Anemia and its associated factors among school-age children living in different climatic zones of Arba Minch Zuria District, Southern Ethiopia

A community-based cross-sectional study was conducted from April to May 2017 among SAC (6 to 14 years old) in Arba Minch Health and Demographic Surveillance Site (AM-HDSS), Arba Minch Zuria district, Southern Ethiopia. Arba Minch Zuria district has a total of 31 kebeles, among which 9 kebeles are included as the site for AM-HDSS. These nine kebeles, which are assumed to be the representative of all 31 kebeles in the district, were selected by stratifying based on their climatic zone (all the three different climatic zones: high land, midland and lowland, are included) and, in addition, one semi-urban kebele was also included. The HDSS has been running by Arba Minch University, Demographic and Health Research Center. The altitude of nine kebeles ranges from 1128 to 2866 above sea level. According to the HDSS report of January 2016, a total of 74, 157 individuals live in the surveillance site, of whom 20,062 (27.1%) were aged between 6 and 14 years.

The sample size was estimated using single population proportion formula through assumptions of the prevalence of anemia among SAC in Jimma, which was 37.6% [16] at 95% confidence level (CI), 5% margin of error and adding 10% for non-responses. Finally, the study recruited a total of 397 children. All children aged 6 to 14 years old living in the nine kebeles of AM-HDSS were included as a source population. To select the study participants, AM-HDSS database, which contains the list of all individuals in the kebeles with their date of birth, individual and household identification, was used as a sampling frame. Then, considering the number of children living in each kebele, study participants were proportionally allocated. Study participants were selected following a simple random sampling technique using Stata version 14 software.

Data were collected using structured, pretested Amharic version questionnaires. The study instruments included two sets of structured questionnaires. One was administered for children, and another for their family (children guardian preferably the mother). Data on family wealth, as a measure of economic status, was collected by asking ownership of selected assets that are common in the local area based on Ethiopian Demographic and Health Survey (EDHS) 2016 wealth index variables [17]. The household food insecurity level was measured using Household Food Insecurity Access Scale (HFIAS), which is a structured, standardized and validated tool that developed by Food and Nutrition Technical Assistance Project (FANTA) [18]. The child dietary intake pattern, for assessing the child dietary diversity, was measured by a qualitative recall of all foods consumed by each child during the previous 24 h based on Food and Agriculture Organization (FAO) Guideline. In addition, to assess the usual nutrient intake of the child, the frequency of consumption of certain nutrients 1 week prior to the survey was recorded [19]. Altitude, where each child resides, was measured to describe the malaria transmission pattern in the study area and to aid in the adjustment of the hemoglobin concentration of the child.

Weights of children were measured to the nearest 0.1 kg using calibrated portable electronic digital scale (Seca, Germany model) and heights were measured to the nearest 0.1 cm using a portable height-measuring board with a sliding head bar following standard anthropometric techniques [20].

A questionnaire was prepared in English and translated into Amharic language and again translated back to English. It was pretested by recruiting 20 children and their guardians in one of HDSS kebeles on households which were not selected for actual data collection. Data collectors were given 2 days of intensive training on the method of identifying the household, anthropometric measurements, blood and stool sample collection and examination, ethical issues and the purpose of the study. Hemoglobin measurement and stool microscopic examination were done following standard protocols. Investigators were supervising all aspects of data collection throughout the data collection period.

Data were entered using Epi Data version 3.1. For further analysis, the data was exported to Statistical Package for Social Science (SPSS) version 22 software. Descriptive analysis was indicated using numerical summary measures. Households who experienced none of the food insecurity (access) conditions or just experiences worry, but rarely in the past 4 weeks were labeled as ‘food secured, while the remaining were labeled as food insecure [18]. To calculate the child dietary diversity score, certain food groups were aggregated and the mean score was used to classify child food intake as adequate or not [19].

The blood hemoglobin cutoff values were adjusted for altitude before determining the children either as anemic or not. The adjustment was made based on the application of formula: Hb adjustment = − 0.32 × (altitude in meters × .0033) + 0.22 × (altitude in meters × .0033)². Then, the adjustment was subtracted from the measured hemoglobin concentration at the relevant altitude to get the sea-level value [23‐25]. The hemoglobin concentration cutoff points for anemia were 11.5g/dl for children age from 6 to 11 years and 12 g/dl for children age from 12 to 14 years. Those children with hemoglobin concentration below 8 g/dl and between 8 and 10.9g/dl were further classified as severe and moderate anemia, respectively. The remaining anemic children with hemoglobin concentration between 11 and 11.4g/dl (for children age from 6 to 11 years) and 11–11.9 g/dl (for children age from 12 to 14 years) were categorized as mild anemia [26].

