Social Determinants of Uncorrected Distance and Near Visual Impairment in an Older Adult Population

Abstract

Purpose: Uncorrected visual impairment (VI) significantly impacts life quality and exacerbates age-related health issues. Social determinants of health (SDOH) are associated with uncorrected VI, but quantitative evidence is limited. This study investigated the link between SDOH and uncorrected VI among aging adults to identify disparities and improve vision care.

Methods: We used data from the Atherosclerosis Risk in Communities (ARIC) study visits 4 and 6 and the ancillary Eye Determinants of Cognition (EyeDOC) study. We included subjects who were >70 years old and extracted their sex, race, residence, household income, education level, having an eye doctor, health insurance status, and Area Deprivation Index (ADI) and vision outcomes. Uncorrected VI was categorized into uncorrected distance (UDVI) or near visual impairment (UNVI). Associations between SDOH indicators and VI were evaluated using logistic regressions.

Results: Among 967 adults (mean ± SD age, 78.6 ± 4.35 years; 37.9% male), UDVI was found in 293 and UNVI in 186. Living in Jackson, MS, was associated with lower odds for UNVI (adjusted odds ratio [aOR] = 0.36; 95% CI, 0.20–0.65). Higher odds for UNVI were associated with male sex (aOR = 2.01; 95% CI, 1.41–2.87), low educational attainment (aOR for not completing high school = 2.32; 95% CI, 1.37–3.92; aOR for high school only = 1.92; 95% CI, 1.26–2.92), no eye doctor (aOR = 1.58; 95% CI, 1.05–2.39), and having government health insurance only (aOR = 1.48; 95% CI, 1.00–2.17). Associations between SDOH factors and UDVI were weaker or non-existent.

Conclusions: This study links SDOH factors to uncorrected VI among older adults.

Translational Relevance: SDOH should be considered when designing interventions to reduce VI in vulnerable communities.

Introduction

It is estimated that over 550 million people globally suffer from distance visual impairment (VI) and 510 million people are living with near VI.¹ Although uncorrected refractive error is the most common reason for distance VI, uncorrected presbyopia is the leading cause of near VI. Given that the burden of VI is the greatest in those 50 years and older, these estimates are predicted to increase significantly by 2050 as the population ages.^2,3 In the United States, 5.0% of the population 40 years or above have inadequately corrected refractive error.⁴ Older adults may face unique challenges due to both near or distance VI from uncorrected refractive error and near VI from uncorrected presbyopia, resulting in functional impairment in social settings, reduced quality of life, and exacerbation of age-related declines in health.^5,6 

Social determinants of health (SDOH) are defined as “conditions in the environments where people are born, live, learn, work, play, worship, and age that affect a wide range of health, functioning, and quality-of-life outcomes and risks.”⁷ According to the Healthy People 2030 framework, domains of SDOH are comprised of “education access and quality,” “economic stability,” “social and community context,” “neighborhood and built environment,” and “healthcare access and quality.”⁸ SDOH have been examined in past studies for their influence in vision and eye health outcomes. Adults with low educational attainment are less likely to undergo annual eye exams and to be able to afford eyeglasses,^9,10 whereas those with high educational attainment are more likely to seek professional eye care.¹¹ Economic stability, reflected in factors such as household income, transportation availability (especially for older adults requiring caregiver accompaniment or wheelchair-accessible vehicles to attend medical appointments), and insurance status, plays a critical role in determining access to vision care. Individuals facing economic instability often experience substantial barriers to accessing eyecare services.¹² Community-level SDOH factors exert an indirect influence on individual socioeconomic status and directly affect access to community resources.^13,14 Vision health disparities can be attributed, in part, to variations in SDOH factors across demographic groups and their impact on health-seeking behaviors and access to healthcare.¹⁵ 

A growing number of studies have examined the impact of SDOH on the prevalence and severity of VI^16,17; however little research has focused on older adults, a demographic that is expanding and may experience unique social drivers for uncorrected VI. For example, many older adults live alone,¹⁸ live on fixed incomes, and have comorbidity and polypharmacy burdens that may contribute to difficulty in obtaining vision care.^19,20 Perceptions that refractive error services are too expensive and the belief that declining visual acuity is normal with aging may contribute to deprioritize vision health compared to other health concerns.^21,22 Existing literature does not characterize well how the different domains of SDOH drive uncorrected VI risk in aging adults. 

