Abstract
Background
Dietary diversity is a critical determinant of children’s nutritional well-being and micronutrient intake, particularly during the complementary feeding period (6–23 months). This study examines geographic disparities in minimum dietary diversity (MDD) among Indian children aged 6–23 months, emphasizing its role in addressing malnutrition. Despite India’s high burden of child undernutrition, less than one-third of children meet the WHO’s MDD standards. The study aligns with Sustainable Development Goal 2 (SDG 2), zero hunger, aiming to identify regional inequalities and inform targeted interventions.
Data and methods
Using data from the National Family Health Survey-5 (NFHS-5, 2019–2021), this study analyzed a final sample of 63,247 children aged 6–23 months. Predictor variables included individual, maternal, and household-level factors, while MDD was defined as the consumption of foods from at least five out of eight food groups. Spatial analysis techniques, including choropleth mapping, Getis-Ord Gi* hotspot analysis, Ordinary Kriging interpolation, and Geographically Weighted Regression (GWR), were employed to explore geographic variations and their determinants.
Results
The prevalence of inadequate MDD was 77.06%, with significant geographic disparities. Districts in the southern and north-eastern states exhibited better dietary practices, whereas most districts in central and northern regions, including Bihar, Uttar Pradesh, Madhya Pradesh and Chhattisgarh showed alarmingly high inadequacy rates (80.10–96.00%). GWR analysis revealed spatially varying relationships between predictors and inadequate MDD across Indian districts. For instance, Southern districts, especially in Tamil Nadu, Kerala, Karnataka, and parts of Andhra Pradesh, showed strong negative coefficients (–0.427 to − 0.250), indicating that better toilet facilities are linked to lower levels of inadequate MDD. Similarly, most districts in states like Uttar Pradesh, Bihar, Madhya Pradesh, Maharashtra, Odisha, Chhattisgarh, West Bengal, Andhra Pradesh, Kerala and Telangana show negative coefficients (–0.253 to 0.000), indicating that greater maternal exposure to mass media is associated with lower inadequate MDD. Furthermore, districts in southern, western, and eastern India, including Tamil Nadu, Karnataka, Maharashtra, Andhra Pradesh, Telangana, Odisha, West Bengal, and the northeastern states, show strong positive associations (coefficients 0.401 to 0.800), indicating that higher prevalence of underweight mothers is linked to poorer child dietary diversity.
Conclusion
This study highlights critical geographic disparities in inadequate MDD among children aged 6–23, emphasizing the need for region-specific interventions. Central and northern regions require urgent attention due to the high clustering of inadequate dietary diversity, while southern and northeastern states demonstrate favourable conditions. Integrated approaches addressing maternal nutrition, sanitation facilities, maternal exposure to mass media, and maternal age at first birth are essential for reducing the level of inadequate dietary diversity.
Keywords: Minimum dietary diversity, Child, Geographically weighted regression, Spatial analysis, Geographical disparity, Child nutrition, NFHS, India
Introduction
Dietary diversity plays a crucial role in determining children’s micronutrient intake and overall nutritional well-being [11]. Dietary diversity is a widely used indicator to evaluate dietary quality, micronutrient adequacy, and food availability [24]. To guarantee minimal dietary diversity (MDD), the World Health Organization (WHO) [38] recommends A child aged 6–23 months was considered to have achieved MDD if they consumed foods from at least five out of eight food groups in the previous day or night. Its include Cereals, legumes, nuts, dairy products, meat products, eggs, vitamin A-rich fruits and vegetables, and other fruits and vegetables are some examples of these categories [31, 36]. The transition period from exclusive breastfeeding to the initiation of a variety of foods as well as breast milk is especially known as complementary feeding, typically occurring between 6 and 24 months of age [6]. This 18-month period is crucial for children, with a critical window of opportunity to address undernutrition and prevent its long-lasting consequences on health and neurocognitive development [21]. Brain development is more rapid in the early years of life compared to the rest of the body [2]. The first two years of life are vital for brain development, with the brain reaching 80% of its adult weight by age two, making proper nutrition essential during this period [5, 17].
Children who lack sufficient dietary diversity after 6 months of age are at risk of stunting, even if they receive optimal breastfeeding [3]. Dietary diversity plays a vital role in influencing the health status of children [1] and is also it is a key protective factor against poor nutritional status [14]. Poor nutritional status becomes apparent in the form of stunting (chronic and long-term undernutrition); wasting (an acute form of undernutrition); and underweight (a combination of both stunting and wasting) among children [19]. In this context, a diversified diet is essential starting from the sixth month, as exclusive breastfeeding alone can no longer fulfill the nutritional requirements of a growing child [20].
Unfortunately, less than one-third (29%) of children aged 6–23 months globally meet the WHO standard for minimum dietary diversity (MDD) [37]. Whereas, just 16% of children in India can fulfill the worldwide criterion for nutritional adequacy, which is even less than the percentage in African nations like Tanzania (21%), Ghana (26%), and Mali (22%) (FAO, [8, 9]). MDD in India shows clear geographical variation, with lower diversity observed in many northern and central states where vegetarian diets are more prevalent. For example, Rajasthan records a dietary diversity of 15.4%, followed by Haryana at 10.9%, and Uttarakhand, Uttar Pradesh, and Madhya Pradesh each at 14.1% and 18.6%. In contrast, the northernmost regions such as Jammu & Kashmir (36.6%) and Ladakh (41.0%) exhibit relatively higher diversity. In the eastern part of the country, Bihar and Jharkhand report low dietary diversity at 18.9% and 21.5% respectively. However, coastal states demonstrate improved diversity, with Odisha at 39.4% and West Bengal reaching 48.8%. Western states like Gujarat (16.4%) and Maharashtra (17.7%) also show lower dietary diversity. In southern India, dietary diversity improves, with Tamil Nadu at 28.8% and Kerala notably higher at 46.3%. These patterns highlight regional differences influenced by food preferences, agricultural practices, and socio-economic condition [12] Consequently, insufficient nutritional status, either directly or indirectly, contributes to 60% of the 10.9 million deaths among children under the age of five [32]. Micronutrient deficiencies impair growth and increase children’s vulnerability to infections and illnesses [15]. The government combats malnutrition through ICDS, Anganwadi Services, and Poshan Vatikas, which provide nutrient-rich foods and promote dietary diversity [26].
