[Skip to Navigation]
Sign In
Figure 1.  Secular Trend of Visual Impairment (VI) Prevalence Among Schoolchildren Aged 7 to 18 Years From 1985 to 2014, Stratified by Sex, Urban vs Rural Location, and Age Group
Secular Trend of Visual Impairment (VI) Prevalence Among Schoolchildren Aged 7 to 18 Years From 1985 to 2014, Stratified by Sex, Urban vs Rural Location, and Age Group
Figure 2.  Geographical Distribution of the Visual Impairment (VI) Prevalence Among Schoolchildren Aged 7 to 18 Years in China From 1995 to 2014
Geographical Distribution of the Visual Impairment (VI) Prevalence Among Schoolchildren Aged 7 to 18 Years in China From 1995 to 2014
Table 1.  Prevalence of Vision Impairment During 30 Years Among Students Aged 7 to 18 Years
Prevalence of Vision Impairment During 30 Years Among Students Aged 7 to 18 Years
Table 2.  Risk of Total VI and Moderate to Severe VI in 2014 Compared With 1995
Risk of Total VI and Moderate to Severe VI in 2014 Compared With 1995
Table 3.  Association of Risk of Total VI and Moderate to Severe VI With Every 100% Increase in Provincial GDP per Capita
Association of Risk of Total VI and Moderate to Severe VI With Every 100% Increase in Provincial GDP per Capita
1.
World Health Organization. Blindness and vision impairment. Key facts. Geneva, Switzerland: World Health Organization; 2018. http://www.who.int/news-room/fact-sheets/detail/blindness-and-visual-impairment. Accessed October 19, 2018.
2.
Resnikoff  S, Pascolini  D, Mariotti  SP, Pokharel  GP.  Global magnitude of visual impairment caused by uncorrected refractive errors in 2004.  Bull World Health Organ. 2008;86(1):63-70. doi:10.2471/BLT.07.041210PubMedGoogle ScholarCrossref
3.
Holden  BA, Fricke  TR, Wilson  DA,  et al.  Global prevalence of myopia and high myopia and temporal trends from 2000 through 2050.  Ophthalmology. 2016;123(5):1036-1042. doi:10.1016/j.ophtha.2016.01.006PubMedGoogle ScholarCrossref
4.
Morgan  IG, Ohno-Matsui  K, Saw  S-M.  Myopia.  Lancet. 2012;379(9827):1739-1748. doi:10.1016/S0140-6736(12)60272-4PubMedGoogle ScholarCrossref
5.
Miao  J, Wu  X.  Urbanization, socioeconomic status and health disparity in China.  Health Place. 2016;42:87-95. doi:10.1016/j.healthplace.2016.09.008PubMedGoogle ScholarCrossref
6.
Sun  H-P, Li  A, Xu  Y, Pan  CW.  Secular trends of reduced visual acuity from 1985 to 2010 and disease burden projection for 2020 and 2030 among primary and secondary school students in China.  JAMA Ophthalmol. 2015;133(3):262-268. doi:10.1001/jamaophthalmol.2014.4899PubMedGoogle ScholarCrossref
7.
Jan  CL, Congdon  N.  Chinese national policy initiative for the management of childhood myopia.  Lancet Child Adolesc Health. 2018;2(12):845-846. doi:10.1016/S2352-4642(18)30318-3PubMedGoogle ScholarCrossref
8.
CNSSCH Association.  Report on the 2000th National Survey on Students’ Constitution and Health [in Chinese]. Beijing: China College & University Press; 2002.
9.
CNSSCH Association.  Report on the 2005th National Survey on Students’ Constitution and Health [in Chinese]. Beijing: China College & University Press; 2007.
10.
CNSSCH Association.  Report on the 2010th National Survey on Students’ Constitution and Health [in Chinese]. Beijing: China College & University Press; 2012.
11.
CNSSCH Association.  Report on the 2014th National Survey on Students’ Constitution and Health [in Chinese]. Beijing: China College & University Press; 2016.
12.
Song  Y, Wang  H-J, Ma  J, Wang  Z.  Secular trends of obesity prevalence in urban Chinese children from 1985 to 2010: gender disparity.  PLoS One. 2013;8(1):e53069. doi:10.1371/journal.pone.0053069PubMedGoogle ScholarCrossref
13.
Song  Y, Wang  HJ, Dong  B, Ma  J, Wang  Z, Agardh  A.  25-year trends in gender disparity for obesity and overweight by using WHO and IOTF definitions among Chinese school-aged children: a multiple cross-sectional study.  BMJ Open. 2016;6(9):e011904. doi:10.1136/bmjopen-2016-011904PubMedGoogle ScholarCrossref
14.
Lougheed  T.  Myopia: the evidence for environmental factors.  Environ Health Perspect. 2014;122(1):A12-A19. doi:10.1289/ehp.122-A12PubMedGoogle ScholarCrossref
15.
Colenbrander  A. Visual Standards: Aspects and Ranges of Vision Loss With Emphasis on Population Surveys. Report for the International Council of Ophthalmology. San Francisco, CA: International Council of Ophthalmology; 2002;2:1-33.
16.
