Drawn from life: Cocreating narrative and graphic vignettes of lived experience with people affected by dementia

The growing literature on Patient and Public Involvement and Engagement (PPIE) and dementia identifies specific problems related to the influence that involvement has on research outcomes, over‐reliance on family members as proxies and lack of representation of seldom‐heard groups. Adaptations to the PPIE process are therefore needed to make possible the involvement of a broader spectrum of people living with dementia.

PPIE is now considered essential in health and social care research, and is increasingly a requirement of research funding bodies and publishers. 1,2 Crucial to the definition of the term PPIE is that those involved are advisors or coresearchers rather than research participants. 3,4 The extent to which PPIE group members have genuine influence on the research process is often questionable, with involvement in aspects such as design, data collection and analysis particularly limited. [5][6][7] These concerns increase when the research relates to conditions such as dementia, which are characterised by cognitive impairment.
Of 54 articles on PPIE in dementia research in a recent scoping review, 8 almost all were published since 2010, indicating how recently dementia has come to the PPIE table. Few articles reviewed reported on the impact of PPIE on the dementia research process and outcomes. 8 Yet, PPIE has significant potential for improving aspects of dementia research such as recruitment from seldom-heard groups, 9 and people living with dementia have said they want more opportunities to act as coresearchers. 10 Coresearch involving family members is now increasingly well established. 11 Despite people living with dementia having published guidance for researchers hoping to involve them in research as long ago as 2014, 12 however, coresearch with people living with dementia is still comparatively rare. 13,14 In one recent study, for example, the three 'people affected by dementia' in a PPIE group were all current or former caregivers. 15 While it is less challenging to recruit family members 16 research findings indicate that people with dementia often have different views and priorities from those expressed by their relatives. [17][18][19] In recent years, direct involvement of people living with earlystage dementia in research processes such as data analysis workshops 11 and coauthored accounts of the research process 20 have become more frequent, showing that it is possible to hear the voices of people who are actually living with the condition being researched. This is not just a matter of inclusive principle, however, since without the perspectives of those living with the condition, research lacks validity and important insights may be missed. 21 Attention has also been drawn to the ethical imperative to involve a more diverse range of people living with dementia in all aspects of research. 22 At present, those who take an active part in PPIE are often recently diagnosed or have relatively mild cognitive difficulties, raising questions about the representativeness of the experience that is being drawn upon. Younger, recently retired members of PPIE groups are, for example, unlikely to have personal experience of the services provided to older people living with dementia who may have more pronounced cognitive difficulties and those from underserved groups are likely to have had different, possibly more extreme, experiences than those who are younger, more recently diagnosed or more socially visible. 23 While recent welcome developments such as the Balanced Participation Model 24 have been developed with people with early-stage dementia, there are currently fewer models for the involvement of people with more severe cognitive difficulties.
It has been suggested that due to the challenge of involving people with more severe cognitive problems, the representativeness of the wider population with dementia may be less important than thoughtful input from lay people without this condition. 25 We suggest, instead, that there is a moral requirement to create conditions under which people with dementia, including those from seldomheard and underserved groups, are able to contribute. 22 While the involvement of people with dementia does present practical and methodological complexities, 26 the onus is on researchers to adapt their methods and processes accordingly. This has already been achieved with some of the more creative approaches to research participation for people living with dementia, 27 and coresearch involving people with young-onset dementia (YOD). 20 Below, we outline a PPIE project in which older people with more advanced dementia also took part, in addition to people living with YOD and family members.

| The research study
The What Works in Dementia Education and Training? study, within which this PPIE project took place, was designed to identify approaches to, and characteristics of, effective training and education on dementia. Ethical approval for this study was given on 24 We do not intend to go into detail in this article about the findings from the broader research, 28,29 but to explain the process adopted in the PPIE arm of the study and its outcomes.

| Aims of the PPIE
The aims of this project were as follows: (1) to draw on the personal testimony of people with dementia and their families to facilitate discussion between researchers and health and social care practitioners in the field of dementia studies and (2) to use an enhanced range of methods to support the involvement in PPIE of people with dementia across a spectrum of cognitive abilities.

