Study Protocol Registration and Reporting
The Preferred Reporting Items for Systematic review and Meta-analyses (PRISMA) was used to develop the review protocol [23], and PRISMA-P 2015 checklist [24] statements will be used to report the findings (Additional file 1). The protocol for this review had registered at PROSPERO with registration number: CRD42020165628
PECO search guide
Population: Adolescent girl (10–19 years) [25]
Exposure: Predictors or associated factors of menstrual hygiene practice. Factors are characteristics or exposures that increase or decrease the likelihood of menstrual hygiene practice among adolescent girls in Sub-Saharan Africa. Those factors include residence, age, maternal educational level, family income, menstrual flow duration, and knowledge on menses.
Comparison: This is the reported reference group for each associated factor in each study: adolescent girls residing in urban versus rural, from educated mothers versus no education, with good knowledge on menses versus poor knowledge.
Outcome: The primary outcome of the study will be the pooled prevalence of good menstrual hygiene practice. The second outcome of the study is associated factors of menstrual hygiene practice among adolescent girls in Sub-Saharan Africa. Good menstrual hygiene practice indicate when the studies report the overall good menstrual hygiene practice for the different measurement of menstrual hygiene practices (type of menstrual items used, maintenance of item if reusable or disposal if one-time use only, changing frequency, using a clean menstrual management material to absorb or collect, washing the body as needed with soap and water, access to facilities to dispose of used menstrual materials etc.).
Searching strategy
For the preparation and presentation, preferred Reporting Items for Systematic Reviews and Meta-Analysis guideline protocol will be used. Online electronic databases PubMed, Google Scholar, CINAHL and unpublished grey literature will be used to search articles from January 1-May/2021. The process of searching and selection of studies will be reported in accordance with the PRISMA diagram (Additional file 2). Cross-reference search will be done to add other related studies from the final included studies. Retrieve studies will be exported to Endnote version 8 reference manager software [26].
Medical Subject Heading (Mesh) text search term will be developed from articles PMID, author’s keywords, titles and abstract. The search string will be developed using different Boolean operators, and modified, depending on the specific requirement of the database and relevant studies (Additional file 3).
Study selection and eligibility criteria
First, duplicate studies will be removed from the Endnote citation manager. Two authors (EW and SB) will independently screen the studies based on inclusion and exclusion criteria. Irrelevant titles and articles will be excluded. However, relevant titles and articles with the full-text will be further screened. In case of articles which are not open access, we will contact the corresponding author. If the authors are not willing to provide full text, the articles will be excluded from the study. Studies that will be approved by both authors in the review processes will be included. During the review of the studies, any disagreement among reviewers will be resolved by discussion to reach a common understanding.
This review will include all observational studies: cross-sectional, analytical cross-sectional, case-control, and cohort studies. Articles published only with the English language, studies that report the overall good menstrual hygiene practice and its associated factors among adolescent girls will be included. However, study that only report overall good menstrual hygiene practice will also be included. If a studies analyzed knowledge, attitude and practice, we will only include the practice results. Institution and community based studies will be included. We will only consider the quantitative results for study that examine both quantitative and qualitative studies. Data that are difficult to extract will be excluded. Moreover, adolescent girls whose age group is not well defined, case reports, conference reports, national survey reports, and expert opinions will be excluded. No restriction will be made to the date of publication.
Quality assessments
Articles assessment using their title, abstract, and a full review of the manuscripts will be done before the inclusion of articles in the final meta-analysis. A critical appraisal will be performed by two authors (EW and SB) using the Joanna Briggs Institute Meta-Analysis of Statistics Assessment and Review Instrument (JBI-MASt-ARI) [27] (Additional file 4). Studies 50% and above of the quality scale will be included and considered for systematic review and meta-analysis. For any scoring disagreements between the authors, the sources of discrepancy will be investigated through revision.
Data extraction and management
After identifying all eligible articles two independent reviewers (EW and SB) will extract the relevant data using an organized format on Microsoft Excel Spreadsheet 2016. Pretesting the data extraction form will be done before the beginning of the actual data extraction. If variations of extracted data exist, the phase will be repeated, and then discrepancies between data extractors will be discussed to reach a consensus. For each included article, we will record the author name, year of publication, the study area, study design, study period, sample size, the response rate, the target population, outcome definition, comparison groups, predictors, and overall good menstrual hygiene practice. For the practice studies, the prevalence, the logarithm of the prevalence, and standard error (SE) of the logarithm of the prevalence will be calculated. Similarly, for associated factors, OR, logarithms of OR, and SE of the logarithms of OR will be calculated. For any difficulties that might be encountered during data extraction, communication will be made with the corresponding author.
Data synthesis and analysis
The extract data will be imported into STATA version 14 software. A narrative description of the study population will be done, and tables and figures will be used to summarize the results.
A random-effect model will be used to estimate the pooled prevalence of good menstrual hygiene practice among adolescent girls in Sub-Saharan Africa [28]. The Freeman Tuckey variant of the arcsine square root transformation of proportions will be used to avoid variance variability when controlling proportions close to one [29, 30]. We will assess heterogeneity by using the chi-squared test on Cochran’s Q statistic with a 5% level of statistical significance [31] and I2 statistic test [32]. Assuming I2 values of 25%, 50%, and 75% is representative of low, moderate, and high heterogeneity respectively. If the heterogeneity is significant (I2>75%), and p-value <0.05 will be declared as the presence of heterogeneity. Hence, subgroup analyses and meta-regression for continuous and categorical data based on region, and sample size will be done to investigate sources of heterogeneity at a p-value< 0.05.
Sensitivity analysis will be done to see the effect of studies on the overall estimation on forest plots, and an individual study is suspected of excessive influence if the point estimate of its omitted analysis lies outside the confidence interval of the combined analysis or differs in significance relative to the combined analysis in forest plot. Publication bias will be examined by the visual inspection of funnel plots [33] and Egger’s test [34]. A p-value <0.05 will be considered indicative of statistically significant publication bias. If evidence of publication bias present, the trim-and-fill (Duval and Tweedie’s) method will be performed [35]. The statistical significance level will be declared at a p-value of less than 0.05.