Article Text


Outdoor activity and myopia in Singapore teenage children
  1. M Dirani1,2,
  2. L Tong1,
  3. G Gazzard3,
  4. X Zhang2,
  5. A Chia4,
  6. T L Young5,6,
  7. K A Rose7,
  8. P Mitchell8,
  9. S-M Saw2
  1. 1
    Singapore Eye Research Institute, Singapore
  2. 2
    Department of Community, Occupational and Family Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
  3. 3
    Glaucoma Research Unit, Moorfields Eye Hospital, London, UK
  4. 4
    Singapore National Eye Center, Singapore
  5. 5
    Duke National University of Singapore-Graduate Medical School, Singapore
  6. 6
    Duke University School of Medicine, Durham, North Carolina, USA
  7. 7
    Discipline of Orthoptics, University of Sydney, Sydney, Australia
  8. 8
    Department of Ophthalmology, University of Sydney, Sydney, Australia
  1. Correspondence to Dr M Dirani, Department of Community, Occupational and Family Medicine, Yong Loo Lin School of Medicine, National University of Singapore, MD3, 16 Medical Drive, Singapore 117597; dirani{at}


Aim: To investigate the relationship of outdoor activities and myopia in Singapore teenage children.

Methods: Teenage children (1249 participants), examined in the Singapore Cohort study Of Risk factors for Myopia (SCORM), during 2006 were included in analyses. Participants completed questionnaires that quantified total outdoor activity, and underwent an eye examination.

Results: The mean total time spent on outdoor activity was 3.24 h/day. The total outdoor activity (h/day) was significantly associated with myopia, odds ratio 0.90 (95% CI 0.84 to 0.96) (p = 0.004), after adjusting for age, gender, ethnicity, school type, books read per week, height, parental myopia, parental education and intelligence quotient. In addition, the total time spent outdoors was associated with significantly less myopic refraction (regression coefficient = 0.17; CI 0.10 to 0.25, p<0.001) and shorter axial length (regression coefficient −0.06 (CI −0.1 to −0.03, p<0.001). Total sports was also significantly negatively associated with myopia (p = 0.008) but not indoor sports (p = 0.16).

Conclusions: Participants who spent more time outdoors were less likely to be myopic. Thus, outdoor activity may protect against development of myopia in children, supporting recent Australian data. As near work did not predict outdoor activity, this can be viewed as an independent factor and not merely the reciprocal of near work.

Statistics from

In Western populations, myopia is found in one in three individuals;12345 however, it is alarmingly higher in selected regions of Asia, where the prevalence is as high as 80%.678 Myopia is a complex eye condition. Although several environmental risk factors, namely urbanisation and educational attainment, may contribute to the development of myopia, these factors only explain a small proportion of the variance in myopia.9

More recently, research has redirected its focus from exploring risk factors for myopia to understanding potential roles for protective parameters involved in myopia. The prevalence of myopia has been shown to be significantly lower in individuals who engage in greater amounts of outdoor activity.101112131415 Jones and coworkers13 examined 514 children and found that children who remained non-myopic (11.65 (SD 6.97) h per week) engaged in significantly greater amounts of sports and outdoor activity than those who became myopic (7.98 (6.54) h per week), p<0.001. A study by Rose and coworkers16 assessed children of two separate age groups in the Sydney Myopia Study, 1765 six-year-olds and 2367 twelve-year-olds, and found that outdoor activity was associated with a lower prevalence of myopia in children aged 12 years and in 6-year-old boys.

In our previous cohort analysis,17 outdoor activity was quantified only as a single parameter in a larger parent administered baseline questionnaire, using only one question “In the past year, how many hours per week did your child usually spend on outdoor games and activities?” Because the quantification of outdoor activity may be subject to misclassification bias, we have now used a more comprehensive approach, namely by capturing information on outdoor activity using an extensive questionnaire that was also administered in the Sydney Myopia Study.16 In this study, we aimed to investigate whether an association existed between outdoor activity (outdoor leisure and outdoor sports) and myopia in Singapore teenage children.


