“It made me feel like I wasn’t alone in the darkness”: exploring dementia care network communication and coordination through a digital health platform

Abstract

Objective

To explore the use of a shared communication and coordination platform—the CareVirtue journal feature—for care networks of people living with Alzheimer’s disease and related dementias to inform the design of care network support technologies.

Materials and Methods

In the primary study, care networks comprised the primary caregiver and other caregivers they invited to participate (eg, family members, in-home aides) used CareVirtue, for 60 days followed by a semistructured interview to explore primary caregivers’ perceptions of usefulness. This secondary analysis focused on use of the shared journal feature of CareVirtue, which allowed care networks to communicate through posts that were shared with the network and to which network members could respond. Journal posts were analyzed using a deductive/inductive content analysis to categorize information behavior. We also conducted a thematic analysis of the interviews to identify primary caregivers’ perceptions of the journal’s usefulness.

Results

Care networks used the journal for: (1) information acquisition, (2) information sharing, (3) strategy development, and (4) information feedback. Thematic analysis revealed that caregivers felt the journal was useful at the individual, care network, and relational levels and that journal integration was influenced by care network structure and relationships.

Discussion

Care networks used the journal to document, share, and acquire information; co-create strategies; and provide support. The usefulness of this shared communication and coordination platform included individual and care network level benefits.

Conclusion

These findings point to the importance of caregiver-centered technologies that support both the individual primary caregiver and the care network.

INTRODUCTION

Society relies on informal caregivers (family and friends) to provide a majority of the long-term care for the nearly 6 million people living with Alzheimer’s disease and related dementias (ADRD) in the United States. Despite the estimated 18.5 billion hours of unpaid labor annually in the United States, ADRD caregivers are undersupported, under-resourced, and report persistent unmet needs.^1–4 The high level of physical and psychosocial strain that is often associated with the caregiver role can contribute to burden, depression, burnout, and other suboptimal health outcomes, resulting in caregivers becoming the “invisible second patient.”^5,6

Research suggests that ADRD caregivers often provide care in collaboration with a care network of both informal (eg, family, friends) and formal (eg, primary care providers, social workers) supports.^7–9 Care networks can range in size, with care network members operating across multiple roles (eg, primary, secondary, tertiary caregiver) and providing a range of supports from functional (eg, facilitating activities of daily living or medical needs) to relational (eg, emotional support for the primary caregiver).^7,9–11 If functioning optimally, the care network has the potential to support positive caregiver outcomes, decrease distress, and improve quality of care.⁹ For the care network to function caregivers must maintain effective communication practices, coordinate, and delegate responsibilities, and use collaborative strategies to close knowledge gaps between care network members.¹² Currently, caregivers report that the communication, coordination, social dynamics, and role ambiguity across a care network increase demands and add complexity to the caregiver role.^7,13–16 These care network-related demands have been recognized as part of the “invisible work” caregivers must perform and have been associated with increased workload and negative health outcomes.^7,17

To make the invisible work of caregivers visible, ADRD caregivers require tools that can support the full extent of their activities, including coordinating across a care network. Health Information Technology (HIT) interventions demonstrate great potential to support care network function due to their increasing accessibility, low-cost implementation, and ability to securely store and retrieve information.¹⁸ To date, most HIT platforms designed for ADRD caregivers have focused on enabling primary caregivers to connect with healthcare professionals, participate in support groups, and engage on caregiver forums.^19–21 These social support features have been recognized as important contributors to the ability of HIT interventions to improve ADRD caregiver outcomes, including mental health and caregiver burden.^22–26 Although valuable, social support features are most often designed for use by individual caregivers and as such do not effectively address the additional communication and coordination needs of the care network. In a recent evaluation of caregiver-facing apps on the mobile marketplace as of 2021, findings indicated that out of 17 available apps, only 2 of the apps included features related to care network communication and coordination and these apps were rated as below minimally acceptable quality based on the Mobile Application Rating Scale.²⁷

In summary, there is evidence to indicate that web-based interventions can support individual caregivers in their work and contribute to improved well-being.^21,28 However, there is little understanding around how such tools can support the care network. Studies suggest the inclusion of functionalities for delegation and coordination on HIT platforms could support information behavior by reducing the burden of information management among family and friends of older adults, but more research is needed to characterize the ways that care networks interact on such platforms to inform design requirements for future interventions.^20,21,29

Objective

The objective of this study was to inform the design of future care network support technologies by exploring the use of a shared communication and coordination platform for ADRD care networks. Specifically, we aimed to: (1) identify and characterize the ways caregivers used the platform to communicate and coordinate across the care network, (2) categorize caregiver’s perceptions of the usefulness of a care network platform.

