COVID-19 Quarantine Reveals Grade-specific Behavioral Modification of Myopia: One-Million Chinese Schoolchildren Study

Study design and participants

As part of our ongoing efforts within the MEIS project, the current survey of refraction-confirmed myopia was conducted among more than one-million schoolchildren in Wenzhou City, Zhejiang Province, China in June 2019, and followed by two examinations in December 2019 and June 2020. To facilitate this comprehensive myopia survey, we developed a semi-automated vision examination, and an information-entry pipeline, the Adolescent Myopia Information Management System (AMIMS), which involves manual inspection of automated refractometry data, automatic data import, and collaboration between clinicians and a statistician (Supplemental Figure S1).

In June of 2019, a census of 1,305 elementary and high schools encompassing different educational systems (key school and non-key school [i.e., public, sport, and martial arts]) in 11 districts of Wenzhou was conducted. There were 1,060,925 students aged 7 to 18 years recruited at the baseline, including 612,648 elementary school students (grades 1-6 in Chinese education system), 264,661 junior high school students (grades 7-9), and 183,616 senior high school students (grades 10-12). All students identified in the census were invited to complete a self-administered questionnaire regarding demographic information and their knowledge of myopia. Certified technicians were trained at the Wenzhou Medical University Affiliated Eye Hospital with respect to standard procedures for determining visual acuity (VA) and auto refraction testing. Each school in the district was equipped with an auto refractometer (GoldEye RM-9000) and electronic logarithmic visual chart (GoldEye CM-1900C) by the Wenzhou Municipal Government for assessing the degree of myopia. These resources were utilized by trained technicians to examine all participants. Among the 1,060,925 students, 18,627 students did not attend the VA and auto refraction testing. Participants with refraction values deemed nonstandard by the optometry staff were also excluded (N = 26,350), as were participants with atypical personal identity information (N = 14,119), resulting in 1,001,749 eligible participants (94.4% of the initial study population) remaining in the study (Figure 1).

• Download figure
• Open in new tab

Figure 1. Flow of participants in the current longitudinal study with 12-month follow-up.

Subsequently, two follow-up examinations were conducted in December 2019 involving 813,755 students (81.2%) and June 2020 involving 768,492 students (76.7%). It was not possible to follow the remaining participants mainly because they had graduated from high school or refused to participate.

The present study was approved by the Ethics Committee of the Wenzhou Medical University Affiliated Eye Hospital (approval numbers Wmu191204 and Wmu191205). All procedures were conducted in accordance with institutional/national research committee ethical standards as well as the 1964 Helsinki declaration. Participant completion of the self-administered questionnaire was considered informed consent.

Measurements and Outcomes

Due to the large sample size, performing the criterion standard of cycloplegic refraction to diagnose myopia was not applicable. Thus, the current survey adopted an alternative method of VA and autorefraction testing to identify myopia. VA was evaluated using an “E-type” standard logarithmic visual chart at a distance of 5 m. All students underwent noncycloplegic refraction testing using an automated refractometer, followed by complement subjective refraction testing for validation using an “E-type” standard logarithmic visual chart. Students who required corrective lenses were examined with and without eyeglasses, and naked eye refraction data were used to calculate the spherical equivalent. Myopia was identified as an uncorrected VA of less than 20/25, and spherical equivalent refraction (SER) was equaled to (sphere + [cylinder/2]) of -0.5 diopters (D) or less³³. Since analysis for right and left eyes yielded a satisfactory correlation (Pearson correlation coefficient = 0.954, Supplemental Figure S2), SER data from the right eye were arbitrarily used for assessing myopia development, as reported in previous studies ^{34, 35}. An SER of -6.0 D or less indicated as high myopia, and an SER between -0.5 D and -6.0 D defined as non-high myopia ³⁶.

From June 2019 to December 2019, the first 6-month interval indicated as “Normal” period without the influence of COVID-19 quarantine, and the second 6-month interval from January 2020 to June 2020 indicated as “COVID-19” period with the influence of COVID-19 quarantine. Since COVID-19 quarantine had a great influence on alterations of schoolchildren lifestyle and study behaviors, it provided a practical intervention model for comparing the differences of risk factors related to myopia between normal period and COVID-19 period. Earlier studies ^37-39 have used the 6-month interval to measure myopia progression for exploring the intervention strategies. In addition, during August 2020, we further used a new-designed questionnaire to survey 12,013 students, which were randomly selected from all grades. The questionnaire consisted of 17 questions, including the basic characteristics of schoolchildren, myopia information of their parents and siblings, and students’ lifestyle or learning approaches such as outdoor activities times and online course times during normal period and COVID-19 period.