Child undernourishment among children aged 6–23 months is a critical public health concern in India. SDG-2 aims to end hunger by 2030, emphasizing access to safe, nutritious, and affordable food (Target 2.1), especially for disadvantaged groups. Assessing dietary diversity’s impact on child growth is essential, aligning with SDG-3: Good Health and Well-being. Understanding the spatial disparities in dietary diversity in India among children aged 6–23 months is crucial to addressing malnutrition effectively. Using NFHS-5 data, this study aims to explore geographic variations in minimum dietary diversity and its determinants. Identifying high-risk regions through spatial statistical methods can lead to targeted interventions and policy decisions, ultimately contributing to improved child nutrition and reducing regional inequalities in dietary practices.
Data and methods
Sample selection procedure
Figure 1 outlines the process used to derive the final analytical sample of children aged 6–23 months included in the dietary diversity study. The initial dataset included 232,920 children aged 0–59 months. The first filtering step removed 144,950 children older than 23 months, focusing the sample on children under 24 months who were living with their mothers, resulting in 87,970 cases. Subsequently, 2,680 duplicate cases (with the same case ID) were excluded, bringing the sample down to 85,290. From this, children under 6 months of age (22,043 cases) were also removed, yielding a final sample of 63,247 children aged 6–23 months for analysis. Within this final group, dietary diversity was categorized into two groups:
Fig. 1.
Analytical study sample selection procedure
Adequate dietary diversity
15,145 children (22.94%).
Inadequate dietary diversity
48,102 children (77.06%).
This flowchart efficiently illustrates the data cleaning and selection process, highlighting the rigorous inclusion criteria to ensure a focused and relevant analytical sample for studying dietary diversity in early childhood.
Variable description
Outcome variable
Minimum Dietary Diversity (MDD) was defined based on the WHO Infant and Young Child Feeding (IYCF) guidelines and the Demographic and Health Surveys (DHS) standards [36]. According to these guidelines, MDD was achieved if a child aged 6–23 months consumed foods from at least five out of eight food groups during the previous day or night. The eight food groups included breastmilk; grains, white/pale starchy roots, tubers, and plantains; legumes and nuts; dairy products such as infant formula, milk, yogurt, and cheese; flesh foods including meat, fish, poultry, and organ meats; eggs; vitamin A-rich fruits and vegetables; and other fruits and vegetables [12], 38]. A binary variable was created, classifying children as having adequate dietary diversity (MDD = 1) if they consumed foods from at least five groups and inadequate dietary diversity (MDD = 0) if they consumed fewer than five groups.
Predictor variable
The predictor variables in this study were categorized into three levels: individual-level (child-related), maternal-level, and household-level characteristics, as detailed in Table 1. These variables were selected based on theoretical relevance and previous literature concerning child nutrition, feeding practices and dietary diversity.
Table 1.
Description of the predictor variables used in the study
| Characteristics | Response description |
|---|---|
| Child characteristics | |
| Sex of the child | Male, Female |
| Age (months) | Grouped as 6–11 months, 12–17 months, and 18–23 months |
| Birth order | Categorized as first born, 2–3, and 4 or higher |
| Birth interval | Categorized as less than 24 months or 24 months and above |
| Child is twin | Single vs. multiple births |
| Perceived size of child at birth | Recorded by the mother as small, average, or large |
| Diarrhoea at last two weeks | In the last two weeks preceding the survey (Yes/No) |
| Maternal characteristics | |
| Mother’s age | Categorized into 15–24 years, 25–34 years, and ≥ 35 years |
| Mothers education | Classified as no education, primary, secondary, and higher |
| Number of antenatal care (ANC) visits | Dichotomized as ≤ 3 ANC visits and ≥ 4 ANC visits |
| Maternal body mass index (BMI) (kg/m2) | Categorized as underweight (< 18.5), normal (18.5–24.9), and overweight (≥ 25) |
| Age of mother at first birth | Grouped into < 20 years, 20–29 years, and 30–49 years |
| Marital status | Married and Unmarried |
| Pregnancy wanted | Classified as wanted or unwanted |
| Mother is anemic | Classified as anemic and non-anemic |
| Mass media | Derived from access to television, radio, or newspapers (Yes/No) |
| Place of delivery | Grouped into institutional and non-institutional births |
| Household characteristics | |
| Wealth index | A composite measure provided by NFHS, categorized as poorest, poorer, middle, richer, and richest |
| Household size | Grouped into 1–4 members, 5–8 members, and more than 9 members |
| Sex of the household head | Male/Female |
| Toilet facility | Classified as improved or unimproved based on WHO/UNICEF Joint Monitoring Programme standards |
| Drinking water facilities | Categorized as improved or unimproved |
| Social caste | Categorized as Scheduled Castes (SC)/Scheduled Tribes (ST), Other Backward Classes (OBC), and Other |
| Religion | Grouped as Hindu and Non-Hindu |
| Residence | Categorized as rural or urban |
These variables were considered to examine their association with the outcome variable, minimum dietary diversity (MDD) and to account for the multidimensional factors influencing child feeding practices in India.
