We used a stratified, purposeful sampling strategy [7] to identify participants using the Mayo Clinic electronic health record. The eligibility criteria were (1) age greater than 18 years as of April 2020; (2) established patient at Mayo Clinic Midwest (Rochester or Mayo Clinic Health System), Florida, or Arizona; and (3) attended f2f appointments at the Mayo Clinic but did not utilize Mayo Clinic video appointment services from April 2020 to December 2020. Our goal was to include an equal number of participants by race, sex, location (Midwest, Florida, and Arizona), and outpatient primary care practice and psychiatry. We chose to focus on both primary care and psychiatry as they are settings where patients visit for health maintenance monitoring and medical management that are nonemergent, are noninvasive, and include a broad number of medical and psychiatric conditions, therefore making them suitable for a video appointment. We were especially interested in psychiatry patients and those seeking mental health care due to mounting evidence showing that anxiety and stress have increased significantly among these patients during the COVID-19 pandemic [8]. We estimated we would need about 36 interviews (12 interviews per Mayo Clinic enterprise location) to achieve data saturation whereby no new themes are being identified [9-11]. Furthermore, the goal of our purposeful sample was to glean rich in-depth information from individual interviews, and we plan to use our results to design and subsequently administer a quantitative survey assessing barriers [12]. To assess saturation, we reviewed interviews completed on an ongoing basis to assess if we were capturing an adequate range of patient experiences with our interviews.

Participants’ resident zip code was used to determine rurality based on Rural-Urban Commuting Area (RUCA) codes. For this study, we defined rural as RUCA 4-10 (vs urban if RUCA=1-3) [13]. Zip code data also allowed us to use the 2015 American Community Survey to estimate median household income as a surrogate marker for participants’ socioeconomic status [14,15].

Eligible patients were recruited via email, the Mayo Clinic Patient Portal, telephone, and mail. Social media ads were also posted on Mayo Clinic’s Facebook and Twitter accounts so patients could call to participate in the study. Patients who responded were enrolled in the study after being screened via phone and deemed eligible. Patient recruitment ended when we met the target sample size of 36 participants. Once consent was obtained, participants completed semistructured interviews via telephone, and interviews were digitally transcribed. Transcriptions were then verified for accuracy by 2 study team members (NMR, IWW). After completing the interview, participants were provided with a US $25 remuneration in the form of a cash card. Interviews were conducted between July 2021 and September 2021.

This study was approved by the Mayo Clinic institutional review board (21-00452).

We developed a semistructured interview guide that was pretested with volunteers (nurses, providers, and study staff) for duration, flow, and content. The interview guide was developed by the authors (PS, PS, TAB, CAP) and sought to elicit the following domains based on our study objectives: (1) COVID-19 experience that shaped the decision to engage in f2f appointments, (2) perceived benefits of f2f appointment and barriers to video visits, and (3) recommendations to increase future use of video appointments (see Multimedia Appendix 1). Interviews lasted approximately 45 to 60 minutes and were conducted in English by 3 study team members (NMR, IWW, and LMW). The interviews were highly flexible with probes for elaboration and clarification to obtain detailed accounts. This type of interview is advantageous when limited knowledge exists about the phenomena of interest [10,11].

QSR NVivo software, version 10 [16], helped facilitate response-theme generation with codes and categories based on themes emerging from the interviews. Two study team members (PS2 and ARS) coded responses together, discussing any coding discrepancies until reaching consensus and consulting a third study member (PS1) when necessary [17]. We extracted themes for analysis with code endorsement or elaboration in several interviews. In addition to open coding, we conducted planned comparisons by sex, age, and clinic site.

The final study sample was composed of 36 participants who attended f2f appointments during the April 2020 to December 2020 period of the pandemic. The median age of participants was 60 (SD 14.31) years. Of the 36 participants, 12 resided in the Midwest, 12 resided in FL, and 12 resided in AZ. Our sample of participants was diverse in gender (16 women and 20 men), race (21 White, 11 Black, 2 biracial, and 2 Asian), and education (15 with a bachelor degree, 13 with a graduate degree, 3 with some college degrees, 2 with associate degrees, and 1 with some school). All the participants had English as their preferred language. Most participants (32/36, 89%) were urban residents, and the estimated median household income was US $75,417 (IQR $61,730-$91,045).