All children who were examined as part of the Singapore Cohort Study of the Risk Factors for Myopia (SCORM)18 during the 2006 visit (participation rate = 79.6%) were included in the analyses. A total of 1249 participants (614 boys and 635 girls) of the 1570 aged 11 to 20 years (mean age 13.7 years) who were asked to participate in the 2006 visit were included in the current study. The mean age was significantly lower for participants (mean age 13.7 years) than non-participants (mean age 14.2 years, p<0.001). There was also a significantly higher proportion of boys among the participants (58.7%) than among non-participants (49.0%, p = 0.003). There was also a significantly lower proportion of ethnic Chinese among non-participants compared with the participant group, p = 0.003.

This research followed the tenets of Declaration of Helsinki. Informed consent was obtained from parents of subjects, and written informed assent was obtained from the teenage children. This study was approved by the Singapore Eye Research Institute (SERI) Institutional Review Board (IRB).

Outdoor activity questionnaire

Each participant completed the outdoor activity questionnaire at the study site. Total outdoor activity was defined into two categories, outdoor leisure and sporting activities. The number of hours spent per day was recorded separately for each category. Separate recordings were also taken for school weekdays and school weekends. Outdoor leisure included playing in the backyard/immediate neighbourhood or walking/strolling in the neighbourhood and activities performed away from the neighbourhood, such as at a park. Sporting activities such as athletics were recorded. Questions also determined whether the sporting activities took place outdoors or indoors.

The average number of outdoor activity hours per day was calculated using the formula: (hours spent on weekday)×5+ (hours spent on weekend day)×2)/7. The total outdoor activity was defined as the sum of outdoor leisure and outdoor sporting activities. Outdoor leisure alone, outdoor sports alone, indoor sports alone and total sports were also assessed separately.

Cycloplegic refraction and biometry measures

Cycloplegia was achieved using three separate instillations of 1% cyclopentolate at 5-minute intervals. Refraction was obtained using a table-mounted autorefractor (model RK5; Canon, Tochigiken, Japan). Axial length (AL) measurements were obtained using contact ultrasound biometry (Echoscan model US-800, probe frequency 10 mHz; Nidek, Tokyo) by trained staff.

Other parameters

The intelligent quotient (IQ) was assessed using the non-verbal Raven Standard Progressive Matrix test.20 Near work activity was defined as the number of books read per week. Education was categorised as no formal education, completed elementary school leaving examination, completed high school, completed junior college or polytechnic diploma, or completed tertiary education.

Data analysis and definitions

Myopia was defined as a spherical equivalent (SE) of at least −0.5 dioptres (D). Multiple logistic regressions were performed with the presence of myopia as a dependent outcome variable, and the hours per day of outdoor activity as the main independent variable, adjusting for other covariates. Two-way interaction effects between outdoor activity and gender, outdoor activity and parental myopia, outdoor activity and nearwork activity, outdoor activity and IQ level, outdoor activity and father’s education, and outdoor activity and ethnic group were evaluated. Multiple linear regression was performed first with SE and then with AL as the dependent variable, and outdoor activity as the main covariate and other confounders as other covariates. The area under the curve was constructed using logistic regression models, and the difference between the fit of different curves was assessed using the log likelihood test. Statistical software Statistical Analysis System (SAS), version 9.1 and R 2.6.1, was used with the significance level set to be 0.05.


Outdoor activities

Chinese teenagers had the highest representation (n = 889), followed by Malays (n = 258) and other ethnicities (n = 102). Myopia was found in 868 (69.5%) of the children, hypermetropia in 58 (4.6%), with the remaining (323, 25.9%) being emmetropic. Overall, the 868 teenage children with myopia were found to participate in significantly less time for total outdoor activities (3.09 h/day, SD 1.92) than the 381 non-myopic children (3.59 h/day, SD 2.03), p<0.001 (table 1). Children without myopia participated in greater amounts of outdoor leisure (2.74 h/day, SD 1.61) than those with myopia (2.38 h/day, SD 1.49), p<0.001. The average period playing outdoor sports was greater among individuals without myopia (0.85 h/day, SD 0.80), than with myopia (0.72 h/day, SD 0.82), p = 0.007 (table 1). The playing of outdoor sports, such as football, baseball, softball and club-related sports, was associated with a reduced prevalence of myopia (p = 0.005, p = 0.0004 and p = 0.01, respectively). Additional analyses grouped sporting activities as either outdoor or indoor. The multivariate-adjusted OR of myopia for each unit (1 h) increase in indoor sports activity was 0.75, 95% CI 0.51 to 1.12, p = 0.16, after controlling for age, gender, ethnicity, school, books read per week, height, parental myopia, parental education and IQ. Moreover, there was no significant difference in the mean total time spent outdoors between children who read less than two books a week (3.26 h/day, SD 2.04) compared with those who read two or more books a week (3.24 h/day, SD 1.91), p = 0.84.

