Association of Patient Characteristics With Delivery of Ophthalmic Telemedicine During the COVID-19 Pandemic

Key Points

Question  During the first year of the COVID-19 pandemic in the US, were there differences in the characteristics of patients who received ophthalmic care via telemedicine compared with in-person care?

Findings  In this cross-sectional study of 1911 patients from a single academic ophthalmology practice, patients who were men, self-identified as Black, did not speak English, had an educational level of high school or less, and were of older age were less likely to receive telemedical care compared with in-person care.

Meaning  These results suggest that disparities in the delivery of ophthalmic telemedical care existed during the COVID-19 pandemic and support prioritizing health equity in future telemedicine programs.

Importance  Telemedicine has been shown to have had reduced uptake among historically marginalized populations within multiple medical specialties during the COVID-19 pandemic. An evaluation of health disparities among patients receiving ophthalmic telemedical care during the pandemic is needed.

Objective  To evaluate disparities in the delivery of ophthalmic telemedicine at Massachusetts Eye and Ear (MEE) during the COVID-19 pandemic.

Design, Setting, and Participants  This retrospective, cross-sectional study analyzed clinical visits at a single tertiary eye care center (MEE) from January 1 to December 31, 2020. Patients who had ophthalmology and optometry clinical visits at the MEE during the study period were included.

Exposures  Telemedicine vs in-person clinical encounters.

Main Outcomes and Measures  Variables associated with use of ophthalmic telemedicine during the study period.

Results  A total of 2262 telemedicine ophthalmic encounters for 1911 patients were included in the analysis. The median age of the patients was 61 (interquartile range, 43-72) years, and 1179 (61.70%) were women. With regard to race and ethnicity, 87 patients (4.55%) identified as Asian; 128 (6.70%), as Black or African American; 23 (1.20%), as Hispanic or Latino; and 1455 (76.14%), as White. On multivariate analysis, factors associated with decreased receipt of telemedical care included male sex (odds ratio [OR], 0.86; 95% CI, 0.77-0.96), Black race (OR, 0.69; 95% CI, 0.56-0.86), not speaking English (OR, 0.63; 95% CI, 0.48-0.81), educational level of high school or less (OR, 0.83; 95% CI, 0.71-0.97), and age (OR per year of age, 0.99; 95% CI, 0.989-0.998). When comparing telephone- and video-based telemedicine visits, decreased participation in video-based visits was associated with age (OR per year of age, 0.96; 95% CI, 0.94-0.98), educational level of high school or less (OR, 0.54; 95% CI, 0.29-0.99), being unemployed (OR, 0.28; 95% CI, 0.12-0.68), being retired (OR, 0.22; 95% CI, 0.10-0.42), or having a disability (OR, 0.09; 95% CI, 0.04-0.23).

Conclusions and Relevance  The findings of this cross-sectional study, though limited to retrospective data from a single university-based practice, suggest that historically marginalized populations were less likely to receive ophthalmic telemedical care compared with in-person care during the first year of the COVID-19 pandemic in the US. Understanding the causes of these disparities might help those who need access to virtual care.

The COVID-19 pandemic has negatively affected the ability of ophthalmologists to care for patients, with ambulatory outpatient visits declining by 79% during the height of the pandemic.¹ By the end of 2020, the volume of outpatient visits was 18% less than expected.² Telemedicine became mandatory in many clinical settings as an alternative model of care for patient evaluation while maintaining strict social distancing standards and decreasing the risk of viral transmission.³ Regulatory changes implemented at the outset of the pandemic, such as payment equivalency for telemedicine and in-person visits as well as the loosening of Health Insurance Portability and Accountability Act compliance standards for virtual visit platforms, all supported the use of telemedicine during the pandemic.^4,5 Some ophthalmologists rapidly implemented telemedicine at the outset of the pandemic to maintain patient care. One study³ showed the proportion of ophthalmic visits using telemedicine peaking at 17% in early April 2020, whereas telemedicine represented only 0.04% of all ophthalmic visits in the preceding 6 months.

However, this rise in telemedicine did not reach all patient groups equally, potentially exacerbating health care disparities, as reported in other fields of medicine.^6-8 We herein assess the demographic characteristics of patients who did and did not use telehealth for ophthalmic care at a large tertiary care eye center during the COVID-19 pandemic.

