Abstract
Objective
To examine the association between weekday sleep duration and visual health among adolescents in low-resource rural settings, and to explore potential behavioral mechanisms.
Methods
We used cross-sectional data from 36,139 rural primary and junior high school students in northwestern China (2012, 2019). Visual acuity was assessed via standardized LogMAR tests. To address potential endogeneity in self-reported sleep, we applied two-stage least squares regression with sunset time as an instrumental variable.
Results
Students averaged 8.65 weekday sleep hours, and 58.3 % slept fewer than 9 h—below age-specific recommendations. Instrumental variable estimates indicated that each additional weekday sleep hour reduced the LogMAR score by 0.070 (p < 0.05), indicating better vision. Mechanism analysis suggested that shorter sleep was linked to greater late-night screen use and near work.
Conclusions
Insufficient weekday sleep is associated with poorer vision among rural adolescents, potentially mediated by digital screen exposure. School-based programs that promote adequate sleep and reduce nighttime screen use may offer a low-cost, scalable approach to support visual health in under-resourced areas.
Keywords: Sleep duration, Visual health, Adolescents, Myopia, Rural China
Highlights
-
•
Longer sleep is linked to fewer vision problems in rural adolescents.
-
•
Sunset time is used as an instrument to predict sleep duration.
-
•
Associations are stronger among students at higher myopia risk.
-
•
Adequate sleep may help reduce eye strain and near work.
-
•
Improving sleep habits may help reduce vision disparities in low-resource settings.
1. Introduction
Sleep is a foundational component of adolescent health and development. Yet, chronic sleep deprivation has become increasingly prevalent among children and adolescents worldwide, especially in countries with competitive education systems such as China, South Korea, and Japan (Rhie et al., 2011; Olds et al., 2010; Li et al., 2023). While a growing body of literature has documented the cognitive and psychological consequences of insufficient sleep—ranging from impaired attention to emotional dysregulation (Beebe et al., 2017; Sadeh et al., 2003; Lufi et al., 2011)—less is known about its implications for physical health outcomes, particularly vision.
Visual health, especially myopia, has emerged as a significant public health concern in East Asia, where myopia rates among school-aged children are among the highest in the world (Guan et al., 2023; Alvarez-Peregrina et al., 2021; Harris and Lord, 2016). A number of studies have explored environmental and behavioral risk factors such as near-work activities, screen time, and limited outdoor exposure. Recently, sleep duration has attracted attention as a potential determinant of myopia (Xu et al., 2019; Jee et al., 2016; Liu et al., 2022), yet empirical evidence remains inconclusive. Some studies suggest that shorter sleep duration may disrupt ocular growth regulation and increase the risk of myopia, while others find weak or no association (Zhou et al., 2015; Lu et al., 2021; Wei et al., 2020; Liu et al., 2020; Flitcroft et al., 2019).
Existing research on the sleep–vision relationship is limited in three important ways. First, many studies rely on small urban samples, overlooking vulnerable rural populations where sleep deprivation and access to eye care may be more severe. Second, prior work often fails to account for the endogeneity of sleep behavior, namely, unobserved confounders or reverse causality. Third, little is known about whether the relationship between sleep and vision varies across sex or educational stages, limiting the applicability of findings for targeted intervention design.
To address these limitations, this study aims to provide new evidence on the relationship between sleep and visual health among adolescents in rural China. Drawing on a large dataset of 36,762 primary and junior high school students and employing an instrumental variables (IV) strategy to address endogeneity, we pursue four specific objectives: First, we provide a descriptive overview of sleep patterns and vision outcomes in the study population, establishing the empirical background for further analysis. Second, we estimate the association between weekday sleep duration and visual health outcomes using both ordinary least squares (OLS) and instrumental variable (IV) methods, addressing potential endogeneity in sleep behavior. Third, we explore a potential behavioral mechanism linking sleep to vision—namely, screen exposure—by investigating how sleep duration correlates with time spent on electronic devices. Together, these analyses offer a more comprehensive understanding of how and for whom sleep matters in preventing adolescent visual impairment in low-resource settings.
2. Methods
2.1. Study setting
This is a cross-sectional observational study conducted among rural Chinese adolescents. The data for this study were collected from two large-scale surveys conducted among rural students in northwest China. The first survey was conducted in the fall of 2019 in Ningxia Province, and the second in the fall of 2012 in Shaanxi and Gansu Provinces. All three provinces are located in underdeveloped regions of China and exhibit significant socioeconomic disadvantages. For instance, Gansu is the second-poorest province in China, while Ningxia's per capita GDP ranked 28th out of 31 provinces in 2019. Despite improvements in educational access—with gross enrollment rates above 99 % in primary and junior high schools—educational inequality and resource shortages persist, especially in rural areas. These regions also experience relatively short daylight hours in autumn, which may affect natural circadian rhythms and sleep patterns.
