Interventions to increase time spent outdoors for preventing incidence and progression of myopia in children

Abstract

Background

Myopia or nearsightedness is a type of refractive error. It causes people to see near objects clearly but distant objects as blurred. Good vision can be obtained if the refractive error is corrected properly but, where this is not possible, impaired vision will remain. The remaining myopia imposes a considerable personal and societal burden. In addition, the progression of myopia is more likely to be accompanied by other ocular diseases such as cataract, glaucoma and retinal detachment.

Myopia has emerged as a significant global public health problem in recent years. The World Health Organization (WHO) reported uncorrected or undercorrected myopia to be a major cause of visual impairment worldwide.

From both an individual and social perspective, it is important to prevent the onset of myopia and slow down its progression.

Observational studies have shown that children who spend more time outdoors have a lower incidence of myopia. Several other non‐Cochrane systematic reviews have focused on the association between increasing children's outdoor activity time and the prevention of myopia. However, none of these systematic reviews were limited to randomised controlled trials (RCTs), as they included all types of study designs, including observational studies and non‐RCTs, in addition to RCTs.

Objectives

To assess the effects of interventions to increase outdoor time on the incidence and progression of myopia in children.

Search methods

We searched CENTRAL, MEDLINE Ovid, Embase Ovid, ISRCTN registry, ClinicalTrials.gov, and the WHO ICTRP with no language restrictions. The databases were last searched on 24 June 2022.

Selection criteria

We included RCTs and cluster‐RCTs in which interventions were performed to increase the outdoor time for children with the aim of preventing the incidence and progression of myopia.

Data collection and analysis

We employed the standard methods recommended by Cochrane and assessed the certainty of the evidence using GRADE. We considered the following outcome measures: mean change in refractive error from baseline, incidence of myopia, mean change in the axial length from baseline, mean change in unaided distance visual acuity from baseline, quality of life and adverse event.

Main results

We included five RCTs in this review, four of which were cluster‐RCTs. The total number of participants was 10,733. The included participants were primary school children, most of whom were in first or second grade (aged six to nine years). Four cluster‐RCTs involved school‐based interventions to encourage children to spend more time outdoors. The interventions included classroom time outdoors, routine for spending recess outdoors, motivational tools for spending time outdoors, and encouragement through electronic information tools.

The intervention groups had less change in refractive errors in the direction of myopia; however, 95% confidence intervals (CIs) included no benefit or both benefit and harm at years one and three, and differences at year two included both clinically important and unimportant benefits (at 1 year: mean difference (MD) 0.08 dioptres (D), 95% CI −0.01 to 0.17; 4 studies, 1656 participants; low‐certainty evidence; at 2 years: MD 0.13 D, 95% CI 0.06 to 0.19; 4 studies, 2454 participants; moderate‐certainty evidence; at 3 years: MD 0.17 D, 95% CI −0.17 to 0.51; 1 study, 729 participants; low‐certainty evidence). Our protocol defined a difference of 0.1 D in the change in refractive error as clinically important. At one year, the difference was less than 0.1 D, but at two and three years it was more than 0.1 D.

The incidence of myopia was lower in the intervention groups compared to the control groups, but 95% CIs included no change or clinically unimportant benefits (at 1 year: 7.1% with intervention versus 9.5% with control; risk ratio (RR), 0.82, 95% CI 0.56 to 1.19; 3 studies, 1265 participants; low‐certainty evidence; at 2 years: 22.5% with intervention versus 26.7% with control; RR 0.84, 95% CI 0.72 to 0.98; 3 studies, 2104 participants; moderate‐certainty evidence; at 3 years: 30.5% with intervention versus 39.8% with control; RR 0.77, 95% CI 0.59 to 1.01; 1 study, 394 participants; moderate‐certainty evidence). Our protocol defined a difference of 3% in the incidence of myopia as clinically important. At one year, the difference was 2.4%, but there were clinically important differences between the two groups at two (4.2%) and three years (9.3%).

The intervention groups had smaller changes in axial lengths in the direction of myopia than the control groups; however, 95% CIs included no benefit or both benefit and harm at years one and three (at 1 year: MD −0.04 mm, 95% CI −0.09 to 0; 3 studies, 1666 participants; low‐certainty evidence; at 2 years: MD −0.04 mm, 95% CI −0.07 to −0.01; 3 studies, 2479 participants; moderate‐certainty evidence; at 3 years: MD −0.03 mm, 95% CI −0.13 to 0.07; 1 study, 763 participants; moderate‐certainty evidence).

No included studies reported changes in unaided distance visual acuity and quality of life. No adverse events were reported.

Authors' conclusions

The intervention methods varied from adopting outdoor activities as part of school lessons to providing information and motivation for encouraging outdoor activities.

The results of this review suggest that long‐term interventions to increase the time spent outdoors may potentially reduce the development of myopia in children. However, although the interventions may also suppress the progression of myopia, the low certainty of evidence makes it difficult to draw conclusions. Further research needs to be accumulated and reviewed.

Plain language summary

What are the benefits and risks of interventions to increase time spent outdoors for preventing the incidence and progression of myopia (nearsightedness) in children?

Key message

An intervention of spending more time outdoors may potentially reduce the onset of myopia. However, we are uncertain if the intervention might reduce the progression of myopia.

What is myopia?

Myopia, or nearsightedness, is a disease in which individuals see near objects clearly, but distant objects appear blurred. In recent years, myopia has become a major public health problem all over the world. Myopia is the main cause of poor sight globally. In addition, as myopia gets worse, other eye diseases that cause poor sight also increase; thus, an increase in myopia means an increase in the number of people with low visual acuity (that is, low clarity of vision resulting in an inability to see small details with precision). So, stopping the onset and worsening of myopia is important.

What did we want to find out?

Research has shown that more time spent outdoors could stop myopia from getting worse or prevent myopia from developing. Therefore, we performed this review to find out whether interventions to increase time spent outdoors could prevent the onset and slow the progression of myopia in children.

What did we do?

We searched the medical literature for randomised controlled trials comparing interventions to increase time spent outdoors with usual lifestyle in children. These are clinical trials in which participants are randomly assigned to treatment and control groups. They are known to provide the most reliable evidence on the effectiveness of a treatment. We compared the results from the literature and summarised the evidence. We assessed the level of confidence for each piece of evidence based on factors such as study methods, sample size and consistency of the results across studies.

What did we find?

Five studies met the criteria. The total number of participants was 10,733. Four of these studies involved school‐based interventions to encourage children to spend more time outdoors. In those studies, the schools, which were selected to be as balanced as possible in terms of area and educational level, were randomly allocated to the intervention and control groups. The interventions included classroom time outdoors, routine for spending breaks outdoors, motivational tools for spending time outdoors and encouragement through messages via electronic media.

In summary, the results of this study suggest that interventions to increase the time spent outdoors may potentially reduce the onset of myopia. Although the results showed that interventions may slow myopia progression, our certainty in the results was low. The onset of myopia was assessed as the incidence of myopia (how frequently it happened), and the progression of myopia was assessed as the change in refractive errors (such as an abnormal‐shaped eyeball, which prevents incoming light from focusing correctly on the retina (the back of the eye) to form a clear image) and axial lengths (the distance from the front to the back of the eye).

What are the limitations of the evidence?

The studies in this review monitored the children in the intervention and control groups for different times. Most school children were in the first and second grades of primary school, so it is unclear if the results are applicable to children younger or older than this. In addition, all studies were conducted in China and Taiwan, so it is not possible to conclude whether the results can be directly applied to other countries.

How up to date is this evidence?

The evidence in is up to date to August 2022.

Summary of findings

Summary of findings 1. Intervention to increase time spent outdoors compared with no intervention at one year.

Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with control group	Risk with intervention group
Intervention group compared to control group for preventing incidence and progression of myopia in children at one year
Patient or population: children under 18 years of age Setting: school or ophthalmology clinic Intervention: intervention to increase time spent outdoors (interventions that encouraged participants to spend time outdoors, for example, scheduling an additional outdoor activity class, urging children to go outside during every recess, or assigning homework that includes outdoor activities during weekends and holidays) Comparison: no intervention, a usual life
Change in refractive error at 1 year (D). The smaller the negative number, the stronger the myopia.	The mean change in refractive error ranged across the control groups from −0.23 D to −1.55 D	The mean change in refractive error in the intervention groups was on average 0.08 D better in the direction of myopia (95% CI 0.01 worse to 0.17 better)	—	1656 (4 RCTs)	⊕⊕⊝⊝ Lowa	—
Incidence of myopia at 1 year Myopia defined as ≥ −0.5 D of spherical equivalent refraction	95 per 1000	78 per 1000 (53 to 113)	RR 0.82 (0.56 to 1.19)	1265 (3 RCTs)	⊕⊕⊝⊝ Lowb	—
Change in axial length at 1 year (mm) Larger values represent greater myopia	The mean change in axial length ranged across the control groups from 0.20 mm to 0.33 mm	The mean change in axial length in the intervention groups was on average 0.04 mm lower (95% CI 0.09 lower to 0)	—	1666 (3 RCTs)	⊕⊕⊝⊝ Lowc	—
Quality of life at 1 year	No studies reported this outcome.
Adverse effects at 1 year	No studies reported this outcome.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; D: dioptre; RCT: randomised controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded one level for risk of bias because the risk of bias for the 'incomplete outcome data' in the included studies were high (weighted total of relevant studies 93.0%). The 'incomplete outcome data' domain was the most important for this review. It is difficult to blind their allocation to participants with the intervention of increasing outdoor activity. This may have had a significant impact on participants who were lost to follow‐up in the intervention and control groups. In addition, downgraded one level for imprecision because the confidence interval contained zero.

^bDowngraded one level for risk of bias because the risk of bias for the 'incomplete outcome data' in the included studies are high (weighted total of relevant studies 80.5%) and downgraded one level for imprecision because all the included results are imprecise that include a range of both no benefit and clear benefit.

^cDowngraded one level for risk of bias because the risk of bias for the 'incomplete outcome data' in the included studies are high (weighted total of relevant studies 68.2%) and downgraded one level for imprecision because the confidence interval contained zero.

Summary of findings 2. Intervention to increase time spent outdoors compared with no intervention at two years.

Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with control group	Risk with intervention group
Intervention group compared to control group for preventing incidence and progression of myopia in children at two years
Patient or population: children under 18 years of age Setting: school or ophthalmology clinic Intervention: intervention to increase time spent outdoors (interventions that encouraged participants to spend time outdoors, for example, scheduling an additional outdoor activity class, urging children to go outside during every recess, or assigning homework that includes outdoor activities during weekends and holidays) Comparison: no intervention, a usual life
Change in refractive error at 2 years (D). The smaller the negative number, the stronger the myopia.	The mean change in refractive error ranged across the control groups from −0.87 D to −2.07 D	The mean change in refractive error in the intervention groups was on average 0.13 D better in the direction of myopia (95% CI 0.06 better to 0.19 better)	—	2454 (4 RCTs)	⊕⊕⊕⊝ Moderatea	—
Incidence of myopia at 2 years Myopia defined as ≥ −0.5 D of spherical equivalent refraction.	267 per 1000	224 per 1000 (192 to 262)	RR 0.84 (0.72 to 0.98)	2104 (3 RCTs)	⊕⊕⊕⊝ Moderateb	—
Change in axial length at 2 years (mm). Larger values represent greater myopia.	The mean change in axial length ranged across the control groups from 0.49 mm to 0.61 mm	The mean change in axial length in the intervention groups was on average 0.04 mm lower (95% CI 0.07 lower to 0.01 lower)	—	2479 (3 RCTs)	⊕⊕⊕⊝ Moderateb	—
Quality of life at 2 years	No studies reported this outcome.
Adverse effects at 2 years	No studies reported this outcome.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; D: dioptre; RCT: randomised controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded one level for risk of bias because the included studies had one or more high‐risk domains.
^bDowngraded one level for imprecision because the included results are imprecise and included both no benefit and clear benefit, or both small benefit and clear benefit.

