, unimproved drinking water sources, the poorest wealth quintiles, STs and individuals from the Muslim religion. Results show that a smaller percentage of children are stunted, wasted and underweight in richer wealth quintile households than in households with poorer wealth quintile. Highest levels of undernutrition are observed in Eastern India.

Tables35 provide the quantile regression results highlighting the varying association between covariates and the different quantiles (0.1, 0.25, 0.5, 0.75 and 0.90) of HAZ, WHZ and WAZ score distribution of under-5 children in India. It is observed that female sex is positively and significantly associated with HAZ meaning higher z-scores compared with compared to male child (β=0.09***CI 0.07 to 0.1 at 50th quantile), WHZ (β=0.09***CI 0.07 to 0.12 at 25th quantile) and WAZ (β=0.12***CI 0.09 to 0.14 at 10th quantile) scores at every quantiles (tables 3–5). At the 10th quantile, females have 0.12 units higher HAZ score (β=0.12***; 95% CI 0.09 to 0.16), and this advantage remains consistent at the 25th (β=0.11***; 95% CI 0.10 to 0.13), 50th (β=0.09***; 95% CI 0.07 to 0.10), 75th (β=0.10***; 95% CI 0.07 to 0.12) and 90th (β=0.10***; 95% CI 0.06 to 0.13) quantiles. Moreover, at most quantiles, child age is negatively associated with HAZ, WHZ and WAZ scores (tables35). Compared with children aged less than 6 months (reference group), older children exhibit significantly lower HAZ scores, particularly from 12 months onward. For instance, children aged 12–23 months have a 1.07 unit lower HAZ at the 50th quantile (β=−1.07***; 95% CI−1.64 to −0.50) and a 0.84 unit lower HAZ at the 75th quantile (β=−0.84***; 95% CI −1.31 to −0.37). Similarly, children aged 2–5 years shows significantly reduction in HAZ score by 1.12 units (95% CI −1.68 to −0.56) at 50th quantile and 1.09 units (95% CI −1.55 to −0.62) reduction at the 75th quantile. With the increase in the quantiles, the coefficient value increased but this was not statistically significant. Children aged 6–11 months have lower WHZ at the 10th quantile (β=−0.68*; 95% CI −1.45 to 0.09). Similarly, children aged 2–5 years show a marginal decline at the 75th quantile (β=−0.67*; 95% CI −1.40 to 0.06). At the 10th quantile, children aged 6–11 months have 1.01 units lower WAZ (β=−1.01**; 95% CI −1.85 to −0.16), those aged 12–23 months had 1.19 units lower WAZ (β=−1.19***; 95% CI −2.03 to −0.34), and children aged 2–5 years had 1.25 units lower WAZ (β=−1.25***; 95% CI −2.10 to −0.40). At the 25th quantile, all older age groups show significant declines in WAZ: −0.45*** (6–11 months), −0.68*** (12–23 months) and −0.86*** (2–5 years). At the 50th quantile, children aged 12–23 months (β=−0.56***; 95% CI −0.99 to −0.12) and 2–5 years (β=−0.80***; 95% CI −1.24 to −0.36) have significantly lower WAZ. At the 75th quantile, only children aged 2–5 years show a marginal significant decline (β=−0.69*; 95% CI −1.25 to −0.13) (table 5). Birth order and size of the child at birth is found to have a statistically significant negative association with HAZ, WHZ and WAZ scores in all quantiles of distribution (except for WAH score birth order at 10th and 25th quantiles for WAH scores) and the coefficients increases from 10th to 75th quantiles and reduces at 90th quantiles in reference to first-order birth and ‘large’ birth size. Compared with first-born children, higher birth order is associated with significantly lower WHZ scores at the middle and upper quantiles. At the 50th quantile, second-born (β=−0.22***; 95% CI −0.38 to −0.05) and third-born children (β=−0.25***; 95% CI −0.41 to −0.09) have shows significantly reduced WHZ scores. Similar declines are observed at the 75th quantile for second-born (β=−0.21**; 95% CI −0.40 to −0.02) and third-born children (β=−0.25**; 95% CI −0.44 to −0.05). At the 90th quantile, the reduction is even stronger, particularly for third-born children (β=−0.38***; 95% CI −0.59 to −0.17). It also reveal that children are more prone to be nutritionally vulnerable with higher order birth (example: 2nd to 3rd or 4th order birth; except for WAH scores) and if children are of ‘small’ size at birth than children of ‘large’ birth size. Birth weight is significantly positively associated with undernutrition indicators. Those with normal birth weight have significantly higher z-scores or simply are at a lower risk of being malnourished than those children with very low birth weight (tables35). For instance, compared with low birth weight children, those with normal birth weight have significantly higher WAZ across all quantiles i.e., 0.31*** (10th), 0.30*** (25th), 0.29*** (50th), 0.27*** (75th) and 0.20*** (90th) with p<0.001. Children with unrecorded birth weight also show increased WAZ, ranging from 0.23*** (10th) to 0.27*** (90th).

Table 3. Adjusted multivariable quantile regression results (95% CI) for height-for-age z-score (HAZ), National Family and Health Survey (NFHS-5) (2019–2021), India.