Here, we examine the contribution of SDOH factors to the prevalence of both distance and near uncorrected VI among adults 70 years and older. Our study drew upon data from the Eye Determinants of Cognition (EyeDOC) study, a study nested in the Atherosclerosis Risk in Communities (ARIC) Study cohort. 

Methods

Study Design and Inclusion/Exclusion Criteria

The ARIC study was established between 1987 and 1989 among men and women ages 45 to 64 years, recruited from four U.S. communities: Washington County, MD; suburbs of Minneapolis, MN; Forsyth County, NC; and the city of Jackson, MS.²³ Study participants were randomly selected from a defined population in their community. They have been followed continuously through repeat clinical examinations, active surveillance of hospitalizations, and annual (semiannual since 2012) telephone interviews. The ARIC cohort follow-up is ongoing. Since its initiation, a comprehensive database, including demographics, social, and medical information for all participants, has been developed. The study population includes ARIC cohort members who participated in the ancillary EyeDOC assessment of vision health, which was conducted between 2017 and 2019 among ARIC participants from the Jackson, MS, and Washington County, MD, field centers.²⁴ The neighboring regions of these two field centers were distinct in their built environments. For example, Jackson, the capital of Mississippi, is an urban-based area characterized by greater population density, better walkability, and a higher percentage of “zero car households” compared to the more rural Washington County in Maryland.²⁵ 

Information on sex, self-reported race, place of residence, and educational attainment was determined at ARIC visit 1 and extracted from the ARIC visit 4 database (conducted between 1996 and 1998). Age, household income, and health insurance status were obtained from ARIC visit 6 (conducted between 2016 and 2017). The Area Deprivation Index (ADI) information was extracted from either ARIC visit 4 or visit 6 for the latest information before the EyeDOC study (Fig. 1). Participants missing any SDOH factors, age, or sex, as well as those who did not complete ARIC visit 6, were excluded from the analyses (n = 98). Non-white participants in the Washington County field center were also excluded due to small numbers (n = 8). Of note, ARIC recruitment in Jackson was restricted to black participants. 

Figure 1.

 

Flow chart of study participants.

SDOH Factors

The SDOH factors analyzed in this study included residence (Jackson or Washington County), annual household income (below $24,999, $25,000 to $49,999, or above $50,000), educational attainment (high school non-completion, high school completion, or college and above), having an eye doctor or not, insurance coverage (1, having private insurance regardless of having government insurance or not; 2, having only government-sponsored insurance, including Medicare, Medicaid, Veterans Health Administration health care, TRICARE, CHAMPVA, state-sponsored health plan, or other government health care programs; or 3, having no insurance), and community resource access as measured by the ADI. The ADI is a continuous variable that represents national percentile rankings of disadvantage, ranging from 1 to 100, where 1 indicates the lowest level of disadvantage and 100 indicates the highest level of disadvantage.²⁶ The ADI was extracted based on participants’ home addresses during ARIC visit 4 or visit 6. Participants were categorized into four groups based on the ADI percentiles: <40, 41 to 60, 61 to 80, and 81 to 100. The original question stems in the ARIC or EyeDOC studies pertinent to the SDOH above are shown in Supplementary Table S1. 

These factors were chosen to cover all domains of SDOH: education access and quality (education attainment), economic stability (annual household income), social and community context (ADI), neighborhood and built environment (Jackson or Washington County), and healthcare access and quality (having an eye doctor or not, insurance coverage). 

Definition of Uncorrected Visual Impairment

Presenting distance visual acuity was measured in each eye using the retro-illuminated Early Treatment Diabetic Retinopathy Study (ETDRS) charts (Precision Vision, La Salle, IL). Participants were asked to wear their habitual corrective lenses during the assessment (if any), and the total number of letters correctly read was documented. Presenting near visual acuity was measured binocularly using MNRead charts (Precision Vision) with habitual reading lenses, if any. Corrected distance visual acuity, a metric of distance vision optimally corrected for refractive error, was measured in each eye during the EyeDOC study by using the built-in Snellen charts from the autorefractor used: KR-800S (Topcon, Tokyo, Japan) for Jackson, MS, and ARK 560A (NIDEK, San Jose, CA) for Washington County, MD.²⁷ Presenting distance, near, and corrected distance visual acuities were converted to logarithm of the minimum angle of resolution (logMAR) values for the analysis. 

