Purpose
To examine differences in visual impairment between immigrants and natives in the United States (US).
Design
This is a cross-sectional study of clinical vision examination data from the 2003–2008 National Health and Nutrition Examination Survey.
Methods
Analyses compare myopia, hyperopia, astigmatism, and visual acuity between noncitizens, naturalized citizens, and US natives. Visual acuity variables included having 20/40 or better corrected vision. Differences in 20/20 vision and legal blindness (20/200 or worse) were also examined. Respondents born in the United States were defined as US natives. Foreign-born respondents were categorized as either naturalized US citizens or noncitizen residents. Multivariate logistic regression of outcomes adjusted for sex, age, race/ethnicity, poverty, insurance, diabetes, and surgical correction for eyesight.
Results
A smaller percentage of noncitizens than US natives had myopia (18.8% vs 30.7%) or astigmatism (22.0% vs 30.9%). However, noncitizens using corrective lenses had an adjusted odds ratio (AOR) of 0.65 of having 20/20 vision compared to US natives (95% confidence interval [CI] 0.48–0.88). Differences in visual acuity for 20/40 and better vision were not statistically significant for corrective lens users. Among nonusers of corrective lenses, noncitizens were significantly less likely than US natives to have 20/40 or better vision (AOR = 0.54; 95% CI 0.39–0.74). Noncitizens also had up to 3.5 times the odds of being legally blind relative to US natives after adjusting for confounding factors (95% CI 1.52–7.83).
Conclusions
Significant differences in visual acuity exist between immigrants and US natives. More research is necessary to identify underlying factors that may explain these disparities in visual impairment.
Prior research has shown that immigrants have the same or better life expectancy, infant mortality rates, and other health outcomes as natives in the United States (US). For example, Mexican immigrants have a 40% lower hazard of mortality than non-Latino whites. Cardiovascular disease mortality rates are 31% lower for black immigrants than for US-born blacks. Outcomes associated with hypertension and obesity are also better for older Latino immigrants than for non-Hispanic whites. Latino and non-Latino immigrant groups tend to be less likely to have poor infant health measures such as low birth weight or mortality. Overall, foreign-born persons are expected to live 3.4 years longer than US natives. Immigrants display better-than-expected health outcomes despite having lower socioeconomic status, language barriers, and economic and legislated barriers to health care—a finding termed the “epidemiologic paradox.”
Although racial/ethnic disparities in visual impairment have been documented in previous studies, research on disparities between immigrants and US natives is limited. In a 1998 paper, Voo and associates presented results on the prevalence of myopia, astigmatism, and visual acuity for first- and second-generation immigrant children examined in a mobile eye clinic. However, these data were limited to immigrant children, did not provide a comparison with US native children, and were not nationally representative data. Therefore, more population-based research is needed to better identify the visual health needs of immigrants.
Another motivation for better understanding of the visual health services needs of foreign-born persons comes from recent policy changes. Under the Affordable Care Act, ophthalmologists may be required to query their patients regarding demographic characteristics such as race/ethnicity and primary language used. There is no requirement to ask patients about their immigration status, which could be a concern if evidence suggests that foreign-born persons need regular vision care and encouragement to seek visual screening. To address these gaps in knowledge, our study used nationally representative data on clinical eye examinations to characterize differences in visual impairment between immigrants and US citizens. We examined clinical measures of myopia, hyperopia, astigmatism, and visual acuity taken in a laboratory setting, and compared differences in these measures for noncitizens, naturalized citizens, and US natives stratified by use of corrective lenses.
Methods
Data and Population
Data from the 2003–2008 National Health and Nutrition Examination Survey (NHANES) were used to measure visual impairment by immigration status. The NHANES is an ongoing, nationally representative, publicly available survey conducted by the National Center for Health Statistics, Centers for Disease Control and Prevention, with approval by a human subjects institutional review board. Written informed consent was obtained from NHANES participants. The NHANES collects comprehensive information by both interviews and medical examinations from the US noninstitutionalized civilian population each year, including detailed demographics, socioeconomic characteristics, health-related questions, medical measurements, and laboratory test results. All interviewees aged 12 years and older were eligible for a vision examination. The 2003–2008 NHANES data originally consisted of 20 211 records with nonmissing values for demographic, socioeconomic, and diabetes covariates; among those, 1993 observations failed to report the survey participant’s vision status (vision surgical procedures performed, use of corrective lens). There were 602 observations with missing values in the vision examination data, resulting in our final analytic sample size of 17 616—a difference of 12.8% compared to the overall sample size of 20 221. The NHANES stopped collecting clinical data on visual impairment after 2008.
