This cross-sectional study investigates the sociodemographic factors associated with visual impairment in adolescent children in the US.
Key Points
Question
What sociodemographic factors are associated with visual impairment in adolescent children in the US?
Findings
In this cross-sectional study of 2833 adolescents from a nationally representative sample, Black and Mexican American adolescent children had 3 times the odds of reporting poor subjective visual function and twice the odds of presenting with visual acuity worse than 20/40 in the better-seeing eye compared with non-Hispanic White adolescent children. Children from low-income households had more than twice the odds of self-reporting poor visual function.
Meaning
Racial, ethnic, and socioeconomic disparities in subjective and objective visual function are evident by adolescence in the US.
Abstract
Importance
Although racial, ethnic, and socioeconomic disparities in visual impairment have been described in adults, few studies have focused on the adolescent population, which may provide insight into the emergence of vision health inequities.
Objective
To describe visual health disparities among adolescent children in the US.
Design, Setting, and Participants
This was a cross-sectional study of adolescents from the 2005 to 2008 National Health and Nutrition Examination Survey. Participants were aged 12 to 18 years with a completed visual function questionnaire and eye examination. Data analyses were conducted from January 19 to July 20, 2022.
Main Outcomes and Measures
Outcomes included subjective (self-reported poor vision) and objective (visual acuity worse than 20/40 in the better-seeing eye) measures of visual function. Multivariable logistic and linear regression analyses were conducted to examine the association between the sociodemographic risk factors and each outcome, adjusting for age, sex, and other covariates.
Results
The 2833 included participants (mean [SD] age, 15.5 [2.0] years; 1407 female participants [49%]) represent a survey-weighted 57 million US adolescent children, of whom 14% were non-Hispanic Black participants (876), 11% were Mexican American participants (828), 63% were non-Hispanic White participants (816), and 11% were other race and ethnicity (313). A total of 5% of participants (266) were not US citizens, and 19% (773) had a family income below the poverty threshold. There were increased odds of self-reported poor vision among Black (odds ratio [OR], 2.85; 95% CI, 2.00-4.05; P < .001), Mexican American (OR, 2.83; 95% CI, 1.70-4.73; P < .001), and low-income (OR, 2.44; 95% CI, 1.63-3.65; P < .001) adolescent children. Similarly, there were increased odds of visual acuity worse than 20/40 in the better-seeing eye among Black (OR, 2.13; 95% CI, 1.41-3.24; P = .001), Mexican American (OR, 2.13; 95% CI, 1.39-3.26; P = .001), and non-US citizen (OR, 1.96; 95% CI, 1.10-3.49; P = .02) participants.
Conclusions and Relevance
In this nationally representative sample from 2005 to 2008, adolescent children identifying as Black, Mexican American, low-income, or non-US citizen were more likely to report poor subjective visual function and perform worse on objective visual acuity testing. A greater understanding of the underlying etiology of these disparities may yield opportunities for improving vision at the population level.
Introduction
Black, Hispanic, and low-income adults are more likely to have visual impairment, often attributable to preventable causes.1,2,3 Understanding the association of race, ethnicity, and socioeconomic status with visual impairment in childhood may provide insight into the emergence of vision health inequities and reveal opportunities for intervention. In this study, we focused on the sociodemographic factors associated with visual impairment in adolescent children in the US. We hypothesized that racial, ethnic, and socioeconomic disparities in visual function may be identifiable by adolescence.
Methods
The National Health and Nutrition Examination Survey (NHANES) selects approximately 5000 US civilian participants each year to undergo health questionnaires and examinations.4 Our cross-sectional analysis of the 2005 to 2008 survey cycles, the most recent to include both vision-related questionnaires and eye examinations of adolescent children, was deemed exempt by the Boston Children’s Hospital institutional review board (IRB) because it does not represent human subjects research. This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.