Statistical Analysis

The prevalence of myopia and high myopia were calculated by all eligible participants in three examinations. For prevalence, 95% confidence intervals (CIs) were calculated according to the procedure reported by Newcombe ⁴⁰. To examine factors associated with the incidence of myopia, the Cox proportional hazard regression analysis was performed on students who attended all three examinations and did not have myopia at the baseline. Similarly, Cox proportional hazard regression analysis was conducted to reveal risk factors associated with the incidence of high myopia on students who attended all three examinations and did not have high myopia at the baseline. These multivariate Cox regression analyses were adjusted for grade, age, gender, birth month, educational system (key/non-key school), and habitation (urban/rural). The survival package in R was applied to perform multivariate Cox regression analyses.

The significance of differences between categorical variables was assessed using the Fisher’s test, and that of differences between continuous variables was assessed using Student’s t test. Two-sided P values were used in all statistical analyses. The Pearson correlation analysis was used to calculate the correlation of SER between right- and left-eye. The R software (ver. 3.6.1) was used for all statistical analyses (http://www.r-project.org).

Role of the funding source

The funders had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data and had final responsibility for the decision to submit for publication.

The prevalence of myopia and high myopia

We enrolled and collected data from a total of 1,060,925 students at the baseline (June 2019). After stringent sample filtering, 1,001,749 students (550,756 males and 450,993 females) were included in the current study (Figure 1 and Supplemental Table S1). There were 580,177 students in elementary schools, 250,839 in junior high schools, and 170,733 in senior high schools, covering almost all schools (1,305/1,336 = 97.8%) in 11 districts of Wenzhou City (Supplemental Figure S3). Of the 1,001,749 students, 813,755 (81.2%) were reexamined after the first six months (Normal period) and 768,492 (76.7%) after the second six months (COVID-19 period) (Figure 1). At the baseline, the myopia prevalence was 52.89% (95% CI, 52.79% to 52.99%), and the high myopia prevalence was 4.11% (95% CI, 4.07% to 4.15%). Table 1 shows the prevalence of myopia and high myopia among schoolchildren across three examinations categorized by gender, educational level, educational system, and habitation. The high prevalence of myopia among schoolchildren is similar with other cities of China and East Asia countries (Supplemental Table S2).

The important role of education burden in myopia development

At the baseline, cross-sectionally, the prevalence of myopia increased with grade level increasing (Table 1), and was 35.41% (95% CI, 35.29% to 35.53%) among elementary school (grades 1-6), 74.01% (95% CI, 73.84% to 74.18%) among junior high school (grades 7-9), and 81.26% (95% CI, 81.07% to 81.44%) among senior high school (grades 10-12). The high myopia prevalence was significantly higher among senior high school (12.03%; 95% CI, 11.87% to 12.18%) than that among elementary school (0.82%; 95% CI, 0.8% to 0.85%). A grade-stratified analysis indicated that myopia development could be categorized into two grade groups: Grade Group I (grades 1-6) representing myopia sensitive stage and Grade Group II (grades 7-12) representing high myopia sensitive stage (Supplemental Figure S4-S6). A high increase in myopia prevalence of 9.9% per advance in grade level was observed in Grade Group I, whereas the rate increase in myopia prevalence slowed to 2.4% per advance in grade level in Grade Group II (Supplemental Tables S3-S5). The prevalence of high myopia increased from 4.46% (95% CI: 4.32% to 4.6%) to 13.25% (95% CI: 12.95% to 13.55%) during advancement through Grade Group II (Supplemental Table S3). Interestingly, the prevalence of myopia increased in a stepwise manner by grade, as determined by sequential analyses stratified by the students’ birth month (Figure 2A, Supplemental Figure S7A-B, and Supplemental Table S6). Notably, an average increase in myopia prevalence of 8.54% (95% CI, 6.77% to 9.93%; P < 0.001) in the Grade Group I and 4.32% (95% CI, 0.69% to 7.42%; P < 0.001) in the Grade Group II was observed in students of the same age born in the August of the year compared with those born in the September (Supplemental Figures S8-S9 and Supplemental Table S6), consistent with policies regarding the age of school enrollment in China that students born in August went to school a year earlier. It provides a direct clue that grade increasing of education burden could play more effects on myopia occurrence rather than age.

• Download figure
• Open in new tab

Figure 2. The prevalence and risk factors evaluation of myopia and high myopia.

