Human technology intermediation to reduce cognitive load: understanding healthcare staff members’ practices to facilitate telehealth access in a Federally Qualified Health Center patient population

Abstract

Objectives

The aim of this study was to investigate how healthcare staff intermediaries support Federally Qualified Health Center (FQHC) patients’ access to telehealth, how their approaches reflect cognitive load theory (CLT) and determine which approaches FQHC patients find helpful and whether their perceptions suggest cognitive load (CL) reduction.

Materials and Methods

Semistructured interviews with staff (n = 9) and patients (n = 22) at an FQHC in a Midwestern state. First-cycle coding of interview transcripts was performed inductively to identify helping processes and participants’ evaluations of them. Next, these inductive codes were mapped onto deductive codes from CLT.

Results

Staff intermediaries used 4 approaches to support access to, and usage of, video visits and patient portals for FQHC patients: (1) shielding patients from cognitive overload; (2) drawing from long-term memory; (3) supporting the development of schemas; and (4) reducing the extraneous load of negative emotions. These approaches could contribute to CL reduction and each was viewed as helpful to at least some patients. For patients, there were beneficial impacts on learning, emotions, and perceptions about the self and technology. Intermediation also resulted in successful visits despite challenges.

Discussion

Staff intermediaries made telehealth work for FQHC patients, and emotional support was crucial. Without prior training, staff discovered approaches that aligned with CLT and helped patients access technologies. Future healthcare intermediary interventions may benefit from the application of CLT in their design. Staff providing brief explanations about technical problems and solutions might help patients learn about technologies informally over time.

Conclusion

CLT can help with developing intermediary approaches for facilitating telehealth access.

Background and significance

Telehealth is now a critical form of healthcare delivery, with usage rising dramatically worldwide following the disruption of in-person care during the COVID-19 pandemic.^1,2 Telehealth involves healthcare delivery via video-conferencing, audio technologies, patient portals, and store and forward methods. Benefits of telehealth visits for patients include reduced travel burden and increased access to care in areas with provider shortages.^3,4

However, with the sustained use of telehealth,^5,6 research attention is needed to address related disparities in access, usage, and quality. Low-income patients, who are typically overrepresented at Federally Qualified Health Centers (FQHCs), face well-documented barriers to video visit and patient portal use, including technology access, and technical skills.^7–10 Relatedly, low-income patients have higher rates of telehealth visits via telephone than higher-income patients, who are more likely to have video visits.¹¹ Telephone visits are clinically insufficient for some clinical conditions.^12,13

One approach to addressing disparities in video visit and patient portal usage among FQHC and other low-income patients is to provide human assistance through trusted intermediaries. Human intermediaries are “…people who act in a ‘middle space’ between people and technologies, and perform the mediating work involved in enabling potential users to access, configure, understand, cope with, troubleshoot, and use patient-facing health technologies.”¹⁴ Human technology intermediaries can be informal (eg, family, friends, and neighbors) or formal (eg, nonprofit organizations, volunteers, or staff).¹⁵

As the number and complexity of patient-facing technologies have expanded, there has been growing interest in incorporating human technology intermediaries into health care. Researchers have observed that, for users with little technology experience, intermediaries can “…shield some of the [user interface] complexity from the beneficiary-user…”¹⁶ In health care, such intermediaries typically occupy such roles as “digital navigators.” People in these roles connect patients to patient portals and video visit technologies and provide support and education for using them.¹⁷ During the COVID-19 pandemic, some healthcare organizations redeployed patient navigators and community health workers into digital navigator-like roles.¹⁷

However, dedicated staff may not be consistently available in healthcare organizations, especially in lesser-resourced settings like FQHCs, where low-income people often seek care.¹⁸ When there is no dedicated digital navigator, other healthcare staff (eg, physicians, nurses, receptionists, and medical assistants) may be required to provide technical support. Theoretical and empirical insights into such intermediaries’ telehealth intermediation approaches can shape healthcare staffing models to formally incorporate intermediary work into staff roles or to create dedicated roles and interventions.

Theoretical framework: cognitive load theory

Completing video visits and navigating patient portals are complex tasks, especially for those with limited technology experience. Cognitive load theory (CLT), a theory of learning complex tasks,¹⁹ assists in understanding how intermediaries successfully provide technology-related assistance by reducing cognitive burden for FQHC patients (see Box 1 for definitions). CLT concerns information processing and storage within human memories. CLT recognizes a distinction between long-term memory and working memory (Figure 1). Long-term memory is where learning occurs, and knowledge is stored; it often involves organizing information into a schema or mental model that can be applied to future tasks.²⁰ Working memory is accessed when completing a task; it involves retrieving information from long-term memory and sorting through incoming information from auditory and visual stimuli. Unlike the unlimited storage capacity of long-term memory, working memory has short-term storage capacity limits. Cognitive load (CL) refers to the amount of working memory resources used to process information while performing a task. Cognitive overload occurs when working memory becomes overtaxed, resulting in unsuccessful task completion and stymied learning. CLT also posits that people have autonomous visual and auditory channels in working memory, each with its own information-processing capacity.

Box 1.

Cognitive Load Theory Definitions.

Concept	Definition
Cognitive load	The amount of working memory resources used to process information while performing a task. To facilitate learning and task completion, cognitive load should be kept to a manageable level.78
Cognitive overload	This occurs when learners who do not have previous knowledge to draw form, resort to resulting random, inefficient attempts that can overtax working memory—resulting in unsuccessful task completion and stymied learning.20
Long-term memory	The place where knowledge is stored, and learning occurs. It has infinite capacity and is accessed to make sense of how to approach a new task. The information is often organized into a schema or mental model that can be applied as a single unit in the working memory.20,79
Working memory	Also known as “short term memory,” it is the only memory that can be monitored.79 It has limited capacity (7 items) and duration (2-20 s).20,79
Sensory memory	The novel, random audio-visual information which is introduced from the external environment and processed through the sense organs. Because it is transient, information is only briefly held before being sent to the working memory for processing.20
Extraneous load	The cognitive load that emerges from phenomena not directly connected to completing a task, including distracting instructions, environmental distractions, stress, and negative emotions.25
Intrinsic load	The cognitive resources needed to complete the task; this is fixed by the nature of the task and increases with task complexity.21,22

Figure 1.

