| Background and objectivesThe prevalence of myopia has increased over several decades at an epidemic rate in East Asia,especially in China.Myopia not only seriously affects individuals’ study,work,and the quality of life,but also greatly increases the medical cost of society,and even imposes a heavy burden on national security and stability.Myopia has become one of the most important public health problems in China and even the world.Changes in behaviors and surroundings,brought about by the technological revolution,were thought to have contributed to the rapid increase in myopia over the past two decades.Despite the increasing number of recent studies on adolescent visual health,there still have some limitations.Firstly,most of the relevant studies used artificial methods to monitor and manage adolescent vision,and lack the use and comprehensive evaluation of emerging Internet technology.Secondly,Wuhan is one of the earliest cities in China to launch the prevention and control of adolescent vision,but the survey of vision,eye care behaviors and visual surroundings of adolescents by Internet platform in this area are not known yet.Thirdly,most of the existing behavioral studies focused on the single risk factors adolescent visual health,ignoring the relevance and aggregation of the respective internal items of behavior and surroundings,and the heterogeneity subgroup of adolescent eye care behaviors and visual surroundings has not been clarified.In the end,the past researches on vision always investigate a single point in time,but visual health,eye care behaviors,surroundings of adolescents would change over time,so a longitudinal study should to launch to verify the relationship between vision,behaviors,and surroundings.In summary,the aims of this study were as follows:1.To build the "Internet+" vision monitoring platform and evaluate its applicability and economy;2.To investigate the visual health and related factors of adolescents in Wuhan based on the platform;3.To reveal the latent classes of adolescents’ eye care behaviors and visual surroundings in Wuhan;4.To explore the latent transition of eye care behaviors and visual surroundings of adolescents,and to investigate their association with the occurrence of myopia.MethodsAfter clarifying the construction of vision monitoring platform,1001 teenagers were randomly selected from 5 primary schools in accordance with the proportion of school students,and the accuracy,authenticity and economic benefit of large-scale vision monitoring of vision monitoring platform were discussed through them.At the same time,the monitoring system of the platform and the light box visual chart were used to monitor the eyesight of adolescents,and the cost data of the two monitoring methods were sorted out.The accuracy and authenticity of the visual acuity test was evaluated by comparing the results of the platform test with the artificial light box visual acuity chart.The cost and health benefit output of the two monitoring methods were calculated respectively to obtain the cost consumed by each unit result.In this part,SPSS25.0 and MedCalc19.6 were used for statistical analysis,and the hypothesis test level was set to 0.05.The vision monitoring platform was used to investigate the vision development level of adolescents in Wuhan in 2019 and 2020.5 demonstration schools and 5 nondemonstration schools were selected from 14 districts in Wuhan.4 classes were selected from each school,and about 10 students from each class were investigated.The visual condition of the students was monitored and the eye care behaviors and surroundings were investigated.In this part,descriptive analysis,analysis of variance and Chi-square test were carried out using SPSS 25.0 software.The hypothesis test level was set to 0.05 in the study.Latent Class Analysis(LCA)was used to reveal the latent classes of adolescents’eye care behaviors and surroundings.According to the conditional probability of different behaviors,the classification characteristics were described and the latent classes were named.Then,Regression Mixture Modeling(RMM)was used to analyze the effects of each factor on the distribution of potential classes.In this part,Mplus7.4 was used for statistical analysis,and the hypothesis test level was set to 0.05.The Latent Transition Model(LTM)was used to investigate the latent transformation of eye care behaviors and surroundings of adolescents.First,the transformation matrix of latent classes of visual behavior of the sample population was constructed,and then the transformation probability of latent classes of vision care behaviors was calculated.Then,a Generalized Estimation Equation(GEE)method was used to analyze the influencing factors of adolescent myopia at the individual level,including time-dependent covariables such as latent classes of eye care behaviors and surroundings,and time-independent covariables such as gender.In this