Factors Informing High-Risk Primary Care Patient Choice around Telehealth Use: a Framework

Abstract

Background

While telehealth’s presence in post-pandemic primary care appears assured, its exact role remains unknown. Value-based care’s expansion has heightened interest in telehealth’s potential to improve uptake of preventive and chronic disease care, especially among high-risk primary care populations. Despite this, the pandemic underscored patients’ diverse preferences around using telehealth. Understanding the factors underlying this population’s preferences can inform future telehealth strategies.

Objective

To describe the factors informing high-risk primary care patient choice of whether to pursue primary care via telehealth, in-office or to defer care altogether.

Design

Qualitative, cross-sectional study utilizing semi-structured telephone interviews of a convenience sample of 29 primary care patients between July 13 and September 30, 2020.

Participants

Primary care patients at high risk of poor health outcomes and/or acute care utilization who were offered a follow-up primary care visit via audiovisual, audio-only or in-office modalities.

Approach

Responses were analyzed via grounded theory, using a constant comparison method to refine emerging categories, distinguish codes, and synthesize evolving themes.

Key Results

Of the 29 participants, 16 (55.2%) were female and 19 (65.5%) were Black; the mean age (SD) was 64.6 (11.1). Participants identified four themes influencing their choice of visit type: perceived utility (encapsulating clinical and non-clinical utility), underlying costs (in terms of time, money, effort, and safety), modifiers (e.g., participants’ clinical situation, choice availability, decision phenotype), and drivers (inclusive of their background experiences and digital environment). The relationship of these themes is depicted in a novel framework of patient choice around telehealth use.

Conclusions

While visit utility and cost considerations are foundational to participants’ decisions around whether to pursue care via telehealth, underappreciated modifiers and drivers often magnify or mitigate these considerations. Policymakers, payers, and health systems can leverage these factors to anticipate and enhance equitable high-value telehealth use in primary care settings among high-risk individuals.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11606-023-08491-y.

INTRODUCTION

As American health care establishes its new normal in the aftermath of the COVID-19 pandemic, many patients and providers have been eager to preserve the enhanced access and flexibility offered by telehealth’s widespread uptake^1–5. Despite growing evidence of telehealth’s effectiveness in multiple clinical scenarios⁶, its net impact on utilization, cost, and equitable outcomes remains incompletely understood^7–9. Evidence that telehealth substitutes for in-person care and reduces access disparities in some situations—while increasing total utilization and exacerbating inequities in others—underscores the importance of understanding the dynamics underlying telehealth use^{6, 10, 11}. This is especially true of telehealth’s impact among elderly, chronically ill, and socially disadvantaged patients, who experience higher rates of avoidable acute care utilization¹². The potential cost savings associated with averting such care has attracted substantial private investments in recent years¹³.

The COVID-19 pandemic underscored the significance of patients’ role as consumers of health care. While prior studies have described population-level preferences¹⁴, the dynamics underlying individual decisions to pursue care via telehealth, in-person or not at all, have not been deeply explored. As public policymakers, payers, and at-risk providers weigh decisions around sustaining tele-access for clinically and socially at-risk populations^15–18, the ultimate impact of these policies may hinge on how well they anticipate and affect how patients pursue care via telehealth¹⁹.

To that end, we leveraged a unique period during the summer of 2020 to study factors informing high-risk primary care patients’ decisions regarding whether to pursue care via telehealth, in-person, or at all. Following a March 2020 transition from exclusively in-office to primarily telehealth-based primary care services, by June 15, 2020, the University of Pennsylvania Health System (UPHS) began allowing primary care patients to schedule care in-office or via telehealth for what was effectively the first time. Understanding the dynamics underlying patients’ subsequent choices stands to inform policymakers, payers, providers, and organizations seeking to scale equitable high-value telehealth use, and avoid perpetuating low-value or inequitable care.

METHODS

Study Design

We used a qualitative cross-sectional research design and grounded theory methodology²⁰ to characterize patients’ considerations when choosing between telehealth (audiovisual or audio-only), in-office care, or no care. Semi-structured telephone interviews were conducted July 13, 2020, through September 30, 2020. We followed the Standards for Reporting Qualitative Research Guidelines²¹. The study protocol was deemed exempt by the University of Pennsylvania IRB.