Summary of findings 3. Intervention to increase time spent outdoors compared with no intervention at three years.

Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with control group	Risk with intervention group
Intervention group compared to control group for preventing incidence and progression of myopia in children at three years
Patient or population: children under 18 years of age Setting: school or ophthalmology clinic Intervention: intervention to increase time spent outdoors (interventions that encouraged participants to spend time outdoors, for example, scheduling an additional outdoor activity class, urging children to go outside during every recess, or assigning homework that includes outdoor activities during weekends and holidays) Comparison: no intervention, a usual life
Change in refractive error at 3 years (D). The smaller the negative number, the stronger the myopia.	The mean change in refractive error in the control group was −1.59 D	The mean change in refractive error in the intervention groups was on average 0.17 D better in the direction of myopia (95% CI 0.17 worse to 0.51 better)	—	729 (1 RCT)	⊕⊕⊝⊝ Lowa	—
Incidence of myopia at 3 years Myopia defined as ≥ −0.5 D of spherical equivalent refraction.	398 per 1000	306 per 1000 (234 to 401)	RR 0.77 (0.59 to 1.01)	394 (1 RCT)	⊕⊕⊕⊝ Moderateb	—
Change in axial length at 3 years (mm). Larger values represent greater myopia.	The mean change in axial length in the control group was 0.98 mm	The mean change in axial length in the intervention groups was on average 0.03 mm lower (95% CI 0.13 lower to 0.07 higher)	—	763 (1 RCT)	⊕⊕⊕⊝ Moderateb	—
Quality of life at 3 years	No studies reported this outcome.
Adverse effects at 3 years	No studies reported this outcome.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; D: dioptre; RCT: randomised controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded one level for risk of bias because the risk of bias for the 'incomplete outcome data' in the included studies was high (weighted total of relevant studies 100%) and downgraded one level for imprecision because the included results were imprecise and included both no benefit and clear benefit.
^bDowngraded one level for imprecision because the included results were imprecise and included both no benefit and clear benefit.

Background

Description of the condition

Myopia or near‐sightedness is a condition that causes close objects to appear clear but distant ones to appear blurry. Myopia has become a significant global public health issue in recent years. The World Health Organization (WHO) reported that uncorrected or undercorrected myopia was a major cause of visual impairment worldwide (Resnikoff 2008). Myopia is increasing in East Asia, Europe and North America, and this trend is particularly remarkable in urban areas. It has been predicted that by 2050, nearly 50% of the world's population will have myopia (defined as a spherical equivalent of −0.5 dioptres (D) or less) and nearly 10% will have high myopia (defined as a spherical equivalent of −5.0 D or less) (Holden 2016). The degree of myopia is expressed as spherical equivalent (D) and the lower the value, the more myopic the eye.

Studies have indicated that myopia progresses faster when it develops in younger children (Gwiazda 2007). Early onset of myopia is associated with high myopia in adult life (Braun 1996; Jensen 1995; Liang 2004). High myopia is a significant problem because it increases the risk of several ocular diseases, including cataract, glaucoma, retinal detachment and myopic retinal degeneration, resulting in visual impairment and even blindness (Pruett 1998; Saw 2005; Saw 2006).

Description of the intervention

Several interventions, such as special multifocal‐like soft contact lenses (Walline 2013), progressive addition lenses (Hasebe 2008), bifocal spectacle lenses (Cheng 2014), overnight orthokeratology (Cho 2012), and atropine (Chia 2012; Chia 2016), have been shown to slow the progression of myopia in children. By contrast, outdoor time could be the only effective factor in postponing the onset of myopia as long as possible (French 2013; Xiong 2017), and may also be effective in controlling progression in eyes that are already myopic (Sherwin 2012).

How the intervention might work

Observational studies have shown that children who spend more time outdoors have a lower incidence of myopia (Jones 2007; Rose 2008). Various theories have been proposed to explain the protective effect of outdoor time. For example, in animal models, exposure to more intense light prevents the onset and progression of myopia (Cohen 2011; Karouta 2014). Some hypotheses about the underlying biological mechanisms also exist, such as release of dopamine from the retina (Smith 2012), and increased vitamin D (Tideman 2016). However, the underlying mechanism remains unclear.

Why it is important to do this review

Myopia is a type of refractive error. Refractive errors occur when a person does not have a normal shaped eyeball, which prevents incoming light from focusing correctly on the retina to form a clear image. Good vision can be obtained if the refractive error is corrected properly, but where this is not possible, impaired vision will remain. The remaining myopia imposes a considerable personal and societal burden. In addition, the progression of myopia is more likely to be accompanied by other ocular diseases such as cataract, glaucoma and retinal detachment. Therefore, it is important to prevent the onset of myopia and slow down its progression.

One Cochrane review updated in 2020 dealt with interventions to slow the progression of myopia in children, but did not mention interventions to increase time spent outdoors (Walline 2020). A few other non‐Cochrane systematic reviews focused on the association between increasing the time spent outdoors by children, and the prevention of myopia (Cao 2020; Deng 2019; Ho 2019; Sherwin 2012; Xiong 2017). These reviews included all types of study designs, including randomised controlled trials (RCTs), observational studies, and non‐RCTs. A systematic review limited to RCTs will provide the highest quality evidence as well as inform future research.

Objectives

To assess the effects of interventions to increase outdoor time on the incidence and progression of myopia in children.

Methods

Criteria for considering studies for this review

Types of studies

We included RCTs, including cluster‐randomised trials (such as school‐based clusters).

Types of participants

We included children under 18 years of age at the start of the trial with or without established myopia at baseline. For studies that included only a subset of relevant participants, we planned to contact the investigators to request the results for the subset. We excluded participants with ocular or systemic pathologies, such as amblyopia and strabismus, which could affect myopia progression.

Types of interventions

We included interventions that encouraged participants to spend time outdoors to slow myopia progression, compared with maintenance of usual lifestyle. Such interventions included: scheduling an additional outdoor activity class, urging children to go outside during every break or assigning homework that included outdoor activities during weekends and holidays. We included interventions regardless of the outdoor time and schedule they required, as long as they encouraged participants to go outside. The intervention could be carried out by anyone; for example, parents, teachers or both. We included any intervention that provided advice to parents and teachers indirectly resulting in the children spending time outdoors.

Types of outcome measures

Primary outcomes

• Mean change in refractive error (myopia progression) (for all participants, regardless of whether they had myopia at baseline or not), assessed as mean change in refraction (measured by spherical equivalent) from baseline to the first year of follow‐up. Any method was used to measure refractive error.

• Incidence of myopia among participants without myopia at baseline at the first year of follow‐up. Myopia was defined as at least −0.5 D of spherical equivalent refraction.

Secondary outcomes

• Mean change in refractive error (myopia progression)from baseline to less than one year or from baseline to more than one year of follow‐up. If a trial reported only data of shorter or longer follow‐up than one year, we dealt with it as the secondary outcome.

• Mean change in axial length (for all participants), from baseline to the first year of follow‐up.

• Mean change in unaided distance visual acuity (for all participants), from baseline to the first year of follow‐up.

• Quality of life, measured as a participant‐reported outcome using any validated instrument, for example, the 25‐item National Eye Institute Visual Function Questionnaire (Mangione 2001). Measured at one‐year follow‐up.

• Adverse effects that were related to the interventions; for example, sunburn that requires treatment.

Search methods for identification of studies

Electronic searches

The Cochrane Eyes and Vision Information Specialist searched the following databases for RCTs and controlled clinical trials. There were no language or publication year restrictions. The date of the search was 24 June 2022.

• Cochrane Central Register of Controlled Trials (CENTRAL; 2022, Issue 6) (which contains the Cochrane Eyes and Vision Trials Register) in the Cochrane Library (searched 24 June 2022) (Appendix 1).

• MEDLINE Ovid (1946 to 24 June 2022) (Appendix 2).

• Embase Ovid (1980 to 24 June 2022) (Appendix 3).

• International Standard Randomised Controlled Trial Number (ISRCTN) registry (www.isrctn.com/editAdvancedSearch; searched 24 June 2022) (Appendix 4).

• US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov; searched 24 June 2022) (Appendix 5).

• WHO International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp; searched 24 June 2022) (Appendix 6).

Searching other resources

We searched the reference lists of the trial reports identified to find additional trials related to this review. The review authors contacted the primary investigators of the identified trials for details if necessary.

Data collection and analysis

Selection of studies

Two review authors (AK and MM) independently assessed the titles and abstracts of all reports identified by the search. We employed the inclusion criteria described under Criteria for considering studies for this review, and used Covidence to undertake the screening. The criteria for including studies in this review were related to the participants, interventions and study design, but not to the outcomes. To avoid the risk of selective outcome reporting bias, we did not exclude studies that did not report relevant outcomes. We classified the titles and abstracts as 'yes', 'maybe', or 'no', and obtained the full‐text of reports classified as 'yes' or 'maybe'. Two review authors independently determined the studies that met the eligibility criteria and classified them as 'include' or 'exclude'. In case of disagreements, a third review author arbitrated. If a study did not contain sufficient information for its inclusion or exclusion, we contacted the study investigators to obtain the necessary information and reassessed the studies upon receiving the information. We permitted the study investigators four weeks to respond. We documented the reason for exclusion of the studies retrieved as full‐text, but assessed as not meeting the inclusion criteria.

Data extraction and management

Two review authors independently extracted the data for the primary and secondary outcomes using an online form developed by Cochrane Eyes and Vision, and resolved any disagreement through discussion. We also extracted the data concerning participant characteristics, interventions, complications, and other relevant information, such as intraclass correlation coefficients (ICC), used in the analysis (Appendix 7). We imported the data directly into Review Manager 5, and one review author checked the accuracy of the imported data (Review Manager 2014).

We extracted the following information.

• Trial characteristics: authors, publication year, journal, sponsorship, funding, etc.

• Trial methods: study design, unit of randomisation, etc.

• Participants: country, total number of participants, sex, age, eligibility criteria, a positive family history of myopia, etc.

• Interventions: content of programmes, intervention methods, programme duration, route of administration, etc.

• Outcomes: refraction, axial length, adverse events reported, measurement methods, etc.

Assessment of risk of bias in included studies

Two review authors independently assessed the risk of bias in the studies according to the methods described in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). We considered the following domains:

• generation of allocation sequence;

• allocation concealment;

• masking (blinding) of study participants and personnel;

• masking of outcome assessors;

• incomplete outcome data (attrition, intention‐to‐treat (ITT) analysis);

• selective outcome reporting;

• other risks of bias specific to cluster‐RCTs.

We classified the risk of bias across studies for each domain as 'low', 'unclear' or 'high'. If the paper reported insufficient information for assessing the risk of bias, we contacted the study investigators and asked for the additional information required to assess the risk. If the study investigators did not respond, we classified the trial based on the available information. The seventh domain, biases specific to cluster‐RCTs, included recruitment bias, baseline imbalance, loss of clusters, incorrect analysis and comparability with individually randomised trials.

