Abstract
Purpose:
Assess rate of visual impairment (VI) from uncorrected refractive error (URE) and associations with demographic and socioeconomic factors among low-income patients presenting to the Michigan Screening and Intervention for Glaucoma and eye Health through Telemedicine (MI-SIGHT) program.
Design:
Cross-sectional study.
Participants:
Adults ≥ 18 years without acute ocular symptoms.
Methods:
MI-SIGHT Program participants received a telemedicine-based eye disease screening and ordered glasses through an online optical shop. Participants were categorized based on refractive error (RE) status: VI from URE (presenting visual acuity [PVA] ≤20/50, best corrected visual acuity [BCVA] ≥20/40), URE without VI (PVA ≥20/40, had ≥2 lines of improvement to BCVA), and no or adequately corrected RE (PVA ≥20/40, <2 line improvement to BCVA). Patient demographics, self-reported visual function, and satisfaction with glasses obtained through the program were compared between groups using analysis of variance (ANOVA), Kruskal-Wallis (KW), Chi-square, and Fisher exact testing.
Main Outcome Measures:
PVA, BCVA, and presence of VI (defined as PVA ≤20/50)
Results:
Of 1171 participants enrolled in the MI-SIGHT program during the first year, average age was 55.1 years old (SD=14.5), 37.7% were male, 54.1% identified as Black, and 1166 (99.6%) had both PVA and BCVA measured. VI was observed in 120 (10.3%); 96 had VI from URE (8.2%), 168 (14.4%) had URE without VI, and 878 (75.3%) had no or adequately corrected RE. A smaller percentage of participants with VI from URE reported having a college degree and a larger percentage reported income <$10,000 compared to participants with no or adequately corrected RE (3.2% versus 14.2%, p=0.02; 45.5% versus 21.6%, respectively, p<0.0001. Visual function was lowest among participants with VI from URE, followed by those with URE without VI, and then those with no or adequately corrected RE (VFQ9 composite score 67.3±19.6 versus 77.0±14.4 versus 82.2±13.3, respectively; p<0.0001). 71.2% (n=830) ordered glasses for an average cost of $36.80±$32.60; 97.7% were satisfied with their glasses.
Conclusions:
URE was the main cause of VI at two clinics serving low-income communities and was associated with reduced vision related quality of life. An online optical shop with lower prices made eyeglasses accessible to low-income patients.
Uncorrected refractive error (URE) is the main cause of visual impairment (VI) in the United States (US), contributing to 80% of VI.1 Approximately 5.5% of US adults aged 40 and older have VI due to URE,2 and by 2050, 16.4 million people in the US are expected to have VI from URE. VI is associated with numerous adverse patient outcomes, including decreased quality of life,3 injurious falls,4 depression,5 dementia,6 motor vehicle accidents,7 and mortality.8 URE has substantial societal impacts, including decreased workplace productivity.9,10
Unlike other causes of VI such as glaucoma or macular degeneration, VI from URE can be effectively treated with a simple pair of eyeglasses. Yet URE remains the leading cause of VI because of limited access to these sight-restoring devices. URE in the US is associated with non-White race, low household income, limited educational attainment, and lack of health insurance.11–14 Many patients with low-income report eyeglasses are unaffordable, and working age adults 45–64 years old are the most likely to have difficulty affording glasses compared to young adults or older adults.15 Although there are numerous studies describing programs to provide affordable eyeglasses to children,16,17 there are few reports of programs addressing URE among US adults.
We aimed to assess rates of VI from URE and associations with demographic and socioeconomic factors among patients presenting to two clinics participating in the Michigan Screening and Intervention for Glaucoma and Eye Health through Telemedicine (MI SIGHT) program. Both are primary care clinics that serve Michigan communities with high rates of poverty. In the MI-SIGHT telemedicine program, ophthalmic technicians in the two community clinics take an ophthalmic history and measure manifest refraction, intraocular pressure, and pachymetry, and obtain fundus photographs and optical coherence tomography images. These findings are then transmitted through the electronic health record to eye care providers at the University of Michigan for remote interpretation. As part of the MI-SIGHT program, treatment for URE was provided by having ophthalmic technicians guide patients through ordering affordable eyeglasses online. The ophthalmic technicians then see the patients at a follow-up visit where they give the patient the eye care provider’s recommendations, help them access appropriate follow-up care and fit the glasses that were obtained online. We hypothesized that there would be high rates of VI from URE in this cohort. We also hypothesized that this novel way of treating URE would lead to high patient satisfaction.
Methods
Description of the program
The MI-SIGHT program is a telemedicine-based eye health screening initiative embedded in two primary care clinics: a federally qualified health center (FQHC) in Flint, MI and a free clinic in Ypsilanti, MI.18 Community residents ≥ 18 years of age were eligible to participate. Exclusion criteria included: 1. Significant eye pain (Likert scale ≥ 8 out of 10); 2. Sudden decrease in vision within 1 week; 3. Binocular diplopia; 4. Cognitive impairment; 5. Pregnancy; 6. Incarceration; or 7. Moving outside of driving distance to the clinic within 6 months. At the free clinic, primary care physicians referred patients to the program, and patients received a call from the study team inviting them to participate. At the FQHC, diabetic patients who were overdue for their annual eye exam received a phone call inviting them to participate. Additional recruitment was conducted via advertisements in both communities in places such as foodbanks, churches, barbershops, health fairs, local radio and buses, and some participants learned about the program through word of mouth. Data were collected over the first year of the program; because the two sites joined the program at different times, the first-year cohort consisted of participants recruited from June 28, 2020 until June 27, 2021 at the free clinic, and from January 27, 2021 until January 26, 2022 at the FQHC. The study was approved by the Institutional Review Board at the University of Michigan and adhered to all tenets of the Declaration of Helsinki. The clinical trial component of the study, not described herein, is registered at clinicaltrials.gov (NCT04274764).
Participants either scheduled a visit with an ophthalmic technician stationed at their local clinic or presented as walk-ins. Following confirmation of study eligibility, written informed consent was obtained by the ophthalmic technician. Consent forms were available in English, Spanish, Arabic, Albanian, Chinese, French, Hindi, Korean, and Tagalog. The technician completed a health history and surveys with the patient, including a survey of demographic information (gender, race, ethnicity, education, income, employment, insurance status and type) and the 9-item National Eye Institute Visual Function Questionnaire (VFQ9) to assess visual function.17 This shortened version of the original questionnaire18 assessed visual function across 7 domains (general vision, near and distance activities, mental health, role difficulties, driving, and peripheral vision). Each question was asked on a Likert scale and coded from 0 to 100 with larger values indicating better visual function. The questionnaire was scored overall and for each of the 7 subscale domains as a mean of items included. Non-English speaking participants completed the surveys with translation support from bilingual companions accompanying them to the appointment.