MATERIALS AND METHODS

Design

The study reported here is a secondary analysis of a larger study to demonstrate the feasibility of a web-based application for caregivers (CareVirtue) and develop an intelligent caregiver assistant (R41AG069607) for caregivers of people with ADRD. We have termed this the “primary study.” The University of Wisconsin-Madison Institutional Review Board (IRB) approved this study.

CareVirtue was developed as a platform for caregivers to store and share information and communicate with members of their care network. A primary caregiver is the account owner and invites members of the care network to use CareVirtue with them. Functions of the CareVirtue platform include a personal care guide, calendar, geolocated resources, and the care journal.³⁰ To achieve our objective to explore care networks’ use of a shared communication and coordination platform for ADRD care networks, the present study focused only on the journal feature of CareVirtue. The journal provides a space for caregivers to document and share with the care network their daily care experiences, ask questions, and have conversations (Figure 1). The journal also includes prompts to provide a mood rating (from “Really Bad” to “Good”) with each post and a category tag to denote the content of the post (eg, “medication,” “mobility”). When a caregiver writes a journal post, it uploads directly to a newsfeed that is visible to the entire care network. Members of the care network can make comments on journal posts and respond to others’ comments.

Figure 1.

CareVirtue dashboard displaying the journal with the most recent journal entries, mood indicators, and category tags.

Primary study setting and sample

We recruited primary and secondary caregivers in Wisconsin and Southern California to participate in this study. We used multiple recruitment strategies in several counties to reach a diverse group of participants, including ADRD-focused nonprofit organizations and Area Agencies on Aging. We also recruited through the Wisconsin Alzheimer’s Disease Research Center. We defined a caregiver as a family member or friend providing any unpaid support (eg, direct care, coordinating support) for a person living with ADRD. We allowed for primary caregiver status to be defined by the participant. Eligible participants were 18 years of age or older, proficient in speaking and reading English, had daily access to the internet, and shared caregiving information with other caregivers (eg, other family members, respite care, in-home paid caregivers).

Primary study data collection

Participation began with a virtual enrollment visit in which participants provided informed consent, were oriented to the CareVirtue platform, and completed a demographic survey. Primary caregivers were encouraged to invite a minimum of one other caregiver to their CareVirtue account. Participants used CareVirtue for 60 days, and we encouraged participants to log in at least once daily. At the end of 60 days, we conducted a semistructured interview to understand barriers and facilitators to CareVirtue’s usability and usefulness. The semistructured interview guide (Supplementary Appendix A) was developed by a senior member of the research team (NW) with training in qualitative research and human factors engineering based on a standard interview guide developed to explore usability and usefulness of technology interventions.³¹ The initial guide was reviewed and revised by the study team who has expertise that includes engineering, caregiving, product development, technology commercialization, and qualitative research. The interview guide included specific questions to understand participants’ experience with CareVirtue (eg, “What parts of CareVirtue did you use on a daily basis?,” “How did CareVirtue fit with the tools and resources that you and your care team were already using?,” “How has CareVirtue influenced the involvement of other care team members?”) as well as broader questions on participants’ experience as a caregiver to understand how CareVirtue fit into the context of their role (eg, “What are the biggest challenges that you’ve been experiencing?,” “If you could build your perfect caregiver solution, what would it look like?”). Interviews were audio recorded and lasted up to 45 min. Participants were given a $150 gift card in recognition of their participation.

