A Visual Approach to Facilitating Conversations about Supportive Care Options in the Context of Cognitive Impairment

Abstract

Background:

Persons with cognitive impairment may experience difficulties with language and cognition that interfere with their ability to communicate about health-related decision making.

Objective:

We developed a visual elicitation technique to facilitate conversations about preferences concerning potential future supportive care needs and explored the utility of this technique in a qualitative interview study.

Methods:

We conducted 15 online interviews with persons with mild cognitive impairment and mild to moderate dementia, using storytelling and a virtual tool designed to facilitate discussion about preferences for supportive care. Interviews were transcribed verbatim and analyzed using an inductive qualitative data analysis method. We report our findings with respect to several main themes. First, we considered participants’ perspectives on supportive care. Next, we examined the utility of the tool for engaging participants in conversation through two themes: cognitive and communicative processes exhibited by participants; and dialogic interactions between the interviewer and the participant.

Results:

With respect to participants’ perspectives on supportive care, common themes included considerations relating to informal caregivers such as availability and burden, and the quality of care options such as paid caregivers. Other themes, such as the importance of making decisions as a family, considerations related to facing these challenges on one’s own, and the fluid nature of decision making, also emerged. Common communicative processes included not being responsive to the question and unclear responses. Common cognitive processes included uncertainty and introspection, or self-awareness, of their cognitive abilities. Last, we examined dialogic interactions between the participant and the interviewer to better understand engagement with the tool. The interviewer was active in using the visualization tool to facilitate the conversation, and participants engaged with the interface to varying degrees. Some participants expressed greater agency and involvement through suggesting images, elaborating on their or the interviewer’s comments, and suggesting icon labels.

Conclusion:

This article presents a visual method to engage older adults with cognitive impairment in active dialogue about complex decisions. Though designed for a research setting, the diverse communication and participant-interviewer interaction patterns observed in this study suggest that the tool might be adapted for use in clinical or community settings.

1. Introduction

Difficulties in speech and language are common in dementia. Early difficulties include word finding; as the disease progresses, persons with dementia can experience impairments in communicating with respect to comprehension, speech fluency, word production, and syntax [1].

Though there is extant research on how to engage participants with cognitive impairment in decisions about everyday life activities [2], less is known concerning the extent that they are able to participate in weightier decisions such as the need for more supportive care and future care planning.

Decisions regarding higher levels of supportive care may include asking for additional help from informal caregivers, hiring formal caregivers or other types of in-home household help (e.g., use of cleaning or meal services, medical transport), engagement in adult day programs, or seeking an appropriate care facility. Shared or supported decision making is preferred by persons living with dementia and their families, but does not always occur [3]. For persons with moderate or severe stage dementia, family caregivers are more likely to make the final decisions without soliciting input from the person with cognitive impairment [4].

Persons with cognitive impairment may have the desire to stay involved in decision making about their own care, but this desire can be overlooked by others around them, such as care staff and care partners [5]. Our prior interviews with older adults with dementia and memory loss, caregivers, and professionals that work with individuals with dementia, indicated that difficulties imagining future decline can hinder efforts to involve a person with dementia in shared decision making and planning [6].

Extant research has argued it is important to provide ‘subtle’ support in decision making and assisting people to make their own decisions [7]. In addition, assessing decision-specific competence, clarifying values, and understanding the significance of relationships and context increase engagement in decision making [8].

Various techniques have been found to facilitate recall in individuals with cognitive impairment, including organizational supports (e.g., calendars), external memory aids, note-taking supports, scripts, environmental adaptations, and caregiver training [9]. Visual supports such as pictograms or photos have also been used to facilitate communication and decision making [10].

In conversation, personalized aids comprised of pictures and text can facilitate on-topic statements, diminish nonproductive utterances, lengthen conversational turns, and/or increase the frequency of turns taken by persons with dementia [11]. Some research has employed icons that are moved around on a mat to facilitate discussions about diverse topics [2,12,13]. Tablets, social robots and computer systems can also facilitate conversation with the following potential benefits: 1) breaking the ice; 2) increasing interaction; 3) improving understanding of the person with dementia; and 4) reducing pressure for the conversation partner [14,15]. Within the context of decision making, Chang and Bourgeois observed that visual aids improved the decision making abilities of participants with mild and moderate stage dementia to comprehend medical information, employ supportive reasoning, and relate this information to his or her own situation [16].