The ophthalmic technician measured participants’ presenting visual acuity (VA) separately in each eye using a Snellen chart set for a distance of 10 feet. VA was defined by the smallest line on which the participant correctly read the majority of letters. If the participant brought eyeglasses or contact lenses to the appointment, corrected VA was measured wearing the participant’s habitual refractive correction. Uncorrected VA was measured if the patient did not present with eyeglasses or contact lenses. Refraction was initially performed using an autorefractor (ARK- Autorefractor & Keratometer, Marco Ophthalmic, Jacksonville, FL) and then refined with subjective refraction using a table-clamped phoropter. The final phoropter refraction was placed into trial frames for the patient to confirm that they were pleased with the prescription and additional refinement was performed if necessary. Best corrected visual acuity (BCVA) was measured with the 10-foot Snellen chart wearing the final eyeglasses prescription. Inter-pupillary distance was measured with a pupilometer (Essilor Digital Pupilometer, Essilor, Chicago, IL). The technician also performed numerous other tests and obtained several images including fundus photographs, external photographs, and optical coherence tomography images. All data were transmitted to remote eye care providers at the University of Michigan for review and interpretation within 4 business days of the visit. Although this report focuses on refractive data from the study, findings from other aspects of the program have previously been published.19
If the participant wished to order eyeglasses, the technician guided the participant through the process of ordering low-cost eyeglasses ($7.95 + $4.95 shipping and handling and up) online from ZenniOptical.com. Each clinic had a physical display of a selection of the different $7.95 frames available from the online optical shop arranged by best fit for each range of interpupillary distance and gender to help participants narrow in on the best fit for their face. Each clinic also had a mirror so that participants could try on the various styles of frames. Once participants selected a style of frames, the technician helped them navigate the website to order the appropriate style, color, size, and prescription. Participants paid out of pocket for the eyeglasses directly to the online optical shop with a personal credit card. For participants without a credit card, the technician paid online using a university credit card and the patient reimbursed the technician using cash. The final order for eyeglasses was placed after the remote eye care provider approved the prescription. Eyeglasses were delivered to the clinic.
Participants returned to the clinic two to four weeks after their initial visit to pick up their eyeglasses. The technicians, who were trained to fit eyeglasses by opticians, performed adjustments to optimize the fit. At this visit, participants were asked whether they were satisfied or dissatisfied with their new eyeglasses, and the technician noted whether a remake was required for any reason. The ophthalmic technician also relayed the results of the participant’s screening exam, shared recommendations for follow up from the remote eye care provider and assisted the participant in scheduling any recommended follow-up appointments.
Definitions
VI was defined as presenting VA ≤20/50, corresponding to the World Health Organization definition.20 Presenting VA was compared with BCVA to create 4 groups of refractive error (RE), including 1. VI due to pathology (presenting VA ≤20/50 and BCVA ≤20/50), 2. VI from URE (presenting VA ≤20/50 and BCVA ≥20/40), 3. URE without VI (presenting VA ≥20/40 and BCVA achieving 2+ lines of improvement), and 4. no or adequately corrected RE (presenting VA ≥20/40 and BCVA achieving <2 lines of improvement). RE groups were based on presenting VA from the better seeing eye of a participant or the eye with the greater improvement in visual acuity for participants whose eyes had the same presenting VA.11 Participants with RE were further sub-categorized into undercorrected RE (those who presented with glasses/contact lenses) and uncorrected RE (those who presented without glasses/contact lenses). Because patients who read the 20/20 line successfully were inconsistently shown the 20/15 line at follow up, presenting VA and BCVA values >20/20 were truncated to 20/20.
The spherical equivalent (SE) was calculated from the manifest refraction by adding the sphere power with half of the cylinder power. SE was categorized into myopic and hyperopic groups based on the eye with the largest absolute SE. Myopic categories were selected to align with prior studies, including high myopia (SE ≤-5 diopters, D), moderate myopia (SE −4.99 to −3D), and mild myopia (SE-2.99 to −1D). Emmetropia was defined as SE ±0.99D. Hyperopia and clinically significant hyperopia have been commonly defined as a hyperopic SE of +3.0 diopters or greater.11,28,29 To study lower degrees of hyperopia, we expanded the definition to include mild hyperopia (SE +1 to +2.99D), moderate hyperopia (SE +3 to +4.99D), and high hyperopia (SE ≥+5D).
Statistical Methods
Characteristics of first year MI-SIGHT participants were summarized with means and standard deviations (SD) for continuous measures and frequencies and percentages for categorical measures, stratified by RE group. Patient demographics, self-reported visual function, and satisfaction with glasses were compared between RE groups using analysis of variance (ANOVA), Kruskal-Wallis (KW), Chi-square, and Fisher exact testing. Due to the dissimilarity of participants who had VI due to pathology, this group was excluded from most analyses. Self-reported visual function was compared between myopic and hyperopic groups using Kruskal-Wallis tests. Due to small sample sizes, moderate and high hyperopia groups were combined for analysis. Multiple comparison adjustment was performed using the Holm procedure. Significant Kruskal-Wallis tests were followed by post-hoc pairwise comparisons using the Dwass, Steel, Critchlow-Fligner (DSCF) multiple comparison procedure. All analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC).
Results
A total of 1171 participants were enrolled in the MI-SIGHT program during the first year. Full sample characteristics have been previously reported.19 Briefly, participants were on average 55.1 years old (SD=14.5), 37.7% male, 54.1% identified as Black, 33.9% identified as White, and 10.4% identified as Hispanic or Latino. Of the 1171 participants, 1166 (99.6%) had both presenting VA and BCVA measured (Figure 1 – flow chart). VI was observed in 120 participants (10.3%), including 24 who had VI due to pathology (2.1% of full sample) and 96 who had VI from URE (8.2% of full sample). VI due to pathology included cases of cataract (n=12/24, 50.0%), diabetic retinopathy (n=3, 12.5%), glaucoma (n=2, 8.3%), age-related macular degeneration (n=1, 4.2%), and other less common pathology (n=6, 25.0%, Supplemental Table 1). VI from URE included cases of undercorrected RE (n=27/96, 28.1%) and uncorrected RE (n=69, 71.9%). Most participants did not present with VI (n=1046, 89.7%). However, 168 participants had ≥2 lines of improvement in visual acuity when refracted (14.4% of the full sample). These cases of RE without VI included those with undercorrected RE (n=59/168, 35.1%) and uncorrected RE (n=109/168, 64.9%). The remaining 878 participants needed no spectacle correction (n=187/878, 21.3%) or minor spectacle correction (<2 lines, n=691/878, 78.7%). Although our analysis focuses on VA measures of the better seeing eye at presentation in accordance with other studies,11 it should be noted that a larger percentage of subjects had VI overall (24.6%) and VI from URE (18.7%) when assessing VA measures from either eye.