Cognitive load theory.

CLT, often used in instructional design, differentiates between intrinsic and extrinsic CL. Intrinsic load is the cognitive resources needed to complete the task; this is fixed by the nature of the task and increases with task complexity.^21,22Intrinsic load also increases when a learner has less prior knowledge of the task.²³ Therefore, CL may be relatively high for low-income people, including those who seek care at FQHCs, who have often had fewer opportunities to access technology than people with higher incomes.²⁴ By contrast, extraneous load occurs when information is introduced that is distracting and not necessary for completing the task.²⁵ Extraneous load is generated by factors associated with the organization of a task or related information²⁰ and factors outside of the task itself. Low-income people who seek care at FQHCs often experience past and present socioeconomic hardship that can tax working memory,^26–29 thus increasing baseline extraneous load. Furthermore, negative emotions such as worry can increase extraneous load both related and unrelated to the task at hand.³⁰ Intrinsic and extraneous CL are additive and together consume working memory resources.

While healthcare staff may not be trained in CLT, due to the universal nature of working memory limits, they may discover CL reduction techniques through trial and error when attempting to facilitate patient technology use.

Research aims and objectives

The aim of this study was to investigate how healthcare staff intermediaries support FQHC patients’ access to telehealth services, how their approaches reflect CLT and determine which support approaches FQHC patients find helpful and whether these perceptions suggest CL reduction.

Methods

Research setting

The FQHC, with 5 clinical sites in Metropolitan Detroit, provides care for over 20 000 patients who are majority female (61%), majority Black (38%), and Hispanic (28%), and who are primarily insured through Medicaid or Medicare (57%). The FQHC rapidly transitioned to offering telehealth services via phone and video due to the COVID pandemic in March 2020. FQHC staff used 2 stand-alone telehealth platforms for conducting video calls, and neither the video call nor the patient portal platform required patients to install software prior to use. The video visit platform was not integrated into the EHR system and patient portal, and this remains the case today. Our prior research indicates that low enrollment in a newly launched patient portal, technical issues, internet disruptions, limited data plans, and busy households were barriers to FQHC patients’ use of telehealth technologies.³¹

Data collection

We conducted in-depth, semistructured interviews with FQHC staff using a staff-focused interview guide (see Appendix SA). Interviews occurred from July to August 2020, and staff participants were asked to complete a follow-up demographic survey. Staff interviews averaged 56.6 min in length, and staff received a $25 incentive for participating. We recruited FQHC staff using a purposive sampling approach to represent different clinical settings and roles. These staff members were not formally trained technology intermediaries; it emerged as a necessary new job function during the transition to telehealth. Thus, our analysis identifies emergent helping behaviors used by healthcare staff.

We also interviewed FQHC patients in English or Spanish about their experiences with telehealth. To recruit, we sent a survey link via a text and/or patient portal “campaign” to all patients who were: (1) ≥18 years; (2) English or Spanish speakers; (3) consented to receive texts from the FQHC or had a patient portal; and (4) had completed a phone or video visit during 10 weeks from July to September 2020. Patients who completed the associated survey, which asked about their telehealth experiences, technology ownership, and demographics, indicated whether they were willing to be interviewed. Follow-up interviews with those who stated a willingness to be interviewed and responded to recruitment calls were conducted from March to May 2021. The patient semistructured interview guides (Appendix SA) focused on technology intermediary assistance from FQHC staff. Since the staff and patient interviews occurred sequentially, findings from the staff interviews shaped some of the questions asked of patients. For example, because the staff interview findings revealed that they were triaging patients based on technology ownership, we asked patients the following question as part of the interview guide: “Did they ask you about the technology you have at home? What did they ask? What did you tell them? Did they do anything in response?” Patient interviews were conducted via phone or Zoom calls and were an average of 53.2 min long and patients received a $25 incentive for participating. English interviews were transcribed verbatim and verified. For the Spanish-language interviews, audio files were translated into English and transcribed verbatim simultaneously by a professional translation service. These Spanish-to-English transcripts were verified separately by 2 bilingual research staff (A.K.W. and G.M. in acknowledgments), who then discussed and resolved any areas of uncertainty together. Data collection continued until after data were saturated.³² We assessed saturation based on the occurrence of codes in the interviews, finding that no new codes emerged after the third staff interview, likely reflecting a relatively homogenous set of staff experiences with providing support to the FQHC’s patient population. Similarly, we found that no new information supporting novel codes emerged after the eighth patient interview. However, we continued to conduct interviews after reaching saturation to ensure that we gathered data from patients with diverse demographics and staff in a wide range of patient-facing roles. Nevertheless, we discovered that the sample was relatively homogenous with respect to the tightly focused scope of behaviors and reactions studied, supporting the ultimate study sample sizes. Notably, the design also allowed for triangulation of findings across the staff and patient interviews. The University of Michigan Institutional Review Board reviewed the study and declared it exempt from ongoing IRB review based on a federal exemption for minimal-risk research with adults who prospectively agree to the research and where the identity of subjects cannot be easily ascertained.

Data analysis

Qualitative data analyses began early in the interview process so any necessary adjustments could be made to interview procedures. Three researchers (A.K.W., a doctoral student in information with qualitative methods training; T.C.V., a PhD-trained health informatics professor and mixed methods researcher; and T.R.D., a PhD-trained human-computer interaction professor and mixed methods researcher) independently read 3 randomly selected transcripts. They then collaboratively developed inductive codes using process coding³³ to identify intermediary actions and evaluation coding³³ to categorize participants’ judgments of the value of intermediaries’ actions. For instance, an inductive process code developed after this initial transcript review was “breaking information down into smaller steps,” a helping action which we defined as intermediaries providing information about telehealth to patients in bite-sized, action-oriented, sequential pieces. The first-cycle coding was led by the first author (A.K.W.), with weekly meetings with 2 other researchers (T.C.V.; M.G.A., a PhD-trained health informatics postdoctoral fellow and mixed methods researcher) to discuss and reach consensus on definitions and additions to the codebook, discuss how to apply codes consistently and review preliminary coding applications. Continuing the previous example, the first author used the “breaking information down into smaller steps” code and others to label chunks of text in all of the transcripts using NVivo qualitative data analysis software. Three researchers (A.K.W., M.G.A., and T.C.V.) developed deductive codes and definitions based on a literature review concerning CLT. Then, in second-cycle coding, the first and last authors (A.K.W. and T.C.V.) mapped inductively derived codes onto deductive codes and met weekly with another researcher (M.G.A.) to confirm the interpretation of the coding of CLT concepts. Keeping the same example, in this process, the 3 researchers reviewed the text coded as “breaking information down into smaller steps” and compared it to the aforementioned list of CLT-based deductive codes. In so doing, the researchers determined that this inductive code was similar to the “segmenting instructions” subcode of the broader “supporting the development of schemas” CLT code. Subsequently, “breaking information down into smaller steps” was made a subcode of “segmenting instructions.” Through our analytic discussions on this code, we identified “staff gave instructions that divided the process into smaller, sequential steps” as an intermediary approach for CL reduction.