We included all children between 12 and 18 years of age with a completed vision questionnaire and examination. The IRB granted a waiver of informed consent owing to the use of deidentified data obtained from a publicly available database. The study outcomes included subjective and objective measures of visual function. Subjective visual function was treated as a binary outcome based on the response to the following survey question: “At the present time, would you say your eyesight, with glasses or contact lenses, if you wear them is—” with the responses “good” or “excellent” considered good subjective vision and “fair,” “poor,” or “very poor” considered poor subjective vision. Visual acuity was measured at distance with the participant’s usual refractive correction (eyeglasses, contacts, or no correction) using the built-in acuity chart of the Nidek Autorefractor, model ARK-760 (Nidek Inc).4 Participants with visual acuity of 20/30 or worse underwent objective refraction using the autorefractor, and repeat measurements were obtained with the updated refraction (eMethods in the Supplement). Objective visual function was treated as a binary outcome with poor vision defined as presenting visual acuity of worse than 20/40 in the better-seeing eye, a threshold previously used in studies of visual impairment with NHANES data.5,6,7 Visual acuity worse than 20/40 in the better-seeing eye after automated refraction was evaluated as a secondary outcome. Sensitivity analyses were performed with visual acuity treated as a continuous outcome (eMethods in the Supplement).
Sociodemographic factors examined included age, sex, race and ethnicity, family income, household size, and US citizenship. Age was treated as a continuous variable. Self-reported race and ethnicity were categorized as non-Hispanic Black, Mexican American, non-Hispanic White, and other, which included other Hispanic ethnicity and all other races/multiracial. Reporting estimates for all Hispanic individuals as a single subgroup is not recommended for the 2005 to 2008 survey cycles owing to changes in sampling design.8 Sensitivity analyses were performed for all Hispanic individuals (including Mexican American and other Hispanic) as a single subgroup in the 2007 to 2008 cycle (eMethods in the Supplement). Low family income was defined as the ratio of family income to poverty threshold (poverty income ratio) of 1 or less. Large household size was defined as 6 or more individuals living in the household. US citizenship was defined as a survey response of citizenship by birth or naturalization.
We performed all analyses using the appropriate survey weights to account for the multistage probability sampling design of NHANES. χ2 tests with Rao-Scott second-order correction were performed. Logistic and linear regression models were used to measure the association of each sociodemographic factor with subjective and objective visual function. The adjusted models included age, sex, race and ethnicity, family income, household size, and citizenship as covariates. Odds ratios (ORs) with 95% CIs were reported from the logistic regression models, and regression coefficients (β) were reported from the linear regression models. All analyses were performed using R, version 4.2.0 (R Core Team), with 2-sided statistical tests and significance defined as P < .05. Data analyses were conducted from January 19 to July 20, 2022.
Results
The 2833 included participants (mean [SD] age, 15.5 [2.0] years; 1407 female participants [49%]; 1426 male participants [51%]) represent a survey-weighted 57 million US adolescent children, of whom 14% were non-Hispanic Black participants (876), 11% were Mexican American participants (828), 63% were non-Hispanic White participants (816), and 11% were other race and ethnicity (313). A total of 5% of participants (266) were not US citizens, and 19% had a family income below the poverty threshold (773) (eTable 1 in the Supplement).
Compared with non-Hispanic White race and ethnicity, the prevalence of poor subjective visual function was greater among participants identifying as Black (11.8% vs 3.8%; difference, 8.1%; 95% CI, 6.1%-10.0%; P < .001) and Mexican American (11.9% vs 3.8%; difference, 8.1%; 95% CI, 4.6%-11.6%; P < .001). In addition, the prevalence of poor subjective visual function was greater among non-US citizens compared with US citizens (13.1% vs 6.0%; difference, 7.1%; 95% CI, 0.9%-13.2%; P = .02), low-income families compared with higher-income families (13.8% vs 4.6%; difference, 9.2%; 95% CI, 6.0%-12.4%; P < .001), and large households compared with small households (8.8% vs 6.0%; difference, 2.8%; 95% CI, 0.2%-5.4%; P = .03) (Table 1).