Setting and Participants

We used purposive convenience sampling to recruit adult clinically complex primary care patients at high risk of poor health outcomes from four urban and suburban UPHS internal medicine faculty medical practices. Patients were eligible if they (1) were English speaking; (2) were an established UPHS patient (e.g., at least one visit in the last 3 years); (3) had not had a primary care clinic visit in the last 6 months; (4) had been hospitalized in the last 12 months or were deemed “high risk” for poor health and utilization outcomes by a proprietary electronic health record risk tool internally validated by Epic Systems based on sociodemographic, medical, and health care utilization data²²; and (5) had received a phone call and/or patient portal invitation in the prior month to schedule a primary care appointment via the modality of their choice (e.g., audiovisual, audio-only, or in-office). Patients were eligible regardless of their choice of visit type, whether they completed the visit, declined the invitation, or did not respond to the scheduling invitation. Patients who had changed primary care sites or were admitted to inpatient or nursing facilities at the time of their invitation were deemed ineligible.

Data Collection

Patients were recruited via phone, at which point eligible participants provided verbal consent and completed semi-structured telephone interviews lasting 15–45 min. Interviews consisted of an orientation to the study and receiving consent, followed by structured questions and administration of a questionnaire. Interviews were completed by three team members (JM, MA, MS) who received training in qualitative interview methods from a team member with extensive qualitative research and interview experience (JS). The interview guide was developed using relevant conceptual frameworks^23–25 to elicit participant perceptions around telehealth, reasons for their choice of visit type, experience with telehealth, and thoughts about future telehealth use (see Appendix 1). The guide was iteratively adapted to explore emerging themes as part of the analytic process outlined below. Interviews were audio recorded, transcribed, and reviewed for accuracy. Recruitment continued until thematic saturation was reached when analysis of existing transcripts revealed no new themes²⁰. Participants received a $20 gift card to compensate their time.

Analysis

We used a grounded theory approach to describe factors influencing participants’ selection of visit type²⁰. After independently reviewing six initial interview transcripts, the research team (JM, MA, MS, JS) constructed a preliminary codebook from emerging themes and elements. Using NVivo 12.1, team members (JM, MS) proceeded to double-code twelve of the twenty nine transcripts, holding regular team meetings (JM, MS, MA) after each to review and reach consensus around code definitions and examples, evolving categories, and to synthesize themes²⁶. Team members memoed regularly to explore and document impressions, test relationships, ask questions, and diagram our evolving theory²⁷. As an overarching framework evolved, the final transcripts were coded individually, with the interview guide and sampling strategy tailored to identify participants and elicit perspectives necessary to inform theoretical gaps in the constructs being refined²⁰. Resulting themes were then linked together to fit a framework that captured the range of participant experiences²⁶.

RESULTS

We contacted 75 patients, of whom 8 were ineligible, 38 declined to participate, and 29 participated in the study, resulting in a 43% response rate (29/67). Participants were predominantly Black (n = 19; 65.5%) with an annual income less than $50,000 (n = 18; 85.7%) (Table 1). Compared to those who participated, those who declined or failed to respond were younger (59.6 vs 64.6 years old; p = 0.12) and more likely to be male (57.9% vs 44.8%; p = 0.29).

Table 1.

Demographic, Access, and Experience Characteristics of High-Risk Primary Care Participants

