Correlates of stunting among under-five children in Bangladesh: a multilevel approach

The data used for the present study has been derived from BDHS, 2014, a nationally-representative survey conducted by the National Institute for Population Research and Training (NIPORT) of the Ministry of Health and Family Welfare [45]. The 2014 BDHS collected data on the nutritional status of children by measuring the height and weight of all children aged between 0 and 59 months in the selected households. The sample for the BDHS-2014 is nationally representative and covers the entire population residing in non-institutional dwelling units in the country. The survey used a sampling frame from the list of enumeration areas (EAs) of the 2011 Population and Housing Census of the People’s Republic of Bangladesh, provided by the Bangladesh Bureau of Statistics (BBS). The primary sampling unit (PSU) for the survey is an EA created to have an average of about 120 households. With the design, the survey selected 18,000 residential households, which were expected to result in completed interviews with about 18,000 ever-married women of age 15–49 years. A total of 17,863 ever-married women of age 15–49 years were interviewed, for a response rate of 98% (Fig. 1). Details about the survey collection method can be found in literature [46]. For the current study, the data has been screened for the under-five children for whom height and weight are available (Fig. 2). Therefore, the study uses a sample of size 6965 extracted from 7886 of BDHS 2014 data.

Three standard indices of physical growth that describe the nutritional status of children are: (i) Height-for-age, (ii) Weight-for-height, and (iii) Weight-for-age. Each of these indices provides different information about growth and body composition that can be used to assess nutritional status of under-five children. The nutritional status of children in the survey population is compared with the World Health Organization (WHO) Child Growth Standards to create z-scores [46] resulting the indicators as stunting, wasting and underweight. Standard deviations below the reference median of the WHO reference population in terms of height-for-age was defined as stunted (− 2 SD) or severely stunted (− 3 SD). Similarly, standard deviations below the reference median of the WHO reference population in terms of weight-for-age was defined as wasted (− 2 SD) or severely wasted (− 3 SD); standard deviations below the reference median of the WHO reference population in terms of weight-for-height was defined as underweight (− 2 SD) or severely underweight (− 3 SD). However, z-scores computed for the three health indicators height-for-age, weight-for-height, and weight-for-age were used in this study.

Size of child at birth or a child’s birth weight was recorded based on mother’s perception [46]. A child’s birth weight less than 2.5 kg was considered as less than average, and lager than 2.5 was considered as larger than average. Mother’s BMI was classified as underweight if it was less than 18.5 kg/m², normal if it was between 18.5–24.9 kg/m², overweight if it was between 25 and 29.9 kg/m², and obese if it was above 30 kg/m². For the educational level, primary level was defined as completing grade 5, secondary level was defined as completing grade 10. The wealth index was created in three steps. In the first step, a subset of household indicators (for example, access to safe drinking water, sanitation facilities, structure of housing, crowdedness of housing, type of fuel used for cooking, household possessions etc.) common to urban and rural areas was used to create wealth scores for households in both areas. Categorical variables were transformed into separate dichotomous (0–1) indicators. Those indicators and those that were continuous were then examined using a principal components analysis to produce a common factor score for each household. In the second step, separate factor scores were produced for households in urban and rural areas using area specific indicators. The third step combined the separate area-specific factor scores to produce a nationally applicable combined wealth index by adjusting area-specific scores through a regression on the common factor scores. The resulting combined wealth index had a mean of zero and a standard deviation of one. Once the index was computed, national-level wealth quintiles (from lowest to highest) were obtained by assigning the household score to each de jure household member, ranking each person in the population by his or her score, and then dividing the ranking into five equal categories, each comprising 20% of the population. Computations for wealth index were done by the authority of the survey and details can be found in survey literature [46]. In this study, stunting was used as the indicator of nutritional status of under-five children and used as the response variable in generalized linear regression analysis.