part,Mplus7.4 and R4.1 was used for statistical analysis,and the hypothesis test level was set to 0.05.Results(1)The theory of chronic care model guided the present study.The "Internet+"vision monitoring platform was composed of two parts:the equipment platform included a display and a special button,and the application platform was used to record results,analyze the data,and summarize the big data.The overall monitoring accuracy rate of the vision monitoring platform was 92.61%,and no statistical differences were noted compared with control group.In terms of authenticity,the sensitivity of vision monitoring platform was 84.91%,the false negative rate was 15.09%,the specificity was 91.63%,the false positive rate was 8.37%,the Youden index was 0.765,the positive likelihood ratio was 10.15,and the negative likelihood ratio was 0.16.(2)The comparison between manual monitoring and platform monitoring showed that,with an average of 50 students in each class,the labor cost of manual monitoring was 870.5 yuan,the working period was 1 month,and the material cost was 300 yuan.The labor cost of platform monitoring was 516.5 yuan,the working period was 1 week,and the material cost was 350 yuan.In terms of working period,the workload input and capital input of "Internet+" vision monitoring management platform for each student decreased by 42.86%and 25.97%,respectively.According to the calculation of 907,042 students in Wuhan in 2018,the vision monitoring platform could save 5.5148 million yuan compared with manual monitoring.(3)A total of 6,130 students participated in the platform survey,of which 53.56%were male and 46.44%were female.The age of the study populations ranged from 6 to 17 years,with an average age of 10.33±2.602 years old.The survey results showed that in 2019,50.03%of teenagers in Wuhan had myopia,51.76%had astigmatism,and 23.90%had glasses.In 2020,the proportion of myopia,astigmatism,and glasses among adolescents in Wuhan was 52.30%,54.41%and 25.86%respectively.The adolescents in Wuhan showed decline in the level of vision-related health,while which still better than the average level of the whole country during the same period.(4)In terms of eye care behaviors,the results of the two time points showed that in "Q1:Long time spent on reading of close distance " and "Q4:Adequate sleep duration" had the highest incidence,61.48%and 64.78%at T1 and 66.36%and 67.59%at T2,respectively.The lowest incidence was "Q10:Do you exercise your eyes regularly(more than 3 times a week)by playing table tennis,badminton and other small ball games",which were 35.22%and 32.41%respectively at 2 time points.And adolescents with the bad behaviors had a relatively high rate of myopia.In the classroom surroundings survey,the acceptance rate of "Q2:Average illumination of classroom blackboard surface","Q5:Illumination uniformity of classroom desk surface" and "Q8:Height of classroom lamp tube" at two time points were lower than 40%,and the acceptance rate of "blackboard lighting fluorescent lamp configuration"and "ratio of Windows to floor area" were higher than 70%.In the survey of home visual surroundings,the number of people who selected "Q1:Learning desks and chairs match with their own height" and "Q4:Reading lamps and roof lights are turned on at the same time when learning in the evening" account for a relatively high proportion,accounting for more than 50%.The environmental states of the two time points have changed in different degrees.By comparing the incidence of myopia with the environmental items,it was found that there were significant differences in the visual health status of adolescents under some different environmental items.(5)LCA results showed that the 3-class model was acceptable to the data fitting results in terms of eye care behaviors at two time points.In 2019,the model fitting results were as follows:AIC was 36698.216,BIC was 36906.565,aBIC was 36808.056,and the Entropy was 0.838.In 2019,the model fitting results were as follows:AIC was 35964.184,BIC was 36172.534,aBIC was 36074.024,and Entropy was 0.907.According to the conditional probability of 3-class model,adolescent eye care behaviors could be divided into the bad behavior class,the moderate behavior class and the healthy behavior class.RMM results showed that there were significant differences among the three classes in the basic demographic characteristics of gender,age,education stage,school type and residence.At the same time,there were significant statistical differences in myopia and mirror wearing state of vision development.In comparison with the healthy behavior class,the proportion of adolescents with high school,non-demonstration school,astigmatism was higher in the bad behavior class.Women,15～17 years old,people without myopia were more likely to be in the moderate behavior class.The LCA results of the visual surroundings showed that the 2-class model was acceptable to the data fitting results.At T1,AIC was 80219.469,BIC was 80593.399,aBIC was 80358.444,and Entropy was 0.776.T2 time point fitting results showed that AIC was 80901.851,BIC was 80975.782,aBIC was 80940.827,and Entropy was 0.843.According to the characteristics of conditional probability of each class in the 2-class model,the visual surroundings of adolescents could be divided into the general surroundings and the good surroundings.Compared with the general surroundings,teenagers from demonstration schools were more likely to be distributed in the good surroundings.