Despite efforts to engage persons with cognitive impairment in communication and decision making, challenges remain in how to do so, particularly with respect to facilitating discussions involving complex, health-related topics [10]. Efforts exist to engage persons with cognitive impairment in arts-based and participatory activities [17], but traditional participatory techniques may assume a high level of abstraction or comprehension abilities, both of which can be challenging for persons with cognitive impairment [18]. Barriers may also arise when remote engagement is required, such as in the recent circumstances of the COVID-19 pandemic. Potential challenges in online interviews with older adults include less familiarity with technology and decreased richness in visual cues [19].

The visual elicitation technique developed and employed in this interview study was part of a stakeholder-engaged approach to create a tool for facilitating decision making in the context of cognitive impairment through the Decision-Making in Alzheimer’s Disease (DMAR) project (https://depts.washington.edu/hprc/projects/dmar/) at the University of Washington [6,20]. Due to the COVID-19 pandemic, we adapted what might traditionally have been conducted in-person to an online setting. With the aid of an online visual tool, interviewer and interviewee engaged in a conversation about how participants might make decisions to meet emerging demands regarding future care. In this subsequent qualitative analysis, first we explore participants’ perspectives on supportive care; and then, to better understand the utility of the tool, how the tool facilitates characterization of cognitive and communicative processes and interactions between participant and interviewer.

2. Methods

2.1. Study Design

We conducted a qualitative interview study using an online visual tool to explore how participants make decisions about transitions in care, such as bringing caregivers into the home or moving to an assisted living facility, due to changes in health or functional status. We then characterized cognitive and communicative processes and participant-interviewer interactions, to better understand and provide insight concerning implementation of this visual elicitation tool in online interviews.

2.2. Data Collection

We conducted semi-structured interviews via a HIPAA compliant version of Zoom. Participants were recruited from the University of Washington Alzheimer’s Disease Research Center and local organizations serving older adults. Potential participants were screened by phone. Inclusion criteria included age 65 years or older, fluency in written and spoken English, ability to respond verbally to questions (Dementia Severity Rating Scale [DSRS] Speech and Language Score 1–3) [21] and diagnosis of mild cognitive impairment (MCI) or mild to moderate stage dementia. Scores 1–3 on the DSRS Speech and Language Score correspond to: 1) Sometimes cannot find a word, but able to carry on conversations; 2) Often forgets words. May use the wrong word in its place. Some trouble expressing thoughts and giving answers; 3) Usually answers questions using sentences but rarely starts a conversation [21].

After confirming eligibility, study staff met with participants via Zoom to obtain verbal consent and collect baseline measures, including demographic characteristics, the Quick Dementia Rating System (QDRS)[22] to determine dementia severity, and the Care Values Scale [23]. A second Zoom meeting was scheduled for the qualitative interview.

As our study focused on the perspectives of persons with cognitive impairment, the involvement of caregivers was limited. The caregivers helped answer the pre-interview questionnaires such as the demographics if the participant was having trouble) and the QDRS. We informed caregivers that the interviews were intended to focus on the participant, but there were times when caregivers may have participated in the interview itself.

Each interview was intended to discuss decision making in the context of three scenarios: 1) past decision; 2) hypothetical scenario #1; and 3) hypothetical scenario #2, ordered based on decreasing cognitive abilities and functional independence, and increasing needs for supportive care. Our interview guide involved probing of different facets of the scenarios to facilitate recall, an approach akin to the use of cognitive interviews to facilitate recall [24–26].

The visual tool (hereafter referred to as “Board”) that we used employed an online “canvas” upon which the interviewer could drag icons or type text to represent concepts. A sample Board for the second hypothetical scenario is depicted in Figure 1. During the interview, the interviewer shared their screen and performed actions on the Board. For example, as an icebreaker, they began by asking participants to tell them about important persons in their lives. Then they would open the appropriate category on the right side and move icons as appropriate to the Board. For example, if a participant mentioned an adult child, the interviewer would move one of the appropriate icon and label it with the name of the child. The interviewer sought to personalize each Board to suit the participants by revising the icon text to mirror the words used by participants.

Figure 1.

Sample Board.

There were six different categories of icons: People, Places, Safety, Emotions, Activities, and Issues. The icon sets and categories were intended to represent concepts pertinent to decision making about care and was informed by our prior work [6]. We developed the Board setup in the online tool, diagrams.net, after a survey of existing tools [10]. Key goals in tool development included a focus on simplicity, minimization of cognitive load, and using concepts and images that would resonate with participants. The icons used in the study were either selected from icon repositories or developed iteratively by a multidisciplinary team and a graphic designer (WO). The interview guide and icons are provided in the Supplementary Material.

