An Untapped Potential in Primary Care: Semi-Structured Interviews with Clinicians on How Patient Portals Will Work for Caregivers in the Safety Net

Introduction

Background

Growing evidence on the benefits of patient engagement has fueled health systems’ focus on patient portals as a central access point for the future of primary care [1,2]. The financial incentives of the Meaningful Use Program, as part of the US health care reform in 2014, spurred a rapid uptake of patient portals across health systems nationwide [3-5]. Emerging national evidence in the last 5 years indicates that the digital divide is shrinking, as 80% of the US population owns internet-accessible smartphones, without differences by race or ethnicity and with few differences across income categories nationally, especially when it comes to mobile use [6,7].

Despite improved health information technology (IT) access, there is a large body of evidence demonstrating significantly lower use of digital health care tools among underserved populations, underscoring the need for more research to understand the contextual factors affecting their use [8,9]. Vulnerable patients within safety-net settings (eg, low socioeconomic status, under- or uninsured, limited English proficient [LEP], aging, and physically or mentally disabled) are interested in accessing their health information on patient portals but are less likely to actually do so. Numerous studies have shown that there is a high level of patient interest in portals, repeatedly documented among LEP populations, Medicaid/Medicare recipients, and older patients [10-14]. A 2014 survey of Spanish-speaking patients in a Los Angeles Federally Qualified Health Center found that the majority of patients had computer access (66%), internet access (78%), a current email account (78%), or a smartphone (71%) and that 75% of patients were interested in using a patient portal [10].

Family and friends are important resources to assist these underserved populations (especially those that are Medicaid recipients, older patients, or LEP) in navigating health systems, and provider support could facilitate their portal engagement. Caregiver engagement with a patient’s portal may range from simply assisting the patient to log on and use the portal to logging onto a patient portal for the patient (unofficial patient surrogate) all the way to acting as a registered proxy for portal access (where caregivers are granted their own registered access to a patient’s health information on the patient portal with the patient’s permission) [15,16]. Older patients, especially those who are LEP, tend to navigate health system processes with family members or close friends who schedule and attend medical visits, coordinate care, manage medications, assist with self-care tasks, and facilitate transitions across care settings [15,16]. Patients with low health literacy similarly benefit from the involvement of trusted family or friends during medical visits [17,18]. Prior studies among older patients have confirmed that family members facilitate patient access and use of the patient portal [19,20] and utilize the patient portal for the patient [21-25]. Given that these patients are interested in sharing electronic health information with family members or close friends [19,20,26,27], tailored strategies are needed to engage family members or friends in accessing health care digitally, especially in safety-net settings with a majority proportion of vulnerable patients and where caregivers play a key role in patient care.

The growing role of caregivers and caregiver/proxy portal use among vulnerable patients presents a potential well-aligned digital strategy for reducing racial/ethnic disparities in health that can be facilitated via already available health IT. However, buy-in from physicians and other health care providers will be important in promoting patient portals on this next level, especially in safety-net settings [28,29]. The aim of this qualitative study was to explore safety-net providers’ perspectives on caregiver/proxy portal use.

Methods

Design and Study Setting

The study was approved by the University of California, San Francisco and University of California, Los Angeles institutional review boards. For this descriptive, qualitative study [30], we conducted individual telephone semistructured interviews with a purposive sample of physicians or providers (n=17), working in a California safety-net setting in Los Angeles, San Francisco, or Alameda County. Los Angeles participants were recruited from the Los Angeles County Department of Health Services (LAC DHS) system. LAC DHS forms the core of the health care safety-net for indigent populations in Los Angeles County, the largest, most ethnically diverse county in the United States. LAC DHS serves more than 10 million residents and provides over two and a half million ambulatory visits every year [31-33]. In northern California, providers were recruited from the San Francisco Health Network (SFHN) and the Alameda Health System (AHS). The AHS is the East Bay’s (Oakland area) public health system with 5 hospitals and 4 primary care clinics, whereas SFHN is San Francisco City’s and county’s public health care system, providing primary care at 14 clinics, youth-focused care at 11 clinics, and care at a hospital serving over 1 million residents. The internet patient portals for the safety-net health systems included in the study allow patients to view clinical data (laboratories and radiology reports), refill medications, request/change appointments, message their health care teams, read about their prior and current diagnoses (links to educational content), and view/download their medical records.