(6)LTA results showed that the probability of the healthy behavior class remaining in the original state was the highest(78.47%).The second was the bad behavior class,with a probability of 45.78%.The moderate behavior class was the most unstable class,with only 32.71%of the cases remaining in the original state.In terms of changing direction,the probability of the bad behavior class changing to the moderate behavior class and the healthy behavior class was 29.10%and 25.12%,respectively.The probability of the healthy behavior class changing to the moderate behavior class and the bad behavior class was 10.62%and 10.91%,respectively.The probability of adolescents’ distribution in the bad behavior class,the moderate behavior class and the healthy behavior class was 23.76%,19.27%and 57.00%,respectively.In terms of the latent transition of visual surroundings,the state of the general surroundings class was relatively unstable,and the probability of remaining in the original class was 78.34%.In terms of the direction of change,the probability of the general surroundings to the good surroundings was 21.66%,while the probability of the good surroundings to the general surroundings was 11.03%.Combined with the initial class and transition probability,the results showed that after 18 iterations,the distribution probability of the general surroundings and the good surroundings were 51.78%.(7)GEE results showed that among time-independent covariates,gender,school type and initial education stage had statistically significant influence on the occurrence of myopia in adolescents.And in the time-dependent covariate,the class of adolescent eye care behaviors and visual surroundings were the influencing factors of adolescent myopia.In the different adolescents,the latent class of the bad behavior was the risk factor of myopia,and the latent class of the healthy behavior was the protective factor of myopia.The specific result was that the regression coefficient of the bad behavior class was 0.197,and the OR was 1.218,that is,the bad behavior class was 1.218 times the possibility of developing myopia of the other classes(P=0.033).The regression coefficient of the healthy behavior class was-0.147,and the OR indicated that the healthy behavior class was 0.863 times the possibility of developing myopia of the other classes(P=0.013).The good surroundings class was the protective factor of myopia.The regression coefficient of the good surroundings class was-0.109,indicating that the good surroundings class was 0.897 times of possibility of developing myopia of the general surroundings class(P=0.049).From the individual level,the possibility of myopia increased to 1.218 times of the original when a student was transferred from other classes into the bad behavior class.The possibility of myopia decreased to 0.863 times of the original when the adolescent was transferred from other classes to the healthy behavior class.And he possibility of myopia decreased to 0.897 times of the original when the visual surroundings of adolescents changed from the general class to the good classConclusionsThis study first constructed and scientifically evaluated the vision monitoring platform,then investigated the development status of vision,behaviors and surroundings of adolescents in Wuhan in 2019 and 2020 based on the platform,and finally explored the correlation between adolescents’ latent behavior,environmental development and the occurrence of myopia.The evaluation results showed that the platform had high accuracy and good authenticity,and was a reliable technical basis for group vision monitoring.The results of the survey on visual health,behavioral and environmental indicators showed that the prevention and control of myopia in Wuhan has achieved remarkable results in recent years.The latent analysis showed that there were heterogenous subgroups in both adolescent behavior and environment,and the conversion of different behavior and environment categories was associated with the occurrence of myopia.In conclusion,the visual health management of adolescents should be based on advanced monitoring technology and on the premise of identifying the health status of the population,fully understand the heterogeneous subgroups of behaviors and visual surroundings,and carry out long-term,continuous,and dynamic management. |
PDF Full Text Request