Association between anthropometric‐based and food‐based nutritional failure among children in India, 2015

Abstract

Inadequate dietary intake is a critical underlying determinant of child undernutrition. This study examined the association between anthropometric‐based and food‐based nutritional failure among children in India. We used the 2015–2016 National Nutrition Monitoring Bureau data where anthropometric outcomes and food intake were both measured for each child. We followed the World Health Organization child growth reference standards to define anthropometric failures (i.e., height‐for‐age z score < −2 SD for stunting, weight‐for‐age z score < −2 SD for underweight, and weight‐for‐height z score < −2 SD for wasting), and the Indian Council of Medical Research recommended dietary allowance (RDA) to define adequacy in intake of calorie, protein, and fat. We used descriptive and regression‐based assessments to test the association between the two indicators of nutritional failure and also computed the area under the receiver operating characteristic curve (AUC). The prevalence of stunting, underweight, and wasting was 28.6%, 24.3%, and 12.8%, respectively, whereas 78.2%, 27.4%, and 50.8% of the children had below RDA norms consumption of calorie, protein, and fat, respectively. We found weak‐to‐null correlation between anthropometric failures and food failures (Pearson correlation ranging from −0.013 to 0.147) and poor discriminatory accuracy (AUC < 0.62), suggesting that in the Indian context, anthropometric failures are not directly associated with food intake. This finding highlights the need for improving adequate intake of macronutrients and draws attention toward adopting a multifactorial approach to improve child nutrition in India. Poor food intake itself merits exclusive policy focus as it is an important nutrition and health concern.

Key messages.

• This study found very weak association and poor discriminatory accuracy between food‐based failures, as per calorie, protein, and fat intake, and anthropometric‐based failures among children in India.

• India continues to experience a high prevalence of inadequate dietary intake, particularly among the disadvantaged population.

• Explicit recognition of dietary intake as part of policy targets can raise socio‐political visibility of this important indicator of nutrition.

1. INTRODUCTION

Nutrition refers to the whole process of ingestion of food, digestion, absorption, assimilation, and utilization of various nutrients by an organism/human being for growth, development, and maintenance of bodily functions. The Oxford dictionary defines nutrition as “the process of providing or obtaining the food necessary for health and growth.” The Merriam‐Webster dictionary defines nutrition as the “the act or process of nourishing or being nourished; specifically: the sum of the processes by which an animal or plant takes in and utilizes food substances.” It also defines nutrition as “foods that are necessary for human nutrition.” The World Health Organization (WHO) describes nutrition as the “intake of food, considered in relation to the body's dietary needs.” These various definitions consistently associate nutrition with food intake. Regular and systematic assessment of food intake levels across various population age groups is, therefore, critical for understanding its distributional patterns and associations with nutritional status of the population.

In the context of low‐ and middle‐income countries, anthropometric failures among children (i.e., stunting, underweight, and wasting) are routinely monitored to understand the burden of nutritional deprivation (Caulfield, Richard, Rivera, et al., 2006). For instance, the National Nutrition Strategy of India and the National Nutrition Mission (POSHAN—Prime Minister's Overarching Scheme for Holistic Nourishment—Abhiyaan) emphasizes on tracking progress in reducing child anthropometric failure (NITI Aayog, 2017). Several important global health and development goals have accorded high priority to anthropometric outcomes as target indicators. In particular, the World Health Assembly Resolution 65.6 calls for achieving a 40% reduction in stunting by 2025 (WHA, 2012).

It is well acknowledged that anthropometric outcomes are multifactorial in nature and are influenced by a range of demographic, socio‐economic, environmental, and contextual factors (Corsi, Mejía‐Guevara, & Subramanian, 2016; Kim, Mejía‐Guevara, Corsi, Aguayo, & Subramanian, 2017; Subramanian, Mejía‐Guevara, & Krishna, 2016). The Global Nutrition Report 2017 highlights that there are important reciprocities between nutrition and Sustainable Development Goals (SDGs; Development Initiatives, 2017). In fact, simultaneous action on various SDGs is necessary as even the most direct interventions have limited efficacy in reducing the global burden of undernutrition (Development Initiatives, 2017). Although anthropometric failures encompass a wide range of determinants (Corsi et al., 2016; Corsi, Subramanyam, & Subramanian, 2011; Kim et al., 2017; Subramanian et al., 2016), inadequate food, in terms of both intake and composition, is invariably an important underlying determinant (Corsi et al., 2016; International Food Policy Research Institute, 2016; Kim et al., 2017). Nevertheless, in low‐income settings, a large proportion of the population is affected by food inadequacies, and thus, poor diet emerges as an important determinant of nutritional health.

