Key Points

QuestionÌý What is the burden of uncorrected refractive error and unmet refractive need in African American individuals?

FindingsÌý This population-based cross-sectional study including 6337 African American individuals documented a prevalence of approximately 1 of 7 individuals with an uncorrected refractive error and confirmed approximately 1 of 20 individuals with an unmet refractive need (5.4%).

MeaningÌý These data suggest that with the relatively low cost and easily implementable intervention of refractive correction (eg, eyeglasses), it is possible to reduce over two-thirds of visual impairment in African American individuals, potentially leading to a reduction in morbidity and improvement in productivity and quality of life.

ImportanceÌý Refractive error remains the largest cause of correctable visual impairment in the US. Correction of refractive error will reduce visual impairment and its associated morbidity but also improve quality of life and productivity.

ObjectiveÌý To determine the burden of and risk factors (RFs) associated with any uncorrected refractive error (UCRE) and unmet refractive need (URN) in a population-based sample of African American adults.

Design, Setting, and ParticipantsÌý This cross-sectional study, conducted from April 2014 to April 2018, included a population-based sample of self-identified African American participants 40 years and older from 30 contiguous census tracts in Inglewood, California. Participants underwent a complete ophthalmic examination and an in-home–administered questionnaire to assess sociodemographic, lifestyle, biological, medical, and health care and eye care usage RFs associated with UCRE and URN. Measurements of visual acuity (VA) were performed using a standard Early Treatment Diabetic Retinopathy Study protocol. Noncycloplegic automated refraction with supplemental subjective refraction was performed. UCRE was defined as an improvement of 2 or more lines with refraction in the better-seeing eye. URN was defined as an improvement of 2 or more lines with refraction in the better-seeing eye in those persons who were visually impaired. Sex- and age-specific burden of UCRE and URN were calculated, and multiple regression analyses were used to identify independent RFs. Study data were analyzed from May 2018 to December 2023.

ExposuresÌý Presence or absence of correctable refractive error.

Main Outcomes and MeasuresÌý Self-reported sex- and age-specific prevalence of and risk indicators of UCRE and URN.

ResultsÌý Of the 7957 eligible participants in the African American Eye Disease Study (AFEDS), 6347 (80%) completed both the in-home interview and the clinical examination. Of these, 6337 participants (mean [SD] age, 61 [11] years; 3997 female [63%]) with complete refractive error data were included in the analysis. Refractive error–related correctable visual impairment was present in over two-thirds of participants with visual impairment (68.7%). The overall prevalence of any UCRE was 14.6% (925 of 6337), and the overall prevalence of any URN was 5.4% (URN1 [those with presenting VA of worse than 20/40 in the better-seeing eye but who could achieve 20/40 or better with correction], 157 of 2893; URN2 [those with presenting VA of worse than 20/40 in the better-seeing eye but who could achieve an improvement of 2 or more lines with refractive correction], 155 of 2891).

Conclusions and RelevanceÌý Results of this cross-sectional study suggest a high burden of refractive error–associated correctable refractive error in African American adults, making it the leading cause of visual impairment in this population. Providing universal coverage for vision care and prescription glasses is an affordable and achievable health care intervention that could reduce the burden of visual impairment in African American adults by over two-thirds and likely raise the quality of life and work productivity, especially in this vulnerable minority population.

Visual impairment and blindness reduced the health and quality of life (QOL) of more than 5.5 million people in the US in 2020,^1,2 and the number of people impacted is projected to increase to 7 million by 2050.^1-5 Uncorrected refractive error (UCRE) remains the largest contributing cause of visual impairment in the US, and correction of refractive error will likely reduce not just the number of adults affected by visual impairment but importantly, the associated risks of falls, fractures, and depression in older persons. In addition, improvement of vision improves QOL and productivity in adults.^6-12 Although the negative impact of UCRE on morbidity and mortality has been well documented,^13-15 there continues to be limited data that address the burden of UCRE in the population.