Secondary analysis

Journal post analysis

Textual data from all journal entries posted on the CareVirtue platform during the 60-day trial period, including replies to original posts, were exported and compiled into Microsoft Excel for coding. Journal posts were analyzed using content analysis and applying a combination of deductive and inductive approaches.³² To contextualize the findings within the field’s understanding of caregiver information behavior, deductive coding was guided by a published personal health information management (PHIM) model, which provides 3 categories of information management that have been identified in dementia caregivers: information acquisition and integration, information maintenance, and information sharing.¹⁴ This model was chosen because of its validation in the target population of dementia caregivers. However, because it was focused on the primary caregiver and not explicitly focused on care network-level information management, an inductive approach was also used to capture network-level communication and coordination activities. We first coded each entry using the categories from Holden’s PHIM model as described. Activities that did not fit within that model were coded inductively to capture aspects of information behavior that did not already appear in the model. Each journal entry was coded independently by 2 members of the research team (AL, DG, and AJ) with expertise in psychology and human factors engineering, and team meetings were held weekly to compare coding, resolve differences, and discuss any proposed changes to the codebook. Coding was also reviewed biweekly with a senior member of the research team (NW).

Post-trial interview analysis

Recordings of the participants’ final interviews were transcribed verbatim by a professional transcription service. We performed a thematic analysis based on Braun and Clarke’s 6-phase analysis process to explore participants’ perceptions of the usefulness of using the CareVirtue journal.³³ First, 2 members of the research team (AL and AJ) performed structural coding of the interview transcripts to identify broad passages related to the care journal, with each coder reviewing all transcripts. The coders met regularly with each other and a senior member of the research team (NW) to compare their coding and come to consensus on passages included. In the second phase of analysis, the goal was to generate initial codes. To do this, the same 2 coders reviewed the coded passages identified in phase 1 to inductively develop initial codes related to the usefulness of the CareVirtue journal feature. Two coders reviewed each transcript to identify the initial codes, which were discussed and refined by a subset of the research team in weekly meetings (AL, AJ, and NW). Each transcript was then coded by one coder using the codebook. Coding was reviewed and discussed by the research team in weekly meetings. In phase 3, the research team (AL, AJ, and NW) met to review the codes and coded excerpts and group them into themes. In phases 4 and 5, the transcripts were reread by the entire research team to confirm and further refine descriptions and rename themes to align with the context of the whole dataset. In phase 6, themes were further refined by the full research team during the manuscript writing process. To facilitate comprehensive coding, transcripts were reviewed in their entirety multiple times. Coding was performed independently, and we managed reliability using regular coder meetings and research team meetings to compare coding and resolve any disagreements through consensus discussion.³⁴

RESULTS

The primary study enrolled 51 primary caregivers (Table 1) and n = 61 care network members. Care network size (number of users invited by primary caregiver) ranged from 1 to 8 members. During the study period, 4 primary caregivers dropped out due to care recipient death (n = 2) and personal situations (n = 2). We were unable to reach n = 6 primary caregiver participants for the post-trial interview.

Table 1.

Primary caregiver and care receiver demographics

Characteristic	Primary caregivers (n = 51)		Care recipients (n = 51)
Female gender, n (%)	38 (74.5)		34 (66.7)
Age in years (mean±SD)	60.3 ± 9.8		79.2 ± 10.6
Race and/or ethnicity, n (%)
Asian	2 (3.9)		2 (3.9)
Black or African American	1 (2.0)		1 (2.0)
Hispanic or Latinx	2 (3.9)		2 (3.9)
Native American or American Indian	1 (2.0)		0
Not reported	1 (2.0)		1 (2.0)
White	44 (86.2)		45 (88.2)
Marital status, n (%)
Married or domestic partnership	37 (72.5)
Divorced	11 (21.6)
Single, never married	2 (3.9)
Widowed	1 (2.0)
Education, n (%)
Postcollege	19 (37.2)
4-year college	17 (33.3)
Technical school, vocational training, community college	10 (19.6)
High school diploma or equivalent	5 (9.8)
Employment, n (%)
Full-time	21 (41.2)
Retired	19 (37.3)
Part-time	7 (13.7)
Not working	4 (7.8)
Income, n (%)
>$100 000	18 (35.3)
$40 000–60 000	8 (15.7)
Do not wish to answer		8 (15.7)
$80 000–100 000		6 (13.7)
$60 000–80 000		4 (11.8)
$20 000–40 000		2 (3.9)
<$20 000		1 (2.0)
Location, n (%)
Wisconsin		29 (56.9)
California		19 (37.2)
Illinois		2 (3.9)
Virginia		1 (2.0)
Location type, n (%)
Urban	42 (82.3)
Rural	9 (17.7)
Relationship of caregiver to person living with ADRD, n (%)
Child			28 (54.9)
Spouse/partner			20 (39.2)
Other relative			3 (5.9)
Distance of caregiver to person living with ADRD, n (%)
In household			34 (66.7)
<20 min			12 (23.5)
>2 h			3 (5.9)
20–60 min			2 (3.9)
Living situation, n (%)
In a house			40 (78.4)
In a nursing home, retirement community, or other assisted living facility			9 (17.6)

