The Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guideline was used to report this review8 (Supplementary file1). The five major databases (PubMed/MEDLINE, CINAHL, EMBASE, Google Scholar, and Science Direct) were searched for relevant studies. Moreover, articles were searched from the reference list of eligible studies. The search was carried out independently by three authors (MYB, GMB, and SSJ). Endnote X9 was used to retrieve and manage studies found through a systematic search. The following terms were used in the above-mentioned databases searches: (((((children[MeSH Terms]) OR (preschool[MeSH Terms])) OR (children[Text Word])) OR (preschool[Text Word]) AND ((y_1[Filter]) AND (fha[Filter]))) AND ((Ethiopia[MeSH Terms]) OR (Ethiopia[Text Word]) AND ((y_1[Filter]) AND (fha[Filter])))) AND (((myopia[Text Word]) OR (refractive error[Text Word])) OR (Myopia[MeSH Terms]) AND ((y_1[Filter]) AND (fha[Filter]))) Filters: Abstract, in the last 1 year.” The searching strategy was lasting from June 28 upto 25 July 2022.

All school based cross-sectional studies were included. All titles/abstracts and full texts found in electronic databases were independently screened by two authors (MYB and GMB). Disagreements were settled through discussion in the presence of third author (CTL) (Fig. 1).

Fig. 1.

PRISMA flow diagram of myopia among Ethiopian schoolchildren.

(0.31MB).

The data were extracted using a Microsoft Excel spreadsheet. To ensure consistency, two authors (MYB and SSJ) extracted data independently using a prepared Microsoft Excel spreadsheet format. During data extraction, the disagreements between data extractor authors were resolved through discussion in the presence of third author (GMB). The data extraction format contains the number of schoolchildren with myopia, the prevalence of myopia in schoolchildren, region, study period, year of publication, sample size, and first author of the primary article to estimate the pooled prevalence of myopia in Ethiopian schoolchildren. As well as, the data extraction checklist contains for factors associated (OR, CI, exposed with disease, nonexposed with nondiseased, exposed with nondiseased, and nonexposed with diseased) with myopia too.

The Newcastle-Ottawa Quality Assessment Scale (NOQAS)9 was used to evaluate the quality of the included primary studies based on study representativeness, adequate sample size, acceptable non-response rate, use of validated measurement tool, comparability of the study, description of outcome assessment, and use of appropriate statistical tests as parameters. High-quality articles received a global rating of seven out of ten10,11 (Supplementary file 2).

In this systematic review and meta-analysis, national estimate of myopia was the main outcome of interest. The second outcome of this systematic review and meta-analysis was to identify factors associated with myopia among schoolchildren in Ethiopia.

The extracted data were exported to Stata™ Version 17.0 software for further management and analysis. The metaprop stata command was used to compute the overall pooled estimate. Using a binominal distribution assumption, the standard errors were calculated from the reported estimates and population denominators. The DerSimonian and Laird Method12 was used to create a random effects meta-analysis model. To identify the possible sources of heterogeneity, additional statistical analyses such as subgroup analyses, publication bias, and meta-regression were performed. We conducted a subgroup analysis using geographical regions, sample size (above mean Vs. below the mean (807)), and study period (before 2015 Vs. after 2015). Furthermore, sensitivity analysis was performed using a random effects model to assess the impact of a single study on the overall pooled estimate. Finally, the results were presented in tables and forest plots.

A total of 589 studies were found through electronic searching of database such as PubMed, EMBASE, Google Scholar, and Science Direct and others like organizational records and website searches from. About 942 articles were excluded due to duplication, 493 articles were excluded due to different in study context,13-17 10 articles were excluded due to different in interest of outcome18-24 and 5 articles were excluded due to different in study population.25,26 Finally, 12 cross-sectional studies were identified as having the potential to be included in the current systematic review and meta-analysis.