This retrospective cross-sectional study of ophthalmic clinical encounters from January 1 to December 31, 2020, was performed at Massachusetts Eye and Ear (MEE), a large academic tertiary eye care center located in Boston. Administrative records were used to extract claims data for ophthalmic clinical encounters that occurred during the study period at MEE. Medical record numbers were used to link claims to individual patients. Encounter types within the electronic medical record system used by MEE (Epic; Epic Systems Corporation) were used to identify telemedicine and in-person visits, which were corroborated through billing practices with programmatic Current Procedural Terminology (CPT) modifier assignment. Approval was obtained from the institutional review board of Mass General Brigham, and the study adhered to the tenets of the Declaration of Helsinki.⁹ Informed consent was waived by the board because it would be costly and time consuming to locate every patient seen at MEE during 2019 and 2020 to obtain consent, and the scientific validity of the study would be compromised if data were limited to individuals from whom we were able to obtain informed consent. All data were protected by privacy safeguards.

Telemedicine visits were either initiated by patients through an urgent hotline or were offered to patients by their ophthalmologist. Institutional guidelines encouraged the use of video visits whenever possible and telephone visits only when video visits could not be conducted. Telemedicine visits included both video visits and telephone visits. Codes from the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10), were used to identify visit diagnoses, which were subsequently classified into clinically meaningful diagnostic categories based on the literature.¹⁰ The Healthcare Common Procedure Coding System and CPT codes were evaluated for each visit. Data from 2019 were used to compare the demographic characteristics of patients seen before and during the COVID-19 pandemic.

The primary outcomes of the study were the demographic characteristics of patients who received telemedical care during the COVID-19 pandemic, including race and ethnicity, primary language spoken, educational level attained, insurance status, and age. Secondary outcomes included use of telemedicine by ophthalmic subspecialty and visit characteristics, including encounter type (telephone vs video), new vs established patients, and visit diagnoses. Patients were categorized as established if they had been seen at MEE within the preceding 3 years. Patients receiving both telemedicine and in-person care were categorized as having telemedicine visits for the purposes of this analysis, regardless of the order of visits. Similarly, any patient who received both a telephone-based and a video-based telemedicine visit was categorized as having a video-based telemedicine visit for the purposes of this analysis, regardless of visit order.

Analyses were performed using R, version 3.6.3 (R Foundation for Statistical Computing). Statistical tests were 2-sided, and α was set to .05. The distribution of continuous numerical data was checked both graphically (histogram and boxplots) and via the Shapiro-Wilk normality test. Normally distributed continuous variables are reported as mean (SD). Continuous variables that were not normally distributed are reported as median (interquartile range [IQR]). Categorical variables are reported as numbers and percentages.

Continuous variables were assessed via the Kruskal-Wallis test. Categorical variables were assessed via the Pearson χ² test with Yates continuity correction. No adjustments were made for multiple analyses or outcomes.

Next, data were assessed using multivariate logistic regression analysis to determine the demographic factors associated with telemedicine vs in-person visits. A separate model was created to determine the demographic factors associated with telephone-only telemedicine vs video-based telemedicine visits through a similar process. The final multivariate model included age, sex, race and ethnicity, educational level, employment status, and primary language.

In 2020, a total of 155 131 ophthalmic clinical visits took place at MEE, 2262 (1.46%) of which were telemedicine visits (Table 1). The 2262 telemedicine visits included 1911 patients with a median age of 61 (interquartile range, 43-72) years; 1179 (61.70%) were women and 732 (38.30%) were men. With regard to race and ethnicity, 87 patients (4.55%) identified as Asian; 28 (6.70%), as Black or African American (herein after referred to as Black); 23 (1.20%), as Hispanic or Latino; and 1455 (76.14%), as White.

The visual rehabilitation service had the highest proportion of telemedicine visits (104 of 898 [11.58%]), followed by neuro-ophthalmology (302 of 5808 [5.20%]) and optometry (372 of 9761 [3.81%]). The retina service had the lowest proportion of telemedicine visits (146 of 41 551 [0.35%]), followed by the uveitis service (20 of 3840 [0.52%]) and oculoplastics service (50 of 6336 [0.79%]). According to visit volume, the cornea (n = 486), comprehensive ophthalmology (n = 447), and optometry services (n = 372) had the most telemedicine visits and the uveitis service (n = 20) and inherited retinal disease service (n = 20) had the fewest. Visit type and diagnostic information were unavailable for 129 telemedicine visits and 4009 in-person visits; these encounters were included in the overall number of telemedicine visits but were excluded from visit characteristic analyses.