We considered the assessment of the fifth domain, 'incomplete outcome data', to be the most important for this review. The reason for this is that the results would be overestimated if dropouts occurred for reasons related to the outcomes. We considered whether or not the reasons for loss to follow‐up were comparable between the treatment arms. It is desirable that the trialists clearly state individual reasons for withdrawal; however, this could be difficult in cluster‐RCTs. Therefore, we assessed the presence of a difference in the number of follow‐up participants in the treatment and control arms. We considered the risk of bias as low if the percentage of the missing outcome data was less than 20% overall and balanced between the arms. We employed a 'high risk' assessment if the dropout rate was unbalanced between the arms. We classified all the other cases as having an unclear risk of bias. In addition, we confirmed whether an ITT analysis was performed. ITT analysis is known to be the least biased way to estimate intervention effects in RCTs (Newell 1992).

In this review, masking‐related domains may not be important, because the outcomes were spherical equivalence refraction and axial lengths, which are objective endpoints that are measured by equipment.

For 'other risks', we included all forms of bias related to cluster‐RCTs, such as recruitment bias, baseline imbalance, loss of clusters, incorrect analysis, and comparability with individually randomised trials, giving priority to recruitment bias among them. Recruitment bias can occur when children are recruited to the trial after the clusters have been randomised, since knowing whether each cluster assigned to 'intervention' or 'control' could affect the types of participants the trialists recruit.

We summarised the risk of bias for each outcome within a study and across studies. We presented the risk of bias assessment on the forest plot for each meta‐analysis.

Measures of treatment effect

We calculated the mean differences (MDs) for the continuous variables and risk ratios (RRs) for the dichotomous variables. Using the same methods as the Cochrane review on myopia control (Walline 2020), we included data as reported by the original study, that is, from one eye, from each eye individually, or the mean of both eyes, and pooled the results, regardless of how the data were analysed in the original study. We reported the continuous variables using standardised mean differences (SMD) when studies used different scales. We anticipated this to only be necessary for quality of life, since most countries express refraction and axial lengths using common scales (dioptres and millimetres, respectively).

Unit of analysis issues

We planned to analyse each outcome at the individual level. When handling the results from cluster‐RCTs, we adjusted them based on the design effect. We calculated the design effect from the reported ICC and the mean cluster size. For continuous variables, we divided the original sample size by the design effect. For dichotomous variables, we divided both the original sample size and the number of events by the design effect. When studies reported results at multiple follow‐up times, we considered the results at one year as primary outcomes and the remaining follow‐ups as secondary outcomes.

We included four cluster‐RCTs, of which two studies reported ICCs (He 2015; He 2022). For these two studies, the ICCs reported in each of them were used for the analysis and for the remaining two studies where ICCs were not reported, we used the ICC reported in He 2022 instead. The reason for adopting the ICC for He 2022 rather than the ICC for He 2015 was that there were several ICCs reported for each outcome in He 2015 and the ICC of He 2022 was the mean value among the reported ICCs.

Dealing with missing data

We used imputed data when the trial investigators computed the data using an appropriate method; however, we did not impute missing data ourselves. We contacted the authors of the trial reports for missing data. When we did not receive a response, we analysed the studies based on the available information.

Assessment of heterogeneity

We assessed trials with statistical heterogeneity using the Chi² test and the I² statistic. We considered P values less than 0.1 to be statistically significant for the test of heterogeneity. We assessed the heterogeneity of the effect estimates across studies using the I² value, according to suggestions provided in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017): 0% to 40%, not important; 30% to 60%, moderate heterogeneity; 50% to 90%, substantial heterogeneity; 75% to 100%, considerable heterogeneity. When the I² value suggested heterogeneity, we explored and described the underlying reasons based on the features of the trial.

Assessment of reporting biases

When the included trials did not report our expected outcomes, such as refraction and axial lengths (which are also used to assess the incidence and progression of myopia in general), we considered the potential for bias. We assessed the outcomes' risk of reporting bias based on communication with the trial authors on any of the unreported assessed outcomes.

Considering publication bias, we constructed a funnel plot (effect size and standard error) and interpreted the results cautiously through visual inspection. Asymmetry of funnel plots is known to be mainly attributed to publication bias; however, relationships between trial size and effect size can also result in asymmetry (Egger 1997; Sterne 2000). Funnel plots were used only when the results of at least 10 trials were available, as described in Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Sterne 2017).

Data synthesis

If the I² statistic was over 50% (indicating a substantial degree of heterogeneity), we did not combine the study results in a meta‐analysis; instead, we employed structured reporting of the available effects using tables and visual displays and following guidance in Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (McKenzie 2019). In the absence of substantial statistical or clinical heterogeneity, we combined the results of the included trials in a meta‐analysis. For meta‐analysis, we employed a random‐effects model, that is, we did not assume the true effect of intervention to be the same in every study. The exception to this was when the data were sparse (e.g. fewer than three studies), where we used a fixed‐effect model since the random‐effects model would yield a less robust estimate in that situation.

If multi‐arm studies were included, we combined all relevant intervention groups of the study into a single group to create a single pair‐wise comparison.

Subgroup analysis and investigation of heterogeneity

We used the test for subgroup differences to examine differences between subgroups, as described in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017). If sufficient trials were available, we planned to analyse each outcome separately (refractive error and axial length) for a subgroup of children with established myopia at baseline, and those without myopia, where myopia was defined as at least −0.5 D of spherical equivalent refraction at baseline. Moreover, if sufficient data were available, we planned to conduct two further subgroup analyses:

• for children with a positive family history of myopia and those without a positive family history; and

• by age group (aged 12 years or less, aged more than 12 years).

Children with a positive family history are more likely to experience myopia progression (Jones 2007). Some studies have reported that interventions given at a lower age are more effective (Hiraoka 2012).

However, there were no studies that separated children with and without myopia at baseline, and it was not possible to obtain additional data, so subgroup analyses could not be performed. Similarly, no studies existed that were separated by the presence or absence of a family history of myopia. As for the age of the children, subgroup analyses could not be performed as all participants were less than 12 years of age.

Sensitivity analysis

We planned to perform the following sensitivity analyses on the primary outcomes:

• excluding studies with high risk of bias in the fifth domain 'incomplete outcome data';

• excluding industry‐funded studies.

However, there was an insufficient number of studies.

Summary of findings and assessment of the certainty of the evidence

We prepared summary of findings tables presenting relative and absolute risks. Two review authors (AK and MM) independently assessed the overall certainty of the evidence for each outcome using GRADE (GRADEpro GDT). We prepared three summary of findings tables listing results after one, two and three years of follow‐up. It was decided from the protocol stage to summarise the results after one year, but we presented the two‐ and three‐year results following discussion among review authors.

• Mean change in refractive error (measured by spherical equivalent)

• Incidence of myopia

• Mean change in axial length of all participants, measured by any method

• Quality of life

• Adverse effects (any time point)

Results

Description of studies

Results of the search

The electronic searches yielded 524 records (Figure 1). We identified a full‐text report of the study by He 2022 which was published after the electronic searches were run. After removal of 146 duplicates, we screened the remaining 379 records and removed 357 records that were clearly irrelevant. We obtained and screened full‐text reports of the remaining 22 reports. After assessment, we excluded 12 reports of 12 studies, of which nine studies had an ineligible study design and three had ineligible outcomes. See Characteristics of excluded studies table for details. For the studies with ineligible outcomes, we considered reporting bias and contacted the researchers but did not receive a response. Therefore, based on the information available, we concluded that they did not assess the relevant outcomes of this review.

1.

We included six reports of five studies (see Characteristics of included studies table for further details). We also identified two ongoing studies, which will be assessed for inclusion when data become available (ChiCTR2100051064; CTRI/2020/02/023382; see Characteristics of ongoing studies table).

We were unable to find published reports of two studies that seem to have been completed (ChiCTR‐IOC‐17010525; ChiCTR‐IOR‐17013868). We contacted the principal investigators for both studies but have not yet had a response. Therefore, these studies are awaiting classification (see Characteristics of studies awaiting classification table).

Included studies

Five studies met the inclusion criteria (He 2015; He 2022; Li 2021; Wu 2018; Yi 2011). Of these, one was a single‐centre, individual‐based RCT (Yi 2011), and four were school‐based cluster‐RCTs (He 2015; He 2022; Li 2021; Wu 2018). See Characteristics of included studies table for further details.

The earliest was a single‐centre study that was reported only in Chinese (Yi 2011). The study included school children aged seven to 11 years who visited the authors' institution and were allocated to an intervention (41 children) and control (39 children) group on an individual basis. The children in the intervention group were described as having at least 14 to 15 hours per week of outdoor activities as well as less than 30 hours per week of near and intermediate vision activities. The specific methods by which outdoor activity time was increased for the intervention group were not described in the paper. The primary outcome of the study was myopic progression of refractive error after two years. Since only school children with myopia were included, the incidence of myopia was not assessed.

The first school‐based cluster‐RCT was conducted in China between October 2010 and October 2013 (He 2015). Twelve schools were selected and allocated to intervention and control groups of six schools each. The first‐grade students (aged six to seven years) from these schools participated in the study. A total of 1903 students were recruited, with 919 in the intervention group and 929 in the control group after exclusion. The children in the intervention group received interventions to increase outdoor activity in two ways: an additional class of a 40‐minute outdoor activity at the end of the school day and the incorporation of items or rewards to encourage outdoor activity on weekends and holidays. The primary endpoint was the three‐year cumulative incidence of myopia, and the secondary endpoints were the changes in the spherical equivalent refraction and axial length. Although only three‐year results, the primary endpoint planned by the researchers, were included in the paper, data from the first and second years were also promptly disclosed and provided upon request by the study authors.

Wu 2018 conducted a school‐based cluster‐RCT in Taiwan between September 2013 and February 2015. Twenty‐four schools were selected and allocated to the intervention and control groups of 12 schools each, with seven and nine groups, respectively, eventually participating in the study. The first‐grade students (aged six to seven years) from these schools took part. There were 365 participants in the intervention group and 565 in the control group. The intervention group was subjected to an intervention programme called ROCT711, which required the students to go outdoors during breaks and while out of school for a minimum amount of time. The primary outcomes were change in the spherical equivalent and axial length at one year, and the intensity and duration of outdoor light exposure. The incidence of myopia was also reported in their paper.

He 2022 conducted a school‐based cluster‐RCT in China from October 2016 to December 2018. This study differed from the other included studies since two types of intervention were planned and compared in three arms. Twenty‐four schools were selected and randomly allocated to groups of eight schools each, one control group and two intervention groups, with first‐ and second‐year students (aged six to nine years) from each school participating in the study. In this review, we treated the two groups with different interventions as one intervention group and compared it to a control group. One intervention group was assigned to have an additional 40 minutes of outdoor activity (implemented either during the midday break or at the end of school day) and the other intervention group was assigned to have an additional 80 minutes of outdoor activity (40 minutes similar to the previous group and 40 minutes over five breaks). There were 4258 (2329 and 1929) participants in the intervention group and 2037 in the control group. The primary outcome was the two‐year cumulative myopia incidence. The secondary outcomes were the changes in the spherical equivalent and axial length at two years. Only the two‐year results were reported in the paper and the one‐year results were not disclosed despite the review authors' request, even though the investigators responded to the review authors' communication.

The most recent study included was a school‐based cluster‐RCT conducted in China between October 2018 and December 2020 (Li 2021). Twelve schools were selected and six schools each were randomly allocated to the intervention and control groups, with first‐year students (aged six to seven years) from each school participating in the study. There were 724 participants in the intervention group and 801 in the control group. In the intervention group, parents were provided with weekly health education information via a social media platform called WeChat. The information included increasing outdoor sun‐exposed activities as well as correcting eye use behaviour and limiting electronic screen time. The primary outcome was the two‐year cumulative incidence of myopia, and the secondary outcomes were the two‐year change in the spherical equivalent refraction and axial length. Only the two‐year results were included in the paper; however, data from the first year were promptly disclosed and provided upon request by the study authors.