Statistical analysis
Bivariate analysis using cross-tabulation and the chi-square test of association was employed to examine the relationship between Minimum Dietary Diversity (MDD) status (adequate vs. inadequate) and various background characteristics of the study population. The MDD variable was dichotomized as per WHO guidelines, with adequacy defined by the consumption of at least five out of eight recommended food groups. Further, spatial analytical techniques were applied to explore geographic disparities in dietary diversity across India. The spatial analysis included:
Choropleth mapping
Choropleth maps were constructed to visually represent the spatial distribution of inadequate dietary diversity at the district level. Each district was color-coded based on the proportion of children with inadequate MDD (MDD < 5 food groups), using a sequential color gradient. Darker shades indicated higher prevalence rates for inadequate minimum dietary diversity. This technique enabled quick identification of regional disparities and served as a foundation for further spatial analysis [18].
Getis-Ord Gi* hotspot analysis
To identify statistically significant spatial clusters of inadequate minimum dietary diversity (MDD) among children aged 6–23 months, the Getis-Ord Gi* (Gi-star) statistic was applied [10]. This local spatial statistic evaluates whether high or low values are clustered in space beyond what is expected under spatial randomness. Specifically, a high positive z-score indicates a hot spot, a cluster of districts with significantly high values of inadequate MDD. Conversely, a low negative z-score indicates a cold spot, districts forming clusters with significantly low values of inadequate MDD. The statistical significance of clustering was determined by the associated p-values, categorized at the 90%, 95%, and 99% confidence levels. Districts with non-significant z-scores were considered not to exhibit statistically significant clustering. This method has been widely applied in nutritional epidemiology to uncover geographic patterns of child undernutrition [16, 29, 33]
Ordinary kriging interpolation
Ordinary Kriging is a geostatistical interpolation method that estimates unknown values at unsampled locations by leveraging the spatial autocorrelation observed among known sample points [16, 33]. The method assumes a constant but unknown mean across the study region. Predictions at unsampled locations are obtained by calculating a weighted average of observed values at nearby sampled locations. The weights assigned to each sampled location are derived from a semi variogram model, ensuring that the sum of the weights equals one to maintain unbiasedness. These weights are optimized to minimize the prediction variance while preserving the spatial dependence inherent in the data. The output is a continuous surface map that depicts predicted values of inadequate Minimum Dietary Diversity (MDD) across all areas, including those not directly surveyed.
Geographically weighted regression (GWR)
GWR allows for the modeling of spatially varying relationships between a dependent variable (inadequate MDD) and its predictors [33, 34]. Instead of fitting a single global regression equation, GWR calibrates a local regression for each observation by giving more weight to nearby data points [4]. A spatial kernel (e.g., Gaussian or bisquare) is used to assign weights to neighbouring observations. The model generates spatial maps of each coefficient, showing how the relationship between variables varies geographically.
All spatial analyses were performed using ArcGIS Pro, whereas statistical analyses of individual-, maternal-, and household-level predictors were conducted in STATA.
Result
Table 2 presents the prevalence of inadequate minimum dietary diversity (MDD) among children aged 6–23 months across various individual, maternal, and household-level characteristics. Overall, 77.06% of children had inadequate MDD, with notable disparities across subgroups. The prevalence of inadequate MDD decreases with the increasing age of the child, varying from 86.55% (6–11 months) to 69.44% (18–23 months), showing a significant association (p < 0.001). Children from multiple births had a significantly higher prevalence of inadequate MDD (85.85%) compared to singletons (76.98%). Maternal education showed a clear variation, with children of mothers with no education having an 81.24% prevalence of inadequate MDD, while those whose mothers had higher education had a lower prevalence (74.85%). Similarly, children whose mothers had fewer than four antenatal care (ANC) visits had a higher rate of inadequate MDD (79.81%) compared to those with four or more visits (74.5%).
Table 2.
Sample characteristics of the study participants selected for the study
| Backgrounds characteristics | Adequate | Inadequate | |
|---|---|---|---|
| Individual-level factors | N (%) | N (%) | P- value |
| Child characteristics | |||
| Sex | |||
| Male | 7817(23) | 24,992(77) | 0.522 |
| Female | 7329(22.88) | 23,154(77.12) | |
| Age (months) | |||
| 6–11 | 2907(13.45) | 18,449(86.55) | < 0.001 |
| 12–17 | 5977(25.33) | 16,138(74.67) | |
| 18–23 | 6262(30.56) | 13,559(69.44) | |
| Birth order | |||
| Firstborn | 5556(22.49) | 19,246(77.51) | < 0.001 |
| 2–3 | 7713(23.96) | 23,022(76.04) | |
| 4 or higher | 1877(20.06) | 5878(79.94) | |
| Birth interval | |||
| < 24 months | 4(30.68) | 12(69.32) | 0.92 |
| ≥ 24 months | 15,142(22.94) | 48,134(77.06) | |
| Child is twin | |||