Participants and Recruitment

From April to November 2018, we recruited and interviewed providers from the LAC DHS, SFHN, and AHS. Providers were sought if they had clinical activity in a safety-net health setting and/or held a leadership role in any safety-net clinical setting. Los Angeles participants were recruited via personal recommendations from the medical director of digital health services in LAC DHS. The SFHN and the AHS providers were recruited via recommendation from study coinvestigators. We conducted 45- to 60-min semistructured interviews with health providers from these 3 large California safety-net health systems. Participants were interviewed via telephone. AC, AGC, WS, and CL conducted the interviews on multiple days until at least three providers were interviewed from each of the 3 health systems. Participants were provided with a US $25 gift card after completing their scheduled interviews.

Data Collection

The semistructured interviews focused on providers’ experiences with caregivers, the roles caregivers play, and how the portal could be leveraged as a tool to support caregivers in their responsibilities. The discussion questions were modified from projects led by study investigators who examined barriers and facilitators to portal use to develop a randomized trial on patient portal training [34]. This prior research highlighted the limited knowledge and understanding of caregivers’ use of the patient portal.

Discussion introductions included participants’ self-describing clinical and administrative roles. Participants were then asked about their experiences with patient caregivers (definition, types of caregivers, and positive and challenging interactions). They were subsequently tasked to describe strategies that they have observed caregivers take on with patient health management. We asked participants about how the patient portal affects these caregiver strategies, what they thought was useful, and how they envisioned the portal could affect these patients’ and caregivers’ care. The discussions wrapped up with a conversation about what the health system could do to help improve caregiver and patient engagement with the portal.

Analysis

Interview discussions were audio recorded and transcribed, reviewed for accuracy, and deidentified. AC and AGC independently read and summarized 6 transcripts over multiple theme meetings, applying inductive and deductive methods to identify the spectrum of themes encountered for each of the interviews with accompanying example quotations. AC and AGC then cross-referenced their independently developed list of themes. Through codebook discussions, they reached a consensus for a more final codebook. AC and AGC then coded 3 transcripts independently with this codebook for validity. The definition for each theme (code), with the final example quote, was developed by AC, with iterative feedback from the AGC and the entire group, until a clear consensus was achieved. Using the established codebook and definitions, a third coder (ML) analyzed the transcripts using Dedoose version 8.2.14. AC and AGC reviewed the analysis by ML and established a consensus regarding any other discrepancies in themes and corresponding quotes [35].

Results

Participant Characteristics and Theme Categories

Of the 16 participants interviewed, 4 specialized in geriatrics, and all held a leadership or administrative roles, as shown in Table 1. All the participants worked in primary care settings, including a speech language pathologist.

We described primary themes highlighting safety-net physicians’ and providers’ recognition of potential strong benefits associated with caregiver portal use and specific barriers to caregiver engagement. The 4 major themes are separated into 2 large categories: (1) positive aspects of portal use by caregivers (Table 2); and (2) challenges to address for portal use by caregivers (Table 3). These show major themes, subthemes, frequencies, and quotes.

Positive Aspects of Portal Use by Caregivers

Table 2 displays themes and subthemes for the category describing positive aspects of portal use. In the theme, caregiver designation in the electronic health record, participants described various potential pathways for identifying caregivers via the electronic health record (EHR) and how this would be useful. Some participants also described the challenges of not having a workflow to readily integrate current caregiver information into the EHR and how having such information available would make it easier for physicians and other providers to engage the caregivers during and outside of the clinic encounter. This was important, as the participants in the study noted multiple caregiver types, which are present in different ways for patients within safety-net settings: relatives, friends, case managers, in-home supportive services, and social workers.