Content analysis of journal posts

We identified 4 categories of communication and coordination activities for which participants used the CareVirtue journal: (1) information acquisition, (2) information sharing, (3) strategy development, and (4) information feedback. Table 2 details each of these categories, along with their subcategories, examples from the data, and frequency of occurrence. Due to varying length of journal posts, it was possible to have multiple categories and subcategories coded in the same post.

Table 2.

Codebook and frequencies for communication and coordination activities from caregivers’ journal posts

Component of communication and coordination	Subcategory	Definition	Examples from the data	Quotes from the data	Number of posts (% total posts)
Information acquisition Posing a direct question to the care team	–	–	Asking about lost items; Asking about nighttime routines; Asking how care recipient is dealing with new stressors?	“How is he after exercising? Does that help his mood?” (P17, secondary); “Did they fix your A/C?” (P65, secondary)	26 (1.2)
Information sharing Reporting information relevant to care recipient’s or caregiver’s own state, for the purposes of sharing information with the care team, maintaining a record, and observing changes over time	Relaying medical information	Reporting to care team on information delivered by medical professionals (CNAs, doctors, physical therapists, etc.)	Giving update on prognosis from doctor; Sharing details from trip to urgent care	“Dr went well, no return for 6 months” (P68, primary); “Took mom to urgent care… they are treating for bacterial infection and itching” (P22, primary)	96 (4.5)
	Tracking behaviors and symptoms	Tracking care recipient’s behavioral patterns and new or ongoing symptoms	Recording sleep patterns and disruptions; Toileting behaviors	“Continues to be confused on where we are, finding way around the house” (P61, primary); “Went to bed at 8:30 but started to get combative at 7:30” (P74, secondary)	968 (45.7)
	Caregiver emotional state	Reflecting on personal reactions to events or generally expressing emotions	Writing down stressors from the day and what strategies (eg., Taking a walk) helped; expressing guilt over placing care recipient in a home; sharing helplessness over dementia prognosis	“It is hard to be disrespected every single day and just sit here with a smile on my face” (P40, primary); “I just worry that I am going to find out when the time comes that I don’t have all the bases covered for Dad. I feel so inadequate” (P5, primary)	190 (9)
	Care recipient emotional state	Reflecting on care recipient’s expressed emotions	Reporting good/bad mood; noting mood swings at certain times of day	“Bad day, extremely crabby, worried about teeth pull tomorrow” (P50, primary); “He seemed in good spirits yesterday, and very engaged” (P127, secondary)	138 (6.5)
	Recap of day’s activities	Describing daily activities not directly related to dementia care	Describing outings with friends; Commenting on the weather	“Had a nice brunch with friends at the house” (P40, primary); “Did some yard cleanup and checked out things around the house” (P77, secondary)	382 (18)
	Adverse health event	Providing details of a fall, emergency hospital visit, heart attack, or other adverse care event	Giving description of when and where care recipient fell; reporting symptoms before and after fall	“[He] had a fall/collapse at the toilet around 2 am; on the floor since then” (P43, primary); “Dad had a nose bleed this morning. Ambulance was called and he went to emergency care” (P92, primary)	11 (0.5)
	Situational update	Replying to an original post with follow-up information or more details	Giving status update on care recipient in hospice after speaking with nurse; Sharing that medical issue was resolved	“The correct name of this med is Buspirone” (P53, primary); “She did let me change her into her pajamas after the first potty break” (P47, primary)	68 (3.2)
	Care planning	Making note of a change in caregiving strategy or planned change in caregiving strategy	Planning to involve care recipient more in care; Noting that having a house camera was useful	“I am trying to include him and make him feel like he has a role” (P40, primary); “5 mg prescribed for situational anxiety, to see how that goes” (P53, primary)	108 (5.1)
Strategy development Collaborating with care team to make decisions regarding caregiving strategy	–	–	Suggesting activities to improve mood; Giving ideas of rewards to use for taking medications	“It is getting late and I am tired. I can’t get [care recipient] to stay in the bedroom.” (P5, primary) REPLY: “Let me know when you need me to ‘tuck him in’ via the phone. That seems to work well sometimes.” (P17, secondary)	119 (5.6)
Information feedback Replying to an original post and responding directly to member of care team	Affirmation	Replying to post to reassure or encourage original poster	Affirming that caregiver is doing an amazing job; telling caregiver that they are supported by the network	“You are doing an amazing job in an extremely hard situation!” (P42, secondary); “That sounds like a sweet evening!” (P1, secondary)	115 (5.4)
	Empathy	Replying to post to offer compassion or console original poster	Empathizing with difficulty of the situation	“I am so sorry you are dealing with this. I just think it is getting too much” (P39, secondary)	32 (1.5)
	Gratitude	Replying to post to express thankfulness for the original poster	Thanking caregiver for the difference they’ve made with care recipient; thanking caregiver for their patience	“How you connected with him was perfect! Thank you!!!!” (P76, secondary); “Thanks for sharing this helpful information” (P65, secondary)	37 (1.7)
	Information exchange	Responding to a question from another care team member	Letting care team member know that care recipient had a good night of sleep following up that behavioral intervention was successful	“Yes, I will definitely pass along whatever the neurologist has to say” (P3, primary)	29 (1.4)
	Confirmation	Noting personal observations of behaviors mentioned by other care team members	Noticing new physical symptoms previously noticed by another care team member	“Yeah, I noticed he hunches when in the elevator or when we stop walking. Kind of a new thing” (P73, secondary)	3 (0.1)