CHARACTERISTICS (n = 29)	n	(%)
DEMOGRAPHICS
Age	64.6	(11.1 SD)
Race
White	11	(37.9%)
Black	19	(65.5%)
American Indian or Alaskan Native	2	(6.9%)
Other	1	(3.4%)
Gender
Female	16	(55.2%)
Male	13	(44.8%)
Education
Did not complete grade 12 or GEDa	1	(3.5%)
Completed grade 12 or GEDa	13	(44.8%)
Completed at least some college	15	(51.7%)
Employment status
Employed	3	(10.3%)
Retired	17	(58.6%)
Other	9	(31.0%)
Household incomeb
< $50,000	18	(85.7%)
$50,000–$100,000	1	(4.8%)
> $100,000	2	(9.5%)
1 + Hospitalization in the last 12 months	19	(65.5%)
DIGITAL ACCESS AND EXPERIENCE
Has home broadband internet accessc	19	(67.9%)
Has access to a computer at homec	16	(57.1%)
Has a video-capable phoned	18	(72.0%)
Has completed any type of video call in the last 3 monthsc	17	(60.7%)
Has completed a telehealth video visit with a provider any time in the pastc	12	(42.9%)
Choice of visit type in response to clinic scheduling invitation:
No visit	5	(17.2%)
Video visit	5	(17.2%)
Audio/telephone visit	12	(41.4%)
In-office visit	7	(24.1%)
Uses transportation assistance for in-office visitsc	13	(46.4%)
Reports being extremely worried about COVID exposure from an in-office visitc	7	(25.0%)

^aGED refers to Tests of General Educational Development

^bThe number of responses available for select questionnaire items was limited to 21

^cThe number of responses available for select questionnaire items was limited to 28

^dThe number of responses available for select questionnaire items was limited to 25

We identified four major themes underlying patient preference for telehealth versus in-office visit types: perceived utility, costs, modifiers, and drivers. We describe each, and how these relate to each other in a framework of patient choice of visit type (Fig. 1). Representative examples of each theme, as well as underlying codes and categories, are included in Table 2.

Figure. 1.

Framework of factors influencing high-risk patient choice around telehealth use. See Table 2, “Themes, Categories, Codes and Illustrative Quotes from High-Risk Primary Care Participants” for an overview of the Themes and Categories included above, details of underlying Codes, and representative examples of each element.

Table 2.

Themes, Categories, Codes and Illustrative Quotes from High-Risk Primary Care Participants

