Types of anthropometric failures among children
An overall CIAF prevalence of 55.32% among children was found as opposed to 38.4% (stunting), 35.7% (underweight) and 21% (wasting) in the NFHS-4 report. Among the CIAF categories, 22.15% of children suffer from only one form of anthropometric failure (groups B, F and Y), whereas 26.55% of children have simultaneous two failures (groups C and E) and 6.62% have all three forms of anthropometric failures. Simultaneous two failures of stunting and underweight (18.33%; group E) is the highest reported type of anthropometric failure. Overall, children of ST Poor appear disadvantaged both in rural and urban areas compared to SC Poor and Other Poor except for ST Poor Female Urban children; ST Poor Male Rural children have the highest proportion of simultaneous three forms of anthropometric failures (12.60%), whereas the ST Poor Male urban children have the highest proportion of children undernourished (71.90%). Both cross-tabulation of CIAF with wealth quintiles and the intersectional analysis shows that economic position is an important factor affecting the nutritional status of children measured by CIAF. There is a large difference in CIAF between children from the poor and non-poor economic position with non-poor advantage irrespective of gender, caste, and place of residences.
Table 1: Distribution of CIAF by intersecting sub-groups of Caste, Economic position, Gender and Place of Residence
Inequality in CIAF and all-three failures by economic position within caste groups
Concentration curve (CC) (figure 1) shows consistently higher economic position inequality in the severe form of undernutrition (simultaneous three failures) than aggregate CIAF index across all caste groups, indicating relatively more unfair clustering of simultaneous three failures across caste groups than with CIAF. While the highest proportion of children from ST suffered simultaneous three failures, the same community reported the lowest rate of inequality in nutritional status by wealth quintiles. The lack of economic position improvements among the ST community limits our ability to assess the role of economic position in improving the child nutritional status among ST. In OBC and general category (other) where the spread of households across all economic positions is seen, the rich-poor inequalities are higher (compare with ST and SC). Interestingly, the CC of SC demonstrated a shifting trend; while the poorest wealth quintile, of both SC and ST groups, coincide, higher up on the economic position gradient, the SC group makes a departure from this trend and shows increasing inequality and coincides with OBC and general categories. This means that the nature of inequality in child nutritional status among the SC poorest wealth quintiles is similar to that of the ST community whereas, among the upper wealth quintiles the inequality increases further indicating differential effects of economic position improvements within the SC category.
Figure 1: Concentration Curve for CIAF by Caste Category
Intersectional Inequality
A cross-tabulation of CIAF and all three failures revealed a significant difference by economic position, caste, gender and place of residence (see additional file 2; Table A). However, when we examined these differences by intersectional sub-groups of caste, economic position, place of residence, and gender, we found that these significant differences were not uniformly distributed. The intersectional sub-group comparison shows that economic position difference in all three failures among ST children is significant only in rural areas, whereas among the urban areas this difference disappears irrespective of gender. Among the SC children, except for urban male children, in all other groups, economic position-based differences appeared significant. Among the other caste groups (OBC and General combined) economic position-based difference is consistently significant irrespective of gender and place of residence. In CIAF, except for ST Female Urban, difference by economic position is consistently significant across all caste group irrespective of gender and place of residence (see Table 2).
Table 2: Economic Position Differences in All Three Failures and CIAF by Intersectional Sub-groups
In the caste-based comparison of all three failures, a significant difference between ST and SC in all three failures and CIAF was observed only among poor male rural. At the same time, this difference between ST and other caste is consistently significant only in rural areas irrespective of gender difference. Except for non-poor male rural, in all other sub-groups, the difference in all three failures between SC and other caste is not significant in all three failures. Whereas in CIAF this difference is significant in poor male rural, non-poor female rural, non-poor female urban, non-poor male rural (see additional file 2; Table B). Significant gender difference in all three failure is found only among rural, and in CIAF this significant differences disappeared in all the groups (see additional file 2; Table C). The significant rural and urban difference in all three failures and CIAF disappeared from all its intersectional sub-groups (see additional file 2; Table D).
District Based Undernutrition Hotspots in India
Based on >1 SD of the mean district prevalence of co-occurring two failures or simultaneous three-failures, (>23% for stunting and underweight, > 11.7% wasting and underweight, >9.5% all-three-failures; see additional file 1; Table A), critical (11), very serious (72) and seriously affected (28) districts were identified (see additional file 1; tables B – G). In all the critical districts, nearly half (>45%) of children reported at least two simultaneous anthropometric failures. Among these, the Dangs district from Gujarat reported the highest proportion (60.1%) of children with at least two failures. From the states of Bihar and Jharkhand two districts each and from Chhattisgarh, Karnataka, Madhya Pradesh, Rajasthan, Uttar Pradesh, and West Bengal one district each was reported as critical (see additional file 1; table B and C). Among the very serious districts, Pashchimi Singhbhum district in Jharkhand with 64.7% of children with at least two anthropometric failures nearly met the criteria for being a critical district (underweight and wasting prevalence 11.5%; all-three-failures 21.2%; and stunting and underweight 32%). The highest number of very serious districts are reported from Madhya Pradesh (19), followed by Jharkhand (13) (see additional file 1; Table D and E). Gujarat has the highest number of serious districts (8) followed by Odisha (4) (see additional file 1; Table F and G). Overall spatial distribution of critical, very serious and serious district-level prevalence shows geographical clustering of these districts in four undernutrition hotspots spanning over 12 high burden states. In south India, a cluster of eight districts in north Karnataka forms the undernutrition hotspot. The second undernutrition hotspot is eleven districts along the state boundaries of Chhattisgarh (4), Odisha (6) and Maharashtra (1). The third hotspot is spread across the regions spanning the borders between West Bengal, Bihar and Jharkhand consisting of 28 districts, of which 11 are from Bihar, 3 from West Bengal and 14 from Jharkhand. The fourth hotspot is 53 districts spanning across Madhya Pradesh (23), Rajasthan (7), Gujarat (17), Maharashtra (4) and Uttar Pradesh (2) (See figure 2).
Figure 2: Map of India showing the undernutrition hotspots; Critical districts = High prevalence in all three failures, stunting and underweight, and underweight and wasting. Very serious districts = High prevalence in all three failures, stunting and underweight or underweight and wasting. Serious districts = High prevalence in all three failures or high prevalence in stunting and underweight, and underweight and wasting.
Figure 2: Undernutrition hotspots in India
The Moran plot shows a linear fit through the point cloud. The slope of this line corresponds to Local Moran’s I values were 0.615 for stunting, wasting and underweight, 0.69 for stunning and Underweight, and 0.63 for wasting and underweight (see figure 3). All the coefficients were statistically significant (see additional file 2; figure A). This indicates that the three-dimensional and two-dimensional anthropometric failures among children in India is not uniformly distributed across Indian districts, rather there is significant clustering of the high prevalence of two-dimensional and three-dimensional failures in India, further strengthening the case for identifying hotspots.
Figure 3: Univariate LISA maps of India showing clustering of undernutrition hotspot and cold spot by two dimensional and three-dimensional anthropometric failures