The wealth index was constructed via Principal Component Analysis (PCA) method [27]. For anthropometric data analysis, standard deviation (Z-scores) scores were obtained by the World Health Organization (WHO) Anthro Plus software. Children whose Height for Age z-score (HAZ) and BMI for Age z-score (BAZ) above-2SD scores were considered as well-nourished and those below -2SD scores as being malnourished (stunted and thin respectively) [28, 29]. Malaria transmission pattern in the study area was described based on the difference in altitude: malaria-free highland areas (above 2500 m altitude); highland fringe areas between 1500 and 2500 m affected by frequent epidemics; and lowland areas below 1500 m with seasonal patterns of transmission [30].

All covariates that were significant at p-value < 0.25 in bivariate analysis were considered for multivariable analysis to control possible confounders. To measure the strength of association between dependent and independent variables, Adjusted Odds Ratio (AOR) with 95% Confidence interval (CI) was calculated. Finally, the level of statistical significance was declared at p-value < 0.05. The fitness of the model was tested by Hosmer- Lemeshow goodness of fit test and the test revealed a p-value of 0.820 showing the model was fit.

A total of 391 SAC with their mothers/caregivers were participated in making a response rate of 98.5%. More than half of respondents, (65.0%) were in the age group of 6–11 years. The mean ± SD of the age of children was 10.1 ± 2.6 years while that of parents/ caregivers was 37 ± 8.7 years. Nearly half (50.1%) of the participating children were male and 223 (57.0%) were enrolled in school. Concerning the educational status of parents, 257 (65.7%) of mothers/ caregivers, had no formal education while 132 (33.8%) of fathers had formal education. A total of 370 (94.6%) mothers were housewives while 133 (34%) of the fathers were farmers.

In the study area, 130 (33.25%) of households were in the poor wealth terciles; 206 (52.7%) of households were food insecure and 159 (40.7%) of surveyed households were residing in kebeles with an altitude < 1500 m where seasonal transmission of malaria is common (Table 1).

Out of all children included in this study, 163 (41.9%) were with height for age < −2Z score and 31 (8.0%) were with BMI for age < −2Z score. The detail on the anthropometric status of the child is found in our previously published article [31]. The stool samples were collected from 391 children among which 182 (46.5%) of children were positive for intestinal parasites and the remaining 209 (53.5) were negative. Ascaris lumbricoides (14.2%) and hookworms (14.1%) being the predominant species. Only 16 (4.1%) of children reported the presence of illness 2 weeks prior to the survey. Nearly all children, 364 (93.1%) had no habit of consumption of meat or meat products 1 week prior to the survey. More than half of children, 238 (60.9%), had a Dietary Diversity Score (DDS) of ≥ 4 food groups (Table 2).

Before the adjustment of children hemoglobin concentration for altitude, the mean ± SD hemoglobin level of study participant was 13.1 ± 1.8. After adjustment, the mean ± SD hemoglobin level of study participant was 12.3 ± 1.8. The overall prevalence of anemia among SAC was 37.3% (146); (95% CI: 32.5, 42.2). Among those who were anemic, 110 (28.1%) (95% CI: 23.8, 33.0) had mild form of anemia, 35 (9%) (95% CI: 6.1, 11.8) moderate form of anemia, and a single case of severe anemia (0.3%) (95% CI: 0.0, 0.8) was identified. The prevalence of anemia was 57 (46.3%), 40 (36.7%) and 49 (30.8%) in an altitude ranges of ≥2500 m, between 1500 and 2500 m and <  1500 m respectively, even though, the difference is not statistically significant when adjusted for other covariates.

The present study identified major modifiable associated factors for anemia. Children who were positive for intestinal parasitic infections were 3.3 times more likely to be anemic than those negative for intestinal parasitic infection (AOR = 3.30, 95% CI: 2.04, 5.35). The likelihood of anemia was also higher among children not-enrolled in schools (AOR = 2.05, 95%CI: 1.26, 3.32) than those children who were enrolled. When compared with children who had the habit of eating vegetables for more than 3 days in 1 week prior to the survey, anemia was less common among children who had no habit of eating vegetables in 1 week prior to the survey (AOR = 0.35, 95%CI: 0.14, 0.84) (Table 3).