African American individuals, the secondary largest minority group in the US, compose 13% of the US population, with approximately 2.3 million African American individuals in California.^16,17 In 2015, African American adults contributed to the second highest proportion of visual impairment cases, accounting for 15% of all visual impairment cases in the US.³ Over time, the number of African American adults experiencing visual impairment is expected to increase, and they are projected to have the largest burden of blindness by 2050. Despite this, African American individuals continue to remain underserved, contributing to a higher burden of visual impairment than non-Hispanic White individuals in the US.^18-20

To quantify the burden of UCRE and its related visual impairment in African American adults, we assessed the prevalence of any UCRE and unmet refractive need (URN) in participants of the African American Eye Study (AFEDS), a population-based study of eye disease in African American adults 40 years and older residing in Los Angeles County, California. The difference between UCRE and URN lies in the fact that any UCRE is measured in all persons with any level of presenting visual acuity (VA; including those with VA of 20/40 or better), whereas URN is measured in those persons whose presenting VA is worse than 20/40 (ie, visually impaired). The importance of assessing UCRE in persons with VA that is either 20/40 or better is that these persons can also benefit and improve their QOL and visual functioning with refractive correction.²¹ The importance of assessing URN (ie, persons with VA worse than 20/40) is that we can measure the number of people with visual impairment (based on presenting VA) who would not be visually impaired after refractive correction.

Specifically in this analysis we (1) estimated the sex- and age-specific prevalence of any UCRE and URN and (2) described the risk associations between selected socioeconomic, biological, ocular, and lifestyle risk factors (RFs) with any UCRE and URN. We also compared the burden of any UCRE and URN in this African American sample with other racial and ethnic groups.

The AFEDS, a cross-sectional study, consisted of noninstitutionalized self-identified African American individuals aged 40 years or older and residents of Inglewood, California, within Los Angeles County. Details of the study design, sampling plan, and baseline data have been described previously.²² Eligible participants were invited to complete a home interview and a comprehensive eye examination. Written informed consent was obtained from each participant during enrollment from April 1, 2014, through August 31, 2018. The study adhered to the tenets of the Declaration of Helsinki and was approved by the institutional review board/ethics committee of the Los Angeles County–University of Southern California Medical Center institutional review board. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology () reporting guidelines.

A detailed in-home interview was conducted. The interviewer collected information on sociodemographic (eg, marital status, employment, education level, annual household income level), lifestyle (smoking history: nonsmoker, ex-smoker, current smoker), and biological and medical factors (weight and height, health and vision insurance, health care and eye care use in the past 12 months). All eligible individuals were then scheduled for a comprehensive eye examination, including VA testing and refraction.

Standard Early Treatment Diabetic Retinopathy Study (ETDRS) protocols with a modified ETDRS distance chart that was transilluminated with a chart illuminator (Precision Vision)²³ were used for each AFEDS participant to measure distance VA in each eye with presenting correction (if any) at 4 m. Using different charts for each participant’s right and left eye, the examiner measured the presenting distance VA for each eye (right eye followed by left eye) with the participant’s existing refractive correction (if any). If the participant read fewer than 55 letters at 4 m in either eye, automated refraction was performed (Humphrey Autorefractor [Carl Zeiss Meditec]), followed by subjective refraction with standardized protocols. After refraction, each eye was retested to measure the best-corrected VA in that eye. If the participant was unable to read letters at 4 m, VA measurement and subjective refraction were performed at 1 m. If the participant was still unable to read the largest letters at 1 m, a semiquantitative estimate of the VA (finger counting, hand motion, light perception, or no light perception) was made. A logarithmic E chart was used for participants who were unable to read the ETDRS charts.

Any UCRE was defined as an improvement of 2 or more lines in VA (irrespective of the presenting VA) in the better-seeing eye after refraction of the same eye.²⁴ We used 2 definitions of URN. URN1 could only be measured in persons who had a presenting VA of worse than 20/40. URN1 was the proportion of participants with habitual correction who had a presenting VA of worse than 20/40 in the better-seeing eye but who could achieve 20/40 or better with correction in that same eye. This definition was used to facilitate comparisons of URN with other previously published studies.²⁵ The foundation of this definition is rooted in the VA standard needed to drive an automobile in many of the states in the US.²⁶ URN2 was defined as the proportion of participants with habitual correction who had a presenting VA of worse than 20/40 in the better-seeing eye but who could achieve an improvement of 2 or more lines with refractive correction in that same eye. This measure was developed to identify the proportion of participants who may not be identified in URN1 as they may not have achieved a VA of 20/40 or better after refraction but nevertheless did benefit from refraction as their VA improved by 2 or more lines (refer to the Discussion for additional details).

The prevalence of any UCRE was the proportion of the total number of individuals who had any uncorrected refraction. The prevalence of URN was calculated for each of the 2 definitions as the ratio of the number of individuals with unmet refractive need to the number of individuals who were visually impaired. Differences in the prevalence across age categories were tested using the Cochran-Armitage trend test.