| Single birth | 15,051(23.02) | 47,692(76.98) | < 0.001 |
| Multiple birth | 95(14.15) | 454(85.85) | |
| Perceived size of a child at birth | |||
| Small | 1469(23.19) | 4935(76.81) | < 0.001 |
| Average | 10,342(21.93) | 34,212(78.07) | |
| Larger | 3150(26.67) | 8523(73.33) | |
| Diarrhoea at last two weeks | |||
| No | 13,469(22.76) | 43,319(77.24) | < 0.001 |
| Yes | 1660(24.49) | 4777(75.51) | |
| Maternal factors | |||
| Mother’s age | |||
| 15–24 | 5556(22.03) | 20,101(77.97) | < 0.001 |
| 25–34 | 8368(23.36) | 25,015(76.64) | |
| >=35 | 1222(26.11) | 3030(73.89) | |
| Mother’s education | |||
| No education | 2540(18.76) | 9670(81.24) | < 0.001 |
| Primary | 1733(21.62) | 5693(78.38) | |
| Secondary | 8342(23.97) | 25,533(76.03) | |
| Higher | 2531(25.15) | 7250(74.85) | |
| ANC visit | |||
| <=3 ANC visit | 5536(20.19) | 19,991(79.81) | < 0.001 |
| >=4 ANC visit | 9305(25.5) | 26,198(74.5) | |
| Maternal BMI (kg/m2) | |||
| < 18.5 | 8357(22.66) | 27,037(77.34) | < 0.001 |
| 18.5 to 24.9 | 2850(21.42) | 10,174(78.58) | |
| 25+ | 2421(26.69) | 6445(73.31) | |
| Age of mother at 1 st birth | |||
| < 20 | 4607(23.52) | 14,912(76.48) | < 0.001 |
| 20–29 | 9901(22.47) | 31,600(77.53) | |
| 30–49 | 638(26.68) | 1634(73.32) | |
| Marital status | |||
| Unmarried | 32(36.2) | 69(63.8) | 0.068 |
| Married | 15,114(22.93) | 48,077(77.07) | |
| Pregnancy wanted | |||
| Wanted | 13,966(22.89) | 44,418(77.11) | 0.848 |
| Un wanted | 1180(23.52) | 3728(76.48) | |
| Mother anemia status | |||
| Non-Anemic | 5956(22.58) | 18,676(77.42) | 0.411 |
| Anemic | 8685(23.46) | 27,669(76.54) | |
| Mass media | |||
| No | 3675(19.61) | 14,246(80.39) | < 0.001 |
| Any | 11,471(24.2) | 33,900(75.8) | |
| Place of delivery | |||
| Non-Institutional | 1763(21.11) | 5377(78.89) | 0.109 |
| Institutional | 13,383(23.12) | 42,769(76.88) | |
| Household factors | |||
| Wealth index | |||
| Poorest | 3846(21.88) | 12,790(78.12) | < 0.001 |
| Poorer | 3365(21.84) | 11,211(78.16) | |
| Middle | 3028(23.46) | 9501(76.54) | |
| Richer | 2702(24.06) | 8130(75.94) | |
| Richest | 2205(24.02) | 6514(75.98) | |
| Household size | |||
| 1–4 | 4217(24.94) | 12,174(75.06) | < 0.001 |
| 5_8 | 8781(22.95) | 27,894(77.05) | |
| > 9 | 2148(19.9) | 8078(80.1) | |
| Sex of the household head | |||
| Male | 12,615(22.72) | 41,036(77.28) | < 0.001 |
| Female | 2531(24.16) | 7110(75.84) | |
| Toilet facility | |||
| Improved | 10,310(23.71) | 31,267(76.29) | < 0.001 |
| Unimproved | 4608(21.51) | 16,352(78.49) | |
| Drinking water facilities | |||
| Improved | 13,633(22.96) | 44,063(77.04) | < 0.001 |
| Unimproved | 1133(22.69) | 2940(77.31) | |
| Social caste | |||
| SC/ST | 6627(23.16) | 18,910(76.84) | < 0.001 |
| OBC | 5130(21.12) | 19,238(78.88) | |
| Other | 2517(24.11) | 7977(75.89) | |
| Religion | |||
| Hindu | 10,222(21.92) | 36,634(78.08) | < 0.001 |
| Non-Hindu | 4924(26.91) | 11,512(73.09) | |
| Residence | |||
| Urban | 3229(24.31) | 9602(75.69) | < 0.001 |
| Rural | 11,917(22.45) | 38,544(77.55) |
Maternal nutritional status was also associated with dietary outcomes; children of underweight mothers (BMI < 18.5) had a 77.34% prevalence of inadequate MDD, compared to 73.31% among children of overweight mothers (BMI ≥ 25). Exposure to mass media emerged as a significant protective factor; children of mothers not exposed to any mass media had an 80.39% prevalence of inadequate MDD, compared to 75.8% among those with media exposure. Sanitation also played a role—children in households with unimproved toilet facilities had a higher prevalence (78.49%) than those with improved toilets (76.29%). Rural residence was associated with higher inadequate MDD (77.55%) compared to urban areas (75.69%). Religious differences were observed as well; Hindu children had a higher prevalence of inadequate MDD (78.08%) than children from non-Hindu backgrounds (73.09%). Nearly all, above discussed factors are significantly associated with inadequate MDD with p < 0.001.
Figure 2 presents the district-level spatial distribution of the prevalence of inadequate dietary diversity among children aged 6–23 months in India, visualized through a choropleth map. The map utilizes a five-class graduated color scale ranging from yellow (lowest inadequacy) to dark blue (highest inadequacy), with the number of districts in each category indicated in parentheses.
Fig. 2.
Spatial distribution of inadequate dietary diversity among children aged 6–23 months in India
Yellow (0.00–45.00%; 22 districts): Represents districts with the lowest levels of dietary inadequacy, suggesting relatively better child feeding practices. These are sparsely located, notably in parts of Kerala, and northeastern districts.
Light green (45.10–60.00%; 69 districts): Indicates the intermediate level prevalence of inadequate dietary diversity. These districts are scattered across southern, northern and north-eastern India.
Green and light blue (60.10–80.00%; 326 districts): Denotes high levels of dietary inadequacy, predominantly observed in central and southern states including Maharashtra, Karnataka, Odisha, Andhra Pradesh, Madhya Pradesh, Chhattisgarh, Rajasthan, and some parts of the northeast.
Dark blue (80.10–96.00%; 290 districts): Indicates the most severely affected districts with extremely high levels of inadequate dietary diversity. These are prominently located in Bihar, Jharkhand, Uttar Pradesh, Madhya Pradesh, Rajasthan, and parts of Gujarat and Assam, marking them as critical regions of concern for child nutrition interventions.