| Recruitment
The opportunity to take part in the PPIE group for the What Works in Dementia Training and Education? study was advertised in a regular Experts by Experience newsletter sent to members of the existing PPIE panel at one of the Universities involved in the research. The circulation list included a resource centre attended by people living with dementia. This was a site where one of the coauthors had conducted a number of previous community participatory research studies, and there were pre-existing good relationships with both staff and people with dementia. The copy of the newsletter sent to the resource centre participants was accompanied by an additional explanatory letter to the manager asking for expressions of interest in taking part.
All those who expressed an interest in taking part, from both the general mailing list and the resource centre, were provided with an information sheet summing up the aims of the research in plain English. For the three resource centre participants considered not to have the capacity to decide whether to take part for themselves, best interests assessments were carried out, involving the manager and family members. All other participants were able to consent for themselves.

| Participants
A total of fourteen participants contributed to the project discussed below. Nine members were living with dementia (six women, of whom one was Black Caribbean and one was White Irish, and three men, of whom one was White Irish). Of these, one man and one woman had YOD. The remaining five participants were family members (three men and two women, all of whom were White British).
Three of the people living with dementia (one accompanied by their spouse) took part on a University campus, together with four family members, all of whom were former carers for a spouse living with dementia.
The remaining six people living with dementia took part at the resource centre that they regularly attended during the daytime.

| The PPIE process
The PPIE group met on a total of 15 occasions during the 2 years of the research study and had a varied remit during that period, including the creation of a lay summary of the literature review findings, reviewing information materials for the research participants, codesign of survey items and the creation of vignettes of their lived experience to be used in research focus groups with care staff. The process discussed in this paper relates specifically to the last of these activities.
The PPIE was cofacilitated by two coapplicants on the research study: one a family member whose spouse has dementia and one an academic researcher and educator. The roles of the cofacilitators were, generally speaking, to organise PPIE meetings, keep records, coordinate discussions, advise PPIE group members on ethical and methodological issues, liaise with the broader research team and report back to the Programme Advisory Group and funders on progress.
Several of the PPIE group members already belonged to groups affiliated to the UK Dementia Engagement and Empowerment Project (DEEP) and one had taken part in an advisory group on the development of the DEEP guidelines on involving people with dementia in steering and advisory groups. 30 We therefore followed these guidelines throughout.
An initial 1.5 h meeting was held at each of the two sites to elicit members' lived experiences of using dementia services. The meeting at the University, attended by eight people, took place in a small seminar room, with a 'round-table' layout. The meeting at the resource centre, in which six people actively participated, took place in the main lounge.
Although this area was also being used by other people at the same time, the room's layout, with a variety of small seating areas, made it possible to conduct discussions relatively privately. Each meeting was recorded using both written notes and audio-recording.
To move away from more traditional methods of conducting PPIE, such as formal consultation exercises, we used a variety of narrative elicitation techniques. Narrative healthcare has attracted renewed interest in recent years, and narrative elicitation is a widely used method for embedding principles of patient-or person-centredness. 31 In the work of Arthur W. Frank, for example, it is recognised that when a person chooses to tell a particular story, he or she is also telling the listener about what matters most, and stories are a powerful way of amplifying the voices of those who are seldom heard. 32 At the University site, oral narrative elicitation was the main method used. The following trigger questions were used: Please think of times when you (or your relative in the case of family members) were 1. Not treated equally as a person.
2. Talked about using negative language.
4. Provided with a really good service. 5. Treated in a way that added to your quality of life.
6. Made to feel you belonged.
We found that participants at the University naturally adopted a storytelling approach, in which they answered the prompt questions by recalling specific incidents in considerable detail. Rather than sticking closely to the prompt questions, we therefore encouraged the participants to elaborate on these experiences, which were often recounted with a degree of dramatic performance.
[My partner] fell on the stairs and had to go to hospital.
She had eleven stitches in her head. And the doctor was going to discharge her. I said, 'Hold on; she told me she can see two of everything'. They should have tested her for concussion, but they didn't because she had de-

