Association of Visual Acuity Improvement With Uncorrected Refractive Error in Patients New to Low Vision Clinics

Key Points

Question  What is the association of visual acuity improvement and uncorrected refractive error in patients with ocular disease who are new to low vision clinics?

Findings  In this cases series, 27.8% of patients with mean visual acuity of 20/100 had at least 2-line acuity improvement; this finding was more likely seen in working-age adults aged 40 to <65 years, African American patients, or those with moderate vision impairment.

Meaning  The findings suggest that refractive evaluation should be emphasized in patients under care for ocular disease.

Importance  There is substantial socioeconomic and individual burden from uncorrected refractive error (URE) and chronic ocular disease. Understanding the association of visual acuity (VA) reduction with URE and the adults most likely to benefit from refraction may help support clinical decision-making in ophthalmologic care and maximize patient outcomes.

Objectives  To assess the magnitude of VA improvement associated with URE among adults under ophthalmic care who obtain low vision rehabilitation (LVR) services and identify the characteristics of the patients who are most likely to experience improvement.

Design, Setting, and Participants  This retrospective case series assessed patients 20 years or older who were new to the LVR clinics from August 1, 2013, to December 31, 2015, and who had habitual VA between 20/40 and counting fingers (not including) and underwent refraction. Data analysis was performed from April 4, 2018, to December 20, 2019.

Exposures  Patient demographics and clinical data, including habitual and refraction VA, refraction, and disease diagnosis. Habitual VA was categorized as mild (VA worse than 20/40 and at least 20/60), moderate (VA worse than 20/60 and better than 20/200), severe (VA 20/200 or worse and better than 20/500), and profound (VA 20/500 or worse) vision impairment (VI).

Main Outcomes and Measures  At least 2-line VA improvement and any VA improvement (≥1-line) by refraction.

Results  Among the 2923 patients new to LVR clinics, 1773 (mean [SD] age, 70 [18.2] years; 1069 [60.3%] female) were included in these analyses. The mean habitual VA was 20/100 (mean [SD], 0.67 [0.36] logMAR). At least 2-line improvement was observed in 493 patients (27.8%), and any VA improvement was seen in 1023 patients (57.7%). At least 2-line improvement was observed in 54 patients (34.8%) with corneal disorders and was more likely seen among patients aged 40 to <65 years compared with those aged 20 to <40 years (odds ratio [OR], 1.57; 95% CI, 1.02–2.41), African American patients compared with white patients (OR, 1.41; 95% CI, 1.08-1.85), or patients with moderate VI compared with mild VI (OR, 1.36; 95% CI, 1.07-1.72).

Conclusions and Relevance  The findings suggest that URE is prevalent among patients with ocular disease and accessing LVR and that refractive evaluation should be considered for patients with ocular disease and reduced VA, especially working-age adults aged 40 to <65 years, African American patients, and those with moderate VI.

Uncorrected refractive error (URE) is the leading cause of vision impairment (VI) worldwide.^1,2 People with URE and VI experience increased challenges with everyday living, such as slower walking speed, worse balance capacity, problems reading, and difficulty seeing faces.^3-6 There is evidence that correcting refractive error is associated with improved quality of life and decreased symptoms of depression.^7,8 Thus, many public health endeavors and leading opinions⁹ have emphasized the importance of measuring and treating URE given that simple and relatively low-cost treatment brings gains in quality of life.^10,11

Much of the work^12-14 on URE has focused on identifying its magnitude and determinants in the general population. Many of these studies^12,15 defined worse than 20/40 visual acuity (VA) as VI and included populations with VA distributions skewed toward the near-normal levels. A recent population-based study¹⁶ among older adults in France with a mean age of 84 years highlighted the unexpected finding of high rates of URE (range, 35%-44%), defined as at least a 1-line improvement of the presenting distance VA in the better eye when using the best optical correction, in patients with ocular disease compared with 2% to 24% in previous population-based studies.^12-14,17 The authors suggest that URE may be overlooked by the treating ophthalmologist because the focus is on medical management in patients with reduced VA and concurrent ocular disease.¹⁶ For the current analysis, URE references refractive error that may be associated with an absent, incorrect, or outdated correction.