Measures
The Nidek Auto Refractor Model ARK-760 (Nidek Co Ltd, Gamagori, Japan) was employed to measure refractive error, including 3 separate averaged measurements—sphere, cylinder, and axis—for both eyes, after removing usual corrective lenses. The instrument also tested visual acuity with usual distance vision correction by built-in charts with the scale of 20/20, 20/25, 20/30, 20/40, 20/50, 20/60, 20/80, and 20/200. The scale is truncated at 20/20 and may include persons with better than 20/20 vision. On the basis of reported visual acuity, we chose the better-seeing eye for individuals throughout our analysis, an approach consistent with prior research. If both eyes tested the same, we randomly selected one of them. We then calculated spherical equivalent (SphEq) value as the sphere plus half the cylinder measurement. Our primary dependent variables were clinically defined as follows: myopia as a SphEq value of −1.0 diopter (D) or less, hyperopia as a SphEq value of +3.0 D or greater, and astigmatism as a cylinder value of +1.0 D or greater. In addition, we used 2 measures to denote visual acuity. First, we dichotomized the visual acuity variable by respondents who had perfect vision for daily life (20/20) with usual correction or who did not (equal to 20/25 or worse). Second, we examined respondents who had 20/40 or better vision vs those who had 20/45 or worse vision. The cutoff in qualifying for an unrestricted driver’s license for most states in the United States is 20/40 or better vision. In addition, we examined differences in legal blindness (20/200 or worse visual acuity in the better-seeing eye) across immigration groups.
We used 2 variables (nativity and citizenship status) from NHANES to categorize immigration status as US native, naturalized citizen, and noncitizen. Respondents born in the United States were defined as US natives. Persons not born in the United States were further defined as either naturalized US citizens or noncitizen residents.
Demographic, socioeconomic, and medical history control variables included sex, age (12–19, 20–39, 40–59, and 60–85 years), race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, and other race/ethnicity), education level (less than high school, high school, college), poverty status (annual family income less than the poverty threshold), insurance status (insured or not), and diabetes condition. The NHANES does not provide further detail on “other race/ethnicity” categories. For diabetes, respondents were asked, “Have you ever been told by a doctor or health professional that you have diabetes or sugar diabetes?” Additionally, respondents answered 3 questions regarding their history of vision care utilization: “Have you ever had eye surgery to treat or prevent nearsightedness or myopia?” “Have you ever had eye surgery to treat cataracts?” “Do you wear glasses or contact lenses for distance vision?” It is not possible to differentiate between wearing glasses vs contact lenses using NHANES data.
Analytical Plan
We conducted a univariate analysis to depict distributions of all explanatory and control variables, stratified by immigration status. The Pearson χ 2 test was used to examine disparities in visual impairment between US natives, naturalized citizens, and noncitizen residents. We then used multivariate logistic regression to model the relationship between visual impairment and immigration status for both those who wore corrective lenses and those who did not for each of the dependent variables. We used Stata 12 (StataCorp, College Station, TX, USA) software to adjust for survey weights and the complex survey design of the NHANES in all analyses. We considered a P value of less than .05 as statistically significant.
Results
Descriptive statistics of respondents’ visual impairment, vision care utilization, and sociodemographic characteristics are presented in Table 1 . Of the 17 616 respondents in our sample, 81.5% (14 366) were US natives, 7.1% (1253) were naturalized citizens, and 11.4% (1997) were noncitizen residents. There were significant differences in myopia, hyperopia, astigmatism, and visual acuity between the 3 groups ( P < .01). US natives had the highest prevalence of myopia (30.7%) and astigmatism (30.9%) among those groups, but they also had the best visual acuity with correction (60.8% had 20/20 vision and 93.7% had 20/40 or better vision). Noncitizens had almost 3 times the prevalence of legal blindness as US natives. We also examined trends over time, but they were not statistically significant and thus are not reported.