Most maternal and child nutrition interventions focus on some version of food supplementation. For example, the Government of India has launched important food supplementation programmes such as Integrated Child Development Services, Midday Meal Scheme for school‐going children, Rajiv Gandhi Scheme for Empowerment of Adolescent Girls—“Sabla,” and Pradhan Mantri Matru Vandana Yojana for pregnant women and lactating mothers to improve nutritional health of the vulnerable populations. Moreover, legislations such as the National Food Security Act 2013 have a direct focus on food and nutrition entitlements to protect, support, and promote nutritional health of children and pregnant and lactating mothers.

Only a few studies have explored the association between food intake and anthropometric failures among children. For example, Corsi et al. (2016) found that failure to meet the minimum dietary diversity plays a statistically significant but only a moderate role in terms of magnitude, in increasing the risk of stunting and underweight in India. Menon, Bamezai, Subandoro, Ayoya, and Aguayo (2015) also found similar associations between dietary diversity and nutritional status of children. Conditional on other important risk factors, dietary diversity and feeding frequency shared similar weak‐to‐null relationships with anthropometric failures across other South Asian countries such as Afghanistan, Bangladesh, Nepal, and Pakistan (Kim et al., 2017). All these studies focus on dietary diversity as a proxy for dietary adequacy. Both indicators are correlated, and they are also sensitive toward the approach and the instrument used for information collection (Arimond & Ruel, 2004; Coates & Monteilh, 1997; Ruel, 2003). Therefore, it is presumably relevant for adopting alternative approaches that examine a direct association between dietary intake and anthropometric outcomes to shed further light on the challenges and concerns underlying the lack of association between these indicators in India.

Estimates from the National Sample Survey household consumer expenditure survey 2011–2012 of India reveal high level of deficiency in calorie, protein, and fat intake among poor households (NSSO, 2014). The population also experiences a lack of dietary diversity as 61% of the caloric intake is derived from cereals alone. Although the per capita calorie intake from cereals has declined between 1983 and 2004–2005 by 295 kcals in rural areas and 156 kcals in urban areas, there has been no commensurate shift in dietary adequacy and diversity (Deaton & Drèze, 2009). Given such concerns, it is important to not only undertake an assessment of dietary intakes in terms of macronutrients (calories, protein, and fat) but also to examine its association with outcome‐focused indicators such as anthropometric indicators. Deaton and Drèze (2009) also indicate that improvements in health environment through improving access of safe water and sanitation facilities can reduce diseases and infections, and this can be instrumental in mediating the effect of a given dietary intake on anthropometric status. These associations can further highlight whether dietary intake in isolation can be an effective factor to curb the multifactorial phenomenon of undernutrition in India.

This paper aims to present an analysis of the association between dietary intake and anthropometric outcomes among Indian children. We use data from the National Nutrition Monitoring Bureau (NNMB), which is the only source in India that provides direct measures of both food and anthropometric failure for each child. We perform descriptive and regression‐based assessments and use discriminatory measures to systematically test the association between dietary intake inadequacies and anthropometric failures that is stunting, wasting, and underweight.