Santos  CA, Fiaccone  RL, Oliveira  NF,  et al.  Estimating adjusted prevalence ratio in clustered cross-sectional epidemiological data.  BMC Med Res Methodol. 2008;8:80. doi:10.1186/1471-2288-8-80PubMedGoogle ScholarCrossref
17.
Barros  AJ, Hirakata  VN.  Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio.  BMC Med Res Methodol. 2003;3:21. doi:10.1186/1471-2288-3-21PubMedGoogle ScholarCrossref
18.
Halekoh  U, Højsgaard  S, Yan  J.  The R package geepack for generalized estimating equations.  J Stat Softw. 2006;15:1-11. doi:10.18637/jss.v015.i02Google ScholarCrossref
19.
National Bureau of Statistics. China Statistical Yearbook. http://www.stats.gov.cn/tjsj/ndsj/. Accessed January 5, 2019.
20.
Liu  J, Bray  M.  Determinants of demand for private supplementary tutoring in China: findings from a national survey.  Educ Econ. 2017;25:205-218. doi:10.1080/09645292.2016.1182623Google ScholarCrossref
21.
Ross  H, Wang  Y.  The college entrance examination in China: an overview of its social-cultural foundations, existing problems, and consequences: guest editors’ introduction.  Chin Educ Soc. 2010;43:3-10. doi:10.2753/CED1061-1932430400Google ScholarCrossref
22.
Morgan  IG, Rose  KA.  Myopia and international educational performance.  Ophthalmic Physiol Opt. 2013;33(3):329-338. doi:10.1111/opo.12040PubMedGoogle ScholarCrossref
23.
Jan  C, Li  SM, Kang  MT,  et al.  Association of visual acuity with educational outcomes: a prospective cohort study.  [published online January 18, 2019].  Br J Ophthalmol. 2019;bjophthalmol-2018-313294. doi:10.1136/bjophthalmol-2018-313294PubMedGoogle Scholar
24.
Mountjoy  E, Davies  NM, Plotnikov  D,  et al.  Education and myopia: assessing the direction of causality by mendelian randomisation.  BMJ. 2018;361:k2022. doi:10.1136/bmj.k2022PubMedGoogle ScholarCrossref
25.
Ashby  R, Ohlendorf  A, Schaeffel  F.  The effect of ambient illuminance on the development of deprivation myopia in chicks.  Invest Ophthalmol Vis Sci. 2009;50(11):5348-5354. doi:10.1167/iovs.09-3419PubMedGoogle ScholarCrossref
26.
Ashby  RS, Schaeffel  F.  The effect of bright light on lens compensation in chicks.  Invest Ophthalmol Vis Sci. 2010;51(10):5247-5253. doi:10.1167/iovs.09-4689PubMedGoogle ScholarCrossref
27.
He  M, Xiang  F, Zeng  Y,  et al.  Effect of time spent outdoors at school on the development of myopia among children in China: a randomized clinical trial.  JAMA. 2015;314(11):1142-1148. doi:10.1001/jama.2015.10803PubMedGoogle ScholarCrossref
28.
Ji  CY, Chen  TJ; Working Group on Obesity in China (WGOC).  Empirical changes in the prevalence of overweight and obesity among Chinese students from 1985 to 2010 and corresponding preventive strategies.  Biomed Environ Sci. 2013;26(1):1-12.PubMedGoogle Scholar
29.
Lu  B, Congdon  N, Liu  X,  et al.  Associations between near work, outdoor activity, and myopia among adolescent students in rural China: the Xichang Pediatric Refractive Error Study report no. 2.  Arch Ophthalmol. 2009;127(6):769-775. doi:10.1001/archophthalmol.2009.105PubMedGoogle ScholarCrossref
30.
China Ministry or Education. Number of Female Students of Schools by Type and Level. http://www.moe.gov.cn/s78/A03/moe_560/jytjsj_2017/qg/201808/t20180808_344694.html. Accessed October 10, 2018.
31.
Chen  ZT, Wang  IJ, Liao  YT, Shih  YF, Lin  LL.  Polymorphisms in steroidogenesis genes, sex steroid levels, and high myopia in the Taiwanese population.  Mol Vis. 2011;17:2297-2310.PubMedGoogle Scholar
32.
Tong  L, Saw  SM, Tan  D,  et al.  Sensitivity and specificity of visual acuity screening for refractive errors in school children.  Optom Vis Sci. 2002;79(10):650-657. doi:10.1097/00006324-200210000-00011PubMedGoogle ScholarCrossref
33.
Leone  JF, Mitchell  P, Morgan  IG, Kifley  A, Rose  KA.  Use of visual acuity to screen for significant refractive errors in adolescents: is it reliable?  Arch Ophthalmol. 2010;128(7):894-899. doi:10.1001/archophthalmol.2010.134PubMedGoogle ScholarCrossref
34.
Lin  LL, Shih  YF, Hsiao  CK, Chen  CJ.  Prevalence of myopia in Taiwanese schoolchildren: 1983 to 2000.  Ann Acad Med Singapore. 2004;33(1):27-33.PubMedGoogle Scholar
35.
Bourne  RRA, Flaxman  SR, Braithwaite  T,  et al; Vision Loss Expert Group.  Magnitude, temporal trends, and projections of the global prevalence of blindness and distance and near vision impairment: a systematic review and meta-analysis.  Lancet Glob Health. 2017;5(9):e888-e897. doi:10.1016/S2214-109X(17)30293-0PubMedGoogle ScholarCrossref
Original Investigation
May 6, 2019