Hotspot analysis of inadequate dietary diversity
Figure 3 presents the Getis-Ord Gi* hotspot analysis, identifying the spatial clustering of inadequate dietary diversity among children aged 6–23 months across Indian districts. The map categorizes districts into statistically significant hotspots (areas with high prevalence) and cold spots (areas with low prevalence) at three confidence levels: 90%, 95%, and 99%. Hotspots (shades of red) represent districts where the prevalence of inadequate dietary diversity is significantly higher than the national average, and where surrounding districts also report high values. The most intense clustering (99% confidence) is seen in the northern and central regions, notably across Uttar Pradesh, Madhya Pradesh, Rajasthan, Uttarakhand and some parts of Haryana. These areas are the most critical zones, where poor dietary diversity is not only widespread but statistically concentrated. Cold spots (shades of blue) are areas where the prevalence is significantly lower, indicating better child feeding practices. These clusters are clearly visible in the southern states like Kerala, Tamil Nadu, Karnataka, and parts of Andhra Pradesh, as well as in northeastern states such as Nagaland, Mizoram, and Sikkim, and Jammu & Kashmir in the north.
Fig. 3.
Hotspot analysis of inadequate dietary diversity among children aged 6–23 months in India
Interpolation mapping using ordinary kriging
Figure 4 presents the spatial interpolation of the prevalence of inadequate dietary diversity among children aged 6–23 months in India using Ordinary Kriging. This geostatistical method estimates unknown values based on spatial autocorrelation among observed data points. The interpolated map uses a five-class color gradient ranging from yellow (low prevalence) to dark blue (high prevalence):
Fig. 4.
Spatial pattern Kriging interpolation of inadequate dietary diversity in India
Yellow (46.32–54.55): These areas reflect the lowest prevalence of inadequate dietary diversity, indicating better nutritional practices. Such regions are observed in parts of northern Jammu & Kashmir and select northeastern states.
Light green and green (54.56–71.02): This range reflects relatively moderate prevalence, seen in southern regions like Kerala, Tamil Nadu, and parts of northeastern, Karnataka, Andhra Pradesh, and the eastern states India.
Light blue (71.02–79.23): Higher prevalence regions appear in large parts of central and northern India, including regions of Maharashtra, Gujarat, Chhattisgarh, and Odisha.
Dark blue (79.25–87.48): These are the critical hotspots with the highest estimated prevalence of inadequate dietary diversity, prominently found in states such as Madhya Pradesh, Uttar Pradesh, Bihar, and Rajasthan, indicating urgent nutritional risk among children.
Geographically weighted regression (GWR) analysis
Figure 5 presents the Geographically Weighted Regression (GWR) analysis of inadequate minimum dietary diversity (MDD) among children aged 6–23 months in India. The map displays the local beta coefficients estimated from the GWR model, indicating the strength and direction of association between various covariates and inadequate MDD across districts. The colour gradation reflects the magnitude of the GWR coefficients rather than residuals. Areas shaded in blue and dark blue indicate stronger positive associations, suggesting that increasing prevalence of the selected covariates increasing the likelihood of inadequate MDD. In contrast, regions in light green to yellow show weaker or even negative associations, implying either a lower or inverse effect of covariate on inadequate MDD in those districts.
Fig. 5.
Geographically Weighted Regression (GWR) analysis of Inadequate Dietary Diversity (IDD) among children aged 6–23 months in India
The spatial distribution of GWR coefficients highlights distinct regional variations in factors associated with inadequate minimum dietary diversity (MDD) among Indian children. Map A reveals strong positive associations between child age (6–11 months) and inadequate MDD across most districts, while parts of Uttar Pradesh show weak negative links, suggesting improved diversity in this age group. Map B indicates that higher birth order (second/third) is positively linked with inadequate MDD in northern, western, and north-eastern India, whereas southern and central regions show weaker effects. Map C reveals positive associations in northern and north-eastern districts, indicating dietary disadvantages for average birth-size children, while southern and central India show negative links suggesting better postnatal care. Map D demonstrates negative associations between maternal age (25–34 years) and inadequate MDD in southern, eastern, and parts of northern India,
Map E shows stronger positive associations between maternal undernutrition and inadequate MDD in southern, northern, and north-eastern regions, emphasizing the critical role of maternal nutritional status. In most other districts, maternal underweight also shows a positive link with inadequate MDD. Map F highlights that even among women aged 20–29 years at first birth, stronger positive coefficients in northern, north-eastern, and some western regions indicate persistent dietary vulnerability, possibly due to other underlying disadvantages. Map G demonstrates that maternal non-anemia is negatively associated with inadequate MDD across most districts, suggesting that better maternal health improves child feeding outcomes. However, exceptions in parts of northern, north-eastern, and western India indicate spatial heterogeneity. Finally, Map H reveals that higher maternal exposure to mass media in southern, eastern, middle, and some north-eastern districts is negatively associated with inadequate MDD, which indicates exposure to the mass media works as protective factor against inadequate MDD.
Map I shows strong positive associations between household size and inadequate MDD in central, western, and eastern India, indicating poorer dietary diversity among children from larger households. Map J reveals predominantly negative associations between improved sanitation (toilet availability) and inadequate MDD across most regions, except parts of northern and north-eastern India indicating that a higher prevalence of improved toilet facilities is linked to a reduction in inadequate MDD. Map K focuses on children belonging to Other Backward Classes (OBCs). In Central and Eastern India, strong positive GWR coefficients indicate that children from OBC households face a significant disadvantage in terms of dietary diversity. In contrast, Southern India reveals negative to near-zero coefficients, suggesting minimal or inverse associations. Map L demonstrates positive associations between rural residence and inadequate MDD across most regions, except some eastern districts, indicating that children in rural areas are more likely to suffer from inadequate dietary diversity.
Discussion
This study highlights significant geographic disparities in minimum dietary diversity (MDD) among Indian children aged 6–23 months, using data from the latest round of the National Family Health Survey (NFHS-5, 2019–2021). Employing advanced spatial analysis methods, including hotspot analysis (Getis-Ord Gi*), Ordinary Kriging interpolation and Geographically Weighted Regression (GWR), we identified distinct regional patterns and key determinants of inadequate MDD. Higher birth order, child age, maternal underweight, larger household size, mass media, sanitation infrastructure, caste, and religion emerged as significant factors influencing dietary diversity. Despite national initiatives such as the Integrated Child Development Services (ICDS), Poshan Abhiyaan (National Nutrition Mission), and Anganwadi Services aimed at combating malnutrition and enhancing child nutrition, our geospatial analyses reveal persistent disparities across regions.