A) The prevalence of myopia and high myopia divided by schoolchildren birth month among all grades at the baseline. Each grade was divided into 12 bars according to birth month, ranging from September to August of next year according to Chinese enrollment policy. B) Forest plot shows risk factors for myopia based on multivariate Cox model. C) Forest plot shows risk factors for high myopia based on multivariate Cox model. In both Cox models, adjusted predictors included grade, age, gender, birth month, educational system, and habitation.

Consistently, by performing multivariate Cox regression analyses with adjustment, we found grade was statistically significantly associated with myopia (HR: 1.1, 95% CI, 1.08 to 1.13, P < 0.001) and high myopia (HR: 1.4, 95% CI, 1.35 to 1.46, P < 0.001), whereas age and birth month showed non-significant association with myopia and high myopia (Figure 2B-C). Furthermore, we found gender, habitation, and educational system showed a statistically significant association with myopia (Figure 2B). Compared with males, non-key school, and rural habitation separately, factors of females (HR: 1.2, 95% CI, 1.19 to 1.22, P < 0.001), key school (HR: 1.04, 95% CI, 1.03 to 1.06, P < 0.001), and urban habitation (HR: 1.11, 95% CI, 1.09 to 1.13, P < 0.001) had a higher risk to develop myopia.

The influence of COVID-19 quarantine on myopia progression

During the COVID-19 pandemic, we found students’ online time (> 2h) substantially increased (Figure 3A) and outdoor activity time (> 1h) largely decreased (Figure 3B). For students in Grade Group I, the online time increased by 3.4 times from 22.7% in normal period to 77.5% in COVID-19 period, and the outdoor activity time decreased from 41.8% in normal period to 36.6% in COVID-19 period. With regard to students in Grade Group II, the online time increased by 2.07 times from 46.9% in normal period to 96.9% in COVID-19 period, and the outdoor activity time decreased by 1.71 times from 40.1% in normal period to 23.5% in COVID-19 period (Figure 3B). These increased online time and decreased outdoor time among school-aged children may lead to occur or worsen myopia.

• Download figure
• Open in new tab

Figure 3. The behavioral modifications of schoolchildren during COVID-19 quarantine and its influence on myopia progression.

A) The differences of schoolchildren’s online time between normal and COVID-19 period categorized by Grade Group I and II. B) The differences of schoolchildren’s outdoor activity time between normal and COVID-19 period categorized by Grade Group I and II. C) The differences of grade-specific half-year myopia progression between normal and COVID-19 period.

To assess the impact of COVID-19-induced lifestyle changes on myopia development, we found myopia progression speed among all schoolchildren increased of approximate 1.5 times from -0.263 D (95% CI, -0.262 to -0.264) during normal period to -0.39 D (95% CI, -0.389 to -0.391) during COVID-19 period (P < 0.001, Supplemental Table S7). Stratified by gender, educational level, educational system, and habitation, we also observed a significant exacerbation of myopia progression during COVID-19 period (P < 0.001, Supplemental Table S7). Notably, compared with normal period, we found that the faster myopia progression of students from Grade Group I than ones from Grade Group II during COVID-19 period (Figure 3C and Supplemental Table S8).

The influence of COVID-19 quarantine on myopia incident

For all schoolchildren, we found that the half-year incidence rate of myopia increased from 8.5% in normal period to 9.87% in COVID-19 period (OR: 1.18, 95% CI, 1.15 to 1.20; Figure 4A). Stratified by female/male, key/non-key school, and urban habitation, schoolchildren had a significantly higher half-year incidence rate during COVID-19 period than that during normal period (P < 0.001, Figure 4A). COVID-19 quarantine showed no significant impact on half-year incidence rate among students from rural habitation (P = 0.62). With regard to educational level, students in Grade Group I showed a remarkably higher risk of developing myopia during COVID-19 period than normal period (OR: 1.36, 95% CI, 1.33 to 1.40; P < 0.001, Figure 4A), whereas students in Grade Group II had a prominently lower risk of incident myopia during COVID-19 period than normal period (OR: 0.45, 95% CI, 0.43 to 0.48; P < 0.001, Figure 4A). Interestingly, we found a distinct pattern that the half-year incidence rate of myopia during COVID-19 period consistently increased with grade increasing in Grade Group I, and the myopia half-year incidence rate during COVID-19 period gradually decreased with grade increasing in Grade Group II (Figure 4B and Supplemental Table S9). These results suggested COVID-19 was a risk factor for developing myopia among students in Grade Group I, whereas it was a protective factor for developing myopia among students in Grade Group II.

• Download figure
• Open in new tab

Figure 4. Significant changes in myopia incidence under the COVID-19 quarantine model.

