Table 3.
Study Outcomes, Part I
| No. | Study | Myopia Prevalence/Progression | Time Outdoors Hours/Day or Hours/Week | Statistical Association | Covariates for Adjustment | Limitations | Effect - Summary |
|---|---|---|---|---|---|---|---|
| Cross-Sectional Studies | |||||||
| 1 | Dirani et al (2009)25 | 69.5% (868/1249). | 3.09 ± SD 1.92h/d vs 3.59 ±SD 2.03h/d, p<0.001 in myopic vs non-myopic. | OR = 0.90 (95% CI 0.84–0.96; p=0.004). | Age, gender, ethnicity, school, books read per week, height and parental myopia, father’s education level, IQ. | Report bias, highly selected population, role of physical activity, no longitudinal data. | Protective |
| 2 | Sun et al (2018)26 | 52.02% (2544/4890). Mean refraction error −1.62(±1.82D). | Between 1.42±0.96h/d (age 15y) and 2.28±1.21h/d (at age 10y). | OR = 0.74 (95% CI 0.53–0.92; p<0.001) (multivariate analysis), OR = 0.67 (0.46–0.78; p=0.03) (univariate analysis). | NS, other variables in the analysis: age, gender, parental myopia, near work distance, near work time. | Report bias, data collection for cycloplegic autorefraction lasted 2 month, no longitudinal data. | Protective |
| Prospective Cohort Studies | |||||||
| 3 | Chen et al (2016)27 | Mean SER at baseline was −0.52D (SD ±1.97D). Annual progression according to clusters: “stable” −0.08±0.08D, “slow” −0.31±0.07D, “fast” −0.58±0.13D. | Cluster “stable”: 0.89(±1.21)h/d, cluster “slow”: 0.97(±1.14)h/d, cluster “fast”: 0.95(±1.11)h/d. | Principal component analysis (PCA), negative association with component 3 (“stabilization”). Late onset of refraction progress was associated with increased outdoor time, p=0.0422. | Age, baseline refraction. | Significant differences in the baseline characteristics among the different clusters. Novel usage of statistical model for evaluation of data. Selective cohort. Report bias. | Late myopia onset |
| 4 | Guggenheim et al (2012)28 | Prevalence at baseline (age 6y): 13.6% (1236/9109). Children who became myopic after age 11y: 14.0% (281/2005). | Grouping “low” time outdoors <3h/d vs “high” >3h/d. | Prediction of incident myopia OR = 0.65 (95% CI 0.45–0.96). (Logistic regression analysis for children who were non-myopic at age 11years (n=2005)). | Parental myopia, time reading, gender, physical activity/sedentary behavior and constant. | Non-cycloplegic autorefraction. Report bias. Loss-to follow up. Unclear how many hours/day outdoor time at aged 11 years when OR was estimated. | Protective |
| 5 | Hsu et al (2017)29 | Grouping according to progression velocity: slow change of SER max. −0.5D/y, moderate SER change of −0.5 to −1.0D/y, fast SER change of more than −1.0D/y. No difference in urban vs Suburban. | Two subcategories: outdoor activities on weekdays <1h/day vs >1h/day; on weekends <2h/day vs >2h/day. | For weekdays in the moderate progression group: OR = 0.98 (95% CI 0.79–1.22). In the fast progression group: OR=1.21 (95% CI 0.95–1.55). | NS, other variables in the analysis: age when starting near work, gender, parental myopia, baseline SE, time spent on near work, cycloplegic treatment, etc. | Cycloplegia treatment during observation time. Variations of eye examination procedures. Report bias. Children with fast annual myopia progression were more myopic at baseline and had a shorter reading distance. | No association |
| 6 | Jones-Jordan et al (2012)30 | Only myopic children included. Average age of myopia onset 10.4years (SD=1.8y), mean SE −1.82D (SD=1.06). Annual rate of progression −0.39D (SD=0.32D). | Threshold effect, at least 9 hours of outdoor activity per week necessary to see an effect. | Outdoor/sports activity did not have significant univariate associations with progression. OR=2.67 (95% CI 1.75–4.06) for at least 9 hours/week threshold. | NS, other variables in the analysis: average hours of reading for pleasure, average hours of TV, average hours of studying, etc. | Report bias. Misclassification bias: unclear how many participants were excluded because of unrealistic questionnaire responses. Missing power analysis. | Not protective following onset |
| 7 | Ma et al (2018)31 | Among 7 year old children prevalence 11.3%, among 11 year old children 52.9%. Number of children with newly developed myopia in 2 years/number of children at baseline was 30% (170/566), 29.2% (182/624) and 33.2%(149/449) for grades 1, 2, and 3. | Three categories defined at baseline: low = <4h/week, moderate = 4–9h/week, high = >9h/week. Average time spent outdoors was 0.76h/day, 0.81h/day and 0.89h/day at weekdays for children of grades 1, 2, 3. On the weekend: 1.91h/day, 1.82h/day and 1.95h/day. | OR = 1.12 (95% CI 0.77–1.64). Logistic regression showed that the 2-year incident myopia was only associated with parental myopia. | Age, gender, parental myopia and baseline SER in multivariate logistic regression analysis | Small variations of time of outdoor activity. Report bias. Questionnaires collected only at baseline. | No association |