We found that nearly all participants adhered to some degree of COVID-19 safety precautions. For example, 1 participant stated the following:

Only 2 participants expressed skepticism about the pandemic: one concerning the severity of COVID-19 infections and the other doubting the information related to COVID-19 through authorities and news media. For example, a White, male participant with some college credit mentioned:

Nearly all participants expressed that they felt safe from COVID-19 infection when attending f2f appointments. However, a few participants stated they would prefer a video appointment if infection rates were high in their area.

All participants conveyed a general trust in the Mayo Clinic, implying that the institutional guidelines limiting the number of persons in the patient waiting rooms and the patient screening of COVID-19 symptoms entering the hospital made them feel safe attending f2f appointments. Participants emphasized that personally adhering to masking and social distancing guidelines also made them feel safe attending in-person appointments at the Mayo Clinic. For example, a White, male participant with a bachelor’s degree stated:

We identified 9 major themes within this domain that are reported subsequently: (1) f2f rapport, (2) f2f diagnostic and therapeutic advantages, (3) habit, (4) privacy and internet security, (5) digital literacy, (6) internet access, (7) bodily intimacy, (8) billing, and (9) patient portal. The corresponding quotes related to themes can be found in Table 1.

Few participants had suggestions to offer. However, the recurrent theme among those who did was that video appointments were inappropriate for the first appointment with a new provider but acceptable for follow-up visits once rapport had already been established. A few participants noted the need for an institutional digital navigator who could immediately provide assistance if they are not able to connect to their provider virtually.

Our qualitative study provides information on personal and environmental factors that affected patients’ choice to not use video appointments during the COVID-19 pandemic within a large multistate institution for nonemergent and noninterventional outpatient clinical care. This preliminary study qualitatively explores patients’ telemedicine perceptions including barriers. Our findings will help inform the development of a large-scale survey that will guide future targeted interventions to increase the utilization and ease of use of telemedicine services.

The accumulated evidence from the early period of the COVID-19 pandemic (when this study was designed) showed that people had several misconceptions about the spread of COVID-19 [18] as a result of misinformation spread through social media platforms and other outlets. For example, some people believed COVID-19 was just a severe common cold without significant mortality [19]. In addition, before the development of vaccines, erroneous claims of achieving herd immunity by exposure to the virus were made [20]. Through the questions within domain 1 of our study, participants were prompted to talk freely about their COVID-19–related beliefs and perceptions and if those beliefs affected their decision to participate in f2f appointments. Our study is the first to explore such an association qualitatively. The results of our study indicate that the COVID-19 misconceptions may not have played a role in participants’ decisions to engage in f2f appointments. All participants in our study followed institutional and federal measures to mitigate the spread of COVID-19 during f2f clinical encounters, which probably motivated them to continue with f2f clinic care.

Within the second domain, we explored barriers associated with video appointments and factors that led participants to continue engaging in f2f appointments during the pandemic. We were interested if historical barriers to video appointments continued to exist during the pandemic. Like studies before the pandemic [5], our results indicate that participants who reported digital access and digital literacy as barriers to engaging in telemedicine were mainly older adults. Studies conducted before the pandemic highlighted that older adults struggled more with a poorly designed end user interface such as small text and computer screens and other digital skill tasks (scrolling down a menu tab, familiarity with a patient portal, or accessing a video conference link via the patient portal), which was supported by the participants in our study. These findings add to the existing body of literature and demonstrate that digital barriers (broadband access, digital literacy, and poor end user interface) continued to exist during the pandemic. Therefore, people with interrupted and limited digital access, such as those living in rural areas and those with limited digital literacy, such as older adults, may have no choice but to rely on traditional health care delivery methods such as f2f appointments. However, prepandemic studies have shown that older adults have demonstrated willingness to engage in technology and participate in telemedicine programs if the noted barriers are addressed [21,22]. Therefore, we propose a clinical practice change of assessing the digital competency of every patient so that digital solutions can then be applied based on the identified problem area. For example, people with no broadband access could be referred to a nearby free public hotspot or Wi-Fi access while also highlighting strategies to reduce privacy and data security concerns. We believe that those with limited digital literacy could be connected with institution digital navigator or support [23].

Lack of rapport has been highlighted in prior studies as a barrier to engaging in telemedicine care across different demographic subgroups [24]. Through the qualitative nature of our research, we present new findings by exploring this barrier in-depth with respect to patient sensory inputs and perceived therapeutic satisfaction with telemedicine care. This granular categorization reflects the diversity in which participants perceive lessened feelings of personability or connection with the clinician in video appointments. The most common subdivision of the “lack of rapport” category was the visual emphasis, followed by personability, then therapeutic, and finally auditory. We anticipate that these subcategories may not work in isolation and often interact with one another to define patient experience and perception. For example, patients who have trouble visualizing or hearing their clinician may report poor personal connection (personability). Future studies using quantitative approaches could explore how these subdivisions interact with one another.