		Quantiles for height-for-age z-scores (HAZ)
		10	25	50	75	90
Sex of child	Male (Ref.)
	Female	0.12*** (0.09 to 0.16)	0.11*** (0.1 to 0.13)	0.09*** (0.07 to 0.1)	0.1*** (0.07 to 0.12)	0.1*** (0.06 to 0.13)
Child age	<6 months (Ref.)
	6−11 months	−0.58 (−2.34 to 1.19)	0.09 (−0.72 to 0.9)	−0.31 (−0.88 to 0.26)	−0.04 (−0.51 to 0.44)	0.04 (−1.78 to 1.86)
	12−23 months	−1 (−2.77 to 0.76)	−0.54 (−1.36 to 0.28)	−1.07*** (−1.64 to −0.5)	−0.84*** (−1.31 to −0.37)	−0.68 (−2.5 to 1.13)
	2−5 years	−0.72 (−2.5 to 1.05)	−0.44 (−1.25 to 0.38)	−1.12*** (−1.68 to −0.56)	−1.09*** (−1.55 to −0.62)	−1.24 (−3.06 to 0.58)
Birth order	1 (Ref.)
	2	0.01 (−0.33 to 0.35)	−0.19** (−0.34 to −0.03)	−0.17* (−0.34 to 0)	−0.36*** (−0.56 to −0.17)	−0.38** (−0.76 to −0.01)
	3	−0.01 (−0.35 to 0.32)	−0.27*** (−0.43 to −0.1)	−0.26*** (−0.42 to −0.09)	−0.44*** (−0.64 to −0.24)	−0.51*** (−0.85 to −0.16)
	≥4	−0.11 (−0.45 to 0.23)	−0.38*** (−0.54 to −0.21)	−0.38*** (−0.55 to −0.21)	−0.56*** (−0.76 to −0.36)	−0.6*** (−0.95 to −0.24)
Size of child at birth	Large (Ref.)
	Average	−0.02 (−0.06 to 0.02)	−0.05*** (−0.07 to −0.03)	−0.06*** (−0.08 to −0.04)	−0.06*** (−0.1 to −0.02)	−0.01 (−0.08 to 0.06)
	Small	−0.05* (−0.1 to 0.01)	−0.11*** (−0.14 to −0.08)	−0.15*** (−0.18 to −0.11)	−0.17*** (−0.22 to −0.13)	−0.13*** (−0.23 to −0.04)
Birth weight	Low (Ref.)
	Normal	0.26*** (0.21 to 0.3)	0.25*** (0.22 to 0.27)	0.27*** (0.24 to 0.3)	0.24*** (0.2 to 0.27)	0.2*** (0.13 to 0.27)
	Not weighted	0.13*** (0.04 to 0.21)	0.17*** (0.1 to 0.23)	0.24*** (0.2 to 0.27)	0.27*** (0.18 to 0.36)	0.38*** (0.28 to 0.49)
Birth interval	1st birth (Ref.)
	<24	−0.23 (−0.57 to 0.11)	−0.06 (−0.22 to 0.09)	−0.1 (−0.27 to 0.07)	0.03 (−0.16 to 0.23)	0.02 (−0.38 to 0.43)
	≥24	−0.05 (−0.38 to 0.28)	0.09 (−0.07 to 0.25)	0.05 (−0.11 to 0.21)	0.22** (0.02 to 0.42)	0.29 (−0.08 to 0.66)
Child anaemia	Not anaemic (Ref.)
	Anaemic	−0.28*** (−0.32 to −0.25)	−0.24*** (−0.26 to −0.23)	−0.2*** (−0.22 to −0.19)	−0.2*** (−0.23 to −0.16)	−0.21*** (−0.24 to −0.18)
Breastfeeding	Never (Ref.)
	Ever	0.11*** (0.05 to 0.17)	0.06*** (0.04 to 0.09)	0.04*** (0.01 to 0.07)	0.01 (−0.05 to 0.06)	−0.08* (−0.17 to 0.01)
	Current	−0.08*** (−0.14 to −0.03)	−0.07*** (−0.1 to −0.04)	−0.03* (−0.06 to 0)	0 (−0.05 to 0.04)	0.06 (−0.04 to 0.17)
Recent illness	No (Ref.)
	Yes	0.03 (−0.01 to 0.06)	0.02 (−0.02 to 0.05)	−0.02** (−0.04 to 0)	−0.06*** (−0.09 to −0.03)	−0.16*** (−0.22 to −0.1)
Mother’s age at birth	<19 (Ref.)
	20−24	0.06** (0 to 0.12)	0.06*** (0.02 to 0.11)	0.07*** (0.04 to 0.1)	0.04* (0 to 0.09)	0.07* (0 to 0.15)
	25−29	0.11*** (0.05 to 0.16)	0.15*** (0.1 to 0.2)	0.16*** (0.12 to 0.2)	0.15*** (0.1 to 0.19)	0.23*** (0.14 to 0.32)
	30+	0.12*** (0.07 to 0.17)	0.17*** (0.12 to 0.22)	0.22*** (0.18 to 0.26)	0.24*** (0.18 to 0.29)	0.29*** (0.21 to 0.38)
BMI	Underweight (Ref.)
	Normal	0.13*** (0.08 to 0.17)	0.16*** (0.13 to 0.18)	0.23*** (0.21 to 0.25)	0.33*** (0.3 to 0.36)	0.54*** (0.49 to 0.6)
	Overweight	0.29*** (0.23 to 0.35)	0.3*** (0.26 to 0.35)	0.32*** (0.29 to 0.36)	0.36*** (0.33 to 0.4)	0.45*** (0.37 to 0.54)
	Obese	0.38*** (0.31 to 0.45)	0.38*** (0.32 to 0.43)	0.38*** (0.33 to 0.44)	0.43*** (0.38 to 0.47)	0.37*** (0.26 to 0.47)
Educational level	No education (Ref.)
	Primary	0.14*** (0.09 to 0.19)	0.05** (0.01 to 0.1)	0.01 (−0.02 to 0.04)	−0.04* (−0.09 to 0)	−0.18*** (−0.25 to −0.11)
	Secondary	0.23*** (0.19 to 0.26)	0.16*** (0.12 to 0.2)	0.11*** (0.09 to 0.14)	0.04** (0 to 0.08)	−0.09*** (−0.16 to −0.02)
	Higher	0.28*** (0.22 to 0.35)	0.23*** (0.18 to 0.27)	0.21*** (0.18 to 0.25)	0.14*** (0.1 to 0.19)	−0.03 (−0.09 to 0.03)
Mother height	<145 cm (Ref.)
	145−149 cm	0.39*** (0.33 to 0.46)	0.4*** (0.36 to 0.44)	0.36*** (0.32 to 0.41)	0.36*** (0.33 to 0.39)	0.37*** (0.24 to 0.49)
	150−154 cm	0.64*** (0.56 to 0.73)	0.67*** (0.63 to 0.71)	0.63*** (0.59 to 0.68)	0.61*** (0.58 to 0.63)	0.62*** (0.5 to 0.74)
	≥155 cm	0.87*** (0.78 to 0.95)	0.94*** (0.9 to 0.98)	0.92*** (0.87 to 0.96)	0.89*** (0.85 to 0.92)	0.88*** (0.76 to 0.99)
Mother anaemia	Not anaemic (Ref.)
	Anaemic	−0.01 (−0.05 to 0.02)	−0.04*** (−0.07 to −0.02)	−0.05*** (−0.07 to −0.03)	−0.07*** (−0.09 to −0.04)	−0.12*** (−0.15 to −0.09)
Place of delivery	Home (Ref.)
	Public health facility	0.04 (−0.02 to 0.1)	0.04** (0.01 to 0.08)	0.05*** (0.01 to 0.08)	0.06** (0.01 to 0.1)	0.07 (−0.04 to 0.17)
	Private health facility	0.16*** (0.1 to 0.22)	0.16*** (0.11 to 0.2)	0.1*** (0.05 to 0.15)	0.04 (−0.01 to 0.09)	0.01 (−0.09 to 0.12)
Delivery by c-section	No (Ref.)
	Yes	0.07*** (0.02 to 0.12)	0.06*** (0.04 to 0.08)	0.05*** (0.03 to 0.07)	0.09*** (0.06 to 0.11)	0.08*** (0.03 to 0.14)
Place of residence	Urban (Ref.)
	Rural	0.06*** (0.03 to 0.09)	0.03** (0.01 to 0.06)	0 (−0.03 to 0.03)	−0.01 (−0.05 to 0.02)	−0.07 (−0.15 to 0.02)
Source of drinking water	Unimproved (Ref.)
	Improved	−0.05 (−0.12 to 0.01)	−0.06*** (−0.1 to −0.03)	−0.06*** (−0.11 to −0.02)	−0.05 (−0.11 to 0.01)	−0.02 (−0.11 to 0.07)
Wealth index	Poorest (Ref.)
	Poorer	0.13*** (0.09 to 0.16)	0.11*** (0.08 to 0.15)	0.09*** (0.05 to 0.12)	0.1*** (0.05 to 0.16)	0.11*** (0.03 to 0.18)
	Middle	0.21*** (0.16 to 0.26)	0.19*** (0.15 to 0.23)	0.17*** (0.13 to 0.21)	0.18*** (0.13 to 0.22)	0.11*** (0.04 to 0.19)
	Richer	0.29*** (0.22 to 0.35)	0.29*** (0.26 to 0.32)	0.27*** (0.24 to 0.3)	0.26*** (0.22 to 0.31)	0.21*** (0.13 to 0.3)
	Richest	0.35*** (0.28 to 0.42)	0.37*** (0.32 to 0.41)	0.35*** (0.3 to 0.39)	0.37*** (0.31 to 0.43)	0.33*** (0.23 to 0.43)
Caste	SC
	ST	−0.03 (−0.09 to 0.02)	0.01 (−0.03 to 0.04)	0.05*** (0.02 to 0.09)	0.11*** (0.06 to 0.16)	0.17*** (0.07 to 0.26)
	OBC	0.06*** (0.02 to 0.1)	0.04*** (0.02 to 0.07)	0.05*** (0.02 to 0.09)	0.05** (0 to 0.09)	0.03 (−0.03 to 0.09)
	Others	0.14*** (0.09 to 0.19)	0.15*** (0.11 to 0.18)	0.14*** (0.11 to 0.18)	0.14*** (0.09 to 0.2)	0.13*** (0.06 to 0.2)
Religion	Hindu (Ref.)
	Muslim	−0.22*** (−0.27 to −0.17)	−0.15*** (−0.17 to −0.12)	−0.07*** (−0.11 to −0.03)	0.05** (0.01 to 0.1)	0.33*** (0.27 to 0.38)
	Others	0.03 (−0.03 to 0.08)	0.01 (−0.02 to 0.05)	−0.01 (−0.05 to 0.03)	−0.02 (−0.06 to 0.01)	0.07 (−0.03 to 0.17)
Region	North (Ref.)
	Central	−0.08*** (−0.14 to −0.02)	−0.07*** (−0.1 to −0.05)	−0.11*** (−0.14 to −0.08)	−0.14*** (−0.18 to −0.11)	−0.17*** (−0.23 to −0.11)
	East	0.12*** (0.07 to 0.17)	0.09*** (0.05 to 0.13)	0.02 (−0.01 to 0.06)	−0.02 (−0.05 to 0.02)	−0.1*** (−0.15 to −0.04)
	North-East	0.01 (−0.05 to 0.08)	0.05*** (0.02 to 0.07)	0.06*** (0.02 to 0.1)	0.09*** (0.04 to 0.14)	0.2*** (0.1 to 0.29)
	West	−0.28*** (−0.36 to −0.2)	−0.23*** (−0.28 to −0.19)	−0.21*** (−0.25 to −0.17)	−0.08*** (−0.13 to −0.03)	0.25*** (0.13 to 0.37)
	South	−0.09** (−0.17 to −0.02)	−0.13*** (−0.17 to −0.1)	−0.2*** (−0.24 to −0.17)	−0.22*** (−0.27 to −0.18)	−0.1* (−0.2 to 0)