According to Table 1, about 12 studies with a total of 9688 primary and secondary schoolchildren met the inclusion criteria. This systematic review and meta-analysis included two regions and one city administration. Among them were Amhara region including six studies27-32 and South nation and nationality of people (n = 4),21,33-35 and Addis Abeba (n = 2).36,37 The study with the smallest and largest sample sizes, 349 and 4238, was conducted in the SNNPR's Hawassa and Sodo, respectively (Table 1).

In Ethiopia, the national estimat of myopia among schoolchildren was 6.49 percent (95%CI: 4.86, 8.12) (Fig. 2). When we looked at it by region, Amhara and South nation and nationality of people had largest (7.48 for both of them) and Addis Abeba city administration had smallest burden (2.21%(95%CI:1.65, 2.77)) of myopia (Fig. 3).

Fig. 2.

The overall pooled prevalence of myopia among schoolchildren in Ethiopia.

(0.35MB).

Fig. 3.

The regional pooled prevalence of myopia in Ethiopia.

(0.41MB).

After determining the presence of heterogeneity between studies, we attempted to address the heterogeneity by conducting subgroup meta-analysis while taking into account region, sample size, and study period. As a result of the subgroup meta-analysis, it was discovered that publication bias study period, and sample size were all identified as sources of heterogeneity with the presence of strong evidence supporting the existence of heterogeneity that means region (Amhara:I2=92.2 with p-value 0.001, SNNP: I2=94.7 with p-value 0.001, and Addis Ababa: I2 = 0.0 with p-value 0.53) (Fig. 4), studies having sample size > 807: I2 = 96.00 with p-value 0.0001 and those studies having sample size < 807: I2= 93.4 with p-value 0.0001 (Fig. 5), study period(after 2015: I2= 93.4 with p-value 0.0001 and before 2015: I2= 93.4 with p-value 0.0001) (Fig. 6).

Fig. 4.

Subgroup analysis using sample size, 807 as cut off point.

(0.35MB).

Fig. 5.

Subgroup meta-analysis using study period after and before 2015.

(0.36MB).

Fig. 6.

s: Funnel plot to check publication bias.

(0.08MB).

The publication year and sample size were used as covariates in random-effects meta-regression. The analysis revealed that sample size (p = 0.72) and publication year (p = 0.99) had no effect on heterogeneity (heterogeneity was not explained by sample size and publication year) (Table 2).

A funnel plot was used to determine the presence or absence of publication bias, and the scatter plot points were asymmetrical (attributed to publication bias), indicating that the small-study had effects on the heterogeneity of myopia estimate among Ethiopian schoolchildren (Fig. 7). The egger linear regression test was used to objectively check the presence or absence of publication bias. As a result, there was statistically significant publication bias (P = 0.016) (Table 3).

Fig. 7.

Association between myopia and family history of myopia in Ethiopia.

(0.19MB).

Factors associated with myopia in schoolchildren was determined based on sex (including seven studies),27-29,31-33,37 family history of myopia (including four studies),28,29,31,33 residence(including eight studies),21,28-31,34,36,37 family educational status (including four studies)[29–31, 34], and use of computer regularly (three studies).28,35,37 Finally, being female and had family history of myopia were the factors which are significantly associated with myopia.

Being rural residence was 1.09 (95% CI: 0.56, 2.13), had no regular computer exposure had a 0.69 (95% CI: 0.45, 1.06), and having educated family had a 1.89 (95% CI: 0.59, 6.06) times larger chance of myopia than its corresponding counterpart. However, these differences were not statistically significant.

The odds of myopia among female schoolchildren was 0.94 (95% CI: 0.50, 0.980) times lower than male schoolchildren (Fig. 6). Schoolchildren who had family history of myopia was 6.48 (95% CI: 3.03, 13.82) times greater than students who had no family history of myopia (Fig. 8).

Fig. 8.

Association between myopia and family history of myopia in Ethiopia.

(0.19MB).

This study's national estimate of myopia may not exactly represent the burden of Ethiopian schoolchildren because it was conducted using primary studies only from Amhara, South nation nationalities and people, and Addis Ababa. In addition to the foregoing, this study only included articles published in English, which may affect the prevalence of myopia and taken as a limitation of this study.