Systematic review registration and reporting
The present review protocol has been registered within the PROSPERO database (registration number CRD42020164240) and is being reported in accordance with the reporting guidance provided in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) statement [23,24] (see checklist in Additional file 1) and Meta-analysis Of Observational Studies in Epidemiology (MOOSE) reporting guideline [25].
Data sources and search strategies
The primary source of literature will be a structured search of electronic databases (from January 2000 onwards, as we are primarily interested in the contemporary literature): PubMed/MEDLINE, Embase, SCOPUS, PsycINFO, PEDro and Ebsco. The secondary source of potentially relevant material will be a search of the grey or difficult to locate literature, including Google Scholar, dissertation databases and other relevant databases (e.g. World Health Organization, Centers for Disease Control and Prevention, Work Place Health Promotion). We will perform hand-searching of the reference lists of included studies, relevant reviews or other relevant documents. The search will include a broad range of terms and keywords related to musculoskeletal disorders, workplace and occupation, prevalence and the geographical area ‘Ethiopia’. A draft search strategy for PubMed/MEDLINE is provided in Additional file 2. Only studies published in English (or with an English version of the document) will be eligible. Moreover, the corresponding authors will be contacted by mail whenever a need arise or for any difficulties faced during data extraction.
Eligibility criteria
Studies will be selected based on the following criteria: participants, condition or outcome(s) of interest, study design, and context.
Participants (population): We will include studies involving adult population (≥ 18 years of age, regardless of sex) working in governmental sectors or private sectors or self-employed in Ethiopia.
Condition or outcome(s) of interest: The primary outcome will be the prevalence (e.g. point prevalence, period prevalence, prevalence rate) of WRMSDs (including neck disorders, low back disorders or other chronic disorders) indicating the number of people that have the disorder divided by the population number at a given point in time, presented often as a (prevalence) proportion. We will use authors’ reported definitions using standardized scales or questionnaires (e.g. Standardized Nordic questionnaire, Dutch Musculoskeletal questionnaire, or Visual Analogue Scale) providing sufficient information to calculate the prevalence. Secondary outcomes will be to identify the risk factors of musculoskeletal disorders (e.g. psychosocial factors, posture, heavy physical work, work place-related factors, and smoking). When outcome results (e.g., proportion and 95% CI) are not directly reported and it is feasible, we will calculate estimates from the number of cases and sample size mentioned in each single study.
Study design and context: Eligible studies will be observational studies (cohort, cross-sectional or health surveys) reporting prevalence data using validated or non-validated tools and conducted in the Ethiopian working population. Cross-sectional studies will be the most appropriate study design to determine the pooled prevalence of the work-related musculoskeletal disorders. For cohort studies, only the first phase (cross-sectional) data will be considered. We will exclude studies published in non-English, reviews, commentaries, conference abstracts, letter to editors, non-human articles, and studies conducted outside of Ethiopia.
Study screening and selection for inclusion in the review
The titles and abstracts of articles retrieved from the search of different databases will be stored and managed in a Zotero version 5 reference manager and we will remove the duplicate records from it. Two reviewers (BJ and KN) will independently review the titles and abstract part of all the articles, the full text of the potentially eligible article will be assessed for compliance and analyzed for inclusion. Disagreements or conflicts throughout the review process will be resolved by consensus and if needed, by requesting the opinion of the third and fourth reviewers. The consistency of the selection process and quality assessment across the reviewers will be ensured by calculating the level of inter-rater agreement (Kappa statistics) [26]. A PRISMA flowchart showing details of the studies included and excluded at each stage of the study selection process will be provided.
Data extraction and management
Once eligible studies are identified, two independent reviewers will extract the data using a prepared standardized data extraction form. Reviewers will not perform data extraction from an article of which they were (co-) authors to eliminate possible bias. Data such as first author’s last name, year of publication, study location within Ethiopia, sample size, response rate, the reason for non-response, number of events, regions of pain reported, data on prevalence, recall period, ascertainment of outcome measures, risk factor or protective factors determined by each study along with their respective odds ratio (OR) and 95% confidence interval, and information needed for the risk of bias assessment will be extracted Table 1. The findings of the review will be illustrated through figures and tables.