Models are fully adjusted multivariable quantile regression models including all child, maternal and household-level covariates simultaneously. Coefficients represent the change in the respective z-score at the specified quantile associated with each predictor category relative to the reference category (Ref.), holding all other covariates constant.

***p<0.001, **p<0.05, *p<0.10.

BMI, body mass index; OBC, other backward class; SC, scheduled caste; ST, scheduled tribe.

Table 5. Adjusted multivariable quantile regression results (95% CI) for weight-for-age z-scores (WAZ), National Family and Health Survey (NFHS-5) (2019–2021), India.

		Quantiles for weight-for-age Z-scores (WAZ)
		10	25	50	75	90
Sex of child	Male
	Female	0.12*** (0.09 to 0.14)	0.09*** (0.08 to 0.11)	0.08*** (0.07 to 0.1)	0.08*** (0.07 to 0.1)	0.09*** (0.06 to 0.12)
Child age	<6 months (Ref.)
	6–11 months	–1.01** (–1.85 to –0.16)	–0.45*** (–0.62 to –0.29)	–0.32 (–0.76 to 0.12)	–0.16 (–0.72 to 0.39)	–0.12 (–0.87 to 0.63)
	12–23 months	–1.19*** (–2.03 to –0.34)	–0.68*** (–0.85 to –0.51)	–0.56** (–0.99 to –0.12)	–0.41 (–0.98 to 0.16)	–0.38 (–1.11 to 0.36)
	2–5 years	–1.25*** (–2.1 to –0.4)	–0.86*** (–1.03 to –0.69)	–0.8*** (–1.24 to –0.36)	–0.69** (–1.25 to –0.13)	–0.69* (–1.43 to 0.04)
Birth order	1 (Ref.)
	2	–0.16 (–0.43 to 0.1)	–0.13* (–0.27 to 0)	–0.25*** (–0.34 to –0.15)	–0.18*** (–0.3 to –0.06)	–0.29** (–0.52 to –0.07)
	3	–0.21 (–0.47 to 0.05)	–0.2*** (–0.34 to –0.05)	–0.33*** (–0.42 to –0.23)	–0.26*** (–0.38 to –0.13)	–0.37*** (–0.58 to –0.16)
	≥4	–0.23* (–0.5 to 0.04)	–0.23*** (–0.37 to –0.09)	–0.37*** (–0.46 to –0.28)	–0.33*** (–0.45 to –0.21)	–0.48*** (–0.71 to –0.25)
Size of child at birth	Large (Ref.)
	Average	–0.03** (–0.06 to 0)	–0.05*** (–0.07 to –0.03)	–0.06*** (–0.07 to –0.05)	–0.07*** (–0.1 to –0.05)	–0.07*** (–0.1 to –0.04)
	Small	–0.11*** (–0.14 to –0.07)	–0.16*** (–0.19 to –0.14)	–0.16*** (–0.19 to –0.13)	–0.19*** (–0.23 to –0.16)	–0.18*** (–0.21 to –0.14)
Birth weight	Low (Ref.)
	Normal	0.31*** (0.28 to 0.34)	0.3*** (0.28 to 0.33)	0.29*** (0.27 to 0.31)	0.27*** (0.25 to 0.29)	0.2*** (0.17 to 0.24)
	Not weighted	0.23*** (0.18 to 0.28)	0.21*** (0.17 to 0.25)	0.24*** (0.21 to 0.28)	0.24*** (0.2 to 0.28)	0.27*** (0.19 to 0.35)
Birth interval	1st birth (Ref.)
	<24	–0.04 (–0.3 to 0.22)	–0.06 (–0.2 to 0.08)	0.06 (–0.03 to 0.16)	–0.03 (–0.15 to 0.1)	0.03 (–0.19 to 0.25)
	≥24	0.08 (–0.18 to 0.34)	0.04 (–0.09 to 0.18)	0.15*** (0.06 to 0.24)	0.06 (–0.06 to 0.18)	0.13 (–0.09 to 0.35)
Child anaemia	Not anaemic (Ref.)
	Anaemic	–0.19*** (–0.21 to 0.16)	–0.13*** (–0.15 to –0.12)	–0.12*** (–0.14 to –0.11)	–0.12*** (–0.14 to –0.1)	–0.15*** (–0.17 to –0.13)
Breastfeeding	Never (Ref.)
	Ever	0.05** (0.01 to 0.1)	0.02 (–0.01 to 0.06)	0 (–0.03 to 0.02)	–0.02 (–0.05 to 0.01)	–0.06*** (–0.11 to –0.02)
	Current	–0.09*** (–0.14 to –0.04)	–0.11*** (–0.14 to –0.08)	–0.12*** (–0.15 to –0.1)	–0.1*** (–0.12 to –0.07)	–0.1*** (–0.15 to –0.05)
Recent illness	No (Ref.)
	Yes	–0.01 (–0.04 to 0.02)	–0.04*** (–0.06 to –0.02)	–0.06*** (–0.07 to –0.04)	–0.08*** (–0.11 to –0.05)	–0.11*** (–0.13 to –0.08)
Mother’s age at birth	<19 (Ref.)
	20–24	0.06** (0.01 to 0.11)	0.03* (0 to 0.06)	0.05*** (0.03 to 0.08)	0.05*** (0.02 to 0.08)	0.06** (0.01 to 0.1)
	25–29	0.05* (0 to 0.11)	0.05*** (0.02 to 0.08)	0.09*** (0.06 to 0.12)	0.11*** (0.07 to 0.14)	0.14*** (0.08 to 0.19)
	30+	0.07*** (0.03 to 0.11)	0.07*** (0.04 to 0.11)	0.12*** (0.09 to 0.15)	0.16*** (0.13 to 0.2)	0.23*** (0.17 to 0.28)
BMI	Underweight (Ref.)
	Normal	0.14*** (0.12 to 0.17)	0.23*** (0.21 to 0.26)	0.31*** (0.29 to 0.33)	0.36*** (0.35 to 0.38)	0.43*** (0.39 to 0.46)
	Overweight	0.41*** (0.37 to 0.46)	0.48*** (0.46 to 0.49)	0.5*** (0.47 to 0.53)	0.55*** (0.52 to 0.57)	0.62*** (0.57 to 0.67)