A) Forest plot indicates the differences of myopia incidence between normal and COVID-19 period stratified by gender, educational level, educational system, and habitation. B) The differences of grade-specific half-year myopia incidence between normal and COVID-19 period. C) Forest plot indicates the differences of high myopia incidence between normal and COVID-19 period stratified by gender, educational level, educational system, and habitation. B) The differences of grade-specific half-year high myopia incidence between normal and COVID-19 period.

As for high myopia, the half-year incidence rate among all schoolchildren decreased from 1.31% during normal period to 0.096% during COVID-19 period (OR: 0.73, 95% CI, 0.70 to 0.75; Figure 4C). Among students in Grade Group I, the half-year incidence rate of high myopia during COVID-19 period was significantly higher than that during normal period (OR: 1.30, 95% CI, 1.22 to 1.39, Figure 4C), whereas there was a lower half-year incidence rate of students in Grade Group II during COVID-19 period than that during normal period (OR: 0.57, 95% CI, 0.54 to 0.59, Figure 4C). Consistent with the identified COVID-19-influenced pattern of myopia, half-year incidence rate of high myopia during COVID-19 period gradually increased with grade increasing in Grade Group I, but decreased with grade increasing in Grade Group II (Figure 4D). In contrast, the half-year incidence rate of high myopia during normal period tended to persistently increase with grade increasing from 0.063% to 3.87% (Supplemental Table S10). Our finding consistently indicated COVID-19 contributed risk to incident high myopia among students in Grade Group I and had a protective effect on developing high myopia among students in Grade Group II.

The COVID-19 quarantine reveals the crucial risk factors and uncovers the sufficiency of behavioral modifications and grade-specific control of myopia development

The ongoing COVID-19 pandemic has caused a public health emergency and global threat ²⁷ and led to the population worldwide experiencing life-altering challenges. The COVID-19 quarantine also offers an unprecedented opportunity to evaluate the effect of behavioral modifications on the incidence and/or the worsening of myopia among schoolchildren on mega scale (such as one-million schoolchildren in the MEIS study). There existed positive and negative impacts of COVID-19 pandemic on various aspects of lifestyle behaviors ^{31, 41, 42}. Hu et al. ⁴¹ reported approximately 90% of the Chinese participants had longer screen time and approximately 40% of the participants had inactive leisure-time physical exercise. Another study ³¹ on Chinese children and adolescents from five schools in Shanghai also reported that screen time remarkably increased by about 30 hours per week on average during the COVID-19 period. Consistently, we found that online time during COVID-19 period considerably increased by 3.4 times among Chinese students in Grade Group I and 2.07 times among ones in Grade Group II. Meanwhile the outdoor activity time substantially decreased among students among both Grade Groups I and II. These two factors (i.e. increased online time and decreased outdoor activity time) have been documented to enhance the risk of developing myopia ^{23, 24, 27}. In light of COVID-19 quarantine led to billions of students are simultaneously exposed to the influence of cumulative risk factors for myopia ³², it is highly valuable to establish the COVID-19 quarantine intervention model in a real world study to uncover the crucial risk factors of myopia development.

First, some early studies ^{37, 38} may show that behavioral modifications for six months can influence myopia development. However, these studies are limited by sample size and by the difficulty in implementing behavioral modification at large scale. As the result, the sufficiency of behavioral modification pattern and time remain unclear. To the best of our knowledge, the EMIS study provides the first human intervention evidence at one-million scale to demonstrate that behavioral modification for six months during COVID-19 quarantine was sufficient to alter myopia progression. This sufficiency time for behavioral modification to be effective is essential in guiding the formation of myopia development and control among schoolchildren. Additional studies are needed to evaluate and demonstrate whether the COVID-19 pandemic exert lasting effects on the progression of myopia or high myopia (i.e. this specific pattern would persist during or after the COVID-19 quarantine) and to identify the shortest time for behavioral modification (such as three months) that is effective in influencing myopia development.

Furthermore, previous studies have demonstrated that myopia onset and peak progression are rapid in childhood, particularly around puberty ⁴³. Compared with normal period, we found the COVID-19 quarantine had grade-specific effects on worsening the progression of myopia among schoolchildren, particularly among students in elementary school (Grade Group I), suggesting that elementary children were more vulnerable to be influenced by altered external environment resulting from COVID-19 quarantine. Interestingly, we uncovered a distinct pattern that COVID-19 quarantine was a risk factor of developing myopia or high myopia among students from Grade Group I, whereas COVID-19 quarantine was a protective factor for the development of myopia or high myopia among students from Grade Group II. Although COVID-19 increased the online time and decreased the outdoor activity time for all schoolchildren, COVID-19 still decreased the incidence rate of myopia and high myopia among students in Grade Group II, which may be due to the largely reduced study pressure.