The association of age, sex, and other risk indicators with any UCRE and URN was explored by using univariate and age-adjusted stepwise multivariable logistic regression procedures. All candidate RFs meeting the criteria of P < .05 were entered into the final model. In the multivariable logistic analyses, the missing-indicator method was used to include observations with missing values for 1 or more predictor variables. Candidate sociodemographic risk indicators included marital status (married, never married, widowed, or separated/divorced), employment status (employed, retired, or unemployed), education level (<12 years vs ≥12 years), annual household income level (<$20 000, $20 000-$40 000, or >$40 000), history of smoking (nonsmoker, ex-smoker, or current smoker), having health and vision insurance, and health care and eye care use in the past 12 months. Candidate clinical risk indicators included body mass index (normal, 24.9; overweight, 25.0-29.9; or obese, 30.0; calculated as weight in kilograms divided by height in meters squared). All analyses, conducted from May 2018 to December 2023, using SAS software, version 9.4 (SAS Institute), included 2-sided P values but were not adjusted for multiple analyses. A P value <.05 was considered statistically significant.

Of the 7957 eligible participants in AFEDS, 6347 (80%) completed both the in-home interview and clinical examination. Of the 6347 participants, 6337 (mean [SD] age, 61 [11] years; 2340 male [37%]; 3997 female [63%]) with complete refractive error data were included in the analysis.

The overall prevalence of any UCRE was 14.6% (925 of 6337). The prevalence of any UCRE was higher for men compared with women (15.9% [373 of 2340] and 13.8% [552 of 3997], respectively; P = .02) (Table 1). There was no age-associated trend in the prevalence of any UCRE (Table 1).

Of the 6337 participants, 2893 (46%) were included in the analysis of URN because they had presenting VA of worse than 20/40 in the better-seeing eye. The prevalence of URN (for URN1 and URN2) was 5.4% (URN1, 157 of 2893; URN2, 155 of 2891). The prevalence of URN was higher for men (6.9% [64 of 927]) compared with women (4.7% [93 of 1966]; P = .02) (Table 1). There was no age-associated trend in URN (Table 1). Overall, 2.4% participants (155 of 6337) would no longer be visually impaired after refraction. When separated by sex, 2.7% of the men (64 of 2340) and 2.3% of the women (93 of 3997) would no longer be visually impaired following refractive correction (data not shown).

These findings suggest that 68.7% of all participants (1987 of 2893) with visual impairment would no longer be visually impaired after refractive correction, assuming those participants would subsequently use the refractive correction habitually.

We found 5 independent risk indicators associated with a higher prevalence of any UCRE. Compared with those aged 40 to 49 years, the prevalence of any UCRE was higher for those aged 50 to 69 years (odds ratio [OR], 1.31; 95% CI, 1.04-1.66; P = .02), 60 to 69 years (OR, 1.60; 95% CI, 1.26-2.03; P < .001), 70 to 79 years (OR, 1.49; 95% CI, 1.13-2.00; P = .005), and 80 years and older (OR, 1.61; 95% CI, 1.11-2.33; P = .01) (Table 2). Higher prevalence of any UCRE was also associated with those participants in the lowest category of annual household income (<$20 000: OR, 1.94; 95% CI, 1.54-2.40; P < .001), those who were current smokers (OR, 1.54; 95% CI, 1.27-1.86; P < .001), those who reported no history of having had an eye examination within the past 12 months (OR, 1.41; 95% CI, 1.19-1.67; P < .001) (Table 2), and those who reported that they did not have vision care insurance (OR, 1.31; 95% CI, 1.11-1.54; P = .001) (Table 2). Univariate associations of potential RFs are presented in the eTable in Supplement 1.

Risk indicators associated with a higher prevalence of URN1 and URN2 were those with low/intermediate category of annual household income, including income less than $20 000 (URN1: OR, 5.24; 95% CI, 3.16-8.69; P < .001; URN2: OR, 7.02; 95% CI, 4.00-12.30; P < .001), income between $20 000 and $40 000 (URN1: OR, 2.01; 95% CI, 1.10-3.67; P = .02; URN2: OR, 2.60; 95% CI, 1.35-5.03; P = .004) (Table 2), and those who did not have vision care insurance (URN1: OR, 1.53; 95% CI, 1.01-2.29; P = .02; URN2: OR, 1.62; 95% CI, 1.11-2.37; P = .01) (Table 2). The prevalence of URN was lower among those participants who had overweight (URN1: OR, 0.45; 95% CI, 0.30-0.68; P <.001; URN2: OR, 0.47; 95% CI, 0.31-0.71; P <.001) or obesity (URN1: OR, 0.36; 95% CI, 0.24-0.54; P < .001; URN2: OR, 0.38; 95% CI, 0.25-0.58; P < .001) (Table 2) compared with those who did not.