2.2. Sample selection
A total of 377 rural schools were sampled across the three provinces using a multi-stage random sampling strategy. In Ningxia, we began with a list of 273 junior secondary schools and excluded urban schools and those with fewer than 40 seventh-grade students. From the remaining schools, one or two classes per school were randomly selected based on logistical feasibility. In Shaanxi and Gansu, we first selected 18 populous counties. Then we sampled 435 rural primary schools based on enrollment size (50–150 students in grades four and five), randomly choosing one class per grade in each selected school. The initial sample included 39,183 students. After data cleaning and listwise deletion of missing values, the final sample included 36,762 students. The inclusion and exclusion process, as well as the derivation of the final analytical sample, is illustrated in Fig. A1 in the Appendix.
2.3. Data collection and measures
2.3.1. Sleep duration
Students reported their usual wake-up and bedtimes on weekdays. Sleep duration was calculated as the time interval between these two responses, reported in hours and minutes. This measure reflects habitual weekday sleep duration, a key modifiable behavior with known physiological implications. Sleep duration was calculated in hours (e.g., 8.5 for 8 h and 30 min).
We focused on weekday sleep duration to capture sleep patterns more directly linked to school demands and routines, while weekend sleep duration was excluded due to potential confounding by varied social and recreational activities.
2.3.2. Visual health
Visual acuity (VA) was measured using a standardized Early Treatment Diabetic Retinopathy Study (ETDRS) chart at a distance of 4 m. The test protocol involved two trained enumerators to ensure accuracy and consistency. Each student was tested on both eyes, and VA was recorded as the smallest line on which the student correctly identified at least 4 out of 5 optotypes. The VA of the better eye was used in the analysis, as it is a validated indicator of overall visual health. For statistical modeling, VA was converted to logarithm of the minimum angle of resolution (LogMAR) using the formula LogMAR = log10 (1/VA). A higher LogMAR value indicates worse visual acuity. In robustness checks, we additionally constructed a binary indicator for myopia status based on measured visual acuity in the better-seeing eye (LogMAR >0.3), consistent with national diagnostic criteria.
2.3.3. Covariates
We controlled for a range of student- and household-level covariates. Student characteristics included age, sex, ethnicity (Han or ethnic minority), only-child status, boarding status, and grade level (Grades 4, 5, 7, or 8). Time-use behaviors encompassed daily nap duration, phone screen time, and outdoor activity during lunch break, all measured in minutes.
Parental background variables included whether either parent had attained at least senior high school education, whether either parent had migrated for work, and the age of both parents. Household socioeconomic status was assessed using a composite asset index based on a 12-item checklist adapted from the National Household Income and Expenditure Survey (Wang et al., 2022; Ma et al., 2021; Li et al., 2024). We also included a binary indicator for prior participation in an eyeglass distribution program to control for baseline vision correction. All covariates were included in both the main and instrumental variable regression models.
To further explore potential behavioral mechanisms linking sleep duration and visual health, we also collected self-reported time (in minutes) spent on four categories of activities on a typical weekday: (1) using mobile phones (e.g., messaging, gaming, or browsing); (2) watching television; (3) using computers; and (4) participating in outdoor sports or play. These variables were used to examine how sleep patterns might substitute or interact with daily screen exposure and physical activity, which are known correlates of visual development.
The overall proportion of missing data in our dataset was low, with the highest missing rate across all variables being 2.5 %. For demographic variables like parental age and education, we used logical imputation where possible. The remaining missing values were addressed through listwise deletion, given the low missingness. This approach ensured that our analyses were conducted on a consistent sample while minimizing potential biases due to missing information (Fox-Wasylyshyn and El-Masri, 2005).
2.3.4. Statistical analysis
We structured our empirical analysis to address four primary research aims. First, we presented descriptive statistics to summarize the distribution of weekday sleep duration and LogMAR visual acuity across the full sample, as well as by sex and school level.
Second, to examine the association between weekday sleep duration and vision health, we estimated linear regression models with robust standard errors clustered at the school level. All models adjusted for student- and household-level covariates and included county fixed effects. The baseline model was specified as:
| (1) |
where Yi is the LogMAR score for student i, Si is the weekday sleep duration, Xi is a vector of covariates, COUNTYi represents county fixed effects, and εi is the error term.
Given concerns about endogeneity arising from omitted variable bias or reverse causality, we employed a two-stage least squares (2SLS) instrumental variable (IV) approach. Following prior research on circadian rhythms, we used average sunset time at the township level as an instrument for sleep duration. The instrument was the average sunset time at the township level, drawn from a national astronomical database (https://richurimo.bmcx.com/), matched by survey month and location. Sunset time affects biological sleep timing but is plausibly unrelated to vision outcomes except through its impact on sleep. All IV models included the same covariates and county fixed effects.
Third, we explored behavioral mechanisms that may link short sleep duration to vision problems. Specifically, we examined whether shorter sleep was associated with increased screen exposure and near work activities during evening hours. These models used the same covariate and fixed-effects specifications as the main regressions.
All analyses were conducted using Stata 18. Additional robustness checks and first-stage regression diagnostics are provided in the Supplementary appendix.