Information acquisition, the least frequent journal post category, occurred when caregivers used the journal to ask questions of other care network members. Primary and secondary caregivers both asked questions to their network, typically information on events that they missed (eg, How did the bedtime routine go, did the care recipient enjoy seeing family, etc.).

Information sharing, which was the most frequent journal post category, occurred when caregivers posted information. We identified 8 subcategories of information sharing including relaying medical information, tracking behaviors and symptoms, caregiver emotional state, care recipient emotional state, recap of day’s activities, adverse health event, situational update, and care planning. Primary caregivers posted most of information sharing posts (93%) for purposes including sharing notes from doctor visits (relaying medical information), noting the care recipient’s declining cognitive state (tracking behavior and symptoms), and sharing their own daily joys or frustrations (caregiver emotional state).

Caregivers used journal posts to work with other caregivers in co-creating care strategies, categorized as strategy development. Strategy development posts were explicitly collaborative and action oriented. Caregivers often called on specific care network members and offered direct suggestions for care including managing behavioral difficulties, administering medications, or making the care recipient more comfortable.

Caregivers also used the reply feature to respond to posts, which we categorized as information feedback. Subcategories of information feedback included affirmation, empathy, gratitude, information exchange, and confirmation. The most common of these subcategories was affirmation (53% of total information feedback posts), which contained positive messages designed to uplift the other caregiver and reassure them of their worth and abilities.

Post-trial interviews: thematic findings

We identified 4 overarching themes regarding caregivers’ perceptions of the journal’s usefulness as detailed in the following sections. Overall, caregivers noted that the journal supported collaboration and awareness across their care network, heightened their own situational and emotional awareness, and made them feel seen and appreciated. We also found that usefulness was influenced by the current nature of care network relationships.

