Role of individual and household level factors on stunting: A comparative study in three Indian states

Abstract

Background: Status of growth especially in early childhood is not only the most important determinant of health of a child but also a reflection of the well-being of the entire society. The extent of malnutrition in India is very high, but the exact magnitude varies considerably depending on which indicator is used. Child health in this paper is measured through chronic malnutrition (termed as stunting). Three states were selected, namely Bihar, West Bengal and Kerala. These three states represent the three stages of development. Bihar is one of the least and Kerala is one of the most developed states in India.

Aim: The present paper aims to investigate the degree of chronic malnutrition in the context of socio-economic, demographic and other characteristics of the children and their households in the three selected states in India.

Subjects and methods: The data for this study were taken from the National Family Health Survey (NFHS-2) conducted by the International Institute for Population Sciences (IIPS), Mumbai, in 1998–1999. The NFHS-2 sample covers ever-married women in the age group 15–49 years from 26 states in India. Besides collecting information on health, the survey collects data on socio-economic and demographic characteristics at individual and household level.

Results: The percentage of stunting of children in Bihar, West Bengal and Kerala was found to be 54, 39 and 23%, respectively. Regression analysis showed that the major factors that significantly influenced the status of health in the children in all three states were women's education and the household condition index. Months of breastfeeding and birth interval also had some association with health status. The effect of the above-mentioned variables was most prominent in Bihar and least in Kerala in terms of statistical significance.

Conclusion: There is a close positive link between the nutritional status of pre-school children and the stages of development of the states. Mothers’ education and household condition are important influences on children's health status irrespective of the stage of development.

• Stunting
• pre-school children
• socio-economic variables
• BMI
• regression
• chi-square test

Notes

Notes

[1] Household condition index (HCI) is measured by an index that reflects the household condition in terms of sanitation, drinking source and family size. Each of these variables are given scores according to the degree of facility available or the amount of the variable as follows. Toilet facility: 0 for no facility, 1 for shared or public pit toilet, 2 for public or shared flush or own pit toilet and 4 for own flush toilet. Source of drinking water: 0 for source other than pipe, hand pump, public tap or well, 1 for public tap, hand pump or well, and 2 for private pipe lines, hand pump or well in residence/yard/plot. Family size: 0 for 11 or more members in the family, 1 for 8–10 members in the family, 2 for five to seven members in the family and 4 for less than five members in the family. The HCI can then be defined as the total score divided by the maximum possible score. In this paper we defined the HCI to be low if the total score is in the interval [0–2]; medium if the total score is in the interval [3–6]; and high if it is in the interval [7 and above].

[2] Since height-for-weight data are continuous it may be thought that the usual multiple linear regression model would be a more appropriate model. The usual linear regression model gives the expected change in the value of the response variable due to one unit change in the explanatory variable regardless of whether the child is malnourished or not. We have, in this paper, tried to find the probability of a child being chronic malnourished given the explanatory variable. Thus the logistic regression is more appropriate in this case.

[3] The NFHS data as shown in the all India report show slightly different figures (Bihar: 34%, Kerala: 7% and West Bengal: 14%). The first reason behind this discrepancy is that the data size has been greatly reduced in the present study. The reduction was necessary because some values were missing or there were outlying observations. One missing or erroneous observation forced the entire unit to be discarded, i.e. other observations of the unit are also deleted. The second reason is that only non-pregnant mothers were considered.

[4] There is a widespread belief that exclusive breastfeeding for the first few months of an infant's life ensures good physical growth. The actual age of supplementary food varies with respect to other factors. But the nature of our data prevents us from doing any fruitful analysis on this relation. We have only the present status of breastfeeding, and the effect of this should be different at different age groups. We do not have enough data to do this analysis. We should not give much weight to the results of statistical tests of feeding practices given in Table IV.

[5] The status of breastfeeding should be taken along with other inputs like supplemented food and age of the child. The interacting nature of these variables can not be analysed in this type of model. For this, one must resort to the Cox regression model.