Screen time and health issues in Chinese school-aged children and adolescents: a systematic review and meta-analysis

Author: Zhang Y, Tian S, Zou D, Zhang H, Pan C.

Geographical coverage: China

Sector: Burden of disease

Sub-sector: Epidemiology

Equity focus: Children.

Study population: Chinese children aged 6 to 18 years.

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background: Many literature reviews document relationships between screen time and child health (with screen time of 2 hours+ per day being associated with negative health impacts), but they only include a few studies conducted in Chinese children and adolescents.

Objectives: To evaluate relationships between screen time and health issues among Chinese school-aged children and adolescents.

Main findings:

In summary, the review found that 59.2% of articles focused on myopia reported a negative association with screen time – although the evidence overall was mixed and of moderate quality.

Authors included 252 articles reporting 268 studies with unique samples in the review. These studies investigated relationships between screen time and health issues of adiposity, myopia, psycho-behavioural problems, poor academic performance, cardiometabolic disease risks, sleep disorder, poor physical fitness, musculoskeletal injury, sub-health, and miscellaneous issues of height and pubertal growth, injury, sick leave and respiratory symptoms. The pooled odds ratio from 19 studies comparing health risks with the screen time cut-off of 2 hours per day was 1.40. The pooled effect size was 1.29 after trimming seven studies for publication bias adjustments.

Out of 76 articles that focused on myopia, 28.4% were included in the study. These comprised nine high-quality studies, 44 of moderate quality, and 23 of low quality. A negative correlation with screen time was reported in 59.2% of the myopia-focused articles. Mixed evidence: Most studies indicated positive relationships, but null and inconsistent results were observed in all three cohort studies and several cross-sectional studies. Interestingly, one study found an inverse relationship, possibly because parents of myopic children tend to limit their screen time to prevent myopia progression. This reciprocal relationship between the diagnosis and management of myopia, which involves screen time restriction, complicates the investigation of causal relationships.

Policy-relevant findings were not stated explicitly by the review authors, although the review has implications for the standard guidance of children having less than 2 hours of recreational screen time per day (in terms of showing evidence of negative health impacts).

Overall, future studies need to investigate health-specific dose effects and mechanisms of screen time. In terms of myopia specifically, the current level of evidence needs to be strengthened with more robust findings from cohort and experimental studies.

Methodology:

The study selection followed the PICOS framework and included peer-reviewed articles in Chinese or English that involved children and adolescents aged 6-18 years or in grades 1-12. The studies were observational or experimental, involved participants without psychological or physical diseases at baseline, and reported on the relationship between screen time and certain health indicators. Excluded were conference abstracts, non-original research articles, studies with participants diagnosed with diseases at baseline, repetitive publications, studies not reporting on screen time and health indicators, cross-sectional surveys associating screen time with internet addiction if addiction measures included excessive screen time, and studies with small sample sizes.

Peer-reviewed articles written in Chinese and English were retrieved from CNKI, Wanfang, PubMed, EMBASE, and Web of Science from inception to June 2020 (earliest article included from 1999). No limits were imposed on the publication date.

Each step of article selection, information extraction, and study quality appraisal involved collaborations among two to three reviewers to ensure data quality.

A study quality assessment was performed using the Downs & Black checklist. Two trained research assistants assessed the included studies independently, and a third researcher compared and resolved discrepancies.

Meta-analyses used random-effects models and mixed-effects models to calculate pooled adjusted odds ratios and 95% confidence intervals. Heterogeneity, sensitivity and publication bias were assessed using Q and I² statistics, “one-study removed” analysis, the funnel plot, trim and fill analysis, and classical fail-safe N, respectively. A mixed effects analysis was applied to generate subgroup effects by health issues using random-effects models and combine effects from subgroups to yield an overall effect using a fixed-effect model.

Applicability/external validity: Part of the premise for this study was to understand the importance of social context in mediating the relationship between screen time and health outcomes. It was envisaged that two elements specific to China (prolonged overall sedentary time and the relatively isolated cyberspace) might be particularly important in this regard.

Geographic focus: Despite the above, authors conclude that “findings from Chinese children and adolescents are generally congruent with those from the English literature, which connects screen time with adverse physical, mental and behavioural health conditions”. There is no consideration of the evidence base in relation to low and middle income countries specifically.

Summary of quality assessment:

The methods used to critically appraise studies were generally robust. However, the literature search itself was insufficiently thorough – with no evidence of reference lists being searched, experts contacted, or language or publication bias being avoided. While the approach to data analysis was generally robust, there was no discussion of which specific factors might explain differences in the results of published studies. In addition, the authors themselves indicate that many of the studies included were of low quality, with associated measurement issues relating to screen time, for example. For these reasons, we would define our confidence in the quality of this study as low.