In 2019, a total of 162 250 ophthalmic clinical encounters took place at MEE (Table 2). A comparison between 2019 and 2020 encounters demonstrates differences in patient demographics. Compared with 2019, there was a small increase in the proportion of female patients (39 233 [57.97%] vs 41 463 [57.93%]; difference [∆], 0.04% [95% CI, –0.47% to 0.57%]; P < .001), but this finding was not clinically significant. Patients who identified as Black (5967 [8.82%] vs 5942 [8.30%]; ∆, 0.52% [95% CI, 0.22% to 0.81%]; P < .001) and other race (includes American Indian or Alaska Native, Native Hawaiian or other Pacific Islander, and other) (5987 [8.85%] vs 5819 [8.13%]; ∆, 0.72% [95% CI, 0.42%-1.01%]; P < .001) represented a larger proportion of patients in 2020 compared with 2019. In contrast, patients who identified as Asian (3830 [5.66%] vs 4368 [6.10%]; ∆, –0.44% [95% CI, –0.69% to –0.19%]; P < .001) represented a smaller proportion of total patients in 2020 compared with 2019. In 2020, there was an increase in the proportion of patients with private insurance (36 589 [54.07%] vs 37 916 [52.97%]; ∆, 1.10% [95% CI, 0.57%-1.62%]; P < .001) and a decrease in the proportion of patients who were insured through Medicare (22 329 [32.99%] vs 24 314 [33.97%]; ∆, –1.01% [95% CI, –1.47% to –0.48%]; P < .001) or were uninsured (1300 [1.92%] vs 1586 [2.22%]; ∆, −0.30% [95% CI, –0.45% to –0.14%]; P < .001). There was an increase in the proportion of employed patients (27 367 [40.44%] vs 27 198 [38.00%]; ∆, 2.44% [95% CI, 1.93%-2.96%]; P < .001) and patients who were students (2174 [3.21%] vs 2149 [3.00%]; ∆, 0.21% [95% CI, 0.03%-0.39%]; P = .02) and a decrease in the proportion of retired patients (17 611 [26.02%] vs 19 784 [27.64%]; ∆, –1.62% [95% CI, –2.08% to –1.15%]; P < .001) in 2020. The proportion of patients who spoke English as their primary language was increased in 2020 (59 729 [88.26%] vs 62 544 [87.38%]; ∆, 0.88% [95% CI, 0.53%-1.22%]; P < .001). In 2020, there was a relative decrease in the proportion of patients who had a postgraduate education (10 635 [15.72%] vs 11 537 [16.12%]; ∆, –0.40% [95% CI, –0.79% to –0.02%]; P < .001) and a relative increase in patients with a college education (21 236 [31.38%] vs 21 646 [30.24%]; ∆, 1.14% [95% CI, 0.65% to 1.62%]; P < .001). All other demographic characteristics were similar between 2019 and 2020.

A total of 1911 patients participated in 2262 telemedicine visits during 2020, and a total of 65 763 patients participated in 147 211 in-person visits (Table 3). The median age of telemedicine patients was 61 (IQR, 43-72) years compared with 63 (IQR, 49-72) years for in-person patients (P < .001). On multivariate analysis, older age was associated with decreased receipt of telemedical care (odds ratio [OR] per year of age, 0.99; 95% CI, 0.989-0.998). Furthermore, male sex was associated with decreased receipt of telemedical care compared with female sex (OR, 0.86; 95% CI, 0.77-0.96), Black patients were less likely to receive telemedical care compared with White patients (OR, 0.69; 95% CI, 0.56-0.86), and those with an educational level of high school or less were less likely to receive telemedical care compared with postgraduate-educated patients (OR, 0.83; 95% CI, 0.71-0.97). Patients who did not speak English as their primary language were also less likely to participate in telemedicine visits compared with those who spoke English (OR, 0.63; 95% CI, 0.48-0.81).

A total of 150 patients (159 visits) had telephone-based telemedicine visits, and 1761 patients (2103 visits) had video-based telemedicine visits (Table 4). On multivariate analysis, increasing age was associated with decreased odds of taking part in a video-based visit compared with telephone visits (OR per year of age, 0.96; 95% CI, 0.94-0.98). Similarly, those with an educational level of high school or less were less likely to participate in video visits compared with patients who had a postgraduate education (OR, 0.54; 95% CI, 0.29-0.99). Patients who were unemployed (OR, 0.28; 95% CI, 0.12-0.68) or retired (OR, 0.22; 95% CI, 0.10-0.42) or had a disability (OR, 0.09; 95% CI, 0.04-0.23) were also less likely to participate in video visits compared with employed patients.

Of the 1911 patients who were seen via telemedicine, 426 (22.29%) were new patients, whereas 1485 (77.71%) were established patients (eTable in the Supplement). On multivariate analysis, established patients were more likely to be older than new patients (OR per year of age, 0.97; 95% CI, 0.96-0.98).