mentia. [family member]
Participants at the resource centre had, in general, more difficulties with memory, comprehension and verbal expression and several also had impaired hearing. As a result, the two cofacilitators used enhanced elicitation methods at this site, including photoelicitation, informal conversation and observation. Around half of the meeting at the resource centre was spent with a small group of three participants discussing the same general subject areas as the University group. The difference here was the use of The written notes and audio-recorded material from both sites were transcribed by the academic cofacilitator, and key passages were discussed at several subsequent PPIE group meetings. In addition to identifying key themes, we were also interested in discovering which formats the group members felt might have maximum impact. The process by which the group decided to cocreate a series of vignettes is explained in Section 3.

| FINDINGS
At both sites, it was easier to elicit responses related to poor care than it was to identify positive experiences. It was, however, gen-

| Narrative elicitation
Experiences of using services that the participants referred to included primary care consultations, homecare visits, memory clinic assessments, admission to hospital for acute care and community facilities such as pharmacies, supermarkets, banks and libraries.
At the University meeting, many of the comments were about microinteractions between healthcare professionals and people living with dementia, such as disparagement and invalidation. It helped a lot in the care home she went into, because they just let her get on with it. To me that was the most important thing, and they need to find ways to do it.

[family member]
At the resource centre, the three people in the small group told us about a number of specific incidents. Brenda (all names have been changed) told us that she did not like to be rushed, and then recounted an experience with the homecare service that she received, where there was no time to choose her own clothes. Brenda pointed out that the clothes she was wearing were not her own choice: Look at these horrible trousers. I didn't want to wear these. She just put them on me.
It was possible to draw inferences about other participants' priorities less by answers to explicit questions than by spontaneous remarks. For example, Peter was very concerned about items going missing, and we noticed that Fiona, who had been widowed some time ago, still made frequent reference to her late husband as a source of authority.
The resource centre members identified a number of experiences of having their direct choices and wishes over-ridden. They valued being treated as individuals, feeling socially included, taking part in meaningful activity and having their concerns taken seriously. Not being allowed to do things that are perceived as dangerous by others was also a repeated theme.

| Informal conversation
We spoke to Valerie, who was sitting alone in a corner of the resource centre lounge. Predetermined questions and prompts were not used, as Valerie preferred to tell us a story about her earlier life.

| Observation
While sitting with the remaining two women at the resource centre, we did have some conversation, but it was not directly related to their experience of dementia services. Instead, we made detailed notes about the women's (Ella and Christine's) interactions with each other, including their nonverbal communication. Both women appeared to find that sitting together added to their existing discomfort, and this seemed to go largely unnoticed by staff, who were getting the other side of the room ready for lunch. Many of Christine's comments were directed to an unseen person who seemed to be a figure of authority, and she spoke about an injury she had received treatment for. When Ella tried to join in politely, Christine tended to respond short-temperedly, clearly upsetting Ella. With the other findings above, this interaction later became the basis of one of the three vignettes.