Table 1. Prevalence of Poor Subjective Visual Function in Adolescent Children by Sociodemographic Factor.
| Characteristic | Poor subjective visual functiona | ||
|---|---|---|---|
| Prevalence, %b | Difference, % (95% CI)c | P value | |
| Sex | |||
| Male | 5.6 | Reference | NA |
| Female | 7.2 | 1.6 (−0.3 to 3.5) | .10 |
| Race/ethnicity | |||
| Non-Hispanic Black | 11.8 | 8.1 (6.1 to 10.0) | <.001 |
| Mexican American | 11.9 | 8.1 (4.6 to 11.6) | <.001 |
| Non-Hispanic White | 3.8 | Reference | NA |
| Otherd | 8.8 | 5.0 (0.8 to 9.2) | .02 |
| Family income | |||
| Poverty income ratio >1 | 4.6 | Reference | NA |
| Poverty income ratio ≤1 | 13.8 | 9.2 (6.0 to 12.4) | <.001 |
| Citizenship | |||
| US citizen | 6.0 | Reference | NA |
| Non-US citizen | 13.1 | 7.1 (0.9 to 13.2) | .02 |
| Household size | |||
| <6 Individuals | 6.0 | Reference | NA |
| ≥6 Individuals | 8.8 | 2.8 (0.2 to 5.4) | .03 |
Abbreviation: NA, not applicable.
Defined as a response of “fair,” “poor,” or “very poor” to the survey question “At the present time, would you say your eyesight, with glasses or contact lenses, if you wear them is—?”
Survey-weighted prevalence accounting for the multistage sampling of the National Health and Nutrition Examination Survey.
Regression least-squares mean difference and corresponding P value.
Other race and ethnicity includes other Hispanic ethnicities and all other races/multiracial.
Compared with non-Hispanic White participants, presenting visual acuity worse than 20/40 in the better-seeing eye was more prevalent among Black participants (15.6% vs 7.2%; difference, 8.4%; 95% CI, 4.4%-12.4%; P < .001) and Mexican American participants (17.9% vs 7.2%; difference, 10.6%; 95% CI, 7.2%-14.1%; P < .001). In addition, presenting visual acuity worse than 20/40 in the better-seeing eye was more prevalent among non-US citizens compared with US citizens (21.3% vs 9.7%; difference, 11.6%; 95% CI, 4.3%-18.8%; P = .002) and participants from large households compared with those from small households (13.9% vs 9.6%; difference, 4.4%; 95% CI, 0.5%-8.3%; P = .03). After objective refraction, the prevalence of vision worse than 20/40 remained greater for Black participants (3.2% vs 0.9%; difference, 2.3%; 95% CI, 0.8%-3.9%; P = .004) and Mexican American participants (2.7% vs 0.9%; difference, 1.8%; 95% CI, 0.7%-2.9%; P = .002) compared with non-Hispanic White participants (Table 2).
Table 2. Prevalence of Visual Acuity Worse Than 20/40 in the Better-Seeing Eye Measured With Baseline and Updated Refractive Correction.
| Characteristic | Poor objective visual functiona | |||||
|---|---|---|---|---|---|---|
| With baseline refractive correction | With updated refractive correction | |||||
| Prevalence, %b | Difference, % (95% CI)c | P value | Prevalence, %b | Difference, % (95% CI)c | P value | |
| Sex | ||||||
| Male | 9.6 | Reference | NA | 1.3 | Reference | NA |
| Female | 11.2 | 1.7 (−1.3 to 4.6) | .26 | 1.7 | 0.4 (−0.6 to 1.3) | .44 |
| Race/ethnicity | ||||||
| Non-Hispanic Black | 15.6 | 8.4 (4.4 to 12.4) | <.001 | 3.2 | 2.3 (0.8 to 3.9) | .004 |
| Mexican American | 17.9 | 10.6 (7.2 to 14.1) | <.001 | 2.7 | 1.8 (0.7 to 2.9) | .002 |
| Non-Hispanic White | 7.2 | Reference | NA | 0.9 | Reference | NA |
| Otherd | 13.9 | 6.6 (2.1 to 11.2) | .004 | 1.6 | 0.7 (−0.9 to 2.4) | .39 |
| Family income | ||||||
| Poverty income ratio >1 | 9.7 | Reference | NA | 1.3 | Reference | NA |
| Poverty income ratio ≤1 | 12.3 | 2.7 (−0.8 to 6.1) | .13 | 2.6 | 1.3 (−0.4 to 3.1) | .13 |
| Citizenship | ||||||
| US citizen | 9.7 | Reference | NA | 1.5 | Reference | NA |
| Non-US citizen | 21.3 | 11.6 (4.3 to 18.8) | .002 | 1.7 | 0.2 (−1.4 to 1.8) | .77 |
| Household size | ||||||
| <6 Individuals | 9.6 | Reference | NA | 1.3 | Reference | NA |
| ≥6 Individuals | 13.9 | 4.4 (0.5 to 8.3) | .03 | 2.4 | 1.1 (−0.6 to 2.8) | .21 |
Abbreviation: NA, not applicable.