2. METHODS

2.1. Data

The study is based on NNMB Urban Nutrition Study 2013–2015. The main objective of this cross‐sectional survey was to provide information on child undernutrition and dietary intake in urban India. The survey was conducted in 16 Indian states and union territories, namely, Andaman and Nicobar Islands, Andhra Pradesh, Assam, Bihar, Delhi, Gujarat, Karnataka, Kerala, Madhya Pradesh, Maharashtra, Odisha, Puducherry, Rajasthan, Tamil Nadu, Uttar Pradesh, and West Bengal. The NNMB survey adopted a multistage random sampling procedure. In the first stage, five cities were selected from each of the 16 NNMB states. Out of five cities, one was the capital city of the state, and the remaining four cities were randomly selected from a list of cities whose population was more than 100,000. In the second stage, 15 wards were randomly selected from each of the selected city. In the third stage, six census enumeration blocks were randomly selected from each of the selected ward. In the final stage, eight households were randomly selected from each census enumeration block. A total of 48 households were covered within each urban ward. All available individuals from the selected households were recruited for various investigations. Thus, a sample of 3,600 households per state (57,600 households overall) was covered in the survey. The data collected height and weight measurements for 12,162 children (below 5 years) and food and nutrient data for a subsample of 2,723 children. The final analytic sample for our complete case analysis included 1,816 children who had information available on the focal indicators of height, weight, and food and nutrient intakes.

2.2. Measures of nutritional failure

We considered two measures of nutritional failure among children: (a) anthropometric‐based measures and (b) calorie‐intake‐ or food‐intake‐based measures. Anthropometric information regarding height and weight of each child was measured by trained field investigators. Height was measured using anthropometric rod for all children of 2 years and above. If the child was less than 2 years, infantometer was used for their measurement of length. Child's height‐for‐age, weight‐for‐age, and weight‐for‐height were expressed in standard deviation units (z scores) from the median of the reference population (WHO, 2006). Children whose height‐for‐age z score was below minus two standard deviations (−2 SD) from the median of the reference population were considered as stunted. Children with weight‐for‐age z score < −2 SD and weight‐for‐height Z‐score < −2 SD were considered as underweight and wasted, respectively.

A 24‐hr dietary recall was conducted in one fourth of the households surveyed for anthropometric measurements in each state by a nutritionist who was trained and standardized in diet survey methods. The nutritionist interviewed the housewife or kitchen‐in‐charge about the type foods prepared and consumed in the previous day by all household members according to their meal patterns (i.e., early morning tea, breakfast, midmorning lunch, lunch, evening tea, dinner, and night drinks tea/coffee/milk/butter milk) Standardized tools were used to capture the quantity and volume of all consumptions. The dietary recall also included information on food consumption outside of the household. After assessing the raw quantity for each ingredient in each menu and its cooked volume, an individual raw equivalent was assessed, and nutrient composition was estimated for each individual by using the nutritive value of Indian foods and Indian food composition tables 2017 (Longvah, Anantan, Bhaskarachary, & Venkaiah, 2017). The average daily intake of different foods by individuals was calculated according to their age/sex, physiological status, and activity status. The mean intake of foods and median intake of various nutrients were compared with the suggested balanced diets provided in recommended dietary intake for Indians and nutrients as per the recommended dietary allowances (RDA) suggested by the ICMR Expert Committee (Gopalan et al., 1990; ICMR, 2010). The data on caloric intake (in kcals), protein intake (in gm), and fat intake (in gm) were categorized in two categories to reflect consumption levels that were below and above the RDAs for Indians (National Institute of Nutrition, 2011). Based on the dietary guidelines for Indian children, the RDA for caloric intake was 80 kcal/kg/day for children below age of 12 months, 1,060 kcal/day and 1,300 kcal/day for 12 to 35 months and 36 to 59 months children, respectively. The RDA for protein intake was 1.69 gm/kg/day for children below 12 months, 16.7 gm/day for 12 to 35 months, and 20.1 gm/day for 36 to 59 months olds. The RDA for fat intake was 19 gm/day for children younger than 12 months, 27 gm/day for 12 to 35 months old children, and 25 gm/day for 36 to 59 months old children.

2.3. Covariates

The analysis also included information on key socio‐economic and demographic variables including age and gender of the child as well as religion (Hindu, Muslim, others), social group (scheduled caste, schedule tribes, other backward classes, others), and per capita income of the household. The analysis also considered maternal characteristics such as mother's height, education, and occupation to adjust for its impact on the association between dietary intake and anthropometric failure. Maternal education was categorized as follows: no schooling, primary, secondary, higher secondary, and college education. Maternal height was categorized as follows: below 145, 145–149.9, 150–154.9, 155–159.9, and 160 cm and above. Lastly, maternal occupation was categorized as follows: labourers and cultivators, services and professionals, and business and homemakers.