Association of Visual Impairment With Economic Development Among Chinese Schoolchildren

Author Affiliations
  • 1Institute of Child and Adolescent Health, School of Public Health, Peking University, Beijing, China
  • 2The George Institute for Global Health, School of Medicine, The University of New South Wales, Sydney, New South Wales, Australia
  • 3Department of Occupational & Environmental Health Sciences, School of Public Health, Peking University, Beijing, China
  • 4Stanford Prevention Research Center, Department of Medicine, Stanford University, Stanford, California
JAMA Pediatr. 2019;173(7):e190914. doi:10.1001/jamapediatrics.2019.0914
Key Points

Question  Is economic development associated with increasing child and adolescent visual impairment?

Findings  This time series analysis of 7 successive cross-sectional surveys of 1 951 084 participants found that the increasing prevalence of visual impairment in Chinese schoolchildren was associated with indicators of economic development. Girls and children living in urban areas had the highest prevalence of visual impairment, although the most rapid relative increases were in rural areas.

Meaning  Future exacerbation of a vision crisis in school-aged children may occur as economic development continues in China, a process that has global implications.

Abstract

Importance  Few studies have reported the association of economic growth with trends of visual impairment in schoolchildren in China or elsewhere.

Objectives  To describe 30-year trends and patterns in visual impairment in China and to explore the association between visual impairment and economic development.

Design, Setting, and Participants  In this time series analysis of 7 successive cross-sectional surveys from 1985 to 2014, a total of 1 951 084 schoolchildren aged 7 to 18 years from all provinces and autonomous regions of mainland China, excluding Tibet, were studied. In 1985, the survey was conducted between March to June; In 1991-2014, the surveys were conducted between September and November. Data analysis was performed from April 1, 2018, to January 31, 2019.

Exposures  The province-level gross domestic product (GDP) per capita was extracted from the China Statistical Yearbook.

Main Outcomes and Measures  Visual impairment, defined as unaided distance visual acuity of worse than 6/7.5 (20/25 Snellen equivalent), and moderate to severe visual impairment, defined as unaided distance visual acuity of worse than 6/18 (20/63 Snellen equivalent), in the worse eye.

Results  A total of 1 951 084 participants (mean [SD] age, 12.6 [3.4] years; 50.5% male) were included in the analysis. Among students aged 7 to 18 years, the prevalence of visual impairment increased from 23.7% (95% CI, 23.6%-23.8%) in 1985 to 35.1% (95% CI, 34.9%-35.3%) in 1995 to 55.0% (95% CI, 54.8%-55.3%) in 2014. In 2014, the prevalence was higher among girls (58.3%; 95% CI, 54.8%-55.3%) vs boys (51.8%; 95% CI, 51.5%-52.1%) (prevalence ratio [PR] girls vs boys, 1.12; 95% CI, 1.11-1.13) and among students living in urban (59.9%; 95% CI, 59.6%-60.2%) vs rural (50.2%; 95% CI, 49.9%-50.5%) areas (PR urban vs rural, 1.17; 95% CI, 1.16-1.18), although a more rapid relative increase in prevalence occurred in rural areas (15.3% in 1985 to 50.2% in 2014) than in urban areas (31.5% in 1985 to 59.9% in 2014). From 1995 to 2014, the GDP in China increased from $1263 to $7594 (in 2014 US$ constant price). After demographic characteristics (cluster effect of school, age, sex, urban vs rural location, and relative socioeconomic status within province) were adjusted for, the regression model revealed that every 100% increase in GDP was associated with a 20% (PR, 1.20; 95% CI, 1.20-1.21) increase in the relative risk of visual impairment and a 27% (PR, 1.27; 95% CI, 1.26-1.27) increase in the relative risk of moderate to severe visual impairment. The association was stronger in male (PR, 1.22; 95% CI, 1.21-1.23 for male [as reference] vs 1.19; 95% CI, 1.18-1.19 for female; P < .001), rural (PR, 1.32; 95% CI, 1.31-1.33 for rural [as reference] vs 1.12; 95% CI, 1.12,-1.33 for urban; P < .001), and younger age groups (PR, 1.27; 95% CI, 1.24-1.29 for 7-9 years of age [as reference] vs 1.34; 95% CI, 1.32-1.36 for 10-12 years of age; P < .001; 1.21; 95% CI, 1.20-1.22 for 13-15 years of age; P < .001; and 1.12; 95% CI, 1.11,-1.13 for 16-18 years of age; P < .001).

Conclusions and Relevance  The rapid increase of visual impairment prevalence and the association between GDP and visual impairment over time suggest that further exacerbation of childhood visual impairment may occur as economic development continues in China.

Introduction

Poor vision accounts for nearly half of all childhood disability in China and most other developing countries.1 Of the 13 million children globally with visual impairment (VI) attributable to uncorrected refractive error, half reside in China,2 and nearly all have myopia.3 These children bear the harms of lost educational opportunity and economic productivity because of poor sight2 and an increased risk of irreversible sight-threatening ocular complications.4

In the past 3 decades, China has undergone unprecedented economic growth. Economic prosperity does not necessarily translate into better health for all across all domains of health. Instead, the health benefits of economic growth may be compromised by the increase of new health issues and increasing health disparities between high-income and low-income populations. More specifically, economic development has fostered numerous social changes that may offset the benefits of reduced poverty. Resulting changes in health behavior include increases in academic pressure and indoor and sedentary lifestyles.5 Although there has been much discussion of the association of development and increased noncommunicable disease, an additional, underrecognized adverse effect is on vision.

The prevalence of VI among China’s children has been increasing,4,6 and the Chinese government is paying more attention than ever to the vision health of children and adolescents.7 However, to our knowledge, there has been no China-wide analysis using original data to investigate the current prevalence and its increase over time. Furthermore, no study to our knowledge has explored the association of economic development with VI in China or globally. This article describes secular trends in the prevalence of childhood and adolescent VI during the past 3 decades, identifies subpopulations most susceptible to VI, and explores the association between economic development and VI over time. We analyzed original individual-level data covering nearly 2 million school-aged children in China, providing an opportunity to study prevalence, patterns, and trends in childhood VI from 1985 to 2014. These patterns may provide useful insights for future trends in China and other low- and middle-income countries. To our knowledge, this article is the first such report on a comprehensive data set to address economic data and its association with VI over time.