2.4. Ethical approval and consent to participate
Research ethical approval for this study was obtained from the Institutional Review Boards (IRBs) of Stanford University (Approval number: 52514) for the research conducted in Ningxia, and from Stanford University Stanford University (Approval number: 24847), and Sun Yat-sen University (Approval number: 2013MEKY018) for the research conducted in Shaanxi and Gansu. Permission to conduct the in-school survey and vision test was granted by each sample school's local education bureaus and principals. Written informed consent was obtained from at least one parent for each child participant, and student participants provided oral consent during the survey. The study adhered closely to the principles outlined in the Declaration of Helsinki.
3. Result
3.1. Descriptive analysis
Table 1 presents summary statistics for the study sample, divided by sex and school-age group (junior versus primary level). The mean age of the adolescent participants was 12 years. Among them, 51.4 % were male, with a slight difference in age distribution between junior (mean age 13.43) and primary (mean age 10.50) groups. Han ethnicity constituted 77.8 % of the sample, and 47.4 % were only children. Notably, 40.9 % had one or both parents who migrated for work, and 28.1 % boarded at school. Only a small proportion of parents had completed education beyond senior high school (7.2 % of mothers and 10.9 % of fathers).
Table 1.
Summary statistics of key variables among rural adolescents in northwest China (2012, 2019).
| Variables |
Total |
Female |
Male |
Junior |
Primary |
|---|---|---|---|---|---|
| (N = 36,709) | (N = 17,846) | (N = 18,863) | (N = 18,845) | (N = 17,864) | |
| Dependent and Independent Variables | |||||
| Weekday sleep duration, hours, mean (SD) | 8.7 (1.1) | 8.6 (1.1) | 8.7 (1.1) | 8.2 (1.1) | 9.01 (1.0) |
| Visual acuity of better eye (LogMAR), mean (SD) | 0.12 (0.3) | 0.2 (0.3) | 0.1 (0.3) | 0.2 (0.3) | 0.1 (0.2) |
| Visual acuity of worse eye (LogMAR), mean (SD) | 0.3 (0.3) | 0.32 (0.4) | 0.2 (0.3) | 0.3 (0.4) | 0.2 (0.3) |
| Myopia, n (%) | 13,297 (36.2) | 7431 (41.6) | 5848 (31.0) | 8420 (44.7) | 4859 (27.2) |
| Individual and Family Characteristics | |||||
| Student age, in years, mean (SD) | 12.0 (1.8) | 12.0 (1.8) | 12.0 (1.8) | 13.4 (1.0) | 10.5 (1.0) |
| Male, n (%) | 18,885 (51.4) | 9670 (51.3) | 9193 (51.5) | ||
| Han Ethnic, n (%) | 28,583 (77.8) | 13,820 (77.4) | 14,736 (78.1) | 11,277 (59.8) | 17,279 (96.7) |
| Boarding student, n (%) | 10,343 (28.1) | 4992 (28.0) | 5333 (28.3) | 6560 (34.8) | 3765 (21.1) |
| Only child at home, n (%) | 17,431 (47.4) | 8563 (48.0) | 8860 (47.0) | 1241 (6.6) | 16,182 (90.6) |
| Nap time, n (%) | |||||
| 0 | 15,469 (42.1) | 7463 (41.8) | 7970 (42.3) | 8981 (47.7) | 6452 (36.1) |
| 30 min | 10,222 (27.8) | 5267 (29.5) | 4938 (26.2) | 6941 (36.8) | 3264 (18.3) |
| 30–60 min | 5325 (14.5) | 2501 (14.0) | 2818 (14.9) | 2365 (12.5) | 2954 (16.5) |
| More than 60 min | 5756 (15.7) | 2615 (14.7) | 3137 (16.6) | 558 (3.0) |
5194 (29.1) |
| Mother's age, mean (SD) | 37.31(4.9) | 37.1 (4.8) | 37.51(4.9) | 38.7 (4.7) | 35.9 (4.7) |
| Father's age, mean (SD) | 39.8 (5.1) | 39.5 (5.1) | 40.0 (5.1) | 41.1 (4.8) | 38.3 (5.0) |
| Mother finished senior high school and above, n (%) | 2639 (7.2) | 1203 (6.7) | 1434 (7.6) | 1122 (6.0) | 1515 (8.5) |
| Father finished senior high school and above, n (%) | 4019 (10.9) | 1870 (10.5) | 2147 (11.4) | 1667 (8.8) | 2350 (13.2) |
| One or both parents out migrated for work, n (%) | 15,049 (40.9) | 7173 (40.2) | 7851 (41.6) | 5888 (31.2) | 9136 (51.1) |
| Tercile of family asset, n (%) | |||||
| Bottom | 11,350 (30.9) | 5667 (31.8) | 5660 (30.0) | 6137 (32.6) | 5190 (29.1) |
| Middle | 12,901 (35.1) | 6355 (35.6) | 6535 (34.6) | 6279 (33.3) | 6611 (37.0) |
| Top | 12,511 (34.0) | 5824 (32.6) | 6668 (35.3) | 6429 (34.1) | 6063 (33.9) |
| Instrumental Variable | |||||
| Average sunset in the month of the study of each county, mean (SD) | 19.1 (0.8) | 19.1 (0.8) | 19.1 (0.8) | 18.3 (0.0) | 19.8 (0.1) |
| Other Variables | |||||
| Province, n (%) | |||||
| Ningxia | 18,899 (51.4) | 9175 (51.4) | 9670 (51.3) | 18,845 (100.0) | 0.0 (0.0) |
| Gansu | 9440 (25.7) | 4704 (26.4) | 4734 (25.1) | 0.0 (0.0) |
9438 (52.8) |
| Shaanxi | 8433 (22.9) | 3967 (22.2) | 4459 (23.6) | 0.0 (0.0) |
8426 (47.2) |
| Grade, n (%) | |||||
| 4 | 8690 (23.6) | 4188 (23.5) | 4502 (23.9) | 0.0 (0.0) |
8690 (48.6) |
| 5 | 9174 (25.0) | 4483 (25.1) | 4691 (24.9) | 0.0 (0.0) |
9174 (51.4) |
| 7 | 9399 (25.6) | 4497 (25.2) | 4881 (25.9) | 9378 (49.8) | 0.0 (0.0) |
| 8 | 9499 (25.8) | 4678 (26.2) | 4789 (25.4) | 9467 (50.2) | 0.0 (0.0) |
