Effects of online family health education on myopia prevention in children by parental myopia: a randomized clinical trial

ABSTRACT Clinical relevance Online family health education may be effective for myopia prevention in children, and the effects may be different between children with myopic and non-myopic parents. Background Myopia is a common cause of vision loss. The aims of this study were to evaluate the effects of online family health education on preventing the development of myopia in children, and to estimate whether the effects vary according to parental myopia. Methods A cluster randomised trial including grade 1 and grade 2 children from 12 primary schools was conducted in Guangzhou, China. Weekly online family health education messages were sent to parents in the intervention group. Data collection included eye examinations of children and questionnaires completed by parents. Results Among the 3123 children included at baseline (1703 boys [54.5%]; mean [SD] age, 6.83 [0.73] years), 2376 completed the follow-up after 3 years. The differences in the incidence of myopia and myopic shift between the study groups were not significant in total. However, the 3-year cumulative incidence rate of myopia in the intervention group (125 of 445 [28.1%]) was significantly lower than that in the control group (225 of 603 [37.3%]; difference, 9.2% [95% CI, −14.9% to −3.5%]; P = 0.001) among children with non-myopic parents. In parallel, among children with non-myopic parents, the mean myopic change in SER was less for the intervention group than for the control group (−1.10 D vs. −1.24 D; difference, 0.13 D [95% CI, 0.03 to 0.23 D]; P = 0.01). Conclusions Compared with children with myopic parents, online family health education was more effective in children with non-myopic parents. The incidence of myopia and myopic shift in refraction have been reduced in children with non-myopic parents. Further studies are needed to assess these differences by parental myopia.


Introduction
Myopia is a common ocular abnormality worldwide, and it was predicted that 49.8% of the world population would be myopic by 2050. 1,2The increasing prevalence of myopia has become an important public health issue in East and Southeast Asia, affecting 80% to 90% of young adults in many communities, and accompanied by 10% to 20% of young adults with high myopia. 3,4Considering the association of high myopia with blinding myopic pathologies, myopia prevention has become an important public health priority, especially in Asia. 4,5ehavioral interventions have been widely used in the prevention of myopia in children. 6,7][10] Most of the current measures to increase outdoor time in children are schoolbased implementation, but outdoor activities beyond school hours are equally important for myopia prevention.
However, strategies to increase outdoors time beyond school hours have been hard to implement, because parents are concerned that increased outdoors time may affect time for studies. 11Therefore, health education for parents to raise parental awareness could be an effective way to control myopia in children.Especially during the COVID-19 outbreak, outdoor activities of children were constrained by a severe lockdown for several months in China, which required them to learn online at home. 12This makes it more necessary to educate parents on myopia prevention in children.In recent years, internet and social media are used frequently in health management. 13,14nline family health education has been initially applied to myopia prevention in children.Our previous study indicated that the 2-year intervention of online family health education could reduce myopia incidence among 1525 grade 1 students. 15Recently, a randomised clinical trial among 268 grade 2 students shown that sending text messages to parents to take their children outdoors for 1 year resulted in lower myopia progression in schoolchildren over 3 years. 16However, the effects in larger population and longer follow-up duration need to be further determined.
Parental myopia was related to parental attitudes and behaviours towards myopia prevention in children, and there was a trend for myopic parents to set rules for visionrelated health behaviours of their children more than nonmyopic parents. 17,18Furthermore, a consistent finding is that children with myopic parents were at a higher risk of myopia, as myopic parents could create more myopia genic environments for their children. 5,19ecause of these behavioural and genetic differences, the effects of online family health education for myopia prevention in children may be different between myopic and nonmyopic parents.The purpose of this study was to evaluate the effects of online family health education on preventing the development of myopia in children, and to estimate whether the effects vary according to parental myopia.

