Menu
Menu
Menu
Menu
Background: Caesarean section (CS) is an important medical intervention for reducing the risk of poor perinatal outcomes. However, CS trends in sub-Saharan Africa (SSA) continue to increase yet maternal and neonatal mortality and morbidity remain high. Rwanda, like many other countries in SSA, has shown an increasing trend in the use of CS. This study assessed the trends and factors associated with CS delivery in Rwanda over the past two decades. Methods: We used nationally representative child datasets from the Rwanda Demographic and Health Survey 2000 to 2019–20. All births in the preceding 3 years to the survey were assessed for the mode of delivery. The participants’ characteristics, trends and the prevalence of CS were analysed using frequencies and percentages. Unadjusted and adjusted logistic regression analyses were used to assess the factors associated with population and hospital-based CS in Rwanda for each of the surveys. Results: The population-based rate of CS in Rwanda significantly increased from 2.2% (95% CI 1.8–2.6) in 2000 to 15.6% (95% CI 13.9–16.5) in 2019–20. Despite increasing in all health facilities over time, the rate of CS was about four times higher in private (60.6%) compared to public health facilities (15.4%) in 2019–20. The rates and odds of CS were disproportionately high among women of high socioeconomic groups, those who resided in Kigali city, had multiple pregnancies, and attended at least four antenatal care visits while the odds of CS were significantly lower among multiparous women and those who had female babies. Conclusion: Over the past two decades, the rate of CS use in Rwanda increased significantly at health facility and population level with high regional and socio-economic disparities. There is a need to examine the disparities in CS trends and developing tailored policy guidelines to ensure proper use of CS in Rwanda.
Rwanda is a low-income, agricultural and landlocked country with approximately 11 million people living in five regions covering an area of 26,338 km2 [21]. It has an average of 4.4 persons per household [22] and a gross domestic product per capita of US $780.80 [23]. About half (48%) of its population is under 19 years of age and 39% live below the poverty line with a life expectancy at birth of 71.1 years for women and an adult literacy rate of 80% among 15–49 years old women. In addition, 87.3% of the population has health insurance and access to health services; spending an average of 47.4 min to reach a health centre [21]. In 2016, CS in Rwanda were conducted in 27 (75%) of the 36 districts, provincial and referral hospitals [21]. The study used the child datasets from the Rwanda Demographic Health Surveys (RDHS) conducted in 2000, 2005, 2010, 2014–15 and 2019–20 using stratified, two-stage cluster sampling [22, 24–27]. Households were stratified into urban or rural and all eligible women 15–49 years in selected households were interviewed using standard DHS questionnaires. All babies born within the preceding 3 years of each survey and with complete data were included in the population-based analysis while only babies delivered at a hospital were included in the hospital-based analysis. Of the 75,777 children born within the 3 years preceding each survey, 34,144 children were included after excluding 41,633 children with missing observations in the outcome variable as explained by the guide to DHS statistics [28] (Suppl. Fig. 1). We conceptualized our study variables using the framework adapted from Kizito and Schuemacher [29] as shown in Fig. Fig.1.1. The outcome variable was delivery by CS, which was categorized into “Yes” or “No”. Women were asked if they had been delivered by CS within the 3 years preceding the survey. Since it is possible for women to have more than one CS in 3 years, we used the participants’ unique identifiers and weighted samples to account for the clustering of CS. The explanatory variables included in the study were identified from a review of literature on factors associated with CS use [12, 14, 30–35]. Supplementary Table 1 operationalises these variables. The explanatory variables were categorised into maternal, child and household characteristics. Intervening variables were a set of variables acting on the explanatory variables and included access to information (Yes or No), place of delivery (private, public and home/others) and antenatal care (ANC) attendance (< 4 and ≥ 4 visits and missing). Maternal characteristics included residence (urban or rural), maternal age at birth (< 20, 20–34 and ≥ 35 years), education status (no formal, primary and secondary or higher), marital status (in-a-union and not-in-a-union), occupation (not working, agricultural and formal employment), parity (1, 2–4, ≥5) and region (East, West, South, North and Kigali City). Child characteristics included the weight of the baby (normal [2500-4000 g], low birth weight [ 4000 g]), low birth weight and big baby) [36], sex of the baby (male or female) and type of pregnancy (singleton or multiple). Household characteristics included household income and partner’s education (no formal, primary and secondary or higher). Conceptual framework on factors associated with CS use (adapted from Kizito 2021) We performed analyses of datasets using Stata version 17.0 (Stata Corporation, College Station, TX). We applied design based analysis using DHS sample weights and adjusted for sample errors using svy command. Participants’ characteristics and trends in the prevalence of CS were analysed using frequencies and percentages. To assess the factors associated with population-based and hospital-based CS, bivariate and multivariable logistic regression models were performed for each of the surveys. Variance inflation factor was used to assess multicollinearity. All variables included in the unadjusted model were hierarchically included in the adjusted model due to their importance in explaining CS and guided by the conceptual framework by Kizito and Schuemacher [29]. We reported both the unadjusted and adjusted odds ratios and considered significance at a p-value of less than or equal to 0.05. Only children with complete data were included in the analyses. The reporting in this study were guided by the STROBE guidelines for cross-sectional studies (Suppl. Table 3) [37]. The study used anonymised open-access secondary data from the RDHS, which received ethical approvals from the Rwanda National Ethics Committee and the Institutional Review Board of ICF International. The data were accessed upon approval of data request to the DHS program and were used as per the data agreement. Additional information on the ethical approvals and processes for the surveys can be obtained from the published reports [21, 23–25].
