Prevalence of Occupational Injury among Workers in the Construction, Manufacturing, and Mining Industries in Africa: A Systematic Review and Meta-analysis
Mitiku Bonsa Debela 1*, Muluken Azage 2, Achenef Motbainor Begosaw 2
1. Master of Public Health in Epidemiology, School of Public Health, College of Medicine and Health Sciences, Bahir Dar University, Ethiopia.
2. PhD, Schools of Public Health, College of Medicine and Health Sciences, Bahir Dar University, Ethiopia
* Corresponding author: Mitiku Bonsa Debela; E-mail: mitikubonsa8@gmail.com
Abstract
Background: Despite the volume of existing literature on the prevalence of occupation-related injuries in Africa, it is fragmented across a broader spectrum and difficult to quickly understand the average magnitude. Besides, there is a lack of empirical shreds of evidence on the regional pooled estimate. Hence, the study aimed to develop the regional pooled estimates of occupation-related injuries among workers in Africa's industries.
Materials and Methods: The study followed preferred reporting items for systematic reviews and meta-analyses guidelines. The used databases included Scopus, PubMed, Science Direct, and Cochrane Library. Further, a modified version of the Newcastle-Ottawa Quality Assessment was used for the critical appraisal of studies. The pooled prevalence of injury was computed using STATA version 14 statistical software. Funnel plot and Egger's tests were conducted to evaluate publication bias. The study assessed the heterogeneity using the I-squared test and Galbraith plot.
Results: Out of 603 accessed studies, 20 that met the eligibility criteria were included. The pooled prevalence of occupational injury in Africa was 57% (95% CI: 48, 67). Totally, 62% (95% CI: 44, 77), 57% (95% CI: 38, 76), and 51% (95% CI: 32, 69) of injuries were identified in the manufacturing, construction, and mining sites, respectively, based on the subgroup analysis.
Conclusions: The rate of occupation-related injuries is dramatically increasing. Such injury is one of the immense concerns for workers' health and safety in Africa. Hence, the stakeholders should carry out rigorous law enforcement to ensure compliance with health and safety measures.
Keywords: Construction, Occupational Injury, Prevalence, Manufacturing, Mining, Systematic Review, Africa.
Introduction
Any personal injury, illness, or death resulting from an occupational accident that constitutes a significant global burden is occupational injury [1, 2]. Such injuries at work pose a significant public health issue and lead to severe social and economic implications. Workplace injury represents a large portion of the global injury burden, accounting for almost 30% of all medically treated injuries to adults aged 18 to 64 years [3].
Globally, occupational injury contributed to over 2.78 million deaths and 374 million non-fatal injuries in 2017 [4, 5]. Besides, it is estimated that the number of non-fatal workplace injuries is considerably rising (started from 2010) [6]. World Health Organization (WHO) recently has reported that 20%-50% of industrial workers worldwide are exposed to various occupational hazards. This figure is likely to be higher in developing counties [7]. Moreover, in 2018, the finding in the U.S showed that 5,250 workers died from work-related accidents. In 2019, non-fatal workplace accidents among the private sector workers were 2.8 cases per 100 full-time equivalent employees with eight median days away from work [8].
In developing countries, where manufacturing is mainly concentrated, health and safety regulation is not well enforced; thus, the burden of workplace injuries is incredibly high [9, 10]. African countries are among developing countries where occupational injuries are more serious [11, 12]. Evidence from the Ghana gold mining industry alone revealed that the occupational injury burden from 2015 to 2016 was 26% [13]. Evidence from Zambia's largest copper mining industry showed that over 165 workers were sustained occupational injuries [14]. In Congo, the prevalence of occupational injury was 24% [15]. Another finding from Tanzania indicated that the prevalence of occupational injury was 58.5% [16].
Moreover, studies from Ethiopia reported inconsistent magnitude of occupational injuries. For example, one study showed that about 1,356 working days were lost, and 35% of workers were absent from work for 15 to 30 days due to injuries [17]. Another study from the southwest region of Ethiopia showed the prevalence of occupational injury to be 45.2% [18]. Similarly, evidence from the northern part of Ethiopia indicated the prevalence of occupational injury to be 58.2% [19]. Besides, findings in Addis Ababa showed the prevalence of occupational injury among workers in manufacturing industries to be 91% [20].
In conclusion, the occupation-related morbidity and mortality rate is becoming a substantial public health concern for many industrial workers worldwide, as reported by the International Labor Organization. Despite the volume of existing literature on the prevalence of occupation-related injuries in Africa, it is fragmented across a broader spectrum and difficult to quickly understand the average magnitude. For instance, findings have shown a range from 9.7% to 97.5% in Ethiopia [21, 22], 19.2% to 98.1% in Kenya [23], 39.25 to 69% in Nigeria [24, 25], 23.7% to 72.2% in Congo [15, 26], and 35.6% to 86.3% in Rwanda [27]. Besides, there is a lack of empirical shreds of evidence on the regional pooled estimate of occupation-related injuries among workers in Africa's construction, manufacturing, and mining industries. Therefore, understanding the pooled prevalence of occupational injury is paramount to design health and safety strategies to reduce such injuries and associated economic costs.
According to the present study, although much has been written on work-related injuries among workers in the construction industry, limited studies have attempted to pool the extent of such injuries in other occupational groups (mining and manufacturing sugar industries). The present review incorporates heterogeneous occupational groups, thus appreciably advancing fundamental understanding or knowledge in the area. In addition, data from prominent sources (occupational groups) were combined, providing an opportunity to develop novel theoretical perspectives.