During and after coding, we used analytic memoing to synthesize findings. Finally, we compared and contrasted the findings from patients and providers to synthesize findings.

Results

Characteristics of participants

Staff participants (n = 9) were all female, aged 21-52, and majority White and non-Hispanic individuals (88.9%) (Table 1). All had provided some intermediary assistance to patients for using video visits and/or patient portals. Most patients (n = 22) were non-Hispanic Black (45%) or Hispanic/Latino (35%). Sixteen (80%) were female, and 6 (20%) were male, and their ages were 23-66 years (Table 1). See Table 1 for participants’ education, employment, technology experience, and access.

Table 1.

Characteristics of study participants (n = 9 FQHC staff and n = 22 patients).^a

	FQHC staff (n = 9)		Patients (n = 22)
	Number	Percentage	Number	Percentage
Gender
Male	0	0	16	72.7
Female	9	100	6	27.3
Age
Mean/median	36.4/31		48.6/49.0
Range	21-52		23-66
Race
White (non-Hispanic) (non-Hispanic)	8	88.9	4	18.2
Black/African American	1	11.1	9	40.9
Hispanic/Latino	0	0	7	31.8
Missing	0	0	2	9.1
Education
Grades 9-12, no diploma	0	0	1	4.5
High school graduate/GED	0	0	4	18.2
Some college	3	33.3	6	27.3
Associate’s degree	0	0	1	4.5
Bachelor’s degree	3	33.3	5	22.7
Graduate degree	3	33.3	1	4.5
Professional degree	0	0	2	9.1
Missing	0	0	2	9.1
Employment
Full time	9	100.0	6	27.3
Part time	0	0	2	9.1
Unemployed	0	0	8	36.4
Disabled	0	0	2	9.1
Retired	0	0	2	9.1
Missing	0	0	2	9.1
Language of interview
English	9	100	18	81.8
Spanish	0	0	4	18.2
Years of internet experience
Mean/median	17.6/18.0		15.4/18.0
Range	10-25
Technology access—has any deviceb	9	100	14	63.6
Has desktop computer	7	77.8	7	31.8
Has laptop computer	9	100	13	59.1
Has smartphone/cellphone	9	100	19	86.4
Has tablet computer	5	55.6	6	27.3
Has e-book reader	3	33.3	4	18.2
Has video game controllerc	3	33.3	2	9.1
Missing	0	0	2	9.1
Frequency of internet use
Several times a day	9	100	14	63.6
About once a day	0	0	5	22.7
3-5 days a week	0	0	1	4.5
1-2 days a week	0	0	0	0
Every few weeks	0	0	0	0
Less often or never	0	0	0	0
Missing	0	0	2	9.1
Health insurance	Not askede
Medicaid	—		7	31.8
Medicare	—		4	18.2
Private	—		3	13.6
Uninsured	—		6	27.3
Missing	—		2	9.1
Staff role			Not askedd
Community health worker	1	11.1	—
Medical assistant	1	11.1	—
Medical receptionist	2	22.2	—
Nurse clinic manager	1	11.1	—
Nurse practitioner	2	22.2	—
Physician assistant	1	11.1	—
Registered nurse	1	11.1	—

^aData were collected via participant self-report from web-based surveys completed via the Qualtrics survey platform.

^bMultiple responses were possible.

^cWe asked about this since some video game consoles are connected to the Internet, and our past research with young people revealed that video game consoles can be a major mode of communication with others via social gaming applications. We also reasoned that video game console access could be indicative of experience with technology, or of having individuals with technology experience in the household.

^dThis was not asked since patients do not have staff roles.

^eThis was not asked since staff members’ personal health insurance was not pertinent to the study.

Abbreviation: FQHC, Federally Qualified Health Center.

Intermediary approaches

As we outline below, staff intermediaries used 4 overall approaches to supporting access to, and usage of telehealth services among patients, all of which mapped onto CLT. Patients experienced each of these approaches as helpful in different ways.

Shielding patients from cognitive overload

Cognitive overload can be prevented by helping people navigate through unnecessary information (ie, extraneous load) and signaling essential information³⁴ for task performance (ie, intrinsic load). To limit cognitive overload from video visit-related tasks, intermediaries performed technological triaging to shield patients from predictable incompatibility problems, applied black box fixes to carry the burden of resolving emergent technical problems themselves, and offered suggestions for reducing environmental distractions. Additionally, intermediaries signaled essential information for completing video visits by providing timely, succinct information.

Technological triaging. Video visits often were not possible due to incompatibility or language issues, or interruptions from technical problems like inconsistent connectivity, disrupted audio, or camera issues. Thus, FQHC staff, primarily medical assistants and receptionists, assessed the potential for patients to successfully conduct a video visit by asking patients about technology ownership. The staff recognized that patients with internet access often owned smartphones (rather than a desktop computer), so they would begin calls by asking patients if they had a smartphone (Table 2). This had the potential to prevent the cognitive overload associated with problem solving within the video visit, only to find that the video visit was not technically feasible. Furthermore, when staff had many Spanish-speaking patients, they were stymied by technological burdens of explaining English-only or partially translated interfaces and communications within telehealth and patient portal platforms. Consequently, these staff’s intermediary efforts involved overcoming untranslated platform communications. For video and telephone visits, this meant switching from the main platform (platform 1) used at the FQHC to one that functioned differently. These staff used a secondary platform because it rang a video call on patients’ smartphones, reducing steps, and bypassing English-only text notifications.