| Treat, n(%) | 11,938 (32.5) |
5822 (32.6) | 6116 (32.4) | 0.0 (0.0) |
11,938 (66.8) |
Note: This table presents descriptive statistics for all variables used in the analysis. For binary variables, responses were coded as 1 for “yes” and 0 for “no”. For example, sex was coded as 1 for male; only-child status as 1 if the student had no siblings; boarding status as 1 if the student resided at school. Treat was coded as 1 if the student had previously received free glasses. Continuous variables are reported as means with standard deviations in parentheses, and categorical variables are reported as percentages. All statistics are calculated from the final analytical sample.
Average weekday sleep duration was 8.65 h, with primary school students sleeping longer (mean 9.08 h) than junior school students (mean 8.24 h). Additionally, 57.9 % reported napping on weekdays, with different durations across subgroups. In terms of visual health, the mean visual acuity for the better eye was 0.18 LogMAR units, and 0.28 LogMAR units for the worse eye, with myopia prevalence at 36.2 %. Myopia rates were higher among females (41.6 %) and junior students (44.7 %) compared to males (31.0 %) and primary students (27.2 %), indicating potential differences across these groups. Further breakdowns by family characteristics, regional sunset times, and provincial distributions are also provided.
According to the National Sleep Foundation, the recommended sleep duration is 9–11 h per night for children aged 6–13 years and 8–10 h for adolescents aged 14–17 years (Hirshkowitz et al., 2015). As shown in Fig. 1, nearly 58.34 % of students reported sleeping less than 9 h per night on weekdays, indicating that a substantial proportion of children and adolescents in our study were not meeting the recommended sleep duration for their age group.
Fig. 1.
Distribution of weekday wleep duration among rural adolescents in northwest China (2012, 2019).
3.2. The association between weekday sleep duration and vision health
Table 2 presents the estimated associations between sleep duration and vision health. The first model reports unadjusted results, while subsequent models incrementally control for individual, family, and sociodemographic characteristics. The final specification includes county-fixed effects to account for geographic variation. All standard errors are clustered at the school level, except for the baseline model.
Table 2.
Association between weekday sleep duration and visual acuity among rural adolescents in northwest China (2012, 2019).
| LogMAR bare vision of better eye |
|||
|---|---|---|---|
| (1) | (2) | (3) | |
| Weekday sleep duration | −0.026*** | −0.011*** | −0.008*** |
| (0.001) | (0.002) | (0.002) | |
| Constant | 0.405*** | 0.078** | −0.047 |
| (0.012) | (0.032) | (0.040) | |
| Control variables | NO | YES | YES |
| County FE | NO | NO | YES |
| R-sq | 0.010 | 0.050 | 0.073 |
| N | 36,762 | 36,762 | 36,762 |
Note: Ordinary least squares (OLS) regressions estimate the association between weekday sleep duration (hours) and LogMAR visual acuity score. All models adjust for student age, sex, ethnicity, only-child status, boarding status, nap duration, phone screen time, outdoor activity, parental education, parental age, parental migration status, household asset index, province, grade and treat. Standard errors clustered at the school level are reported in parentheses. Column (1) reports unadjusted estimates. Column (2) includes student- and household-level covariates, and column (3) additionally controls for county fixed effects.
* significant at 0.1; ** significant at 0.05; *** significant at 0.01.
Table 2 reveals a substantial positive correlation between weekday sleep duration and vision health across all regressions. Longer sleep duration is associated with lower LogMAR values, indicating better visual acuity (p < 0.01). Estimates from the adjusted model suggest that for every additional hour of sleep duration, there is a 0.008 LogMAR units decrease in the vision acuity of the better eye (p < 0.01).
In additional analyses, we examined sleep timing variables, including bedtime and sleep midpoint, alongside sleep duration (see Table A1). We find that later bedtime and later sleep midpoint are both significantly associated with worse visual acuity. Notably, sleep duration continues to show a negative and significant association with LogMAR scores even after adjusting for sleep timing, indicating that both the quantity and timing of sleep matter for adolescent visual health. All regressions control for individual and family covariates as well as county fixed effects.