During the interview session, the interviewer sought to balance the amount of information displayed to maintain simplicity and avoid cluttering the Board. Some concepts were deemed to be represented as text rather than as a graphic (e.g., “wasn’t a decision… more of a natural, fluid change”). Rather than depicting everything that was said by the participant, the interviewer focused on capturing concepts that seemed most important to the participant. To make the discussion points more salient and personalized, the participant’s own phrases were used to accentuate and flesh out these concepts.

2.3. Interview Setup

In this study, the intention was to explore participants’ preferences for additional supportive care through a collaborative exploration of three scenarios, one involving a past change that they had made to their life routines, and two hypothetical future scenarios. Scenarios were developed based on findings from earlier in-depth interviews [6]. In the past change scenario, the interviewer asked participants to share an experience that they may have made in after they began experiencing memory loss or changes in cognition (for the full prompt, please see the supplemental file). Examples of past scenarios included having their caregiver help with managing medications, organizing their appointments, and overseeing finances (Figure 2). As participants told their story, the interviewer would ask clarifying questions as needed, and move icons to the Board to illustrate salient aspects of the story.

Figure 2.

Past change example (P04).

The interviewer next asked participants to consider two hypothetical scenarios of differing needs for support. In the first scenario, participants were generally asked what additional support options they might consider if, in the future, they had problems with cooking and driving. If one of the problems did not apply (e.g., participants had already given up driving), the interviewer would substitute something that did apply. The second hypothetical scenario was around increased need for support, where the individual had trouble with eating, using the bathroom, and getting dressed in the morning (Figure 3). The interviewer often changed the original icon labels to facilitate greater personalization of the Board to the participants’ life situations.

Figure 3.

Example of Hypothetical Scenario #2 (P06).

All interviews included a scenario in which participants and interviewers explored a past decision for change in supportive care. About half of the interviews are missing at least one of the hypothetical scenarios (n=8). There were various reasons hypothetical scenarios were omitted. For example, a scenario may not have been relevant to a participant, or a scenario may have been skipped in the interest of time. The interviewer made these decisions as appropriate based on the pace of the interview and content covered up to that point.

2.4. Data Analysis

Recorded interviews were transcribed verbatim and analyzed in Dedoose version 9.0.107, a qualitative data analysis software package [27]. As there was related and/or prior work relating to some themes of interest but not all, we analyzed the data using a hybrid inductive-deductive method, which seeks to leverage the strengths of each respective approach [28]. We drew upon related work as appropriate to develop an initial coding scheme, but complemented this with inductive qualitative data analysis method to revise and refine of the category system [29].

The focus of our analysis was on three main themes of interest: 1) participants’ perspectives on supportive care; 2) cognitive and communicative processes; and 3) interactions between the participant and the interviewer. With respect to (1) participants’ perspectives on supportive care, we drew from our own prior work on decision making in the context of cognitive impairment, which resulted codes concerning decision making. We added codes as needed and ensured that these were applied to the entire dataset through an iterative process [29].

With respect to (2), the systematic coding of utterances between participant and interviewer is a useful method for studying interview processes [30]. We adapted the method here to better understand the utility of visual elicitation techniques in facilitating conversation. We sought to code cognitive processes that are commonly examined in the assessment of decision making capacity. In clinical contexts, these capacities are: understanding the decision to be made, appreciating how the decision relates to the self, comparing the benefits and risks of options by integrating and analyzing information rationally, and communicating a choice [16,31]. Consistent with an inductive approach, we reviewed the data and adapted the capacities that we coded to suit the data. As a result, we focused on four capacities: appreciating the situation and its consequences, communicating an option, reasoning about options, and communicating a choice. Each cognitive process was coded in terms of whether participants were able to or had difficulty performing the process.

As communicating about decisions involves cognitive and communicative abilities in addition to decision making capacity, we also developed a coding scheme to identify these other processes to afford a more holistic view of the data. In doing so, we incorporated communicative acts such as requests for clarification, which have been used in prior research with persons with cognitive impairment [32], and other cognitive processes, such as self-awareness/introspection and distancing oneself from negative experiences, which has been recognized as an adaptive coping mechanism [33,34]. We include a list of codes with descriptions and examples in the Supplementary Material.

Each transcript was coded independently by two coders, and differences were resolved through discussion. We also employed memos to record our thoughts and questions about codes, facilitate communication between analysts, and serve as a record of our analysis process [35].