Methods

The Chinese National Surveys on Students’ Constitution and Health (CNSSCH)8-11 is the largest nationally representative sample of school-aged children in China, providing an opportunity to study prevalence, patterns, and trends in children’s and adolescents’ VI in the past 30 years spanning from January 1, 1985, to December 31, 2014. Data were obtained from 7 national successive cross-sectional surveys from the 1985, 1991, 1995, 2000, 2005, 2010, and 2014 cycles of the CNSSCH.8-10 In 1985, the survey was conducted between March to June; in 1991-2014, the surveys were conducted between September and November. Data analysis was performed from April 1, 2018, to January 31, 2019. This project was approved by the Medical Research Ethics Committee of Peking University Health Science Center. With data collected from schools across China, approved by the Medical Research Ethics Committee of Peking University Health Science Center, the school principals determined the process for gaining informed consent (ie, written vs verbal, active vs passive) from both children and their parents. Participants’ information was anonymized before data analysis to protect participants’ privacy.

The sampling procedures, as previously described,12,13 were the same in all CNSSCH administrations at each time point. In brief, these surveys used a multistage sampling method to select students from each of the mainland provinces. Within each province, equally sized samples were selected from 3 regions at different relative socioeconomic statuses (SESs) (categorized into upper, moderate, and low based on 5 metrics: gross domestic product [GDP] per capita, total yearly income per capita, mean food consumption per capita, natural growth rate of the population, and the regional social welfare index).13 Each region was classified into 2 area groups (urban and rural) according to residential locations. An equal number of students was selected in each of the cells defined by province, regional SES class within the province, urban vs rural, sex, and age by year.

Only participants of Han ethnicity were included, who account for 92% of the total Chinese population, from all 26 provinces and 4 municipalities of mainland China except Tibet (where the Han ethnicity is a minority). All participants were selected by stratified cluster sampling; sampling took place in classes randomly selected from each grade in the selected schools among students who had lived in the same area for at least 1 year. All eligible participants completed a medical screening by a physician before survey measurement. Exclusion criteria were as follows: (1) serious organ disease (eg, heart, lung, liver, or kidney); (2) abnormal physical development (eg, pygmyism, gigantism); (3) physical impairment or deformity (eg, severe scoliosis, pectus carinatum, limp, genu valgum, and genu varum); and (4) acute disease, high fever or diarrhea during the past month and not yet recovered.

Measuring VI and Myopia

Unaided distance visual acuity (VA) for each eye was measured by experienced eye care professionals using a retroilluminated logMAR chart with tumbling-E optotypes (Precision Vision).14 In this study, we followed the definitions for population surveys set by the International Council of Ophthalmology in cooperation with the World Health Organization and the International Agency for the Prevention of Blindness.15 Mild VI was defined as unaided distance VA of worse than 6/7.5 (20/25 Snellen equivalent) and that of better or equal to 6/18 (20/63 Snellen equivalent) in the worse eye (ie, logMAR vision <0.1 and ≥0.5), moderate to severe VI was defined as unaided distance VA worse than 6/18 (20/63 Snellen equivalent, ie, logMAR vision <0.5), and the presence of VI was defined as the unaided distance VA worse than 6/7.5 (20/25 Snellen equivalent) in the worse eye.

Because of the large sample size, conducting cycloplegic refraction (the criterion standard to diagnose myopia) was not feasible. Therefore, the survey used an alternative method to identify myopia. For children with unaided distance VA less than 6/6 (<20/20 Snellen equivalent), a simple subjective refraction was used to detect crude refractive status (myopia vs hyperopia), and then their distance VA was measured again wearing positive or negative diopter spherical lens of ±0.75 diopters. Compared with the unaided distance VA, if the distance VA while wearing a positive lens reduced 1 line or more and the distance VA while wearing a negative lens improved 1 line or more, the examined eye was defined as myopic.

Other Data

We used GDP per capita as the indicator for economic development. To make data in different survey years comparable, all GDP data were adjusted to 2014 constant price in US$ based on the Consumer Price Index and average exchange rate in 2014. The GDP, Consumer Price Index, and exchange rate were sourced from the China Statistical Yearbook. From the China Statistical Yearbook, we also obtained provincial population data to ensure that all regression analyses were weighted by population.

Statistical Analysis
Descriptive Analyses

We used the χ2 test for trend to test the overall trends across years. The overall trend analyses without stratification by province were based on survey samples in 1985, 1991, 1995, 2000, 2005, 2010 and 2014, and a total of 1 951 084 students aged 7 to 18 years with VI data were included. Because the provincial data in 1985 and 1991 were unavailable, we did not weight the data by provincial population for the national trends. However, we performed analyses stratified by province from 1995 to 2014 when the provincial data were available.

Quantifying the Trend From 1995 to 2014

We used Poisson regression with robust variance and generalized estimated equation (Poisson-GEE) to quantify the prevalence ratio (PR) with 95% CI of VI in 2014 compared with 1995 to present the overall trend of VI from 1995 to 2014. In this statistical model, GEE was used to adjusted for the cluster effect of schools in the CNSSCH. The model also adjusted for sex, age, urban vs rural location, and regional SES within province. The PR is an unbiased estimator of relative risk in cross-sectional studies; it avoids the likely exaggeration in relative risk estimated by the odds ratio (OR) from logistic regression when the prevalence of outcome is high (≥10%).16,17 Stratified analyses were performed for sex, age group, and urban-rural location to test whether the PRs (2014 vs 1995) were different between subgroups (eg, female vs male, urban vs rural) by introducing a group × year interaction term to the model.