Sunness et al¹⁸ were among the first to examine the importance of URE in a group of patients new to low vision (LV) clinics with identified eye disease and found that 11% of the 739 patients had at least 2-line VA improvement after refraction. In patients with chronic VA loss who are accessing LV care, it is particularly important to maximize optical correction because better VA is associated with greater visual ability and a reduced demand for magnification.^19,20 In contrast to rehabilitation intervention, which focuses on improvement at the task level, improving VA in a patient with concurrent ocular disease has the potential to improve a person’s ability across all activities and, thus, is more restorative and potentially more meaningful. In addition, understanding the effects of refractive correction alone in patients with ocular disease may modify management and VA outcome expectations of medical and surgical interventions and when examining retrospective VA data in clinical research.

We conducted a comprehensive evaluation of the association of refraction with changes in VA in adult patients new to LV clinics in a large academic medical center. In this analysis, we examined changes in VA and describe age and racial/ethnic characteristics among all new patients seeking low vision rehabilitation (LVR) services at multiple clinic sites examined by multiple physicians.

In this retrospective case series, we extracted patient data from the Johns Hopkins Hospital electronic health record (EHR) system Epic. All patients with at least 1 completed LV office visit at the Wilmer Eye Institute from August 1, 2013, to December 31, 2015, were identified. Data analysis was performed from April 4, 2018, to December 20, 2019. The study was approved by the Johns Hopkins University School of Medicine institutional review board with an informed consent waiver. There was minimal risk because the project only collected existing EHR data. In addition, this research was conducted with a waiver of informed consent because there was no direct patient contact and all data were deidentified.

Quiz Ref IDPatients were defined as new if they had no previous LV clinic visit within the past 3 years. Patients were excluded if they (1) were younger than 20 years (because their refractive profiles may differ from those of older patients), (2) had habitual VA (HVA) of 20/40 or better or counting fingers (CF) or worse, and (3) had not undergone refraction (Figure 1). These patients had less potential to yield 2-line VA improvement after refraction (primary study outcome), were less likely to have refraction performed, or were likely to have less reliable change because the stimuli were not standardized (eg, CF) or did not involve measurement of the angular size of critical detail (eg, light perception). LVR services were provided by 6 specialists (J.E.G. and others) in the main urban hospital center (East Baltimore, Maryland) and 4 suburban satellite clinics.

Refraction was performed during patients’ initial office visits by an LVR specialist (J.E.G. and others). Clinic visit information extracted from their first LV visit during the study period included service date and location (urban hospital or suburban clinic). Demographic information included (1) date of birth, (2) sex, (3) race (white, African American, or other), and (4) ethnicity (Hispanic or not Hispanic). Age was grouped into (1) 20 to younger than 40 years, (2) 40 to younger than 65 years, (3) 65 to younger than 80 years, and (4) 80 years or older.

Distance VA data from the following EHR fields were included for analysis: right and left eye uncorrected, corrected, and manifest refraction and binocular uncorrected and corrected. HVA was defined as the best of the uncorrected and corrected VA before refraction and grouped into 4 categories: (1) mild VI (HVA worse than 20/40 and at least 20/60), (2) moderate VI (HVA worse than 20/60 and better than 20/200), (3) severe VI (HVA 20/200 or worse and better than 20/500), and (4) profound VI (HVA 20/500 or worse). Manifest refraction VA was generated using the same approach. All VA measures in fraction form were converted into logMAR values.

VA improvement from refraction was determined by the difference between HVA and manifest refraction VA on eye and person levels, and each 0.1-logMAR decrease was equivalent to 1-line improvement on the VA chart. Improvement by line was categorized as 1 line, 2 lines, 3 lines, and 4 lines or more. Any VA improvement was defined as at least 1-line improvement. Considering test retest variability in VA measurement,^21,22 the primary analysis outcome was defined as at least 2-line VA improvement, consistent with a previous clinical trial.²³ Although acceptable for clinical use, our analyses did not evaluate partial lines read because letters gained or lost is not consistently recorded as part of routine VA documentation practices and is less reflective of function (eg, hemianopia).

Manifest refraction data were extracted, and spherical equivalent was calculated as spherical power added to half cylindrical power for each eye. Refractive correction was considered prescribed when the final refraction EHR data element was available. However, information was unavailable for whether the patient filled the prescription.