Participants also discussed caregiver use and potential use of the portal, detailing current uses of patient portals by caregivers and aspirational ideas about what the patient portals could potentially do for safety-net caregivers in terms of facilitating patient care. This theme was organized into 3 subthemes: for subtheme 1, portal as a tool to assist caregivers with standard health care tasks, participants mentioned ways that the current versions of their patient portal supported or eased the job of the caregiver by facilitating simple health care tasks for the patient via the portal (eg, obtaining medication refills or making appointments), making it easier for the caregiver to care for the patient (as they might have to spend less time on the phone making calls and possibly decreasing the need for face-to-face visits).

For subtheme 2, portal as a tool to directly support the caregiver, participants discussed how portals could provide well-being/resources specifically tailored for the caregiver. For example, they mentioned that patient portals could be a way to provide explicit support to caregivers (eg, preventing caregiver burnout) and connect caregivers directly to community resources (including information about self-care and programs that assist caregivers in the safety-net) that make the caregiving job more manageable. They also noted that such resources could include a tutorial on caregiver education/navigation. The portal could provide caregivers access to tools on how to care for patients and/or how to navigate a complicated safety-net health care and social service system.

For subtheme 3, expanding portal functionality for caregiver use, discussions centered on a wish list of what the portal could do in the future (Portal 2.0) for caregivers. Participants mentioned how technology could help paint a better picture of the patient and caregiver via telehealth (access to home context) and more frequent communication. Such modalities could facilitate the implementation of evidence-based initiatives such as advanced care planning in the safety-net.

Challenges to Portal Use by Caregivers

Table 3 displays themes and subthemes related to challenges to portal use by caregivers. The major theme, portal privacy and security in the caregiver-provider-patient triad, encompassed unresolved issues that these participants noted in communicating with the patient and caregiver through the portal. They raised several challenges that may have not been fully thought through yet as patient portals are being marketed to caregivers in the safety-net. These were summarized in 3 subthemes.

For subtheme 1, portal triad relationship, participants described the nuances of building or managing a relationship with the caregiver, as well as the patient, through the portal. For example, many times throughout these discussions, physicians cited the many nuanced ways in which they would validate caregiver reporting during clinic visits (talk to patient and caregiver separately), check for caregiver abuse of the patient by examining the patient or reading body language, and simply also assess how the caregiver was doing with the burden. They noted that this subtle art of relationship management and checking in with a caregiver, which are essential features of the triad relationship [24,36,37], would be very difficult to achieve via digital communication and expressed concern that some of these needed in-person checks and balances would be lost via the current version of our patient portals.

For subtheme 2, control of shared access, participants mentioned concerns about proxy access. For example, what if a patient allows a caregiver to log in as a proxy into their patient portal, but the patient does not want the caregiver to see their entire medical record? Or what if the proxy sends messages to the provider on the portal as the patient, but the provider is not aware that they are communicating with someone that is not their patient? This is an important concern in a safety-net population where patients may have sensitive diagnoses (such as sexually transmitted infections, illicit drug use, addiction, or mental illness) that have not been disclosed to the caregiver or other close friends or family because they carry some social stigma. Sensitive information on the portal could also reveal risks to family caregivers that they may not have been made aware of (genetic diseases or genetic risk factors). Participants also noted the simple workflow barriers to creating proxy login for caregivers. Although most systems have the potential to do this, most clinical settings do not outright offer it and/or facilitate the registration process.

For subtheme 3, general portal security and technology concerns, safety-net physicians made general comments about security or privacy issues with technology in health care. Providers worried that their patients and caregivers would not use a portal secondary to fears about their private data being accessed by outside parties (be hacked).

In the last major theme, barriers to caregiver enrollment and use, participants noted the challenges to enrolling caregivers as proxies into their patient’s portal. Many of the barriers noted paralleled the general challenges to portal use among patients in the safety-net: decreased access to and familiarity with technology (low digital health literacy), lack of easy access to the internet at home, being LEP (when most of the portal is in English), fear of signing up because of immigration documentation status, lack of awareness about a patient portal in the safety-net, or lack of understanding of what a patient portal can do.