The primary outcome of our study was the presence of uncorrected VI, which was classified into two categories: uncorrected distance visual acuity impairment (UDVI) and uncorrected near visual acuity impairment (UNVI). UDVI was defined as a distance presenting visual acuity of 20/40 or worse on the Snellen scale (equivalent to 0.3 or greater on the logMAR scale) with a corrected distance visual acuity better than 20/40 on the Snellen scale (equivalent to less than 0.3 on the logMAR scale) from the better eye. UNVI was defined as near presenting visual acuity 20/40 or worse on the Snellen scale (equivalent to 0.3 or greater on the logMAR scale) where corrected distance visual acuity was better than 20/40 on the Snellen scale (equivalent to less than 0.3 on the logMAR scale) from the better eye. The better eye was chosen as it better captures the impact of uncorrected VI on functioning.²⁸ 

Other Covariates

Other variables in the analysis included sex (male, female) and age at the EyeDOC visit (71–74, 75–79, 80–84, and 85 years old or above). Race was not included in the regression models due to the perfect collinearity between race and place of residence in our study after exclusion of non-white participants in Washington County. As such, we refer to a single variable of residence hereafter, which captures various contextual and community-level differences between study sites, as well as aspects of structural racism and bias underlying race categories. 

Statistical Analysis

Descriptive statistics stratified by age were used to summarize the demographics and prevalence of UDVI and UNVI of the two population samples of Jackson and Washington County participants. In the overall sample, we used univariate logistic regressions to explore the marginal association between each SDOH factor and UDVI or UNVI. To assess the independent effects of different SDOH factors, we regressed each outcome (UDVI or UNVI) on community, sex, age, annual household income, education level, having an eye doctor, health insurance coverage status (with private health insurance, with only government health insurance, or without health insurance), and ADI category. The reference groups were defined as the most conceptually resourced groups in each category when applicable: Washington County (community), female (sex), ages 71 to 74 years (age), income ≥ $50,000 (annual household income), college and above education (education level), having an eye doctor (with/without an eye doctor), having private health insurance (insurance coverage), and ADI < 40 percentile (ADI category). We used the generalized variance inflation factors (GVIFs) to assess collinearity among the variables in each model. The GVIF values were low, indicated minimal collinearity. 

The overall sample included individuals with normal distance or near visual acuity on naked eyes; however, a person was considered at risk for “uncorrected” VI if they exhibited distance or near VI without corrective lenses. To better evaluate the associations, we conducted additional analyses focusing on subsets that included only participants at risk for either UDVI or UNVI. Specifically, individuals who presented with distance or near visual acuity of 20/40 or worse that could be corrected or who wore corrective lenses during the test were classified as at risk for UDVI or UNVI and included in their subset analysis. 

Given the extreme distribution of races and distinct environment builds across the two communities, we performed stratified analyses examining participants in each community, in addition to the overall analysis. Statistical significance was defined as a two-sided P < 0.05 for the odds ratio (OR). All analyses were conducted using R 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria). 

Results

Characteristics of the Study Sample

We included a total of 967 adults: 175 who were 71 to 74 years old (37.1% male), 455 who were 75 to 79 years old (37.4% male), 223 who were 80 to 84 years old (37.2% male), and 114 who were 85 years of age or older (43.0% male). The mean age of the whole population was 78.6 ± 4.35 years, and 37.9% were male. The proportion of participants living in Jackson decreased as age increased (P = 0.012). (Table 1) Younger participants had higher annual household income (P < 0.001) and higher education levels (P = 0.036) compared to older participants. The proportion of having an eye doctor remained stable across all age groups. More than 70% of participants had government health insurance, and less than 60% had private health insurance coverage. The distribution of ADI National Rank categories was similar across different age groups. Regarding visual function and eye conditions, 85% to 90% of participants were considered at risk of UDVI, and 77% to 81% of participants were considered at risk of UNVI across different age groups. Of note, the prevalence of self-reported cataracts (including current or past cataract and those who underwent cataract surgery) increased with age, from 77.1% among those 71 to 74 years old to 91.2% among those 85 years or older (P = 0.003) (Table 1). 