Table 2.

Quotes regarding intermediary approach 1.

Intermediary approach 1. Shielding patients from cognitive overloada
Technological triaging	Asking patients about technology ownership “They asked me that I have a newer cell phone and not a flip phone.” [P104] “It's mostly patients that don't have a good phone…some [patients], they don't have good phones for them. So we just do a phone call, because we can't be wasting time.” [S6] Overcoming untranslated platform communications “…a…barrier is having apps and information …in Spanish …to explain [Platform 1] in Spanish …interpret that and go through it, it would have taken forever just to get somebody set up on the video. So I did not use [Platform 1]… what I found that worked really well was [Platform 2].” [S1] “…[Platform 1] had too many English instructions and [Platform 2] was click and then super easy…We could tell them over the phone, “You're going to get the link. It'll be in English in your text message. Click on it and it'll pop up through a video.” And that was just so easy versus [Platform 1]… It wasn't accessible for our Spanish speaking population.” [S4] Some patients found technical issues unavoidable even if they did have a smartphone “I think it’s more so the signal in my house that’s bad. That’s why I’m on the porch now.” [P101] “If it didn’t have a battery you were in trouble.” [P107]
Applying black box fixes	Staff resolved emergent technical problems themselves “The thing I found that's persistent throughout the visit is the sound of it. Sometimes it'll go on and off, so I'll have to tell them to turn their video on and off and then rejoin the meeting or turn off their audio on, and then return it back on and I'll have to do that too…that's basically the only guidance throughout if we can't hear each other or see each other.” [S3] “We…go to a second [video visit] platform or when it's not working, I'm like, ‘Restart, restart, restart, try a different platform,’ and we end up just doing a phone call [if those fail].” [S4] Patients did not understand the technical problem or the solution “…[there] was a time where the link wasn't working…Not exactly sure [why the link worked later].” [P106] “I could hear [Nurse Practitioner] when she was [cutting out]. She was like, ‘Hold on, I need to go to another link.’ What that means, I don't know. But whatever she did made it so it worked.” [P101]
Reducing environmental distractions (cognitive load theory: weeding principle)	Staff suggested where to conduct the video visits “Well, I need you to leave the store. You can't be in Walmart and having an appointment with me.” [S2] “…to…be in a quiet place by themselves and not driving. Definitely not at a store anywhere…somewhere they can be listening and have their appointment.” [S7] Patients’ abilities to reduce home distractions was limited “[Said to doctor]‘Sorry that you see so many people in here and you hear people. It’s not a quiet place.’” [P106]
Providing timely, succinct information (cognitive load theory: weeding principle and signaling principle)	Highlighting essential information that patients needed to hold in their working memory “…they told me…that they would send me a link a few minutes before the call” [P121]. “[Staff told me to] Just click the page I send and that's all… it was very good.” [P118] “…when I scheduled they told me that the nurse was going to call me prior to my… About 20 to 30 minutes prior to my appointment and then give me instructions at that point.” [P105] “[My patients] can't read English…when they get the text message about the waiting room in [Platform 1], they don't…know that they're supposed to click on it… we usually have to call them and tell them, ‘You haven't come into the waiting room.’ And they're like, ‘Oh, we didn't know what that meant.’” [S5] Some patients found connecting to telehealth straightforward “They explained to me then that it will be either via the computer or the cell phone. They would send me a link and I would click the link and I would be connected to a doctor…Everything was fine. They explained everything I needed to know, so I was okay.” [P114] “…it was really easy. They send you the link, they let you know when it’s time to go in…they called beforehand, the medical assistant called…” [P103] “…they called me 30 minutes before, to assure me that I would indeed get the call, and that if I needed a translation or something, they would provide it…everything was very clear.” [P122] Some patients wanted more instructions “…the communication could be a little stronger…what mechanism we're using to conduct the telehealth would've been helpful to me… they just told me, “Oh, well the nurse is going to call you.” Okay, well the nurse is going to call me with what? The link? Are they going to tell me go to this link?…[it would have been helpful to have]… Like instructions I guess basically. Like the nurse is going to call you an hour before, she's going to email you the link, you log in to the link and maybe you can describe what that looks like. Is it going to ask you for your name? You have to put a little bit of information in, whatever that looks like, like walk you through it.” [P105]

^aParticipants identification codes beginning with the letter S refer to staff members, and those beginning with the letter P refer to patients.

While providers saw a benefit from this tactic, some patients found technical issues unavoidable even if they did have a smartphone. Many patients mentioned unstable access to high-quality cell phone signals and Wi-Fi connections. Patients also spoke about electricity blackouts, cable/internet outages, or battery limitations they experienced (Table 2).

Applying black box fixes. Staff resolved emergent technical problems themselves, often without explaining what they were doing to patients. Staff members’ tactics most commonly included turning video or audio devices off and on, logging out and back into telehealth programs, re-sending links, or even switching video visit platforms (Table 2). Patients considered black box fixes helpful as they allowed them to focus on the visit. However, when intermediaries resolved telehealth issues, patients did not understand the technical problem or solution. This approach shielded patients from expending their cognitive resources with problem solving, but it also meant that patients did not typically learn new skills for future use of these technologies.

Reducing environmental distractions. Patients were not always accustomed to the best environments in which to have a successful video visit. For instance, some patients attempted to have video visits in public places like stores. After a few such instances, staff suggested where to conduct the video visits to limit patients being distracted by their environment (Table 2). This allowed people to focus their attention on both using the technology and the subsequent clinical consultation. Accordingly, this method applied CLT’s weeding principle, or excluding extraneous information (see Box 2 for CL reduction principles).³⁵ However, advice about how to reduce distractions during the COVID-19 pandemic was less helpful since patients’ abilities to reduce home distractions was limited due to the presence of other family members.

Box 2.

Cognitive Load Reduction Principles: Definitions.