Current data on any UCRE in the general US population are limited, including data in African American people, a population that makes up 13% of the entire US population. In our sample of 6337 African American participants 40 years and older, the prevalence of any UCRE was high at 14.6%. Two previous population-based studies (the Los Angeles Latino Eye Study [LALES] and the Chinese American Eye Study [CHES]) have provided data on the prevalence of any UCRE in detail in Latino individuals and Chinese American individuals in the US. Compared with any UCRE prevalence in African American participants in the AFEDS study, Latino individuals in the LALES study and Chinese American individuals in the CHES study had a higher prevalence (14.6% vs 5.1% vs 17.9%, respectively).^27-29 This difference may be attributed to the slightly older population in the LALES and CHES studies compared with the AFEDS trial.

When comparing risk indicators for any UCRE, both studies (LALES and AFEDS) had common risk indicators: older age (50 years and older), lower annual household income ($20 000-$40 000), lack of vision insurance, and not having had an eye examination within the past 12 months. In addition, Latino individuals in the LALES study were also more likely to have any UCRE if they were unemployed, never married, had low acculturation scores, had less than high school education, and had been born outside the US. It is clear from both of these studies that a lack of access to vision correction is responsible for the higher burden of any UCRE. Furthermore, both African American and Latino participants who had overweight or obesity were less likely to have any UCRE compared with those who did not have overweight or obesity. One explanation for this reduced burden is that since individuals with overweight or obesity likely had additional comorbidities such as diabetes, high blood pressure, and joint disease, they would be more likely to access the health care system and obtain vision care and vision correction in addition to general medical care. One reason for this is that internal medicine/family practice physicians and endocrinologists are encouraged and incentivized by the Centers for Medicare and Medicaid Services to obtain annual diabetic eye examinations on all patients with diabetes as part of the Clinical Quality Measures of the Merit-Based Incentive Payment System.

Unmet refractive need in African American participants in the AFEDS study was lower (5.4%) than that in Latino individuals in the LALES study (8.9%) and Chinese American participants (12.9%) (Table 3).^22,28-31 Two other studies (Baltimore Eye Survey³⁰ and the Proyecto Ver Study³¹ conducted in 1985-1988 and 1999-2000, respectively) have provided data on URN in the US. In the Baltimore Eye Survey, the prevalence of URN was 6.5% in African American individuals and 7.83% in White individuals³⁰ (Table 3). In the Proyecto Ver study,³¹ the prevalence of URN in Hispanic individuals was 8.0% (Table 3). Clearly, the burden of URN in African American participants is lower in the AFEDS study (5.4%) compared with African American individuals in the Baltimore Eye Survey (6.5%). This lower burden of URN in African American individuals between participants in the AFEDS study compared with the Baltimore Eye Survey could be attributed to improved vision care services and perhaps better understanding of the benefits of vision correction over the past 3 decades between when the 2 studies were conducted (Baltimore Eye Survey, 1985-1988, and AFEDS, 2014-2018). In the LALES and CHES studies, the burden of URN was higher; hence, the participants were older than participants in the AFEDS study. One explanation may be a higher prevalence of cataracts in older Latino individuals and Chinese American individuals.

We developed the measure URN2 because, based on previous data, a 2-line improvement in VA was associated with a significant improvement in vision-related QOL (vision-specific function).³² Although the overall prevalence of URN is similar using the 2 definitions, the prevalence of URN2 was consistently higher than URN1 in all age groups except those 80 years and older, signifying a higher sensitivity in identifying people who may benefit from refraction when using the definition for URN2. The persons identified by URN2 are at least partially different from those identified in URN1 because the individuals identified in URN2 also include those who, despite an improvement of 2 or more lines after refraction, may not be able to achieve a corrected VA of 20/40 or better. In essence, URN2 identified those with worse-presenting or best-corrected VA who may have substantial ocular morbidity from other ocular conditions such as diabetic macular edema and age-related macular degeneration but could nevertheless benefit from refraction than those identified by URN1. Compared with URN2, individuals identified by URN1 were those aged 80 years and older who most likely had lens opacities that could be corrected to better than 20/40 VA or those who did not have any other ocular disease that would limit their best-corrected VA.

