This study comprehensively analyses and maps the spatial variability and spatial spread (inequality) in nutritional outcomes among young children residing in the 277 high-focus districts in the EAG region. We used data from the fourth update of National Family Health Survey (NFHS-4), from 2015 to 2016, and adopted a Bayesian distributional structured additive regression (STAR) approach to quantify the variations in the mean level and dispersion in the undernutrition indicators across space and based on other demographic and socioeconomic indicators available to the child, mother, and household. Markov random field and Bayesian P-splines were used as prior distributions for the spatial and nonlinear effects, respectively. Substantial intra-district heterogeneity in childhood nutritional deficiency found in EAG region; the districts where dispersions are huge, are predominantly located in three states: Uttar Pradesh, Madhya Pradesh, and Bihar. Our findings provide evidence of strong spatial structure that suggests neighbouring districts from different states have similar pattern of distribution in terms of undernourished children. The effects of child's age, mother’s age and her BMI were found to be non-linear. Furthermore, we find significant linear effects of several categorical covariates, such as, gender, birth order, birth size, maternal schooling, household wealth and sanitary condition, on indicators of undernutrition in young children. Findings strongly recommend that nutrition improvement programmes in the EAG states need to account for internal spatial heterogeneity evident across the districts of this region. Efforts are needed in assessing the district-specific causes of nutritional deficiency among the young children.

本研究全面分析并绘制了 EAG 地区 277 个重点地区幼儿营养结果的空间变异性和空间分布（不平等）图。我们使用 2015 年至 2016 年第四次全国家庭健康调查 (NFHS-4) 更新的数据，并采用贝叶斯分布结构化加性回归 (STAR) 方法来量化各地区营养不良指标的平均水平和离散度的变化。空间并基于儿童、母亲和家庭可获得的其他人口和社会经济指标。马尔可夫随机场和贝叶斯 P 样条分别用作空间和非线性效应的先验分布。 EAG 地区儿童营养缺乏存在显着的地区内异质性；分散程度较大的地区主要位于三个邦：北方邦、中央邦和比哈尔邦。我们的研究结果提供了强有力的空间结构的证据，表明不同州的邻近地区在营养不良儿童方面具有相似的分布模式。孩子的年龄、母亲的年龄和她的体重指数的影响被发现是非线性的。此外，我们发现几个分类协变量（例如性别、出生顺序、出生大小、母亲受教育程度、家庭财富和卫生条件）对幼儿营养不良指标的显着线性影响。研究结果强烈建议 EAG 各州的营养改善计划需要考虑到该地区各地区明显的内部空间异质性。需要努力评估幼儿营养缺乏的地区特有原因。