Disease diagnosis codes (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-10]) were extracted from the identified clinic visits. Each office visit was associated with 1 or more disease diagnosis codes, and all ICD-10 codes were obtained. The following disease diagnosis categories were defined to be consistent with a previous report¹⁸: (1) age-related macular degeneration, (2) diabetic retinopathy, (3) retinal conditions other than age-related macular degeneration or diabetic retinopathy, (4) glaucoma, (5) neuro-ophthalmic conditions, and (6) corneal conditions. The first 3 categories were grouped as any retinal condition.

Descriptive statistics were used to compare patient demographics and service location by HVA categories. Distributions of VA improvement associated with refraction by HVA categories were plotted using percentage bar charts, and comparisons were made using the χ² test. Logistic regression models adjusting for age, sex, race, service locations, HVA, and refraction were used to determine the odds ratios (ORs) and associated 95% CIs of having a 2-line or greater VA improvement at the eye level after refraction. The absolute values of spherical equivalent regardless of the myopic or hyperopic nature of the refractive error were used for the regression analysis. For each disease category, logMAR VA improvement and refraction were shown as mean (SD). Two-sided P < .05 was considered statistically significant. All analyses were performed using Stata, version 15 (StataCorp).

A total of 3867 patients had at least 1 completed visit to the LV clinic; among them, 2923 were considered to be new patients. The analytic sample included the initial clinic visits for 1773 new patients (mean [SD] age, 70 [18.2] years; 1069 [60.3%] female) who met inclusion criteria (Figure 1). Of the patients included in the analysis, 777 (43.8%) had mild VI, 632 (35.6%) had moderate VI, 295 (16.6%) had severe VI, and 69 (3.9%) had profound VI in the better eye (Table 1). The mean HVA was 20/100 (mean [SD], 0.67 [0.36] logMAR). A total of 43 patients (62.3%) in the profound VI group were of working age (20 to <65 years) compared with 208 patients (26.8%) in the mild VI group, 202 (32.0%) in the moderate VI group, and 122 (41.4%) in the severe VI group (P < .001). Similarly, greater proportions of patients in the profound VI group were African American (31 [44.9%]) compared with patients in the mild VI (162 [20.9%]), moderate VI (172 [27.2%]), and severe VI (76 [25.8%]) groups (age-adjusted P = .05). No differences were found regarding the distributions of sex, ethnicity, or service locations.

Quiz Ref IDAt least a 2-line VA improvement was observed in 493 patients (27.8%) and any (≥1-line) VA improvement in 1023 patients (57.7%) after they underwent refraction. At least a 2-line VA improvement was seen in 199 patients (25.6%) with mild VI, 197 patients (31.2%) with moderate VI, 74 patients (25.1%) with severe VI, and 23 patients (33.3%) with profound VI. The greatest percentage of patients who gained at least a 4-line improvement was in the profound VI group (16 [23.2%]) (Figure 2). In addition, a 6-line or greater VA improvement was observed in 22 patients (1.2%). After refraction, 596 patients (33.6%) had corrected VA improvement to at least 20/40 (eTable in the Supplement). Any improvement was observed among 1006 of 1773 patients (56.7%) in the better eye and 562 of 1474 patients (38.1%) in the worse eye; a 2-line or greater VA improvement was observed among 477 of 1773 patients (26.9%) in the better eye and 302 of 1474 patients (20.5%) in the worse eye. Of the 1479 patients who had refraction performed in both eyes, the eye with the worse HVA became the better eye after manifest refraction in 86 patients (5.8%). Specifically, after refraction, the worse HVA eye became 1-line superior to the fellow eye in 56 patients (3.8%), 2-line superior to the fellow eye in 22 patients (1.5%), 3-line superior to the fellow eye in 5 patients (0.3%), and 4-line superior to the fellow eye in 3 patients (0.2%). Overall, the mean (SD) VA improvement associated with refraction was 0.11 (0.14) logMAR (Table 1).

Documented spherical and cylindrical data were available for 1745 patients (98.4%) who underwent refraction. Spherical equivalent refraction ranged from −28 to +30 diopters (D). With increasing VA loss, there was an increasing range of refractive error in the mild to severe impairment groups. Refractive error of at least 2.0 D was seen in 482 patients (27.6%), with higher proportions among patients with more severe VI. Overall, 1027 of 1773 patients (57.9%) were prescribed refractive corrections and were more frequently seen in the less-severe VI groups (eTable in the Supplement). In addition, 383 patients (77.7%) with a 2-line or greater VA improvement were prescribed refractive corrections compared with 644 (50.3%) of those without a 2-line or greater VA improvement (P < .001).