	Obese	0.63*** (0.56 to 0.69)	0.66*** (0.6 to 0.71)	0.7*** (0.65 to 0.75)	0.72*** (0.67 to 0.77)	0.76*** (0.68 to 0.83)
Educational level	No education (Ref.)
	Primary	0.1*** (0.05 to 0.15)	0.05*** (0.02 to 0.08)	0 (–0.03 to 0.03)	0.01 (–0.03 to 0.04)	–0.01 (–0.06 to 0.04)
	Secondary	0.17*** (0.13 to 0.2)	0.13*** (0.1 to 0.15)	0.08*** (0.06 to 0.11)	0.05*** (0.03 to 0.07)	0.05* (0 to 0.1)
	Higher	0.25*** (0.19 to 0.31)	0.2*** (0.16 to 0.23)	0.18*** (0.15 to 0.21)	0.16*** (0.13 to 0.2)	0.14*** (0.08 to 0.19)
Mother height	<145 cm
	145–149 cm	0.21*** (0.16 to 0.26)	0.23*** (0.22 to 0.25)	0.25*** (0.23 to 0.28)	0.27*** (0.25 to 0.29)	0.21*** (0.16 to 0.26)
	150–154 cm	0.38*** (0.34 to 0.41)	0.41*** (0.39 to 0.42)	0.42*** (0.4 to 0.44)	0.44*** (0.41 to 0.46)	0.36*** (0.32 to 0.41)
	≥155 cm	0.51*** (0.46 to 0.56)	0.6*** (0.58 to 0.62)	0.64*** (0.61 to 0.66)	0.65*** (0.63 to 0.67)	0.58*** (0.53 to 0.63)
Mother anaemia	Not anaemic (Ref.)
	Anaemic	0.03* (0 to 0.06)	–0.01 (–0.02 to 0.01)	–0.02*** (–0.03 to –0.01)	–0.04*** (–0.05 to –0.02)	–0.04*** (–0.07 to –0.02)
Place of delivery	Home (Ref.)
	Public health facility	0.02 (–0.04 to 0.07)	0.01 (–0.02 to 0.04)	0 (–0.02 to 0.01)	0 (–0.05 to 0.04)	0 (–0.05 to 0.05)
	Private health facility	0.12*** (0.04 to 0.2)	0.09*** (0.05 to 0.13)	0.08*** (0.05 to 0.1)	0.04 (–0.01 to 0.09)	0.03 (–0.02 to 0.09)
Delivery by c-section	No (Ref.)
	Yes	0.07*** (0.03 to 0.11)	0.05*** (0.03 to 0.08)	0.07*** (0.05 to 0.09)	0.09*** (0.07 to 0.12)	0.1*** (0.07 to 0.13)
Place of residence	Urban (Ref.)
	Rural	0.08*** (0.04 to 0.12)	0.03** (0.01 to 0.06)	0.01 (–0.01 to 0.04)	0 (–0.03 to 0.02)	–0.01 (–0.05 to 0.02)
Source of drinking water	Unimproved (Ref.)
	Improved	0.01 (–0.05 to 0.06)	–0.02 (–0.06 to 0.02)	0 (–0.02 to 0.03)	0.02 (–0.01 to 0.04)	0.03* (0 to 0.06)
Wealth index	Poorest (Ref.)
	Poorer	0.15*** (0.11 to 0.18)	0.13*** (0.12 to 0.15)	0.11*** (0.09 to 0.13)	0.1*** (0.08 to 0.12)	0.11*** (0.07 to 0.15)
	Middle	0.21*** (0.16 to 0.26)	0.21*** (0.18 to 0.24)	0.19*** (0.16 to 0.21)	0.19*** (0.16 to 0.22)	0.2*** (0.16 to 0.23)
	Richer	0.29*** (0.24 to 0.33)	0.27*** (0.25 to 0.3)	0.26*** (0.24 to 0.28)	0.25*** (0.21 to 0.29)	0.28*** (0.23 to 0.33)
	Richest	0.34*** (0.28 to 0.41)	0.36*** (0.32 to 0.4)	0.33*** (0.3 to 0.36)	0.36*** (0.32 to 0.41)	0.38*** (0.33 to 0.44)
Caste	SC
	ST	–0.07*** (–0.11 to –0.03)	0.01 (–0.02 to 0.04)	0.04*** (0.01 to 0.06)	0.06*** (0.03 to 0.09)	0.12*** (0.09 to 0.16)
	OBC	0.02 (–0.02 to 0.06)	0.03* (0 to 0.06)	0.03** (0.01 to 0.05)	0.01 (–0.01 to 0.04)	–0.01 (–0.04 to 0.02)
	Others	0.08*** (0.02 to 0.14)	0.12*** (0.09 to 0.15)	0.13*** (0.1 to 0.15)	0.12*** (0.1 to 0.15)	0.14*** (0.1 to 0.17)
Religion	Hindu (Ref.)
	Muslim	–0.24*** (–0.29 to –0.18)	–0.16*** (–0.18 to –0.13)	–0.11*** (–0.13 to –0.09)	–0.04** (–0.07 to 0)	0.07*** (0.02 to 0.12)
	Others	0.16*** (0.12 to 0.2)	0.16*** (0.12 to 0.19)	0.17*** (0.14 to 0.21)	0.15*** (0.12 to 0.19)	0.12*** (0.06 to 0.17)
Region	North (Ref.)
	Central	–0.13*** (–0.18 to –0.08)	–0.1*** (–0.13 to –0.07)	–0.15*** (–0.17 to –0.13)	–0.2*** (–0.23 to –0.17)	–0.2*** (–0.23 to –0.16)
	East	–0.09*** (–0.15 to –0.04)	–0.11*** (–0.14 to –0.08)	–0.17*** (–0.2 to –0.15)	–0.22*** (–0.26 to –0.19)	–0.25*** (–0.29 to –0.21)
	North-East	0.11*** (0.07 to 0.15)	0.13*** (0.09 to 0.17)	0.09*** (0.05 to 0.12)	0.09*** (0.05 to 0.12)	0.09*** (0.04 to 0.14)
	West	–0.48*** (–0.55 to –0.4)	–0.46*** (–0.49 to –0.42)	–0.47*** (–0.51 to –0.44)	–0.47*** (–0.51 to –0.43)	–0.41*** (–0.47 to –0.34)
	South	–0.22*** (–0.27 to –0.17)	–0.21*** (–0.25 to –0.18)	–0.3*** (–0.33 to –0.27)	–0.37*** (–0.4 to –0.34)	–0.36*** (–0.42 to –0.31)