3. Results

3.1. Sample

The sample (N=15) was largely white (n=13) and male (n=9) (Table 1). Participants had varying degrees of education, although all had completed high school; there was a wide range of incomes. Most participants lived in their own private residences (n=9) and with a partner (n=9), but there were other residential situations as well, including living in a retirement community (n=6) and living alone (n=3). Participants rated their own health as being good to excellent (Table 2). The Quick Dementia Rating System (QDRS) was used to assess dementia severity (normal, mild cognitive impairment, mild dementia, moderate dementia, and severe dementia) [22]. Most participants experienced mild dementia (n=11) as identified by QDRS score.

Table 1.

Demographic characteristics

Demographic Characteristics	Statistic
	M/SD
Age	77.8 ± 7.1
	n (%)
Gender
Male	9 (60)
Female	6 (40)
Race
White	13 (86)
Black/African American	1 (7)
More than one race	1 (7)
Ethnicity
Not Hispanic/Latinx	15 (100)
Education
Some college but no degree	4 (27)
Associate degree or technical degree	1 (7)
Bachelor’s degree	7 (46)
Graduate degree	3 (20)
Income
$0 – 24,999	2 (13)
$25,000 – 49,999	2 (13)
$50,000 – 74,999	3 (20)
$75,000 – 99,999	3 (20)
$100,000 – 124,999	1 (7)
$125,000 and up	2 (13)
Living situation
Own private residence	9 (60)
Retirement community	6 (40)
Shared living space
Spouse/partner	9 (60)
Family member	2 (13)
Multiple people	1 (7)
No one	3 (20)

Table 2.

Health-related characteristics

Health Status	n (%)
Good	8 (53)
Very good	5 (33)
Excellent	2 (13)
QDRS Category
Normal	1 (7)
MCI	2 (13)
Mild	11 (73)
Moderate	1 (7)

3.2. Participants’ Perspectives and Considerations Regarding Supportive Care Options

There were three main additional support options that participants mentioned: receiving additional support for instrumental activities of daily living, incorporating or increasing in-home supportive care, and moving to a more supportive environment. Participants mentioned different considerations that might affect their decisions. Factors that were brought up by approximately a third of the participants, equally across the three scenarios, were those relating to the caregiver (burden, availability, suitability), cost, and distance/proximity to family. These participants recognized that with disease progression, an increased burden of care would fall on their caregivers, and that this burden may decrease their caregiver’s quality of life.

P02: I don’t want to be a burden to anyone. I don’t want anybody of my family to have to deal with me seriously for a month or a year or something like that. I want them to go on and live their better life… They could possibly have the same thing down the line.

A third of the participants also brought up past experiences or knowledge that influenced their decisions, such as seeing their parents’ experiences of living in communities, or what they heard from others. P12 was anchored on negative perceptions about the quality of nursing home care from their own career in the medical field, and also referenced secondhand experiences from their daughter’s job at a nursing home:

P12: Well, the problem is being in the medical field, I know what’s going on in nursing homes and with the problem of caregivers. And I wouldn’t want to be there really because I know all about nursing homes. And some of them are quality, and all of them are pressed for help…. A lot of them are really understaffed…. my daughter works in a skilled nursing facility… A lot of these places are hurting, as far as help, finances. Anyway, I don’t need to go there.

Quality of care came up often with respect to hypothetical scenario #2, which indicated greater need for support. Individuals were asked about hypothetical decisions that would be made if they experienced trouble with eating, using the bathroom, and getting dressed in the morning. Participants described different concerns and preferences that they might have with respect to the support offered in communities, in care facilities, and by caregivers:

P01: You take somebody who’s 80 years old, been doing something one way for 80 years, expect them to suddenly like brussel sprouts if they’ve never ate them before in their life… That’s just an illustration, but you have to meet them where they’re at. Not trying to change things for them. Say we’re going to make you healthy so you’re going to eat tofu every day. They have to be considerate to what the people want not force a lifestyle or something.

Some themes were prominent in only a few interviews. For example, in many of the interviews, participants said that others such as family members that lived with them help them with scheduling or remind participants of their appointments (n=6). In contrast, P10 lived alone; he made use of various organizational tools such as printed calendars and medicine boxes to help with scheduling and medication management.

A few participants also spoke of family expectations and the fundamental meaning of family (n=3):

Interviewer: I’m curious, because it sounds like there are expectations that your daughters will help you in the future. Are they aware of those expectations?