Study design
This school-based, cluster-randomised trial was conducted in Guangzhou, China, since 2018, with three years of follow-up so far, to evaluate the effects of online family health education on preventing the development of myopia. 15welve schools were randomly selected from 2 districts (Panyu and Huadu districts) with similar socioeconomic levels in Guangzhou, and then randomly assigned to the intervention and control group in each district.Finally, there were 6 schools in the intervention group (3 from district 1 and 3 from district 2) and 6 schools in the control group.
All grade 1 and grade 2 children from the selected schools were invited to participate in this study, with a response rate of 76.3% (3433/4502).Measurements on visual status of children performed by ophthalmologists who were blinded to intervention conditions, and questionnaires completed by parents were collected at baseline and follow-up.
Details of the study design and follow-up are shown in Figure 1.Of 3123 children at baseline (1498 in the intervention group and 1625 in the control group), 2376 (76.1%) completed the final assessment after the 3-year follow-up (1138 [76.0%] in the intervention group and 1238 [76.2%] in the control group).Baseline myopia prevalence, axial length, and the proportion of both parents with college degree in children who completed the follow-up was lower than those who were lost to follow-up.There were no other statistical differences between the 2376 participants who attended follow-up assessments and the 747 who missed follow-up visits (eTable 1 in Supplement 1).
For all participants, parental consent was obtained prior to baseline examination.This study was approved by Sun Yatsen University, School of Public Health Institutional Review Board.The trial registration is available at http://www.chictr.org.cn (identifier, ChiCTR1900022236).The protocol is shown in Supplement 2.

Intervention and supervision of implementation
The intervention was that each class teacher sent health education messages compiled by the researchers to parents via WeChat group every week.After that, teachers fed back the screenshots of the messages forwarded by each class to the researchers to ensure the completion of the intervention.Health education messages mainly included increasing outdoor activities, cultivating good eye habits, and limiting screen time.The flowchart of intervention is shown in Figure 2.
In China, because the class WeChat group is used by teachers to communicate with parents about performance of their children at school, assign daily homework, etc., parents pay more attention to this WeChat group, which ensures that parents can check health education messages on time.

Measurements
Eye examinations were performed in schools by a group of trained optometrists and ophthalmologists.Axial length was measured by optometrists and ophthalmologists using an IOL Master biometer (Carl Zeiss Meditec).Cycloplegic autorefraction was performed by optometrists and ophthalmologists using a autorefractor (Topcon 8800K).Cycloplegia was induced with 3 drops of 1% cyclopentolate hydrochloride administered at 0, 5, and 20 minutes to both eyes.
Cycloplegia and pupil light reflex were checked after 15 minutes to determine whether complete cycloplegia (pupil dilated ≥6 mm and pupil light reflex was absent) was achieved.All measurements were performed 3 times in each eye, and the mean value was calculated for each eye.The follow-up eye examinations were the same as those at baseline and were performed by the same group of ophthalmologists and equipment.
The electronic questionnaire included basic information about children and parents, parental attitudes towards myopia prevention and vision-related health behaviours of children, which was answered by parents (or guardians).Academic performance was measured by parents based on the ranking of the overall examination results.The teachers reminded parents with children participating in eye examinations to complete the questionnaire.The questionnaire is shown in Supplement 3.
Eye examinations and questionnaires were performed annually in November and December at both baseline and follow-up.In the third follow-up, examinations in 2 schools (1 in the intervention group and 1 in the control group) were delayed until March 2022 because of COVID-19.

Outcomes
The primary outcome of this study was the 3-year cumulative incidence rate of myopia.Myopia was defined as a spherical equivalent refractive error (sphere +0.50D cylinder) of at least −0.50D in either eye. 20Only children without myopia from baseline cycloplegic refraction data were included in the analysis for the 3-year incidence rate of myopia.
Secondary outcomes were the 3-year change of spherical equivalent refraction (SER) and the elongation of axial length (AL) at 3-year follow-up.Because the refraction and biometry of the right and left eyes were highly correlated, data from the right eye were used for analysis.