Hotspot mapping identified statistically significant clusters of inadequate MDD in northern and central India particularly in Bihar, Uttar Pradesh, Madhya Pradesh, and Rajasthan. In contrast, coldspots were observed in the southern states (Kerala, Tamil Nadu) and parts of the northeast, indicating better infant and young child feeding practices in these regions [13]. The kriging interpolation map further reveals a pronounced north-south gradient in dietary diversity, with the highest prevalence of inadequacy (79.25–87.48%) concentrated in a belt spanning Bihar, Uttar Pradesh, Madhya Pradesh, and Rajasthan. Moderate prevalence levels (71.02–79.23%) were observed in Jharkhand, Odisha, Gujarat, and Maharashtra, while transitional zones (62.79–71.02%) were identified in southern Maharashtra, Andhra Pradesh, and West Bengal. The lowest levels of inadequate MDD (46.32–54.55%) were found in Kerala, Tamil Nadu, coastal Karnataka, and northeastern states such as Mizoram and Sikkim, reflecting relatively better infant and young child feeding practices in these regions.
Geographically Weighted Regression (GWR) results indicate that in northern and central India, particularly states such as Punjab, Haryana, Uttar Pradesh, Bihar, Madhya Pradesh, Rajasthan, and Jharkhand, we observed a strong convergence of multiple risk factors. Younger child age (6–11 months), maternal underweight status, higher birth order [35], younger maternal age at first birth [39], younger maternal age, limited maternal media exposure, poor sanitation, large household size, and social disadvantage (particularly among OBC) and rural residents were all significantly associated with inadequate dietary diversity. In some regions, children of mothers in the biologically favorable age group (20–29 years) still experienced inadequate dietary diversity likely due to systemic poverty, food insecurity, or weak health infrastructure, despite the age advantage [27]. Although maternal anemia was not significantly associated with inadequate MDD at the national level, GWR revealed regionally varied effects in areas with structural health challenges. While some media exposure exists, its impact is dampened by prevailing patriarchal norms, low maternal education, and restricted decision-making autonomy, especially in western Bihar and Jharkhand. Improved toilet access was generally associated with lower inadequate MDD; however, in a few northern and northeastern districts, the impact was limited, suggesting that infrastructure alone may not be sufficient without complementary behavioral and nutrition interventions. Furthermore, children from rural residents in these areas were significantly more likely to experience dietary inadequacy, highlighting the role of a vegetable-based household diet in child nutrition.
In contrast, southern and northeastern states including Kerala, Tamil Nadu, Karnataka, Andhra Pradesh, Telangana, Assam, and Mizoram displayed a remarkably different profile, with most of these variables showing weak or no association with inadequate dietary diversity. Here, higher maternal education levels, effective Anganwadi and health outreach systems, greater mass media exposure, smaller household sizes, and more equitable feeding practices mitigate the risks posed by child age, maternal undernutrition, or birth order. Notably, maternal age, anemia, and caste background showed minimal or no association with inadequate MDD in these states, suggesting a more inclusive and effective nutrition environment. Mass media exposure, in particular, showed strong positive associations with better dietary diversity. Similar patterns have emerged across India and South Asia, where nutrition-focused media campaigns and regular media exposure have influenced child feeding practices [7, 22, 30]. Improved sanitation infrastructure in these regions also correlated with higher dietary adequacy. This result is consistent with findings from Uganda, where poor sanitation increased the risk of food insecurity by 3.4-fold [25], reflecting successful integration of health and nutrition initiatives. Even within disadvantaged social groups, such as OBC communities, dietary diversity outcomes were comparatively better, highlighting the influence of more equitable development frameworks in these regions.
Western India presented a mixed picture. Urban areas and better-developed districts in Maharashtra and Gujarat demonstrated moderate to high dietary adequacy, likely benefiting from better access to health services, media exposure, and potentially fortified foods. However, tribal belts and desert districts in Rajasthan and parts of Maharashtra showed strong associations between inadequate diets and factors such as maternal underweight, younger maternal age at first birth, larger household size, and caste disadvantage. These observations mirror findings from Ethiopia [28] and Uganda [25], as well as from within South Asia [7], where overlapping social and environmental vulnerabilities shape child nutrition outcomes.
Building on these regional profiles, our geospatial analysis further confirms significant geographic disparities in dietary diversity outcomes. Central and northern regions—including Bihar, Uttar Pradesh, Rajasthan, Madhya Pradesh, Jharkhand, and Odisha—emerge as critical zones requiring urgent nutritional interventions due to the high clustering of minimum dietary diversity failure (MDDF), which exceeds 80% in many districts. In contrast, southern and northeastern districts such as Kerala, Tamil Nadu, Andhra Pradesh, coastal Karnataka, West Bengal, Sikkim, Nagaland and Mizoram consistently demonstrate more favorable conditions for child nutrition, with cold spots indicating improved feeding practices. These regional patterns echo findings from previous NFHS rounds and recent studies [23], suggesting that long-standing structural disadvantages persist in the central belt. The spatial patterns reveal that southern and northeastern states benefit from synergistic advantages, including higher maternal education, improved sanitation, and favorable household characteristics, which together create an enabling environment for child nutrition. Conversely, central and northern regions face compounded disadvantages where no single intervention suffices, demanding integrated approaches that simultaneously address education, sanitation, nutrition, and caste-based inequities. These findings underscore the limitations of uniform national policies and advocate for decentralized, regionally tailored strategies that respond to localized nutritional needs.