Table 1

Outdoor activity patterns for myopic and non-myopic children

Demographics and outdoor activity

The mean number of hours spent on total outdoor activity was 1.06 (SD 0.95) h/day, with no significant gender difference (boys 1.04 h/day, SD 0.96, girls 1.08 h/day, SD 0.94), p = 0.46. The mean number of hours spent playing outdoors was 1.42 (SD 0.89) h/day. However, boys spent a significantly greater amount of time playing outdoors (1.49 h/day, SD: 0.90) compared with girls (1.35 h/day, SD 0.87), p = 0.01. Similarly, boys were found to spend greater amounts of time playing outdoor sports, 0.96 h/day (SD 0.97), compared with girls, 0.56 h/day (SD 0.57), p<0.001, with the combined mean number of hours playing outdoor sports 0.76 h/day (SD 0.82).

Malay participants spent more time in mean total outdoor activity (3.94 h/day, SD 1.90), compared with Chinese (3.05 h/day, SD 1.89) and children from other ethnicities (3.21 h/day, SD 2.41), p<0.001 (table 2). Older children (p<0.001) and boys separately (p<0.001) also spent more time outdoors. Moreover, taller teenage children (height ⩾1.60 cm) spent a greater mean total outdoor activity time (3.56 h/day, SD 2.02) than shorter teenage children (height <1.59 cm) (2.89 h/day, SD 1.84), p<0.001 (table 2). In multiple logistic regression models, with myopia as the dependent variable, height was no longer associated with outdoor activity once age and gender were included. In similar multiple logistic regression models, the effect of parental myopia on outdoor activity diminished once ethnicity was included as a covariate.

Table 2

Total outdoor activity by age, gender, race, father’s education, reading, IQ level, parental myopia and height

Outdoor activity and myopia

Children who spent more time outdoors were 0.90 (95% CI 0.84 to 0.96; p = 0.004) times likely to have myopia, after adjusting for all covariates (table 3). In addition, children who engaged in more sports (indoor and outdoor) were 0.81 (95% CI 0.69 to 0.95, p = 0.008) times likely to have a myopia. Similar models were also developed adjusting for near work assessed as the number of hours per day instead of books read per week, and similar results were found. The interaction terms, outdoor×gender, outdoor×ethnicity, outdoor×IQ and outdoor×parental myopia were not found to be significant, p = 0.23, p = 0.99, p = 0.14 and p = 0.23, respectively. In a multiple linear regression model, time spent on outdoor activity was significantly associated with refractive error (regression coefficient = 0.17; CI 0.10 to 0.25) and AL (regression coefficient −0.06; CI −0.03 to 0.1), after adjustment for the same factors. Similarly, when children of Chinese ethnicity were analysed separately, the risk of myopia was 0.89 (95% CI 0.81 to 0.97) times higher for each unit increase in outdoor activity, p = 0.02. In these Chinese children, the relationships between outdoor activity (p<0.001) and SE or AL (p = 0.016) remained significant.

Table 3

Outdoor activity and myopia, refractive error or axial length

In receiver operator curve (ROC) analyses (fig 1), we found that areas under the curve were 0.632, 0.678 and 0.688 for Model 1 (age, gender, ethnicity and school), Model 2 (Model 1 plus, IQ level, number of books read per week, height and parental myopia) and Model 3 (Model 2 plus time spent outdoors) respectively. For Models 2 and 3, the likelihood ratio test suggested that adding the variable “total outdoor activity” to Model 2 improved the fit of the model (p value = 0.008).

Figure 1

Receiver operating curves of multiple logistic regression models for myopia. Model 1: age, gender, ethnicity and school. Model 2: model 1 plus intelligence quotient, number of books read per week, height and parental myopia. Model 3: model 2 plus number of hours spent in outdoor activity.


