Modelling Under Nutrition Among Underfive Children in Uttar Pradesh : a MULTILEVEL APPROACH
Mishra, S.; Pandey, S. K.
International Journal of Agricultural and Statistical Sciences 14(2): 483-490
2018
ISSN/ISBN: 0973-1903 Document Number: 255858
Undernutrition continues to be a major public health issue across the globe. At least half of the deaths worldwide are attributed to malnutrition. According to 2017 Global Hunger Index (GHI) Report, published by International Food Policy Research Institute (IFPRI), India ranks 100th out of 119 countries with a serious hunger problem. Nutrition-related factors were responsible for about 35% of child deaths and 11% of the total global disease burden. Three standard indices based on anthropometric measurements viz. weight and height, that describe nutritional status of children are: height-for-age (stunting), weight-for-age (underweight) and weight-for-height (wasting) [WHO (2006)]. According to National Family Health Survey-4 (NFHS-4), about half of the children under five years of age are stunted in the state, forty percent children are underweight and eighteen are wasted (low weight-for-height). Owing to the hierarchical structure of the NFHS data, multi-level random intercept logistic regression model has been applied to determine factors affecting undernutrition among children. Data on 36,036 children aged 0-59 months taken from the National Family Health Survey (NFHS-4) 2015-16 has been used for the analysis. Three-level random intercept logistic regression model [Marini and Gragnolati (2006), Longford (1993), Goldstein (2003)] was fitted for stunting, underweight & wasting, where level 1 were children, level 2 households and level 3 were the clusters which could be a complete Primary Sampling Unit (PSU) or its part, as defined by NFHS-4. For all the three nutritional indicators, Likelihood Ratio (LR) test comparing the multilevel against the standard logistic model was significant favouring the use of multilevel model for the used nested data set. The results also showed that although there was some reduction in the intra-class correlation coefficient across levels after adjusting for maternal and household level factors, however, there was not much reduction in the percent variance explained at each level indicating need to further explore more factors which remain unexplained like environmental etc. not considered in the given study. Maternal characteristics like nutritional status, educational status showed significant impact on child's nutritional status, consistent with the findings of other studies.