Models are fully adjusted multivariable quantile regression models including all child, maternal and household-level covariates simultaneously. Coefficients represent the change in the respective z-score at the specified quantile associated with each predictor category relative to the reference category (Ref.), holding all other covariates constant.

***p<0.001, **p<0.05 and *p<0.10.

BMI, body mass index; OBC, other backward class; SC, scheduled caste; ST, scheduled tribe.

Table 4. Adjusted multivariable quantile regression results (95% CI) for weight-for-height z-scores (WHZ), National Family and Health Survey (NFHS-5) (2019–2021), India.

		Quantiles for weight-for-height z-scores (WHZ)
		10	25	50	75	90
Sex of child	Male (Ref.)
	Female	0.13*** (0.09 to 0.17)	0.09*** (0.07 to 0.12)	0.06*** (0.04 to 0.07)	0.04*** (0.02 to 0.07)	0 (–0.03 to 0.03)
Child age	<6 months (Ref.)
	6–11 months	–0.68* (–1.45 to 0.09)	–0.21 (–0.86 to 0.44)	0.14 (–0.77 to 1.05)	–0.34 (–1.06 to 0.38)	–0.3 (–2.39 to 1.79)
	12–23 months	–0.6 (–1.37 to 0.18)	–0.09 (–0.75 to 0.56)	0.2 (–0.7 to 1.11)	–0.44 (–1.16 to 0.28)	–0.64 (–2.72 to 1.45)
	2–5 years	–0.54 (–1.31 to 0.24)	–0.13 (–0.78 to 0.52)	0.04 (–0.86 to 0.94)	–0.67* (–1.4 to 0.06)	–0.96 (–3.04 to 1.13)
Birth order	1 (Ref.)
	2	0.04 (–0.23 to 0.32)	0.03 (–0.13 to 0.18)	–0.22*** (–0.38 to –0.05)	–0.21** (–0.4 to –0.02)	–0.3*** (–0.51 to –0.08)
	3	0.06 (–0.22 to 0.34)	0.01 (–0.16 to 0.19)	–0.25*** (–0.41 to –0.09)	–0.25** (–0.44 to –0.05)	–0.38*** (–0.59 to –0.17)
	≥4	0.16 (–0.14 to 0.46)	0.07 (–0.1 to 0.24)	–0.22*** (–0.38 to –0.05)	–0.25** (–0.44 to –0.06)	–0.38*** (–0.59 to –0.16)
Size of child at birth	Large (Ref.)
	Average	–0.06*** (–0.1 to –0.02)	–0.07*** (–0.1 to –0.03)	–0.05*** (–0.07 to –0.02)	–0.05*** (–0.07 to –0.02)	–0.04** (–0.08 to –0.01)
	Small	–0.11*** (–0.17 to –0.05)	–0.15*** (–0.18 to –0.11)	–0.15*** (–0.19 to –0.11)	–0.15*** (–0.19 to –0.11)	–0.15*** (–0.23 to –0.06)
Birth weight	Low (Ref.)
	Normal	0.18*** (0.14 to 0.22)	0.21*** (0.18 to 0.24)	0.2*** (0.17 to 0.23)	0.19*** (0.16 to 0.21)	0.13*** (0.07 to 0.19)
	Not weighted	0.08* (0 to 0.17)	0.14*** (0.08 to 0.19)	0.17*** (0.13 to 0.21)	0.17*** (0.12 to 0.22)	0.17*** (0.11 to 0.24)
Birth interval	1st birth (Ref.)
	<24	–0.08 (–0.34 to 0.19)	–0.06 (–0.22 to 0.1)	0.14* (–0.02 to 0.31)	0.11 (–0.09 to 0.3)	0.12 (–0.1 to 0.34)
	≥24	–0.07 (–0.35 to 0.21)	–0.09 (–0.24 to 0.07)	0.14* (–0.03 to 0.3)	0.1 (–0.09 to 0.3)	0.13 (–0.07 to 0.34)
Child anaemia	Not anaemic (Ref.)
	Anaemic	–0.03 (–0.07 to 0.01)	0 (–0.02 to 0.02)	0.01 (–0.01 to 0.03)	0.03** (0.01 to 0.05)	0.03** (0.01 to 0.06)
Breastfeeding	Never (Ref.)
	Ever	0.06** (0 to 0.11)	0.02 (–0.02 to 0.06)	–0.04** (–0.08 to 0)	–0.05*** (–0.08 to –0.02)	–0.07** (–0.12 to –0.01)
	Current	–0.15*** (–0.22 to –0.09)	–0.13*** (–0.19 to –0.08)	–0.14*** (–0.18 to –0.1)	–0.1*** (–0.14 to –0.06)	–0.11*** (–0.18 to –0.05)
Recent illness	No (Ref.)
	Yes	0 (–0.04 to 0.04)	–0.04** (–0.07 to –0.01)	–0.06*** (–0.08 to –0.05)	–0.09*** (–0.11 to –0.07)	–0.1*** (–0.14 to –0.06)
Mother's age at birth	<19 (Ref.)
	20–24	–0.01 (–0.07 to 0.06)	–0.02 (–0.06 to 0.02)	0.02 (–0.02 to 0.05)	0.02 (–0.03 to 0.08)	0.02 (–0.04 to 0.09)
	25–29	–0.08* (–0.16 to 0)	–0.04* (–0.1 to 0.01)	0 (–0.04 to 0.05)	0.02 (–0.04 to 0.07)	0.04 (–0.03 to 0.11)
	30+	–0.11*** (–0.19 to –0.04)	–0.04 (–0.1 to 0.02)	0.03 (–0.02 to 0.07)	0.07** (0.01 to 0.13)	0.12*** (0.05 to 0.19)
BMI	Underweight (Ref.)
	Normal	0.01 (–0.05 to 0.06)	0.18*** (0.15 to 0.22)	0.27*** (0.25 to 0.29)	0.32*** (0.29 to 0.35)	0.38*** (0.33 to 0.43)
	Overweight	0.35*** (0.28 to 0.41)	0.43*** (0.4 to 0.45)	0.47*** (0.45 to 0.5)	0.49*** (0.46 to 0.53)	0.54*** (0.48 to 0.6)