3.3. Instrumental variable estimates
To address potential endogeneity issues between sleep duration and outcomes related to visual health, we applied IV estimation using sunset time as our instrument. Table 3 presents the IV results for vision health. Firstly, we confirmed the strength of sunset time as an IV. The first-stage results demonstrate a strong correlation, with estimated coefficients significant at the 1 % level and F-statistics of 53.33, well above the commonly accepted threshold of 10 (Greene, 2018).
Table 3.
Two-stage least squares instrumental variable estimates of the association between weekday sleep duration and visual acuity among rural adolescents in northwest China (2012, 2019).
| (1) | (2) | |
|---|---|---|
| Weekday sleep duration | LogMAR bare vision of better eye | |
| Weekday sleep duration | −0.070* | |
| (0.040) | ||
| Average sunset | −2.148*** | |
| (0.293) | ||
| Control variables | YES | YES |
| county FE | YES | YES |
| F-first stage | 53.669 | – |
| N | 36,762 | 36,762 |
Note: Two-stage least squares (2SLS) regressions use sunset time as an instrumental variable for weekday sleep duration. Estimates are adjusted for student age, sex, ethnicity, only-child status, boarding status, daily nap duration, phone screen time, outdoor activity during lunch break, parental education, parental migration status, parental age, household asset index, province, grade, and treat. County fixed effects are included. Column (1) reports the first-stage regression results, and column (2) reports the second-stage results. Robust standard errors are clustered at the school level.
* p < 0.10, ** p < 0.05, *** p < 0.01.
Secondly, the results, the second-stage estimates for the impact of sleep duration on visual health with endogeneity controlled, remain significant at the 1 % level, with larger coefficients than those in Table 2. This indicates that the positive relationship between sleep duration and vision health is robust to endogeneity correction. These findings support our hypothesis and align with research underscoring the relationship between sleep and adolescent health.
3.4. Robustness check
Our primary findings suggest a positive association between sleep duration and visual health. To validate the robustness and reliability of these results, we conducted multiple robustness tests that align with our research hypotheses, ensuring the consistency of the observed relationships across different conditions and subgroups. These tests were selected to address potential sources of variability and control for nonlinear effects, outliers, and sample-specific influences that could affect the results. The results of these robustness checks are presented in Table 4.
Table 4.
Robustness checks of the association between weekday sleep duration and visual acuity among rural adolescents in northwest China (2012, 2019).
| LogMAR bare vision of worse eye |
Myopia |
LogMAR bare vision of better eye |
||||
|---|---|---|---|---|---|---|
| (1) | (2) | (3) Myopia = 1 | (4) Boarding student = 0 | (5)Low-asset group (bottom 33 %) |
(6)High-asset group (top 33 %) |
|
| Weekday sleep duration | −0.008*** | −0.029*** | −0.007*** | −0.007*** | −0.008*** | −0.012*** |
| (0.002) | (0.008) | (0.002) | (0.002) | (0.003) | (0.003) | |
| Constant | −0.001 | −1.216*** | 0.205*** | −0.077* | 0.006 | −0.024 |
| (0.045) | (0.179) | (0.052) | (0.046) | (0.060) | (0.059) | |
| Control variables | YES | YES | YES | YES | YES | YES |
| County FE | YES | YES | YES | YES | YES | YES |
| N | 36,762 | 36,762 | 13,297 | 26,419 | 11,350 | 12,511 |
Note: Estimates are adjusted for student age, sex, ethnicity, only-child status, boarding status, daily nap duration, phone screen time, outdoor activity during lunch break, parental education, parental migration status, parental age, household asset index, province, grade, and treat. County fixed effects are included. Columns (1), (3), and (4) use OLS regression, and column (2) uses Probit regression. *p < 0.10, **p < 0.05, ***p < 0.01.
First, we substituted our primary measure of visual health with the visual acuity of the worse eye, a method commonly used to assess vision in medical research (Finger et al., 2013; Hirneiss, 2014; Brown et al., 2001). The results demonstrate that an additional hour of sleep is associated with a significant improvement in visual acuity (p < 0.01), indicating consistency with our main results.
Additionally, we examined the probability of myopia as an alternative outcome variable. The results show that longer sleep duration reduces the probability of myopia (p < 0.01), suggesting a protective effect of sleep on visual health. We also performed separate regressions on subsamples of myopic and non-boarding students to confirm that the observed association remains robust across subpopulations (p < 0.01).
To further assess whether the relationship between weekday sleep duration and visual acuity varies by family socioeconomic status (SES), we conducted subgroup analyses by family asset level. As shown in Table 4, the negative association between weekday sleep duration and poor visual acuity remained significant among both the lowest 33 % and highest 33 % of the asset distribution. However, the magnitude of the coefficient was slightly larger for students from poorer households, suggesting that adequate sleep may be particularly protective for students in more disadvantaged family environments.
