Regular antenatal care visits were associated with low risk of low birth weight among newborns in Rwanda: Evidence from the 2014/2015 Rwanda Demographic Health Survey (RDHS) Data

Background: Low birth weight (LBW) remains the global unfinished agenda in most countries of the world especially in low- and middle-income countries. LBW subsequently has harmful effects on the lifestyle, psychosocial and physiological development of the child. Although it is known that antenatal care (ANC) visits are important interventions contributing to prediction of newborn birth weight, little has been conducted on effect of ANC visits on birth weight in Rwanda. This study aimed at determining the association between regular ANC visits and risk of LBW among newborns in Rwanda. Methods: A cross-sectional study design was conducted to analyse the effects of ANC on LBW using the 2014/2015 Rwanda Demographic Health Survey. Associations of socio-demographic, socio-economic, and individual factors of the mother with LBW newborns were performed using bivariate and multiple logistic regression analyses. Results: Prevalence s of LBW and macrosomia were 5.8% and 17.6%, respectively. Newborns delivered from mothers attending fewer than four ANC visits were at almost three-times greater risk of having LBW [aOR=2.8; 95%CI (1.5-5.4), p=0.002] compared to those whose mothers attending four or more ANC visits. Residing in a rural area for pregnant women was significantly associated with LBW [aOR=1.1; 95%CI (0.7-1.6), p=0.008]. Maternal characteristics, such as anemia, predicted an increase in LBW [aOR=3.5; 95%CI (1.5-5.4),p<0.001]. Those who received no nutritional counseling [aOR=2.5; 95%CI (2-8.5), p<0.001] and who were not told about maternal complications [aOR=3.3; 95%CI (1.5-6.6), p=0.003] were more prone to deliver newborns with LBW than those who received them. Pregnant women who received iron and folic acid were less likely to have LBW newborns [aOR=0.5; 95%CI (0.3-0.9), p=0.015]. Conclusion: ANC visits significantly contributed to reducing the incidence of LBW. This study underscores the need for early, comprehensive, and high-quality ANC services to prevent LBW in Rwanda. The fifth RDHS 2015 was utilized as a nationally representative sample implemented by the National Institute of Statistics of Rwanda (NISR) and Ministry of Health of Rwanda. The study design was a secondary analysis of cross-sectional survey data from RDHS 2014/2015 that was retrospectively carried out for investigating the effects of antenatal care visits on birth weight of the newborn in Rwanda. The RDHS data collection fieldwork was conducted from November 9, 2014, to April 8, 2015. The data entry, editing, and cleaning was completed by May 15, 2015, and the final survey report was completed in March 2016. A total of 8,004 pregnant women who were to receive antenatal care interventions before delivery were recruited. The interviewed women were of reproductive age (15–49 years). This study was conducted in Rwanda, a small country located in the Central and Eastern Africa bordered by the Republic Democratic of Congo to the West, Uganda to the North, Tanzania to the East and Burundi to the South. This country lies a few degrees south of the equator and is landlocked. Concerning ANC visits, accessibility to ANC services is increasing due to the improvement of the health system and health financing 1 . This health system contributes to the achievement of SDG-III those targets reducing morbidity and mortality of mothers and children worldwide specifically in LMICs. The total area of Rwanda is approximately 26,338 km 2, the Rwandan population density around 416 people per km 2 and the total population is roughly 10.8 million. The majority (43%) of the Rwandan population is aged 15 years or less. Women accounted for about 52.6% of the population, 84% of Rwandans resided in the rural setting, and 71% participant in agricultural activities 40 . RDHS was a national survey conducted to assess the birth weight of newborns. To collect the data of this household-based survey, mothers who had the youngest children, age five years or less, were interviewed to provide data related to birth weight for their children. The data for this survey were collected using a two-stage sampling strategy for enrolling participants. These stages were cluster sampling design and the sampling frame. The sampling frame was composed of the list of the enumerators’ areas (EAs) that covered the entire country. All residents in selected households were eligible to be interviewed. At the first stage of this study, 492 clusters were randomly selected (113 in urban and 379 in rural areas). At the second stage of this study, the systematic sampling technique that focused on selecting the households was applied. Then, a fixed number of 26 households were selected randomly from each cluster and a total of 12,792 households were selected for the final sample for this study. Additionally, the proportional sampling technique was used in the survey where the sample for each cluster was equal. The study included women aged 15–49 years who were permanent residents of the households or visitors who stayed in the recruited household the night before the survey. Instead, the mothers were interviewed about the size of their children at birth because this determinant was found to be a proxy for the weight of the newborn. Therefore, 8,004 mothers with 15–49 years of reproductive age were interviewed for reporting the actual weight in kilograms using the written information about birth weight or recalling the weight at birth