Background
If a newborn weighs less than 2500 g at birth, it is termed as low birth weight (LBW) [
1]. A low birth weight carries an increased risk of death on the newborns early in life or exposes to multiple health and development challenges later. The burden of immediate health problems on low birth weight newborns has been relatively widely documented in many low income countries with national demographic surveys [
2].
An estimated 13% all babies each year in Sub-Saharan Africa (SSA) are born LBW [
3‐
6]. According to the Ethiopian Demographic and Health Surveys an estimated 13–11% of all babies in the country are born low birth weight [
7,
8]. The proportions have been documented with a wide variation across the different settings in the country. The prevalence has so far documented as low as 6.3 to 28.3% across Ethiopia [
7‐
13]. Thus the national prevalence of low birth weight has remained high in Ethiopia.
In Ethiopia, most commonly used data on birth weight of the newborns are often based on national survey indirect indicators. However, the national data is known for its limitations including sampling strategies which would not be assumed to be representative of the socio-economic disparities and cultural diversities across the country. Though the proportion of skilled childbirth attendance is increasing in Ethiopia, we lack adequate evidence on maternal factors which would potentially predict birth outcomes of the newborns. This is so limited particularly in South Ethiopia which is characterized by a diverse culture and dietary practices. Thus the proportion of low birth weight in health facilities has been least documented in South Ethiopia.
A densely populated Wolaita area in south Ethiopia [
14‐
19], many households often suffer from chronic food insecurity [
20]. Thus even at times of crop harvest, many children in the area remain undernourished. However, we have very limited evidence on child health and nutritional status predictors particularly attached to maternal health during pregnancy and child births. Thus the main aim of this study was to estimate the proportion of low birth weight among term singletons that were born without congenital malformations in Wolaita Sodo town in South Ethiopia. It was also further aimed to identify socio-demographic and maternal factors associated with the outcome. The findings might help for local community based interventions to improve birth outcomes, and also expected to add on the existing knowledge.
Methods
Study setting
Wolaita Sodo town is an administrative capital for the Wolaita zonal administration in South Ethiopia. It is located at 380 km south from Addis Ababa. The town has 3 sub-cities; 11 lower administrative units. It had a census projected population of 110,659; 48% were female.
One government hospital and another private one, 3 health centres, 11 health posts and 21 other private health facilities deliver health services to the population in the town. Out of all facilities; only health centres and hospitals are staffed to attend skilled childbirth [
21]. Based on the local health service administrators’ report for 2014/15, about 5000 childbirths were attended in government’s health facilities at Wolaita Sodo town.
Study design and period
We did a facility based cross-sectional study involving postpartum mothers and their newborns through birth records from March to April in 2016.
Population and sampling
All postpartum mothers and their newborns in government health facilities in Wolaita Sodo town were the source population. Consecutively recruited 432 mother-newborn-pairs were studied. Term newborns (born at 37 weeks or later of gestation) and singletons were included. Birth outcomes such as congenital malformations, multiple births, and stillbirths, and postpartum mothers with unknown last normal menstrual period were excluded.
A sample size of 451 was calculated with the following assumptions: 22.5% expected proportion of low birth outcome which was taken from similar Ethiopian study [
13], 95% confidence level, 5% margin of error, and 10% none-response consideration. It was proportionally allocated to government health facilities in Wolaita Sodo town based on number of previous childbirth attendance. Finally all eligible postpartum mother- to-child pairs were consecutively recruited until subsamples for the facilities and the total sample for the study were achieved.
Variables and measurements
Outcome
Birth weight of a term singleton newborn was an outcome measure. It was considered low birth weight if the newborn weighed less than 2500 g. Birth weight measurement was taken within an hour after birth [
1].
Exposure variables and covariates
Socio-demographic and socio-economic: maternal age, educational status, occupation, residence, family income and newborn’s sex [
22].
Diets and supplements: consumption of any additional meals and supplements such as iron and folic acid during pregnancy [
23].
Maternal morbidity and previous obstetric histories: history of hypertension, diabetes mellitus and infections during pregnancy [
24], pregnancy intervals, number of childbirth, low birth weight, abortions, and stillbirths [
24].
Maternal health service utilization: antenatal care (ANC), gestational age at first ANC visit, and dietary counselling in ANC [
22,
23]. Whereas, a woman needs to make at least four ANC visits during pregnancy [
25,
26].
Maternal nutritional status: Mid Upper Arm Circumference (MUAC) of mothers. Whereas, MUAC less than 23 cm centimetres defined the women as undernourished [
27].
Data collection
We adopted a structured questionnaire from relevant articles and related literatures (Additional file
1) [
7,
13]. The questionnaire was pretested in 23 respondents which were later not included to the main study. Four data collectors were trained on: the different modules of the questionnaire, participants’ selection, maternal MUAC and birth weight measurements, and ethics.
Anthropometric measures were standardised for Technical Errors against an expert measurer. Weights of newborns were measured within an hour after birth with a digital weight scale. The MUAC of mothers was measured by using a none-stretchable tape.