Table 1.

 

Demographic Characteristics of the Study Population

 

Prevalence of UDVI and UNVI

Of the 967 participants, UDVI was found in 293 (30.3%), and UNVI was found in 186 (19.2%). The prevalence of UDVI was 34.3% (293 out of 853) among participants who were at risk of UDVI and 24.4% (186 out of 761) among participants who were at risk of UNVI. The UDVI rates were higher in males (27.7%, 25.3%, 44.6%, and 32.7% among the 71–74, 75–79, 80–84, and 85 years or above age groups, respectively) compared to females (24.5%, 30.5%, 32.9%, and 29.2%, respectively) across all age groups except for the 75 to 79 years group. (Fig. 2) Similarly, UNVI was more common across all age groups in males (26.2%, 25.9%, 24.1%, and 32.7%, respectively) than in females (13.6%, 11.2%, 22.1%, and 16.9%, respectively). 

Figure 2.

 

Prevalence of uncorrected VI across different age groups stratified by sex.

SDOH Factors Associated With UDVI

Overall Dataset Analysis

Results from models estimating the association between SDOH factors and UDVI are shown in Table 2. In univariate analyses, we found that participants 80 years or older had higher odds for UDVI compared to those 71 to 74 years old (OR = 1.71; 95% confidence interval [CI], 1.11–2.64 for participants 80–84 years old). There was a borderline positive association for participants not completing high school and a greater risk of UDVI than those with college and above education (OR = 1.42; 95% CI, 0.96–2.09). Likewise, there was a borderline-positive association for participants with government health insurance only and increased odds of UDVI when compared to those with private health insurance (OR = 1.32; 95% CI, 0.98–1.77). Living in Jackson, male sex, low annual household income, not having an eye doctor, or a low ADI were not associated with the odds of UDVI. After adjusting for other SDOH factors, the associations between UDVI and individuals 80 to 84 years old remained significant. However, the associations of UDVI with education attainment and health insurance status became non-significant. 

Table 2.

 

Univariable and Multivariable Logistic Regressions Among SDOH Factors for UDVI (Overall Sample)

 

Subset Analysis

The result of subset analysis is summarized in Table 3. Mostly similar findings were observed in the subset of participants who were at risk of UDVI. Of note, the OR for UDVI among participants living in Jackson compared to Washington County was 1.57 overall and was 1.88 (95% CI, 1.16–3.05) among participants 80 to 84 years of age. 

SDOH Factors Associated With UNVI

Overall Dataset Analysis

Results from models estimating the association between SDOH factors and UNVI are shown in Table 4. In the univariable model, we found that living in Jackson was associated with lower odds for UNVI (OR = 0.35; 95% CI, 0.25–0.50). Males had significantly higher risks of UNVI compared to females (unadjusted OR = 2.06; 95% CI, 1.49–2.85). There were significantly higher ORs for participants with lower education levels; the unadjusted OR for participants not completing high school was 2.70 (95% CI, 1.71–4.27), and the unadjusted OR for participants completing high school but no college degree was 2.25 (95% CI, 1.54–3.28). Older age, lower annual household income, not having an eye doctor, or no private health insurance coverage were not associated with increased odds for UNVI. 

Table 3.

 

Multivariable Logistic Regressions Among SDOH Factors for UDVI and UNVI (Subset Samples)

 

After adjusting for other SDOH factors, the OR for UNVI among participants living in Jackson or male sex remained consistent; the adjusted OR (aOR) for participants living in Jackson: 0.36 (95% CI, 0.20–0.65), and aOR for male sex was 2.01 (95% CI, 1.41–2.87). The ORs among participants with lower education levels attenuated slightly: aOR for participants not completing high school was 2.32 (95% CI, 1.37–3.92), and aOR for participants completing high school but no college degree was 1.92 (95% CI, 1.26–2.92). Of note, not having an eye doctor became significantly associated with a higher odds for UNVI (aOR = 1.58; 95% CI, 1.05–2.39). In addition, participants with government health insurance had increased OR only after adjustment (aOR = 1.48; 95% CI, 1.00–2.17). Participants without health insurance were also at greater risk of UNVI, although the results were not statistically significant. 