Visual acuity worse than 20/40 in the better-seeing eye obtained with participant’s current refractive correction at the time of exam (glasses, contacts, or no correction) and using objective refraction obtained from Nidek Autorefractor, model ARK-760.
Survey-weighted prevalence accounting for the multistage sampling of the National Health and Nutrition Examination Survey.
Regression least-squares mean difference and corresponding P value.
Other race and ethnicity includes other Hispanic ethnicities and all other races/multiracial.
Multivariable regression analysis revealed increased odds of poor subjective vision in Black participants (OR, 2.85; 95% CI, 2.00-4.05; P < .001), Mexican American participants (OR, 2.83; 95% CI, 1.70-4.73; P < .001), and low-income participants (OR, 2.44; 95% CI, 1.63-3.65; P < .001) after adjusting for age, sex, and other covariates. Similarly, there were increased odds of low presenting visual acuity for Black participants (OR, 2.13; 95% CI, 1.41-3.24; P = .001), Mexican American participants (OR, 2.13; 95% CI, 1.39-3.26; P = .001), and non-US citizens (OR, 1.96; 95% CI, 1.10-3.49; P = .02). These differences were not significant after objective refraction (Table 3). Similar associations were observed in the sensitivity analysis treating visual acuity as a continuous outcome and among all Hispanic adolescent children (including Mexican American and other Hispanic ethnicity) in the 2007 to 2008 survey cycle (eTable 2 and eTable 3 in the Supplement).
Table 3. Logistic Regression Models Describing the Association of Subjective and Objective Visual Function With Patient Sociodemographic Factors.
| Characteristic | Poor subjective visual function | Poor objective visual function | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Unadjusted | Adjusteda | With baseline refractive correction | With updated refractive correction | |||||||||
| OR (95% CI) | P value | OR (95% CI) | P value | Unadjusted | Adjusteda | Unadjusted | Adjusteda | |||||
| OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | |||||
| Age (per 1-y increase) | 1.04 (0.95-1.14) | .34 | 1.06 (0.97-1.15) | .16 | 0.89 (0.81-0.97) | .01 | 0.87 (0.79-0.96) | .006 | 0.94 (0.80-1.10) | .42 | 0.97 (0.81-1.16) | .70 |
| Female sex (reference male sex) | 1.31 (0.92-1.85) | .13 | 1.29 (0.89-1.87) | .17 | 1.20 (0.87-1.65) | .27 | 1.20 (0.85-1.70) | .29 | 1.29 (0.65-2.55) | .45 | 1.28 (0.61-2.67) | .49 |
| Race/ethnicityb | ||||||||||||
| Non-Hispanic Black | 3.42 (2.50-4.69) | <.001 | 2.85 (2.00-4.05) | <.001 | 2.37 (1.58-3.56) | <.001 | 2.13 (1.41-3.24) | .001 | 3.66 (1.39-9.66) | .01 | 2.93 (0.99-8.71) | .05 |
| Mexican American | 3.44 (2.23-5.28) | <.001 | 2.83 (1.70-4.73) | <.001 | 2.79 (2.02-3.85) | <.001 | 2.13 (1.39-3.26) | .001 | 3.05 (1.30-7.17) | .01 | 2.61 (0.90-7.52) | .07 |
| Otherc | 2.46 (1.27-4.08) | .01 | 2.39 (1.29-4.45) | .008 | 2.06 (1.31-3.24) | .003 | 1.67 (1.09-2.57) | .02 | 1.82 (0.49-6.78) | .36 | 1.69 (0.42-6.78) | .45 |
| Low family income | 3.34 (2.30-4.84) | <.001 | 2.44 (1.63-3.65) | <.001 | 1.31 (0.93-1.86) | .12 | 0.94 (0.66-1.34) | .73 | 2.08 (0.88-4.91) | .09 | 1.49 (0.59-3.76) | .39 |
| Non-US citizenship | 2.35 (1.31-4.21) | .006 | 1.33 (0.68-2.61) | .38 | 2.51 (1.59-3.97) | <.001 | 1.96 (1.10-3.49) | .02 | 1.17 (0.43-3.16) | .75 | 0.78 (0.25-2.46) | .66 |
| ≥6 Individuals living in household | 1.45 (1.00-2.11) | .049 | 1.12 (0.73-1.73) | .60 | 1.53 (1.06-2.21) | .03 | 1.33 (0.91-1.92) | .13 | 1.86 (0.81-4.28) | .14 | 1.63 (0.67-3.93) | .26 |
Abbreviation: OR, odds ratio.