P03: They damned well better…

Interviewer: Have you already talked about this with your daughters?

P03: No, not as of this time, but it’s just sort of, that’s the way we go. That’s what we do. It’s not something one-off time… you do that all the time if you’re doing it right…

Another participant responded as follows:

P09: Like I said, it would be a group issue. I think the thing that you’re not really looking at is the fact that if there’s a need within the greater family of any member of the family, even when you’re mad at them, but we would have a strong sense that they would be … If they have need that’s part of what it is to be a part of our family.”

P03 expressed a strong conviction of the expectation of assistance from the family, and P09, the need to pull together and make decisions as a group, with some expressed frustration due to the focus on individuals in the study.

There was also the sentiment that one was less concerned about receiving help, but rather spending time together:

P02: …as long as I have [Caregiver], I probably don’t need somebody else. But it would be nice for the others to come and visit now and then, not to help me, but just to remember that we are family.

In considering these themes, it is perhaps important to weigh the significance of non-tangible support in addition to more concrete forms of assistance, in the process of decision making.

3.3. Cognitive and Communicative Processes

3.3.1. Decision Making Processes

We characterized cognitive processes concerning decision making through four themes: appreciating the situation and its consequences, communicating an option, reasoning about options, and communicating a choice. Overall, participants were able to appreciate the situation and its consequences:

P07: Well, if it’s very serious, I’m going to have to stop driving and I know that’ll happen at some point in the future. That makes me angry just because I used to do a lot of things and those things are getting closer and closer and closer… I do depend now on [Caregiver] doing most of things that I used to do. It makes me sad.

However, there were also situations when participants did not fully appreciate the gravity of a situation, such as this one involving taking an incorrect dose of medication:

P04: It wasn’t really a medical emergency. Well, I guess it really was because I was on the floor and I couldn’t get up.

There were times when participants had difficulty articulating options:

Interviewer: So, we thought about one option to help with the problem of not being able to eat, use the bathroom or get dressed by yourself. Is there another option that you would consider if you were to need help with those things?

P01: Need help with those things? I don’t know what an option would be.

All participants were able to suggest options to some extent, though about a third had difficulty either communicating or reasoning about options, or both. There were times when participants drew upon their prior or current experiences to present options. For example, P11, who lived in a retirement community, was able to easily pose resources for supportive care that existed within her current living situation (e.g., increasing hours of paid caregiving, moving into more supportive care). Prior experience also appeared to play a role in participants’ perspectives towards their openness to living in a retirement community, and also in communicating a choice. Some participants who lived in retirement communities had visited them previously, knew of person(s) who lived in them, or had known someone who lived in them previously, so were open to the idea of moving into them (n=5):

P08: … both of my parents who are long gone now lived in retirement communities and [Caregiver]’s one parent did when we were married, so we were familiar with them many years ago. But this one… we knew people here, it was close to where we’ve been living for many years, just made sense.

On the other hand, another participant had formed a negative impression of communities based on what she had observed with her mother:

P02: I never wanted to go there. I just watched my mother just fading away. And I don’t want that to happen to me.

3.3.2. Other Cognitive and Communicative Processes

We also performed an inductive analysis of cognitive and communicative processes exhibited by participants. There were three main categories of cognitive and communicative processes: requests, response quality and structure, and cognition and attitude. Requests for clarification of what was being asked was the most common process, and performed by almost every participant (Figure 4).

Figure 4.

Cognitive and Communicative Processes.

In terms of response quality and structure, potential barriers to effective communication included utterances that were not directly responsive to questions (n=9), unclear responses (n=6), or words that did not appear to fit the statement being made (n=5):

P03: Make sure that you get to your people, the people that you like about. And I see a lot of people who have problems with their cognizance in their brains, is they really don’t do very much other than just eat bad food and they don’t get enough sleep.

With respect to cognition and attitude, one common behavior was being uncertain when asked a question (n=9):

Interviewer: Okay. And then on the flip side, what do you think would work well? Or what do you think would be good about hiring a housekeeper?

P04: I guess it would… I don’t know.

Some participants had trouble imagining themselves in the future state posed as a hypothetical scenario (n=7):

Interviewer: Some people, when they think about getting more care, they might consider moving somewhere else in order to get that kind of support, or they might consider bringing in a caregiver into their home… or they might have someone else move into their home or something like that. Are any of those options something that you would consider if you were to need help with eating, toileting, and dressing in the future?