2.4. Statistical analyses

In order to assess the association between anthropometric outcomes and dietary intake levels, we first conducted descriptive analysis for both continuous measures and categorical measures. Prevalence of stunting, underweight, and wasting and mean z scores for height‐for‐age, weight‐for‐age, and weight‐for‐height were calculated for children with adequate versus inadequate intake of total calorie, protein, and fat. We additionally applied logistic regression analysis to assess the strength of association between the two failures. For sensitivity analysis, we additionally adjusted for socio‐economic and demographic correlates. The results from all logistic regressions are presented in the form of odds ratio (OR) along with 95% confidence interval (CI). An OR is a measure of association between an exposure and an outcome. The OR represents the odds that an outcome will occur given a particular exposure, compared with the odds of the outcome occurring in the absence of that exposure. However, OR is a measure of average association that does not necessarily provide information on how well the predictor of interest actually discriminates cases with outcomes versus noncases (Merlo, Wagner, Ghith, & Leckie, 2016). Hence, we also present estimates on the basis of the measure of discriminatory accuracy. Specifically, we estimate the area under the receiver operating characteristic curve (AUC) that quantifies the accuracy of using individual‐level information alone for identifying those with the outcome (Merlo et al., 2016). The AUC is based on postestimation‐predicted probabilities obtained from the age–sex adjusted logistic regression models. The AUC captures the ability of the model to correctly categorize children with or without a given outcome as a function of the predicted probabilities. Specifically, for different binary classification thresholds of the predicted probabilities, the AUC plots the association between true positive fraction (TPF or sensitivity) on the y‐axis against the false positive fraction (FPF or 1‐specificity) on the x‐axis. The AUC ranges from 0.5 to 1 with higher values reflecting greater accuracy of the model to discriminate the outcome. For interpretation purposes, AUC of 1 reflects perfect discrimination, whereas AUC of 0.5 implies no predictive power. All analyses were performed in statistical software, Stata 15 (StataCorp, LP, 2007).

2.5. Ethics

The study protocol was approved by the National Institute of Nutrition (NIN)—Scientific Advisory Committee (Protocol number: CR10/I/2014). Written informed consent was obtained from head of the household and mother of under 5‐year‐old children. The study was reviewed and approved by the NIN Institutional Review Committee.

3. RESULTS

It can be observed that 28.6% of the children in our final analytic sample were stunted, 24.3% were underweight, and 12.8% were suffering from wasting (Table 1); 78% of the children had low calorie consumption (below RDA), 27% had low protein intake, and 51% had low fat intake (Table 2). Further, only 20% of the children had adequate consumption of all three macronutrients as per the RDA norms. There was a weak association between anthropometric failures and food failures in terms of RDA levels. It may be noted that prevalence of inadequate consumption of calorie, protein, and fat was equally high for children with and without anthropometric failures. Table 2 shows that 76% of nonstunted children and 77% of nonunderweight children had low calorie intake. Also, 83% of stunted and underweight children were reported to have low caloric intake. The levels of anthropometric failure were much lower than the proportion of children reporting food intake below the RDA norms.

Table 1.

Prevalence (%) of stunting, wasting, and underweight among children (6 to 59 months) by macronutrient intake, NNMB

	Prevalence (%)			Mean z score			N
Dietary intake	Stunting	Wasting	Underweight	Height‐for‐age	Weight‐for‐height	Weight‐for‐age
Calorie (kcal)
Below RDA	30.3	13.5	25.7	−1.30	−0.72	−1.20	1,421
Above RDA	22.2	10.6	19.5	−0.92	−0.62	−0.94	395
Protein (gm)
Below RDA	38.6	12.8	29.6	−1.52	−0.72	−1.33	497
Above RDA	24.7	12.7	22.4	−1.10	−0.68	−1.01	1,319
Fat (gm)
Below RDA	33.6	13.3	27.0	−1.37	−0.70	−1.23	923
Above RDA	23.4	12.4	21.6	−1.06	−0.69	−1.06	893
Calorie, protein, and fat
Below RDA	30.4	13.3	25.6	−1.30	−0.72	−1.20	1,457
Above RDA	21.1	11.1	19.2	−0.91	−0.62	−0.93	359
All	28.6	12.8	24.3	−1.22	−0.69	−1.14	1,816

Note. % SD of z scores are based on WHO Reference Group Norms for child anthropometric measurements.