Adjusted models include all covariates in the table.
Non-Hispanic White is the reference for the race and ethnicity categories.
Other race and ethnicity includes other Hispanic ethnicities and all other races/multiracial.
Discussion
In this cross-sectional study using a nationally representative sample from NHANES 2005 to 2008, Black and Mexican American adolescent children had approximately 3 times the odds of poor subjective visual function and twice the odds of low objective visual acuity than non-Hispanic White adolescent children after adjusting for socioeconomic status indicators, including family income. Participants from low-income households had more than twice the odds of poor self-reported vision. Racial, ethnic, and socioeconomic differences in subjective and objective visual function can be identified by adolescence in the US.
Previous research has linked poor vision in children to sociodemographic factors such as race, ethnicity, family income, and access to health insurance. Our findings likely represent underlying social and economic inequities in access to vision care services.9,10,11 Children from disadvantaged backgrounds may have fewer ocular diagnoses, lower vision care utilization, and a higher chance of being lost to follow-up.11,12,13 By adolescence, the combination of underdiagnosis and undertreatment may contribute to the observed differences in subjective and objective visual function. Black, Hispanic, and low-income individuals are less likely to wear appropriate refraction or be able to afford glasses.8,9 Much of the visual impairment in this study appears correctable, which emphasizes the importance of access to refractive correction in adolescence, a time when poor vision may affect academic performance, future employment, and economic opportunities.14,15
Strengths and Limitations
This study leveraged a large cohort representative of the US adolescent population and combined self-reported data with objective visual acuity measurements to provide a deeper understanding of the racial, ethnic, and socioeconomic disparities in visual impairment. Limitations included the use of survey data from 2005 to 2008, which may not reflect the visual function of the current US adolescent population; however, as many of these children enter early adulthood, investigation of historical disparities may help inform the practice of physicians caring for this population.
Conclusions
Results of this cross-sectional study suggest that adolescent children identifying as Black, Hispanic, low-income, or non-US citizen were more likely to report poor subjective visual function and perform worse on objective visual acuity testing. Physicians caring for adolescent children should be aware of the racial, ethnic, and socioeconomic disparities in vision. Improving access to vision care services may decrease the burden of preventable visual impairment extending into adulthood.