Subset Analysis

The result of subset analysis is summarized in Table 3. The relationship of not having an eye doctor (OR = 1.80; 95% CI, 1.17–2.77) versus UNVI was further strengthened through subset analysis. Furthermore, increases in the odds ratios for developing UNVI were observed when comparing individuals with private insurance to those with government insurance only (OR = 1.57; 95% CI, 1.05–2.35) or no insurance at all (OR = 1.91; 95% CI, 0.95–3.84) (Table 4). 

Table 4.

 

Univariable and Multivariable Logistic Regressions Among SDOH Factors for UNVI (Overall Sample)

 

Stratified Analyses by Community

The results of the stratified analyses are presented in Supplementary Tables S2 and S3. The association between male sex and UNVI was primarily observed in Washington County. The association between educational attainment and UNVI appeared in both Jackson and Washington County, whereas the association between having an eye doctor and UNVI was evident only in Washington County. 

Discussion

This research revealed SDOH factors that contributed to the risks of uncorrected VI among older adults residing in two U.S. communities of contextual distinction. Overall, the prevalence of uncorrected VI was high in this population but varied between the two communities, with Jackson having a significantly lower prevalence of UNVI. In addition, we found that males were at higher risk of having UNVI than females. Low education attainment, not having an eye doctor, or lack of private health insurance were also positively associated with UNVI. However, the associations between SDOH factors and UDVI were either considerably weaker or non-existent, with a significantly higher risk of UDVI observed only in participants 80 to 84 years of age compared to those 71 to 74 years old. 

Our findings underscored the importance of educational attainment in relation to visual and overall health outcomes in aging populations. The association between lower educational attainment with higher odds of UNVI was apparent not only in the overall and subset analysis but also in each community in the stratified analysis. This demonstrated the relevance of education attainment across different community settings. Previous studies in the United States have indicated that higher education levels are correlated with improved eye care and decreased VI.²⁹ Education is often linked to health-conscious behavior, including attention to visual conditions,³⁰ and is a surrogate for health literacy.³¹ Among older adult patients, higher education levels correspond to better healthcare compliance, as those with lower educational attainment often require tailored recommendations to understand their treatment plans.³² Furthermore, higher education leads to improved job prospects during the working years, providing individuals with greater financial stability and support for their health needs as they age.³³ 

Not unexpectedly, we found that having an eye doctor reduced the risk of UNVI. The community-stratified analysis revealed that this relationship was present only in Washington County and not in Jackson. An established relationship with an eye doctor enhances personal motivation and independent actions that promote health consciousness and address visual conditions—key components of healthcare quality.³⁴ In suburban or rural areas such as Washington County where eyecare services may be less accessible, having an established eye doctor may be even more critical for patients to obtain timely vision correction. 

Moreover, having either government insurance only or no insurance was associated with increased risk of UNVI in this population. Government insurance programs appear to be less flexible and sometimes do not adequately address the vision health needs of older adults. Medicare, for example, does not cover refraction, which is considered part of a routine exam and not a medical need.³⁵ Similarly, Medicare (Parts A and B) does not help pay for prescription sunglasses, contact lenses, or eyeglasses unless they are needed following cataract surgery or another medically necessary procedure.^10,36 Individuals with private health insurance tend to receive more frequent visual testing (e.g., for glaucoma) than those who rely solely on Medicare,^37,38 and they experience shorter waiting times for appointments.³⁹ 

We considered ADI as a measure of community-level socioeconomic factors. High ADI was not independently associated with the likelihood of uncorrected VI after adjusting for other variables. The relationship between community-level factors and vision health remains an underexplored area. In terms of relationships with overall health, neighborhood deprivation has been linked with poor health outcomes, likely capturing health disparities that result from neighborhood differences in education quality or economic stability,⁴⁰ though greater access to government-funded health care programs, including eyecare services, can make relationships with indicators of community level factors complicated. Community also served as an indicator of the neighborhood context. Similarly, we failed to identify an association between household income and uncorrected VI. Older adults in Jackson exhibited higher odds of UDVI and significantly lower odds of UNVI compared to older adults in Washington County; however, this disparity could also be attributed to the differing racial compositions of the two communities. Elam et al.⁴¹ highlighted various strategies to expand access to refractive correction services in communities to reduce vision health disparities, such as including vision in chronic disease metrics, incorporating vision into health care in lower socioeconomic status communities, and providing vision health messaging in general health. 