Our study findings support a protective role in outdoor activities for myopia in Singapore teenage children. Each category of outdoor activity was found to be associated with lower risks of myopia. Moreover, it had been suggested that outdoor activity may at best reflect the opposite effect of near work activity. However, our study supports an independent effect of outdoor activity, which is not merely the reciprocal of near-work activity.

Our findings concur with previous studies that showed a protective role for outdoor activity in myopia in children (6 to 12 years).1012131521 Our study findings are closely in line with those from the Sydney Myopia Study,19 which suggested a protective role for outdoor activity in myopia for 12-year-old children. However, when they assessed for this association in their younger age group (6-year-olds), it held true only for boys (p = 0.01). A gender-specific association between outdoor activity and myopia has been supported in previous studies.21 Furthermore, no direct comparisons can be made with our study findings, as our sample population consisted of predominantly Chinese Singaporean teenage children.

Our current findings differ from our previous cohort analyses in SCORM17 that reported no significant association between outdoor activity and incident myopia (RR = 1.01; 95% CI 0.98 to 1.04). However, the previous cohort analyses in SCORM were completed by parents of the participants who were aged 7 to 9 years at baseline, and only one question was used to quantify outdoor activity. The potential misclassification bias may have brought the relative risks closer to the null. The current study has also used a comprehensive outdoor activity questionnaire that was completed by participants who are now older, aged 11 to 20 years, at the time of examination.

Total sports was also significantly negatively associated with myopia in Singapore children. Because outdoor sports is a component of both total outdoor and total sports, it is difficult to definitely tease out the individual effects of sports alone or outdoor activity alone. Indoor sports alone was not associated with myopia, but outdoor leisure (outdoor activity minus sports) was associated with myopia, suggesting that outdoor activity overall may play a larger role than sports. It is possible that physical activity per se may be a surrogate marker of outdoor activity or may have an independent effect on myopia. A recent longitudinal study of 143 Danish medical students found that the number of hours of weekly physical exercise undertaken in the last 6 months was associated with reduced myopia progression over a 2-year period.22 However, the change in refractive error for those with and without myopia at baseline was not separately assessed. It remains unclear whether this physical activity preceded the refractive change or vice versa. Nonetheless, consistent with our findings, the Sydney Myopia Study[29] reported that the total time spent outdoors, rather than sports alone, played a greater protective role for myopia.

Our study disagrees with the findings from two earlier studies that did not support a protective role for outdoor activity in myopia.1423 These studies, however, were limited by small sample size, lack of control for confounding factors and inadequate quantification of outdoor activity.

The mechanisms underlying a “protective effect” of outdoor activity remain unclear. It is postulated that the effects of ultraviolet radiation exposure24 may partly explain the protective nature of outdoor activity against myopia. Moreover, with the increased intensity of light outside, pupils may be more constricted, and a greater depth of field may be achieved with less attendant image blur.16 It may also be hypothesised that biochemical changes related to increased physical activity could have an inhibitory role on eye growth.

Our study has included a large sample of children over a broad age range and obtained reproducible refractive measures; however, it is not spared from methodological limitations. These include recall bias, whereby children with myopia could underestimate the amount of time spent outdoors and participation bias whereby participants differed significantly from non-participants.

With our current knowledge and future research, the relationship between outdoor activity and myopia will also be of clinical importance, with increased outdoor activity recommended for young children to assist in preventing the onset or progression of myopia.

In conclusion, our study found that a greater number of hours spent in outdoor activities by teenage children was protective for myopia independent of the extent of reading undertaken. Future cohort studies should be conducted to evaluate the temporal relationship between outdoor activity and the development of myopia, as well as the relative effects of outdoor activity and sports. Nonetheless, future interventions to address the growing problem of myopia may be tackled in part by the promotion and advocacy of outdoor activity during childhood years.


The current study was funded by the National Medical Research Council, Singapore (NMRC/0975/2005).


View Abstract


  • Competing interests None.

  • Ethics approval Ethics approval was provided by the Singapore Eye Research Institute (SERI) Institutional Review Board (IRB).

  • Patient consent Obtained from the parents.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.