In the AFEDS study, the independent risk indicators for URN (both definitions) were similar to those for UCRE (annual household income <$20 000, $20 000-$40 000, and lack of vision insurance). Similar RFs were identified for URN in the LALES study.²⁶ However, in the Proyecto Ver study³¹ older age, less than 13 years of education, low acculturation, lack of insurance coverage, and not having been evaluated by an eye care practitioner were independent RFs. Similar to the RFs for UCRE, lack of access to vision care contributed to a greater burden of URN.

Similar to UCRE, obesity and overweight contributed to a lower burden of unmet refractive error. This is likely due to more frequent visits to health care and vision care practitioners.

These RFs highlight the need to consider universal vision coverage particularly in the more vulnerable segments of our society—older individuals with lower income and no health insurance.³³ The lack of resources in these vulnerable people possibly contributes to a higher burden of visual impairment that makes them less likely to be productive and contributing members of society.³³ Furthermore, visual impairment in older individuals contributes to an increase in injuries and falls, which leads to higher morbidity and mortality.³⁴ Provision of vision correction (with glasses/contact lenses) is also likely to reduce morbidity and mortality in older individuals. Furthermore, it could contribute to a higher QOL and likely could reduce the cost of care both personal and societal,³⁵ assuming that the individuals would use the vision correction device. The provision of universal vision correction was highlighted in a recent report from the National Academies (Institute of Medicine), which called for eliminating correctable vision impairment in the US by 2030.³³

These findings support coordinated efforts in the public and private sector, at the federal, state, and local levels, to expand public health capabilities. Based on this study, we recommend programs and payment policies that provide access to appropriate screenings, comprehensive eye examinations, and follow-up care (including coverage of corrective lenses). Such policies and programs that promote overall population eye health are likely to reduce and potentially eliminate correctable vision impairment. Furthermore, if left unchecked, this problem is likely to grow, given the aging and increased life expectancy of the population.

Given our participation rate was 80%, which is comparable to other population-based studies in the US, and our participants’ socioeconomic status and demographic characteristics are similar to those of African American individuals in Los Angeles County, California, and the US,²² we are able to generalize our findings to other African American adults in the US. However, this study does not have any ancestry admixture data, which are needed for a better understanding of the burden of eye disease and refractive error in this population.

Our study has some limitations. For instance, because we did not perform cycloplegic refraction, some latent myopia may not have been uncovered; however, because our participants were 40 years and older, they were more likely to have presbyopia, and it is unlikely that cycloplegic refraction would uncover substantial amount of latent myopia. Another potential limitation is the higher number of female participants compared with male participants in our sample. However, sex was not an independent RF; thus, it is unlikely to impact our findings.

Results of this cross-sectional AFEDS trial suggest that there were approximately 1 of 7 individuals with an uncorrected refractive error and approximately 1 of 20 individuals with an unmet refractive need (5.4%) in a sample of African American individuals. These data suggest that given the relatively inexpensive and easy intervention of refraction and provision of refractive correction, it may be possible to reduce this leading cause of visual impairment in this vulnerable segment of the US population, assuming these individuals subsequently used the refractive correction habitually.^36-39

Accepted for Publication: November 20, 2023.

Published Online: February 22, 2024. doi:10.1001/jamaophthalmol.2023.6781

Corresponding Author: Rohit Varma, MD, MPH, Southern California Eye Institute, CHA Hollywood Presbyterian Medical Center, 1300 N Vermont Ave, Ste 101, Los Angeles, CA 90027 (rvarma@sceyes.org).

Author Contributions: Drs Varma and McKean-Cowdin had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Dhablania, Torres, McKean-Cowdin, Varma.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Dhablania, Torres, Varma.

Critical review of the manuscript for important intellectual content: All authors.

Statistical analysis: Dhablania, Burkemper.

Obtained funding: McKean-Cowdin, Varma.

Administrative, technical, or material support: Torres, McKean-Cowdin, Varma.

Supervision: Torres, Varma.

Conflict of Interest Disclosures: Drs Burkemper, McKean-Cowdin, and Varma reported receiving grants from the National Eye Institute during the conduct of the study. No other disclosures were reported.

Funding/Support: This work was supported by grant U10 EY023575 from the National Eye Institute.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: Members of the African American Eye Disease Study Group appear in Supplement 2.

Meeting Presentation: This study was presented at the Association for Research in Vision and Ophthalmology Meeting; May 4, 2022; Denver, Colorado.

Data Sharing Statement: See Supplement 3.