THEME	CATEGORY	CODE	REPRESENTATIVE QUOTE
UTILITY	Clinical Utility	Diagnostic Efficacy	“But with technology, I don’t think it’s gonna show you everything you need or maybe you gonna lose something that could be vital to the situation.” (AP78)
		Focus	“And when you’re talking to somebody, you forget things or you overlook things. When you’re there personally, you have a chance to review the things… You remember things that you don’t normally remember, you could bring it up.” (FP87)
		Communication	“It’s just a certain way of communicating with an individual. I’m able to sense them, just like they are able to sense me up and see just what the doctor is– where he or she is coming from, to evaluate their bedside manner.” (AP78)
	Non-Clinical Utility	Altruism	“I would be more than happy to learn that which I need to learn to do my part in alleviating, you know, because I am sure it’s a very difficult time for all of us, including the doctors. So as long as that’s not algebra, I can’t see why I couldn’t [use telemedicine].” (MP013)
		Social Connection	“I get to see lots of people. Lots of different people. And like I said everyone’s pleasant. I get to even eat something that I enjoy eating there.” (JP27)
COSTS	Time		“It’s great. I don’t have to figure out a ride to get down to the doctor’s office…So, it’s easy. I can just sit in my living room, take the call, and go take a nap afterwards.” (MN3375)
	Money		“On technology, when I think about it, it should be cheaper because they’re not really doing that much.” (AP78)
	Effort		“Once the pandemic is over, I would want to go into the office. There’s no point of the video, because you’ll still have to get blood work, so it’s kind of like a double thing you have, like a double thing, might as well see them if they’re going to take the blood from me anyway.” (JP821)
	Safety	Loss of Privacy	“It [video visit] doesn’t bother me, I pretty much, I kind of like it because you get that one-on-one setting and you know, there is nobody there looking over your shoulders. It’s really helpful because you can be in the privacy of your own home.” (JP57)
		Loss of Health	“The positive in this particular case is that I didn’t have to go into an environment where the COVID virus might be floating about thereby my risk of contracting it was reduced by virtue of this telemedicine, teledoc or telemedicine visit. That was a good thing.” (DE313)
		Emotional Distress	“And I’m sitting there, I had my computer open, I’m looking for a phone number: why can’t I see this? why isn’t it happening? … It’s nerve wracking if you never did it.” (DP83)
MODIFIERS	Clinical Situation	Acuity	“It all really depends on how I feel, you know what I mean? If I felt like it was something that would need to be, an emergency thing, then I would want [my PCP] to see it [in-person], or whatever.” (JP57)
		Uncertainty	I was concerned about a couple of things that had been ongoing and I think he needed not only to look at them but investigate, which is something you can’t do through a video. (JT47)
		Exam Relevance	Well, if I’m going through an issue that requires her to examine me, you can’t do that over the phone. (JP66)
	Choice Availability	Authorization	“And [PCP] gave me that option. She had told me, ‘If you feel something going on or if something is going on with you,’ and if I had that, like she told me, it could be over the phone or I could come. So she gave me that choice.” (JP57)
		Awareness	“You know, I think there– I think I was invited to do a televisit and it never occurred to me to ask for it otherwise.” (MV27)
		Access	“I don’t have that type of phone that I can do a video on, and I don’t have a computer. So that’s all I could do was a phone chat.” (FP81)
	Decision Phenotype	Assertiveness	“It’s not a big deal for me, if that’s what’s needed I’m pretty easy to go. I’m – whatever I need, that’s what I’ll get. If [the health system] is offering it, I’m taking it… I’m real easy, I’m telling you. Whatever is good for them, and if I don’t need anything I’m fine with it. I have good doctors so, if I’m having a problem they take care of it immediately.” (JP27)
		Biases	“First of all, I’m an adventurous person and this is something new and I always like to try new things. When this was presented to me as an option, I thought, sure. Why not? Let’s try that, see how it works.” (DE313)
DRIVERS	Background Experiences		“You know since I’ve been in and I do have the rapport with him [PCP], I would feel comfortable [having an audiovisual visit] if something really was worrying me.” (AB413)
	Digital Environment	Tech Familiarity	“It’s hard for me to do the texting, and as far as the computer itself, like going online, I am really computer illiterate.” (FP81)
		Access Logistics	“I don’t deal with the phones and stuff like that…. I’d get somebody to teach me how to get the person on the phone or something…I would have to hook up a family friend to get me to do that.” (JP47754)
		Use Logistics	“I could see the video of the doctor, but the [video visit platform]…wouldn’t come up and I got tired of trying.” (JP821)

See Fig. 1, “Framework of the Factors Influencing High-Risk Patients’ Choice of Telehealth Use” for a graphic representation of the Themes and Categories outlined above

Utility

The most immediately evident determinant of patients’ preference of visit type was the relative utility they associated with each option. This included perceptions of their relative clinical and non-clinical utility.

Clinical utility represented participants’ beliefs about how likely a particular visit type was to identify and address their medical needs. It was the primary consideration for most participants. Most perceived in-office visits as more effective than telehealth visits, particularly with respect to accurate and comprehensive diagnoses. Telehealth care was often associated with inferior non-verbal communication, a loss of visual cues prompting patient recall of symptoms or needs, and an inability to pose “I almost forgot…” comments or questions once telehealth calls ended. Despite these general sentiments, several participants associated the absence of office-based distractions as improving provider attention during telehealth visits.

Non-clinical utility also factored into participants’ choice of visit type, but were often seen as secondary to perceived differences in clinical utility. Some expressed a preference for telehealth because they felt it contributed to provider convenience or the public good (e.g., decreasing the spread of COVID-19). Others preferred office visits given the personal and social interactions they offered.

Costs

The second major factor influencing patient choice of visit type was the relative cost associated of available options. These included costs in the form of time, money, effort, and safety.

Participants often described telehealth as requiring less time to complete than in-office visits, primarily due to reduced travel and transportation coordination requirements. However, telehealth’s net time savings were often reduced by the need for complementary lab work or return visits for in-person examination. Even when telehealth was associated with net time savings, it was often seen as of secondary importance relative to the perceived clinical utility of visit options.