Statistical analysis
Data were entered, cleaned and analyzed by using SPSS version 20. Descriptive statistics were done for the main variables. Through binary logistic regression analysis we selected exposure variables with a crude association to the outcome. The multivariate candidate variables were prioritized by using a yardstick cut-off point for statistical significance on bivariate analysis. Thus exposures and covariates with p-value less than 0.25 were taken for multivariate analysis. Finally, a multivariate logistic regression analysis was done to control for potential confounders and identify independent predictors of the outcome. Accordingly, we reported AORs as effect measures with 95% CIs, and statistical significance declared at p-value < 0.05.
Discussion
Evidence shows that birth weight is a good summary measure of multifaceted public health problems. Thus often used to indicate long-term maternal malnutrition, ill health, and poor health care during pregnancy [
8]. This study documented that about 8.1% of term childbirths in government health facilities in Wolaita Sodo town ended up with low birth weight during the study period.
We observed a slightly lesser proportion of LBW in current study as compared to similar studies in Northern Ethiopia (10%) and elsewhere (10.6%) [
28,
29]. It was also lower than similar study findings from Pakistan (9.9%) and Kenya (12.3%) [
5,
30]. The main possible reason for the slight variation might be due to our exclusion of preterm newborns; whereas, the above studies included preterm babies. Conversely, we found a higher proportion of Term LBW compared to other studies in Ethiopia (6.3%), Iran (6.8%), and Egypt (7.3%) [
10,
11,
31,
32].
In consistent with studies from India, Tanzania, and Sudan, lower level of maternal education was documented in this study as a positive predictor of low birth weight [
4,
6,
33]. A better formal education might have improved mothers’ perceptions and dispositions in proper dietary habits during pregnancy and health service utilization. On the other hand, it might be assumed that the less educated would tend to be less informed on key dietary recommendations and health care during pregnancy [
34].
Unlike many other study findings, this study documented that rural residents had a lower risk to have low birth weight babies as compared to those from nearby urban settings. This finding disagrees with similar study findings from Tigray (Northern Ethiopia) and Bale (South east Ethiopia) hospitals in Ethiopia [
10,
11,
35] and also findings from Ghana [
36].These might possibly be due to less representation of our sample for rural and urban groups which made our finding with a higher uncertainty (a very wide confidence interval for AORs). It might also be due to socioeconomic and other differences. The current finding agrees with a study in Jimma west Ethiopia which could be homogenous in many aspects to our settings [
13].
Housewives had much higher odds of giving birth to low birth weight babies as compared to those who had some occupational engagements. This finding is in difference with similar studies South East in Ethiopia as well as Tanzania [
6,
34,
35]. On the other hand, in favour of our finding, a study from Iran revealed an increased risk of low birth weight among housewives as compared to employees [
37]. This might imply that housewives are mostly less educated, and often disadvantaged in accessing relevant health messages [
7,
8,
11]. This might have affected their perceptions in dietary preferences and health service utilization. Furthermore, our study setting, Wolaita area is known for asset poverty; many households in the area are often food insecure [
17,
20].
Mothers who did not consume fruits daily during pregnancy had much higher odds of giving birth to low birth weight as compared to those who did so. This was in agreement with a similar study in Nepal [
38]. This might be due to a complementary nutritional benefits of fruits as main food sources for micronutrients; thus if particularly consumed in the first trimester, they could enhance organ development for the foetus. Some studies elsewhere in Ethiopia, Bangladesh, Indonesia and Nepal have documented a significant association between low MUAC measures of mothers and term low birth weight in their newborns [
11,
39]; however our data did not imply such association. We recommend further studies to evaluate the effect of maternal undernutrition on newborn birth weight.
Study limitations
This study was limited to health facilities in scope and used a crossectional data. Thus mainly maternal side predictors of term low birth weight were assessed. Moreover, the time allotted for recruiting participants was limited though we assumed standard procedures of sample size calculation. The estimate might be better representative if a longitudinal follow-up data were used. Certain level of recall bias was expected with regard to menstrual dates and dietary habits; health workers conscious on cultural issues collected the data to reduce recall bias. Though all mothers attending delivery service with ANC follow up were ahead screened for HIV, we could not access this information due to limitations out of our scope. We have no data on other covariates such as food security and a reliable wealth data.
Conclusions
Though it would or would not be possible to bring all child bearing or potential mothers with less education to school; alternative opportunities of empowering women could be done. Occupational engagements of the women might also contribute to empower them on decision making capacities particularly on diet and health care.
Skilled counselling to diet during pregnancy including intake of fruits would benefit the women. Formal education when feasible and alternatively equivalent approaches targeting on diet and health care utilization would benefit. Moreover, we recommend community based data of birth weight through longitudinal studies for a better estimation of the prevalence of low birth weight.
Acknowledgements
We would like to thank the SENUPH_NORHED project for granting the study. We are grateful to the School of Public Health at WSU for facilitating the research programme. We would like to thank Mr.Wondimagegn Paulos (a focal person for SENUPH project at WSU) for facilitating the research grant. We thank Wolaita Zone health department and the local authorities for official permissions and administrative supports. We acknowledge our responsible data collectors and supervisors. We are grateful to the consent and commitment of mothers involved in the study just after their childbirth without which this study would not be possible.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (
http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.