Methodological quality of the review: Medium confidence
Author: Carla Lanca and Seang-Mei Saw
Region: China, USA, India, Vietnam
Subsector: Myopia risk factor
Equity focus: No
Study population: Children
Type of programme: Community based
Review type: Other review
Quantitative synthesis method: Meta-analysis
Qualitative synthesis method: Not applicable
Background: Myopia is a major contributor to the global burden of eye disease due to high prevalence, especially in East Asia. The high prevalence of myopia in East Asian cities is due to several risk factors, including environmental factors, such as education, increased near work and less time spent outdoors. However, consistent evidence of an association between screen time and myopia development is lacking. Education and near work have played major roles in the rise in myopia rates even before the introduction of digital devices. A recent report from the World Health Organization (WHO) provides guidelines to restrict sedentary screen time for children under 5 years old, as some evidence suggests that screen time may increase sedentary behaviour with negative impact for children’s health. Norms to limit screen time to prevent myopia need to be supported with evidence-based studies.
Objectives: The objective of this review was to conduct a systematic review to summarise the evidence for the association between screen time and myopia or myopia progression.
Main findings: A total of 15 references met the criteria for inclusion in this meta-analysis, with a total of 49,789 children aged between 3 and 19 years old. The majority of the participants were from China (62%) and India (20%). Use of computers (either categorical variable or hours per week [h/week]) was significantly (all Ps<0.05) associated with the prevalence of myopia (SE≤-0.5 D, SE≤-3.00 D and SE≤-6.00 D) and more myopic refractive error in children 5-16 years old. A positive association between screen time and prevalent myopia was reported in some studies. In 7,681 children aged 5-16 years old, myopia (adjusted OR=1.17, 95% CI 1.03-1.32; P=0.015) and high myopia prevalence (adjusted OR=2.31, 95% CI 1.17-4.57; P=0.016) were associated with computer use. In older children, the association emerged, with children aged 5-15 years old showing a positive association between prevalent myopia and h/week playing computer, video or mobile games (OR=4.5, 95% CI 2.33, 8.98; P<0.001 for 1-4 h/week; OR=8.1, 95% CI 4.05, 16.2; P<0.001 for >4 h/week). Similar results were observed in children (n=2,238) aged 12-15 years old with h/week of using computers being associated with prevalent myopia (OR=1.02, 95% CI 1.00, 1.04; P=0.022). A pooled OR of 1.02 (95% CI: 0.96-1.08; P=0.48) suggests that screen time is not associated with prevalent and incident myopia in this group of five studies.
Overall, authors note that the results for screen time and myopia are mixed. Further studies with objective screen time measurements are necessary to assess evidence of an association between screen time and myopia.
Methodology: Inclusion criteria consisted of studies that: 1) reported screen time; 2) reported myopia (prevalence or incidence), myopia progression rate or axial length progression as the outcome measure; and 3) reported the association either with an effect estimate (95% confidence interval [CI]), standard error or raw data that allows for the calculation in the meta-analysis. Myopia considered in the article did not involve treatment or therapeutic intervention. Systematic strategies were used to search key electronic databases, including PubMed, ScienceDirect, the Cochrane Library, and Science Citation Index databases. The review was undertaken using Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. English language articles published up to July 2019 were retrieved using the keyword search terms, such as myopia, screen time, computers and video games. The following types of studies were included in the review: cohort, case-control, cross sectional or intervention trials. We also reviewed the references of all retrieved articles to identify articles not captured by the initial electronic search. Conference transactions/proceedings, which are not electronically listed, were hand searched. The full papers of any studies considered potentially relevant were considered, and the selection criteria applied independently by the two authors of this review. Pre-designed data extraction forms were used to gather information on authors, year of publication, country of study, study design, sample size and age, definition of myopia, effect estimate (95% CI) and/or standard error and adjustment of confounders. Data was extracted and documented by one of the authors (CL) and verified by the other (SMS). Data about quality appraisal and risk of bias was also extracted. Methodological quality of included studies was evaluated using the tool developed by Hoy and co-workers. A score of 1 (yes) or 0 (no) was assigned for each of the 10 items of the list, and total scores summed to generate the quality score (0-10). Only studies with low risk of bias were included. Odds ratios (ORs) collected from the studies were reported as dichotomous outcomes. Unadjusted data was included if available. The I2 statistic was used to show the percentage of observed total variation across studies that is due to real heterogeneity rather than chance. A P value of less than 0.05 was accepted as statistically significant. RevMan v. 5.3 software was used.
Applicability/external validity: The authors reported that generalisation of the results should be done with caution due to the relatively few studies involved. Additionally, to the limitation of low number of studies included, the studies cover a period in which there has been a rapid change in patterns of digital devices usage, and hence the failure to find any association.
Geographic focus: The results of this review are specific to Asia and America where rapid change in patterns of digital devices is happening. The results may not be applied in LMIC, such as Africa.
Summary of quality assessment:
Overall, there is medium confidence in the conclusions about the effects of this study. Authors did not conduct a thorough search of the literature to ensure that all relevant studies were included in the review; and a summary quality assessment of included studies was not reported in the review.