Excluded studies

For further details, see Characteristics of excluded studies table.

We excluded nine records since they were not RCTs and three since the outcomes were ineligible. We contacted the principal investigators via email if the available information was insufficient, particularly for deciding whether the outcomes were ineligible. For example, the study that we excluded for the reason of ineligible outcome considered the increase in outdoor activity itself as an outcome and did not consider the onset of myopia or refractive values as an outcome.

Studies awaiting classification

Two studies are awaiting classification. Although they appear to have been completed, there are no published results (ChiCTR‐IOC‐17010525; ChiCTR‐IOR‐17013868; see Characteristics of studies awaiting classification table).

Ongoing studies

Two studies are ongoing and will be assessed for inclusion when data become available (ChiCTR2100051064; CTRI/2020/02/023382; see Characteristics of ongoing studies table).

Risk of bias in included studies

The risk of bias in the included studies is shown in Figure 2 and Figure 3.

2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

In the only individual‐based RCT, randomisation occurred by coin toss, which we considered at unclear risk of bias (Yi 2011). With the exception of this study, the other studies were cluster‐RCTs (He 2015; He 2022; Li 2021; Wu 2018). Cluster‐RCTs were assessed according to the random generation method of cluster assignment. All the cluster units were schools. All the studies used a method of stratified randomisation, with schools matched according to study‐specific criteria, such as geographical location, socioeconomic level and distribution of visual acuity, and randomisation was subsequently carried out. Reports from four trials described methods of random sequence generation that we considered at low risk of bias.

With regard to allocation concealment, cluster‐RCTs were assessed according to the allocation concealment method of cluster assignment. Four studies were at low risk, as random allocation took place after the participating schools had been selected (He 2015; He 2022; Li 2021; Wu 2018). The one individual‐based RCT did not specify whether random allocation took place after the study participant's declaration of participation in the study (Yi 2011). Therefore, it was possible that study participants could have made a decision to participate after knowing their allocation, which we judged at high risk of bias.

Blinding

No studies implemented masking, which we assessed at high risk of performance bias (He 2015; He 2022; Li 2021; Wu 2018; Yi 2011). The lack of masking in the intervention and control groups could have effects beyond the intervention itself; that is, concerning the presence of performance bias. However, masking is not possible for the current research question, an intervention to increase time spent outdoors to control myopia, owing to its inherent nature. Therefore, we considered that the high risk of performance bias was unavoidable.

Since the objective outcomes were based on machine measurements, whether the outcome assessor was masked or not did not have a significant impact on the results. Three studies masked outcome assessors, which we considered at low risk of bias (He 2022; Li 2021; Wu 2018). He 2015 stated that the assessors were not masked, and the study was at high risk of bias. Yi 2011 did not include sufficient information in the paper for accurate assessment, so it is likely allocation was not concealed. We considered it at high risk of bias.

Incomplete outcome data

We considered two studies at low risk of bias for incomplete outcome data because the proportion of participants who were lost to follow‐up was less than 20% and balanced between the two groups (He 2015; He 2022). The remaining three studies were at a high risk of bias for incomplete outcome data: two because the proportion of participants lost to follow‐up in the intervention and control groups was unbalanced (Yi 2011; Li 2021); and one study experienced dropouts in clusters following randomisation, and the percentage of participants lost to follow‐up was also more than 20% (Wu 2018).

Selective reporting

We considered three studies at low risk of bias for selective reporting since the outcomes described in the results were consistent with those in the trial registers (He 2015; Li 2021; Wu 2018). Two studies only reported the results of the year that each study regarded as its primary outcome (at two‐year or three‐year follow‐up), whereas the results for one‐year follow‐up were promptly disclosed in response to our request (He 2015; Li 2021). We needed to assess one study at high risk of bias for selective reporting, since it only reported the results at two‐year follow‐up and those at one year were not disclosed despite our request (He 2022). One study was at unclear risk of reporting bias because there was no published protocol prior and it could not be determined whether all the planned outcomes were reported (Yi 2011).

Other potential sources of bias

Other risks included all forms of bias associated with cluster‐RCTs: recruitment bias, baseline imbalance, loss of cluster, incorrect analysis, and comparability with individual RCTs. Of these biases, we considered recruitment bias to be the most important since it could have the greatest impact on the results. Therefore, we assessed this item by the severity of the recruitment bias, as described in the protocol.

Recruitment bias is determined based on whether the informed consent of individual participants was obtained before or after they were informed of their school allocation. None of the studies specified this process in the paper. Therefore, judging from the information available, the possibility of the existence of recruitment bias could not be ruled out in any of the studies and the four cluster‐RCTs were assessed at unclear risk of bias (He 2015; He 2022; Li 2021; Wu 2018). In He 2015, the informed consent of the participants may have been obtained before they were informed of their school assignment; however, recruitment bias could not be ruled out owing to imbalanced baseline data (parental myopia). In the remaining three studies, informed consent of the participants was obtained after they had been informed of their school assignment; thus, recruitment bias could have been present (He 2022; Li 2021; Wu 2018). However, the proportion of participants who did not consent was balanced between the two groups. In addition, Yi 2011 was an individual RCT and was not subject to the risk of bias assessment associated with cluster‐RCTs.

Effects of interventions

See: Table 1; Table 2; Table 3

See: Table 1; Table 2; and Table 3.

Change of refractive error

One of the relevant outcomes of this review is the change in the refractive error. The change at one year was our primary outcome and those assessed at other times were secondary outcomes.

The change of refractive error at one year was 0.08 D less in the direction of myopia in the intervention group than in the control groups (95% CI −0.01 to 0.17; P = 0.09, I² = 0; 4 studies, 1656 participants; low‐certainty evidence; Analysis 1.1). At two years, the intervention groups had 0.3 D less of a refractive error in the direction of myopia than the control groups (95% CI 0.06 to 0.19; P = 0.0002, I² = 0; 4 studies, 2454 participants; moderate‐certainty evidence). Only one study evaluated the changes at three years, and the intervention group had 0.17 D less change in the direction of myopia than the control group (95% CI −0.17 to 0.51; 1 study, 729 participants; low‐certainty evidence; Analysis 1.1).

1.1. Analysis.

Comparison 1: Interventions to increase time spent outdoors compared with no intervention, Outcome 1: Refractive error (D)

The protocol defined a difference of 0.1 D in the change in refractive error as clinically important. At one year, the difference was less than 0.1 D, but at two and three years, the difference was more than 0.1 D.

The intervention groups had smaller changes in the refractive errors; however, the 95% CIs included no benefit or both benefit and harms at years one and three, and differences at year two included both clinically important and unimportant benefits.

Incidence of myopia

Similar to the change in the refractive error, the incidence of myopia at one year was considered the primary outcome, while those assessed at other times were treated as secondary outcomes.

The incidence of myopia at one year was lower in the intervention group compared to the control group (7.1% with intervention versus 9.5% with control; RR 0.82, 95% CI 0.56 to 1.19; P = 0.29, I² = 0; 3 studies, 1265 participants; low‐certainty evidence; Analysis 1.2). The incidence of myopia was similarly lower in the intervention group at two and three years (at 2 years: 22.5% with intervention versus 26.7% with control; RR 0.84, 95% CI 0.72 to 0.98; P = 0.03, I² = 0; 3 studies, 2104 participants; moderate‐certainty evidence; at 3 years: 30.5% with intervention versus 39.8% with control; RR 0.77, 95% CI 0.59 to 1.01; P = 0.05; 1 study, 394 participants; moderate‐certainty evidence; Analysis 1.2).

1.2. Analysis.

Comparison 1: Interventions to increase time spent outdoors compared with no intervention, Outcome 2: Incidence of myopia

The protocol defined a difference of 3% in the incidence of myopia as clinically important. At one year, the difference was 2.4%, but there was a difference of more than 3% between groups at two (4.2%) and three years (9.3%).

The incidence of myopia was lower in the intervention groups compared to the control groups, but 95% CIs included no change or clinically unimportant benefits.

Change of axial length

We evaluated the changes in the axial length as secondary outcomes. The intervention groups had smaller changes in the axial lengths in the direction of myopia than the control groups (at 1 year: MD −0.04 mm, 95% CI −0.09 to 0; P = 0.04, I² = 0; 3 studies, 1666 participants; low‐certainty evidence; at 2 years: MD −0.04 mm, 95% CI −0.07 to −0.01; P = 0.009, I² = 0; 3 studies, 2479 participants; moderate‐certainty evidence; at 3 years: MD −0.03 mm, 95% CI −0.13 to 0.07; P = 0.55; 1 study, 763 participants; moderate‐certainty evidence; Analysis 1.3).

1.3. Analysis.

Comparison 1: Interventions to increase time spent outdoors compared with no intervention, Outcome 3: Axial length (mm)

Change of unaided distance visual acuity

Visual acuity was sometimes reported at baseline, but none of the studies treated it as an endpoint, and changes after intervention were not reported as an outcome.

Quality of life

No studies reported quality of life.

Adverse events

Possible adverse events include health problems associated with sunburn. However, the study authors reported no adverse events in any participants.

Discussion

Summary of main results

This review focused on whether interventions to increase the time spent outdoors in school children prevent the incidence and progression of myopia. We identified 379 research reports, and five studies met the inclusion criteria and were included in this review. Owing to the nature of the relevant interventions and the target population being children, four of the five studies were school‐based cluster‐RCTs. The main findings can be summarised as follows.

Interventions to increase the time spent outdoors may potentially reduce the incidence of myopia in children. The results from years one to three were in the same direction but there was imprecision that impacts on the certainty of the evidence.

Interventions to increase the time spent outdoors may prevent the progression of refractive errors in the direction of myopia in children. There was a change in the direction of weaker myopia progression in the intervention group, but the certainty of the evidence, particularly of the one‐year and the three‐year results, was low and further RCTs are needed to draw conclusions.

Interventions to increase the time spent outdoors might prevent the progression of the axial lengths in the direction of myopia in children. However, similar to the change in refraction error, there was a change in the direction of weaker myopia progression in the intervention group, but there was low‐certainty evidence.

No studies reported unaided distance visual acuity and quality of life and this could not be validated.

In summary, the results of this review suggest that long‐term interventions to increase the time spent outdoors may potentially reduce the development of myopia in children. The interventions may reduce the progression of myopia, but the low certainty of the evidence obtained makes it difficult to draw conclusions. Further research needs to be accumulated and reviewed.

We planned to conduct a subgroup analysis classified by the existence of myopia at the beginning of the study; however, this could not be carried out owing to insufficient information. Similarly, subgroup analyses classified by age or parents' history of myopia could not be performed.

Overall completeness and applicability of evidence

We believe that we were able to retrieve all relevant articles using our search strategy. The selected RCTs directly addressed our research questions; thus, we consider the evidence obtained to be highly comprehensive. However, the completeness and applicability of the latest evidence could be reduced for two reasons. First, most of the school children included in this review were in the first and second grades of primary school. The intended target population was all school children under the age of 18 years; however, none of the studies included older children. The study population could be limited to primary school children since children of the primary school age group are most vulnerable to the onset and progression of myopia. Therefore, it is unclear whether the results can be applied to children older than the primary school age group. Second, all the included studies were conducted in Asian countries. This bias could be attributed to the fact that Asia has one of the highest prevalence of myopia in the world, and the increase in myopia is a major social problem in Asia, which is a prime motivation for conducting these studies. It is not possible to conclude whether the results of this review can be directly applied to countries outside Asia.