3.5. Mechanism exploration: sleep duration and screen exposure
To examine potential behavioral mechanisms linking sleep and visual health, we analyzed the relationship between weekday sleep duration and different types of time use, including screen-related and outdoor activities. Table 5 reports these results. We found that weekday sleep duration was significantly negatively associated with time spent on phones (β = −0.092, p < 0.01) and computers (β = −0.063, p < 0.01). However, there was no statistically significant association between sleep duration and time spent watching TV or engaging in outdoor activities.
Table 5.
Mechanism analysis of the association between weekday sleep duration and screen use among rural adolescents in northwest China (2012, 2019).
| (1) |
(2) |
(3) |
(4) |
|
|---|---|---|---|---|
| Time spent on phone | Time spent on TV | Time spent on computer | Time spent on outdoor sports |
|
| Weekday sleep duration | −0.092*** | −0.013 | −0.063*** | −0.002 |
| −0.009 | −0.01 | −0.01 | −0.008 | |
| Control vars | YES | YES | YES | YES |
| County FE | YES | YES | YES | YES |
| N | 36,709 | 36,709 | 36,660 | 36,709 |
Note: Ordinary least squares (OLS) regressions estimate the association between weekday sleep duration and time spent on mobile phones, computers, television, and outdoor activities. All models adjust for student age, sex, ethnicity, only-child status, boarding status, daily nap duration, phone screen time, outdoor activity during lunch break, parental education, parental migration status, parental age, household asset index, province, grade, and treat. Robust standard errors are clustered at the school level.
* p < 0.10, ** p < 0.05, *** p < 0.01.
These findings suggest that shorter sleep duration may reflect higher engagement in screen-related activities, mainly smartphone and computer use, during pre-sleep hours. Since screen exposure is a well-established risk factor for myopia, this pathway offers a plausible behavioral mechanism through which sleep deprivation may impair visual health. This mechanism complements our main findings and highlights the importance of addressing both sleep hygiene and digital behaviors in youth vision protection strategies.
4. Discussion
Adolescents in China and other developing countries frequently experience sleep deprivation, which can have significant implications for their physical development and long-term health. This study used data from a large-scale survey in rural northwest China to investigate the relationship between weekday sleep duration and adolescents' visual health. By focusing on objectively measured visual acuity and using a robust instrumental variable strategy, we offer strong empirical evidence linking shorter sleep duration to an increased risk of myopia.
Our findings indicate that adolescents who sleep less than 8 h on weekdays are more likely to experience visual impairment, as indicated by higher LogMAR scores. These results align with existing literature suggesting that adequate sleep, consistent with natural circadian rhythms, supports ocular recovery and healthy eye growth (Jee et al., 2016; Saara et al., 2022; Cai et al., 2022). In contrast, sleep deprivation may disrupt the physiological processes that regulate eye development, contributing to myopia progression (Wang et al., 2023). Additionally, emerging evidence indicates that socioeconomic disparities are linked not only to sleep duration but also to neurodevelopmental outcomes in children (Hansen et al., 2023; Hansen et al., 2024). Our findings contribute to this literature by exploring vision outcomes in a rural, low-resource context.
Our analysis of behavioral mechanisms found that shorter sleep duration was significantly associated with increased use of mobile phones and computers, especially during pre-sleep hours. These behaviors are recognized risk factors for myopia and may crowd out time that could otherwise be spent outdoors. Our regression results support this mechanism, showing that longer sleep duration is negatively associated with phone and computer use but not significantly correlated with outdoor activity. These findings suggest that digital screen exposure may mediate the relationship between inadequate sleep and myopia risk.
Several limitations warrant discussion. First, although self-reported sleep measures are prone to recall bias and may overestimate actual sleep duration compared to actigraphy, they remain widely used and feasible in large-scale epidemiological studies and school-based surveys. Moreover, we focused on weekday sleep duration to reflect school-related sleep patterns and did not analyze weekend sleep, which may involve additional social and recreational factors. Second, our data are limited to rural China, and the results may not be generalizable to urban settings or international populations. Third, while our analysis included key behavioral covariates, other environmental or physiological factors affecting myopia (e.g., genetics, lighting conditions, or diet) were not available in the dataset. Future research should explore these additional mechanisms. Additionally, given the cross-sectional design, causal interpretations should be made cautiously, although the instrumental variable approach helps address potential endogeneity.
Despite these limitations, our study contributes to a growing body of evidence underscoring the importance of sleep for adolescent health. By combining objective vision measures with robust statistical methods, we demonstrate that sleep duration is not merely a matter of rest but a significant determinant of visual development. Our findings highlight the importance of promoting sleep health not only to improve learning and well-being but also to reduce the rising burden of myopia among school-aged children.
5. Conclusions
This study provides new evidence that insufficient weekday sleep is associated with poorer visual health among rural Chinese adolescents. Students sleeping less than 8 h on school nights are significantly more likely to develop myopia, suggesting that adequate sleep plays a protective role in maintaining visual health. Mechanism analysis indicates that increased nighttime screen exposure may partially explain this relationship. These findings highlight the need for policies that promote healthy sleep habits, regulate late-night screen use, and alleviate academic pressures, offering a low-cost strategy to protect vision and improve adolescent well-being in low-resource settings.
Clinical trial registration
Not applicable.
