4.1 Reporting of the review findings
The protocol for this review is registered at PROSPERO with registration number CRD42020158224. We will use the Preferred Reporting Items for Systematic review and Meta-analyses (PRISMA–2009) [27] (additional file 3) and (PRISMA-P 2015) [28] (Additional file 1) statements to report the findings. Moreover, we have used the guideline of the Prospero for registration.
4.2 Study design
A systematic review and meta-analysis will be conducted to determine the pooled prevalence of birth asphyxia and its determinants among newborns in Ethiopia.
4.2 Eligibility Criteria
4.2.1 Inclusion criteria:
Study Setting: This review will include studies conducted in Ethiopia.
Study Population: the review include only studies involving newborns with.
Study design: All observational studies (i.e. cross-sectional, case-control, and cohort) those have reported on prevalence of birth asphyxia and its determinants and have been published in peer-reviewed journals.
Language: studies reported in the English language will be considered.
Publication year: no restriction on year of publication
4.2.2 Exclusion criteria:
We will exclude studies published in languages other than English. We will exclude studies that include perinatal birth asphyxia. Additionally, we will exclude studies other than observational studies. case reports, conference reports, national survey reports and expert opinions.
4.3 PECO search guide
Population: Live newborns
Exposure: Predictors/determinants of birth asphyxia. The determinants are characteristics or exposures that increase the likelihood of birth asphyxia. These may be related to maternal residence, maternal age, educational status, newborn weight, and antenatal exposure.
Comparison: The reported reference group for each predictor/determinant in each study (e.g. birth asphyxia in mother living in urban area versus mothers living in rural area).
Outcome: We will include studies that assess perinatal birth asphyxia, if the studies have excluded stillbirth from perinatal mortality.
Searching strategy
This meta-analysis will be prepared and presented according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis [PRISMA, 2009]. We will develop an appropriate and comprehensive search strategy with relevant search terms and pilot test it before the final search. We will search MEDLINE, EMBASE, CINAHL, Scopus, Web science, HINARI, Google Scholar, and Cochrane electronic Databases. We will include articles published from start of indexing until May 30, 2020. We will use Medical Subject Heading (Mesh), keywords, and free text search terms. As the search terms, we will include alternative terms for neonatal mortality, and will combine them using Boolean operators.
Search terms (Additional file 2&3):: ((birth asphyxia OR “perinatal asphyxia” OR breathing difficulty at birth” OR low APGAR score) AND (determinants OR predictors OR “risk factors” OR causes) AND (newborn OR perinatal OR Neonate) AND (Ethiopia OR Ethio OR Etopia)
Moreover, we will utilize snowballing to screen the references of identified articles for potentially relevant studies. Furthermore, Studies identified by our database searching strategy will be retrieved and managed using Endnote X8 (Thomson Reuters, Philadelphia, PA, USA) software.
4.5 Outcome measurement
This systematic review and meta-analysis will have two main outcomes. The first outcome is estimating the pooled prevalence of birth asphyxia among the neonates. Birth asphyxia is defined as newborn with APGAR score less than 7 at the 1st and 5th minutes of birth. Pooled prevalence will be calculated by dividing the number of neonates with birth asphyxia to the total number of neonates who have been included in the study (total sample size) multiplied by 100. The second outcome of the study is to identify maternal and neonatal factors associated with birth asphyxia. The predictors will be determined in the form of the log odds ratio.
4.6 selection of studies
Two authors (AWT, SB and GB) will review the studies, based on inclusion and exclusion criteria. The review will follow three stages. In the first stage, reviewers will assess the titles of the studies identified from the search. In the second stage, abstracts of these selected titles will be included for the final stage of full-text screening. In the third stage, full-text screening, we will screen the full texts selected in the previous stage. If the articles are not open access, we will contact the corresponding author at least for three times. If the authors are not willing to provide the full text, we will exclude that specific article.
