We conducted a cross-sectional observational study among adult inpatients at the University of Chicago Medical Center, as part of the University of Chicago Hospitalist Project, a larger ongoing quality-of-care study [18]. We surveyed participants from August 2020 through March 2022.

Eligibility criteria included hospitalization on general medicine services, being 18 years or older, being English speaking, and consenting to the primary study.

The University of Chicago Biological Sciences Division Institutional Review Board approved this protocol (#IRB16-0763). The participants were consented for primary data collection and secondary analysis of research data. Study data are deidentified.

The primary outcome variables were self-reported audio-only and video-enabled telehealth usage and willingness to use video telehealth, captured by binary survey responses. The participants were asked if they had a telehealth appointment prior to the pandemic (defined as before March 2020), if they had a telehealth appointment since the pandemic (defined as after March 2020), and whether they were willing to use video telehealth visits.

As mentioned, technology access and home internet access were self-reported and captured by binary survey responses. To assess technology capabilities, the participants were asked if they knew how to perform given web-based tasks, including the capability to use video without needing help.

The participants were categorized as having either low eHL or adequate eHL based on the eHealth literacy scale (eHEALS), a validated 8-item questionnaire with each item scored on a 5-point Likert scale [16]. The resultant composite score is between 8 and 40. Following other relevant studies, low eHL was defined as a score of <26, and adequate eHL was defined as a score of ≥26 [19,20].

Health literacy (HL) was based on the brief health literacy screen, which is a 3-item survey scored on a Likert scale from 0 to 4 [21]. Low HL is defined by scoring 2 or less on any item in the brief health literacy screen.

We assessed self-reported telehealth usage (audio-only versus video-enabled), willingness to use video visits and technology access, as well as demographic characteristics, across eHL levels using descriptive statistics. Chi-squared tests were used to calculate P values. Statistical significance was set at P<.05.

We then used a multivariable binary logistic regression to determine the odds ratio for being unwilling to use video telehealth, adjusted for age, sex, race or ethnicity, educational level, eHL, HL, technology access, internet access, and technology capability. All analyses were performed using STATA version 16.1 (StataCorp LLC).

Of the 297 adults included in this study, the majority identified as Black (n=185, 62%), half identified as female (n=149, 50%), and the median age was 58 years (25th percentile to 75th percentile: 42-68; Table 1). The median eHEALS score was 30 (25th percentile to 75th percentile: 21-33; Table 1).

Participants with adequate eHL had significantly higher rates of college educational attainment (79/196, 40% vs 18/101, 18%), smartphone ownership (186/196, 95% vs 64/101, 63%), laptop ownership (122/196, 62% vs 20/101, 20%), tablet ownership (79/196, 40% vs 20/101, 20%), and video capability (185/196, 91% vs 48/101, 48%) and fewer households without internet access (26/196, 13% vs 40/101, 40%) compared with participants with low eHL (all P<.001; Table 1).

Low eHL was associated with lower rates of web-based skills, such as the ability to use video without help (P<.001; Figure 1). Participants with low eHL were more likely to need help with using video technology than those with adequate eHL (53/101, 52% vs 11/196, 6%; P<.001; Figure 1).

Fewer participants with low eHL reported a video telehealth appointment compared with those with adequate eHL (34/101, 34% vs 120/196, 61%; P=.007; Figure 2). In contrast to video visits, audio-only telehealth usage during the pandemic did not differ significantly between eHL subpopulations (63/101, 62% vs 129/196, 66%; P=.56).

Using a multivariable binary logistic regression, we found that unwillingness to use video technology for telehealth visits was significantly associated with older age (P=.03), low eHL (P=.003), lack of video capability (ie, needing help using video; P=.04), and lack of home internet access (P=.03; Table 2). In contrast to eHL, HL was not a significant predictor for unwillingness to use video technology (P=.48; Table 2).

Telehealth adoption increased across the board during the pandemic. However, our study found significant disparities in telehealth usage when comparing the 2 telehealth modalities. Patients with low eHL reported greater participation in audio-only telehealth over video telehealth, of which the former may lose its flexible pandemic reimbursement policy. This has the potential to widen the existing health disparities as older adults and patients with low eHL face challenges in transitioning to video telehealth services. The use of audio-only visits was similar across most demographic subgroups; however, the rates of video visits and self-reported ability to use video technology varied significantly by eHL level.

Our findings are consistent with prior studies showing disparities in audio-only versus video telehealth use during the pandemic [2,8,9]. Video telehealth requires additional access to computers and home broadband or smartphones. However, Black and Hispanic adults and adults from low-income households are less likely to report access to home broadband and computers [10,22,23]. Some reasons for low use of video visits are rooted in historical redlining policies that denied services to low-income communities and communities of color, and, more recently, the “digital redlining” of broadband as internet service providers build new fiber networks [24,25]. The ramifications of this on equitable digital patient care is of particular importance in an increasingly digital environment.

Beyond the gaps in broadband access and technology ownership, importantly, we show that gaps in digital skills, specifically low electronic health literacy, were also a driver in video telehealth disparities. Low eHL and lack of video capability reduced the likelihood of being willing to engage in video-enabled telehealth. After multivariable adjustment, the willingness to use video telehealth was significantly lower among participants who were older, lacked home internet access, had low eHL, and needed help using video technology. These results are consistent with prior literature that demonstrate how patients still face challenges in knowing how to use video technology to improve their health despite having access to smartphones and internet [14,17,26].

There are several limitations to this study. This was an urban, single-site study; therefore, the results may not be generalizable to rural locations where a greater proportion of patients may lack broadband access. The study includes only English-speaking patients, so the findings may differ among patients with limited English proficiency who may face additional challenges in telehealth use. Additionally, the study uses patient self-reported measures for telehealth use and technology ownership. While this could introduce recall and social desirability bias, our study design has several strengths. The self-reported nature of the study allows for additional information on patient telehealth preferences that is not captured in claims data. Moreover, the survey was administered over the phone, which reduces the chance of bias from internet surveys that self-select for patients with greater technology capability.

Our study contributes to the literature by showing the positive association of adequate eHL with the patients’ willingness to engage in video visits, and conversely, the negative relationship between low eHL and video engagement. The study results raise the question of how to improve video capabilities and skills among patients who have access to smart devices but face challenges in accessing the internet and operating video equipment. Moving forward, additional research should investigate what interventions will increase digital skills, particularly eHealth literacy, in patients who encounter barriers to using video telehealth.