Quiz Ref IDLogistic regression analyses were used to assess the associations between patient characteristics and 2-line or greater VA improvement by refraction in the better HVA eye (Table 2). At least a 2-line improvement was more likely to be seen in patients aged 40 to <65 years (OR, 1.57; 95% CI, 1.02-2.41) compared with those aged 20 to <40 years. African American patients had greater odds (OR, 1.41; 95% CI, 1.08-1.85) of having such improvement compared with white patients. Compared with patients with mild VI, those with moderate VI were more likely to have a 2-line or greater VA improvement (OR, 1.36; 95% CI, 1.07-1.72). With every diopter increase of spherical equivalent refractive error, there was 1.05 odds (95% CI, 1.01-1.09) of gaining a 2-line or greater VA improvement. There was no association with sex and service location.

A total of 1269 patients (71.6%) had retina-related diagnoses, and 438 (24.7%) had glaucoma diagnoses. Patients were not exclusive to 1 disease category because they may have had more than 1 disorder diagnosis associated with an encounter. Overall, mean HVA was 20/100 for patients with retina-related diseases, glaucoma, and neuro-ophthalmic disease and 20/125 for patients with corneal disease. Despite the worst HVA, patients with corneal disease had the greatest improvement associated with refraction. Other than neuro-ophthalmic diseases, a 1-line or greater VA improvement was observed in more than half of patients with defined disease categories. At least 2.0-D spherical equivalent refractive error was most common among patients with corneal diseases (58 [37.9%]) and those with retinal diseases other than age-related macular degeneration or diabetic retinopathy (155 [37.7%]) (Table 3).

In this large, urban and suburban clinic population, a significant number of adult patients new to the LVR clinics with ocular disease and VA worse than 20/40 had URE. After refraction, more than one-fourth gained at least a 2-line improvement. Such improvement was more likely to be observed in patients aged 40 to <65 years, patients with HVA worse than 20/60 and better than 20/200, African American patients, and those with a higher degree of refractive error. In addition, 1 in 3 patients with corneal disease diagnoses had at least a 2-line improvement after refraction.

A total of 57.7% of the patients with VA worse than 20/40 gained at least 1 line, and 27.8% gained at least a 2-line improvement after refraction in our study. Previously, Sunness et al¹⁸ reported on the association of VA improvement with refraction in an LV clinic and found that 11% of new patients had VA improvement of 2 lines or more. Both our study and the study by Sunness et al¹⁸ evaluated patients new to the LV clinics in academic institutions in Maryland; however, the current study population was younger (mean age, 70 vs 73 years), had better presenting VA (median, 20/70 vs 20/80), and was less likely to have LV associated with a retinal disease diagnosis (71.6% vs 80.2%). The data from the study by Sunness et al¹⁸ only included patients with a 2-line or greater improvement after refraction by 1 practitioner using manual record review. The current study extracted clinic data from an EHR system for all new patient encounters with at least mild VI at an LV clinic that represented care by 6 practitioners. Although small, our study and the study by Sunness et al¹⁸ found proportions (5.8% in the current study and 3.0% in the study by Sunness et al¹⁸) of patients obtaining at least a 4-line improvement. This finding underscores the importance of considering URE in patients seeking LVR services.

Our findings showed that compared with patients aged 20 to 40 years with reduced VA, adults aged 40 to <65 years and African American patients were more likely to have URE and to benefit from refraction, with at least a 2-line improvement. Sunness et al¹⁸ did not find an association with age and did not assess the associations with sex or race. Our findings reveal the importance of checking refraction and maximizing VA in working-age adults to enhance performance in meeting occupational and personal demands. In addition, a change in refractive correction may be associated with simpler rehabilitation solutions, reduced visual fatigue, ongoing employment, and improved quality of life.

In regard to race, the 2005-2008 National Health and Nutrition Examination Survey found that even after adjusting for household income, educational level, and health insurance coverage, non-Hispanic African American individuals were 2.1 times as likely to have inadequate refractive corrections than white individuals.²⁴ Although not completely comparable, in our stratified regression analysis models that evaluated only patients with glaucoma diagnoses or patients with corneal disorder diagnoses (and maybe with other disease diagnoses), the associations for greater odds of African American patients gaining at least a 2-line VA improvement after refraction were retained. These associations may stem from inequality in eye care accessibility, differences in health insurance coverage, and living conditions.^25-27 More detailed examination of these associations is beyond the scope of this retrospective review.