Figure 1.
Flow chart of first year MI-SIGHT patient cohort
MI-SIGHT, VA, Visual Acuity; BCVA, Best-Corrected Visual Acuity; AMD, Age-Related Macular Degeneration; DR, Diabetic Retinopathy
Demographic differences between the 3 main RE groups (VI from URE, URE without VI, and no or adequately corrected RE, based on VA of the better seeing eye at presentation) are presented in Table 1. Participant education and household income were significantly associated with RE status (both p<0.0001). Specifically, a larger percentage of participants with VI from URE reported having less than a high school education than those participants with no or adequately corrected RE (21.3% versus 8.1%, respectively, post-hoc pairwise p=0.0005; all post-hoc pairwise comparisons presented in Supplemental Table 2). Similarly, a smaller percentage of participants with VI from URE reported having a college degree compared to participants with no or adequately corrected RE (3.2% versus 14.2%, respectively, p=0.0154). In terms of income, a larger percentage of participants with VI from URE reported income <$10,000 than those with no or adequately corrected RE (45.5% versus 21.6%, respectively, p<0.0001). No significant differences between RE groups were observed for age, sex, race, ethnicity, diabetes status, employment status, and health insurance status or type (all p>0.05, Table 1). Driving was also associated with RE status, such that a smaller percentage of participants with VI from URE reported driving themselves to their screening appointment than those with URE without VI or those with no or adequately corrected RE (34.7% versus 65.3% or 72.5%, respectively, both pairwise comparisons p<0.0001). Additionally, a larger percentage of participants with no or adequately corrected RE presented with glasses than those with VI from URE or those with URE without VI (60.3% versus 28.1% or 35.1%, respectively, both pairwise comparisons p<0.0001). RE status was also associated with SE type and severity such that a smaller percentage of participants with VI from URE were categorized as having no myopia/hyperopia that those who with URE without VI or those with no or adequately corrected RE (24.0% versus 58.3% or 60.1%, respectively, both pairwise comparisons p<0.0001); a larger percentage of those with VI from URE had mild myopia or mild hyperopia than those with no or adequately corrected RE (mild myopia: 28.1% versus 15.2%, p=0.0165; mild hyperopia: 27.1% versus 11.2%, p=0.0001). Lastly, a smaller percentage of those participants with no or adequately corrected RE reported their last eye exam was ≥1 year ago compared to those with VI from URE or those with URE without VI (67.1% versus 85.2% or 78.3%, respectively; pairwise comparison p=0.0039 and p=0.0100, respectively). No significant differences between RE groups were found for reasons for having an eye exam ≥1 year ago. However, a marginally significant difference in reasons why people had not had a dilated eye exam in the last year was that those with VI from URE reported more frequently that cost was a barrier to eye care than those with URE without VI (22.5% versus 4.4%, p=0.08).
Table 1.
Comparison of participant characteristics between refractive error categories in first year MI-SIGHT participants
| Visual Impairment due to Un/Undercorrected Refractive Error (n=96) | Un/Undercorrected Refractive Error Without Visual Impairment (n=168) | No or Adequately Corrected Refractive Error (n=878) | Holm-Adjusted P-value | |
|---|---|---|---|---|
| Continuous Variable | Mean (SD), Median | Mean (SD), Median | Mean (SD), Median | p-value* |
| Categorical Variable | #/total (Column %) | #/total (Column %) | #/total (Column %) | p-value** |
| Sex - Male | 40/95 (42.1) | 71/166 (42.8) | 319/870 (36.7) | 0.8770 |
| Race | ||||
| White | 32/89 (36.0) | 59/152 (38.8) | 273/827 (33.0) | |
| Black | 51/89 (57.3) | 80/152 (52.6) | 449/827 (54.3) | 0.8664 |
| Asian | 0/89 (0.0) | 3/152 (2.0) | 42/827 (5.1) | |
| Other | 6/89 (6.7) | 10/152 (6.6) | 63/827 (7.6) | |
| Ethnicity - Hispanic | 8/82 (9.8) | 18/140 (12.9) | 71/721 (9.9) | 0.8770 |
| Education | ||||
| <HS | 20/94 (21.3) | 24/164 (14.6) | 70/866 (8.1) | |
| HS or Equivalent | 38/94 (40.4) | 60/164 (36.6) | 265/866 (30.6) | |
| Some College | 28/94 (29.8) | 55/164 (33.5) | 300/866 (34.6) | <0.0001 |
| College Degree | 3/94 (3.2) | 10/164 (6.1) | 123/866 (14.2) | |
| Graduate Degree | 5/94 (5.3) | 15/164 (9.2) | 108/866 (12.5) | |
| Income | ||||
| <$10k | 35/77 (45.5) | 44/149 (29.5) | 161/744 (21.6) | |
| $10k-$19,999 | 19/77 (24.7) | 40/149 (26.9) | 177/744 (23.8) | |
| $20k-$29,999 | 11/77 (14.3) | 34/149 (22.8) | 154/744 (20.7) | <0.0001 |
| $30k-$49,999 | 11/77 (14.3) | 21/149 (14.1) | 139/744 (18.7) | |
| $50k-$69,999 | 0/77 (14.3) | 5/149 (3.4) | 55/744 (7.4) | |
| $70k+ | 1/77 (1.3) | 5/149 (3.4) | 58/744 (7.8) | |
| Employed (Full, Part, Self) | 32/95 (33.7) | 59/162 (36.4) | 361/865 (41.7) | 0.8770 |
| Diabetes | 33/92 (35.9) | 74/160 (46.3) | 321/859 (37.3) | 0.4392 |