	Obese	0.57*** (0.48 to 0.66)	0.62*** (0.56 to 0.69)	0.65*** (0.6 to 0.7)	0.7*** (0.65 to 0.75)	0.79*** (0.66 to 0.92)
Educational level	No education (Ref.)
	Primary	0.07** (0.01 to 0.12)	0.02 (–0.03 to 0.07)	0.01 (–0.02 to 0.03)	–0.02 (–0.06 to 0.02)	–0.04 (–0.1 to 0.01)
	Secondary	0.13*** (0.08 to 0.19)	0.05*** (0.03 to 0.08)	0.02 (–0.01 to 0.04)	0 (–0.02 to 0.03)	–0.04* (–0.08 to 0)
	Higher	0.15*** (0.08 to 0.22)	0.1*** (0.05 to 0.16)	0.09*** (0.05 to 0.12)	0.08*** (0.03 to 0.13)	0.05 (–0.03 to 0.13)
Mother height	<145 cm (Ref.)
	145–149 cm	0 (–0.06 to 0.05)	0.03 (–0.01 to 0.07)	0.03** (0 to 0.06)	0.04* (0 to 0.08)	0.01 (–0.06 to 0.07)
	150–154 cm	0.03 (–0.02 to 0.08)	0.06*** (0.02 to 0.1)	0.07*** (0.04 to 0.1)	0.07*** (0.03 to 0.1)	0.04 (–0.02 to 0.11)
	≥155 cm	0.08** (0.01 to 0.16)	0.14*** (0.1 to 0.18)	0.14*** (0.11 to 0.18)	0.14*** (0.1 to 0.19)	0.14*** (0.07 to 0.21)
Mother anaemia	Not anaemic (Ref.)
	Anaemic	0.07*** (0.03 to 0.11)	0.02 (–0.01 to 0.05)	0 (–0.02 to 0.02)	0 (–0.02 to 0.03)	0 (–0.04 to 0.03)
Place of delivery	Home (Ref.)
	Public health facility	–0.06 (–0.13 to 0.01)	–0.04* (–0.09 to 0)	–0.02 (–0.06 to 0.02)	0 (–0.03 to 0.03)	0.03 (–0.04 to 0.09)
	Private health facility	0.03 (–0.07 to 0.13)	0.02 (–0.03 to 0.08)	0.04 (–0.01 to 0.09)	0 (–0.04 to 0.04)	0.01 (–0.07 to 0.09)
Delivery by c-section	No (Ref.) (Ref.)
	Yes	0.08*** (0.03 to 0.13)	0.05*** (0.02 to 0.09)	0.05*** (0.03 to 0.08)	0.07*** (0.04 to 0.09)	0.1*** (0.06 to 0.14)
Place of residence	Urban (Ref.)
	Rural	0.08*** (0.03 to 0.13)	0.04*** (0.01 to 0.06)	0.02 (–0.01 to 0.05)	0 (–0.03 to 0.03)	–0.03 (–0.07 to 0.02)
Source of drinking water	Unimproved (Ref.)
	Improved	0.1*** (0.04 to 0.16)	0.04* (–0.01 to 0.09)	0.03 (–0.01 to 0.07)	0.03 (–0.01 to 0.07)	0.1*** (0.03 to 0.17)
Wealth index	Poorest (Ref.)
	Poorer	0.08*** (0.02 to 0.14)	0.1*** (0.07 to 0.13)	0.06*** (0.02 to 0.09)	0.05*** (0.02 to 0.08)	0.06*** (0.02 to 0.1)
	Middle	0.13*** (0.08 to 0.19)	0.15*** (0.11 to 0.19)	0.11*** (0.07 to 0.16)	0.1*** (0.07 to 0.13)	0.09*** (0.03 to 0.16)
	Richer	0.17*** (0.11 to 0.23)	0.14*** (0.11 to 0.18)	0.14*** (0.1 to 0.17)	0.12*** (0.1 to 0.15)	0.12*** (0.06 to 0.19)
	Richest	0.22*** (0.13 to 0.32)	0.2*** (0.14 to 0.25)	0.19*** (0.15 to 0.23)	0.18*** (0.14 to 0.23)	0.2*** (0.14 to 0.26)
Caste	SC
	ST	–0.09*** (–0.14 to –0.04)	–0.03* (–0.07 to 0)	0.01 (–0.03 to 0.05)	0.06*** (0.03 to 0.09)	0.11*** (0.06 to 0.16)
	OBC	0 (–0.05 to 0.04)	0 (–0.02 to 0.02)	0 (–0.03 to 0.03)	0 (–0.03 to 0.03)	–0.03 (–0.07 to 0.02)
	Others	0.06* (–0.01 to 0.13)	0.06*** (0.02 to 0.1)	0.05*** (0.02 to 0.08)	0.07*** (0.03 to 0.1)	0.12*** (0.06 to 0.17)
Religion	Hindu (Ref.)
	Muslim	–0.33*** (–0.39 to –0.28)	–0.18*** (–0.21 to –0.15)	–0.06*** (–0.08 to –0.03)	0.01 (–0.03 to 0.05)	0.1*** (0.05 to 0.15)
	Others	0.18*** (0.13 to 0.23)	0.2*** (0.16 to 0.23)	0.22*** (0.17 to 0.26)	0.18*** (0.15 to 0.22)	0.17*** (0.12 to 0.22)
Region	North (Ref.)
	Central	–0.07** (–0.13 to 0)	–0.1*** (–0.14 to –0.07)	–0.11*** (–0.14 to –0.07)	–0.14*** (–0.18 to –0.1)	–0.14*** (–0.18 to –0.1)
	East	–0.19*** (–0.26 to –0.12)	–0.26*** (–0.29 to –0.23)	–0.26*** (–0.29 to –0.23)	–0.28*** (–0.31 to –0.25)	–0.28*** (–0.32 to –0.24)
	North-East	0 (–0.06 to 0.06)	–0.01 (–0.04 to 0.02)	0.07*** (0.02 to 0.11)	0.11*** (0.06 to 0.16)	0.22*** (0.16 to 0.28)
	West	–0.62*** (–0.67 to –0.56)	–0.55*** (–0.59 to –0.51)	–0.48*** (–0.53 to –0.44)	–0.42*** (–0.46 to –0.37)	–0.27*** (–0.34 to –0.21)
	South	–0.17*** (–0.23 to –0.1)	–0.25*** (–0.29 to –0.2)	–0.28*** (–0.32 to –0.24)	–0.31*** (–0.36 to –0.25)	–0.29*** (–0.34 to –0.25)