In the review, we will only include those studies approved by both authors. The authors will resolve disagreements through discussion or consultation with a third reviewer (KU). We will provide reason for exclusion for all excluded studies. Finally, we will prepare a final list of articles for data extraction
4.7 Data extraction
Three authors (AW, SE and GB) will independently extracted all necessary data using a standardized data extraction format, which is adapted from the JBI data extraction format for observational studies [29]. We will pretest the data extraction form on three studies of each type, to ensure that it adequately facilitates the collection of all necessary data required for an effective systematic review and meta-analysis. Discrepancies between data extractors will be discussed to reach consensus. If a consensus cannot be reached, the authors will consult a third reviewer (KU). For each included articles, we will record the first author’s last name, year of publication, the setting where the study was conducted, study design, study period, sample size, the response rate, the population, outcome definition, comparison groups, and the effect estimate.
4.8 Quality assessments
Four authors (AWT, SB, KU and GB) will independently conduct quality assessment of included studies, by using the checklist of the JBI appraisal tool for cross-sectional, cohort and case-control studies (Additional file 3).. In customising the scale to fit this study, we will took into account the study sampling methods and similarities between the study groups regarding adjustment for confounding factors, the ascertainment of exposure and outcomes, and study design. The abstracting tool will include different questions based on the study designs. The four investigators independently will perform the quality assessment while abstracting the data for the meta-analysis. The quality scores of the four abstractors will be averaged. Finally, the studies with higher scores (>50%) will be included into meta-analysis.
4.9 Data synthesis and analysis
The extracted data will be entered into a Microsoft Excel Database and then imported into STATA version 14.0 (Stata Corp LLC, Texas, USA) software with packages of Meta-analysis for further analysis. The researchers will perform a narrative description of the study population, the studies included, the risk factors identified, and the cause for birth asphyxia as well as the outcome characteristics. We will use tables and figures to summarize the selected studies and results.
The pooled prevalence of birth asphyxia in Ethiopia will be demonstrated. We will use fixed effect model if the studies have similar methodology, same population, and study design. If not, we will utilize the random effect model [30]. Then, the “metaprop” command with random-effects models at 95% confidence intervals will be used to calculate the pooled prevalence of birth asphyxia at national level [31]. The Freeman Tuckey variant of the arcsine square root transformation of proportions will be fitted to avoid variance instability when handling proportions close to one [32]. We will assess heterogeneity by using chi-squared test on Cochran’s Q statistic with a 5% level of statistical significance [26] and I2 statistic test [33], assuming that I2 value of 25%, 50%, and 75% being representative of low, moderate, and high heterogeneity, respectively [33]. If the heterogeneity is significant (I2 > 75%), then, we will conduct subgroup analyses and meta-regression to investigate sources of heterogeneity.
Publication bias will be examined by the visual inspection of funnel plots [34] and Egger’s test [35]. A p value < 0.10 will be considered indicative of statistically significant publication bias. Thus, if there is evidence of publication bias, we will use Duval and Tweedie’s trim-and-fill method [36].
For factors associated with birth asphyxia; two-by-two tables will be constructed (if possible), the odds ratio with 95% confidence interval will be calculated. Then, the statistically significance level will be declared at a p-value less than 0.05. However, if the meta-analysis is not possible, we will conduct narrative synthesis.
4.10 Subgroup and sensitivity analyses
Sub-group analysis will be performed based on study design, sample size, regions or state, year of publication, quality of studies, and study settings of included studies. by years of publication, study setting, study design, and sample size of the studies. Finally, to conduct sensitivity analysis, we will assess the stability or robustness of the pooled estimates to outliers and the impact of individual studies [36].
4.11 Discussion
This review will provide a detailed summary of the evidence on prevalence and factors of birth asphyxia in Ethiopia. This review will be the first to synthesis available data on birth asphyxia. Findings of the review will fill an evidence gap in understanding the burden, risk factors, and causes of birth asphyxia in the country. The result from this review will inform health policy planners and researchers up-to-date data on birth asphyxia and provide direction in which risk factors the policy should focus to reduce burden of birth asphyxia in the country.
4.12 Dissemination of results
The results of this systematic review and meta-analysis will be published in a peer-reviewed journal and presented at national and international research conferences.