| Health Insurance - Yes | 72/93 (77.4) | 139/164 (84.8) | 687/857 (80.2) | 0.8770 |
| Health Insurance Type | ||||
| None | 21/90 (23.3) | 25/161 (15.5) | 170/835 (20.4) | |
| Private | 23/90 (25.6) | 56/161 (34.8) | 301/835 (36.1) | 0.3978 |
| Medicare | 10/90 (11.1) | 24/161 (14.9) | 138/835 (16.5) | |
| Medicaid | 36/90 (40.0) | 56/161 (34.8) | 226/835 (27.1) | |
| Presented with Glasses | 27/96 (28.1) | 59/168 (35.1) | 529/878 (60.3) | <0.0001 |
| Refractive Error Type | ||||
| High Myopia (SE ≤ −5) | 7/96 (7.3) | 2/168 (1.2) | 43/877 (4.9) | |
| Moderate Myopia (−5 < SE ≤ −3) | 9/96 (9.4) | 4/168 (2.4) | 58/877 (6.6) | |
| Mild Myopia (−3 < SE ≤ −1) | 27/96 (28.1) | 30/168 (17.9) | 133/877 (15.2) | <0.0001 |
| No Myopia/Hyperopia (SE ± 0.99) | 23/96 (24.0) | 98/168 (58.3) | 527/877 (60.1) | |
| Mild Hyperopia (+1 ≤ SE <+3) | 26/96 (27.1) | 29/168 (17.3) | 98/877 (11.2) | |
| Moderate/Higha Hyperopia (SE ≥ +3) | 4/96 (4.2) | 5/168 (3.0) | 18/877 (2.1) | |
| Drove Self to Appt | 33/95 (34.7) | 109/167 (65.3) | 632/872 (72.5) | <0.0001 |
| Last Eye Exam ≥1 year | 76/91 (85.2) | 126/161 (78.3) | 562/838 (67.1) | 0.0024 |
| Reason Eye Exam ≥1 year | ||||
| Cost | 17/62 (27.4) | 12/102 (11.8) | 72/451 (16.0) | |
| No Insurance | 13/62 (21.0) | 20/102 (19.6) | 136/451 (30.2) | |
| Don’t have eye doctor | 1/62 (1.6) | 6/102 (5.9) | 29/451 (6.4) | |
| Cannot get to office/clinic | 3/62 (4.8) | 1/102 (1.0) | 6/451 (1.3) | 0.2000 |
| Could not get appt | 1/62 (1.6) | 3/102 (2.9) | 5/451 (1.1) | |
| No reason to go | 10/62 (16.1) | 24/102 (23.5) | 99/451 (22.0) | |
| Have not thought of it | 5/62 (8.1) | 16/102 (15.7) | 39/451 (8.7) | |
| Other | 12/62 (19.4) | 20/102 (19.6) | 65/451 (14.4) | |
| Last Dilated Eye Exam ≥1 year | 73/78 (93.6) | 114/136 (83.8) | 577/693 (83.3) | 0.4392 |
| Reason Dilated Eye Exam ≥1 year | ||||
| Cost | 11/49 (22.5) | 3/69 (4.4) | 52/339 (15.3) | |
| No Insurance | 13/49 (26.5) | 12/69 (17.4) | 99/339 (29.2) | |
| Don’t have eye doctor | 2/49 (4.1) | 6/69 (8.7) | 10/339 (3.0) | 0.0308 |
| Cannot get to office/clinic | 3/49 (6.1) | 1/69 (1.5) | 3/339 (0.9) | |
| Could not get appt | 0/49 (0.0) | 3/69 (4.4) | 5/339 (1.5) | |
| No reason to go | 8/49 (16.3) | 17/69 (24.6) | 85/339 (25.1) | |
| Have not thought of it | 5/49 (10.2) | 14/69 (20.3) | 41/339 (12.1) | |
| Other | 7/49 (14.3) | 13/69 (18.8) | 44/339 (13.0) |
MI-SIGHT, Michigan Screening and Intervention for Glaucoma and Eye Health through Telemedicine; SD, Standard Deviation; HS, High School; k, $1000; SE, Spherical Equivalent; P-value using
ANOVA (continuous variables) and
Chi-square (categorical variables with cell frequency ≥5) or Fisher exact test with Monte-Carlo simulation (categorical variables with cell counts <5); All p-values adjusted for multiple comparisons using the Holm method;
Moderate Hyperopia is defined as SE +3 to +4.99, n=25; High Hyperopia is defined as SE≥+5, n=3; categories were combined due to the small sample size of high hyperopes
Self-reported visual function was significantly associated with RE status (Table 2). Participants with VI from URE reported the lowest visual function, followed by those with URE without VI. Those with no or adequately corrected RE reported the highest visual function. All three groups were significantly different from each other with respect to VFQ9 composite score (mean±SD, 67.3±19.6 versus 77.0±14.4 versus 82.2±13.3, respectively; p<0.0001), and subscales related to near vision (66.9±25.2 versus 76.1±19.4 versus 82.4±173; p<0.0001), distance activities (72.6±29.2 versus 83.1±23.0 versus 88.2±19.3; p<0.0001), and mental health (41.4±30.7 versus 54.5±30.8 versus 62.2±28.6; p<0.0001). All post-hoc pairwise comparisons were significant at p<0.05 after adjustment for multiple comparisons. Participants with VI from URE and those with URE without VI reported significantly worse visual function than those with no or adequately corrected RE with respect to general vision (59.4±22.5 and 64.0±18.6 versus 72.0±17.0; p<0.0001) and peripheral vision (79.0±26.5 and 87.7±18.7 versus 91.4±16.0; p<0.0001); both post-hoc pairwise comparisons p<0.05 after adjustment for multiple comparisons. Visual function with respect to driving was significantly lower for those with VI from URE (79.0±26.5) than those with URE without VI (87.7±18.7) and those with no or adequately corrected RE (91.4±16.0); p<0.0001 and both post-hoc pairwise comparisons p<0.05. A total of 17 participants reported they stopped driving due to their vision (1.5% of the full sample), with 29.2% of those with VI due to pathology reporting this, followed by 4.2% of those with VI from URE, 0.7% of those with no or adequately corrected RE, and 0% of those with URE without VI (p<0.0001; Supplemental Table 3). Lastly, visual function with respect to role difficulties was significantly worse in those with VI from URE than those with no or adequately corrected RE (79.2±30.9 versus 88.4±25.6; p=0.0012 and post-hoc pairwise comparison p<0.05).
Table 2.