Principle	Definition
Signaling principle	Cues that provide direction to the essential information needed to complete a task.34,35
Weeding principle	Sorting through and removing extraneous information that may distract from completing a task.35
Expertise principle	Tailoring instruction to a person’s level of expertise by providing and appropriate balance of instruction and activation of their long-term memory.36
Segmentation principle	Providing time between segments of information to allow for better information processing.35
Pretraining principle	Providing an overall introduction to the terms and components and system behaviors, involved in the task prior to attempting that task.35
Off-loading principle	Moving information provision from the visual to the audio channels (or vice versa) to support better information processing.35

Providing timely, succinct information. Staff provided telehealth instructions that told patients what they needed to know when they needed to know it—often in a phone call just before the visit. In line with CLT’s, signaling principle,³⁵ staff highlighted essential information that patients needed to hold in their working memory to connect to the call successfully. Staff also implemented the weeding principle by excluding extraneous information beyond that which was immediately relevant.³⁵ Suggesting that this may have assisted with cognitive processing, some patients found connecting to telehealth straightforward after a previsit call in which they received succinct information about how to connect. Specifically, such patients described the telehealth process as “fine,” or “easy.” Despite this, some patients wanted more instructions before the previsit call (Table 2). Furthermore, sometimes Spanish-speaking patients only received this succinct information after not logging into the video visit waiting room when expected (Table 2). In these cases, intermediary actions involved a follow-up telephone call to explain the process in Spanish to overcome untranslated platform communications.

Drawing from long-term memory

Staff assessed and activated patients’ prior knowledge of digital technologies to help them use video visits and patient portals. Since prior knowledge would be stored in long-term memory, its use can reduce demands on working memory.³⁶ This approach enacts CLT’s expertise principle, which holds that instruction should be tailored to a person’s expertise level by providing an appropriate balance of instruction and activation of long-term memory.

Confirming prior knowledge. Some staff reported gauging patients' prior knowledge, experience, and comfort with telehealth video calls. This approach helped patients with prior knowledge feel prepared. Additionally, it assisted staff in identifying patients for whom video visits were not possible, as well as those with prior experience with them (Table 3). Stated in terms of the expertise principle (Box 2), this approach helped staff identify those for whom prior knowledge may be available to facilitate the visit.

Table 3.

Quotes regarding intermediary approaches 2 and 3

Intermediary approach 2. Drawing from long-term memory
Confirming prior knowledge (cognitive load theory: expertise principle)	Gauging patients' telehealth-related knowledge, experience, and comfort “…because I have some experience with Zoom and others…they asked something to know if you feel comfortable using it, and I was like, ‘Yeah, I've got the ability to do that.’” [P110] “‘Have you already done this?’ I said, ‘Yeah,’ that was it. So, no need for anybody to explain it to me.'” [P111] “[Front desk staff] They are pretty good about asking… [PARTICIPANT NAME] is probably going to be doing a video. Is that okay with you? And then they'll say yes or no. If they say no, then my appointment note when I call them will say phone visit.” [S3] “…“if they haven't [used] FaceTime or Zoom or anything, use any sort of other video app on their phone, it's more difficult for them.” [S3] Helped Patients with Prior Knowledge Feel Prepared “I told him that I had my phone and that I was allowed to use the microphone and the camera. So she was like, ‘Great.’ She was like, ‘Okay. Then when he sends you a link, right away, you're able to connect and you're going to be able to see him…’ I already knew what I was getting myself into…” [P106]
Explaining technology skill transfer (cognitive load theory: expertise principle)	Verbally equating video visits and patient portals to other technologies patients already used “It'll set up and be kind of like a FaceTime between you two you'll both be able to see one another.” [S8] “I see that they have a smartphone. I ask them, ‘Do you have a Facebook?’ And they tell me, ‘Yes.’ That's when I just tell them, ‘It's the same thing. It's the same as Facebook, if you know how to use it. It's the same.’” [S6] “We try to explain that it's really easy and we might say, ‘Oh, do you use text messaging to communicate with your friends or family?’ And if they say yes, then we say, ‘Okay, this is super easy. You’ll get a text. It has the doctor's name on it, so you know that it's from them.'” [S2] Increased patient comfort with technologies “I just use real-life experiences, like Facebook and WhatsApp. I just tell them examples. And from there, they feel more confident because they know how to use Facebook.” [S6] “I got more comfortable doing a Zoom meeting kind of thing, so it worked.” [P108] “Yeah because it's really, once you learn that's something you'll really like. It's like doing FaceTime.” [P112] “…he would just ring me like if it was a call on WhatsApp or on Messenger, like that but I would answer and it was him.” [P119]
Intermediary approach 3. Supporting the development of schemas
Detailing what to expect (cognitive load theory: pretraining principle)	Explaining what telehealth is, and how it compares to usual visits “I…let them know that it's like a video call with a doctor where she can see you, you can see her. Like you in person, you can still tell her your concerns, and you guys can talk to her. It's like a regular visit.” [S11] “There has been a few times where I use the word telehealth like, ‘Okay, your next visit will be telehealth.’ And they're like, ‘What's telehealth?’ And I'm like, ‘It's where I call you.’ ‘Oh, okay.’ Or like, ‘We'll do videos.’” [S1] Patients compared their experiences of video and in-person clinical consultations “…the medical assistant said…That it's going to be the telehealth over…video…they explained that the medical assistant… was going to call me and go over my history, and any concerns I want to relay to the doctor. And that link will be sent to you via text message. And when you get to that link, click it, and she'll be right on…it's different because when you're used to going into a doctor's office and talking to them, it's different. I never had telehealth visits before 2020…[I was] Curious, really, and see how it was going to work…It worked out pretty good…” [P104] “…the word telehealth. I mean, that didn't quite… I knew what it meant but initially I was like, ‘Oh, I never really use that word[telehealth]’…some of those words, were new to me… video was…pretty cool because she got to see me and I got to talk to her… I feel like the interaction, it was as natural as being in the office. That actually surprised me. I thought it would be kind of weird but it really wasn’t. It was fine…video feels like you’re just in the office. It feels the same to me.” [P105]
Segmenting instructions	Staff gave instructions that divided the process into smaller, sequential steps “…this is going to be a video appointment, so you’ll get a text message when the provider’s ready to start their visit. It will say, ‘hi, this is doctor whoever.’ And there’s a link in the text message. You…need to click on the link to start the connection process. It’ll ask you then to type your name in, and then you have to enable the camera on your phone. And from there, it’ll complete the connection process on its own. And then you’ll see them pop up on your screen and they’ll see you on their screen.” [S2] “…they explained it detail by detail…” [P116]. “They gave me the instructions. The website to log into. How to type in the doctor that I want to see. How to type in the password or whatever and I put in the password. I had to type in my name and things like that. “[P115] Offered Clear Understanding and Created a Sense of Readiness “I didn’t [have questions] because the information they sent me, to me it was clear cut, straight to the point…one step, two step, three step. That helped me.” [P101]. “Someone from their own front office, called me and explained to me a couple of days before, of what was going to happen, how to do it. ‘You’re going to do it like this…’ So …when she called I was ready.” [P107]

