Telemedicine and the Exacerbation of Health Care Disparities

The COVID-19 pandemic has increased health disparities in vulnerable and marginalized populations (eg, people with disabilities, immigrants, and people who are non–English speaking) across all areas of medicine. Some of the direct health effects of the pandemic include the disproportionate risk racial and ethnic minority groups face in contracting COVID-19 because of their higher representation in jobs that do not allow for work from home. Similarly, given the higher rates of comorbid diseases like diabetes and obesity, racial and ethnic minority populations are at increased risk of mortality once infected.¹ While these headlines have rightly received the most attention, the pandemic shutdown has also increased health disparities in other subtle and unanticipated ways. The promise of telemedicine has always been in its convenience for patients, allowing them to more easily fit a “visit” to the physician into an otherwise busy schedule without suffering the drawbacks of an in-person visit like parking or having to wait in a waiting room. During the pandemic, an additional benefit was the ability to obtain care without risking infection by breaking home quarantine. Although telemedicine has been promoted as a way to improve access, we now have evidence that like many other aspects of medicine, use of telemedicine has not spread equally across the population.

It is in this context in which Aziz and colleagues² present their findings from a large tertiary care practice reporting which patients were more or less likely to complete a telemedicine visit during the strictest measures of the United States 2020 pandemic shutdown. They found that patients who were older, African American, or did not speak English had a much lower odds of completing a telemedicine visit using video. They also found that patients who had less education, were unemployed, were retired, or had a disability were more likely to complete a telephone visit as compared with a video visit among telemedicine service users.² While certainly not the intention of telemedicine proponents, this is yet another data point in how racial and ethnic minority and other vulnerable populations can be disproportionately impacted by unintended consequences.

This study is consistent with other recent findings by Elam and colleagues³ who examined the telemedicine rates of a different academic tertiary medical center and reported that older patients were significantly less likely to use telemedicine. They also found a concerning trend that non-White patients were less likely to use telemedicine.³ Fortunately, the 2 studies diverged with regards to in-person visits during 2020, with Elam et al finding a disparity in the rates non-White patients attended in-person visits compared with White patients. However, Aziz et al actually reported slight increases in Black and Hispanic or Latino patients attending an in-person visit during 2020.^2,3 While evidence that disproportionate care occurred in any single practice is not a good thing, the fact that differences exist across practices suggests the problem is not uniform throughout the country. Efforts should be made to understand the underlying differences between the 2 practices that may have led to the disparate results.

While the Aziz et al study represents an excellent step forward in highlighting a disparity, a goal for future studies should be to understand how the characteristics of the practice being evaluated may impact the results. Are all physicians within a practice participating equally in telemedicine? If not, are the specialties/physicians that are participating more likely to see disease states that are more common in one race or another (ie, glaucoma, age-related macular degeneration), thereby altering the underlying population of who truly has an opportunity to be seen via telemedicine? In a similar manner, when comparing across practices, are some using a hybrid model that includes ancillary testing in addition to the virtual physician visit? A hybrid setup may give more access to patients with nonanterior diseases, again altering the population of who would be eligible for telemedicine. Portney and colleagues⁴ recently reported from a large national cohort that the most common diagnoses associated with telemedicine visits during the US shutdown were cornea or external disease, which accounted for 48.0% of all telemedicine visits (chalazia specifically was diagnosed on 9.4% of all visits). In the Aziz et al study, they found that of the 2262 telemedicine visits in 2020, 1355 (59.9%) happened either in the cornea, optometry, oculoplastics, or comprehensive service, again suggesting anterior or external disease is likely to be more heavily weighted in any assessment of telemedicine.² Given this and the increased prevalence of diseases more suited to being monitored via an eye examination among racial and ethnic minority patients (eg, diabetic retinopathy or glaucoma), evaluations looking to assess disparities in care should concurrently measure in-person visits alongside telemedicine visits.

Ophthalmology was in a difficult position during the shutdown in 2020 as it was not only the single most negatively impacted specialty with regards to patient volume loss, but also the specialty with the fewest telemedicine users prior to the shutdown.^5,6 Ophthalmologists across the country were forced to change course quickly and, in many cases, create telemedicine programs de novo. While rates of telemedicine visits have decreased as society has reopened, it is likely that more encounters will move toward telemedicine in the future as artificial intelligence (a technology with its own concerns over inherent biases) progresses and becomes a tool used in the daily practice of ophthalmology. As this transition happens, we need to be vigilant to not only prevent the exacerbation of existing health care disparities, but to identify and reduce them wherever they exist.

Corresponding Author: Brian L. VanderBeek, MD, MPH, MSCE, Scheie Eye Institute, 51 N 39th St, Philadelphia, PA 19104 (brian.vanderbeek@pennmedicine.upenn.edu).

Published Online: September 23, 2021. doi:10.1001/jamaophthalmol.2021.3735

Conflict of Interest Disclosures: None reported.