eMethods
eReferences
eTable 1. Distribution of Model Covariates by Race and Ethnicity
eTable 2. Secondary Analyses of Objective Visual Function With Acuity Treated as a Continuous Outcome
eTable 3. Sensitivity Analysis With All Hispanic Participants (Including Mexican American and Other Hispanic Ethnicity Categories) Included as a Single Subgroup in the 2007-2008 NHANES Survey Cycle
References
- 1.Varma R, Vajaranant TS, Burkemper B, et al. Visual impairment and blindness in adults in the US: demographic and geographic variations from 2015 to 2050. JAMA Ophthalmol. 2016;134(7):802-809. doi: 10.1001/jamaophthalmol.2016.1284 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ko F, Vitale S, Chou CF, Cotch MF, Saaddine J, Friedman DS. Prevalence of nonrefractive visual impairment in US adults and associated risk factors, 1999-2002 and 2005-2008. JAMA. 2012;308(22):2361-2368. doi: 10.1001/jama.2012.85685 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Zambelli-Weiner A, Crews JE, Friedman DS. Disparities in adult vision health in the US. Am J Ophthalmol. 2012;154(6):S23-30.e1. doi: 10.1016/j.ajo.2012.03.018 [DOI] [PubMed] [Google Scholar]
- 4.Centers for Disease Control and Prevention . NHANES questionnaires, datasets, and related documentation. Accessed July 16, 2022. https://wwwn.cdc.gov/nchs/nhanes/default.aspx
- 5.Vitale S, Cotch MF, Sperduto RD. Prevalence of visual impairment in the US. JAMA. 2006;295(18):2158-2163. doi: 10.1001/jama.295.18.2158 [DOI] [PubMed] [Google Scholar]
- 6.Rein DB, Lamuda PA, Wittenborn JS, et al. Vision impairment and blindness prevalence in the US: variability of vision health responses across multiple national surveys. Ophthalmology. 2021;128(1):15-27. doi: 10.1016/j.ophtha.2020.06.064 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Qiu M, Wang SY, Singh K, Lin SC. Racial disparities in uncorrected and undercorrected refractive error in the United States. Invest Ophthalmol Vis Sci. 2014;55(10):6996-7005. doi: 10.1167/iovs.13-12662 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Centers for Disease Control and Prevention . Analytic note regarding 2007-2010 survey design changes and combining data across other survey cycles. Accessed August 18, 2022. https://wwwn.cdc.gov/nchs/data/nhanes/analyticguidelines/07-10-analytic-note.pdf
- 9.Zhang X, Cotch MF, Ryskulova A, et al. Vision health disparities in the US by race/ethnicity, education, and economic status: findings from 2 nationally representative surveys. Am J Ophthalmol. 2012;154(6)(suppl):S53-62.e1. doi: 10.1016/j.ajo.2011.08.045 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Dotan G, Truong B, Snitzer M, et al. Outcomes of an inner-city vision outreach program: give kids sight day. JAMA Ophthalmol. 2015;133(5):527-532. doi: 10.1001/jamaophthalmol.2015.8 [DOI] [PubMed] [Google Scholar]
- 11.Stein JD, Andrews C, Musch DC, Green C, Lee PP. Sight-threatening ocular diseases remain underdiagnosed among children of less affluent families. Health Aff (Millwood). 2016;35(8):1359-1366. doi: 10.1377/hlthaff.2015.1007 [DOI] [PubMed] [Google Scholar]
- 12.Ganz ML, Xuan Z, Hunter DG. Prevalence and correlates of children’s diagnosed eye and vision conditions. Ophthalmology. 2006;113(12):2298-2306. doi: 10.1016/j.ophtha.2006.06.015 [DOI] [PubMed] [Google Scholar]
- 13.Shoshany TN, Chinn RN, Staffa SJ, Bishop K, Michalak S, Hunter DG. Identifying characteristics predictive of lost-to-follow-up status in amblyopia. Am J Ophthalmol. 2021;230:200-206. doi: 10.1016/j.ajo.2021.05.002 [DOI] [PubMed] [Google Scholar]
- 14.Davidson S, Quinn GE. The impact of pediatric vision disorders in adulthood. Pediatrics. 2011;127(2):334-339. doi: 10.1542/peds.2010-1911 [DOI] [PubMed] [Google Scholar]
- 15.Wittenborn JS, Zhang X, Feagan CW, et al. ; Vision Cost-Effectiveness Study Group . The economic burden of vision loss and eye disorders among the US population younger than 40 years. Ophthalmology. 2013;120(9):1728-1735. doi: 10.1016/j.ophtha.2013.01.068 [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
eMethods
eReferences
eTable 1. Distribution of Model Covariates by Race and Ethnicity
eTable 2. Secondary Analyses of Objective Visual Function With Acuity Treated as a Continuous Outcome
eTable 3. Sensitivity Analysis With All Hispanic Participants (Including Mexican American and Other Hispanic Ethnicity Categories) Included as a Single Subgroup in the 2007-2008 NHANES Survey Cycle