Abbreviations: NNMB, National Nutrition Monitoring Bureau; RDA, recommended dietary allowance.

Table 2.

Percentage distribution of children by macronutrient intake and anthropometric outcomes, NNMB

Dietary intake	Stunting[Link]		Wasting		Underweight[Link]		All % (N)
	Yes % (N)	No % (N)	Yes % (N)	No % (N)	Yes % (N)	No % (N)
Calorie
Below RDA % (N)	83.0 (431)	76.0 (990)	82.1 (192)	77.7 (1,229)	82.6 (365)	76.9 (1,056)	78.2 (1,421)
Above RDA % (N)	17.0 (88)	23.6 (307)	17.9 (42)	22.3 (353)	17.4 (77)	23.1 (318)	21.8 (395)
Protein
Below RDA % (N)	37.0 (192)	23.5 (305)	27.3 (433)	27.3 (64)	33.3 (147)	25.5 (350)	27.4 (1,319)
Above RDA % (N)	63.0 (327)	76.5 (992)	72.6 (1,149)	72.6 (170)	66.7 (295)	74.5 (1,024)	72.6 (497)
Fat
Below RDA % (N)	59.7 (310)	47.3 (613)	52.6 (800)	50.6 (123)	56.3 (249)	49.0 (674)	50.8 (923)
Above RDA % (N)	40.3 (209)	52.7 (684)	47.4 (782)	49.4 (111)	43.7 (193)	51.0 (700)	49.2 (893)
Calorie, protein, and fat
Below RDA % (N)	85.4 (443)	78.2 (1,014)	82.9 (194)	79.8 (1,263)	84.4 (373)	78.9 (1,084)	80.2 (1,457)
Above RDA % (N)	14.6 (76)	21.8 (283)	17.1 (40)	20.2 (319)	15.6 (69)	21.1 (290)	19.8 (359)
All % (N)	100 (519)	100 (1,297)	100 (234)	100 (1,582)	100 (442)	100 (1,374)	100 (1,816)

Pearson chi‐square shows significant difference for bivariate tabulation of the concerned anthropometric indicators vis‐à‐vis all the three dietary intake variables.

The correlation coefficient (r) for calorie intake and height‐for‐age z score was much lower at −0.129 (P < .000; Figure 1). Protein intake (r = 0.147; P < .000) and fat intake (r = 0.126; P < .000) also had a weak‐to‐null correlation with height‐for‐age z scores. A similar weak‐to‐null correlation was observed between weight‐for‐age z scores and calories (r = 0.071; P < .002), protein (r = −0.101; P < .000), and fat intake (r = 0.076; P < .001). The correlation with respect to weight‐for‐height z scores was nonsignificant (calories r = −0.011; P = .639, protein r = 0.014; P = .545, fat r = 0.013; P = .583).

Figure 1.

Association of height‐for‐age, weight‐for‐height, and weight‐for‐age z scores with energy, protein, and fat intake, National Nutrition Monitoring Bureau

From our logistic regressions, we found that children with caloric intake lower than the RDA norms had 54% (95% CI [1.17, 2.01]) higher chances of stunting than children with adequate intake (Table 3). The odds of being underweight were also higher for children with below RDA caloric intake (OR: 1.47; 95% CI [1.11, 1.95]). Further, stunting among children with protein intake below RDA levels was two times higher than their counterparts (OR: 1.93; 95% CI [1.53, 2.42]). A similar effect was observed for inadequate protein intake and underweight (OR: 1.66; 95% CI [1.31, 2.11]). After adjusting for important socio‐economic covariates and maternal characteristics (Model 2), the risk of stunting and underweight was still significantly higher for children with below RDA protein intake (OR: 1.54; 95% CI [1.17, 2.02] for stunting and OR: 1.37; 95% CI [1.03, 1.82] for underweight). However, in these adjusted models, caloric and fat intake did not show any significant association with anthropometric failure.

Table 3.