Comparison of self-reported visual function between refractive error categories in first year MI-SIGHT participants
| Visual Impairment due to Un/Undercorrected Refractive Error (n=96) |
Un/Undercorrected Refractive Error Without Visual Impairment (n=168) |
No or Adequately Corrected Refractive Error (n=878) |
|||||
|---|---|---|---|---|---|---|---|
| VFQ Subscale | n | Mean (SD), Median | n | Mean (SD), Median | n | Mean (SD), Median | p-value* |
| Composite | 96 | 67.3 (19.6), 68.1 | 165 | 77.0 (14.4), 78.9 | 874 | 82.2 (13.3), 84.4 | <0.0001a,b,c |
| General Vision | 95 | 59.4 (22.5), 60.0 | 164 | 64.0 (18.6), 60.0 | 868 | 72.0 (17.0), 80.0 | <0.0001a,b |
| Near Activities | 96 | 66.9 (25.2), 75.0 | 165 | 76.1 (19.4), 75.0 | 873 | 82.4 (17.3), 83.3 | <0.0001a,b,c |
| Distance Activities | 92 | 72.6 (29.2), 75.0 | 160 | 83.1 (23.0), 100.0 | 851 | 88.2 (19.3), 100.0 | <0.0001a,b,c |
| Mental Health | 96 | 41.4 (30.7), 50.0 | 165 | 54.5 (30.8), 50.0 | 874 | 62.2 (28.6), 75.0 | <0.0001a,b,c |
| Role Difficulties | 95 | 79.2 (30.9), 100.0 | 163 | 86.7 (27.0), 100.0 | 866 | 88.4 (25.6), 100.0 | 0.0012a |
| Driving | 73 | 77.1 (30.5), 100.0 | 145 | 92.6 (14.2), 100.0 | 814 | 92.4 (17.3), 100.0 | <0.0001a,c |
| Peripheral Vision | 94 | 79.0 (26.5), 100.0 | 164 | 87.7 (18.7), 100.0 | 864 | 91.4 (16.0), 100.0 | <0.0001a,b |
MI-SIGHT, Michigan Screening and Intervention for Glaucoma and Eye Health through Telemedicine; VFQ, Visual Function Questionnaire; SD, Standard Deviation
Kruskal-Wallis test
Post-hoc pairwise comparisons with Dwass, Steel, Critchlow-Fligner multiple comparison procedure showed significant differences between:
Visual Impairment due to Un/Undercorrected refractive error versus No/Minor refractive error,
Un/Undercorrected refractive error versus No/Minor refractive error,
Visual Impairment due to un/undercorrected refractive error versus Un/Undercorrected refractive error
Most participants had no myopia or hyperopia, as defined by the eye with the largest absolute SE (n=657, 56.4%). However, many patients had mild hyperopia (n=158, 13.6%), with less having moderate hyperopia (n=25, 2.2%) or high hyperopia (n=3, 0.3%). Alternatively, more patients had mild myopia (n=195, 16.8%), moderate myopia (n=71, 6.1%), and high myopia (n=55, 4.7%). Only visual function with respect to general vision (p=0.0014) and driving (p=0.0093) showed significant differences between myopic and hyperopic groups (Table 3). Specifically, participants with mild hyperopia reported significantly lower general vision function than those with moderate and high myopia (65.7±20.1 versus 74.6±17.17 and 76.0±19.0, respectively; p<0.05). Participants with mild myopia reported significantly worse driving visual function compared to those with no myopia or hyperopia (87.0±21.3 versus 91.5±19.4; post-hoc pairwise comparisons p<0.05).
Table 3.
Comparison of self-reported visual function between spherical equivalent categories in the MI-SIGHT participants
| High Myopia SE ≤ −5 (n=55) | Moderate Myopia −5 < SE ≤ −3 (n=71) | Mild Myopia −3 < SE ≤ −1 (n=195) | No Myopia/Hyperopia SE ± 0.99 (n=657) | Mild Hyperopia 1 ≤ SE <3 (n=158) | Moderate/High* Hyperopia SE ≥ 3 (n=28) | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| VFQ Subscale | n | Mean (SD), Median | n | Mean (SD), Median | n | Mean (SD), Median | n | Mean (SD), Median | n | Mean (SD), Median | n | Mean (SD), Median | P-value** |
| Composite | 55 | 83.8 (14.7), 88.9 | 70 | 82.5 (13.8), 86.7 | 194 | 78.9 (15.1), 81.7 | 652 | 79.6 (15.1), 82.5 | 157 | 78.1 (16.6), 81.7 | 28 | 81.9 (16.7), 88.0 | 0.0570 |
| General Vision | 55 | 76.0 (19.0), 80.0 | 70 | 74.6 (17.7), 80.0 | 193 | 69.7 (18.7), 80.0 | 645 | 69.1 (18.0), 60.0 | 157 | 65.7 (20.1), 60.0 | 28 | 71.4 (12.7), 80.0 | 0.0014a,b |
| Near Activities | 55 | 85.2 (17.6), 91.7 | 70 | 83.1 (17.2), 83.3 | 193 | 79.7 (19.4), 83.3 | 652 | 79.5 (19.1), 83.3 | 157 | 77.3 (21.7), 83.3 | 28 | 79.8 (21.2), 87.5 | 0.1445 |
| Distance Activities | 53 | 87.7 (18.1), 100.0 | 69 | 87.3 (19.5), 100.0 | 188 | 84.2 (22.7), 100.0 | 633 | 85.9 (21.8), 100.0 | 152 | 85.0 (21.8), 100.0 | 28 | 85.7 (23.0), 100.0 | 0.9210 |
| Mental Health | 55 | 62.7 (29.6), 75.0 | 70 | 58.9 (28.2), 50.0 | 194 | 59.3 (27.2), 50.0 | 652 | 58.6 (30.7), 50.0 | 157 | 58.0 (30.8), 50.0 | 28 | 63.4 (29.3), 75.0 | 0.8255 |
| Role Difficulties | 55 | 88.6 (27.6), 100.0 | 70 | 88.6 (22.8), 100.0 | 193 | 84.7 (28.3), 100.0 | 643 | 87.0 (26.4), 100.0 | 156 | 86.4 (28.2), 100.0 | 28 | 95.5 (19.3), 100.0 | 0.2496 |
| Driving | 52 | 93.3 (14.9), 100.0 | 66 | 92.0 (17.1), 100.0 | 181 | 87.0 (21.3), 100.0 | 587 | 91.5 (19.4), 100.0 | 134 | 91.2 (20.4), 100.0 | 23 | 93.5 (21.6), 100.0 | 0.0093c |
| Peripheral Vision | 54 | 90.7 (18.4), 100.0 | 69 | 92.4 (15.0), 100.0 | 190 | 88.6 (17.6), 100.0 | 647 | 89.1 (18.5), 100.0 | 155 | 88.9 (20.2), 100.0 | 28 | 92.0 (18.1), 100.0 | 0.4997 |
MI-SIGHT, Michigan Screening and Intervention for Glaucoma and Eye Health through Telemedicine; VFQ, Visual Function Questionnaire; SE, Spherical Equivalent; SD, Standard Deviation
Moderate Hyperopia is defined as SE 3–4.99, n=25; High Hyperopia is defined as SE≥5, n=3; categories were combined due to the small sample size of high hyperopes;
Kruskal-Wallis test
Post-hoc pairwise comparisons with Dwass, Steel, Critchlow-Fligner multiple comparison procedure showed significant differences between:
High myopia versus Mild hyperopia,
Moderate myopia versus Mild hyperopia,
Mild myopia versus No myopia/hyperopia
Most participants ordered glasses through the MI-SIGHT program (n=830, 71.2%; Table 4). Participants who ordered glasses were on average 55.8 years old (SD=14.2), 65.2% were female, 59.5% were Black, and 11.1% were Hispanic ethnicity A larger percentage of participants with VI from URE and those with URE without VI ordered glasses than participants with no or adequately corrected RE (81.3% and 82.1% versus 68.3%, respectively; p=0.0001). The average cost of glasses was $36.80 (SD=$32.6) and this cost was similar across RE groups (ANOVA p=0.9). Of those participants who ordered glasses, 97.7% reported they were satisfied with their glasses, 1.3% were dissatisfied, and 1.0% required a remake of their glasses. Satisfaction with glasses was similar across RE groups (p=0.9) and age deciles (p=0.5). Of the 336 participants who did not order glasses, the majority (82.7%) had no or adequately corrected RE and a minority (3.0%) had pathology where correction could not be obtained. It is unknown why the remaining participants (14.3%) did not order glasses, which included 5.4% who had VI from URE and 8.9% who had URE without VI.