Explaining technology skill transfer. This approach involved verbally equating video visits and patient portals to other technologies patients already used. When discussing skill transfer, patients and providers mentioned various technologies, including FaceTime, Zoom, WhatsApp video, Facebook Messenger video, Skype, and text messaging. Two staff explicitly told patients that they could apply that knowledge to telehealth if they knew how to use other video-conferencing technology. Others described how conducting a video call on other platforms would be similar to conducting a telehealth call. In line with the expertise principle, this approach supported the activation of previously acquired knowledge, which might be available for patients to organize any new information³⁶ about how telehealth visits were performed specifically at the study FQHC.

While most of the evidence for explaining technology skill transfer came from the staff interviews, patients sometimes referred to telehealth calls as “Skype” or “Zoom” calls, showing familiarity with parallels between these platforms. This approach increased patient comfort with technologies, especially if patients were nervous about conducting an initial video visit.

Supporting the development of schemas

Long-term memory is organized into schemas, “stable patterns of relations between [knowledge] elements.”³⁷ In CLT, acquisition of schemas is central to learning and can be facilitated by implementing the pretraining principle and segmentation.Detailing what to expect and segmenting instructions for connecting to video visits helped patients create an overall understanding of how telehealth visits worked visits before their visit.

Detailing what to expect. One aspect of building patients’ expectations was staff explaining what telehealth is, and how it compares to usual visits (Table 3). These approaches align with the pretraining principle that involves helping patients build a mental model (“schema”) of video visits prior to that visit. Some patients compared their experiences of video and in-person clinical consultations, in terms of what they initially expected and their postvisit assessment of that experience, which suggests their mental connection between experiences.

Segmenting instructions. Staff gave instructions that divided the process into smaller, sequential steps via email or phone before the visit (Table 3). According to the segmenting principle in CLT, “…allow[ing] time between successive bite-size segments,”³⁵ may help in reducing CL. Patients largely found that this approach offered clear understanding and created a sense of readiness for them.

Reducing the extraneous load of negative emotions

Many intermediaries reported trying to help patients access and use telehealth technologies, especially video visits, while conveying emotional warmth, encouragement, attunement, and presence. The intermediaries’ approaches may thus have reduced extraneous CL that can emerge from negative emotions.³⁰

Walking patients through the process. Both patients and staff spoke about patients being “walked through” the steps of having a video visit. This referred to a “hands-on” interaction in which staff and patients followed the process together in time (Table 4). To accomplish this, some staff mirrored the patient's view of the patient portal or telehealth prompts on their own devices to see exactly what the patient saw on their screen. The interaction occurred synchronously—sometimes screen-by-screen and click-by-click. This interaction thus centered the patient in any problem-solving, with the staff acting as immediate intermediary support. Patients also found that this staff accompaniment resulted in them not feeling “alone” (Table 4) and reduced negative emotions like worry. Accordingly, this may have reduced extraneous load for patients.³⁰ At the same time, for users overwhelmed by the visual information on the screen, this approach aligned with the CLT’s off-loading principle, which operates by “…mov[ing] some essential [cognitive] processing from visual channel to auditory channel.”³⁵ For some, this approach resulted in successful visits despite patients’ initial apprehension or challenges (Table 4).

Table 4.

Quotes regarding intermediary approach 4.

Intermediary approach 4. Reducing the extraneous load of negative emotions
Walking patients through the process	“Hands-on” interaction with staff and patients following the process together in time “Mr. [Name], you are on the screen where it shows such and such?’ And I say, ‘Yes.’ They said, ‘Okay, click down here or enter this and it will take you to the next point,’ which it did. And then when I got to the point where it says, ‘You are in the waiting room. Your Zoom meeting will start. The person has not entered the chat yet.’” [P113] “…when the patient needs help figuring out how to get onto their portal account, or figuring out how to log into the visit…I do it with them at the same time on my computer. So that I can visually see it. I have a really hard time when the patient’s like, okay, I’m on this screen now, what do I click? I’m like, I have no idea.” [S8] “If you can have them call you prior to their appointment, so I can join in and walk you through the process. Once the call starts, I wouldn’t be on the phone with them, but I’m just walking them through the process, until it’s effectively where you check in to see the doctor. So, I’m on the phone with them, all the way up until they check in…Sometimes, some patients need that.” [P115, a patient who also worked at a clinic] Not feeling “alone” “…[have someone available to walk patients through] so that they don’t have to feel alone when they’re going through, even though it’s simple, but it’s just overwhelming.” [P103] “…let [patients] know that they’re going to be walked through the process. They’re never going to be on their own. They don’t have to worry…If anything should happen and they need assistance, they could call back and they’ll be walked through again.” [P112] Successful visits despite patients’ initial apprehension or challenges “I expressed that I’m not really in for all that technology, so it’s going to take me time to make sure I’m looking at what I’m looking at and it was easy after that…they asked me, “What did I do?” and where exactly was I at on my screen and I told them…They were able to guide me through it.” [P113] “…they were specific. At which one to click, but click. And then when I got to the next page, they didn’t want me doing anything, trying to go ahead. Which I had to backtrack because I did it. You know, sometimes I’m like a child. I’m impatient. But I learn quickly to stop and listen. And then they would walk me through it. And then they would say, “Well can you see yourself?” Well no. “okay, well you’re not in the right one. We need you to go to your settings.” And so yeah, it was wonderful. …Once she walked me through it and made sure I got to the link, to the site where I could actually see myself…as long as I could see myself to stay on the line until the doctor came on…” [P112]
Encouraging and reassuring patients	Emboldening patients to try and persist with using technology “…[the portal is] something that we try to really ask that they do because it’s good for them. Their results, everything is on there. They can talk to their doctor and send messages.” [S7] “I was just telling her that she can do … You can do what I said. That she can continue on. Because when she wanted to give up, I'm like, ‘You know, you can go ahead. You can continue on. You can just let me know what it is where you stopped at, where you're having a[n] issue…and I'll try to explain it to her.’” [S11]. “Some of the encouraging words that she was telling me and some of the techniques that she was telling me to use, when I was beginning to feel a certain type of way.” [P115] Enhanced patient self-efficacy “…when we explain to them that it's really just a matter of one click and two other pieces of information, they feel more confident that they can actually do that. They're like, ‘Oh, that's not that difficult. I can do that.’” [S2] “In the beginning of the process, still uneasy, but after the process was over, I was like, ‘Wow. This [telehealth] is cool. I can do this all the time.’” [P114] “He just said that, ‘It's so easy, don't worry.’ And I went, ‘Okay.’ And it was.” [P111]
Displaying positive affect	Modulating tone of voice and facial expressions “…[I] couldn't feel judgment simmering in his voice. Just very down to earth and talked to me in a respectful way…I really liked this person's voice.” [P11] “[When I logged into a video visit with my doctor] It was his usual smiling face…” [P11] Reducing negative emotions that may be present due to health or technologies “I have really bad anxiety to the point where my doctor always asks, ‘Why is your heart rate so high?’ ‘It's because I'm nervous.’…the staff… answer questions if I have them. They're confirming it. ‘Okay, I'm just going to wait until the doctor initiates.’” [P108] “[I was more comfortable] because [of] the tone… even though we aren't face to face, the tone says a lot.” [P115]