Hence, the study aimed to develop the regional pooled estimates of occupational-related injury among workers in Africa's construction, manufacturing, and mining industries and provide the necessary information for the scientific communities and policymakers who intervene in the problem.
Materials and Methods
This systematic review and meta-analysis protocol estimated the pooled prevalence of occupational injury among workers in Africa's construction, manufacturing, and mining industries. The whole review project was conducted from 10/1/2020 to 10/2/2021. According to the latest United Nations estimates (2021), the population of Africa is 1,361,684,609. The population density and the total land area in Africa are 45 persons per km2 and 29, 648, 48 km2, respectively. Nigeria is the most populous African country, with over 206 million inhabitants as of 2020 [28, 29].
The databases of Cochrane Library, Joanna Briggs Institute (JBI), and PROSPERO were checked for ongoing review projects related to the prevalence of occupational injuries in Africa. Preferred reporting items for systematic review and meta-analysis (PRISMA-P 2015) guidelines were followed to show accessed, screened, rejected, and included articles systematically or as per predesigned searching strategies.
The Cochrane Library, Joanna Briggs, SCOPUS, PubMed, Science Direct, Cochrane Library, and African journals online databases were systematically searched from January 1/2010 to February 10/2021 using the following keywords: (prevalence) OR (epidemiology) OR (burden) OR (magnitude) OR (distributions) AND (workplace injuries) OR (occupational accident) OR (occupational illness) OR (work-related injury) OR (work-related accidents) AND Africa. The key terms were combined using Boolean operators like "OR" or "AND." The review was restricted to full texts, free articles, and English language publications. It was used in all fields and Mesh words during the advanced PubMed search. The first reviewer performed the initial search and completed it on 10/02/2021. Then, the literature was scanned for updates.
The review considered all primary studies conducted in Africa on the prevalence of occupational injury. It incorporated an observational study design published from a time frame of 1/1/2010 up to 10/2/ 2021. In addition, the study included both published and unpublished studies in English. It also included studies with the outcome of interest (occupational injury).
This review omitted primary studies not entirely accessed during the searching process, as well as those without a report on the outcome of interest and with methodological problems. Besides, studies with low quality as the pre-settled parameters were also removed. The full-text review was limited to studies that met the requirements for inclusion.
Duplicates were removed, and database search results merged using Endnote (version X8). A modified version of the Newcastle-Ottawa quality assessment scale was also adapted to assess the methodological qualities of the included studies [30]. The three authors (MB, MA, and AM) independently evaluated the quality of included research articles. Disagreements were solved by discussion among the reviewers. Primary studies that scored ≥7 out of 10 were considered as high quality, thus being included in the final meta-analysis.
Data were extracted using a data extraction tool adapted from a meta-analysis of statistics assessment and review instruments. The corresponding author of the original research was contacted for incomplete or ambiguous information or to clarify method details as needed. First, the title and abstract of the paper were critically reviewed; duplicated data and articles whose titles were irrelevant to the study were excluded. Then, full documents of the papers were read and reread. The abstracts were extracted. Data were defined and extracted by MB and double-checked by a second reviewer. Also, two authors (MB and MA) independently extracted all the required data. The outcome of interest (prevalence) data extraction format consisted of the first author's name, publication year, study location, analysis design, sample size, number of participants with the outcome (case), a sub-region of the study, site
of injury, scale or scope of the industry, and response rate.
The PROSPERO registration number (CRD42021230787) was obtained. The study reported the prevalence of occupational injury as a percentage or as the number of cases (n)/total number of participants in the sample (N). Therefore, the prevalence rate was determined by dividing the number of individuals injured by the total number of participants in the study (sample size) multiplied by 100.
The necessary data were extracted from each article using the Microsoft Excel spreadsheet to estimate the pooled prevalence. Then, the analysis was conducted using STATA 14 version software. First, the pooled prevalence of occupational injury was computed with Metaprop on Stata command and presented in a forest plot with a corresponding 95% confidence interval.
Next, publication bias was checked by funnel plot (subjectively), as well as Begg's and Egger's tests (objectively). The study used a cutoff point of a p-value of less than 0.05 for Begg's and Egger's tests to declare the existence of publication bias. Then, the Galbraith and forest plots were used to visualize the presence of heterogeneity. Further, Higgins I-Squared (I2) and Cochran's Q statistic were applied objectively to figure out the heterogeneity. Finally, the I-square statistics used a cutoff point of 50% to declare significant heterogeneity.
A random-effect model that accounts within and between-study variability was used to estimate the pooled effect size. The results were presented via forest plot with the corresponding odds ratio and 95% confidence intervals. Also, the prevalence rate, the prevalence logarithm, and the standard error (SE) of the prevalence logarithm were computed. An output in meta-analyses was double-checked for internal consistency by the same person.
Results
A total of 603 papers were accessed from PubMed databases, SCOPUS, Science direct, and other sources. Among them, 326 studies were rejected due to duplication. After reading the title and the abstract, 48 studies were omitted since they did not align with this review's purpose and had a methodological deficit. Sixty studies were screened for full-text review, 20 of which were included for the systematic review and meta-analysis. The detailed steps of the screening process are shown in the study selection PRISMA flow map (Fig. 1).