Of the 2262 telemedicine visits, the 5 most common visit diagnoses based on ICD-10 codes were open-angle glaucoma (n = 187), hordeolum/chalazion (n = 159), lacrimal gland dysfunction (n = 122), postoperative care (n = 80), and cataract (n = 64). For video-based telemedicine visits, the 5 most common visit diagnoses were open-angle glaucoma (n = 163), hordeolum/chalazion (n = 153), lacrimal gland dysfunction (n = 110), postoperative care (n = 74), and cataract (n = 58). For telephone telemedicine visits, the 5 most common diagnoses were open-angle glaucoma (n = 24), lacrimal gland dysfunction (n = 12), giant cell arteritis (n = 9), secondary glaucoma (n = 7), and cataract (n = 6).

Before the COVID-19 pandemic, almost all clinical encounters at MEE were in-person visits, with a weekly mean of 3844 (range, 2831-4333) visits (Figure). However, beginning in March 2020, this number declined, with a nadir of 583 visits in the week of March 23 to 27. This was accompanied by a corresponding rise in the number of weekly telemedicine visits, which peaked from May 10 to 16 (181 telemedicine visits per week). There was a subsequent decline until the end of August 2020, when weekly telemedicine visits plateaued at 22.4. The number of in-person visits began to increase in April 2020 and reached near prepandemic levels in July 2020, with 3700 in-person visits weekly.

We found disparities in the demographics of patients who used telemedicine for ophthalmic care in 2020 in a retrospective cross-sectional study at a university-based practice. Although telemedicine is increasingly considered to be 1 approach to improve access to care and decrease health care disparities, reports of even greater disparities in the use of telemedicine have surfaced across multiple fields of medicine^6-8 and are confirmed herein.

Patients who are men, identify as Black, have an educational level of high school or less, and do not speak English as a primary language were less likely to receive telemedical care during the pandemic. Other medical specialties^6-8,11 report similar disparities in care among these groups, including decreased use of telemedicine among patients who are Black, are men, have a rural residence, and have a lower income. Our findings highlight the fact that implementation of telemedical care does not necessarily improve access to care for all populations, and overreliance on telemedicine using current approaches may inadvertently increase health disparities for historically marginalized populations. We hypothesize that barriers to the use of telemedicine in those populations who did not access such care include limitations in the availability and affordability of high-speed broadband, lack of access to suitable electronic devices, and lower technological and health literacy.^12-15

Recognizing that access to video technology may be an important barrier to using telemedicine, we further evaluated the use of telephone-based vs video-based care. Once again, similar disparities were seen, with older, less educated, and poorer populations using video less often. Moreover, we noted that hordeolum/chalazion, a diagnosis that benefits from examination of the ocular adnexa, was common during video visits but was not among the top diagnoses for telephone visits. This may be owing to the benefit of external ophthalmic evaluation using video-based telemedicine, which can be used to confirm diagnoses or rule out more concerning pathologic features. Most video- and telephone-based visits were provided to established patients. When comparing established vs new patients in a multivariate analysis, being younger was the only factor independently associated with increased odds of being a new patient. This finding suggests that older patients may be less likely to initiate care through telemedicine visits.

Our study found that certain subspecialties, such as vision rehabilitation (11.58%), neuro-ophthalmology (5.20%), and optometry (3.81%), saw a larger proportion of their patients through telemedicine visits. This may be due to a combination of factors. First, it is possible to evaluate external and neuro-ophthalmic conditions via telemedicine video visits, which is supported by the literature.¹⁶ Vision rehabilitation often works to improve patient visual outcomes through counseling and assistive devices, which can be performed in a video-based setting and do not require detailed clinical examinations. Indeed, 2 of the 5 most common diagnoses made via virtual visits involved the exterior of the eye, allowing for easy examination without a slitlamp. In contrast, the retina service saw the smallest proportion of telemedicine visits (0.35%). This can be attributed to the need for posterior segment imaging and evaluation that is very difficult to perform via video- or telephone-based telemedicine visits. Although prior literature shows that oculoplastic virtual eye examinations are feasible,^17-19 there was not a marked uptake of telemedicine visits in our oculoplastics department (50 of 6336 visits [0.79%]).