Perceptions of lower resource requirements, lesser clinical utility, and experience with free digitally based resources meant that even equivalent out-of-pocket monetary costs for telehealth, relative to in-office care, was sensed as excessively and disproportionately costly to several patients.

While some participants reported that telehealth required less physical effort than in-office care, many felt it required as much or more overall effort than in-office care. For example, an inability to “cluster” related tasks (e.g., receive assistance coordinating referrals, authorize the transfer of medical records, or address billing inquiries) after their interactions with the primary provider left some participants feeling that telehealth was a less efficient way of addressing everything customarily completed during an in-office visit.

Several participants were aware of visit type’s impact on their safety. Infectious exposure risk during an in-office visit was the most commonly cited concern. Few participants expressed concerns that telehealth would contribute to a loss of privacy. Conversely, some participants associated telehealth with greater privacy than in-office care (Table 2).

Modifiers

The importance of utilities and costs when selecting visit types varied based on participants’ clinical situation, choice availability, and decision phenotype.

The acuity, diagnostic uncertainty, and exam-dependence of participants’ unique clinical situation altered utility and cost priorities. In high-acuity situations, participants preferred in-office care, consistent with their heightened sensitivity to optimal clinical outcomes and perception that in-office care was diagnostically superior and more efficient when fulfilling anticipated lab or care coordination needs. The same generally held true when participants were unsure of the cause of their symptoms, even when providers may not have deemed said symptoms as particularly severe. The opposite was also true: despite the apparent limitations, many participants were comfortable with “physicals” done via telehealth. Lastly, even in low-acuity, low-uncertainty situations, participants were likely to prefer in-office care when needs were seen as especially exam-relevant (e.g., musculoskeletal complaints).

Choice availability—awareness of visit type options along with differential access to technology—had an outsized impact on participant decisions. While all participants had permission to schedule in-office, audiovisual telehealth or audio-only telehealth appointments, awareness of their options and access to enabling technologies varied. Even when choice technically existed, limited awareness or ability to take advantage of options effectively silenced utility or cost preferences.

Lastly, decision phenotypes reflected the assertiveness and biases that impacted how patients acted on their utility or cost preferences. For example, some participants exhibited an assertive consumer-like insistence around their choice of visit type, based on their utility or cost preferences. Others tended to defer to their provider teams’ recommendations, regardless of the utility or costs they perceived. Separately, many participants adopted common heuristics—particularly familiarity bias—leading them to favor in-office care in the absence of compelling reasons otherwise.

Drivers

Participants’ background experiences informed their utility and cost perceptions as well as the modifiers they experienced. The impact of background experiences on other constructs was often mediated via trust. For example, positive individual- and community-level experiences with health care and health care providers were associated with more deferential decision phenotypes—meaning more likely to accept a recommendation from the provider. Additionally, past clinical experiences, and how patients viewed themselves medically, directly impacted perceptions. For example, patients who identified as medically complex often emphasized the diagnostic uncertainty of their clinical situations, placing a premium on the perceived clinical utility of visit types when deciding where and how to receive care.

Participants’ digital environment also influenced the determinants underlying their preferences. For example, socioeconomic and demographic barriers to digital fluency and access to devices or broadband restricted choice availability. Similarly, participant’s past experience with telehealth technology (e.g., inability to download apps, difficulty hearing providers, inconsistent video streams) shaped the costs (e.g., time and effort) they associated with telehealth.

DISCUSSION

Policymakers, payers, practices, and providers are actively attempting to understand and direct patient use of telehealth from primary care settings²⁸. We describe factors underpinning high-risk, medically complex patients’ decisions around pursuing primary care via telehealth, in-office, or not at all. Our study offers three primary findings that can inform the delivery of patient-centered, equitable, high-value telehealth services.

Enhancing High-Value Telehealth Use

First, while preferences varied by individual and situation, participants’ choice of visit modality was strongly influenced by perceptions of utility and cost. While participants unsurprisingly associated in-office care with superior utility, telehealth’s anticipated time and effort cost savings was often offset by its inability to allow patients to address complementary tasks at the same time and place (e.g., lab testing, follow-up coordination)²⁹.