Table 4.
MI-SIGHT program satisfaction with glasses service
| Overall (n=1166) | Visual Impairment due to Un/Undercorrected Refractive Error (n=96) | Un/Undercorrected Refractive Error Without Visual Impairment (n=168) | No or Adequately Corrected Refractive Error (n=878) | ||
|---|---|---|---|---|---|
| Categorical Variable | #/total (Column %) | #/total (Column %) | #/total (Column %) | p-value* | |
| Ordered Glasses | 830/1166 (71.2) | 78/96 (81.3) | 138/168 (82.1) | 600/878 (68.3) | 0.0001a,b |
| Satisfaction w/ Glasses | |||||
| Satisfied | 803/822 (97.7) | 75/76 (98.7) | 135/138 (97.8) | 579/594 (97.5) | |
| Dissatisfied | 11/822 (1.3) | 1/76 (1.3) | 1/138 (0.7) | 9/594 (1.5) | 0.9001 |
| Dissatisfied - remake required | 8/822 (1.0) | 0/76 (0.0) | 2/138 (1.5) | 6/594 (1.0) | |
| Continuous Variable | Mean (SD), Median | Mean (SD), Median | Mean (SD), Median | Mean (SD), Median | p-value** |
| Cost of Glasses | 36.8 (25.2), 32.6 | 34.3 (19.5), 31.7 | 36.9 (22.8), 32.9 | 37.3 (26.6), 33.0 | 0.8843 |
MI-SIGHT, Michigan Screening and Intervention for Glaucoma and Eye Health through Telemedicine; SD, Standard Deviation
Chi-square (categorical variables with cell frequency ≥5) or Fisher exact test with Monte-Carlo simulation (categorical variables with cell counts <5)
ANOVA (continuous variables)
Post-hoc pairwise chi-square or Fisher exact comparisons with Holm-adjustment showed significant differences between:
Visual Impairment due to un/under-corrected refractive error versus No/Minor refractive error,
Un/Undercorrected refractive error versus No/Minor refractive error
Discussion
Approximately 1 in 10 participants in this cohort of patients from an FQHC and a free clinic had VI, and 80% of VI was due to URE. URE was associated with having less education and lower income. Participants with URE were less likely to drive themselves to their appointment and they were more likely to have had their last eye exam more than a year ago. Participants with VI due to URE reported the lowest visual function. Participants with mild hyperopia reported significantly lower general vision function than participants with moderate and high myopia. Participants with mild myopia reported significantly worse driving visual function compared to those with no myopia or hyperopia. The majority (70%) of participants ordered eyeglasses through the MI-SIGHT program with an average cost of $36.80. Nearly all participants were satisfied with their new eyeglasses.
URE was the most common cause of VI in this study. URE was also the main cause of VI in the nationally-representative National Health and Nutrition Examination Study survey, in which an autorefractor containing built-in VA charts was used to measure BCVA from 1999–2008.1,21 URE is detrimental to patients’ health and well-being. As our participants’ VRQL scores showed, VI from URE has a significant negative impact on visual function: MI-SIGHT participants with VI from URE had worse VFQ9 scores across all domains compared to participants without VI. Participants with milder URE that did not meet the definition of VI also had decreased VFQ9 scores compared to those with no or adequately corrected RE. There are abundant data from other studies showing VI is associated with numerous undesirable outcomes such as decreased workplace productivity,9,10 injurious falls,4 depression,5 dementia,6 cognitive decline, motor vehicle accidents,7 and mortality.8 In a seminal 2016 report, the National Academies of Science, Engineering, and Medicine called for eliminating correctable and avoidable vision impairments by 2030.22 Because RE is completely treatable with a simple pair of eyeglasses, it is discouraging that URE remains the main cause of VI in this sample of adults from two clinics serving populations with low income. To reduce the numerous harms associated with RE, addressing RE must become a public health priority in the US.
Lower income and education were significantly associated with presenting with URE, while other sociodemographic characteristics including race, ethnicity, employment, and health insurance were not. These findings differ from prior studies using data from the nationally-representative NHANES survey, which found that Mexican Americans and non-Hispanic Black participants had higher odds of inadequate refractive correction compared to non-Hispanic White participants across all age groups.1,11 Social determinants of health are key drivers of health disparities and inequities in the US, so it is critical to examine the societal factors shaping vision health.23 It is increasingly accepted that race and ethnicity represent social constructs as opposed to biologic determinants, and therefore racial and ethnic differences in health outcomes are likely due to socioeconomic determinants of health as wealth is strongly associated with race and ethnicity in the US.24 For example, recent studies have challenged a previously identified association between race and severe retinopathy of prematurity by showing that an association between race and ethnicity and gestational age was no longer significant after adjusting for household income and insurance status.25 Some authors have found that racial and ethnic health disparities among adults of similar incomes are small compared to disparities identified between high- and low-income populations within racial and ethnic groups.26 This could explain why there was no association between URE and race and ethnicity among the predominantly low-income participants in this study. More than 70% of participants in this study reported less than $30,000 of household income per year,19 which is less than half the median household income in the US ($79,900).27
Hyperopia has been defined inconsistently in prior literature on RE. Hyperopia and clinically significant hyperopia has been commonly defined as a hyperopic SE of +3.0 diopters or greater.11,28,29 Thus, patients with SE measures of +1.0 or +2.0 diopters are not defined as hyperopes in many studies. We found that mild hyperopes (SE +1.00 to +2.99 diopters) had significantly lower general vision function than those with moderate and high myopia. This suggests that mild hyperopia may have a larger impact on vision and daily life than previously thought. Going forward, definitions of hyperopia should be broadened to include mild hyperopes.