P02: Well, I don’t have problems with any of those right now…

Most participants also demonstrated moments of introspection, or self-awareness of their cognitive abilities (n=9):

P12: But partly is that I just can’t make decisions. I can’t focus on things that… I’m still doing the finances, but having a little bit more trouble with that… at some point, I’m going to have to have help with finances.

P13: I don’t feel safe driving. And then just trying to remember where to go, what to do. And people drive very quick, and I don’t think my mind can work that fast anymore.

Many participants demonstrated introspection or awareness of one’s abilities and limitations (n=10). These statements might also include affective and social concerns:

P15: I’m trying to think, “Well, what would I normally say?” Or something like that. But I just feel… I’ve always had real real bad anxiety. That’s actually improved, but still, I think that’s a major source of my anxiety is concern about relationships and not spoiling them. I’ve got a real dim view of my social aptitude and that was something that was long before the memory loss kicked in, but it does make it more of a concern.

We observed that there was diversity in participants’ reasonings and attitudes, reflecting the role of individuality in participants’ experiences, perspectives, and ways of dealing with life. In this example, the participant seems to be able to clearly articulate what might occur if the hypothetical situation came to pass, and he explains his attitude using his personal tendency to take things as they come:

Interviewer: If you were not able to do the cooking or the housekeeping, and if [Caregiver] also was not available to help you with these things…

P05: We would probably have a healthcare lady or whatever they are that supports old people, I mean, that come in and clean your house or cook your dinners or whatever… I would deal with it when it comes around… I don’t sit around and plan a lot, what am I going to do in this situation… when it comes up to that point, I will decide what to do.

3.4. Participant-Interviewer Interactions

We now explore whether the visual technique was effective for facilitating conversation and decision making by examining interactions between participants and interviewers (Figure 5). Actions that the interviewer often performed included explaining the visual setup and talking/confirming with the participant about text or icons that they were adding in response to participants’ testimonies. The interviewer also summarized the current Board for participants and then asked if there was anything missing. Other actions included asking participants if there was anything in a given image container, e.g., “Activities” that fit their situation, and changing icon labels to personalize the Board.

Figure 5.

Participant-Interviewer Interactions

Participants confirmed the state of the interface more than any other action, but they also referred to the images (either by referring to their usage or incorporating the icons in their answers) often, suggesting that the icons and the canvas were helpful in facilitating discussion. For example, in this exchange, the interviewer shows the participant some images, which then facilitates discussion about a topic that may not have come up otherwise:

Interviewer: I just opened up a box that’s labeled, “Activities.” These are different tasks or activities that some people might say they’ve made changes in or had to make decisions about. I think what might be relevant to what you described is this image of working.

P13: Using technology, I’ve always had a hard time, because we didn’t even have electric typewriters when I graduated.

We also observed other forms of engagement, such as selecting an image (n=8), elaborating with more detail (n=6) and making corrections (n=5). After seeing the images, a participant was able to select an image representing their spouse as a source of support:

Interviewer: …in these different set of images here, do any of these look like people who have supported you because of your memory loss?

P01: Well, of course the spouse would be the… main one…

After the interviewer reviewed the current board representing Hypothetical #1, a participant elaborated on his responses:

Interviewer: So, is there anything that you think is missing from these… from this board?

P05: No, not really. I mean, I think it’s done a lot of good talking here… It shows me how my life projects. It will go the road to whatever. And so far, it looks like it’s [going] pretty well. I got an idea how I want to live the rest of my life, but we’ll see what… It’s just throwing up suddenly or whatever. I mean, you cannot make definite plans. It just doesn’t work that way.

During review of a board representing a past change of ceasing driving, a participant requested to make a correction:

Interviewer: Another big factor, you said the biggest one was that the insurance rates had gone up. So you did not want to keep paying for such a big expense.

P06: Okay. I would back off on that statement because… The insurance rates were a factor in my decision-making… not the primary factor… The business of getting lost and et cetera, et cetera. It’s still going to make the needle move when you put it up on the counter to weigh it.

Interviewer: Okay. So it’s a factor, but not the biggest one?

P06: Yeah.

These examples illustrate that the Board enabled participants to elaborate on their stories, as well as delve a bit deeper in thinking through a topic.

Some participants exhibited little or no engagement with the Board, while others engaged with it more actively. One participant was particularly engaged during the interview, both in terms of reading off parts of the interface, as well as confirming what the interviewer said.

Interviewer: Oh, nice. So I’m just going to open up this box that says activities.