Simple logistic regression for association of macronutrient intake with stunting, wasting and underweight outcomes

Correlates	Model 1	Model 2	Model 3	Model 4
	OR	OR	OR	OR
Stunting
Calorie: Above RDA. ®	1.00	1.00
Calorie: Below RDA	1.54 *** [1.17, 2.01]	1.10 [0.79, 1.94]	—	—
Protein: Above RDA. ®	1.00	1.00
Protein: Below RDA	1.93 *** [1.53, 2.42]	1.54 *** [1.17, 2.02]	—	—
Fat: Above RDA. ®	1.00	1.00
Fat: Below RDA	1.56 *** [1.26, 1.53]	1.14 [0.86, 1.50]	—	—
Calorie, protein, & fat: Above RDA ®			1.00	1.00
Calorie, protein, & fat: Below RDA®	—	—	1.64 ** [1.24, 2.18]	1.45 ** [1.07, 1.97]
Wasting
Calorie: Above RDA®	1.00	1.00
Calorie: Below RDA	1.32 [0.92, 1.98]	1.33 [0.86, 2.04]	—	—
Protein: Above RDA®	1.00	1.00
Protein: Below RDA	1.06 [0.78, 1.46]	0.95 [0.65, 1.37]	—	—
Fat: Above RDA®	1.00	1.00
Fat: Below RDA	1.18 [0.89, 1.57]	1.08 [0.76, 1.54]	—	—
Calorie, protein, & fat: Above RDA®			1.00	1.00
Calorie, protein, & fat: Below RDA®	—	—	1.23 [0.85, 1.78]	1.23 [0.83, 1.82]
Underweight
Calorie: Above RDA®	1.00	1.00
Calorie: Below RDA	1.47 *** [1.11, 1.95]	1.15 [0.81, 1.61]	—	—
Protein: Above RDA®	1.00	1.00
Protein: Below RDA	1.66 *** [1.31, 2.11]	1.37 ** [1.03, 1.82]	—	—
Fat: Above RDA®	1.00	1.00
Fat: Below RDA	1.50 *** [1.19, 1.87]	1.14 [0.86, 1.51]	—	—
Calorie, protein, & fat: Above RDA®			1.00	1.00
Calorie, protein, & fat: Below RDA®	—	—	1.49 ** [1.12, 2.00]	1.38 ** [1.01, 1.90]

Note. % SD of z scores are based on WHO Reference Group Norms for child anthropometric measurements. All the models are adjusted for age and sex of the child. Model 2 and 4 further adjusts for monthly per capita income, religion and social group of the household and maternal covariates like education height, BMI and occupation. OR: odds ratio; [.] 95% confidence intervals.

^*** P < .01.

^** P < .05.

^* P < 0.

Models 3 and 4 further examine the association between anthropometric outcomes and combined intake of calorie, protein, and fat as per the RDA norms. Model 3 (adjusted for age–sex only) suggests higher chances of both stunting (OR: 1.64; 95% CI [1.24, 2.18]) and underweight (OR: 1.49; 95% CI [1.12, 2.00]) among children who fail to meet the RDA norms for the combined intake of calorie, protein, and fat. When adjusted for other demographic and socio‐economic correlates (Model 4), the magnitude of associations attenuated but were statistically significant for both stunting (OR: 1.45; 95% CI [1.07, 1.97]) and underweight (OR: 1.38; 95% CI [1.01, 1.90]). Importantly, across all models, none of the indicators for food‐based failure displayed significant association with wasting.

Finally, discriminatory accuracy was ascertained using AUC on the basis of the age/sex adjusted estimates from Model 1 (Figure 2). The AUC values were low (all AUCs <0.62), implying that the individual level information on all three food‐based failures, either alone or jointly, was insufficient to identify and categorize children across the three indicators of anthropometric failure. The low discriminatory accuracy score suggests a poor association between dietary intake indicator and anthropometric outcomes. This finding indicates the multifactorial nature of the phenomenon and the potential role of complementarities across determinants in influencing anthropometric outcomes.

Figure 2.