In this study, participants were very satisfied with the affordable glasses ordered online; dissatisfaction and re-makes were extremely rare. Recent investigations have found that the accuracy of how online eyeglasses are manufactured has increased significantly over the past 10 years; over 90% of eyeglasses obtained online met national quality standards for prescription accuracy.30 Yet despite the improved accuracy of eyeglasses obtained online, many patients may have concerns about obtaining eyeglasses online. A study of Canadian adults found that concerns over fit and quality were barriers to obtaining eyeglasses online,31 and a study among low-income adults at the free clinic served by this program found that nearly all participants were skeptical about obtaining eyeglasses online, most commonly due to concerns over inaccurate fit or prescription.32 The high level of satisfaction with eyeglasses ordered online in this study shows that eye care programs can use online optical shops to improve access to eyeglasses when delivered in a way that overcomes these common patient concerns.
Several features of our program likely contributed to its success in having high patient satisfaction with the online optical shop experience. Having a display of sample frames for patients to try on, having a technician guide the patient through the process of purchasing eyeglasses online and input the prescription values for the patient so as to minimize transcription error, having the interpupillary distance measured with a pupillometer, and having the technician make minor adjustments to optimize fit likely assuaged common patient concerns about ordering eyeglasses online. Telehealth programs have historically not been optimized for older33 or low-income patients,34 or those with limited internet access35 or digital literacy.36 This novel program is an example of how online optical shops and telehealth programs can successfully address RE in adults, even among those who may not have access to the internet, do not feel comfortable navigating healthcare online, or do not have a credit card with which to purchase eyeglasses online.
This study has notable strengths, including a large sample size. By recruiting patients from a free clinic and FQHC, we drew a sample of predominantly low-income people who face barriers to getting eyeglasses.32,37 Thus, we were able to target programmatic resources toward populations with the greatest need. While this is one of the study’s key strengths, it also introduces ascertainment bias and limits the generalizability of the findings. These results cannot be extrapolated to reflect the population prevalence of inadequate refractive correction. Some demographic variables contained a moderate amount of missing data; this may be due to social desirability bias. Cycloplegic refraction was not performed in this study, even for young adults who may have high accommodative amplitudes. Some authors have suggested that cycloplegic refraction should be the gold standard in refraction of young adults38,39 up to age 50.40 However, satisfaction with the eyeglasses dispensed was very high and did not differ by age decile. Participants were asked whether they were satisfied or dissatisfied with their new eyeglasses at the time of dispensing, so any dissatisfaction that may have arisen later was not captured in this study.
In conclusion, URE was the main cause of VI at two clinics serving communities with high levels of poverty and was associated with significantly reduced vision related quality of life. Yet, URE is completely treatable with a pair of glasses, making it a perfect target for public health interventions. The MI-SIGHT program successfully made low-cost eyeglasses accessible to low-income patients by utilizing an online optical shop and telemedicine. This program may be a good model for treating URE in other low-income settings.
Supplementary Material
Financial support:
This work was supported by the Centers for Disease Control and Prevention (CDC U01 DP006442) and Research to Prevent Blindness (MAW, PANC). This work was also supported by the University of Michigan National Clinician Scholars Program (OJK). The funding organizations had no role in the design or conduct of this research. The findings are that of the authors and do not represent the views of the CDC.
Footnotes
Conflict of interest statement: No conflicting relationship exists for any author.
Meeting presentation: This manuscript was presented orally at the 2023 ARVO Annual Meeting, April 23–27, in New Orleans, LA.
References
- 1.Vitale S, Cotch MF, Sperduto RD. Prevalence of Visual Impairment in the United States. JAMA 2006;295:2158–2163. [DOI] [PubMed] [Google Scholar]
- 2.Chou C-F, Frances Cotch M, Vitale S, et al. Age-Related Eye Diseases and Visual Impairment Among U.S. Adults. American Journal of Preventive Medicine 2013;45:29–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Tahhan N, Papas E, Fricke TR, et al. Utility and Uncorrected Refractive Error. Ophthalmology 2013;120:1736–1744. [DOI] [PubMed] [Google Scholar]
- 4.Ehrlich JR, Hassan SE, Stagg BC. Prevalence of Falls and Fall-Related Outcomes in Older Adults with Self-Reported Vision Impairment. J Am Geriatr Soc 2019;67:239–245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Killeen OJ, Xiang X, Powell D, et al. Longitudinal Associations of Self-Reported Visual, Hearing, and Dual Sensory Difficulties With Symptoms of Depression Among Older Adults in the United States. Frontiers in Neuroscience 2022;16. Available at: https://www.frontiersin.org/articles/10.3389/fnins.2022.786244 [Accessed July 15, 2022]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rogers MAM, Langa KM. Untreated poor vision: a contributing factor to late-life dementia. Am J Epidemiol 2010;171:728–735. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Piyasena P, Olvera-Herrera VO, Chan VF, et al. Vision impairment and traffic safety outcomes in low-income and middle-income countries: a systematic review and meta-analysis. The Lancet Global Health 2021;9:e1411–e1422. [DOI] [PubMed] [Google Scholar]
- 8.Ehrlich JR, Ramke J, Macleod D, et al. Association between vision impairment and mortality: a systematic review and meta-analysis. The Lancet Global Health 2021;9:e418–e430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Naidoo KS, Fricke TR, Frick KD, et al. Potential Lost Productivity Resulting from the Global Burden of Myopia: Systematic Review, Meta-analysis, and Modeling. Ophthalmology 2019;126:338–346. [DOI] [PubMed] [Google Scholar]
- 10.Marques AP, Ramke J, Cairns J, et al. Global economic productivity losses from vision impairment and blindness. eClinicalMedicine 2021;35. Available at: https://www.thelancet.com/journals/eclinm/article/PIIS2589-5370(21)00132-2/fulltext [Accessed June 15, 2022]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.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;55:6996–7005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Muñoz B, West SK, Rodriguez J, et al. Blindness, visual impairment and the problem of uncorrected refractive error in a Mexican-American population: Proyecto VER. Invest Ophthalmol Vis Sci 2002;43:608–614. [PubMed] [Google Scholar]
- 13.Varma R, Wang MY, Ying-Lai M, et al. The Prevalence and Risk Indicators of Uncorrected Refractive Error and Unmet Refractive Need in Latinos: The Los Angeles Latino Eye Study. Investigative Ophthalmology & Visual Science 2008;49:5264–5273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Varma R, Vajaranant TS, Burkemper B, et al. Visual Impairment and Blindness in Adults in the United States: Demographic and Geographic Variations From 2015 to 2050. JAMA Ophthalmology 2016;134:802–809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Saydah SH, Gerzoff RB, Saaddine JB, et al. Eye Care Among US Adults at High Risk for Vision Loss in the United States in 2002 and 2017. JAMA Ophthalmol 2020;138:479–489. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Neitzel AJ, Wolf B, Guo X, et al. Effect of a Randomized Interventional School-Based Vision Program on Academic Performance of Students in Grades 3 to 7: A Cluster Randomized Clinical Trial. JAMA Ophthalmol 2021;139:1104–1114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Killeen OJ, Zhou Y, Musch DC, et al. Access to eye care and prevalence of refractive error and eye conditions at a high school–based eye clinic in southeastern Michigan. Journal of American Association for Pediatric Ophthalmology and Strabismus 2022. Available at: https://www.sciencedirect.com/science/article/pii/S1091853122001355 [Accessed July 22, 2022]. [DOI] [PubMed] [Google Scholar]
- 18.Newman-Casey PA, Musch DC, Niziol LM, et al. Michigan Screening and Intervention for Glaucoma and eye Health through Telemedicine (MI-SIGHT): Baseline Methodology for Implementing and Assessing a Community-Based Program. J Glaucoma 2021;30:380–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Newman-Casey PA, Niziol LM, Elam AR, et al. Michigan Screening & Intervention for Glaucoma and Eye Health through Telemedicine Program: First Year Outcomes and Implementation Costs. Am J Ophthalmol 2023:S0002–9394(23)00097–1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Anon. Vision impairment and blindness. Available at: https://www.who.int/news-room/fact-sheets/detail/blindness-and-visual-impairment [Accessed April 12, 2023].