Models are fully adjusted multivariable quantile regression models including all child, maternal and household-level covariates simultaneously. Coefficients represent the change in the respective z-score at the specified quantile associated with each predictor category relative to the reference category (Ref.), holding all other covariates constant.

***p<0.001, **p<0.05 and *p<0.10.

BMI, body mass index; OBC, other backward class; SC, scheduled caste; ST, scheduled tribe.

Moreover, the mother’s age, BMI and education level have significant positive association with the z-score of height-for-age and weight-for-age of the child. An increment in the mother’s age improved the z-score in all quantiles. For instance, as we move from normal BMI mothers to obese mothers, the HAZ score increases from 0.13*** and 0.33*** points at the 10th and 75th quantiles, respectively, to 0.38*** and 0.43*** in the obese category in the respective quantiles for overweight mothers. It is also observed that the children’s HAZ, WHZ and WAZ coefficient values increased when their mother’s education increases from lower education to higher education. It indicates that, as the mothers get more educated, their children get less vulnerable to stunting, wasting and underweight. Having primary education show small gains in the WAZ scores at the 10th (0.10***) and 25th (0.05***) quantiles compared to mother with no education. Likewise, secondary education increased WAZ from 0.17*** (10th) to 0.05*** (90th). Higher education showed the strongest association, with WAZ increasing by 0.25*** (10th), 0.20*** (25th), 0.18*** (50th), 0.16*** (75th) and 0.14*** (90th) points. Children born to anaemic mothers are less likely to be nutritionally secure (except for the WAH score). Furthurmore, moving from lowest to highest quantile, the negative association between mother’s anaemic status and HAZ and WAZ scores becomes sharp. Children of anaemic mothers have lower HAZ scores, particularly at higher quantiles. The decline was by −0.04*** points at the 25th, −0.05*** at the 50th, −0.07*** at the 75th and −0.12*** at the 90th quantile. The place of residence is positively related to z-score at the lower tail of the conditional distribution; however, the coefficient values decreased as we moved from the 50th to the 90th quantile of the distribution (tables35). A negative association is observed between improved water sources and the HAZ scores. Improved drinking water is associated with slightly lower HAZ at the 25th (β=−0.06***; 95% CI −0.10 to −0.03) and 50th (β=−0.06***; 95% CI −0.11 to −0.02) quantiles. Furthermore, children from richer and richest households are likely to have higher HAZ, WHZ and WAZ scores compared to the children from poorest households. It is observed that the contribution of wealth to the z-score gradually increases with the different tails of the distribution. For example, Children from poorer households have higher WAZ by 0.15 (10th) to 0.11 (90th) than children from poorest households. Those in the middle quintile have increased by 0.21 units (10th) to 0.20 units (90th). Children from richer households show gains by 0.29 units (in 10th quantile) to 0.28 units (in 90th quantile), while the richest households have the strongest association, with WAZ increases by 0.34 (10th), 0.36 (25th), 0.33 (50th), 0.36 (75th) and 0.38 (90th) units. As evident, caste is significantly negatively associated with HAZ, WHZ and WAZ scores (in the 25th to 90th quantiles) while Muslim religion is negatively associated with z scores in the lower quantiles. As we move to the higher quantiles, the coefficient values reduces and the positive association between Muslims and other religious groups increases. The quantile regression reveals the heterogeneous HAZ, WAZ and WHZ scores across different regions of the country.

The following characteristics are positively associated with stunting (HAZ): female child, birth weight, <12 months of breastfeeding, mother’s age at delivery, mother’s BMI, Secondary or higher level of education, wealth quintile and caste; and negatively associated with stunting (HAZ): child age, birth order, size of child at birth, anaemia and improved water source.

The following characteristics are positively associated with WAZ score: female sex, normal birth weight, higher mother’s age at delivery, mother’s BMI, improved water source, other religious groups; and negatively associated with WAZ score: high birth order, small size of child at birth, current breastfeeding and western, southern, central and eastern regions.

The following characteristics are positively associated with WHZ score: female sex, birth weight, mothers age at delivery, mother’s BMI, education, wealth quintile, caste, others religious group; and negatively associated with WHZ score: birth order, size of child at birth, breastfed for 1 year or more, anaemic mother, Muslim religion and all regions except for north-eastern region. The distribution of z-scores of anthropometric measures among children under age 5 years are presented in online supplemental appendix figure A1.

Also, coefficient plot for HAZ, WAZ and WHZ by various determinants are provided in the appendix (online supplemental appendix figures A2–A5).

Discussion

The present study focuses on the association between individual, maternal and household factors and different child under-nutrition indicators using the quantile regression approach.