Quantifying the Association Between GDP per Capita and VI Prevalence

At the province level, we used simple linear regression to quantify the association of logGDP and VI prevalence at the province level separately for each survey year from 1995 to 2014. At the individual level, we used the Poisson-GEE model to estimate the PR and its 95% CI associated with every 100% increase in GDP per capita by adding logGDP into the model as an independent variable. The model adjusted for cluster effect of school, age, sex, urban-rural location, regional SES within province, and survey year. Stratified analyses were performed by sex, age group, urban-rural location, and survey year. We introduced a group × logGDP interaction term to the model to compare the trends between subgroups.

All regression analyses were weighted by provincial population. All analyses were conducted in R, version 3.3.1 (R Foundation for Statistical Computing), and the Poisson-GEE models were performed using the geeglm function of the geepack package, version 1.2-1 in R.18 Two-sided P < .05 was considered to be statistically significant.

Results

A total of 1 951 084 participants (mean [SD] age, 12.6 [3.4] years; 49.5% female and 50.5% male) were included in the analyses. The sample sizes in the CNSSCH for different years ranged from 5789 to 36 480 in each sex and age subgroup, and the boy to girl ratio or urban to rural ratio approximately equaled 1:1 in each survey. The participation rates were high (99.94%-100%), and even at its lowest in the year 2000, there were only 128 children sampled but not included. The sample size in each subgroup was larger in 1985 than in the subsequent years. It was determined that sample sizes could be reduced without greatly sacrificing the survey’s statistical analyses power. To ensure national representation, the surveys after 1985 selected the same schools as in 1985 but fewer students in each school; thus, more than 85% of the schools sampled were identical in each survey.

Trends in VI Prevalence From 1985 to 2014

Among students aged 7 to 18 years in the CNSSCH during 30 years, the prevalence of VI (unaided distance VA<6/7.5 [20/25 Snellen equivalent] in the worse eye) increased almost 2.5-fold from 23.7% in 1985 to 55.0% in 2014 in an almost linear fashion (P < .001 for trend). The prevalence of mild VI more than doubled from 7.3% in 1985 to 16.5% in 2014 (P < .001 for trend). The prevalence of moderate and severe VI likewise increased from 16.4% in 1985 to 38.6% in 2014 (P < .001 for trend) (Table 1).

In both sexes and in both the urban and rural areas, VI increased linearly over time (eTable 1 in the Supplement). In all surveys over time, girls had higher rates of VI than boys (Figure 1A and eTable 1 in the Supplement), and children living in urban areas had higher rates of VI compared with those living in rural areas (Figure 1B and eTable 1 in the Supplement). The increase in VI prevalence was more rapid in rural areas (15.3% in 1985 vs 50.2% in 2014) than urban areas (31.5% in 1985 vs 59.9% in 2014). Regression models adjusting for demographic characteristics (cluster effect of school, age, sex, urban-rural location, and relative SES within province) also revealed that children in 2014 had higher risk of VI than children in 1995 (overall PR for total VI, 1.57; 95% CI, 1.56-1.58 in 2014 vs 1995; overall PR for moderate to severe VI, 1.85; 95% CI, 1.83-1.87 in 2014 vs 1995) (Table 2) regardless of sex, urban-rural location, or age group. After further adjustment for group × time interaction to compare the trend between subgroups, the relative risk of VI and the moderate to severe VI increased more rapidly over time in male, rural, and younger age groups. The association was stronger in male (PR, 1.62; 95% CI, 1.60-1.63 for male [as reference] vs 1.53; 95% CI, 1.51-1.54 for female; P < .001), rural (PR, 1.79; 95% CI, 1.76-1.81 for rural [as reference] vs 1.42; 95% CI, 1.41-1.43 for urban; P < .001), and younger age groups (PR, 2.55; 95% CI, 2.46-2.65 for 7-9 years of age [as reference] vs 2.27; 95% CI, 2.22-2.32 for 10-12 years of age; P < .001; 1.53; 95% CI, 1.51-1.55 for vs 13-15 years of age; P < .001; and 1.21; 95% CI, 1.20-1.22 for 16-18 years of age; P < .001) (Table 2).

Trends in Myopia Prevalence From 1995 to 2014

No data on myopia were collected in 1985 and 1991. The prevalence of myopia increased in all age groups from 1995 to 2014 and increased as children aged (eTable 2 in the Supplement). In all survey years, the VI prevalence tripled between the 7- to 9-year age group vs the 16- to 18-year age group (Figure 1C). The prevalence of myopia was correlated with the prevalence of VI from 1995 to 2014 (total r = 0.990, P < .001; female r = 0.990, P < .001; male r = 0.991, P < .001; rural r = 0.992, P < .001; urban r = 0.988, r<0.001) (eTable 1 and eTable 2 in the Supplement). Subsequent results will focus only on VI.

Association of VI Prevalence and Economic Development From 1995 to 2014

There was an increase in VI prevalence nationwide in all provinces from 1995 to 2014. The top 3 largest regional increases in VI prevalence were seen in Tianjin (22.3% to 64.6%), Beijing (32.1% to 62.2%), and Gansu (35.6% to 60.7%), whereas the bottom 3 were in Fujian (49.6% to 56.2%), Hainan (26.8% to 37.3%), and Guizhou (24.4% to 37.2%) (eTable 3 in the Supplement). Four eastern coastal provinces, Shanghai, Zhejiang, Shandong, and Jiangsu, had a consistently higher prevalence of VI than other provinces during the 2 decades, all with a prevalence of VI of higher than 60% in 2014. The VI prevalence in Tianjin, Beijing, and Gansu also reached this level in association with rapid economic growth during those 2 decades (Figure 2 and eFigure 1 in the Supplement).