| DEVELOPING VIGNETTES OF LIVED EXPERIENCE
The decision to cocreate a series of vignettes (or 'scenarios' as PPIE group members preferred to describe them) evolved during the process of transcribing and organising the recorded material into key themes. The findings seemed to lend themselves naturally to the development of a set of stories about the experiences that the PPIE group members had recounted. We agreed that we did not want these stories to exist solely in the form of text. Some members of the group had difficulties in processing blocks of text, and some had worked in fields where an understanding of different learning styles was important (e.g., staff development and educational welfare), so we felt it was important to use images as well.
The term vignette refers to text, images or other forms of stimuli to which people are invited to respond, 33 an approach widely used in education, training and staff development. It has been noted that vignettes offer a means of unpacking complex, nuanced material and making it more accessible. 34 Vignettes appear to promote high levels of engagement and empathy, with respondents often placing themselves and their experiences within the scenario. 35 While a fictional vignette (VIG-Dem) has previously been used to assess clinical skills in dementia, 36   The three scenarios produced are outlined in Table 1. Scenario 1, set in an Emergency Department, was based largely on a former family caregiver's account of his wife's hospital admission following a fall that had resulted in a head injury. Male and female roles were reversed in the Emergency Department scenario, and to introduce issues related to cultural diversity, the characters were also given names (Constantin and Ava) suggesting Eastern-European origin (for the full scenario in narrative format; see Table 2).
For each scenario, a list of key points for practice improvement was also produced by the PPIE group members. For example, in Scenario 1, among 10 points in total that it was hoped health and social care practitioners taking part in the research would identify were 1. Constantin appears to have double vision, and yet he is not sent for a scan.
2. There seems to be a lack of recognition that his behaviour in the toilet is the result of fear.
3. He is left for more than 24 h with nothing to eat.  Table 3 shows how this was done.

| Graphic storyboarding
All three scenarios were produced in both narrative and graphic forms. This dual format was decided on by the PPIE group members primarily as a way to engage different types of learners. As previously noted, there can be advantages to using both text-and image-based scenarios depending on the groups to whom they are to be presented. 33 The dementia care workforce is both multicultural and cross-generational. Some of those encountering the scenarios would not have English as a first language, and graphical vignettes may be preferred among cultures and age groups familiar with forms such as Japanese Manga or graphic novels. There are other notable exemplars of the use of graphic novels and cartoon strip format to convey the lived experience of dementia 39 and to support learning about homecare for people living with dementia. 40 Although textual representation is privileged in social sciences research, it has been suggested that the engagement of other senses, including sight, is more likely to promote empathy. 34 To facilitate the storyboarding process, each narrative was first broken down by the group members into six 'mini-scenes' capturing its key action points. This was done first in a one-page, six-panel, verbal form. Table 4 demonstrates the key elements from Scenario 2 in which Nosheen visits her GP. A graphic artist whose parent was living with dementia, and who was related to a member of the PPIE group, was commissioned to carry out the corresponding artwork. Figure 1 shows the graphic novel-style storyboard for Nosheen's GP appointment.
The key elements of the stories of the two PPIE group members with YOD were kept intact in both the narrative and the graphic At 9 PM, Constantin has 11 stitches put in his head. Ava is told he can now go home, but tells the nurse, 'I'm worried that he says he can see two of me'. As a result, Constantin is kept in the hospital overnight to undergo a brain scan the following day.
The next morning, on the ward, Constantin is given porridge, which he does not eat. He is taken for his scan at 10 AM, but it does not take place until 12.30. When he is brought back to the ward, he has missed lunch. Ava has gone home, and while she is away, Constantin is brought his teatime meal, which is left on a tray next to his bed. Constantin cannot eat without help, even though the food is something he likes this time, and 30 min later, it is taken away again. When Ava comes back, Constantin says he is hungry and thirsty. Ava tells the charge nurse, who says he will make sure Constantin gets some tea and toast as soon as visiting is over. The nurse is then sent to deal with an emergency on another ward and when he comes back at 7 PM, Constantin is asleep.
| 1895 version: These include discriminatory assumptions made solely on the basis of diagnosis; attributing all Nosheen's symptoms to dementia (diagnostic overshadowing); taking for granted that a 'carer' will attend appointments along with the person with dementia; and overuse of medical terminology.
Purely fictional life history material related to people with dementia often fails to take into account the impact of national and social events, or aspects of care practice that can trigger unpleasant memories of such events. Although some details of Scenario 1 were changed, it was based on the similar experience of a real person with dementia and her reactions to such a trigger factor. Violence, imprisonment and sexual abuse are among those 'off-limits' experiences of which, it has been suggested, vignettes may help to promote discussion. 33 In response to Scenario 2, a primary care practitioner reflected, in response to Nosheen's visit to the GP: You don't know how cognitively impaired she is really.
You should direct the information to her anyway and make sure that someone else has the information so that …. At the same time do we know that she wants to bring someone else to her appointment? She might not.
Here, we see how a vignette might help with the development and assessment of clinical skills. 36 The practitioner checks his or her own first assumptions ('so that…At the same time') in the process of engaging with the scenario, reflecting that passing personal information to a third party may not be appropriate without more information about Nosheen's own mental capacity.