Conclusion
This study reveals significant geographic disparities in Minimum Dietary Diversity (MDD) among Indian children aged 6–23 months, with central and northern states such as Bihar, Uttar Pradesh, Madhya Pradesh, and Rajasthan exhibiting alarmingly high rates of inadequate MDD (80.10–96.00%), while southern and northeastern states like Kerala, Tamil Nadu, Mizoram, and Sikkim demonstrate relatively better dietary practices. The most important variables influencing MDD include child age, birth order, maternal nutrition, age at first birth, maternal anemia, sanitation infrastructure, mass media exposure, social caste, and rural residents. However, the effects of these factors vary significantly across regions showing strong associations in central and northern states but weak or minimal associations in the south and northeast. These spatial variations indicate that the need for region-specific interventions tailored to local contexts. Policy recommendations should focus on targeted nutrition programs, expansion of Anganwadi services, Rural residents, improved sanitation infrastructure, and contextualized mass media campaigns, particularly in high-risk areas mainly north and central region. Decentralized, multidimensional strategies that address, caste inequities, household size, and maternal health are essential for improving child nutrition outcomes and achieving SDG 2 zero hunger in India.
Limitation
This study has several limitations. The cross-sectional design limits the ability to establish causal relationships between dietary diversity and associated factors. The use of secondary data restricted the inclusion of important behavioral and environmental variables like maternal nutrition knowledge and seasonal food availability. Dietary diversity was based on a 24-hour recall, which may not reflect long-term feeding practices and is prone to recall bias. Spatial analysis techniques like GWR describe patterns but do not explain underlying mechanisms and are sensitive to model assumptions. Results from GWR and hotspot analysis may be influenced by spatial scale and bandwidth selection. District-level aggregation may mask smaller-area variations, leading to ecological fallacy. Some key factors influencing dietary diversity, such as food preferences and cultural practices, were not fully captured. The MDD indicator does not account for nutrient quality or portion sizes, limiting its sensitivity. The study assumes uniformity in data collection across regions, which may not always be accurate. Despite these limitations, the study provides valuable insights into geographic disparities in child nutrition across India.
Author contributions
SB and MS contributed to the conceptualization of the study, with SB leading the data analysis. MS, AS and SB contributed to the interpretation of the data and the initial draft of the article. All authors critically reviewed the final version of the manuscript and approved it for submission.
Funding
Authors did not receive any funding to carry out this research.
Data availability
No datasets were generated or analysed during the current study.
Declarations
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Aboagye RG, Seidu A-A, Ahinkorah BO, Arthur-Holmes F, Cadri A, Dadzie LK, et al. Dietary diversity and undernutrition in children aged 6–23 months in sub-Saharan Africa. Nutrients. 2021;13(10):3431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Benton D. The influence of dietary status on the cognitive performance of children. Mol Nutr Food Res. 2010;54(4):457–70. [DOI] [PubMed] [Google Scholar]
- 3.Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, De Onis M, et al. Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet. 2013;382(9890):427–51. 10.1016/S0140-6736(13)60937-X. [DOI] [PubMed] [Google Scholar]
- 4.Brunsdon C, Fotheringham AS, Charlton ME. Geographically weighted regression: a method for exploring spatial nonstationarity. Geogr Anal. 1996;28(4):281–98. [Google Scholar]
- 5.Bryan J, Osendarp S, Hughes D, Calvaresi E, Baghurst K, Van Klinken J-W. Nutrients for cognitive development in school-aged children. Nutr Rev. 2004;62(8):295–306. [DOI] [PubMed] [Google Scholar]
- 6.Dewey KG, Brown KH. Update on technical issues concerning complementary feeding of young children in developing countries and implications for intervention programs. Food Nutr Bull. 2003;24(1):5–28. [DOI] [PubMed] [Google Scholar]
- 7.Dhawan D, Pinnamaneni R, Viswanath K. Association between mass media exposure and infant and young child feeding practices in India: a cross-sectional study. Sci Rep. 2023;13(1):19353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Fao E (n.d.), editor. AUC. 2020. Africa Regional Overview of Food Security and Nutrition 2019. Accra.FAO, ECA and AUC. 2021. Africa regional overview of food security and nutrition 2020: Transforming food systems for affordable healthy diets. Accra, FAO. 10.4060/cb4831en
- 9.FAO, ECA and AUC. 2021. Africa regional overview of food security and nutrition 2020: Transforming food systems for affordable healthy diets. Accra, FAO. 10.4060/cb4831en
- 10.Getis A, Ord JK. The analysis of spatial association by use of distance statistics. Geogr Anal. 1992;24(3):189–206. [Google Scholar]
- 11.Habte T-Y, Krawinkel M. Dietary diversity score: A measure of nutritional adequacy or an indicator of healthy diet? 2016.
- 12.Iips I. National Family Health Survey (NFHS-5): 2019-21 India. Mumbai: International Institute for Population Sciences (IIPS). 2021.