Consistent with previous studies, the prevalence of undernutrition is found to be marginally higher among male children.^{23 24} This could be attributed to biology, with epidemiological studies placing male children at a higher risk of morbidity as compared with their female counterparts.²⁵ Additionally, behavioural patterns such as favouritism towards female children in certain communities could also be one of the causes.²⁶

A strong negative and statistically significant association between birth order and child nutritional status is observed. This could be because of the high probability of higher-order births being unwanted. This results in reduced antenatal and postnatal care utilisation for subsequent births.²⁷ Additionally, studies have reported a decrease in the intra-household allocation of food and resources with an increasing number of child births in the household. As a result of the above, higher-order children might suffer from both resource constraints and health hazards. This relationship has been abundantly established in previous studies conducted in India and several other countries as well.^{28 29} This study showed a positive association between birth weight and WAZ, HAZ, WHZ scores. This finding is consistent with similar studies conducted in other parts of the world.²⁹

Our findings show that breastfeeding is positively associated with HAZ scores but is negatively associated with WAZ and WHZ scores. Several previous studies report a null association between wasting and breastfeeding practices.³⁰ The possible reasons behind this could be maternal recall bias and troubles in defining and measuring complex variables such as exclusive breastfeeding and duration of breastfeeding. The positive association of breastfeeding on child stunting, however, has been abundantly recorded.31,33 The breastmilk is known to boost a child’s immunity, reduce diarrheal episodes and other infectious diseases, and reduce exposure to other unsafe foods and dirty drinking water that are known risk factors for child stunting.^{31 32}

Our findings on maternal age show that an increase in maternal age increases the z-scores in all quantiles for all three nutrition indicators. This is in line with findings from previous studies.^{26 34} This could be because young mothers are themselves nutritionally vulnerable and their nutritional needs might interfere with those of the child. Additionally, food shared between the infant and the mother might not be adequate for the proper growth of the child. Furthermore, young mothers might not have proper knowledge and might be unaware of healthcare practices for children.³⁵ Our findings suggest that z-scores for children are higher when their mothers’ BMI is normal or above. This relationship has also been previously documented by several studies.³⁶ This stresses the trans-generational nature of adverse environmental outcomes, meaning that environmental conditions leading to adverse nutritional and health outcomes for mothers are passed on to the offspring.³⁷

Another important relationship that is reinforced by our findings is that between maternal education and children’s nutritional status. Our results show that maternal education level has a significantly positive association with child’s z-score of height-for-age and weight-for-age. Maternal education acts as a protective shield against all diseases and undernutrition indicators.^{24 35} This is because educated mothers are more aware of their children’s health. They are also more exposed to mass media and are likely to have better childcare knowledge and adopt practices that may benefit their children.³⁵

Our findings on anaemia show that maternal anaemia is negatively associated with child HAZ and WHZ scores. Low maternal iron levels lead to low haemoglobin levels in infants and this condition may persist for up to 1 year, thus causing anaemia in children.³⁸ Iron is essential for neural metabolism and functioning. Infants suffering from anaemia are at risk of facing development difficulties, including cognitive, social-emotional and adaptive functions.³⁹ Additionally, iron intake among mothers leads to a reduction in the prevalence of low-birth-weight babies.⁴⁰ Moreover, birth weight is a crucial determinant of stunting risk in children.⁴¹ Additionally, low birth weight infants are more likely to suffer from different comorbidities that might lead to stunting.⁴²

India has implemented a gamut of programme-based interventions to reduce undernutrition among infants and mothers over the last five decades. With the inception of Integrated Child Development Services (ICDS) and the pan-India implementation of the national Food Security Act, 2013 and the mid-day meal scheme, India has managed to buttress nutrition-based outcomes for women and children, especially in rural and remote areas of the country. Additional schemes targeted at improving children’s and mothers’ health include Anganwadi Services, Poshan Abhiyaan and the Pradhan Mantri Matru Vandana Yojana under the ICDS.

The prevalence of undernutrition being higher among children in rural areas has been endorsed by several previous studies.⁴³ In all such previous studies, this rural–urban disparity has been attributed to improved healthcare services and facilities and better food and dietary practices in urban areas.^{44 45} Our findings also suggest that water and sanitation have important ramifications for child nutrition, that is, risk of contamination of non-improved water sources, which exacerbates the risk of water-borne diseases such as diarrhoea.³⁵

Findings that children from richer households have better z scores than their counterparts from poorer households are in line with studies done in different countries.^{46 47} This is primarily because poorer households are food insecure and usually have a larger number of children. This makes it difficult to take adequate care of each child, leading to low dietary intake and, thus, nutritional deficiencies. Caste-based findings suggest that undernutrition prevalence was highest among the ST community. This finding has been endorsed by several previous studies.⁴⁸ The practice of untouchability hinders access to a basket of antenatal and postnatal health services to mothers and children belonging to such marginalised communities. Religion is also an important determinant of the nutritional status of children, as the findings of our study suggest. Higher fertility and low levels of education and awareness among some religious groups contribute to these differences.⁴⁸ Our findings on regional heterogeneity suggest that the prevalence of all the child undernutrition indicators was highest in Eastern India. Studies have reported clusters of high stunting, wasting and underweight in these states.⁴⁹ States in the eastern region have been prone to floods. Studies suggest that children, especially in rural settings, who are exposed to seasonal fluctuations such as droughts and floods in utero have a higher probability of being underweight.⁵⁰ Additionally, these states are also agriculturally not resilient enough and literature suggests that agriculture improves nutritional outcomes by providing nutritious, healthy and affordable foods.⁵¹ Agricultural production has important ramifications on child feeding practices, which, in turn, are associated with children’s nutritional status.⁵² Areas where agricultural production is diversified are expected to have better feeding practices and, hence, improved nutrition.⁵³

Although the study presents important policy implications, it has certain limitations. Birth size and certain maternal characteristics were self-reported and may be subject to recall bias or social desirability bias. The use of nationally representative NFHS data supports the generalisability of the findings to the Indian population. However, the cross-sectional study makes it difficult to establish causality between child nutrition scores and individual and household-level determinants. Additionally, since the data being used has been collected using a two-stage stratified sampling frame, multi-level regression could be employed to obtain optimal results.

Conclusion

Findings from our study suggest that individual, maternal and household level factors are strongly associated with the nutritional status of children. The factors that came out prominently in this study are the child’s birth weight, birth order, maternal education, maternal nutrition, household wealth, caste and religion. These findings stress the importance of cumulative efforts to tackle undernutrition among children in India. It is also crucial to acknowledge that to have healthy babies, we must have healthy mothers. Thus, interventions to address child nutrition in India should also focus on maternal health, education and increasing the agency and empowerment of women. In addition to the above, it will also be beneficial if nutrition-related education is introduced in the curriculum of schools and colleges. Thus, concerted efforts addressing the manifold determinants of children’s nutritional status are required to bolster nutrition outcomes for children in India.

Data availability statement

Data are available in a public, open access repository.