Contribution statement
H.G. contributed to the conception, design of the work, the methodology, analysis, drafting the work, and obtaining funding for the research. X.C. contributed to the methodology interpretation of the data and substantively revised the study. W.L. contributed to the methodology and interpretation of the data. L.Z. designed the study and contributed to substantively revising the study. Y.D. contributed to the interpretation of the data and substantively revised the study. All authors read and approved the final manuscript.
CRediT authorship contribution statement
Hongyu Guan: Writing – review & editing, Writing – original draft, Funding acquisition, Conceptualization. Xiangzhe Chen: Writing – review & editing, Methodology, Formal analysis, Data curation. Wenting Liu: Methodology, Data curation. Lidong Zhang: Writing – review & editing, Conceptualization. Yuxiu Ding: Writing – review & editing, Data curation.
Ethical approval and consent to participate
Research ethical approval for this study was obtained from the Institutional Review Boards (IRBs) of Stanford University (Approval number: 52514) for the research conducted in Ningxia, and from Stanford University Stanford University (Approval number: 24847) for the research conducted in Shaanxi and Gansu. Permission to conduct the in-school survey and vision test was granted by each sample school's local education bureaus and principals. Written informed consent was obtained from at least one parent for each child participant, and student participants provided oral consent during the survey. The study adhered closely to the principles outlined in the Declaration of Helsinki.
Submission declaration
The work described has not been published previously except as a preprint, an abstract, a published lecture, an academic thesis, or a registered report. All authors and the responsible authorities where the work was carried out tacitly or explicitly approve the article's publication. If accepted, the article will not be published elsewhere in the same form, in English or any other language, including electronically, without the written consent of the copyright holder.
Declaration of Generative AI and AI-assisted technologies in the writing process
During the preparation of this work, the author(s) used ChatGPT to improve the readability and language of the manuscript. After using this tool, the authors reviewed and edited the content as needed and took full responsibility for the content of the published article.
Funding
This work was supported by the Higher Education Discipline Innovation Project (Grant Number: B16031).
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.pmedr.2025.103217.
Appendix A. Supplementary data
The supplementary material include Figure A1, which presents the flowchart of sample selection among rural adolescents in northwest China (2012, 2019), and Table A1, which shows the association between sleep timing and visual acuity outcomes among rural adolescents in northwest China (2012,2019).
Data availability
The dataset used and/or analyzed during the current study is available from the corresponding author on reasonable request.
References
- Alvarez-Peregrina C., Villa-Collar C., Andreu-Vazquez C., Sánchez-Tena M.Á. Influence of vision on educational performance: a multivariate analysis. Sustainability. 2021;13(8):4187. [Google Scholar]
- Beebe D.W., Field J., Milller M.M., et al. Impact of multi-night experimentally induced short sleep on adolescent performance in a simulated classroom. Sleep. 2017;40(2) doi: 10.1093/sleep/zsw035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Brown M.M., Brown G.C., Sharma S., et al. A utility analysis correlation with visual acuity: methodologies and vision in the better and poorer eyes. Int. Ophthalmol. 2001;24(3):123–127. doi: 10.1023/a:1021171000838. [DOI] [PubMed] [Google Scholar]
- Cai T., Zhao L., Kong L., et al. Complex interplay between COVID-19 lockdown and myopic progression. Front. Med. 2022;9 doi: 10.3389/fmed.2022.853293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Finger R.P., Fenwick E., Hirneiss C.W., et al. Visual impairment as a function of visual acuity in both eyes and its impact on patient reported preferences. PloS One. 2013;8(12) doi: 10.1371/journal.pone.0081042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Flitcroft D.I., He M., Jonas J.B., et al. IMI - Defining and Classifying Myopia: A Proposed Set Of Standards for Clinical and Epidemiologic Studies. Invest Ophthalmol. Vis. Sci. 2019;60(3) doi: 10.1167/iovs.18-25957. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fox-Wasylyshyn S.M., El-Masri M.M. Handling missing data in self-report measures. Res. Nurs. Health. 2005;28(6):488–495. doi: 10.1002/nur.20100. [DOI] [PubMed] [Google Scholar]
- Greene W. Econometric Analysis. 8th ed. Pearson Education; London: 2018. [Google Scholar]