The increasing trend over time was most obvious in affluent provinces (eTable 3 in the Supplement). The GDP varied significantly across all provinces in all survey years (Figure 2). By contrast, the variance of VI prevalence consistently increased during the most recent 2 decades, strengthening the association between logGDP and VI prevalence, with r2 increasing from 0.046 in 1995 to 0.353 in 2014 and P decreasing from .28 to <.001 (eFigure 2A in the Supplement). The association between moderate to severe VI prevalence and GDP mirrored that of total VI prevalence and GDP (eFigure 2B in the Supplement and Table 3). In other words, the socioeconomic disparity in VI at the province level widened after 1995. From 1995 to 2014, the GDP in China had increased 6-fold from $1263 to $7594 (in 2014 US$ constant price).19 After adjustment for demographic characteristics (cluster effect of school, age, sex, urban-rural location, and relative SES within province), the regression model revealed that every 100% increase in GDP was associated with a 20% (PR, 1.20, 95% CI, 1.20%-1.21%) increase in the risk of VI and a 27% (PR, 1.27; 95% CI, 1.26%-1.27%) increase in the risk of moderate to severe VI. The association was stronger in male (PR, 1.22; 95% CI, 1.21-1.23 for male [as reference] vs 1.19; 95% CI, 1.18-1.19 for female; P < .001), rural (PR, 1.32; 95% CI, 1.31-1.33 for rural [as reference] vs 1.12; 95% CI, 1.12,-1.33 for urban; P < .001), and younger age groups (PR, 1.27; 95% CI, 1.24-1.29 for 7-9 years of age [as reference] vs 1.34; 95% CI, 1.32-1.36 for 10-12 years of age; P < .001; 1.21; 95% CI, 1.20-1.22 for 13-15 years of age; P < .001; and 1.12; 95% CI, 1.11-1.13 for 16-18 years of age; P < .001), which is generally consistent with eFigure 2 in the Supplement (Table 3).

Discussion

Increasing child and adolescent VI prevalence in all Chinese provinces over time was highly variable and positively associated with economic growth. These patterns are consistent with prior research,6 but multilevel modeling was possible because our data consist of individual-level student data.

There are several possible explanations for this association. Economic development is associated with behavioral and lifestyle changes that are detrimental to eyesight. For example, modern industrialization requires more close, indoor vision compared with agricultural or even some older industrial occupations. China’s rapid economic development and urbanization have created numerous environmental risk factors for the increasing myopia rates. Two major environmental risk factors for myopia have been identified: near work and outdoor activity. China, like all other countries undergoing a myopia epidemic, has intense educational schemes at primary and secondary schools.20-23 Mendelian randomization analysis has confirmed the association between educational level and myopia.24 The protective effects of increased time outdoors on myopia development have been confirmed by animal studies25,26 and a randomized clinical trial in Chinese children.27 Unfortunately, with increased academic pressure and a near work–focused sedentary lifestyle associated with new information technology, less time may be available for playing outdoors.27,28

Geographically, VI prevalence was not evenly distributed across China (Figure 2 and eTable 3 in the Supplement). More economically developed regions, especially the 4 eastern coastal provinces (Shanghai, Zhejiang, Jiangsu, and Shandong), had a higher prevalence of VI than less economically developed regions regardless of survey year. Uniform trends across provinces within each economic quartile make the association between economic development and rapidly increasing VI even more compelling. The greater association noted most recently between log GDP and VI prevalence suggests that the socioeconomic disparities in VI at the province level have widened since 1995 (eFigure 2 in the Supplement).

Girls are more likely to develop myopia, probably because of cultural factors. In traditional Chinese culture, girls are encouraged to be quiet and boys to be active. Chinese girls may also care more about white skin than Chinese boys. These factors may predispose girls to stay indoors more and to avoid sun exposure.29 Girls may study harder than boys, as indicated by longer homework time29 and higher university entrance rates.30 Biological factors may also contribute to the observed differences between boys and girls.31

Strengths and Limitations

The major strength of our study is our large, nationally representative data collected using the same study protocol over time, which illuminated patterns and associations that were not possible with smaller samples. This research provides high-quality evidence of a cohort effect of deteriorating vision in children over time across China.

Our study also has limitations. Because of its serial cross-sectional nature, we cannot establish a causal relationship between increasing GDP and VI; we can only postulate that economic development is associated with VI. Cycloplegic refraction (the gold standard for assessing myopia) was not feasible given the survey’s scope. However, reduced VA appears to be a reasonable surrogate for myopia,6 because even without simple subjective refraction, our definition of VI measured by a simple vision test achieved a sensitivity of 84.0% to 97.8% and a specificity of 88.0% to 97.1% in screening for myopia.32,33 Vision testing plus a simple subjective refraction method has been used as an efficient way to screen for myopia in China for more than 20 years.6 In the Chinese school setting, where 80% to 90% of vision impairment is attributable to uncorrected myopia,34 the strategy of using uncorrected VA as a surrogate for myopia is justified. Our definition of VI is distinct from the World Health Organization definition of VI, which is based on presenting VA. We based our definitions on unaided distance VA in the worse eye following the guidelines set by International Council of Ophthalmology.15 This definition allows us to detect the emergence of VI before any treatment (such as spectacles) was given but limits comparison between our results and the previous literature.35

Conclusions

Our findings may have considerable implications for global health policies and practices. The findings suggest that China’s economic growth has been associated with increased childhood prevalence of myopia that may increase in the future. China’s national high rate of myopia, 55% in 2014, included highly urbanized areas with myopia rates greater than 60%, as well as many lower-rate rural areas undergoing economic growth. These findings suggest a need for policy strategies to provide adequate eye care and broad myopia prevention.7 China’s long-term patterns, shown using our unique national data sources, may also have implications for other developing countries now experiencing rapid economic growth.