| Strengths and limitations of the PPIE process
Direct involvement of people with dementia in coresearch is still rare, 8 so the fact that 9 of the 14 members of the PPIE group for this project were living with dementia was a positive aspect of this study.
There was also a significant degree of equality between group members in the cocreation of the scenarios, with both people living with dementia and family members able to develop scenarios based on their own experiences. Membership was reasonably diverse in terms of age, gender, ethnicity and living arrangements, going some way towards addressing the need for diversity in PPIE advocated by recent sources. 3,21 Perhaps partly because of this, however, rates of attrition were also relatively high during the 2 years of the study. One of the five family members died, one person living with dementia moved to formal care and another moved out of reasonable travelling distance to be able to continue taking part by the end of the process.
The resource centre also underwent some changes, and none of the six participants who took part there were still attending by the end of the wider research study.
The inclusion of older people with dementia, who also had more pronounced cognitive difficulties and were recruited via outreach work from a formal care service, remains a novel aspect of this PPIE project, as do the adapted methods that led to the scenario based on Ella and Christine. At the same time, it is possible to question whether Ella and Christine's own involvement took the form of genuine cocreation, and despite the outcome of their best interests assessments, concerns may be raised about their awareness of the nature of the process they were taking part in.

| CONCLUSIONS
Three key problems identified in the literature on PPIE involving people with dementia are the lack of direct influence of people with dementia and their families on the research process or its outcomes, the lack of direct involvement of people with dementia by comparison with proxies and the lack of representativeness of the wider population with dementia, including those who belong to seldomheard groups. The PPIE work reported here attempted to address these problems in three ways: through cocreation of resources based on the personal testimony of people living with dementia and their families, by direct involvement of people living with dementia rather than reliance on proxy accounts from family members alone and by using an outreach approach and adapted methods to include people with dementia whose voices are less often heard.
Work of this nature can, however, create new ethical issues. The involvement of four of the six people at the resource centre seems to be relatively unproblematic from an ethical point of view. For the three participants who took part in the discussion group, their contribution was not a great deal different from those who took part at the University. The one-to-one narrative approach adopted with the fourth reflected her own preferred communication style; it demonstrated appropriate flexibility on the day and a willingness to adapt PPIE activities to individual preferences. 20,22 The involvement of the two remaining participants does, however, raise additional ethical issues. On balance, we believe that including Ella and Christine's lived experience in a fully anonymized form was the right thing to do. While we could have decided to exclude our observations from the time we spent with them, their insights into what it is like to spend time sitting in a communal lounge as a person with dementia seemed particularly valuable in the context of staff development-the focus of the research study within which this PPIE was carried out. Moreover, people like Ella and Christine rarely have any kind of voice in dementia research.
The major ethical objection is that the two women were not aware of the purpose of the exercise. This is, of course, often the case in dementia research, particularly where this relates to medical treatment. It might well be argued that we should expect higher standards for informed consent when the person's role is to influence research rather than merely to take part in it. This would, however, risk further social exclusion of those with more significant levels of cognitive impairment, whose experiences are likely to be among the most extreme and where there are currently the greatest deficits in understanding.
Research into the use of vignettes suggests that they have more contributing to drafts. Michelle Drury, the researcher who used the vignettes in data collection, was responsible for sharing data and reading and contributing to drafts. Sahdia Parveen, co-applicant on the funding bid, was involved in reading and contributing to drafts.
Claire Surr, Principal Investigator, was involved in reading and contributing to drafts. Lesley Healy, a member of the Patient and Public Involvement and Engagement group, contributed to the development of the vignettes. Cara Sass, the researcher who used the vignettes in data collection, was involved in reading and contributing to drafts.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.