Quality of the evidence

We assessed the certainty of evidence for the incidence of myopia and the changes in the refractive errors and axial lengths (see Table 1; Table 2; Table 3).

The main reasons for downgrading the certainty of the evidence were risk of bias and imprecision. When considering risk of bias, the most important item in this study was incomplete outcome data. A downgrading due to risk of bias was unavoidable since the difficulty in masking the participants probably influenced the loss to follow‐up in the intervention and control groups. If the loss to follow‐up in the intervention group had been related to the progression of myopia in the participants, then we would have underestimated the progression of myopia in the intervention group and the actual effect in favour of the intervention group would be expected to be smaller than the present results. In contrast, if the control group's loss to follow‐up had been related to the progression of myopia, we would have underestimated the progression of myopia in the control group and the actual effect in favour of the intervention group would be expected to be greater.

Potential biases in the review process

We followed standard Cochrane methods to minimise possible bias in the review process. However, the following limitations must be noted. The first is that none of the studies included in this review reported on adverse effects. It is possible that there were no truly adverse effects due to the nature of the intervention, but they may not have been detected because the review was not focused on adverse effects. The second was the completeness of the data collection process. We contacted the authors of the studies to request results that were not included in the published papers, most of which were disclosed upon our request. However, some studies did not disclose their results despite our requests. The possibility that the direction of the results may have factored into the willingness to disclose needs to be considered. Finally, this review was based on a literature search on 24 June 2022. However, as of May 2024, the two ongoing studies remain ongoing.

Agreements and disagreements with other studies or reviews

One systematic review of the overall myopia prevention treatments has been carried out and reported in line with the Cochrane methodology (Walline 2020). However, the review did not assess the studies addressing the impact of interventions to increase outdoor activity. Although, several systematic reviews have been conducted on the same research question as this review (time spent outdoors and prevention of myopia progression), none of these reviews were limited to RCTs (Dhakal 2022; Karthikeyan 2022). Therefore, we designed and carried out a systematic review limited to RCTs following Cochrane methodology. Their reviews concluded that interventions to increase time spent outdoors are associated with a reduction in the incidence of myopia, but that it is uncertain whether they reduce the progression of myopia. The conclusions of these studies are consistent with those of our present review.

Authors' conclusions

Implications for practice.

In the included randomised controlled trials (RCTs), the intervention methods varied from adopting outdoor activities as part of school lessons to providing information and motivation for encouraging outdoor activities.

The results of this review suggested that long‐term interventions to increase the time spent outdoors have the potential to prevent children from developing myopia. However, regarding the progression of myopia, although the interventions may have an inhibitory effect from the perspective of refractive errors and the axial lengths, we could not draw definitive conclusions based on the results of this review.

Implications for research.

The present review only included RCTs conducted in Asia (China and Taiwan) and did not include RCTs conducted outside Asia. The effects of sunlight exposure are likely to differ at different latitudes; thus, similar RCTs need to be conducted and validated in other countries around the world. Moreover, the RCTs were conducted with young primary school children; thus, the effects on older primary school children and older school children are unknown. Furthermore, the effects on sunlight exposure may also differ in terms of whether the child had myopia to begin with and whether the child has a family history of myopia. Subgroup analyses were planned but could not be carried out owing to lack of appropriate studies. Moreover, the longer the duration of the intervention, the greater the effect size was suggested for refraction; however, for axial length, effect sizes were similar regardless of intervention period. Different effects of sun exposure could exist on the refraction and axial length. Further RCTs are warranted to address these unexamined issues.

History

Protocol first published: Issue 3, 2020

Appendices

Appendix 1. CENTRAL search strategy

#1 MeSH descriptor: [Child] explode all trees
#2 MeSH descriptor: [Adolescent] explode all trees
#3 MeSH descriptor: [Pediatrics] explode all trees
#4 boy* or girl* or child* or minor* or offspring or prepubescen* or pubescen*
#5 adolescen* or juvenile* or teen or teens or teenage* or youth or youths or underage
#6 young near/1 (people or person*)
#7 paediatric* or pediatric*
#8 (primary or elementary or high or secondary) near/1 school*
#9 schoolchild* or school child* or schoolage or highschool* or daycare
#10 #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9
#11 MeSH descriptor: [Refractive Errors] this term only
#12 MeSH descriptor: [Myopia] this term only
#13 MeSH descriptor: [Astigmatism] this term only
#14 refract* near/3 error*
#15 myopia* or myopic* or myope* or astigmati*
#16 shortsight* or short sight* or nearsight*
#17 #11 or #12 or #13 or #14 or #15 or #16
#18 MeSH descriptor: [Leisure Activities] explode all trees
#19 outdoor* or out door*
#20 outside or out side
#21 #18 or #19 or #20
#22 #10 and #17 and #21

Appendix 2. MEDLINE Ovid search strategy

1. randomized controlled trial.pt.
2. (randomized or randomised).ab,ti.
3. placebo.ab,ti.
4. dt.fs.
5. randomly.ab,ti.
6. trial.ab,ti.
7. groups.ab,ti.
8. or/1‐7
9. exp animals/
10. exp humans/
11. 9 not (9 and 10)
12. 8 not 11
13. exp child/
14. exp adolescent/
15. exp pediatrics/
16. (boy$ or girl$ or child$ or minor$ or offspring or prepubescen$ or pubescen$).tw.
17. (adolescen$ or juvenile$ or teen or teens or teenage$ or youth or youths or underage).tw.
18. (young adj1 (people or person$)).tw.
19. (paediatric$ or pediatric$).tw.
20. ((primary or elementary or high or secondary) adj1 school$).tw.
21. (schoolchild$ or school child$ or schoolage or highschool$ or daycare).tw.
22. or/13‐21
23. refractive errors/
24. myopia/
25. astigmatism/
26. (refract$ adj3 error$).tw.
27. (myopia$ or myopic$ or myope$ or astigmati$).tw.
28. (shortsight$ or short sight$ or nearsight$ or near sight$).tw.
29. or/23‐28
30. exp Leisure Activities/
31. (outdoor$ or out door$).tw.
32. (outside or out side).tw.
33. or/30‐32
34. 22 and 29 and 33
35. 12 and 34

The search filter for trials at the beginning of the MEDLINE strategy is from the published paper by Glanville 2006.

Appendix 3. Embase Ovid search strategy

1. exp randomized controlled trial/
2. exp randomization/
3. exp double blind procedure/
4. exp single blind procedure/
5. random$.tw.
6. or/1‐5
7. (animal or animal experiment).sh.
8. human.sh.
9. 7 and 8
10. 7 not 9
11. 6 not 10
12. exp clinical trial/
13. (clin$ adj3 trial$).tw.
14. ((singl$ or doubl$ or trebl$ or tripl$) adj3 (blind$ or mask$)).tw.
15. exp placebo/
16. placebo$.tw.
17. random$.tw.
18. exp experimental design/
19. exp crossover procedure/
20. exp control group/
21. exp latin square design/
22. or/12‐21
23. 22 not 10
24. 23 not 11
25. exp comparative study/
26. exp evaluation/
27. exp prospective study/
28. (control$ or prospectiv$ or volunteer$).tw.
29. or/25‐28
30. 29 not 10
31. 30 not (11 or 23)
32. 11 or 24 or 31
33. exp child/
34. exp adolescent/
35. exp pediatrics/
36. (boy$ or girl$ or child$ or minor$ or offspring or prepubescen$ or pubescen$).tw.
37. (adolescen$ or juvenile$ or teen or teens or teenage$ or youth or youths or underage).tw.
38. (young adj1 (people or person$)).tw.
39. (paediatric$ or pediatric$).tw.
40. ((primary or elementary or high or secondary) adj1 school$).tw.
41. (schoolchild$ or school child$ or schoolage or highschool$ or daycare).tw.
42. or/33‐41
43. refraction error/
44. myopia/
45. (refract$ adj3 error$).tw.
46. (myopia$ or myopic$ or myope$ or astigmati$).tw.
47. (shortsight$ or short sight$ or nearsight$ or near sight$).tw.
48. or/43‐47
49. exp recreation/
50. (outdoor$ or out door$).tw.
51. (outside or out side).tw.
52. or/49‐51
53. 42 and 48 and 52
54. 32 and 53

Appendix 4. ISRCTN search strategy

(refractive error OR myopia) AND (outdoor OR out door OR outside OR out side)

Appendix 5. ClinicalTrials.gov search strategy

(refractive error OR myopia) AND (outdoor OR out door OR outside OR out side)

Appendix 6. WHO ICTRP search strategy

Title =outdoor OR out door OR outside OR out side AND Condition = refractive error OR myopia

Appendix 7. Data on study characteristics

Mandatory items		Optional items
Methods
Study design	· Parallel group RCTi.e. people randomised to treatment · Cluster RCTi.e. communities randomised to treatment · Other, specify	Exclusions after randomisation Losses to follow up Number randomised/analysed How were missing data handled? e.g., available case analysis, imputation methods Reported power calculation (Y/N), if yes, sample size and power Unusual study design/issues
Unit of randomisation/ unit of analysis	· people or cluster, if cluster RCT, what is cluster
Eyes	· One eye included in study, specify how eye selected · Two eyes included in study, both eyes received same treatment, briefly specify how analysed (best/worst/average/both and adjusted for within person correlation/both and not adjusted for within person correlation) and specify if mixture one eye and two eye
Participants
Country		Setting Ethnic group Equivalence of baseline characteristics (Y/N)
Total number of participants	This information should be collected for total study population recruited into the study. If these data are only reported for the people who were followed up only, please indicate.
Number (%) of boys and girls
Average age and age range
baseline prevalence of myopia
Number (%) of children with a family history of myopia
Inclusion criteria
Exclusion criteria
Interventions
Intervention (n = ) Comparator (n = )	· Number of people/clusters randomised · Cluster size, if cluster RCT · Content of program (schedule, duration) · Intervention methods
Outcomes
Primary and secondary outcomes as defined in study reports	Primary · Refractive error at one year · Prevalence of myopia at one year Secondary · Refractive error less than and more than one year · Axial length at one year · Unaided distance visual acuity at one year · Quality of life at one year measurement methods for each outcome Adverse events reported (Y/N)	Planned/actual length of follow up
Notes
Date conducted	Specify dates of recruitment of participants mm/yr to mm/yr	Full study name: (if applicable) Reported subgroup analyses (Y/N) Were trial investigators contacted?
Sources of funding
Declaration of interest
Included on trials registry	Y/N including registration number if available

Data and analyses

Comparison 1. Interventions to increase time spent outdoors compared with no intervention.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Refractive error (D)	5		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.1.1 At 1 year	4	1656	Mean Difference (IV, Random, 95% CI)	0.08 [‐0.01, 0.17]
1.1.2 At 2 years	4	2454	Mean Difference (IV, Random, 95% CI)	0.13 [0.06, 0.19]
1.1.3 At 3 years	1	729	Mean Difference (IV, Random, 95% CI)	0.17 [‐0.17, 0.51]
1.2 Incidence of myopia	4		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.2.2 At 1 year	3	1265	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.56, 1.19]
1.2.3 At 2 years	3	2104	Risk Ratio (M‐H, Random, 95% CI)	0.84 [0.72, 0.98]
1.2.4 At 3 years	1	394	Risk Ratio (M‐H, Random, 95% CI)	0.77 [0.59, 1.01]
1.3 Axial length (mm)	4		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.3.1 At 1 year	3	1666	Mean Difference (IV, Random, 95% CI)	‐0.04 [‐0.09, ‐0.00]
1.3.2 At 2 years	3	2479	Mean Difference (IV, Random, 95% CI)	‐0.04 [‐0.07, ‐0.01]
1.3.3 At 3 years	1	763	Mean Difference (IV, Random, 95% CI)	‐0.03 [‐0.13, 0.07]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

He 2015.