These findings are consistent with recent trends and reports of telehealth use inside and without primary care³. Areas where telehealth’s use have been greatest include behavioral health and medical specialties (e.g., endocrinology) where utility is less dependent on in-person examination, and/or the need for complementary tasks (e.g., lab testing) is lower².

Our framework provides a roadmap for payers and practices seeking to improve uptake of high-value primary care (e.g., preventive services, chronic care management) by shedding light into high-risk member decisions around seeking care^{15, 30, 31}. While many practices have focused on IT assistance to facilitate telehealth use among digital newcomers, we identify additional ways to reduce barriers to high-value telehealth use (see Table 3). First, practices can minimize telehealth’s task clustering inefficiencies by arranging direct televisit “hand-offs” with scheduling, financial, and clinical support staff (e.g., social workers, patient navigators) following provider calls. Second, practices can extend efficiencies of in-office care to patients at home via at-home phlebotomy and direct staff phone lines to circumvent cumbersome phone-trees³². Third, payers can minimize cost sharing for high-value telehealth services given individuals’ sensitivity to financial costs for digital care and natural checks created by non-monetary utility/cost telehealth considerations³³.

Table 3.

Policy Implications

Maximizing the value of telehealth services

*Replicate in-office task-clustering via televisit “hand-offs” to complementary staff (e.g., schedulers, social work) immediately following patients’ telehealth visit with the primary provider

*Incorporate traditional efficiencies of in-office care into telehealth services (e.g., home phlebotomy, direct clinician phone line)

*Minimize cost-sharing for telehealth services, particularly around preventive care and chronic care management uses (e.g., insulin or hypertensive titration as part of diabetes or hypertension management)

Promoting high-value and minimizing low-value telehealth uses

*Use caution around relying on telehealth visits when patients: (a) perceive a need for physical examination, (b) believe their condition is not of low acuity, or (c) feel unsure about what could be causing their symptoms

*Leverage nudges (e.g., defaults), framing techniques (e.g., highlighting advantages of visit types), and social proofs (e.g., reinforcing normal uses of visit types) to align patients’ choice with the type of visit most likely to address their needs

*Consider alternating telehealth and in-office care to maximize potential telehealth conveniences while offsetting some of telehealth’s perceived limitations

Promoting equitable access and outcomes

*Continue to authorize home as an originating site for telehealth services

*Make audio-only access available for a wide range of medical and behavioral health services

*Educate staff, including physicians, nurses, and service representative, on how to support patients in making informed decisions around where and how they receive their care (e.g., shared decision making, bias recognition training)

Minimizing Low-Value Telehealth Use

Second, modifiers play an underappreciated role in modifying the utility and cost considerations underlying patients’ selection of visit type. For example, the clinical scenarios where patients tended to be amenable to telehealth shared common characteristics: (a) patients perceived no need for in-person examination, (b) patients deemed their conditions to be of low acuity, and (c) they anticipated little diagnostic uncertainty (e.g., post-hospital discharge and chronic care follow-up)⁴.

The impact of modifiers on patient choice aligns with decision-making and behavioral economic findings³⁴ and trends in health care utilization and preferences². These include high rates of in-person follow-up after telehealth visits with meaningful diagnostic uncertainty, even when these have involved low-acuity scenarios (e.g., management of upper respiratory infections)³⁵. Difficulty substituting telehealth for in-office care in these scenarios parallels certain challenges replacing emergency department services with urgent care³⁶.

These findings are valuable to entities not only interested in facilitating patient choice of telehealth services considered high-value (per above), but doing so while minimizing telehealth use that supplements without substituting for or enhancing care. Rather than imposing hard limits on which services are or are not eligible for telehealth, payers can collaborate with health systems and practices to leverage heuristics that align patient choice with the modality of care most likely to address their needs. These can include nudges (e.g., clinical scenario-based scheduling defaults), framing techniques (e.g., highlighting the conveniences of simultaneous lab-work), and social proofs (e.g., sharing with patients the ‘typical’ visit modalities used for certain scenarios) (see Table 3). Additionally, providers might consider alternating telehealth with in-office visits to leverage telehealth’s legitimate, if situational, cost advantages while offsetting its diagnostic and communication limitations. Assuring patients of opportunities to address office-dependent needs—such as physical examination—may help minimize duplicative in-office follow-up³⁷.

