Prevalence of myopia in Indian school children: Meta-analysis of last four decades

Authors: Agarwal D, Saxena R, Gupta V, Mani K, Dhiman R, Bhardawaj A, Vashist P.

Geographical coverage: India

Sector: Burden of disease

Sub-sector: Epidemiology

Equity focus: School-aged children

Study population: Indian school children aged 5-15.

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background:

The increasing prevalence of myopia, particularly among school children and young adults, is a global concern. This common refractive error is becoming an epidemic in some regions. Despite India having a significant population under 18 years (41% or 492 million), the country lacks adequate data on myopia prevalence and trend studies over decades. This lack of data results in poor representation of India in global myopia trend predictions.

Objectives: Understand the prevalence of myopia, regional variations and prediction of trend, using all the published literature of the last four decades from India.

Main findings:

Authors included a total of 77 studies in the review and in the sensitive analysis. Eighteen studies were deemed as having high risk of bias; therefore, 59 studies were included in the meta-analysis. 37 studies included data from urban regions, 12 included data from rural regions and 10 studies showed both urban and rural data. Region-wise representation of studies is as follows: North India (12), Northeast India (4), Central India (6), East India (7), West India (8) and South India (22). All studies were cross-sectional in nature. Most of the studies were school-based, with only four being population-based. Eighteen studies were graded high risk of bias, 21 moderate risk of bias and 37 low risk of bias.

The overall pooled prevalence of myopia among the 5-15-years age group over the last four decades reported by the authors was 7.5% (95% CI, 6.5-8.5%). The overall pooled prevalence of myopia in urban children was 8.5% (95% CI, 7.1-9.9%) and in rural settings was 6.1% (95% CI, 4.5-7.7%) in past four decades. Authors note an increase of prevalence over the years from 4.6% to 6.8% between 1980-2019 in children from rural areas of India. The sensitivity analysis, which excluded high risk of bias studies, showed similar trends as the main analysis, with 7.4% (95% CI, 7-7.8%) overall pooled prevalence of myopia over last four decades in the 5-15-years age group.

The prevalence of myopia in children aged between 11-15 years over the last four decades was 10.7% (95% CI, 9-12.4). Authors note an increase of prevalence in rural children from 6.9% to 12.3% between 1980-2019.

Authors report high heterogeneity in the pooled analyses as well as between the urban-rural subgroups analysis.

Based on the findings of the review, authors conclude that there is a rapid rising trend of myopia in rural school children compared to their urban counterparts. Therefore, this finding should result in adoption of urgent preventative and curative measures among various stakeholders to tackle this menace on time. Authors also note that future prospective studies should be planned among various regions of India to elucidate the trend of myopia and analyse various local epidemiological risk factors involved.

Methodology:

Authors searched PubMed, Medline, EMBASE, OVID, Web of Science, CINAHL and Cochrane Library databases from 1 January 1980 to 31 March 2020.

Authors included cross-sectional studies, including population-based as well school-based studies in the review. These were included if estimated the prevalence and/or incidence of refractive errors and/or myopia among children aged between 5-15 years and covered urban and rural settings. Qualitative studies, review articles, articles published in languages other than English and articles which did not have relevant information available online were excluded. A data extraction form was developed by the authors and included various characteristics such as study design and location. Quality assessment was conducted by three reviewers independently using the critical appraisal checklist developed for prevalence studies by Hoy et al.

The data was combined and stratified by five year age groups: 5-10 and 11-15, where possible. Meta-analysis used a random-effects model using DerSimonian and Laird methods to calculate pooled effect sizes and its 95% CI. The variation in the magnitude of the effect was examined and heterogeneity was quantified using the I² statistic (although this does not appear to have been specifically reported). The funnel plot was used to detect potential reporting bias and small/large study effects and the Egger method was used to assess asymmetry. Studies which were categorised as ‘high risk’ based on assessment of methodological quality described above, were excluded from the final analysis. All studies (low, moderate and high risk) were included in a sensitivity analysis. Urban and rural data was analysed separately. The studies which represented both urban and rural population were later subdivided into separate datasets based on study settings (urban or rural) for detailed analysis. Rural-urban and time stratified estimates of prevalence of myopia across included studies were obtained.

Applicability/external validity:

Findings of this review may be generalisable to children aged between 5 and 11 years from rural and urban areas of India.

Geographic focus:

Authors focus on different regions of India only; they are interested in prevalence in India specifically and it is not intended that their findings should be extrapolated to other LMICs.

Summary of quality assessment:

This systematic review clearly described the criteria used for deciding which studies to include in the review, the search covered a sensible time period and used appropriate tools and methods to assess the risk of bias of included studies in the review. Furthermore, authors appropriately pooled data for statistical analyses and excluded high risk of bias studies in the sensitivity analysis. However, important limitations were identified which can impact in the validity of the findings. The searches did not cover a comprehensive set of databases and were conducted to identify published studies written in English only. Important literature, including unpublished studies, in other languages may have been missed. Moreover, the review does not state if there was independent screening of studies and data extraction by two or more researchers (although it is clear that studies were assessed by multiple individuals). Therefore, a low confidence was attributed to the findings of this review.