Table 1 Data items that will be extracted
1.
|
Publication details: Title, Journal, Author, Year, City/region in Ethiopia, where the study was conducted, Type of publication, and Source of funding, if any.
|
2.
|
Design: Type of study design (observational studies, cross-sectional, cohort, case-control, others); aims or objective of the study, method of data collection, outcome variable (measures and operational definition), response rate, recruitment and sampling method, eligibility (inclusion and exclusion criteria)
|
3.
|
Study participants information’s: Sample size, population characteristics including settings, role in the institution, age, gender, ethnicity, education level, demographic information’s, behavioral information’s, physical measurements, comorbidities, other health-related characteristics, Employment details, work and work-environment related characteristics, bodily distribution of pain
|
4.
|
Data for outcome variables: all reported estimates, or sufficient data to calculate an estimate of the point prevalence, cumulative incidence, and incidence rate of musculoskeletal disorders in categories of bodily distribution like low back pain, neck pain, shoulder pain, elbow pain, wrist and hand pain, hip pain, knee pain, ankle pain, etc
|
5.
|
Limitations: Bias (selection, response, information), limitation of outcome measures/tools, and limitation reported by authors
|
Risk of bias and quality assessment
Two review authors (BJ and KN) will independently assess the quality of all included studies using the Newcastle-Ottawa Quality Assessment tool adapted for cross-sectional studies [27,28]. The tool will be further adapted for use in this review (Additional file 3). Discrepancies of aggregate or total scores will be resolved by the third reviewer (TG) after a detailed evaluation of the source of the discrepancy. The tool contains three domains; selection of participants (3 items), quality of data (4 items), and definition of work-related musculoskeletal disorders (3 items). For the purpose of this review, all the items in the appraisal tool will be equally weighted and so the total score will be 10. There will be no subminimum score criteria for inclusion of studies.
Data synthesis
The prevalence rate, the logarithm of prevalence, and standard error (SE) of the logarithm of prevalence will be computed. Correspondingly, for the factors associated, the logarithm of OR and SE of the logarithms of OR will be calculated. The pooled prevalence (proportion) of musculoskeletal disorders or musculoskeletal pain-related work and the pooled odds ratios (OR) of associated factors with a 95% confidence interval will be calculated using random-effects and quality effects model. The quality-effects meta-analysis [29] will be used to examine how the quality of each study influenced the pooled estimate compared with the results from the random effects [30]. The quality scores of each included study will be incorporated in the calculation of study weight to improve the robustness and help minimize the estimator variance and subjectivity in quality assessment. The presence of heterogeneity among studies will be quantified by estimating variance using both Cochrane’s Q statistics and the I2 statistics. I2statistics is the proportion of the variation in the estimates of prevalence due to genuine variation in prevalence rather than sampling error [31, 32]. Funnel plots will be used to assess small study effects or publication bias [33] by inspection of asymmetry and in addition Egger’s regression test (p<0.05) [34] and Beggs equations will be computed to declare publication bias. Double arsine transformation will be used in the case of variance instability [35].
Possible subgroups will be identified based on the study characteristics and population characteristics. Subgroup analysis will be performed to determine the source of heterogeneity attributed to gender, sample size, place of study, study setting, study design, year of publication, outcome tool used, type of occupation, and region of pain. Sensitivity analysis will be performed after excluding each study one by one and the pooled estimate will be calculated for the remaining studies. All statistical analyses will be performed using Meta XL version 5.3 [36] and STATA 15 Metaprop package [37]. Quantitative data from different designs, variability in prevalence (e.g. cohort, case-control studies) will not be combined, nor unadjusted and adjusted models. Transformation of rates to their natural logarithm, with corresponding standard errors (SEs) will be entered into meta-analyses. When quantitative synthesis is not feasible, we will retain the studies for narrative synthesis.
Timeline, presenting and reporting of the results
The review process will commence only after the final peer-review comments are received, and the protocol accepted for publication. The study selection step by step process will be outlined in a flow diagram and the reasons for exclusion will also be mentioned. The study characteristics, risk of bias, and quality assessment of the included studies will be presented in tables. Forest plots will be used to display the pooled estimates of prevalence proportions.