| US Natives (%) (N = 14 366) | Naturalized Citizens (%) (N = 1253) | Noncitizens (%) (N = 1997) | P Value | |
|---|---|---|---|---|
| Myopia a | 30.7 | 23.6 | 18.8 | <.01 |
| Hyperopia b | 1.9 | 3.0 | 1.1 | <.01 |
| Astigmatism c | 30.9 | 27.4 | 22.0 | <.01 |
| Visual acuity d | ||||
| 20/20 or better | 60.8 | 57.0 | 53.8 | <.01 |
| 20/40 or better | 93.7 | 91.8 | 88.3 | <.01 |
| Legally blind e | 0.7 | 1.0 | 1.9 | <.01 |
| Independent variables | ||||
| Sex | ||||
| Male | 48.3 | 50.2 | 55.2 | <.01 |
| Female | 51.7 | 49.8 | 44.8 | |
| Age | ||||
| 12–19 years | 15.4 | 6.2 | 10.1 | <.01 |
| 20–39 years | 31.4 | 28.2 | 56.7 | |
| 40–59 years | 34.6 | 44.3 | 28.3 | |
| 60–85 years | 18.6 | 21.3 | 4.9 | |
| Race/ethnicity | ||||
| Non-Hispanic white | 78.5 | 29.6 | 20.4 | <.01 |
| Non-Hispanic black | 12.2 | 7.9 | 4.9 | |
| Hispanic | 6.2 | 36.2 | 60.5 | |
| Other | 3.1 | 26.3 | 14.2 | |
| Education | ||||
| Less than 12 years | 25.1 | 26.1 | 49.2 | <.01 |
| 12 years | 23.8 | 19.7 | 18.1 | |
| More than 12 years | 51.1 | 54.2 | 32.7 | |
| Poverty f | 12.1 | 13.1 | 31.5 | <.01 |
| Insured | 85.4 | 81.9 | 48.0 | <.01 |
| Diabetes | 6.4 | 7.8 | 5.0 | .12 |
| Surgery for myopia | 2.0 | 3.2 | 1.1 | <.05 |
| Surgery for cataracts | 4.4 | 4.5 | 1.1 | <.01 |
| Wear glasses | 50.0 | 44.1 | 24.9 | <.01 |
a Myopia is clinically defined as a SphEq refraction value of −1.0 D or less.
b Hyperopia is clinically defined as a SphEq value of 3.0 D or greater.
c Astigmatism is clinically defined as a cylinder value of 1.0 D or greater.
d Visual acuity is defined in 2 ways: (1) 20/20 vs less than 20/20 vision in respondent’s better-seeing eye; and (2) 20/40 or better vs less than 20/40 vision in respondent’s better-seeing eye.
e Legal blindness is defined as 20/200 vision or worse in respondent’s better-seeing eye.
f Poverty is defined as annual family income less than the poverty threshold.
The distribution of SphEq values by immigration status is shown in the Figure . US natives and naturalized citizens shared similar patterns in the distribution of SphEq. In contrast, noncitizen immigrants were less likely to have high or low values of SphEq; half of noncitizens had SphEq values between −0.5 and 0.5, compared to only 36% for US natives. Kolmogorov-Smirnov equality-of-distributions tests rejected the null hypotheses that noncitizens and US natives had the same distribution of SphEq values and that naturalized citizens and US natives had the same distribution of SphEq values.
Tables 2 and 3 present the association between visual impairment and immigration status for those respondents who wore corrective lenses and those who did not, respectively. Adjusted odds ratios (AORs) for both age and multivariate regression, 95% confidence intervals (CIs), and P values are reported, with US natives serving as the reference group for nativity and citizenship status variables. For corrective lens wearers, age-adjusted odds ratios indicated that naturalized citizens were more likely than US natives to have hyperopia (AOR 1.71, 95% CI 1.09–2.66) and less likely to have perfect visual acuity (AOR 0.78, 95% CI 0.61–0.99) ( Table 2 ). Noncitizens were less likely than US natives to have better visual acuity (20/20 AOR 0.54, 95% CI 0.41–0.71) but were also less likely to have myopia (AOR 0.62, 95% CI 0.45–0.85). After adjusting for demographic characteristics, socioeconomic factors, diabetes, and vision care utilization, differences between naturalized citizens and US natives were no longer statistically significant. However, noncitizens continued to have lower odds of myopia than US natives (AOR 0.65, 95% CI 0.45–0.94). Differences between noncitizens and US natives in perfect visual acuity decreased slightly with multivariate vs age adjustment but remained highly statistically significant (20/20 AOR 0.65, 95% CI 0.48–0.88). Differences in visual acuity for 20/40 and better vision were not statistically significant for users of corrective lenses. Noncitizens were also more likely to be legally blind (AOR 2.25, 95% CI 1.14–4.45).