Study characteristics
Methods	Study design: cluster‐RCT Country: Guangzhou, China Cluster unit: school Unit of randomisation: school Unit of analysis: individual (only right eye in principle, if not available, left eye) Study date: October 2010 to October 2013 Definition of myopia: spherical equivalent refractive error (sphere +½ cylinder) of ≥ −0.50 D Outcome measurement methods Refractive error: cycloplegia was then induced with 3 drops of 1% cyclopentolate administered to each eye at 0, 5 and 20 minutes. Pupil light reflex and pupil dilation were checked after an additional 15 minutes and recorded. Full cycloplegia was assumed if the pupil dilated to ≥ 6 mm and the light reflex was absent. Autorefraction using the Topcon 8800K was then performed, with 3 measurements taken on the right eye and 3 on the left eye. The mean value of 3 valid measurements was calculated. Axial length: axial length and corneal curvature were measured by non‐contact partial‐coherence laser interferometry using the IOL Master (Carl Zeiss Meditec). 3 measures were taken for the right eye only. Visual acuity: assessed by following standard procedures using tumblingE ETDRS charts (Precision Vision). QoL: N/A Time point of outcome assessments Refractive error: 0, 12, 24, 36 months Axial length: 0, 12, 24, 36 months Prevalence of myopia: 0, 12, 24, 36 months
Participants	Total number of participants: 1903 students (952 + 951) before exclusion Baseline characteristics Intervention group Number of participants: 919 Age: 6.61 (SD 0.33) years Gender (male/female): 489/430 Prevalence of myopia: 16 (1.84%) Spherical equivalence refraction: 1.30 (SD 0.97) D Axial length: 22.60 (SD 0.71) mm Number of participants with a family history of myopia: 435 (53.64%) Number of clusters: 6 Unaided distance visual acuity: 0.80 Control group Number of participants: 929 Age: 6.57 (SD 0.32) years Gender (male/female): 509/420 Prevalence of myopia: 14 (1.89%) Spherical equivalence refraction: 1.26 (SD 0.81) D Axial length: 22.66 (SD 0.70) mm Number of participants with a family history of myopia: 406 (59.79%) Number of clusters: 6 Unaided distance visual acuity: 0.80 Inclusion criteria All children in grade 1 (aged 6–7 years) of the selected schools Exclusion criteria Ocular conditions (e.g. amblyopia, tropia) With other systemic conditions (e.g. mental retardation) Parental consent
Interventions	Intervention characteristics Intervention group Intervention method: 1. an additional 40‐minute outdoor activity class; 2. interventions to increase children's engagement Details of Intervention: increasing time spent outdoors was implemented in 2 ways. First, an additional 40‐minute outdoor activity class was scheduled at the end of each school day throughout the school year in the intervention schools (which was approximately 9.5 months per year, with 2.5 months as school holiday). The intervention started at the beginning of September 2009. Participation in these classes was compulsory whether or not there was consent for assessment. An outdoor activity programme brochure was distributed to grade 1 classes. The supervising teacher and headteacher were asked to fill in forms to report the outdoor activities. To maximise compliance, study staff went to 2 of the 6 intervention schools each day to inspect the outdoor classes without prior notice. The frequency of school visits was reduced to 1 day per week during the third year of the study. Details of compliance were recorded. The second part of the intervention was aimed at increasing the engagement of children in outdoor activities after school hours, especially during weekends and holidays. This was promoted to parents and children by providing items such as school bags, umbrellas, water bottles, and hats with outdoor activity logos. Children were rewarded for completing a diary of weekend outdoor activities and a regular newsletter was distributed to parents. Children and parents in control schools continued their usual patterns of activity. Control group Intervention method: none Details of intervention: N/A
Outcomes	Primary outcome 3‐year cumulative incidence rate of myopia Secondary outcomes Changes in mean spherical equivalent Changes in axial lengths
Identification	Sponsorship source: Fundamental Research Funds of the State Key Laboratory in Ophthalmology and by grant 81125007 from the National Natural Science Foundation of China. Dr He receives support from the University of Melbourne Research at Melbourne Accelerator Program Professorship. The Centre for Eye Research Australia receives operational infrastructural support from the Victorian Government. Authors' names: Mingguang He, MD, PhD; Fan Xiang, MD, PhD; Yangfa Zeng, MD; Jincheng Mai, BSc; Qianyun Chen, MSc; Jian Zhang, MSc; Wayne Smith, MD, PhD; Kathryn Rose, PhD; Ian G Morgan, PhD Institution: State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat‐sen University, Guangzhou, China, Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, East Melbourne Email: mingguang_he@yahoo.com Address: Zhongshan Ophthalmic Center, Guangzhou 510060, China Registry number: NCT00848900
Notes	Although only 3‐year results, the primary endpoint planned by the researchers, were included in the paper, data from the first and second years were also promptly disclosed and provided upon request by the authors.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation process performed with a simple random sampling using SAS version 9.2.
Allocation concealment (selection bias)	Low risk	Randomisation took place after the declaration of participation in the study, as determined from the statements in the text and the flowchart.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Due to the nature of the intervention, masking of participants was not possible.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Stated in the text that the outcome assessors were not masked.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Proportion of participants lost follow‐up < 20%, balanced between the 2 groups. Proportion of students lost to follow‐up 5.3% (49/919) in the intervention group and 4.1% (38/929) in the control group.
Selective reporting (reporting bias)	Low risk	The planned outcomes were reported without concealment. Outcomes not mentioned in the paper were also promptly disclosed at the request of the authors.
Other bias	Unclear risk	Recruitment bias was likely to be low as the entire student body from the selected schools participated and the proportion of participants who did not consent was almost balanced. But baseline imbalance existed (parental myopia).

He 2022.

Study characteristics
Methods	Study design: cluster‐RCT Country: China Cluster unit: school Unit of randomisation: school Unit of analysis: individuals (only right eye data were analysed) Study date: October 2016 to December 2018 Definition of myopia: cycloplegic spherical equivalent in right eye ≤ −0.50 D Outcome measurement methods Refractive error: assessed using KR‐8900, Topcon, Tokyo, Japan. Cycloplegia was induced with 2 (3 if cycloplegia was insufficient after 2) drops of 1% cyclopentolate (Cyclogyl; Alcon, Fort Worth, Texas, USA) 5 minutes apart and refractive error assessment was conducted 40 minutes later when pupils were > 6 mm with no light reflex. Axial length: assessed using IOL Master, Carl Zeiss Meditec, Germany Visual acuity: retro‐illuminated ETDRS chart, Guangzhou Xieyi Weishikang, Guangzhou, China QoL: N/A Time point of outcome assessments Refractive error: 0, 12, 24 months Axial length: 0, 12, 24 months Prevalence of myopia: 0, 12, 24 months
Participants	Total number of participants: 2037 participants from control group and 4258 (2329+1929) from intervention group enrolled Baseline characteristics Intervention group 1 Number of participants: 2329 Age: 7.3 (SD 0.7) years Gender (male/female): 1232/1097 Prevalence of myopia: 180 (7.7%) Spherical equivalence refraction: 1.02 (SD 1.02) D Axial length: 22.89 (SD 0.77) mm Number of participants with a family history of myopia: 1185 (55.4%) Number of clusters: 8 Unaided distance visual acuity: N/A Intervention group 2 Number of participants: 1929 Age: 7.2 (SD 0.7) years Gender (male/female): 1039/890 Prevalence of myopia: 150 (7.8%) Spherical equivalence refraction: 1.00 (SD 0.99) D Axial length: 22.86 (SD 0.74) mm Number of participants with a family history of myopia: 929 (51.3%) Number of clusters: 8 Unaided distance visual acuity: N/A Control group Number of participants: 2037 Age: 7.2 (SD 0.7) years Gender (male/female): 1075/962 Prevalence of myopia: 181 (8.9%) Spherical equivalence refraction: 0.98 (SD 1.02) D Axial length: 22.89 (SD 0.76) mm Number of participants with a family history of myopia: 1134 (60.6%) Number of clusters: 8 Unaided distance visual acuity: N/A Pretreatment: included children and those excluded children totally and stratified by groups were comparable in terms of demographic and other factors. Inclusion criteria Students from grades 1 and 2 (aged 6–9 years) were recruited from each of the selected schools Exclusion criteria With strabismus or amblyopia Using any myopia control treatment strategies (including but not limited to atropine, orthokeratology lens) Refused cycloplegia
Interventions	Intervention characteristics Intervention group 1 Intervention method: an additional outdoor time of 40 minutes per school day Details of Intervention: an additional outdoor time of 40 minutes per school day (scheduled either during the mid‐day break or at the end of school day) Intervention group 2 Intervention method: an additional outdoor time of 80 minutes per school day Details of Intervention: an additional outdoor time of 40 minutes per school day (scheduled either during the mid‐day break or at the end of school day) and another 40 minutes over 5 recesses per school day Control group Intervention method: continued with their usual outdoor activities Details of Intervention: none
Outcomes	Primary outcome 2‐year cumulative incidence rate of myopia Secondary outcomes Changes in mean spherical equivalent Changes in axial lengths Time outdoors in each group
Identification	Sponsorship source: National Key R&D Program of China (No.2021YFC2702100, No.2021YFC2702104, No.2019YFC0840607); three‐year Action Program of Shanghai Municipality for Strengthening the Construction of the Public Health System (2015–2017) (No.GWIV‐13.2); Brien Holden Vision Institute; Excellent Discipline Leader Cultivation Program of Shanghai Three Year Action Plan on Strengthening Public Health System Construction (No.GWV‐10.2‐XD09); National Natural Science Foundation of China (No.82003562). Authors' names: Xiangui He, PhD, Padmaja Sankaridurg, PhD, Jingjing Wang, PhD, Jun Chen, PhD, Thomas Naduvilath, PhD, Mingguang He, PhD, Zhuoting Zhu, PhD, Wayne Li, MD, Ian G Morgan, MD, Shuyu Xiong, PhD, Jianfeng Zhu, MD, Haidong Zou, MD, Kathryn A Rose, MD, Bo Zhang, MD Institution: Shanghai Eye Disease Prevention and Treatment Center, Shanghai Eye Hospital, Shanghai Vision Health Center & Shanghai Children Myopia Institute, Shanghai 200030, China. Department of Ophthalmology, Shanghai General Hospital, Shanghai Jiao Tong University Email: drxuxun@sjtu.edu.cn Address: No. 1440 Hongqiao Road, Shanghai, 200030, China Registry number: NCT02980445
Notes	Only the 2‐year results were reported in the paper and the 1‐year results were not disclosed despite the authors' request, even though the investigators responded to the authors' communication.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation process was performed using a simple 48 random sampling package in SAS.
Allocation concealment (selection bias)	Low risk	Based on the flowchart, it was considered that the randomisation was conducted after 24 schools had declared their participation.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Due to the nature of the intervention, masking of participants was not possible.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The outcome examiners, including technicians, optometrists and statisticians, were masked to the allocations.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Proportion of participants lost to follow‐up was < 20%, balanced between the 2 groups. Proportion of students at follow‐up was 85.4% ((1985 + 1653)/(2329 + 1929)) in the intervention group and 83.6% (1702/2037) in the control group.
Selective reporting (reporting bias)	High risk	The planned secondary endpoints had changed. In addition, although we were successful in contacting the researchers, they did not disclose results at 1 year.
Other bias	Unclear risk	Since informed consent of participants was obtained after they had been informed of their school allocation, there may have been a recruitment bias. However, the proportion of participants who disagreed with their participation was balanced between the 2 groups and was, therefore, considered 'unclear risk'.