Statistical analysis
The sample size was calculated according to a formula commonly used for counting material in experimental epidemiological study designs. 21The 3-year cumulative incidence of myopia among primary school students in the control group was estimated to be 39.5%, 9 the expected reduction in the myopia incidence was set at 6% during the 3-year study.A 2-sided α of 0.05 and a power level of 90% were assumed.
The assumption of a 3-year loss to follow-up of less than 20% and a baseline participation rate of 80% led to a total sample size of 4221 children.Therefore, this study consisted of students from 12 primary schools, considering that each public primary school in Guangzhou had a mean of 350 children in grade 1 and grade 2. 22 The data were presented as mean ± standard deviations (SD) for continuous variables and percentages for categorical variables.In the descriptive analysis of baseline characteristics, the χ 2 tests and the independent sample t-tests were used to compared characteristics between different groups.For the follow-up data, distributions of parental attitudes were compared between the 2 groups using χ 2 tests, and covariance analysis was conducted to compare the differences of eye behaviours of children time between the 2 groups and adjusted the baseline values.
The cumulative incidence of myopia in children was compared between the intervention and control groups using Logistic regression, and adjusted for age, sex, sleep time, television time, computer time, and outdoor time at baseline.Cumulative variances in SER and AL were analysed by the linear mixed model to correct for the random effects of cluster design and adjusted for age, sex, and the baseline value.All statistical analyses were performed using SPSS, version 25.0 (IBM Corporation).Two-tailed P values less than 0.05 were considered statistically significant.

Effects of online family health education on myopia prevention in children in total
As shown in Table 1, a total of 3123 children in 12 schools (1498 in the intervention group and 1625 in the control group) were included at baseline.Overall, the mean (SD) age was 6.83 (0.73), and 1703 (54.5%) were boys.At baseline, 184 students (5.9%) were myopic.The mean (SD) age of the children was 6.86 (0.71) years in the intervention group and 6.81 (0.74) years in the control schools (P = 0.046).
The proportion of myopic parents was higher in the intervention group than the control group (59.3% vs. 50.0%;P < 0.001), and the proportion of both parents with college degree was higher in the intervention group than in the control group (45.0% vs. 29.2%;P < 0.001).Children in the intervention group spent less time watching television (0.51 h/d vs. 0.56 h/d; P = 0.01).Other characteristics of the students in the intervention and control groups at baseline were similar.
After the 3-year intervention, a higher proportion of parents in the intervention group compared with the control group limited television time of their children (83.6 vs. 80.1%;P = 0.03) and computer time of their children (86.6  vs. 82.2%;P = 0.003), and arranged family outdoor activities more than 2 times per week (50.7% vs. 45.7%,P = 0.02).There was no difference in computer time and outdoor time of children between the intervention and control groups (eTable 2 in Supplement 1).
Table 2 shows the outcome measures in the 2 study groups at the 3-year follow-up.The 3-year cumulative incidence rate of myopia was 41.5% (449 of 1081) in the intervention group, and 42.6% (506 of 1187) in the control group.The mean myopic change in SER was −1.33 (95% CI, −1.38 to −1.28) D in the intervention group, and −1.35 (95% CI, −1.40 to −1.30) D in the control group.The cumulative axial elongation was 0.87 (95% CI, 0.84 to 0.89) mm in the intervention group, and 0.85 (95% CI, 0.82 to 0.87) mm in the control group.These differences in outcomes between the study groups were not significant.

Effects of online family health education on myopia prevention in children, stratified by parental myopia
The baseline characteristics of children in the intervention and control groups by parental myopia were the same as the total (eTable 3 in Supplement 1).The baseline characteristics of children with myopic and non-myopic parents are shown in eTable 4 in Supplement 1.A higher proportion of children with myopic parents had above-average academic performance than children with non-myopic parents (47.6 vs. 42.4%;P = 0.01).The proportion of both parents with college degree was higher in children with myopic parents than nonmyopic parents (51.9% vs. 18.7%;P < 0.001).Children with myopic parents had more myopic refractive error than children of non-myopic parents (P < 0.001).
Parental myopia is associated with parental attitudes and behaviours towards myopia prevention and vision-related health behaviours of children.At baseline, a higher proportion of myopic parents compared with non-myopic parents limited time spent watching television of their children (83.9% vs. 79.1%;P = 0.001) and using computer (83.8% vs. 79.3%;P = 0.001).Children of myopic parents spend less time watching television than children of non-myopic parents (0.51 h/d vs. 0.57 h/d; P = 0.002).In terms of outdoor activities, there was no difference between myopic and non-myopic parents.
Table 3 shows parental attitudes and behaviours of children at the 3-year follow-up for the two groups by parental myopia.Among non-myopic parents, a higher proportion of parents in the intervention group compared with the control group arranged family outdoor activities more than 2 times per week (51.9% vs. 45.8%;P = 0.046).Children in the intervention group spent more time outdoors compared with the control group among children with non-myopic parents (0.87 h/d vs. 0.80 h/d; P = 0.04).
Among myopic parents, a higher proportion of parents in the intervention group compared with the control group limited time spent watching TV and using computer of their children.However, there was no difference between the study groups in parental attitudes towards outdoor activities and outdoor time of children among myopic parents.
The outcome measures in the 2 groups by parental myopia at the 3-year follow-up appear in Table 4.The 3-year cumulative incidence rate of myopia in the intervention group (125 of 445 [28.1%]) was significantly lower than that in the control group (225 of 603 [37.3%]; difference, 9.2% [95% CI, −14.9% to −3.5%]; P = 0.001) among children with non-myopic parents.Among children with non-myopic parents, the mean myopic change in SER was less for the intervention group than for the control group (−1.10 D vs. −1.24D; difference, 0.13 [95% CI, 0.03 to 0.23] D; P = 0.01), but cumulative axial elongation was not significantly different.The differences in outcomes between the study groups were not significant among children with myopic parents.Expressed as a percentage and calculated using simple asymptotic method.§ Calculated as follow-up value minus the baseline value, and adjusted for age, sex, the baseline value, and the random effects of cluster design by using line mixed models.