Possible explanations for the lack of observed associations between most SDOH factors and UDVI could be linked to differences in the development, progression, and access to correction for UDVI and UNVI. Most UDVI conditions (e.g., refractive error including myopia) develop during childhood but tend to stabilize before age 30 years.⁴² In contrast, UNVI conditions (e.g., presbyopia) are more likely to worsen with age.³ Older individuals may adapt to the distance VI and become less likely to seek treatment, causing unconscious functional decline.⁴³ As individuals adjust to chronic illnesses in addition to the mobility limitations experienced by many older adults, differences in SDOH could become less important.⁴⁴ The prevalence differences of the two types of VI across various age groups in our cohort (Fig. 2) might also contribute to this discrepancy. 

There are several strengths of this study. To our knowledge, this is the first study to explore the association of SDOH factors with both uncorrected near and distance VI among older adults. Data from the ARIC and EyeDOC studies allowed us to explore indicators of SDOH across several domains. Further, the EyeDOC study included participants from Jackson and Washington County field centers of the ARIC study, reflecting the diversity of aging communities across the United States. Our study highlights the need for researchers and implementation scientists to consider various social drivers when designing and implementing interventions to reduce the burden of uncorrected VI in aging adult communities. For example, older adults with lower educational attainment face greater risks for having uncorrected VI; thus, public health initiatives targeting this community may have to promote both increased awareness of visual health as well as regular screenings. Additionally, insurance programs should consider expanding coverage options and increasing flexibility in visual health services, including corrective lenses for older adults. 

We acknowledge the limitations of the cross-sectional nature of this study, which constrains our ability to infer causality. The SDOH factors available in ARIC may not be the best indicators in each SDOH domain, and unmeasured confounders are possible. For example, neighborhood and built environment refer to much more than administrative and geographic location. The built environment (traffic, air pollution, light at night, parks, etc.) may vary significantly even within the same geographic location, and other indicators for these domains (e.g., abundance and availability of medical resources, availability of support systems) should be explored in future studies. As a result, the categorization of Jackson or Washington County may not be the best representation of this domain. In addition, “having an eye doctor” was selected as an indicator of healthcare access and quality, but we did not have data directly indicating whether and how often the participants utilized these eyecare services. Finally, the lack of association between annual household income and either UDVI or UNVI outcomes may suggest that these metrics were not the most suitable indicators for measuring “economic stability” and “social and community context,” and alternative indicators could be examined in future work. Nevertheless, the findings of this study provide important insights into the potential role of several SDOH factors in the development of uncorrected VI. 

Conclusions

This study provides new insights into the association of SDOH factors with uncorrected VI among older adults using data from ARIC study and EyeDOC study. Our findings highlight the need for researchers and implementation scientists to consider social drivers when testing and implementing interventions to reduce the burden of correctable VI in aging adult communities. 

Acknowledgments

The authors thank the staff and participants of the ARIC study and the ancillary EyeDOC study for their important contributions. 

Supported by grants to the ARIC study from National Heart, Lung, and Blood Institute, Grant/Award Numbers: HHSN2682017, HHSN268201700001I, HHSN268201700002I, R01-HL70825, HHSN268201700004I, HHSN268201700005I, HHSN268201700003I; National Institutes of Health, Grant/Award Numbers: 2U01HL096814, 2U01HL096902, 2U01HL096812, 2U01HL096899, 2U01HL096917; and a grant to the EyeDOC study from the National Institute on Aging, Grant/Award Number: 1R01AG052412. 

Disclosure: P.-J. Lin, None; A.G. Abraham, None; P. Ramulu, None; A. Mihailovic, None; A. Kucharska-Newton, None; X. Guo, None