- 13.Jain A, Wang W, James KS, Sarwal R, Kim R, Subramanian SV. Small area variations in dietary diversity among children in India: a multilevel analysis of 6–23-month-old children. Front Nutr. 2022;8:791509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kennedy GL, Pedro MR, Seghieri C, Nantel G, Brouwer I. Dietary diversity score is a useful indicator of micronutrient intake in non-breast-feeding Filipino children2. J Nutr. 2007;137(2):472–7. [DOI] [PubMed] [Google Scholar]
- 15.Khan J, Das SK. The burden of anthropometric failure and child mortality in India. Sci Rep. 2020;10(1):20991. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kitaw TA, Abate BB, Tilahun D, B., Haile RN. Geospatial pattern of level of minimum acceptable diet and its determinants among children aged 6–23 months in Ethiopia. Spatial and multiscale geographically weighted regression analysis. Front Public Health. 2024;12:1348755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Lenroot RK, Giedd JN. Brain development in children and adolescents: insights from anatomical magnetic resonance imaging. Neurosci Biobehav Rev. 2006;30(6):718–29. [DOI] [PubMed] [Google Scholar]
- 18.Mandadi R, Tripathi N, Pal I, Mozumder C, Gonzales A. Geostatistical exploratory analysis on child malnutrition and its determinants in India. Int J Geoinformatics. 2023;19(6):77–90. [Google Scholar]
- 19.McDonald CM, Olofin I, Flaxman S, Fawzi WW, Spiegelman D, Caulfield LE, et al. The effect of multiple anthropometric deficits on child mortality: meta-analysis of individual data in 10 prospective studies from developing countries. Am J Clin Nutr. 2013;97(4):896–901. 10.3945/ajcn.112.047639. [DOI] [PubMed] [Google Scholar]
- 20.Molla W, Adem DA, Tilahun R, Shumye S, Kabthymer RH, Kebede D, et al. Dietary diversity and associated factors among children (6–23 months) in Gedeo zone, Ethiopia: cross - sectional study. Ital J Pediatr. 2021;47(1):233. 10.1186/s13052-021-01181-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Nyaradi A, Li J, Hickling S, Foster J, Oddy WH. The role of nutrition in children’s neurocognitive development, from pregnancy through childhood. Front Hum Neurosci. 2013;7:97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Patel A, Pusdekar Y, Badhoniya N, Borkar J, Agho KE, Dibley MJ. Determinants of inappropriate complementary feeding practices in young children in India: secondary analysis of National family health survey 2005–2006. Matern Child Nutr. 2012;8:28–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rai RK, Kumar SS, Kumar C. Factors associated with minimum dietary diversity failure among Indian children. J Nutr Sci. 2022;11:e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Rathnayake KM, Madushani P, Silva K. Use of dietary diversity score as a proxy indicator of nutrient adequacy of rural elderly people in Sri Lanka. BMC Res Notes. 2012;5:1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Rukundo PM, Rukooko B, Andreassen BA, Iversen PO. Housing, water and sanitation implications on food insecurity and diet diversity in landslide affected communities: a cross-sectional survey of two districts in Uganda. Clin Nutr ESPEN. 2019;33:47–56. [DOI] [PubMed] [Google Scholar]
- 26.Saha J, Chouhan P, Malik NI, Ghosh T, Das P, Shahid M, et al. Effects of dietary diversity on growth outcomes of children aged 6 to 23 months in India: evidence from National family and health survey. Nutrients. 2022;15(1):159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Saxena NC. Hunger, under-nutrition and food security in India. Springer. 2018.
- 28.Seboka BT, Hailegebreal S, Yehualashet DE, Gilano G, Kabthymer RH, Ewune HA, et al. Exploring spatial variations and determinants of dietary diversity among children in Ethiopia: spatial and multilevel analysis using EDHS (2011–2016). J Multidiscip Healthc. 2021. 10.2147/JMDH.S327456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Seboka BT, Hailegebreal S, Mamo TT, Yehualashet DE, Gilano G, Kabthymer RH, et al. Spatial trends and projections of chronic malnutrition among children under 5 years of age in Ethiopia from 2011 to 2019: a geographically weighted regression analysis. J Health Popul Nutr. 2022;41(1):28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Senarath U, Godakandage SS, Jayawickrama H, Siriwardena I, Dibley MJ. Determinants of inappropriate complementary feeding practices in young children in Sri Lanka: secondary data analysis of demographic and health survey 2006–2007. Matern Child Nutr. 2012;8:60–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Solomon D, Aderaw Z, Tegegne TK. Minimum dietary diversity and associated factors among children aged 6–23 months in Addis Ababa, Ethiopia. Int J Equity Health. 2017;16:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Temesgen H, Negesse A, Woyraw W, Mekonnen N. Dietary diversity feeding practice and its associated factors among children age 6–23 months in Ethiopia from 2011 up to 2018: a systematic review and meta-analysis. Ital J Pediatr. 2018;44:1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Tessema ZT, Azanaw MM, Bukayaw YA, Gelaye KA. Geographical variation in determinants of high-risk fertility behavior among reproductive age women in Ethiopia using the 2016 demographic and health survey: A geographically weighted regression analysis. Archives Public Health. 2020;78:1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Tigabu S, Liyew AM, Geremew BM. Modeling spatial determinates of teenage pregnancy in Ethiopia; geographically weighted regression. BMC Womens Health. 2021;21(1):254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wang A, Scherpbier RW, Huang X, Guo S, Yang Y, Josephs-Spaulding J, et al. The dietary diversity and stunting prevalence in minority children under 3 years old: a cross-sectional study in forty-two counties of Western China. Br J Nutr. 2017;118(10):840–8. [DOI] [PubMed] [Google Scholar]
- 36.WHO, Usaid U, A., AED U. Indicators for assessing infant and young child feeding practices. Geneva: World Health Organization; 2008. [Google Scholar]
- 37.Worku T, Gonete KA, Muhammad EA, Atnafu A. Sustainable under nutrition reduction program and dietary diversity among children’s aged 6–23 months, Northwest Ethiopia: comparative cross-sectional study. Int J Equity Health. 2020;19(1):14. 10.1186/s12939-019-1120-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.World Health Organization. (2008). Indicators for assessing infant and young child feeding practices: Part 1: definitions: conclusions of a consensus meeting held 6–8 November 2007 in Washington D.C., USA. Indicadores Para Evaluar Las Prácticas de Alimentación Del Lactante y Del Niño Pequeño: Conclusiones de La Reunión de Consenso Llevada a Cabo Del 6 al 8 de Noviembre de 2007 En Washington, DC, EE.UU, 19.
- 39.Zongrone A, Winskell K, Menon P. Infant and young child feeding practices and child undernutrition in Bangladesh: insights from nationally representative data. Public Health Nutr. 2012;15(9):1697–704. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
No datasets were generated or analysed during the current study.