Discussion
In this 3-year school-based randomised clinical trial, the online health education led to changes in attitudes and behaviours towards myopia prevention in children, which further influenced vision-related health behaviours of children with non-myopic parents; however, the effects were not significant in children with myopic parents.This finding was consistent with previous studies that demonstrated the association between family health education and behavioural change of children in different health topics. 23,24The above findings supported that increasing public education to raise parental awareness could prevent myopia in children, although its effectiveness was limited. 11It may provide a scientific basis for future myopia prevention and control strategies.
After the 3-year intervention, the online family health education has resulted in a 9.2% reduction in the incidence of myopia and myopic shift in refraction by 0.13D among primary school children with non-myopic parents.Nonmyopic parents in the intervention group had more positive attitudes towards outdoor activities of their children than the control group, and children in the intervention group spent more time outdoors compared with the control group among children with non-myopic parents.The online family health education in the intervention group appears to have no effect on parental attitudes towards screen time of children among non-myopic parents.
These findings provided further evidence that the most effective way to control myopia is to increase outdoor activity.A cluster randomised trial of children in grade 1 in Guangzhou reported that the addition of 40 minutes of outdoor activity at school could reduce the incidence of myopia by 9.1% and myopic change in SER by 0.17D over the next 3 years, which was similar with this study. 9A 2013 trial conducted in Taiwan by Wu et al. reported that 80 minutes of outdoor activities during class recess in school reduced the incidence of myopia in children aged 7 to 11 years by 9.24% and myopic shift by 0.13D after 1 year. 25Another trial in grade 1 schoolchildren conducted by Wu et al. in 2018 showed that children who spent 200 minutes or more time outdoors every week reduced the incidence of myopia by 2.93% and myopic change in SER by 0.12D per year. 8ompared to the study in Taiwan, the reduced incidence and progression of myopia in this study was relatively small.One possible explanation is that family health education increases outdoor time of children less than school-mandated measures to increase the outdoor activities of children.This suggests that school-based measures to increase outdoor activity have a greater impact on myopia prevention in children.
Another possible explanation is that the COVID-19 pandemic led to a nationwide school closure in China from January to May 2020, and Chinese children were required to learn online at home during this period, which may have an impact on the eyesight of students. 268][29] In this unfavourable visual environment during the COVID-19 pandemic, children with non-myopic parent in the intervention group spent more time outdoors than the control group, which proved the effectiveness of online home health education.This study suggests that the key to preventing myopia in children is to increase outdoor time.However, differences in the incidence of myopia and myopic shift between the study groups were not significant in total, mainly because there were more myopic parents in the intervention group than in the control group.The differences in the incidence of myopia and myopic shift in refraction between the study groups were not significant among children with myopic parents.This was mainly because there was no significant improvement in parental attitudes towards outdoor activities of children among myopic parents, and no further difference in outdoor time for children with myopic parents.
One possible explanation is that, faced with the double pressure of the education of children and visual health concerns of their children, more myopic parents choose education of their children.These parents are reluctant to sacrifice study time of their children to increase outdoor time.The results of this study showed that a higher proportion of children with myopic parents had above-average academic performance than children with non-myopic parents.
Previous studies have shown that myopic parents tend to make their children spend more time reading and less time outdoors. 30,31Therefore, the improvement in parental attitudes towards outdoor activities was not significant among parents with myopia in this study, and further increases in outdoor time and reduction in myopia incidence were also not present in children with myopic parents.