Theme 1: Usefulness at a network level

Enable collaboration

Caregivers explained that the journal enhanced care network collaboration in multiple ways including shared documentation and asynchronous communication. Maintaining a centralized information store allowed care networks to develop a shared understanding of the care recipient’s progression and the primary caregiver’s daily experiences. The ability to comment on posts enabled network members to make informed care decisions together: “The interactions with my dad … was just so frustrating and so tiring. And so I put it in [CareVirtue], and my sister had recommendations. And we had little dialogue going back and forth about what we could do to make it better… and it did help.” (Participant 19)

Caregivers also noted that the asynchronous communication allowed them to share more candidly and frequently and be more open to suggestions. Caregivers mentioned not wanting to “interrupt” members care network members with synchronous updates (eg, phone calls), and felt more willing to share if they had one central place that care network members could access at their convenience. As one caregiver explained, “One of the things that we both knew we should be more careful about and weren’t is discussing how things were going within earshot of my husband…using this [CareVirtue], then we can communicate a whole lot more without having to say much of anything.” (Participant 43)

Maintain network-level situational awareness

Caregivers used the journal to keep care team members “in the loop,” which facilitated the ability of other members to take on care responsibilities with awareness of the care recipient’s status and history. As one caregiver shared, “the app was really helpful, because before a person came in for their ‘shift time’, they can have advance information about how things were before they came, so they could be kind of prepared…did the client have bowel …did you not sleep all night, or did she, was she agitated last night.” (Participant 35)

Theme 2: Usefulness at an individual level

Track information

Caregivers found the journal useful to record the care recipients’ status and changes over time (eg, symptoms, medications). This allowed them to explore patterns of progression or decline that were not previously observable. One caregiver explained, “it helped me in terms of measuring my spouse’s progression over those two whole months. Because, you know, when you’re with someone every day, 24/7 almost, you may not the differences other people do. So… I found that useful to go back and look at what I wrote, you know, a month ago… it seemed like the progression was increasing and just help me quantify it to some degree in terms of what her capabilities were.” (Participant 61)

Documentation in the journal also supported caregivers in communicating with medical professionals. As one caregiver explained, having written documentation of symptoms and behaviors validated their experience: “Well, I tend to think the doctor doesn’t take me seriously sometimes. So I think the fact that I can say, I wrote it down … this is proof that it happened the same day. So I needed a place that I had a proof, record.” (Participant 2)

Therapeutic benefits

Caregivers described using the journal to identify and explore their own emotions. They found that having a designated space to vent frustrations and examine their own reactions to events were therapeutic: “When journaling, I seem to be able to problem solve a lot better than just talking it through with somebody. I can kind of set it aside, come back to it later, look over what I was writing, and figure more things out than I could normally. When I would use it, my day would go better. Just because I was able to release what I needed to…And I hope that by me getting a better start to my day, led to becoming a better caregiver for that day.” (Participant 31)

Theme 3: Relational benefits

Journal posts were described as providing dual value by making the primary caregiver feel “seen” and appreciated by their support system, as well as making other members of the care network feel involved in the care process. As one caregiver described, “In the journaling, I would put down what kind of day we had, often, you know, what were the good parts and what were the hard parts, and it surprised the kids when they saw that. You know, they didn’t realize the things their mom was going through on a day-by-day basis.” (Participant 47) Posting seemed to create a sense of social support for caregivers: “It made me feel like I wasn’t alone in the darkness.” (Participant 45)

Theme 4: Journal integration

The structure and nature of relationships in a care network influenced journal use, and key barriers to journal integration included care network members’ willingness to be involved as well as primary caregivers’ willingness to share information with the network. One caregiver stated, “I wasn’t confident enough to get my aide… to [use the journal]…I didn’t think she was qualified to handle it.” (Participant 2). Care network members’ preferences for established communication routines and channels also posed a challenge to the network’s use of the journal. One caregiver explained how certain family members did not engaged with the CareVirtue journal and instead preferred the method of communication they had historically used to share caregiving updates: “And out of all of my brothers, probably two, maybe three, would actually hit the notification and go look at it… so just recognizing that some people need that direct text message, because they just don’t use another app, kind of thing” (P21).

DISCUSSION

ADRD care networks used the CareVirtue journal in distinct ways to document and share information, acquire information, co-create strategies, and lend support. These findings suggest that the usefulness of this shared communication and coordination platform spanned individual and care network levels. Combined, these results point to the importance of HIT that supports the care network in collaborative information activities and have implications for the design of such care network HIT.