Encouraging patients. Providers emphasized the importance of emboldening patients to try and persist with using technology in the face of challenges. Staff offered encouragement by using positive descriptors to build patients’ confidence in the simplicity of completing a video visit. This typically involved adjectives such as “straightforward,” “simple,” “easy,” “quick,” “not difficult,” or “just a few clicks” when explaining video visits. Patients and staff also mentioned an increase in patients’ confidence in using technologies; that is, enhanced patient self-efficacy following this tactic. Prior research has shown that increased task-related self-efficacy is associated with higher working memory resources and lower CL.^38,39

Displaying positive affect. From patients’ perspectives, staff displayed positive affect by modulating tone of voice and facial expressions (Table 4). Patients described this behavior as reducing their anxiety or increasing comfort, thus reducing negative emotions that may be present due to health or technologies. Negative emotions such as worry and frustration can increase CL.^30,40

Discussion

Staff intermediaries used 4 approaches to support access to and usage of telehealth services for FQHC patients: (1) shielding patients from cognitive overload; (2) drawing from long-term memory; (3) supporting the development of schemas; and (4) reducing the extraneous load of negative emotions. There were also 11 subapproaches that enacted these broad approaches. Notably, many approaches could contribute to CL reduction and successful performance of telehealth-related tasks. Indeed, most of these approaches were viewed as helpful to at least some patients, alternatively resulting in: clear understanding; mapping of video visit processes onto prior knowledge; perceptions of readiness and comfort for telehealth visits; perceptions that video visits were straightforward; and successful visits despite some patients’ initial apprehension or challenges. Psychologically, intermediation also illustrated how to enhance patient self-efficacy and reduce negative emotions from either using technologies or the health issue at hand. Patients who lacked appropriate technology were automatically referred to phone visits, avoiding potentially taxing efforts to participate in an impossible task. Additionally, staff explained processes in Spanish despite English-only platform communications and switched to telehealth software that overcame language barriers for their Spanish-speaking patients. Together, these practices likely reduced confusion among these patients. However, there were limitations to some approaches since patients could not always control their home environments, device capabilities, or Internet access.

Our findings highlight staff intermediaries’ central role in making telehealth work for FQHC patients. Their importance is underscored by the fact that FQHC staff began providing intermediary support emergently; that is, they offered it in response to novel patient needs that surfaced upon rapid telehealth implementation. The need for this support may have been magnified by the nature of the patient population, which was primarily low-income and had varying technology ownership and experience. Furthermore, some staff also had large Spanish-speaking patient populations, necessitating workarounds due to English-only platform communications. Similarly, previous computing research has explored the use of intermediaries to bridge difficulties in access to and usage of technologies among people with limited digital access and skills.^41–44 Sociologists and communication scholars have also documented the importance of intermediaries, especially family and community members,⁴⁵ in using technologies among people with little prior technology experience, particularly for older adults.^46–49 However, access to intermediaries is variable, and healthcare-based technology support remains important since low-income people may have less technology expertise in their communities.⁵⁰ Our findings suggest that human intermediaries are core infrastructure for addressing telehealth access inequities.

Notably, without being trained to do so, FQHC staff discovered approaches that helped patients gain access to telehealth even though both they and patients had little prior experience with these technologies. Critically, these strategies aligned with several established “principles” for reducing CL, indicating that CL reduction can facilitate equitable access to and use of telehealth, especially among FQHC patients. Drawing from these emergent intermediation strategies, future healthcare intermediary interventions may benefit from application of CLT in their design, as we have done in related work with this partner FQHC.³¹ This conceptual framework may complement commonly used frameworks such as “digital/e-health literacy”⁵¹ by focusing more on the process of instruction alongside imparting specific content.

Although technology intermediation may initially appear to be a cognitively focused task, we found that emotional support was crucial. Human intermediaries' ability to accompany apprehensive patients by “walking them through” technology use resulted in successful visits. The emotional impact of these interactions (“not feeling alone”) resonates with social support research that shows that the simple presence of other people—even strangers—can “buffer” or reduce stress.⁵² Additionally, intermediaries responded to patients with positive affect, encouragement, and reassurance. This echoes sociological research on “warm experts”: people with relatively advanced knowledge of technologies who are pleasant to deal with and readily available to people in their daily lives.⁵³ Warm experts can informally assist people with little computer experience accessing the Internet.⁵³ Extending warm expert theory, our findings suggest that emotional support may help patients use telehealth technologies, partly by helping patients to reduce extraneous CL and focus working memory on the task at hand.