| Visual Impairment | Naturalized Citizens | Noncitizens | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P Value | OR | 95% CI | P Value | |
| Myopia a | ||||||
| Age adjusted | 0.81 | 0.56–1.15 | .23 | 0.62 | 0.45–0.85 | <.01 |
| MV adjusted b | 0.77 | 0.52–1.14 | .19 | 0.65 | 0.45–0.94 | .02 |
| Hyperopia c | ||||||
| Age adjusted | 1.71 | 1.09–2.66 | .02 | 1.40 | 0.87–2.27 | .16 |
| MV adjusted b | 1.49 | 0.75–2.96 | .25 | 1.05 | 0.56–1.96 | .87 |
| Astigmatism d | ||||||
| Age adjusted | 0.85 | 0.67–1.07 | .16 | 1.04 | 0.82–1.34 | .73 |
| MV adjusted b | 0.86 | 0.67–1.11 | .24 | 0.98 | 0.74–1.30 | .89 |
| Visual acuity e | ||||||
| 20/20 vs <20/20 | ||||||
| Age adjusted | 0.78 | 0.61–0.99 | .04 | 0.54 | 0.41–0.71 | <.01 |
| MV adjusted b | 0.82 | 0.64–1.05 | .11 | 0.65 | 0.48–0.88 | <.01 |
| ≥20/40 vs<20/40 | ||||||
| Age adjusted | 0.66 | 0.44–1.00 | .05 | 0.57 | 0.38–0.86 | <.01 |
| MV adjusted b | 0.79 | 0.52–1.20 | .26 | 0.83 | 0.54–1.27 | .37 |
| Legal blindness f | ||||||
| Age adjusted | 0.70 | 0.29–1.68 | .41 | 2.85 | 1.32–6.13 | <.01 |
| MV adjusted b | 0.60 | 0.24–1.55 | .29 | 2.25 | 1.14–4.45 | .02 |
a Myopia is clinically defined as a SphEq refraction value of −1.0 D or less.
b Multivariate logistic regression adjusts for sex, age, race/ethnicity, education, poverty, insurance, surgical correction for myopia or cataracts, diabetes, and survey year.
c Hyperopia is clinically defined as a SphEq refraction value of 3.0 D or greater.
d Astigmatism is clinically defined as a cylinder value of 1.0 D or greater.
e Visual acuity is defined in 2 ways: (1) 20/20 vision vs less than 20/20 vision in respondent’s better-seeing eye; and (2) 20/40 vision or better vs less than 20/40 vision in respondent’s better-seeing eye.
f Legal blindness is defined as 20/200 vision or worse in respondent’s better-seeing eye.
| Visual Impairment | Naturalized Citizens | Noncitizens | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P Value | OR | 95% CI | P Value | |
| Myopia a | ||||||
| Age adjusted | 0.86 | 0.53–1.39 | .53 | 1.38 | 1.02–1.87 | .04 |
| MV adjusted b | 0.73 | 0.46–1.16 | .18 | 0.98 | 0.72–1.34 | .88 |
| Hyperopia c | ||||||
| Age adjusted | Omitted d | Omitted d | Omitted d | 0.54 | 0.11–2.52 | .42 |
| MV adjusted b | Omitted d | Omitted d | Omitted d | 0.31 | 0.10–1.01 | .05 |
| Astigmatism e | ||||||
| Age adjusted | 0.76 | 0.58–0.99 | .05 | 0.94 | 0.76–1.16 | .54 |
| MV adjusted b | 0.78 | 0.59–1.04 | .09 | 0.91 | 0.70–1.19 | .48 |
| Visual acuity f | ||||||
| 20/20 vs <20/20 | ||||||
| Age adjusted | 0.94 | 0.74–1.18 | .57 | 0.65 | 0.56–0.75 | <.01 |
| MV adjusted b | 0.97 | 0.75–1.27 | .83 | 0.72 | 0.60–0.86 | <.01 |
| ≥20/40 vs <20/40 | ||||||
| Age adjusted | 0.77 | 0.53–1.11 | .16 | 0.39 | 0.31–0.50 | <.01 |
| MV adjusted b | 0.83 | 0.57–1.23 | .35 | 0.54 | 0.39–0.74 | <.01 |
| Legal blindness g | ||||||
| Age adjusted | 4.67 | 1.67–13.09 | <.01 | 5.48 | 2.97–10.12 | <.01 |
| MV adjusted b | 4.01 | 1.54–10.46 | <.01 | 3.45 | 1.52–7.83 | <.01 |
a Myopia is clinically defined as a SphEq refraction value of −1.0 D or less.