Network communication and coordination

Previous studies have indicated that a significant aspect of primary caregiver information behavior is the acquisition and sharing of information.^7,14 Our findings confirm the importance of these activities at the primary caregiver level and expand on them to include information sharing and acquisition at the care network level. Further, results from the present study indicate that caregivers used the journal to co-create caregiving strategies and delegate responsibilities across their care network, demonstrating a distinct and novel form of caregiver information behavior. By using a shared platform, caregivers were able to use the information shared by their care network through the journal to suggest new approaches to challenges, iterate on previous suggestions, and offer insights from their own personal experience with the care recipient. In addition to real-time collaboration, caregivers maintained a network-level situational awareness through the ability to keep a shared record of status updates, day-to-day progression markers, and previous care network decisions. The variety of strategies employed by caregivers indicates the importance of supporting asynchronous as well as synchronous information exchange. These results point to the potential benefits of caregiver HIT that supports and engages the care network and its unique needs.

Social support for caregivers

We also found that caregivers used the journal to give and receive social support. Caregivers confirmed that they felt validated and appreciated by replies to their posts, indicating that incorporating features to support care network dialogue may counteract the negative psychological effects of caregiver social isolation.³⁵ Recent literature suggests that a caregiver’s perception of how much social support they receive can be a significant predictor of subjective burden, regardless of the amount of support provided.³⁶ In interview theme 3 (“relational benefits”), caregivers expressed an increased sense of perceived social support facilitated by the journal feature and its ability to give care network members a window into each other’s experiences. The social and emotional benefits of the journal feature are congruent with studies on the protective impact of social support on the chronic stress experienced by caregivers, as well as studies on the benefits of online health communities and information sharing.^37–40 Where these results differ from prior studies is the integration of social support and collaborative work; rather than focusing on the care network as a community or the care network as a team, the journal provides space for the network to operate as both, depending on the caregivers’ needs. Another novel element of the study is the incorporation of observational data and qualitative data to connect actual information activities with caregivers’ experiences. Systematic reviews of HIT interventions for older adults reveal that very few studies considered end-user perceptions of feature effectiveness when evaluating the design of mobile health applications.⁴¹ This study considers both actions and perceptions as the basis for future design recommendations.

Therapeutic support for caregivers

Although this study’s focus was primarily on care network communication and coordination, we found that posting in the journal also provided individual benefits including the ability to process emotions and facilitating awareness of patterns that the caregiver may not have been aware of or was actively avoiding thinking about. Caregivers used the journal to give updates about medical information and care recipient symptoms, but also their own emotional state, sometimes in the same entry. In this way, the journal acted as a therapeutic exercise for many caregivers, allowing them to observe patterns in their own emotional reactions and troubleshoot by working through their thoughts in real time. This unexpected benefit of the journal feature demonstrates the value of HIT that gives caregivers a space to work within their role as a caregiver while expressing themselves as an individual.

Challenges to journal integration

The benefits of a shared communication and coordination platform require that all members of the care network engage with the HIT. Our results suggest that the care network structure and nature of network relationships influenced whether the journal feature was successfully integrated into care network communication. This finding is reflective of previous studies, which indicate that care networks vary widely in terms of the number of members, demographic composition, and relationships to the primary caregiver and care recipient.^7,11 Future research should consider how communication activities vary across different care network types. For example, mapping the size and structure of different care networks and how the frequency and type of information behavior changes during their use of HIT platforms can provide deeper insight into the design features needed for different types of networks.

Implications for technology design

Our results can be used to inform the design of future care network technologies in several ways (Table 3). The categories of communication and coordination activities provide a framework for identifying actions that require support in HIT platforms, while interview themes offered additional insight into how caregivers found the existing features useful or in need of adaptation or expansion. For example, our results indicate that beyond information sharing and acquisition, caregivers used the journal to collaborate with their care network—developing strategies and receiving feedback. HIT platforms should support these activities by designing for greater shared situational awareness: caregivers’ moods can be tracked through the use of emoticons and key words, allowing for tailored notifications that prompt care network members to offer encouragement or suggestions when needed. In addition, offering the capability to catalogue past care strategies as well as rate them or make notes on their effectiveness can facilitate collaborative decision-making across the network.

Table 3.