Li 2021.

Study characteristics
Methods	Study design: cluster‐RCT Country: Guangzhou, China Cluster unit: school Unit of randomisation: school Unit of analysis: data from the right eye were used for analysis. If the data for the right eye were missing, the left eye data were used. If data were missing for both eyes, the participant was excluded from the analysis. Study date: October 2018 to December 2020 Definition of myopia: spherical equivalent refractive error (sphere of +0.5 cylinder) of ≥ −0.50 D Outcome measurement methods Refractive error: cycloplegia was induced with 3 drops of cyclopentolate hydrochloride 1%, administered at 0, 5, and 20 minutes to both eyes. Pupil dilation and pupil light reflex were checked after 15 minutes to determine whether full cycloplegia (pupil dilated ≥ 6 mm and pupil light reflex absent) was achieved. Autorefraction was measured using a commercially available device (KR‐8800; Topcon). All measurements were performed 3 times in each eye, and the mean value was calculated for each eye. Axial length: IOL Master; Carl Zeiss Meditec. Visual acuity: students' visual acuity was tested by a standard logarithmic visual acuity meter (LCD visual acuity meter JT‐8800, Shanghai Gatton Company) at a distance of 5 m until the smallest visual mark could be seen. For students wearing glasses, the bare eye visual acuity was tested first, followed by corrected visual acuity testing with glasses. For students who did not wear glasses, only bare eye vision was tested. QoL: N/A Time point of outcome assessments Refractive error: 0, 12, 24 months Axial length: 0, 12, 24 months Prevalence of myopia: 0, 12, 24 months
Participants	Total number of participants: 1525 participants (724 + 801) before exclusion Baseline characteristics Intervention group Number of participants: 724 Age: 6.3 (SD 0.5) years Gender (male/female): 395/329 Prevalence of myopia: 36 (5.0%) Spherical equivalence refraction: 1.1 (SD 1.0) D Axial length: 22.7 (SD 0.7) mm Number of participants with a family history of myopia: 426 (58.9%) Number of clusters: 6 Unaided distance visual acuity: N/A Control group Number of participants: 801 Age: 6.3 (SD 0.5) years Gender (male/female): 440/361 Prevalence of myopia: 49 (6.1%) Spherical equivalence refraction: 1.0 (SD 1.0) D Axial length: 22.7 (SD 0.7) mm Number of participants with a family history of myopia: 454 (56.7%) Number of clusters: 6 Unaided distance visual acuity: N/A Overall Number of participants: 1525 Age: 6.3 (SD 0.5) years Gender (male/female): 835/690 Prevalence of myopia: 85 (5.6%) Spherical equivalence refraction: 1.1 (SD 1.0) D Axial length: 22.7 (SD 0.7) mm Number of participants with a family history of myopia: 880 (57.7%) Number of clusters: 12 Unaided distance visual acuity: N/A Inclusion criteria All grade 1 students of the selected schools who meet the 3 following criteria Primary school students who participated in the cross‐sectional survey Completing relevant checks in the study (including eye examination and the questionnaire survey) Those who voluntarily participate in this study and whose parents or guardians voluntarily sign the informed consent form Exclusion criteria Children with myopia who were detected in cross‐sectional studies History of ophthalmic diseases (including having or having had a previous serious ophthalmic disease such as glaucoma, strabismus, trachoma, ortrauma) Other medical histories (including serious physical and organic illnesses, or other conditions that make participation in the study inadvisable) Relevant medical history: previous surgery such as excimer laser surgery formyopia
Interventions	Intervention characteristics Intervention group Intervention method: consisted of sending health education for parents by head teachers through WeChat, including increasing outdoor sun‐exposed activities, correcting eye‐use behaviour, and limiting electronic screen time. The intervention messages were sent every Monday from 1 December 2018 to 28 December 2020 (except for the month of the Chinese Lunar New Year, when parents and teachers were celebrating the holiday and co‐operation would have been low). Details of Intervention: this study used the natural WeChat group that already existed in each classroom to push the intervention messages edited by the study staff. The frequency of interventions was once a week, in order to avoid triggering negative emotions among parents by frequent tweets. The content of the interventions includes correcting children's bad eye habits, increasing children's time spent outdoors and sleeping, and limiting electronic screen using. The content of the tweets was short and simple so that parents could easily remember them. Classes' WeChat group was used by class teachers to communicate with parents about their children's performance at school and to assign daily homework, so parents would pay more attention to these online groups naturally. The intervention messages were sent every Monday from December 2018 to December 2020 (except for the month of the Chinese Lunar NewYear, when parents and teachers were celebrating the holiday and co‐operation would have been low). Teachers were asked to provide screenshots of each retweet to ensure that the intervention messages are effectively conveyed to the parents. Control group Intervention method: no intervention Details of intervention: N/A
Outcomes	Primary outcome 2‐year cumulative incidence rate of myopia Secondary outcomes Changes in mean spherical equivalent Changes in axial lengths Changes in parental awareness, children's screen time, outdoor activities, learning tools and indoor activities
Identification	Sponsorship source: grant 2017A030313465 from the Natural Science Foundation of Guangdong Province and grant 201803010062 from the Guangzhou Municipal Science and Technology Project. Authors' names: Qian Li, MD; Lan Guo, PhD; Jiayu Zhang, MD; Feng Zhao, PhD; Yin Hu, PhD; Yangfeng Guo, BD; Xueying Du, BD; Sheng Zhang, MD; Xiao Yang, PhD; Ciyong Lu, PhD Institution: Sun Yat‐Sen University Email: luciyong@mail.sysu.edu.cn, yangx_zoc@163.com Address: 54 Xianlie South Rd, Guangzhou 510060, China Registry number:ChiCTR1900022236
Notes	Only the 2‐year results were included in the paper; however, data from the first year were promptly disclosed and provided upon request by the authors.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation process performed using SPSS.
Allocation concealment (selection bias)	Low risk	Based on the flowchart, it can be concluded that randomisation was carried out after the participating schools had been selected.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Due to the nature of the intervention, masking of participants was not possible.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	It was stated in the paper that examiners were masked to the randomisation.
Incomplete outcome data (attrition bias) All outcomes	High risk	Proportion of follow‐up was 79.1% (544/(724 − 36)) in the intervention group and 87.4% (700/(801 − 49)) in the control group. Proportion of participants lost to follow‐up was imbalanced between the intervention and control groups.
Selective reporting (reporting bias)	Low risk	All the planned outcomes, which were published in the registry were reported (ChiCTR1900022236).
Other bias	Unclear risk	All first‐grade students from the selected schools were included, but eventually the proportion of participants who refused the eye examinations and questionnaires was not balanced between the intervention and control groups, which could indicate the presence of recruitment bias.

Wu 2018.

Study characteristics
Methods	Study design: cluster‐RCT Country: Taiwan Cluster unit: school Unit of randomisation: school Unit of analysis: individual Study date: September 2013 to February 2015 Definition of myopia: ≥ 0.5 D of SER on cycloplegic autorefraction performed using an autorefractometer Outcome measurement methods Refractive error: assessed using autorefractometer (KR‐8100; Topcon, Tokyo, Japan). For cycloplegia, 1 drop of 0.5% proparacaine was followed by 1 drop of 1% tropicamide (Mydriacyl; Alcon, Puurs, Belgium) and 1% cyclopentolate hydrochloride (Cyclogyl; Alcon Laboratories, Fort Worth, Texas) administered 5 minutes apart. Measurements were obtained 30 minutes after the initial drop was administered and the pupil size was > 6 mm in diameter. 5–8 consecutive readings were obtained for each child. Axial length: measurements of ocular biometric parameters (axial length and keratometry) were performed with a non‐contact ocular biometry system (Lenstar LS 900; Haag‐Streit AG, Köniz, Switzerland). This instrument works on the principle of optical low‐coherence reflectometry. Visual acuity: N/A QoL: N/A Time point of outcome assessments Refractive error: 0, 12 months Axial length: 0, 12 months Prevalence of myopia: 0, 12 months
Participants	Total number of participants: 930 Baseline characteristics Intervention group Number of participants: 365 Age: 6.35 years Gender (male/female): 201/164 Prevalence of myopia: 51 (14.01%) Spherical equivalence refraction: 0.36 (SD 1.14) D Axial length: 22.78 (SD 0.7) mm Number of participants with a family history of myopia: 257 (82.11%) Number of clusters (participated/allocated): 7/12 Unaided distance visual acuity: N/A Control group Number of participants: 565 Age: 6.34 years Gender (male/female): 284/281 Prevalence of myopia: 89 (15.81%) Spherical equivalence refraction: 0.30 (SD 0.99) D Axial length: 22.81 (SD 0.76) mm Number of participants with a family history of myopia: 381 (78.88%) Number of clusters (participated/allocated): 9/12 Unaided distance visual acuity: N/A Overall Number of participants: 930 Age: 6.34 years Gender (male/female): 485/545 Prevalence of myopia: 140 (15.10%) Spherical equivalence refraction: N/A Axial length: N/A Number of participants with a family history of myopia: N/A Number of clusters (participated/allocated): 16/24 Unaided distance visual acuity: N/A Pretreatment: the groups were fairly comparable, and there was no difference between various baseline factors (all P > 0.05). Inclusion criteria All students in Grade 1 who agreed to participate in the study Exclusion criteria None (however, the following were excluded probably at the time of statistical analysis; children with best‐corrected visual acuity not achieving 20/25 or those diagnosed with amblyopia; those undergoing orthokeratology treatment or atropine eye drop treatment)
Interventions	Intervention characteristics Intervention group Intervention method: encouragement, education according to existing policies to prevent myopia Details of intervention: ROCT711 intervention programme; 1. recess outside classroom programme, 2. outdoor‐oriented school activities, 3. weekend sun‐time passport assignment, 4. booklet for teacher–parent communication, 5. outdoor learning assignments in summer vacation, 6. eye health education for teachers and students, promote outdoor activity and 30/10 rule for myopia prevention, 7. Sport & Health 150: an initiative to promote an additional 150 minutes of exercise per week. This initiative was started during the late period of this study, 8. Tien‐Tien 120: an initiative that promotes outdoor activities for 120 minutes daily. Control group Intervention method: education according to existing policies to prevent myopia Details of intervention: 1. eye health education for teachers and students, promote outdoor activity and 30/10 rule formyopia prevention, 2. Sport & Health 150: an initiative to promote an additional 150 minutes of exercise per week. This initiative was started during the late period of this study, 3. Tien‐Tien 120: an initiative that promotes outdoor activities for 120 minutes daily.
Outcomes	Primary outcomes Changes in mean spherical equivalent Changes in axial lengths Secondary outcome Intensity and duration of outdoor light exposures
Identification	Sponsorship source: the author(s) had no proprietary or commercial interest in any materials discussed in this article. Supported by the Bureau of Health Promotion, Department of Health, Taiwan (grant numbers: C1010603, C1020515 and C1020515‐103) Authors' names: Pei‐Chang Wu, MD, PhD; Chueh‐Tan Chen, MS; Ken‐Kuo Lin, MD; Chi‐Chin Sun, MD, PhD; Chien‐Neng Kuo, MD; Hsiu‐Mei Huang, MD; Yi‐Chieh Poon, MD; Meng‐Ling Yang, MD; Chau‐Yin Chen, MD; Jou‐Chen Huang, MD; Pei‐Chen Wu, MD; I‐Hui Yang, MD; Hun‐Ju Yu Institution: Kaohsiung Medical University Email: yihsya@kmu.edu.tw, stchiou@ym.edu.tw Address: Kaohsiung Medical University, Division of Medical Statistics and Bioinformatics, Kaohsiung Medical University Hospital, No. 100 Shih‐Chuan 1st Road, Kaohsiung 807, Taiwan Registry number:NCT02082743
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The random allocation sequence was generated by a computer‐based random number‐producing algorithm and completed by a researcher not involved in the project to ensure an equal chance of a school being allocated to each group.
Allocation concealment (selection bias)	Low risk	According to the flowchart presented, randomisation was considered to be carried out after the school had declared its participation.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Due to the nature of the intervention, masking of participants was not possible.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	It was stated in the paper that measurements were performed by ophthalmologists and trained research assistants who were masked to intervention conditions.
Incomplete outcome data (attrition bias) All outcomes	High risk	Schools withdrew after randomisation – unbalanced between groups. Participants lost to follow‐up were balanced but the percentage of missing was > 20%.
Selective reporting (reporting bias)	Low risk	All the planned outcomes, which were published in the registry (NCT02082743), were reported.
Other bias	Unclear risk	Since informed consent of participants was obtained after they had been informed of their school allocation, there may have been a recruitment bias. However, the proportion of participants who disagreed with their participation was balanced between the 2 groups and was, therefore, considered 'unclear risk'.