b Multivariate logistic regression adjusts for sex, age, race/ethnicity, education, poverty, insurance, surgical correction for myopia or cataracts, diabetes, and survey year.
c Hyperopia is clinically defined as a SphEq refraction value of 3.0 D or greater.
d Omitted because of an insufficient number of observations for hyperopia outcome of naturalized citizens.
e Astigmatism is clinically defined as a cylinder value of 1.0 D or greater.
f Visual acuity is defined in 2 ways: (1) 20/20 vision vs less than 20/20 vision in respondent’s better-seeing eye; and (2) 20/40 vision or better vs less than 20/40 vision in respondent’s better-seeing eye.
g Legal blindness is defined as 20/200 vision or worse in respondent’s better-seeing eye.
Among those who did not wear corrective lenses, after adjusting for age, noncitizens were more likely than US natives to have myopia (AOR 1.38, 95% CI 1.02–1.87), and naturalized citizens were less likely to have astigmatism (AOR 0.76, 95% CI 0.58–0.99) ( Table 3 ). After adjusting for demographic characteristics, socioeconomic factors, diabetes, and vision care utilization, noncitizen nonusers of corrective lenses continued to be less likely than US natives to have visual acuity of 20/20 (AOR 0.72, 95% CI 0.60–0.86). For the 20/40 vision cutoff, the AOR was 0.54 (95% CI 0.39–0.74). Both naturalized citizens and noncitizens were more likely than US natives to be legally blind (naturalized citizens: AOR 4.01, 95% CI 1.54–10.46; noncitizens: AOR 3.45, 95% CI 1.52–7.83). As a sensitivity analysis, we replicated the results using clinical outcomes for the worse-seeing eye instead of the better-seeing eye (results not shown).
Regression analyses for respondents reporting Mexican American ethnicity are provided as Supplemental Material (available at AJO.com ). Mexican noncitizen immigrants who did not wear corrective lenses were less likely to have 20/40 or better vision (AOR 0.61, 95% CI 0.40–0.92). Although Mexican noncitizens who wore corrective lenses were less likely to have myopia (AOR 0.57, 95% CI 0.38–0.85), they had 3.8 times higher odds than US native Mexican Americans of being legally blind (95% CI 1.86–7.83).
Because the observed disparity across immigrant groups may have been attributable to differences in health status aside from diabetes, we performed additional analyses adjusting for smoking, alcohol use, exercise, basal metabolic index, comorbid medical conditions, diabetic retinopathy, and self-reported general health status. A description of these measures and analytical results are provided as Supplemental Materials (available at AJO.com ). Sample sizes for these analyses were substantially smaller as a result of list-wise deletion. Despite the addition of these covariates, however, disparities between noncitizens and US natives persisted. For example, noncitizens who wore corrective lenses had significantly lower odds than US natives of having 20/20 or better vision (AOR 0.68, 95% CI 0.48–0.97). This immigrant group also had 3.1 times higher odds than US natives of being legally blind (95% CI 1.37–6.84). There were no statistically significant differences between naturalized citizens and US natives. We also restricted the sample to immigrants only and examined length of residency among immigrants (less than 5 years vs at least 5 years), but this variable was not a statistically significant predictor of vision outcomes.
Results
Descriptive statistics of respondents’ visual impairment, vision care utilization, and sociodemographic characteristics are presented in Table 1 . Of the 17 616 respondents in our sample, 81.5% (14 366) were US natives, 7.1% (1253) were naturalized citizens, and 11.4% (1997) were noncitizen residents. There were significant differences in myopia, hyperopia, astigmatism, and visual acuity between the 3 groups ( P < .01). US natives had the highest prevalence of myopia (30.7%) and astigmatism (30.9%) among those groups, but they also had the best visual acuity with correction (60.8% had 20/20 vision and 93.7% had 20/40 or better vision). Noncitizens had almost 3 times the prevalence of legal blindness as US natives. We also examined trends over time, but they were not statistically significant and thus are not reported.