Back to top
Article Information

Accepted for Publication: February 28, 2018.

Published Online: May 6, 2019. doi:10.1001/jamapediatrics.2019.0914

Open Access: This article is published under the JN-OA license and is free to read on the day of publication.

Corresponding Authors: Jun Ma, MD, PhD (majunt@bjmu.edu.cn), and Yi Song, MD, PhD (songyi@bjmu.edu.cn), Institute of Child and Adolescent Health and School of Public Health, Peking University, Room 306, Building of School of Public Health, 38 Xueyuan Rd, Haidian District, Beijing, China 100191.

Author Contributions: Ms Jan and Mr Xu contributed equally. Ms Jan and Mr Xu had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Jan, Xu, Song, Stafford.

Acquisition, analysis, or interpretation of data: Jan, Xu, Luo, Xiong, Ma, Stafford.

Drafting of the manuscript: Jan, Xu, Stafford.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Jan, Xu, Luo, Stafford.

Obtained funding: Ma.

Administrative, technical, or material support: Song, Ma.

Supervision: Song, Ma.

Conflict of Interest Disclosures: None reported.

Funding/Support: The present study was supported by grant 81673192 from the National Natural Science Foundation (Dr Ma) and grant 201202010 from the Research Special Fund for Public Welfare Industry of Health (Dr Ma). Ms Jan is supported by the Chinese Government Scholarship from the China Scholarship Council for her research at Peking University.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Lisa Keay, BOptom, PhD (University of New South Wales) provided helpful comments on the manuscript. Dr Y (who prefers to remain anonymous) provided inspiration and encouragement. Wenke Liao, Wenhua Xing, and Xing Zhang gave permission to access the Chinese National Survey Students' Constitution and Health (CNSSCH) data. None of these individuals was compensated for their help. The following institutes helped with the implementation of the CNSSCH: Ministry of Education, Ministry of Health, Ministry of Science and Technology, the State Ethnic Affairs Commission, and the State Sports General Administration of the People's Republic of China. We thank all the unnamed students who participated in the surveys.

