The prevalence and causes of pediatric uncorrected refractive error: Pooled data from population studies for Global Burden of Disease (GBD) sub-regions

Authors: Cao H, Cao X, Cao Z, Zhang L, Han Y, Guo C.

Geographical coverage: Global

Sector: Burden of disease

Sub-sector: Prevalence and causes

Equity focus: None specified

Study population: General population aged less than 20 years old.

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background:

The World Health Organization’s (WHO) Global Action Plan for 2014 to 2019 has identified human resources for refractive error as a priority to reduce avoidable blindness globally. Uncorrected refractive error (URE) is the most common cause of visual impairment (VI) in children – the WHO estimates 19 million children and adolescents aged 5 to 15 years old suffer from VI among which, approximately 12.8 million cases (67%) are due to URE. Uncorrected refractive error may reduce educational opportunities, productivity and overall quality of life. Identification of the prevalence and causes of visual impairment are crucial for the establishment of local programmes and supra-national, continental and world prevention strategies.

Objectives: This review aimed to summarise the prevalence and causes of paediatric uncorrected refractive error (URE) from studies in the Global Burden of Disease (GBD) sub-regions.

Main findings:

Authors included a total of 58 studies in the review. These were population-based studies from different regions, including Africa, Asia, Americas, Eastern Mediterranean, Europe, Southeast Asia and Western-Pacific. Twelve studies were included from WPRO, including two from Australia, and seven from China, one each from Vietnam and two studies from Cambodia. Ten studies were included from the Americas region analysis, four from the United States of America, and two each from Brazil, Paraguay and Mexico. Estimates of prevalence of URE in the EMRO were reported from 10 studies from Iran, two each from Tehran and Tunisia, and one each from Saudi Arabia and Egypt. Fourteen studies were grouped into the SEARO region. Authors note not being able to calculate the prevalence of URE from Europe region and Africa region due to the lack of data.

The overall quality assessment of 23 studies were graded a score of 4, 19 studies were graded a score of 3 and 13 studies scored 5. The quality assessment of three studies were not reported.

Authors found the overall prevalence of URE in the WPRO region of 3.41 per 1,000 (95% CI: 1.53~7.62), however, high heterogeneity was identified among the studies in this analysis (I² = 99.6%). In the Americas region, the prevalence of URE was estimated at 5.85, however, authors note the imprecision of included studies. Prevalence of URE in the EMRO region reported was 4.4 (95% CI 3.0*6.45), again, authors note high heterogeneity among these studies (I² = 99.6%). In the SEARO region, prevalence was reported at 2.26 (95% CI: 0.85~6.01) with high heterogeneity (percentage not reported).

Based on the pooled prevalence data, authors report for children aged younger than 20 years old in WPRO A region, the overall estimated prevalence of URE was 21.1% in females and 6.6% in males, respectively. In WPRO B region, authors report higher prevalence of URE among Chinese male children than female children. In contrast, authors note that in the Americas region, Brazil and Mexico female children represented larger prevalence of URE (15.2 per 1,000), followed by American and Canadian male children (5.3 per 1,000). Most URE cases remained higher prevalence among African female children but to a lower ratio among male children (15.2 vs 3.8).

The Southeast Asian children with lower URE cases was projected to shift from 1.4 to 2.1 per 1,000 in male children. African female children were projected from 0.75 to 1.96 per 1,000.

When authors compared across the different regions between the year of 2000 to 2021, an increase of prevalence of URE was reported. In the Americas region, the prevalence of in children increased from approximately 10.5% to 31.9% between 2016 and 2018. In the WPRO region, authors report a steady increase between 2008 and 2012 from 3.5% to 6.5%. Authors also note a transient increasing shift in 2010 (percentage not reported).

This review found the prevalence of myopia to be higher in male children aged between 18 and 20 with no significant difference between the different age groups. The prevalence of astigmatism and hyperopia was reported to be highest in children aged between 6 and 11 years old. Authors found significant difference on the prevalence of URE among female children.

Authors note the development of refractive services should focus on the provision of affordable spectacles to address URE.

Methodology:

The review included studies that were population-based, representative of the sampled area and country, with an adequate sample size (ranging from 1,100 to 1.4 million). The studies needed to have a response rate of 80% or higher, report prevalence with a 95% confidence interval (or allow its calculation from raw data) and provide data for individuals with definitions of uncorrected refractive error (URE) and visual impairment consistent with this study. They also had to report presenting visual acuity (PVA) less than 6/18 with its causes and provide standard World Health Organization categories of visual acuity. For children, refractive diagnostics were determined by objective refraction under cycloplegia plus subjective refraction. Only studies written in English were included.

Authors conducted a search on Medline, EMBASE, Web of Science, Cochrane Library, and abstracts from the Association for Research in Vision and Ophthalmology. The bibliographies of the full-text articles that were reviewed were searched for relevant references. In addition, the reference lists from all identified studies were examined.

Two reviewers independently extracted the following data of studies included in the review and assessed the quality of included studies using de Weerd et al and Sophie Rogers et al criteria. A score of 3 or higher was considered adequate quality.

The authors conducted a meta-analysis on the prevalence of Uncorrected Refractive Error (URE), categorizing data by various global regions. They estimated URE prevalence rates using age-gender-standardisation for all age groups under 20, divided into four age categories for both genders. The 95% confidence intervals for these rates were calculated using a normal approximation and Breslow-Day standard errors, modified for a binomial assumption for the variance of the crude stratum-specific rates. Crude prevalence rates per 1,000 and Agresti-Coull modified Wald CIs were also calculated. The initial analysis included data from 15 studies. Publication bias was assessed using the Egger regression asymmetry test and the Begg’s test.

Applicability/external validity:

Authors do not generalise findings of this review across the sub-regions analysed due to the high heterogeneity among studies.

Geographic focus:

Authors include studies from different regions including high medium and low-income settings. The lowest and highest prevalence of URE was seen in Southeast Asian and American children, respectively.

Summary of quality assessment:

This review uses clear inclusion criteria, a rigorous statistical data analysis stratified by the different geographic regions, age groups and gender. Authors assessed for publication bias and heterogeneity among included studies. Authors appropriately categorised the geographic location of included studies according to the WHO regions. Review authors employed appropriate methods to extract data, assess the quality and pool data of included studies. Authors acknowledged high heterogeneity among studies included in the statistical analysis, and that “low confidence” should be given to findings of the pooled analysis. Furthermore, authors note presence of bias in the review as studies written in English only were included, and they assess how limitations of studies had an impact on the findings of the review. However, some important limitations, not raised by the authors in the review, were identified. Although authors assessed the quality of included studies, statistical analysis was not conducted by study quality. Search strategy was not comprehensive enough to ensure that all eligible studies were included in the review, and it is not clear if authors avoided bias whilst screening studies for included in the review. Therefore, there is medium confidence in the conclusions about the effects of this study.