P08: Okay. Activities there. I see at least nine, but it keeps going up, okay.

Interviewer: Yeah.

P08: Can we start with driving?

P08: Well, it’s a peculiar thing you’ve got, because you’ve got six items… excuse me, nine items, and one of them has an X through it. I have no idea privacy, why that’s there. Does that mean something?

Interviewer: Well-

P08: No other item has an X through it. Why does that have an X?

As indicated in Figure 5, in all interviews there were a few technical issues related to manipulating the visual display. At times, the screen would freeze or the interviewer would have trouble manipulating the icons, which prevented the interviewer from concentrating completely on the participants’ stories, as well as the potential need to re-orient participants after interruptions.

4. Discussion

The findings from this study enhance our knowledge of technology use and visual elicitation techniques to support complex decision making in the context of cognitive impairment. Although there is no one-size-fits-all solution for how to facilitate dialogue in the context of cognitive impairment, our visual elicitation technique can be helpful for the important and sometimes sensitive issue of articulating preferences for additional supportive care needs. Specifically, the visual tool helped participants focus on the topic of interest, decision making about supportive care, and think of relevant concepts.

Prior work has used a visual tool employing icons and text labels, called Talking Mats, to enhance communication about everyday topics (Personal Care; Getting Around; Housework; Activities) [2], Voice My Choice employed iconic pictures from Google Images relating to food, activities, daily living, socializing/communication, and pain, to facilitate a preference sorting task [36], the SHARE dyadic care planning intervention [13] has also used to enhance communication in the early stages of dementia. However, these tools have less dynamic flexibility to be adjusted based on participant input. Other concerns, such as the complexity of images and the difficulties of communication about abstract concepts, have also been raised as issues with visual materials [37]. Our study expanded upon this earlier work by incorporating simple icons that could further be customized by the interviewer to reflect participants’ perspectives. Similar to prior work, the interviewer in our study adopted an empathic and curious stance, giving the participant the time and space to articulate their thoughts [38]. Supporting interviewers through training may also enhance their ability to customize icons and labels in realtime, thereby improving participant-interviewer interactions and increasing interviewer effectiveness within our flexible visual elicitation technique.

However, there were also challenges. As mentioned in the Results section, there were some technical issues. Prior work has also reported that the use of technology such as tablets can often be accompanied by technical issues, which in turn can become the subject of substantial conversation [39]. However, in section 3.5, we observed that both interviewer and participant engaged actively in dialogue, with participants commenting and elaborating on the Board’s state. Some participants also demonstrated additional agency and critical thinking in the form of making suggestions for icon labels, correcting, or elaborating on what had been presented on the Boards. Nevertheless, in future studies, it would be important to further consider platform characteristics such as responsiveness and speed, as well as connectivity and device characteristics, all of which may affect participant and interviewer experiences.

In addition, another potential concern would be that a participant focuses on pre-defined concepts in the visual setup, when instead it might be more appropriate for them to generate novel or more personalized responses. In this study, we also observed situations of discordance, in which a few participants expressed frustration with the emphasis on individual preferences. Some of the responses illustrated that it can be difficult to strike a balance between encouraging individual preferences, and also respecting the desire of the individual to make decisions as a family. In this study, our goal was to elicit the decision making preferences of the person with cognitive impairment in order to reduce the tendency for them to be excluded from decision making [5]. In addition, we asked participants to describe discrete examples of decision making, but at times were provided with stories that did not appear to fit this model. For example, past changes were sometimes depicted as more fluid, as a continual process of identifying and executing ways to adapt life to accommodate memory loss.

We observed substantial variability in terms of participants’ cognitive and communicative styles; providing flexibility in conversational support, both in terms of the design of the visual setup and interview approach, can facilitate smoother interactions that also enable participants to express their own voices and perspectives. Important considerations are to incorporate flexibility for participants to express their own attitudes, bring in life experience, and discuss family dynamics, which can play an important role in decision making. A previous concept analysis of ‘meaningful activity’ in the context of dementia has noted the importance of personally relevant goals and identity [40]; doing so can perhaps also result in a more enjoyable, engaging, and meaningful experience for the participant. More generally, user-centered design research conducted on developing a tool to facilitate decision making about medications among older adults also concluded that flexibility and personalization were important design considerations [41].