Discussion

Principal Findings

Patient portals and caregiver engagement in primary care both represent clear ways to improve our communication processes with vulnerable patients by making care more convenient and coordinated. However, even in large integrated delivery systems with established track records of portal use, there is very little research on the role that caregivers can play in terms of digital communication via the EHR. Our study echoes some of the prior literature recognizing the clear need for better functionality of portals for caregivers (specific content for them and better ways for patients to designate what types of information or access to share) [22-24]. We also shed light on some specific needs in this population, such as the needed support for caregivers with communication or language barriers (which is of high importance in safety-net settings).

Beyond the health care setting, there is already a team around many of our patients—family members and friends who serve as caregivers and trusted confidants when making health decisions. Health leaders must recognize the need to make it easier to connect with these trusted team members, in addition to the patient, when creating care plans together, especially as systems focus on the patient portal as a primary health management tool moving forward. The next version of portals can potentially improve the integration of telehealth options into the interface and provide resources that are specific to assisting caregivers, such as local resources, to reduce or prevent caregiver burnout. These innovations may be important for patients and caregivers in safety-net settings.

Participants brought up many important privacy and communication issues to resolve and improve to make portals work for safety-net caregivers and patients. Although most safety-net EHRs have the technical capability to create proxy account functionality [38], limited information exists regarding the use of shared access. Available data indicate that health system implementation of shared access functionality has been variable [39], but national estimates are that up to 30% of portal users have used the portal as a proxy for someone else (with about 50% of them as a formal caregiver/proxy login) [40,41]. Prior studies have evaluated patients’ experiences with accessing and using their own patient portal account [21] and preferences for sharing their electronic health information with others [27,42,43]. These studies suggest that some patients would like the option of selecting the level at which a proxy caregiver has access to their complete medical record. As such, we found that addressing the privacy and security nuances of building a digital relationship with a caregiver in the safety-net was a major barrier to physician and provider participants being completely supportive of proxy patient portal use. In the safety-net, physicians have reported the need to provide limited portal views of sensitive information (eg, HIV results and prior intravenous drug use history) for patients who rely on family or friends as informal or formal caregivers. Many times, the option to use the portal is foregone by patients because of this current all-or-nothing access approach to patient information via a proxy login. One solution to this dilemma is to allow patients the ability to choose the level of proxy access a caregiver or family member will have. EHR vendors should be incentivized via health information policy to innovate, and safety-net health systems should advocate for more patient-centered options around level of access to health data via proxy relationships, especially when patient privacy and security remain a top level of concern among safety-net patients. This is a feature that should be available to all patients and will resonate with several groups, including adolescents, older adults, and patients with some sort of impairment or disability, in other health care settings.

Limitations and Strengths

Among the limitations of this formative study are the small sample size and generalizability to other safety-net health care systems, which are different from the California settings included in the study. However, the purpose of this qualitative study is to generate initial insights about provider buy-in regarding portal potential for caregivers in the safety-net.

This is one of the few studies to probe into provider perspectives around the role of caregivers in portal use within the safety-net. We found that providers recognized the potential for portals to improve information delivery and communication, benefiting caregivers who are assisting socially and medically complex patients in the safety-net. These safety-net health care providers and leaders focused on expanding the functionality of the portal, so that it does more for caregivers.

Conclusions

Overall, safety-net systems that seek to maximize the use of portals in their communities must develop specific strategies to uphold patient preferences and innovatively support caregivers while maintaining privacy and security. Further work is needed to assess and support diverse caregivers’ access to portals by addressing pivotal barriers, such as health or digital literacy and English proficiency [44]. Safety-net health systems provide health care for our most medically and socially fragile patients: populations that include patients with multiple morbid conditions, limited English proficiency, cognitive impairment, physical and mental disabilities, low literacy, homelessness, substance use disorder, justice involvement, and immigrant and refugee status. As they serve the most vulnerable, safety-nets are the ideal places to develop and refine the next patient and caregiver-centered iterations of the EHR and the patient portal. However, to make this tool work for our most vulnerable patients, we must take intentional steps to ensure that the patient portal can be effectively and efficiently deployed by their caregivers in the systems that serve these high-risk patients.