Habitual VA and improvement after refraction showed variability by disease diagnosis. Patients with corneal or glaucoma diagnosis were most likely to have at least a spherical equivalent of 2.0-D refractive error. This finding may reflect a greater likelihood of irregular or increased astigmatism observed in corneal disorders and after glaucoma operations.²⁸ Patients with neuro-ophthalmic disease had the least improvement after refraction, suggesting that visual field loss (eg, nonarteritic anterior ischemic optic neuropathy with 20/25 VA) more than VA loss is the source of the impairment.^29,30 A greater proportion of patients with diabetic retinopathy gained at least a 1-line and 2-line VA improvement after refraction, which may be reflective of the refractive fluctuations experienced with fluctuating glucose levels.³¹

We observed a high prevalence of URE in patients with ocular disease seeking care in LV clinics. Understanding how patients were connected to LVR care (eg, via physician, self, or friend) would be meaningful to design outreach programs to improve refractive delivery. However, limitations on referral data documentation in the current EHR (eg, primary care physician default as the referral source) preclude reliable examination. Regardless, in an era of increasing specialization managing ocular disease, refraction is no longer considered a part of the domain of the ophthalmologic subspecialist and may easily be overlooked with a focus on disease management. Therefore, understanding the magnitude of the underlying URE and identifying patients whose VA may improve from refraction are important first steps in the coordination of care. As the evidence regarding the importance of vision and healthy aging increases, we believe that ophthalmologists caring for patients with reduced VA should emphasize and encourage refractive evaluations periodically to maximize patient function. Future work could explore implications of refractive correction on patient-centered outcome measures in those with concurrent ocular disease.

Quiz Ref IDThe strengths of this work include the large sample of all patients under ophthalmic care newly seeking treatment for LV, which represents a large spectrum of the patient population; clinical practice patterns in both urban and suburban settings; and varying demographic characteristics, with refractions performed by multiple LV specialists. Clinical measurements of VA, refraction, and concurrent ocular diseases were derived from 1 electronic platform despite multiple care locations, providing consistency in data collection and recording practices. The study design allowed us to examine any level of VA improvement, thereby providing a comprehensive profile of patients with VI, and facilitated comparisons with epidemiologic studies using multiple criteria. Moreover, examination of a 2-line or greater improvement in this group of patients with worse than 20/40 VA allowed reliable detection of a change in VA while minimizing the ceiling effect. However, limitations should be taken into consideration when interpreting the outcomes. Although this retrospective analysis leveraged EHR data and reflects real-world clinic practice, no masking of measurements was applied, and VA recordings may be influenced by physician and patient expectations. Refraction was less likely to be performed for patients without form perception VA, such as hand motion and light perception, because negligible improvements in this subgroup would be expected. We did not assess changes in quality of life in patients with VA improvement and recognize that patients with different disease diagnoses may benefit differently from VA improvement.³² For example, patients with corneal disease may perceive less benefit despite equivalent VA improvement after refraction given loss in contrast sensitivity; in addition, contact lens correction may be more meaningful than spectacle correction in this subgroup. Although 57.9% of the study population were prescribed changes in refractive correction, data on prescription filling were not available.

The findings suggest that URE is prevalent among patients treated for ocular disease and accessing LVR and that refractive evaluation should be considered for patients with ocular disease and reduced VA, especially adults aged 40 years or older, African American patients, and those with moderate VI.

Accepted for Publication: April 5, 2020.

Corresponding Author: Judith E. Goldstein, OD, Vision Rehabilitation Services, Wilmer Eye Institute, Johns Hopkins University, 600 N Wolfe St, Baltimore, MD 21287 (jgolds28@jhmi.edu).

Published Online: May 21, 2020. doi:10.1001/jamaophthalmol.2020.1677

Author Contributions: Drs Guo and Goldstein had full access to all the data and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: All authors.

Acquisition, analysis, or interpretation of data: Guo, Goldstein.

Drafting of the manuscript: All authors.

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

Statistical analysis: All authors.

Obtained funding: Goldstein.

Administrative, technical, or material support: Goldstein.

Supervision: Goldstein.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was funded by the Reader’s Digest Partners for Sight Foundation (Dr Goldstein).

Role of the Funder/Sponsor: The funding source 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.