Often, learning about technologies occurs in informal settings outside of schools or training sessions.⁵⁴ Thus, we hold that learning to use telehealth technologies allows patients to gain technical knowledge and skills critical to accessing healthcare today. While many intermediary approaches may have facilitated learning about video technologies and patient portals, other approaches may have limited potential learning opportunities. Specifically, applying black box fixes did not allow patients to develop comprehension. This is understandable since it solved the issue quickly and allowed visits to proceed. However, staff providing brief explanations about problems and solutions might help future patients develop troubleshooting schemata that might help them in other situations.

While staff intermediation helped patients, it is important to recognize that intermediation work was not in these staff members’ job descriptions, and intermediation was added to already-busy workloads. Additionally, providers were not trained to facilitate telehealth access and were learning to use technologies alongside patients. Looking forward, we foresee a need to formalize human intermediary roles and account for them as a core part of telehealth implementation at FQHCs and other sites with low-income or underserved patients. Ideally, a dedicated staff member could provide these types of digital support, such as a digital navigator.^55–57 Clinical researchers have recommended⁵⁸ that healthcare organizations work with community-based organizations and public libraries to develop and implement digital navigation efforts. In the United States, FQHCs and other nonprofit healthcare organizations may be able to apply for funding of digital navigators through state grants made available under the Infrastructure Investment and Jobs Act (IIJA). The Community Health Network of Washington provides an example of a recently funded IIJA project for digital navigation.^59,60 However, if funding for a dedicated digital navigator is not feasible, intermediation should be written into staff job descriptions, and adequate training to perform these functions should be provided. Such training could include embedding telehealth education into healthcare curricula,^61,62 reinforcing telehealth core competencies upon onboarding,⁶³ and continuing education.⁶⁴

Our application of CLT illustrates the existence of multiple theoretically informed, real-world approaches that could inform the design for intermediary telehealth training. Beyond training, providing clinical staff with standardized tools such as a patient telehealth prescreening and a telehealth troubleshooting checklist⁶⁵ may help providers to assess patient needs for telehealth connectivity and provide a structured approach to resolving common issues. Findings from this study could also inform the design of workflows within telehealth platforms and patient portals. For instance, videos or tutorials could be embedded in visit request modules so that patients better understand what to expect, and improved status indicators could help patients keep track of where they are in the telehealth visit process. Workflows embedded in patient portals and telehealth platforms should also segment activities into small, action-oriented chunks to reduce CL. Furthermore, as in this study, previous research has identified language barriers as an impediment to telehealth use for patients whose preferred language is not English.⁶⁶ There is a critical need for telehealth and patient portal platform vendors to provide non-English-language interfaces and instructions to meet the needs of patients in our increasingly multilingual society.

Limitations

There are a few study limitations. The interviews for this study were conducted at 1 FQHC in 1 region of the Midwestern US. While this may impact the generalizability of intermediation strategies, there was a diverse representation of patients across age and race/ethnicity. Patient participants may have had higher levels of digital literacy than the overall FQHC patient population. Despite this, intermediaries still played critical roles in facilitating their telehealth access. Additionally, these interviews occurred during the first year of the COVID-19 pandemic, and digital barriers may have been exacerbated. However, given the speed of technological change, we anticipate an ongoing need to support patients in using new technologies.

Although the staff study population was demographically homogenous, participants reflect the national demographics for these occupations as reported in the 2022 Labor Force Population Survey, where the majority of medical assistants, receptionists, and nurses were female^67–70 and White individuals.^71–73 A strength of our study is that our sample size aligns with qualitative study guidelines for 15-30 overall participants,⁷⁴ the need for fewer individuals when the population is homogenous^75,76 (as reflected by the staff participants), and the quality of our data by conducting interviews with 2 subgroups and triangulating findings from each. However, the limited demographic diversity represented by staff in our study, along with participants’ challenges with the Spanish-language translation emphasizes the need for digital navigators who are representative of patients’ cultural background.

Because patient inclusion criteria required that they had a prior telehealth visit via video or telephone, study participants may have included patients with a stronger preference or motivation for telehealth use than those without telehealth visits in the same timeframe. Thus, study findings do not show skills acquisition among patients who had never tried a telehealth visit—perhaps due to a lack of motivation or opportunity to do so. Data collection involved gathering participant demographic and technology ownership data via a survey separate from the interviews. While all staff completed these separate surveys, 2 patients did not provide the requested survey responses at follow-up. Future research with FQHC and other socioeconomically marginalized patients may benefit from collecting such data during in-person data collection, such as verbally at the end of a qualitative interview. Lastly, while we used CLT to analyze qualitative interviews, future intervention research should examine potential CL reduction due to human intermediation.

Conclusion

Healthcare staff played an essential intermediary role in assisting FQHC patients in successful telehealth participation. Without prior training, staff used approaches central to CLT. These approaches were viewed as helpful to at least some patients, with beneficial impacts on learning, perceptions about the self, perceptions about the technology, and emotions. Furthermore, intermediation resulted in successful visits despite some patients’ initial concerns or difficulties. Further investments are needed for ongoing intermediary support, especially at FQHCs and in limited-resource areas, to facilitate the equitable participation of patients in telehealth. As prior literature has identified, without intermediaries, “those left behind…will face additional, perhaps insurmountable, barriers.”⁷⁷

Author contributions

A.K.W.: conceptualization, data curation, formal analysis, methodology, investigation, visualization, writing—original draft, writing—review and editing. M.G.A.: conceptualization, methodology, formal analysis visualization, writing—review and editing. S.D.: conceptualization, methodology, formal analysis, resources, writing—review and editing. V.K.: investigation, writing—review and editing. T.R.D.: conceptualization, methodology, formal analysis, supervision, writing—review and editing. L.R.B.: conceptualization, methodology, writing—review and editing. T.C.V.: conceptualization, formal analysis, validation, funding acquisition, project administration, methodology, supervision, writing—original draft, writing—review & editing.

Supplementary material

Supplementary material is available at Journal of the American Medical Informatics Association online.

Funding

This research was supported by the National Science Foundation (NSF) RAPID-COVID-19-#2031662.

Conflict of interest

One author is an employee of the FQHC where the research was conducted. The other authors report no conflict of interest.

Data availability

The qualitative data gathering during this study is not publicly available and cannot be shared due to confidentiality, since participants are potentially identifiable from the information contained in the data. Furthermore, ethical restrictions imposed by the consent process used with participants prevent data sharing and data requests. Any questions about this can be directed to the corresponding author.