Areas under the receiver operating characteristic curve for anthropometric‐based nutrition failure adjusted for macronutrient intake‐based nutrition failure, National Nutrition Monitoring Bureau

4. DISCUSSION

Our study has two salient findings. First, the prevalence of food‐based nutritional failure as per the RDA norms was more than twice of the prevalence of anthropometrics‐based nutrition failure among children in India. Second, there was an extremely weak association between dietary intake and anthropometric outcomes, especially after adjusting for socio‐economic and maternal characteristics. Although results from logistic regressions suggest a statistically significant association between protein intake and combined intake of all three macronutrients (calorie, protein, and fat) on stunting and underweight, the discriminatory accuracy tests show that anthropometric‐based failures failed to discriminate children who were experiencing food‐based nutritional failures versus those who were not, and vice versa. These findings suggest that dietary intake alone can have only a limited influence to improve child anthropometric outcomes in India. At the same time, it also draws attention to the high levels of food‐based deprivation among Indian children. Because food is regarded as an important determinant of anthropometric status, it is critical to not only strengthen existing food supplementation policies and programmes but also enhance community knowledge, awareness, and practices to improve dietary intake among the population. Alternative approaches for diet assessment such as dietary diversity scores, food insecurity scales, and dietary intake estimates should be further elevated as policy level indicators for greater socio‐political focus and visibility. In fact, determinants such as safe sanitation (or open defecation) have witnessed considerable improvements in India after a focus on these indicators as important targets for government policies and programmes (NSSO, 2016). Thus, a systematic focus on diet‐related indicators such as dietary diversity or dietary intake could be elevated from a programmatic indicator to a policy level target for greater visibility and action as an important determinant of nutritional health. This would invariably require greater emphasis on measurement, data collection, and policy communication of these indicators.

This study has four major limitations. First, the analysis was based on cross‐sectional data and the results do not necessarily reveal a causal association. Second, because of data paucities, the analysis does not adjust for potential environmental factors such as water, sanitation, and health care practices that can shape the impact of dietary intake on anthropometric outcome. Third, use of only one 24‐hr recall period for dietary assessment may be sensitive to the individual and household context of the last day, whereas anthropometric indicators encompass a comparatively longer period of deprivation. Also, given the wide geographic and agroclimatic variations in India, it is feasible to expect seasonal variations in dietary intake, which further affects the suitability of the variable as a proxy for consistent or long‐term dietary intake. This may partly be a reason for the lack of association with wasting. Finally, information on diet and other socio‐economic and environmental contexts was collected on the basis of self‐reported data.

Our findings have important implications for ongoing programmatic interventions to catalyse improvements in nutritional outcomes in India at both the individual and household levels. Further efforts are warranted to improve the implementation and uptake of programmes such as Integrated Child Development Services, particularly services concerning complementary feeding (IIPS & ICF, 2017; Jain, 2015; Marriott, White, Hadden, Davies, & Wallingford, 2012; Ruel & Menon, 2002). To achieve this, it is important to understand and meet the dietary requirements for under‐five children, especially among the poor households where more concerted engagement is necessary to ensure food security and dietary diversity (Deaton & Drèze, 2009). Food‐based failure continues to be an important concern in India, and ensuring sufficient food consumption, especially for disadvantaged populations, is an important developmental priority. Indicators of adequate dietary intake, in conjunction with anthropometric failures and by their own, should be measured and monitored for greater socio‐political visibility and action.

5. CONCLUSION

This study examined the strength of association and discriminatory accuracy between food intake, as per calorie, protein, and fat intake as per the RDA norms, and anthropometric outcomes and found very weak correspondence. This suggests that dietary intake alone may have limited impact on curbing the complex phenomenon of anthropometric failure. Nevertheless, adequate dietary intake, both in terms of diversity and quantity, should be monitored and improved. Future studies should assess the burden of child malnutrition in terms of different measures of dietary assessments in order to strengthen the evidence base for policies concerning child growth and development.

CONFLICTS OF INTEREST

The authors declare that they have no conflicts of interest.

CONTRIBUTIONS

SVS conceptualized and designed the study. WJ contributed to the conceptualization and led the data analysis, wrote the first draft, and led the revision. SR and RK contributed to the analysis, interpretation of the results, and writing. AL, RH, NA, IIM, NB, MR, and SS contributed to the interpretation of the results and critical revisions. SVS provided overall supervision to the manuscript. All authors approved the final submission of the study.

Joe W, Rajpal S, Kim R, et al. Association between anthropometric‐based and food‐based nutritional failure among children in India, 2015. Matern Child Nutr. 2019;15:e12830 10.1111/mcn.12830