- 21.Swenor BK, Ramulu PY, Willis JR, et al. The prevalence of concurrent hearing and vision impairment in the United States. JAMA Intern Med 2013;173: 10.1001/jamainternmed.2013.1880. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.National Academies of Sciences, Engineering, and Medicine, Health and Medicine Division, Board on Population Health and Public Health Practice, Committee on Public Health Approaches to Reduce Vision Impairment and Promote Eye Health. Making Eye Health a Population Health Imperative: Vision for Tomorrow. (Welp A, Woodbury RB, McCoy MA, Teutsch SM, eds.). Washington (DC): National Academies Press (US); 2016. Available at: http://www.ncbi.nlm.nih.gov/books/NBK385157/ [Accessed August 11, 2022]. [PubMed] [Google Scholar]
- 23.Elam AR, Tseng VL, Rodriguez TM, et al. Disparities in Vision Health and Eye Care. Ophthalmology 2022;129:e89–e113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Anon. Misrepresenting Race — The Role of Medical Schools in Propagating Physician Bias | NEJM. Available at: https://www.nejm.org/doi/10.1056/NEJMms2025768?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%20%200pubmed [Accessed December 19, 2022]. [DOI] [PubMed] [Google Scholar]
- 25.Karmouta R, Altendahl M, Romero T, et al. Association Between Social Determinants of Health and Retinopathy of Prematurity Outcomes. JAMA Ophthalmology 2022;140:496–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Dubay LC, Lebrun LA. Health, behavior, and health care disparities: disentangling the effects of income and race in the United States. Int J Health Serv 2012;42:607–625. [DOI] [PubMed] [Google Scholar]
- 27.Anon. Estimated Median Family Incomes for Fiscal Year (FY) 2021.
- 28.Vitale S, Ellwein L, Cotch MF, et al. Prevalence of Refractive Error in the United States, 1999–2004. Archives of Ophthalmology 2008;126:1111–1119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kempen JH, Mitchell P, Lee KE, et al. The prevalence of refractive errors among adults in the United States, Western Europe, and Australia. Arch Ophthalmol 2004;122:495–505. [DOI] [PubMed] [Google Scholar]
- 30.Gordon A, Twa MD, Cutter G, Kleinstein RN. Accuracy and Repeatability of Internet-ordered Spectacle Lenses. Optometry and Vision Science 2021;98:1340–1347. [DOI] [PubMed] [Google Scholar]
- 31.Sivak AM, Spafford MM, Irving EL. Patient views about online purchasing of eyewear. Ophthalmic and Physiological Optics 2020;40:249–253. [DOI] [PubMed] [Google Scholar]
- 32.Killeen OJ, Cho J, Newman-Casey PA, et al. Barriers and Facilitators to Obtaining Eyeglasses for Vulnerable Patients in a Michigan Free Clinic. Optom Vis Sci 2021;98:243–249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Lam K, Lu AD, Shi Y, Covinsky KE. Assessing Telemedicine Unreadiness Among Older Adults in the United States During the COVID-19 Pandemic. JAMA Internal Medicine 2020;180:1389–1391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Khoong EC, Rivadeneira NA, Hiatt RA, Sarkar U. The Use of Technology for Communicating With Clinicians or Seeking Health Information in a Multilingual Urban Cohort: Cross-Sectional Survey. Journal of Medical Internet Research 2020;22:e16951. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hirko KA, Kerver JM, Ford S, et al. Telehealth in response to the COVID-19 pandemic: Implications for rural health disparities. Journal of the American Medical Informatics Association 2020;27:1816–1818. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Shaw J, Brewer LC, Veinot T. Recommendations for Health Equity and Virtual Care Arising From the COVID-19 Pandemic: Narrative Review. JMIR Formative Research 2021;5:e23233. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ballouz D, Cho J, Woodward MA, et al. Facilitators and Barriers to Glaucoma Screening Identified by Key Stakeholders in Underserved Communities: A Community-engaged Research Approach. J Glaucoma 2021;30:402–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Sun Y-Y, Wei S-F, Li S-M, et al. Cycloplegic refraction by 1% cyclopentolate in young adults: is it the gold standard? The Anyang University Students Eye Study (AUSES). British Journal of Ophthalmology 2019;103:654–658. [DOI] [PubMed] [Google Scholar]
- 39.Mimouni M, Zoller L, Horowitz J, et al. Cycloplegic autorefraction in young adults: is it mandatory? Graefes Arch Clin Exp Ophthalmol 2016;254:395–398. [DOI] [PubMed] [Google Scholar]
- 40.Morgan IG, Iribarren R, Fotouhi A, Grzybowski A. Cycloplegic refraction is the gold standard for epidemiological studies. Acta Ophthalmologica 2015;93:581–585. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.