- Guan H., Wang Z., Ding Y., Zhang Y., Du K., Shi Y. Association between visual impairment and body mass index in students from rural China. Singapore Med. J. 2023;66(7):362–367. doi: 10.4103/singaporemedj.SMJ-2021-387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hansen M., Simon K.R., Strack J., He X., Noble K.G., Merz E.C. Socioeconomic disparities in sleep duration are associated with cortical thickness in children. Brain Behav. 2023;13(2) doi: 10.1002/brb3.2859. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hansen M., Simon K.R., He X., Steele N., Thomas M.L., Noble K.G., Merz E.C. Socioeconomic factors, sleep timing and duration, and amygdala resting-state functional connectivity in children. Front. Psych. 2024;15:1373546. doi: 10.3389/fpsyt.2024.1373546. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harris J., Lord C. Mental health of children with vision impairment at 11 years of age. Dev. Med. Child Neurol. 2016;58(7):774–779. doi: 10.1111/dmcn.13032. [DOI] [PubMed] [Google Scholar]
- Hirneiss C. The impact of a better-seeing eye and a worse-seeing eye on vision-related quality of life. Clin. Ophthalmol. 2014;2014:1703–1709. doi: 10.2147/OPTH.S64200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hirshkowitz M., Whiton K., Albert S.M., et al. National Sleep Foundation’s updated sleep duration recommendations: final report. Sleep Health. 2015;1(4):233–243. doi: 10.1016/j.sleh.2015.10.004. [DOI] [PubMed] [Google Scholar]
- Jee D., Morgan I.G., Kim E.C. Inverse relationship between sleep duration and myopia. Acta Ophthalmol. 2016;94(3):e204–e210. doi: 10.1111/aos.12776. [DOI] [PubMed] [Google Scholar]
- Li L., Liu H., Kang Y., et al. The influence of parental involvement on students’ non-cognitive abilities in rural ethnic regions of Northwest China. Stud. Educ. Eval. 2024;81 [Google Scholar]
- Li D.L., Nie X.Y., Li J., et al. Factors Associated with Sleep Disorders among Adolescent Students in Rural Areas of China. Front Public Health. 2023;11 doi: 10.3389/fpubh.2023.1152151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu Y.Q., Chen Y., Li P., et al. Sleep status and its relationship with poor vision among junior high school students in Zhoushan. Chinese Prim. Health Care. 2022;36(02):82–85. [Google Scholar]
- Liu X.N., Naduvilath T.J., Wang J., et al. Sleeping Late Is a Risk Factor for Myopia Development Amongst School-Aged Children in China. Sci. Rep. 2020;10(1) doi: 10.1038/s41598-020-74348-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lu X., Guo C., Xu B., et al. Association of Myopia in Elementary School Students in Jiaojiang District, Taizhou City, China. J. Ophthalmol. 2021;((no pagination):):3504538. doi: 10.1155/2021/3504538. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lufi D., Tzischinsky O., Hadar S. Delaying School Starting Time by One Hour: Some Effects on Attention Levels in Adolescents. J. Clin. Sleep Med. 2011;7:137–143. [PMC free article] [PubMed] [Google Scholar]
- Ma X., Wang H., Shi Y., et al. Improving learning by improving vision: evidence from two randomized controlled trials of providing vision care in China. J. Dev. Eff. 2021;13(1):1–26. [Google Scholar]
- Olds, T., Blunden, S., Petkov, J., et al., 2010. The Relationships between Sex, Age, Geography and Time in Bed in Adolescents: A Meta-Analysis of Data from 23 Countries. Sleep Med. Rev, 14 (6): 371–378. [DOI] [PubMed]
- Rhie, S., Lee, S., Chae, K. Y, 2011. Sleep Patterns and School Performance of Korean Adolescents Assessed Using a Korean Version of the Pediatric Daytime Sleepiness Scale. Korean J. Pediatr, 54 (1): 29–35. [DOI] [PMC free article] [PubMed]
- Saara K., Swetha S., Subhiksha R., et al. Steep increase in myopia among public school-going children in South India after COVID-19 home confinement. Indian J. Ophthalmol. 2022;70(8):3040–3046. doi: 10.4103/ijo.IJO_40_22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sadeh A., Gruber R., Raviv A. The Effects of Sleep Restriction and Extension on School-Age Children: What A Difference an Hour Makes. Child Dev. 2003;74:444–455. doi: 10.1111/1467-8624.7402008. [DOI] [PubMed] [Google Scholar]
- Wang H., Cousineau C., Fan Y., et al. Gender equity in vision care seeking behavior among caregivers: evidence from a randomized controlled trial in rural China. Int. J. Equity Health. 2022;21(1):26. doi: 10.1186/s12939-022-01625-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang X.X., Liu X., Lin Q., et al. Association between sleep duration, sleep quality, bedtime and myopia: a systematic review and meta-analysis. Clin. Exp. Ophthalmol. 2023;51(7):673–684. doi: 10.1111/ceo.14277. [DOI] [PubMed] [Google Scholar]
- Wei S.F., Li S.M., Liu L., et al. Sleep duration, bedtime, and myopia progression in a 4-year follow-up of Chinese children: the Anyang Childhood Eye Study. Invest. Ophthalmol. Vis. Sci. 2020;61(3) doi: 10.1167/iovs.61.3.37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Xu X., Wang D., Xiao G., et al. Sleep less, myopia more. Theory Clin Pract. Pediatr. 2019;1:11–17. [Google Scholar]
- Zhou Z., Morgan I.G., Chen Q., et al. Disordered Sleep and Myopia Risk among Chinese Children. PLoS One. 2015;10(3) doi: 10.1371/journal.pone.0121796. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
The supplementary material include Figure A1, which presents the flowchart of sample selection among rural adolescents in northwest China (2012, 2019), and Table A1, which shows the association between sleep timing and visual acuity outcomes among rural adolescents in northwest China (2012,2019).
Data Availability Statement
The dataset used and/or analyzed during the current study is available from the corresponding author on reasonable request.