Study design and setting
A community-based unmatched case-control study was conducted among children from 6–59 months of age residing in Wonago town, Gedeo Zone, SNNPR, Ethiopia. This town is located 370 km from Addis Ababa (the capital of Ethiopia) and 12 km from Dilla town (Gedeo Zone administrative center). The town is the administrative center of Wonago rural wereda. The town has a longitude of 382667 (3816′0.120″ E) and latitude of 63167 (619′ 0.012″ N) with altitude of 1776. The dominant ethnic group in the area is Gedeo and most of the people speak Gedeo-Offa language[31]. The town has two sub-divisions. This study was conducted from June 01-25, 2017
Population
In this study, cases (stunted children) were children with height/length –for- age z-score less than -2 SD and controls (non-stunted children) were children having height/length –for- age z-score greater than or equal to -2 SD according to WHO standards. Cases were randomly selected stunted children from among all stunted children of 6-59 months in the study area and controls were randomly selected non-stunted children among all non-stunted children of 6-59 months in the study area. Children with known underlying chronic illnesses and congenital or chromosomal abnormalities were excluded from the study.
Sample size determination and Sampling procedures
The sample size was determined using two population proportion formula by the Epi-Info 7.1 software, with the following assumptions:- p = 28%, which estimates the proportion of non-stunted children who are the outcome of an unintended pregnancy, α = 1.96 critical value at 95% confidence interval of certainty, OR= 2.1 , from literature review, stunted children were expected to come from unintended pregnancy when compared to non-stunted children, power of 80%, r = 1 which is, the ratio of cases to controls 1:1. Using the above assumptions, the sample size became 302, meaning is 151 for cases and 151 for controls.
With regard to the sampling procedure, primarily, a survey was conducted to identify the number of children under the age of five within the study area, and their respective nutritional status. The survey was carried out by trained health professionals who have previous similar experience on anthropometric measurement. They enumerated the children, coded them and measured height and weight of the children. Based on the nutritional status index derived from the height and weight measured during the survey, the total children in the study area were labeled into two groups (stunted and non-stunted) and each children were coded. Using simple random sampling method, 151 stunted children (cases) were selected from among all stunted children and 151 non-stunted children (controls) were selected from among all non-stunted children identified during the survey. This was done using table of random numbers for both groups (stunted and non-stunted children) separately. Fortunately, all of the respondents identified by simple random sampling from both group were successfully responded for the enquiry.
Data collection Instruments, personnel and quality assurance
The questionnaire was adapted from Ethiopian health and demographic survey[32] questionnaire employed for assessment of child malnutrition. It was further developed using peer reviewed published literatures to include determinants of malnutrition, including unintended pregnancy. These consists of socio-economic, socio-demographic, child characteristics, child caring practices, pregnancy intention and environmental health condition/sanitation. The questionnaire was further modified after a pretest was conducted. Finally, the tool was translated into local language for field work purposes and back to English for checking language consistency.
Weight was measured to the nearest 0.1 digits in kilograms with minimum clothing and no shoes. The salter spring scale with the capacity of measuring 25 Kg was used for younger children and the battery powered digital scale (SECA, UNICEF, Copenhagen) was used for older children. Weighing scales were calibrated with known weight objects regularly. The ace of scale indicator was checked against a zero reading after weighing every child. Height and length were measured using a standardized measuring board to the nearest 0.1 cm. Fourteen data collectors, who are diploma nurses, were recruited from the study area and priority was given to those with previous experience with data collection. Two supervisors were assigned to control the overall field work. The data collection was done through face-to face interview by trained data collectors for all participants (parents of the selected children).
Measurements
The weight and height of the children was converted into weight-for-age and height-for-age standard deviation units (z-scores) using ENASMART software based on WHO Child Growth Standards. The children were classified as stunted if their height for age z-score was below -2. Maternal pregnancy intention was determined by asking women to retrospectively recall their feelings at the time of conception for each birth within the past five years. Women were asked whether the pregnancy had been planned (wanted at that time), mistimed (wanted later), or unwanted (not wanted at any time). To measure pregnancy intention, the respondents were asked: “At the time you became pregnant with [name of last-born child], did you want to become pregnant then, did you want to wait until later, or did you want no more children at all?” Intended pregnancies were defined as those pregnancies to mothers who wanted the pregnancy at that time. Unwanted pregnancies were those to mothers who did not want to have more children; and mistimed pregnancies referred to mothers who wanted to become pregnant eventually, but at a later time. The questions and definitions used here are standard for reproductive health surveys conducted worldwide.
Underweight was defined as children with weight for age z-score less than -2 and wasting was defined as weight for height z-score less than -2. For measurement of deworming status, children who got age specific dose of albendazole or mebendazole within the last six months according to deworming protocol in Ethiopia.
Meal frequency was classified as adequate or inadequate based on the age specific number of meal frequency recommendation taken from WHO guideline of indicators for assessing infant and young child feeding practices [35] and OCHA indicators registry for older children [36].
Data analysis
Data entry was done using EpiData version 3.1 by single data entry method and exported to statistical package for social sciences (SPSS) version 20 for analysis. Descriptive statistics were computed for nutritional status, pregnancy intention, socio-demographic and socio-economic characteristics. The proportion of unintended pregnancy and other important variables were compared among case children and control children using chi-square test. Assessment of crude association between stunting and each independent variables at a time was conducted using univariate logistic regression analysis.
Multivariable logistic regression was used to assess the effect of independent variables including the pregnancy intention on the outcome variable (stunting). Crude and adjusted odds ratios with their corresponding 95% confidence intervals were computed. A p-value ≤ 0.05 were considered statistically significant in this study. Efforts were made to assess the fulfillment of the necessary assumptions for the application of multiple logistic regression. In this regard, the Hosmer and Lemeshow's goodness-of-fit test was done to check the model fit. Interaction between different predictor variables was checked using variable inflation factor (VIF). Particularly interaction between the educational level of family and household wealth was checked