Design implications for future care network technologies

Component of collaboration and coordination	Caregiver needs	Design implications
Information acquisition	Ability to easily find information about care recipient’s symptoms and behavior Accessible documentation to refer to when it’s needed	Robust search functions for caregivers to find posts and replies based on keywords, time and date posted, emoticon, or care network member Option to create personalized tags for journal posts to track specific topics
Information sharing	Ability to share updates with the entire care network Safe space for individual journaling and processing emotions	Ability to link PDFs, images or other documents to journal posts to provide additional detail and context Expanded privacy controls for caregivers to specify which care network members can see certain posts, comment, or react
Strategy development	Support for collaborative decision-making	Printable summaries of care network decisions made over time to create body of shared knowledge and support future decisions Smart tracking features that allow care network members to rate or give feedback on strategies and suggest improvements
Information feedback	Ability to communicate directly with network members Options for offering different types of social support tailored to caregiver preferences	Regularly updated journal prompts that encourage caregivers to self-reflect on their experience as well as ask their network questions about their social and emotional health Sentiment analysis and related machine learning methods that can track caregiver emotional state indicators (eg, emoticons) to predict burnout and encourage members of the care network to reach out Notification settings that alert caregivers to a new post or remind them to interact with the journal feed after a certain period of inactivity

Under the theme of “usefulness at a network level,” caregivers emphasized that a unique requirement of working within a care network is the ability to communicate asynchronously, due to caregivers not sharing the same physical space or work hours. Technologies can support asynchronous information exchange by providing customized summaries and printable reports not only of the care recipient’s symptoms and behaviors over time, but also decisions and strategies that have been employed by the care network. In this way, care network members will be able to access information and maintain situational awareness when and where it is convenient for them, reducing the need for additional communication activities.

Further, the “usefulness at an individual level” theme highlighted that the benefits of information sharing did not always require communication with the entire care network but could include the act of sharing as writing down emotions to process them individually or with a smaller group of people. Thus, future designs should include space for caregivers to keep a personal journal and a space to publicly share information within the same application, harnessing the therapeutic as well as relational benefits of the tool. To ensure privacy between individual and shared entries, privacy controls that seamlessly facilitate either privacy or sharing should be included.

Limitations

Certain limitations of this study should be considered. First, although our study was diverse in terms of educational attainment and rurality, the study was conducted with majority white participants. Future work should include a diverse range of perspectives to ensure that care network interventions are effective for a broad base of users. Additionally, to participate in the study, participants needed consistent access to the internet and to have family and friends who were willing to participate. When calculating frequencies of codes for the journal entries, each entry was treated as an independent excerpt, rather than aggregated by participant, which may have given more weight to the activity of participants who wrote a higher number of posts. Some caregivers reported that the platform’s restriction to the English language posed a barrier for Spanish-speaking members of their care network. Hispanic adults are already less likely to use or trust online in health information resources, and a next step is to make CareVirtue available in Spanish.⁴² Finally, interviews were conducted with primary caregivers only, so future studies should explore the perspectives of other members in the care network to get a holistic view of the network’s needs and processes.

CONCLUSION

To inform the design of future care network HIT, we explored the use of a shared communication and coordination platform for ADRD care networks. We found that caregivers documented and shared information, and also used that information to collaborative create strategies and provide and receive support. These activities provided individual, network-level, and relational benefits. Where previous reviews on HIT interventions for caregivers focused on the benefits of psychoeducational programs and peer support, our results point to the supportive nature of a platform designed for direct collaboration across the members of a care network. Our study suggests that a shared journal feature in a web-based application such as CareVirtue may be able to address key caregiver outcomes, such as isolation and burden, by supporting network level communication and coordination. This work provides a foundation for the design of effective and sustainable interventions that consider caregivers’ needs as an individual and member of a collaborative care network.

FUNDING

This work was supported by grant number R41AG069607 from the NIH National Institute on Aging and by the Wisconsin Alzheimer’s Disease Research Center P30AG062715 (PI Asthana) through funding from the NIH National Institute on Aging. The research presented was also supported under NIH award TL1TR002375. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

AUTHOR CONTRIBUTIONS

NEW, CE, MZ, AL, and PL conceptualized the study and led data collection efforts. AL, NEW, AJ, and DG led data analysis efforts with support from CE and MZ. AL and NEW led writing efforts with contributions from everyone else. All authors critically reviewed and approved the final manuscript.

SUPPLEMENTARY MATERIAL

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

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.