Yi 2011.

Study characteristics
Methods	Study design: RCT Unit of randomisation: individual Unit of analysis: individuals (mean of both eyes) Study date: December 2007 to October 2008 Definition of myopia: spherical equivalent < −0.5 D Outcome measurement methods Refractive error: assessed after 3 days from the installation of 1.0% atropine ointment Axial length: N/A Visual acuity: assessed after 20 days from the installation of 1.0% atropine ointment QoL: N/A Time point of outcome assessments Refractive error: 0, 6, 12, 18, 24 months Prevalence of myopia: 0, 6, 12, 18, 24 months
Participants	Total number of participants: 39 participants from control group and 41 from intervention group were enrolled before exclusion. Finally, 29 and 37 participants took part, respectively. Baseline characteristics Intervention group Number of participants: 37 Age: 8.8 (SD 1.5) years Gender (male/female): 20/17 Prevalence of myopia: N/A Spherical equivalence refraction: −1.12 (SD 0.42) D Axial length: N/A Number of participants with a family history of myopia: N/A Number of clusters: – Unaided distance visual acuity: N/A Control group Number of participants: 29 Age: 8.9 (SD 1.7) years Gender (male/female): 15/14 Prevalence of myopia: N/A Spherical equivalence refraction: −1.02 (SD 0.30) D Axial length: N/A Number of participants with a family history of myopia: N/A Number of clusters: – Unaided distance visual acuity: N/A Pretreatment: no difference between groups Inclusion criteria Students aged 7–11 years Best corrected visual acuity ≥ 1.0 Exclusion criteria Refractive error of 2.0 ≥ D History of ocular or systemic diseases wearing contact lenses including amblyopia and strabismus
Interventions	Intervention characteristics Intervention group Intervention method: guidance on controlling close working hours and increasing outdoor activities Details of intervention: children did near‐ and middle‐vision activities for < 30 hours per week and more outdoor activities than 14–15 hours per week Control group Intervention method: none Details of intervention: –
Outcomes	Primary outcome Changes in mean spherical equivalent Secondary outcome Time in near‐vision, middle‐vision and outdoor activity
Identification	Sponsorship source: N/A Authors' names: YI Jun‐Hui, LI Rong‐Rong Institution: Department of Ophthalmology, Third Xiangya Hospital of Central South University Email: yijunhui@hotmail.com Address: Changsha, 410013, China Registry number: N/A
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation conducted by coin toss, as stated in the paper.
Allocation concealment (selection bias)	High risk	It is not possible to determine from the information in the paper whether participants could have known their allocation before they made their declaration of participation in the study. But with a method of coin toss for randomisation, it is likely the allocation was not concealed unless they mentioned how they concealed it.
Blinding of participants and personnel (performance bias) All outcomes	High risk	The specific methods of intervention were not specified, but the nature of the intervention suggested that it would not be possible to conceal the allocation.
Blinding of outcome assessment (detection bias) All outcomes	High risk	We were unable to determine the conduct of masking of outcome assessments, as this was not specified in the paper.
Incomplete outcome data (attrition bias) All outcomes	High risk	The proportion of students lost to follow‐up was 25.6% (10/39) in the intervention group and 2.4% (1/41) in the control group. The proportion of missing outcome data was out of balance between the 2 arms.
Selective reporting (reporting bias)	Unclear risk	It was unclear as to how the study was planned at the planning stage and it is not possible to determine whether reporting bias existed.
Other bias	Low risk	None.

D: dioptres; N/A: not available; QoL: quality of life; RCT: randomised controlled trial; SD: standard deviation.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Chen 2015	Ineligible outcomes
ChiCTR‐TRC‐12002347	Ineligible study design
Drury 2013	Ineligible study design
French 2016	Ineligible study design
Galvis 2014	Ineligible study design
Galvis 2018	Ineligible study design
Guo 2019	Ineligible study design
Jin 2015	Ineligible study design
NCT01388205	Ineligible outcomes
NCT03552016	Ineligible study design
Ngo 2014	Ineligible outcomes
Wu 2013	Ineligible study design

Characteristics of studies awaiting classification [ordered by study ID]

ChiCTR‐IOC‐17010525.

Methods	Intervention study, cluster randomly sampling
Participants	All participants recruited from the Anyang Childhood Eye Study (ACES), which is a school‐based cohort study mainly designed to longitudinally observe the occurrence and development of myopia as well as other ocular diseases in school‐aged children. Grade 3 students from 5 classes of 1 elementary school were recruited.
Interventions	Short message service
Outcomes	Time outdoors, light intensity
Notes	Registered in the registry but no published reports. Based on the information in the registry, we contacted the principal investigator by email but received no response. Judging from the information available, the study will be excluded at review update due to ineligible study design and outcomes.

ChiCTR‐IOR‐17013868.

Methods	Cluster‐RCT
Participants	Inclusion criteria Children in mid‐class of kindergarden No strabismus, amblyopia or other ocular diseases No systemic diseases that may affect the visual development process Exclusion criteria Unable to co‐operate in ocular examination Unwilling to participate in this study Unable to complete the 2‐year follow‐up
Interventions	Intervention group: 45 minutes of outdoor activities Control group: habitual condition
Outcomes	Year of emmetropisation, change in refraction, change in axial length
Notes	Registered in the registry but no published reports. Based on the information in the registry, we contacted the principal investigator by email but received no response. Judging from the information available, the study will be included at review update.

Characteristics of ongoing studies [ordered by study ID]

ChiCTR2100051064.

Study name	Protocol for a cluster randomized trial to evaluate the effectiveness of intervention on myopia (the ES‐SCI study intervention section)
Methods	Cluster‐randomised controlled trial
Participants	Inclusion criteria No other severe eye diseases Chinese Han ethnicity Ability of parents/guardians to provide informed consent Aged 7.5–9.5 years Exclusion criteria Serious eye diseases, including nystagmus, glaucoma, cataract, photosensitivity, retinal detachment, etc. Age of students not 7.5–9.5 years Parents or guardians did not provide informed consent
Interventions	4 types of intervention are available in this cluster‐RCT, and the 2 interventions involved in this review are as follows. Intervention method 1: population in the intervention group and the control group simultaneously implemented recess activities, and the intervention increased an additional 40 minutes of outdoor activities. The intervention is called locked the door to increase the time for outdoor activities. Intervention method 2: intervention group was arranged in the school where the government transformed lighting and desks and chairs, while the control group is set in another entire school. The intervention is called indoor lighting of light source equipment transformation, desks and chairs adjustment.
Outcomes	Outcome measurements Primary outcomes: refractive error, uncorrected visual acuity, axial length Secondary outcomes: biological samples, height, weight, blood pressure, overweight and obesity, elevated blood pressure, cost‐effectiveness Outcome time points Baseline and after 3, 6, 9, 12, 24, 36, 48 and 60 months
Starting date	October 2021
Contact information	Wei Du duwei@seu.edu.cn Southeast University
Notes	Trial registration: ChiCTR2100051064. Registered on 11 September 2021

CTRI/2020/02/023382.

Study name	A study to understand the effect of eye exercises over outdoor play in controlling the onset and progression of "short sight" in young children.
Methods	Cluster‐randomised controlled trial
Participants	Inclusion criteria Children aged 3–5 years with suspected myopia Parental history of myopia With spectacle correction at the time of study Exclusion criteria With hypermetropia, astigmatism, strabismus, and amblyopia at time of study With other eye diseases
Interventions	Vision therapy group: Vision therapy 5 exercises for 10 minutes twice a day Outdoor activities group: outdoor activities 30 minutes play twice a day, mornings: 10.00 a.m. to 10.30 a.m. and evenings: 3.00 p.m. to 3.30 p.m. Control group: no intervention
Outcomes	Outcome measurements Control in the onset and progression of myopia Outcome time points Complete screening, baseline, implementation of intervention, 6, 12, 18, 24, 30, 36, 42, 48, 54 and 60 months.
Starting date	February 2020
Contact information	Venkataramana Kalikivayi, Ahalia School of Optometry, Associate Professor, Ahalia School of Optometry, Room No: 1186, First Floor, Department of Optometry, Palakkad, KERALA, 678557, India kalikivayi@yahoo.com
Notes	Brief summary Myopia is a global burden and various studies are being conducted across the globe to find the various methods to control the progression of myopia. Since there are few studies on the effect of vision therapy in controlling the onset and progression of myopia, this longitudinal cluster randomised study was designed to evaluate the effect of vision therapy in controlling the onset and progression of myopia at an early age in comparison to outdoor activities and no intervention groups.

Differences between protocol and review

The conduct of this review was carried out according to the planned protocol (Kido 2020). For the subgroup and sensitivity analyses, sufficient information for the analysis could not be obtained and hence, it was could not be implemented according to the protocol.

With regard to the summary of findings tables, only outcomes determined at the protocol stage had been included in the table. However, several review authors suggested that the results of two and three years should also be incorporated in the tables, so we decided to add separately tables summarising the results at two and three years.

Contributions of authors

AK was involved in all stages of the review, including conception, protocol development, study selection, data extraction and analysis, assessment of risk of bias, data interpretation and writing of the review draft.

MM was responsible for the selection of studies, extraction of data and assessment of risk of bias, independently of AK. MM also reviewed the review draft.

NW, as supervisor, supported AK at all stages of the process. Specifically, this consisted of reviewing the protocol and the original review draft, and advising on the data extraction phase.

Sources of support

Internal sources

• No sources of support provided

External sources

• Public Health Agency, UK

  The HSC Research and Development (R&D) Division of the Public Health Agency funds the Cochrane Eyes and Vision editorial base at Queen's University Belfast

Declarations of interest

AK: none.

MM: received a consulting fee for reviewing a document on eye drops for preventing myopia progression from Nevakar LLC (a private pharmaceutical company). MM received payment for lectures from companies, not relating to myopia or myopia progression. MM's institution received grant funding for MM's effort from a company, not relating to myopia or myopia progression.

NW: has research funds from the Japanese Ministry of Health Labor and Welfare and the Japanese Ministry of Education, Science, and Technology. He has also received royalties from Sogensha, Medical Review and Akatsuki for publications. The results in the review are completely independent of the intention of these works.

New