Study setting and design
A systematic review and meta-analysis was conducted to determine the pooled prevalence and associated factors of C-section in Ethiopia. Ethiopia is found in East Africa and currently, the Ethiopian population is estimated to be 112,555,079.
This review was prepared according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) (13). To identify eligible articles, all-embracing search was performed with English language restriction from the following international electronics database: PubMed, MEDLINE, EMBASE, CINAHL, Google Scholar, and Science Direct. In addition to this, national database like an Ethiopian medical journal, Addis Ababa University Digital Library and Haramya University Digital Library were searched. Assuming that such language restriction does not alter the outcome of the systematic reviews and meta-analyses (14). Snow ball searching was done from the list of references of the eligible studies to included missed articles from our search strategy.
First, we have searched studies by writing a title “prevalence and associated/determinant factors of caesarean section in Ethiopia” from each database. For instance, we searched using the following keywords: “prevalence”, “associated factors”, “determinant factors”, “cesarean section”, “caesarean section”, “delivery”, “mothers” and “Ethiopia”. The search terms were used separately and in combination using Boolean operators like “OR” or “AND”. We have searched articles from the above electronic database between August 1, 2019 and October 1, 2019. All papers published until August 1, 2019 was included. After identifying potentially relevant studies using our search strategy, studies were retrieved and managed using Endnote X9 software.
Only studies conducted in Ethiopia.
Only studies involving on mothers who have at least one birth.
Both published and unpublished full text accessible articles were included. Articles reported national survey or duplicated also excluded.
All observational prevalence study designs (i.e. cross-sectional, case-control and cohort) reporting the prevalence of C-section were eligible for this review.
Only articles reported in the English language were considered.
This review considered two main outcomes. The primary outcome variable of this study was the prevalence of C-section. It is defined as a prevalence of C-section, cesarean section, Caesarean delivery. The second outcome of this study was to identify factors associated with C-section of mothers among a reproductive age. We determined the association between C-section and associated factors in the form of the log odds ratio. Three major factors, which assessed by each studies were selected to explore their association with C-section. For each factor, the odds ratio was calculated or extracted based on the data reported by each study. The factors assessed in this review were delivery place type (private or public), previous history of C-section (Yes versus no) and gestational age (< 37, ≥ 37 week).
All necessary data were extracted by all authors (AG and AM) independently using a standardized data extraction form, which was adapted from the JBI data extraction format (15). Any disagreements during data collection were resolved through discussion. For the prevalence outcome; primary author, publication year, regions where the study was conducted, study area, sample size, study design and prevalence with 95% confidence intervals were extracted. For the latter outcome, data were extracted in a format of the odds ratio for each factor was extracted on the reports of original studies.
Risk of bias assessment
All authors (AG and AM) independently assessed the quality of each original study using the quality assessment tool. Joanna Briggs Institute (JBI) critical appraisal checklist for prevalence study was used which have nine questions (15, 16). Quality assessment was done for included studies in this review. The quality of each study categorized into three categories as “low risk of bias” (the sum of quality score ≥ 6), “medium risk of bias” (sum of quality score between 3 and 5, inclusively) and “high risk of bias” (sum of quality score less than 3). All published and unpublished studies having greater or equal to five quality score was included. Any disagreement between the two authors was resolved through discussion.
Data processing and synthesis
After essential data extracted, it was entered to Microsoft Excel, and then it exported to STATA Version 14.1 Statistical software to analyze. The existence of heterogeneity was assessed using a Cochrane Q test and I2 test statistics. For results with statistically significant heterogeneity (I2 ≥ 50, P–value < 0.05), Dersimonian and Liard random effect model was used (17). We used the Egger test statistics to assess publication bias. We used forest plot to present estimated value with the 95% confidence interval (CI). We conducted subgroup analyses by region, sample size, publication year, the risk of bias and study design. Additionally, association of C-section with previous history of C-section, place of delivery and gestational age was conducted.