for their newborns. But our study inclucted 7381 women (92%) whose their newborns were measued weight. Therefore, 8% of the women whose newborns were not measured weight at birth were not enrolled in this study. Futher, all records on birth weight, number of ANC visits and BMI were available in the RDHS. RDHS 2014/2015 collected data at the national level using household-based survey data on birth weight retrospectively collected from the mothers. The data collection was completed by trained data collectors who used face-to-face interviews, asking mothers eligible for this study to provide a detailed birth history for children born in the preceding five years. Recruitment included stratified sampling, two stages of cluster sampling design. The first stage was characterized by selecting the participants from the samples frame constructed from enumeration whereas the second stage involved the systematic sampling of the households. These were listed from each cluster to ensure that an adequate number of the completed individuals were obtained 41 . Participants were interviewed based on the measurement of the DHS program. Birth weight was recorded in the RDHS using metric measurement (in kilograms) for all participants from the entire stratum of the country. Data from mothers with stillbirths were excluded from this study. Bias refers to any tendency or deviation from the truth in study design, data collection, recruiting participants, data analysis, and results interpretation. Generally, bias may occur at any stage of the research. To manage the bias for the data from RDHS, the authors systematically did data cleaning and removed the missing variables. All authors checked several times the selected variables to include in the analysis for minimizing all possible systematic errors that could occur in the study. Dependent variable. The outcome variable of the current study was birth weight of the newborns. As per World Health Organization (WHO) classification, newborns weighing <2,500 grams were categorized as LBW while newborns weigh >2,500 grams were categorized as not having LBW 42 . Independent variables. Based on the literature review and the structure of the RDHS 2014/2015 dataset, the independent variables were found. The main independent variable was the number of the ANC visits for the pregnant women. Although we expected to use a cut-off of 8 ANC visits as recommended by the WHO, a low prevalence (1%) of utilising performing 8 recommended ANC visits did not allow to use this appropriate recoommendation. Thus, we considered a cut of 4 ANC visits and considered that the pregnant women who attended less than 4 ANC visits and those who attended 4 and above ANC visits were inadequate and adequate respectively. As recommended by WHO in 2010, the pregnant women who attended 4 ANC visits were considered to have obtained extremely adequate healthcare that effectively contributes to the health of the mother and unborn 43, 44 . This study used different covariate variables selected based on the previous epidemiological studies, reviewing the suitable published studies and the available information provided in the demographic health survey (DHS) datasets with the consideration of the potential confounders. Based on the insights from the literature and availability in the datasets, such factors are socio-demographic data such as maternal maternal age, residence, educational attainment, household wealth status, place of delivery, marital status, maternal occupation, gender of the child, sex of household head. In additional to independent variables, we also had linear variables that compromise the variables such as body mass index (BMI), anemia and nutritional supplements including tetanus injection during the pregnancy, iron folic supplementation, and nutritional counseling during pregnancy. Before analysis, the observations with missing data were dropped. Statistical analysis was performed using descriptive (such as frequency, percentage) and analytical analyses. In the analytical analysis, bivariate logistic regression analyeses were performed and all significant explanatory variables at p<0.25 were included in the multivariate logistic regression models based on the odd ratios to determine the associated factors of LBW, presenting adjusted odd ratios with a consideration of 95% for the confidence intervals. Further, all determinants in the multivariate logistic regression models were assessed for collinearity, which was considered present if the study variables had a variance inflation factor (VIF) higher than 3. Therefore, we adjusted sampling based on the RDHS data that were widely used and consistent data for assessing maternal and child health statistics at the national level using STATA software version 13 (RRID:SCR_012763) 45 . In this cross-sectional study design, we respected the guidelines outlined in the Strengthening the Reporting of Observational Studies in Epidemiology statement in writing the manuscript 46 . Data used were electronically accessed. To get full access, the first registration was completed on the DHS website. The permission to use the 2014/2015 RDHS data was granted by DHS using its website and the prior approval was maintained. In the prior approval, the women of reproductive age who were age 18–49 years provided oral and written informed consent forms to take part in the survey. In the cases on the minor participants (those women aged 15–17 years); the assent form was obtained from them while written informed consent were simultaneously provided by their guardians or parents who were adults.