Third, genetic factors play an important role in development of high myopia ^{22, 44}. As result the extent of high myopia can be modified by environmental factors such as behavioral modification in COVID-19 quarantine is largely unknown. Interestingly, we found that behavioral modification increased the risk of developing high myopia in Grade I group and decreased the risk of incident high myopia in Grade II group. This study provide the clearest data yet in support of the environmental factors contribution to high myopia, which is collaborated by previous studies ^{22, 45}.

The EMIS study provide the first human behavioral intervention data at the one million-scale to identify the grade-specific causal factors for myopia prevention and control

Within our EMIS study population, the overall prevalence of myopia and high myopia were 52.89% and 4.11%, respectively. The prevalence of myopia among students of senior high school were 81.26%, which was similar to that reported among students of high school in 2010 in Shandong Province, China (84.6%) ⁴⁶, Beijing, China (80.07%) ^{2, 47}, Singapore (74.2%) ⁴⁸, and South Korea (80.2%) ⁴⁹. The prevalence of high myopia among all myopia cases at the end of senior high school was up to 13.25%, which is consistent with other Asian countries or regions. The prevalence of high myopia in Taiwan college freshmen ⁵⁰ represented 26% of all myopia cases in 1988; by 2005, high myopia represented 40% of all myopia cases. Lin et al. ⁵¹ reported that 21% of students in Taiwan aged 18 years had high myopia in 2000, up from 10.9% in 1983. Notably, the prevalence of high myopia increased rapidly in Grade Group II of the EMIS study, indicating that although the onset of new cases of myopia slowed in high school, existing cases progressed in severity.

While several studies have evaluated and showed positive correlation between age and grade level with myopia development ^{18, 52, 53}. However, these studies are limited by sample size and lack of the detailed analysis of these risk factors, rendering the importance of these risk factors uncertain.

In our survey of over one-million students, over 5,000 participants for each birth month, enabling stratification of the results by birth month. In students of the same age, differences in myopia were observed between those born in August and September, indicating that educational grade is more closely associated with development of myopia than age. Multivariate Cox regression analyses further validated the significant association between grade and myopia as well as high myopia after adjusting for other risk factors. Thus, our MEIS study with one-million participants establishes that educational burden could play key roles in myopia and high myopia.

Based on our multivariate Cox models, we found that female students had a higher risk of developing myopia or high myopia than male students, which was consistent with previous studies reporting a positive correlation between female sex and increased myopia risk in China ^{47, 54, 55}. This difference could have a biological explanation related to hormonal differences, as postulated in a review of myopia in opposite-sex twins ⁵⁶. In addition, traditional cultural factors might influence the behaviors of Chinese girls, such as Chinese girls prefer to white skin and avoid sun exposure, and are encouraged to be gentle and quiet, which may predispose girls to stay indoors more ⁵⁷. We also observed that students from key school or urban habitation yielded a significantly higher risk of developing myopia or high myopia than those from non-key school or rural habitation. Consistently, Jan et al. ⁵⁵ reported the prevalence of visual impairment was higher among Chinese students living in urban than ones living rural areas. More economically developed regions in China, such as Zhejiang, Jiangsu, and Shanghai, had a higher risk of visual impairment than less developed regions regardless of survey year ⁵⁵.

The limitations of this study should be noted. First, the study participants were schoolchildren and did not include adults at universities and middle-aged or quinquagenarian individuals; hence our findings are limited to younger population. Furthermore, student’s online time and outdoor activity time changes during COVID-19 period were based on self-reported, thus these measurements might exist recall bias. Second, only Chinese schoolchildren were participants. In view of the ongoing COVID-19 spread worldwide, more studies are warranted to explore the influence of COVID-19 on myopia progression of schoolchildren based on more different ethnicities.

In summary, the present study is the largest longitudinal survey to assess the impact of COVID-19 quarantine-related lifestyle changes on myopia development among schoolchildren. We found an interesting COVID-19-induced pattern of developing myopia among children and adolescents, indicating different intervention strategies should be applied to control myopia among elementary and high schools. Our results should help inform educational and health policy decision making and the development of programs aimed at the prevention and control of myopia. Over the next decade, the MEIS will continue to support surveys of myopia in students twice annually, with a focus on discerning the effects of genetics, nutrition, and learning environment using questionnaire surveys and medical intervention studies to promote nationwide strategies for the prevention and control of myopia.