Another area for future study relates to ways of presenting information to facilitate decision making. In this study, it was clear that previous experience affected participants’ attitudes towards the options for additional support. Participants who exhibited a positive attitude toward assisted living communities tended to have had prior positive exposure through people they knew. In addition, when asked to problem solve concerning a hypothetical scenario, participants tended to draw upon firsthand or secondhand life experience. Thus, one approach to bolster decision making about additional support options is to solicit or present stories of supportive care to facilitate re-orienting of pre-existing negative impressions of supportive care.

4.1. Limitations

This study sought to examine the potential to employ a visual interview elicitation technique to support conversations about increased supportive care options in the context of cognitive impairment. Despite potential benefits of the visual elicitation technique in terms of encouraging participants to express their perspectives, there are also limitations, such as the effort required to train interviewers. Interviewers would need instruction in the potential benefits of elicitation techniques and how to perform them and engage in practice of the technique, prior to engaging with participants.

In addition, given its exploratory nature, the sample was limited in terms of the number of participants, and there is a need to explore the utility of the technique with a larger and more heterogenous population. In addition, it may be useful to explore the potential for expanding the experimental techniques employed in this study in longer-term research to foster dialogue along the trajectory of cognitive decline. Previous research has argued that interventions including novelty could enhance broad cognitive abilities, by utilizing highly variable tasks which overlap in targeted processing demands [42]. Though employed in a different context, the hypothetical scenarios in this study did afford participants an opportunity to engage in thinking through a problem on a topic that they might not have considered extensively. An alternate approach might be to engage participants in conversation using this tool more regularly but on different topics and with different tasks, to see if they might respond more readily to conversations on cognitively and emotionally challenging and/or complex topics. Last, as most individuals in this study experienced mild dementia, it is unclear whether this technique might be employed with individuals experiencing greater levels of cognitive impairment.

4.2. Practice Implications

Extant literature on conversations between nursing home residents and nursing aides has reported that there are differences in discourse and utterance level characteristics related to the degree of cognitive impairment, and that additional cues and repetition are needed for individuals with late-stage dementia [43]. In this study, we observed that participants, the majority with mild stage dementia, were able to discuss a variety of difficult topics with the aid of the Board (comprised of the canvas and icons), interviewer, and scenario structure.

In the context of our research, the study paved the way for the development of a decision making tool by facilitating the prioritization of important dimensions of supportive care preferences. Although this visual technique was designed for research use, in the future, one might consider adapting it for use in clinical and/or community settings. For example, in a situation where a social worker or case manager is asked to work with a person to determine future care preferences, a tool which accommodates individuals with cognitive impairment would enable initiation of discussions for complex personal decision making. Use of this technique would not need to be confined to persons suspected of having dementia; instead, it would allow initiation of difficult discussions without requiring that a specific diagnosis or cognitive assessment has been obtained – thereby reducing barriers to individualized future care planning.

In addition, the participant-interviewer interactions identified in Figure 5 might serve as a framework through which to examine dyadic conversation using interview elicitation techniques. Although many are specific to this tool, they could be generalized to represent actions that we might expect participants and interviewers to take in general, and thus be used to evaluate the effectiveness of conversation tools, as well as to train interviewers and others who work with persons with cognitive impairment.

5. Conclusion

In this study, we proposed and characterized the effectiveness of a visual technique to facilitate conversations about preferences concerning future supportive care needs in online interviews with persons with varying levels of cognitive impairment. In our results, we focus on three main themes: 1) participants’ perspectives on decision making; 2) cognitive and communicative processes; and 3) dialogic interactions between the interviewer and the participant, to better understand the utility of the visual elicitation technique. Common themes relating to considerations for decision making included the impact of the supportive care options on the caregiver, quality of care, and the influence of past experiences. Other, less common, themes included the importance of making decisions as a family, facing these challenges on one’s own, and the fluid nature of decision making. These could be potential ways to expand upon, and make more inclusive, the proposed approach for facilitating discussions with individuals with memory loss and dementia regarding their decision making preferences for supportive care options.

Statement of Significance.

Problem or Issue:

There is a need for more knowledge concerning how to facilitate conversations involving decision-making in the context of cognitive impairment.

What is Already Known:

Visual and participatory techniques can help improve engagement among persons with cognitive impairment, but challenges remain in how to effectively implement these techniques to support high level decision making.

What this Paper Adds:

We propose a visual approach to facilitate these conversations, as well as a method of analysis to characterize interviewer and participant interactions. The visual elicitation tool can facilitate conversations in a variety of ways, including prompting participants to think through complex topics, as well as providing opportunities for different forms of engagement.