In addition to parental attitudes towards outdoor activities, another reason for the absence of differences in the Expressed as a percentage and calculated using simple asymptotic method.§ Calculated as follow-up value minus the baseline value, and adjusted for age, sex, the baseline value, and the random effects of cluster design by using line mixed models.
incidence of myopia and myopic shift in refraction between study groups among myopic parents is that myopic parents have paid more attention to their health behaviours related to screen time of children at baseline than non-myopic parents.The results of this study showed that a higher proportion of myopic parents compared with non-myopic parents limited screen time of their children.A cross-sectional study reported similar results, with myopic parents viewing myopia as more of an optical inconvenience, an expense and a cosmetic inconvenience, and limiting screen time use in their household more than non-myopic parents. 17nother study also showed that myopic parents had more knowledge on myopia prevention in children. 18For this reason, there may be less room for attitudinal improvement in families with myopic parents.In addition, a genetic contribution to myopia has been supported by many studies, and parental myopia is an important risk factor for myopia in children. 19,32,33nvironmental factors play a heavier role in families without myopic parents or with mildly myopic parents, than in families with highly myopic parents. 34Whether the effects of online family health education on myopia prevention in children can resist the genetic effects remains to be further determined.
This study had the strength of its design as a randomised clinical trial, long follow-up duration, and outcomes based on cycloplegic refraction.However, this study has some limitations.First, some students were lost to follow-up because they did not complete the refraction examination or questionnaire, or transferred to another school; however, the proportion of children completing follow-up in the intervention and control groups was balanced (1138 [76.0%] in the intervention group and 1238 [76.2%] in the control group).
The characteristics between children who attended followup and those who missed it were compared (eTable 1 in Supplement 1).Although baseline myopia prevalence and AL were lower in children who completed follow-up than in children who did not, there was no difference in SER.There were no other statistical differences between children who attended follow-up and those who missed it.
Second, intervention information was forwarded to parents by teachers, in which teachers are also educated, but the role of teachers in behaviour change of children was not assessed.Third, the control group may receive news and articles related to myopia prevention on the internet, which cannot be strictly controlled.However, both intervention and control group parents were likely to receive such information in their daily lives, which was considered to be balanced between the two groups by randomisation.The intervention group additionally received weekly health education messages.
Finally, the results showed that the effects of intervention on parental attitude towards outdoor activities and outdoor time of children were modest, but the effects on outcomes were larger.It is possible that parents who did not have attitude changes may have still increased outdoors time of their children.Therefore, it is important to measure outdoor time of children objectively.However, the estimates of outdoor time in this study came from questionnaires.Further research is needed to accurately assess the time.

Figure 1 .
Figure 1.Flowchart of the online family health education study design.

Figure 2 .
Figure 2. Flowchart of intervention for the online family health education study.

Table 1 .
Baseline characteristics of children in the intervention and control groups.

related health behaviours of children at baseline, mean (SD), h/d
†Calculated using the χ 2 tests for qualitative variables, and the independent sample t-tests for quantitative variables.‡ P = 0.046.

Table 2 .
Refractive and biometric outcomes at 3-year follow-up.SER, spherical equivalent refraction; AL, axial length.† Calculated using the Logistic regression for qualitative variables (adjusted for age, sex, sleep time, television time, computer time, and outdoor time at baseline), and line mixed models for quantitative variables.‡

Table 3 .
Parental attitudes towards behaviours related to myopia prevention of children and vision-related health behaviours of children at 3-year follow-up by parental myopia.
†Calculated using the χ2 tests for qualitative variables, and the covariance analysis adjusted the baseline value for quantitative variables.‡ P = 0.046.

Table 4 .
Refractive and biometric outcomes at 3-year follow-up by parental myopia.Calculated using the Logistic regression for qualitative variables (adjusted for age, sex, sleep time, television time, computer time, and outdoor time at baseline), and line mixed models for quantitative variables.‡