Maximizing Equity

Third, our framework highlights how demographic, socioeconomic status, and geography tend to alleviate or exacerbate telehealth’s impact on inequities in care through their impact on patients’. On the one hand, geography, frailty, age and poverty can fuel costs (e.g., transportation barriers, limited availability during business hours) and clinical situations (e.g., increased need for chronic disease management)³⁸ that telehealth is well positioned to address. At the same time, these same drivers can be associated with decreased choice availability (e.g., limited broadband) and behavioral phenotypes that limit telehealth use (e.g., socially disadvantaged populations’ loss of trust in healthcare institutions may make them less likely to be early adopters of new technologies or accept care in less resource-intensive settings)^{39, 40}. Beyond impacting equitable access, telehealth risks exacerbating inequity if these same factors make vulnerable populations more likely to be directed to telehealth when in-office care would have been objectively superior.

Additional reports have reinforced telehealth’s complex relationship with access and equity^{8, 41}. While some have highlighted disproportionately low uptake of audiovisual telehealth visits among disadvantaged populations (e.g., among federally qualified health center patients)⁴², others have associated age, Black race, and public insurance with increased total telehealth use relative to younger, White, and privately insured patients^{31, 43, 44}.

Our framework identifies three opportunities to ensure telehealth advances equity instead of exacerbating inequity. First, policymakers and payers can continue investments in digital access (e.g., broadband coverage of rural communities, payer provision of phones for members without cellular devices). While such steps are foundational, they are insufficient on their own. Second, to mediate disparities in choice availability, payers should authorize audio-only delivery of a wide range of medical and therapy services. Third, practices should train administrative and clinical staff in shared decision-making and recognition of implicit age, racial, and cultural biases to maximize choice availability and ensure visit type selections truly align with patient preferences⁴⁵. Without addressing these complementary factors, investments in digital access alone will fail in their attempt to enhance equitable access.

Limitations

Our study has several limitations. The experiences of English-speaking participants from urban and suburban internal medicine practices affiliated with a single urban academic medical center may not be generalizable to all patients and may vary across the subgroups studied (e.g., those opting not to pursue care versus those pursuing in-office care). Patient preferences and the relative importance of their considerations are likely to change over time, particularly as patients become more familiar with telehealth offerings. There may have been an element of self-selection in who agreed to participate in the study, compounded by the fact that interviews were completed via telephone. Last, although approaches were taken to enhance trustworthiness, the themes we identified are only potential interpretations of our data. Future research should not only validate the presence of these considerations—including in other patient populations—but confirm and quantify the relative importance of each to guide actionable decision-making.

CONCLUSION

In conclusion, multiple factors influence patients’ choice of whether to pursue care via telehealth, in-office or at all. These include the relative utility and cost associated with various modalities, as well as underlying modifiers and drivers. Our framework provides direction for researchers, policymakers, payers, and providers interested in exploring how to use telehealth to advance high-value, equitable health care for vulnerable populations.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

John Morgan’s position was funded by the Department of Veterans Affairs (VA) at the Philadelphia Corporal Michael J. Crescenz VA Medical Center, with additional funding for Dr. Morgan and Mandy Salmon provided by the John Eisenberg Scholar Research Award (University of Pennsylvania).

Data Availability

As study participants did not give written consent for their data to be shared publicly, research supporting data is not available.

Declarations:

Disclaimer:

The content of this article is solely the responsibility of the authors and does not reflect the views of the Department of Veterans Affairs, University of Pennsylvania, the Commonwealth of Virginia, or the authors’ current or former employers.

Conflict of Interest:

Nwamaka Eneanya holds Somatus stock. Judy Shea has received grant funding from the National Institute on Aging and National Institute of Mental Health (NIMH), and honoraria from the Association of American Medical Colleges (AAMC).