| 8 | Saxena et al (2017)32 | Prevalence 1297 (13.1%, 1297/9884) at baseline. 8200 with normal UCVA at baseline, of these 275 developed myopia after one year: Incidence 3.4% (SE 0.2, 95% CI 3.0–3.8), myopic progression in 49.2% of children. | Mean outdoor activity 13.95±1.9 h/week, range 8–24h/week. | OR = 0.54(95% CI 0.37–0.79; p=0.002) for ≤ 14h/week versus > 14h/week, slower progression in the group with higher increased outdoor activity. | Age, gender, type of school, socioeconomic status, parental use of distance spectacles, hours of reading/writing at school and home, etc. | Misclassification bias: cycloplegic refraction only in children with VA <6/9.5. Report bias. Rather short follow-up time of one year. | Protective |
| 9 | Wu, L-J. et al (2015)33 | Mean refractive error after one year was −1.13±1.57D, changed by a mean of −0.52±0.73D. Myopic progression in 51.0% (n=2170) by ≤-0.5D. | Time outdoor for leisure (h/day) in children without myopic shift: 1.03±0.65 vs 1.40±0.70 for children with myopic shift (p=0.001). But for time outdoors for sports 0.89±0.61 vs 0.86±0.66, no difference p=0.154. | OR = 0.87, (95% CI=0.78–0.97 p<0.013) for outdoor leisure (adjusted for all covariates), but not for outdoor sports OR= 1.09 (95% CI = 0.97–1.22, p=0.135) (adjusted only for sex and age). | Age, gender, region of habitation, parental myopia, refractive status at baseline. | Noncycloplegic refraction. Report bias. Chinese length measurement in unit of “Chi”. Loss-to follow-up. | Protective |
| Study Outcomes, Part II | |||||||
| Controlled Intervention Trials | |||||||
| 10 | He et al (2015)22 | 3-year cumulative incidence rate of myopia (< −0.5D) in the intervention group: 30.4% vs in the control group: 39.5%. | Additional 40-minute class of outdoor activities was added to each school day. Similar amounts of time spent outdoors of school hours in both groups (intervention vs control) and in every grade (eg 68.04min/day vs 66.42min/day). | OR=0.73 (95% CI, 0.57–0.92, p=0.01) for the 3-year incidence rate of myopia. Significant change also in spherical equivalent. But elongation of axial length was not significantly different between the intervention group and the control group. Successful implication of the outdoor class observed in 83.5%. |
NS, other variables in the analysis: age, gender, weight, height, uncorrected visual acuity, etc. | No masking of examiners. Proportion of parents with myopia was lower in the intervention group. Sample size estimation was based on 50% reduction of incident myopia. Incomplete participation due to refusal of cycloplegic refraction. Change of myopia definition for primary outcome. No objective outdoor time measurement. | Protective |
| 11 | Jin et al (2015)23 | Significant changes in UCVA after 1-year of follow up between intervention and control group, but not between myopia and non-myopia suspected. Subgroup analysis: Incidence of new myopia onset lower in the intervention group than in the control group 3.70% vs 8.50%. Also changes in axial length were significantly lower in the intervention group (0.16+-0.3mm/year vs 0.21±0.21mm/y, p= 0.034). | According to questionnaire similar outdoor activity between intervention and control group. Intervention additional 2x20min outdoor recess. | Comparison of mean UCVA between groups. Multivariate analysis of variance of mean uncorrected visual acuity during the 1-year follow-up period: showed statistical significance with the intervention group having a better UCVA. Subgroup analysis showed lower cumulative incidence of myopia in the intervention group. | NS, other variables in the analysis: age, gender, region of habitation, parental myopia, parental education, family income. | Intervention group had more myopic parents, higher parental education and higher individual income. Differences between groups. Only small subgroup underwent cycloplegic refraction examination. No objective outdoor time measurement. No information on performance of the additional recess time. | Protective |
| 12 | Wu et al (2018)24 | Incidence of myopia in the intervention group 14.47% vs 17.40% in the control group. The intervention group showed significantly less myopic shift and axial elongation compared with the control group (0.35D vs 0.47D; 0.28 vs 0.33mm; p= 0.002 and p= 0.003) and a 54% lower risk of rapid myopia progression. | In the intervention group significant higher percentage of participants who spent more than 11h outdoor per week. | OR= 0.65 (95% CI 0.42–1.01). For progression: the intervention group had slower myopic shift than the control group (21.7% vs 31.0%), OR = 0.46 (95% CI, 0.28–0.77; p= 0.003). Average classroom clearance in the intervention group was 81.29%, in the control group 61.1%. |
NS, other variables in the analysis: age, gender, area, parental myopia, total sun hours during light meter wearing week. | Short observation time. 8 schools withdrew from study after randomization. Objective light measurement only for one week. Report bias. Other programs at the same time ongoing for myopia prevention. Classroom clearance only 20% more in the intervention group compared to control. | Slower progression |