References
1.
World Health Organization. Blindness and vision impairment. Key facts. Geneva, Switzerland: World Health Organization; 2018. http://www.who.int/news-room/fact-sheets/detail/blindness-and-visual-impairment. Accessed October 19, 2018.
2.
Resnikoff  S, Pascolini  D, Mariotti  SP, Pokharel  GP.  Global magnitude of visual impairment caused by uncorrected refractive errors in 2004.  Bull World Health Organ. 2008;86(1):63-70. doi:10.2471/BLT.07.041210PubMedGoogle ScholarCrossref
3.
Holden  BA, Fricke  TR, Wilson  DA,  et al.  Global prevalence of myopia and high myopia and temporal trends from 2000 through 2050.  Ophthalmology. 2016;123(5):1036-1042. doi:10.1016/j.ophtha.2016.01.006PubMedGoogle ScholarCrossref
4.
Morgan  IG, Ohno-Matsui  K, Saw  S-M.  Myopia.  Lancet. 2012;379(9827):1739-1748. doi:10.1016/S0140-6736(12)60272-4PubMedGoogle ScholarCrossref
5.
Miao  J, Wu  X.  Urbanization, socioeconomic status and health disparity in China.  Health Place. 2016;42:87-95. doi:10.1016/j.healthplace.2016.09.008PubMedGoogle ScholarCrossref
6.
Sun  H-P, Li  A, Xu  Y, Pan  CW.  Secular trends of reduced visual acuity from 1985 to 2010 and disease burden projection for 2020 and 2030 among primary and secondary school students in China.  JAMA Ophthalmol. 2015;133(3):262-268. doi:10.1001/jamaophthalmol.2014.4899PubMedGoogle ScholarCrossref
7.
Jan  CL, Congdon  N.  Chinese national policy initiative for the management of childhood myopia.  Lancet Child Adolesc Health. 2018;2(12):845-846. doi:10.1016/S2352-4642(18)30318-3PubMedGoogle ScholarCrossref
8.
CNSSCH Association.  Report on the 2000th National Survey on Students’ Constitution and Health [in Chinese]. Beijing: China College & University Press; 2002.
9.
CNSSCH Association.  Report on the 2005th National Survey on Students’ Constitution and Health [in Chinese]. Beijing: China College & University Press; 2007.
10.
CNSSCH Association.  Report on the 2010th National Survey on Students’ Constitution and Health [in Chinese]. Beijing: China College & University Press; 2012.
11.
CNSSCH Association.  Report on the 2014th National Survey on Students’ Constitution and Health [in Chinese]. Beijing: China College & University Press; 2016.
12.
Song  Y, Wang  H-J, Ma  J, Wang  Z.  Secular trends of obesity prevalence in urban Chinese children from 1985 to 2010: gender disparity.  PLoS One. 2013;8(1):e53069. doi:10.1371/journal.pone.0053069PubMedGoogle ScholarCrossref
13.
Song  Y, Wang  HJ, Dong  B, Ma  J, Wang  Z, Agardh  A.  25-year trends in gender disparity for obesity and overweight by using WHO and IOTF definitions among Chinese school-aged children: a multiple cross-sectional study.  BMJ Open. 2016;6(9):e011904. doi:10.1136/bmjopen-2016-011904PubMedGoogle ScholarCrossref
14.
Lougheed  T.  Myopia: the evidence for environmental factors.  Environ Health Perspect. 2014;122(1):A12-A19. doi:10.1289/ehp.122-A12PubMedGoogle ScholarCrossref
15.
Colenbrander  A. Visual Standards: Aspects and Ranges of Vision Loss With Emphasis on Population Surveys. Report for the International Council of Ophthalmology. San Francisco, CA: International Council of Ophthalmology; 2002;2:1-33.
16.
Santos  CA, Fiaccone  RL, Oliveira  NF,  et al.  Estimating adjusted prevalence ratio in clustered cross-sectional epidemiological data.  BMC Med Res Methodol. 2008;8:80. doi:10.1186/1471-2288-8-80PubMedGoogle ScholarCrossref
17.
Barros  AJ, Hirakata  VN.  Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio.  BMC Med Res Methodol. 2003;3:21. doi:10.1186/1471-2288-3-21PubMedGoogle ScholarCrossref
18.
Halekoh  U, Højsgaard  S, Yan  J.  The R package geepack for generalized estimating equations.  J Stat Softw. 2006;15:1-11. doi:10.18637/jss.v015.i02Google ScholarCrossref
19.
National Bureau of Statistics. China Statistical Yearbook. http://www.stats.gov.cn/tjsj/ndsj/. Accessed January 5, 2019.
20.
Liu  J, Bray  M.  Determinants of demand for private supplementary tutoring in China: findings from a national survey.  Educ Econ. 2017;25:205-218. doi:10.1080/09645292.2016.1182623Google ScholarCrossref
21.
Ross  H, Wang  Y.  The college entrance examination in China: an overview of its social-cultural foundations, existing problems, and consequences: guest editors’ introduction.  Chin Educ Soc. 2010;43:3-10. doi:10.2753/CED1061-1932430400Google ScholarCrossref
22.
Morgan  IG, Rose  KA.  Myopia and international educational performance.  Ophthalmic Physiol Opt. 2013;33(3):329-338. doi:10.1111/opo.12040PubMedGoogle ScholarCrossref
23.
Jan  C, Li  SM, Kang  MT,  et al.  Association of visual acuity with educational outcomes: a prospective cohort study.  [published online January 18, 2019].  Br J Ophthalmol. 2019;bjophthalmol-2018-313294. doi:10.1136/bjophthalmol-2018-313294PubMedGoogle Scholar
24.
Mountjoy  E, Davies  NM, Plotnikov  D,  et al.  Education and myopia: assessing the direction of causality by mendelian randomisation.  BMJ. 2018;361:k2022. doi:10.1136/bmj.k2022PubMedGoogle ScholarCrossref
25.
Ashby  R, Ohlendorf  A, Schaeffel  F.  The effect of ambient illuminance on the development of deprivation myopia in chicks.  Invest Ophthalmol Vis Sci. 2009;50(11):5348-5354. doi:10.1167/iovs.09-3419PubMedGoogle ScholarCrossref
26.
Ashby  RS, Schaeffel  F.  The effect of bright light on lens compensation in chicks.  Invest Ophthalmol Vis Sci. 2010;51(10):5247-5253. doi:10.1167/iovs.09-4689PubMedGoogle ScholarCrossref
27.
He  M, Xiang  F, Zeng  Y,  et al.  Effect of time spent outdoors at school on the development of myopia among children in China: a randomized clinical trial.  JAMA. 2015;314(11):1142-1148. doi:10.1001/jama.2015.10803PubMedGoogle ScholarCrossref
28.
Ji  CY, Chen  TJ; Working Group on Obesity in China (WGOC).  Empirical changes in the prevalence of overweight and obesity among Chinese students from 1985 to 2010 and corresponding preventive strategies.  Biomed Environ Sci. 2013;26(1):1-12.PubMedGoogle Scholar
29.
Lu  B, Congdon  N, Liu  X,  et al.  Associations between near work, outdoor activity, and myopia among adolescent students in rural China: the Xichang Pediatric Refractive Error Study report no. 2.  Arch Ophthalmol. 2009;127(6):769-775. doi:10.1001/archophthalmol.2009.105PubMedGoogle ScholarCrossref
30.
China Ministry or Education. Number of Female Students of Schools by Type and Level. http://www.moe.gov.cn/s78/A03/moe_560/jytjsj_2017/qg/201808/t20180808_344694.html. Accessed October 10, 2018.
31.
Chen  ZT, Wang  IJ, Liao  YT, Shih  YF, Lin  LL.  Polymorphisms in steroidogenesis genes, sex steroid levels, and high myopia in the Taiwanese population.  Mol Vis. 2011;17:2297-2310.PubMedGoogle Scholar
32.
Tong  L, Saw  SM, Tan  D,  et al.  Sensitivity and specificity of visual acuity screening for refractive errors in school children.  Optom Vis Sci. 2002;79(10):650-657. doi:10.1097/00006324-200210000-00011PubMedGoogle ScholarCrossref
33.
Leone  JF, Mitchell  P, Morgan  IG, Kifley  A, Rose  KA.  Use of visual acuity to screen for significant refractive errors in adolescents: is it reliable?  Arch Ophthalmol. 2010;128(7):894-899. doi:10.1001/archophthalmol.2010.134PubMedGoogle ScholarCrossref
34.
Lin  LL, Shih  YF, Hsiao  CK, Chen  CJ.  Prevalence of myopia in Taiwanese schoolchildren: 1983 to 2000.  Ann Acad Med Singapore. 2004;33(1):27-33.PubMedGoogle Scholar
35.
Bourne  RRA, Flaxman  SR, Braithwaite  T,  et al; Vision Loss Expert Group.  Magnitude, temporal trends, and projections of the global prevalence of blindness and distance and near vision impairment: a systematic review and meta-analysis.  Lancet Glob Health. 2017;5(9):e888-e897. doi:10.1016/S2214-109X(17)30293-0PubMedGoogle ScholarCrossref
×