Similar to previous studies,³ our study identified a rapid uptake in telemedicine use at the onset of the COVID-19 pandemic, which hit its peak within the first few months and tapered off to a steady state by September 2020. In March 2020, rapidly evolving policies were made at the national, state, and local levels, starting with the declaration of a global pandemic on March 11 by the World Health Organization and culminating in multiple regulations to promote the implementation of telemedicine.^4,5 Notably, the American Academy of Ophthalmology published guidelines to cease all nonemergent ophthalmic treatment on March 18, 2020.²⁰ There was a subsequent sharp decrease in the number of in-person visits at MEE (although all clinical services remained available) and a commensurate increase in the number of telemedicine visits. On June 1, 2020, MEE announced its reopening plan, after which the number of telemedicine visits decreased.

Although our study focused on disparities in the use of ophthalmic telemedicine, we also found that the most vulnerable groups were less likely to use in-person care during the pandemic compared with 2019. This includes patients insured through Medicare, uninsured patients, those who were unemployed or retired, and those who did not speak English as their primary language. It is notable that the year-to-year difference in proportions of patient visits is in many cases small and possibly not clinically relevant, and that these changes may be attributable to random fluctuation. However, disparities in access to eye care were known to exist even before the pandemic.^21,22 Outreach to vulnerable populations is important because they may have increased difficulty accessing both in-person and virtual care during such a crisis.

Our study has several limitations. First, its retrospective nature may introduce information bias because the data may contain errors, including coding errors and inconsistencies. In addition, some claims had missing information and were therefore excluded. Moreover, clinical information such as disease severity, visual and clinical outcomes, and disease management were not evaluated. Systemic and ocular comorbidities have higher prevalence rates among historically marginalized patient populations, which may have influenced who accessed in-person care.^23-29 Patients with systemic comorbidities may have opted for telemedicine over in-person care owing to the risk of coronavirus infection; conversely, patients who received in-person care for other conditions may have felt more comfortable attending in-person ophthalmic visits. Patients with ocular comorbidities requiring in-person management (eg, intraocular injections for age-related macular degeneration) may have been more likely to attend in-person visits and forgo telemedical care. The study may also be subject to selection and ascertainment bias, because telemedicine visits were not offered to all patients in a standardized fashion. The clinical judgment of ophthalmologists dictated whether a telemedicine visit was offered to a patient, and not all ophthalmologists used telemedicine during the pandemic. Other patients had to request an urgent telemedicine visit through an established hotline. Last, the study involved a single practice at a university in New England and may not be generalizable to other practices in New England or the US or to other practice settings including community-based practices. Future studies should address clinical outcomes related to changes in clinical care resulting from the pandemic, especially among diseases like diabetic retinopathy and glaucoma that are more prevalent in historically marginalized populations. Studies addressing best practices in ophthalmic telemedicine use and patient education are also needed.

In this retrospective cross-sectional study, we identified disparities in the delivery of ophthalmic telemedical and in-person care during the pandemic. Telemedicine was used less often by patients who were older, were men, were non-English speakers, had an educational level of high school or less, and identified as Black. In addition, video-based telemedicine visits, as opposed to telephone-based visits, were used less often by older patients; patients who were unemployed, retired, or disabled; and patients with an educational level of high school or less. The potential exacerbation of health inequalities through the use of ophthalmic telemedicine highlights the importance of focusing on equitable health care delivery through telemedicine in the future.

Accepted for Publication: July 30, 2021.

Published Online: September 23, 2021. doi:10.1001/jamaophthalmol.2021.3728

Corresponding Author: Grayson W. Armstrong, MD, MPH, Department of Ophthalmology, Massachusetts Eye and Ear, 243 Charles St, Boston, MA 02114 (grayson_armstrong@meei.harvard.edu).

Author Contributions: Dr Armstrong and Ms Moon had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Aziz, Moon, Parikh, Lorch, Miller, Armstrong.

Acquisition, analysis, or interpretation of data: Aziz, Moon, Parikh, Lorch, Friedman, Armstrong.

Drafting of the manuscript: Aziz, Parikh, Friedman, Armstrong.

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

Statistical analysis: Moon, Armstrong.

Administrative, technical, or material support: Parikh, Lorch, Miller, Armstrong.

Supervision: Aziz, Parikh, Lorch, Miller, Armstrong.

Conflict of Interest Disclosures: Dr Parikh reported receiving personal fees from Anthem Blue Cross Blue Shield during the conduct of the study. Dr Miller reported receiving personal fees from Alcon, Allergan PLC, ZEISS, Genentech, Inc, Sunovion Pharmaceuticals Inc, and Radius Health, Inc, outside the submitted work. Dr Armstrong reported receiving personal fees from Kriya Therapeutics, Ocular Technologies Inc, and the American Medical Association outside the submitted work. No other disclosures were reported.