More importantly, the unique nature of each subcategory, as highlighted in our results, has practical implications and suggests that attempts to alleviate the “lack of rapport” barrier cannot be one-size-fits-all. Accordingly, we recommend that digital access solutions be investigated to ameliorate rapport-related concerns. For example, digital navigators could provide and help patients use loaner devices with higher resolution screens, superior audio quality, or better internet connectivity to reduce latency and delays during video visits. Yang and colleagues [25] studied the implementation of loaner smartphone devices to patients who did not have smartphones; however, only 72% returned the loaned smartphone within a 30-day window. Future research is necessary to assess the feasibility of this approach for health care delivery. Together, these efforts could help reduce the awkwardness or degree of disruption to the flow of a conversation between patients and providers and increase the degree to which patients feel they can generate social rapport with their providers over video appointment.

Our results indicate that another common barrier to engaging in video appointments was perceived f2f therapeutic advantages. Participants believed their clinician would render better care in person (f2f) than over video. Participants pointed out that video appointments cannot replicate their perceived benefits of physical exams during f2f appointments. Assessing “lack of rapport” and “f2f therapeutic advantages” themes together, it can be inferred that patients perceive a video appointment as subpar (ie, of lower value either monetarily or therapeutically or diagnostically than an f2f interaction). Without providing exact reasons, several participants believed that psychiatry was a specialty well suited to video appointments. Participants could be inferring that psychiatry, unlike many other specialties, does not use physical exams to the same extent and thus would translate more effectively into a virtual format. The participants, however, did not elaborate if they would be comfortable discussing sensitive and personal topics in psychiatry via video.

On the contrary, participants were hesitant to engage in video appointments when they had dermatological, gynecological, or urological concerns. Several participants communicated that they needed to be examined in person and struggled with the possibility of using uploaded photographs of problem areas or exposing such sites to their webcams for providers to view. As highlighted by a few participants, this barrier could be due to concerns about privacy and security as well as awkwardness and discomfort relating to exposing more intimate body parts over video or uploading a photograph of the intimate body area. Although it seems probable that awkwardness and discomfort are the reasons driving the expression of this barrier, it is also possible the security or privacy barrier is at least compounding the concern if not occasionally replacing it. Moreover, concerns about security are potentially compounded by patient discomfort with the possibility, no matter how remote, of others gaining access to images of their bodies. Many patients expressed data privacy and security concerns when engaging in video visits, which has been highlighted as a barrier in the prepandemic literature [26]. Health care institutions need to vet their technological collaboration to assure patients by demonstrating the best practices regarding patient information privacy, data transfer, and storage [27]. A patient concerned about internet security for telemedicine from their end (user Wi-Fi or internet safety) could be connected to the digital navigator within the institution to assist patients.

As this was a qualitative study, our findings cannot be generalized beyond our purposeful sample. However, we opted for a purposeful selection to ensure that we interviewed a diverse group of patients who would express a wide range of views. The study sample was from Mayo Clinic patients; therefore, the participant’s perception of a “safe environment” may not be transferred to other institutions. Additionally, we asked patients to recall a health care experience that had occurred almost a year prior. Hence, it is possible that their recollections and details about events could be biased or incorrect. To help ease such a concern, we did verify eligibility and the existence of an f2f appointment via the medical record. Even though we tried to enroll the participants from diverse backgrounds, most of the participants in our study had bachelor degrees and higher education, which adds to education bias.

Our study’s primary strength was the use of qualitative means to gain a thorough, rich, in-depth understanding of participants’ perceptions relating to barriers to health care and telemedicine during the COVID-19 pandemic. We underscore that such qualitative exploration is a necessary precursor to further investigation by more rigorous, quantitative means. As a next step, we plan to develop a quantitative survey using a larger and more representative sample to determine the extent to which our findings can be generalized

Our study provides an in-depth investigation into barriers to engaging in video appointments for nonemergent clinical care. Limited f2f rapport, poor digital access and literacy, and concerns about privacy and security continued to be significant factors for patients not engaging in video appointments during the pandemic. Most importantly, this study highlighted that rapport-related concerns need to be looked at and problem-solved based on individual needs. Considerable clinical practice changes are required in the future at the institutional and policy level to encourage patients to engage in virtual care.