Adding to the knowledge on Patient and Public Involvement: Reflections from an experience of co‐research with carers of people with dementia

Abstract Background Patient and Public Involvement (PPI) in research ensures that publicly funded research reflects the priorities of the people who will be affected by its results. Co‐research, a branch of PPI, is equal partnership between academic researchers and members of the public, who steer and conduct research together. Objectives To propose a model for good practice in co‐researching with carers of people with dementia, by reporting and synthesizing the personal reflections of the academic and lay researchers around the methodological issues, benefits, and challenges of co‐research. Design An academic researcher and two lay researchers with lived experience of caring with someone with dementia collaborated in all stages of a qualitative research study, including development of the research protocol and topic guide, data collection, analysis and synthesis, and dissemination of findings. Throughout the study, the academic and lay researchers annotated reflections of their experience in personal diaries. Data from the diaries were synthesized and mapped out in a model for good practice in co‐research. Results Co‐research yielded benefits for all those involved and on research outputs. There were practicalities and challenges that required extra resources, in order to make the involvement of lay researchers meaningful and effective. Discussion The model for good practice illustrates overarching and stage‐specific guidelines, which can inform researchers and members of the public wishing to undertake good practice in co‐research.


| INTRODUC TI ON
Patient and Public Involvement (PPI) is involvement in research of members of the public who have lived experience of the phenomenon under investigation. In PPI, research is carried out 'with' or 'by' members of the public, as opposed to 'to', 'about', and 'for' them. 1 PPI is grounded in the principle of 'knowledge by experience', as opposed to 'knowledge by expertise', which deconstructs the notion of hierarchical knowledge, passed by academic experts to lay members. 2,3 Knowledge, through a PPI perspective, is co-built by 'experts by training' and 'experts by experience'. 4 PPI has gained momentum in several countries and many funders, such as the United Kingdom National Institute of Health Research (NIHR) require that PPI plans are included in social and health care research grants.
There are different forms of PPI. 5 These range from consultation, where members of the public provide research input (eg evaluating research proposals) through inclusion in advisory committees, to leading the whole research cycle by members of the public. 6 Collaboration, or co-research, lies in the middle of the 'spectrum', and indicates equal partnership between academic researchers and members of the public, who steer and conduct research together. 6 Ideally, co-research occurs in all stages of the research cycle, including the development of study design and questions, research materials, data collection and analysis, and report and dissemination of findings. 7 In dementia research, lay researchers can be people living with the condition who have the ability to undertake a research role, as well as carer (henceforth defined 'carers'). 6 Co-research enables people with lived experience of dementia to have a voice in identifying research priorities and direct the research process. 8 This responds to the call for research to be qualified, ethical, and relevant to the primary stakeholders. 9 Co-research is also concerned with equitability (ie redistributing power in the research process) and challenges the power in the hands of the experts-by-training, typical of academic-led research. 10 Compared to other vulnerable populations, such as people with Intellectual Disability, 11 co-research in dementia is still limited practice. However, recent publications have evidenced benefits for academic researchers, research participants, lay researchers with dementia, and the public. 8,12 There is a growing emphasis on co-research in dementia, as demonstrated by the recent release of a special issue on the topic in the journal 'Dementia'. However, there is currently only a handful of empirical studies based on co-research with carers of people with dementia [13][14][15] and only a couple have reported on methodological issues. 16,17 Given the challenges of involving people with deteriorating cognition in research and that the method of co-research is still in its evolutionary stage, it is timely to experiment with different types of collaborators with lived experience of dementia. It is also crucial to recognize the central role that carers have in the production and delivery of care 18 and the burden that a caring role may generate, 19 to the point that carers have been defined 'the invisible patients'. 20 This renders carers experts-by-experience and requires that their voices are represented in research that affects their lives.
The present study was co-produced with PPI members with lived experience of dementia. It is ethically grounded in the Alzheimer Europe Strategic Plan 2016-2020, 21

| ME THODS
This study adheres to the consolidated criteria for reporting qualitative research (COREQ) checklist (Appendix 1). 23 It is based on an experiment of co-research with carers of people with dementia embedded in the Promoting Activity, Independence and Stability in Early Dementia (PrAISED) RCT. The PrAISED was designed to promote activity and independence among people with mild dementia or mild cognitive impairment ( Figure 1). 24 Embedded in the PPI of PrAISED, co-research occurred in the context of the process evaluation, 25 a sub-study aiming to investigate the experience of the participants with dementia and their carers ( Figure 1).

| Setting
This section describes how co-research was set up and carried out throughout the stages of the research cycle identified by Mockford et al. 17

| Thinking and planning
The experiment of co-research was not initially considered/ costed in the PrAISED proposal/protocol. When the main author (CDL) joined PrAISED in 2018 to lead on the process evaluation package, having previous experience of co-researching, he proposed to embed an experiment of co-research in the process evaluation package. CDL was able to utilize the financial and staff resources of the PrAISED RCT to set up and carry out co-research. The co-researchers' role was not pre-imposed by the academic team, but it was co-designed by the lay and academic researchers through discussions around the tasks that the lay researchers felt most comfortable undertaking. It was agreed that they would be involved in all stages of the process evaluation research cycle, including designing the study protocol, developing the topic guide, collecting and analysing data, and disseminating research findings.
This would ensure a genuine partnership in areas of research (eg data analysis) which have been traditionally 'academia-dominated'. 26 Being one of the co-applicants in the PrAISED RCT, lay researcher MG was involved in budget planning allocated for PPI contribution, which was based on INVOLVE guidelines. 27 It was agreed that the lay researchers would be compensated for their time spent working on the process evaluation in any of the stages of the study.
In relation to ethics, while the approval received for the PrAISED RCT covered the process evaluation sub-study too, we were unsure whether involvement of PPI members as lay researchers required further ethics approval. As prescribed in the Alzheimer Europe's pamphlet on PPI, 8  In relation to protocol design, the main author of the manuscript (CDL) developed a draft, which was passed to all co-authors (including the lay researchers) to comment and provide feedback on content and language. The lay researchers' input was crucial in ensuring that the study was empowering for the participants with dementia.
For example, although it was initially envisioned that the qualitative interviews would be conducted with the carers and the participants with dementia separately, the lay researchers suggested that asking to the participant with dementia about their preference would be a more ethical approach.
The development of the interview schedule also involved a collaborative effort. The topic guide consisted of open-ended questions, designed to stimulate reflection in the participants about the factors that had contributed to their experience in the PrAISED F I G U R E 1 PrAISED activities. *Includes informant and participant-reported measures on sociodemographics, medical history, medications, frailty, mobility, personality, cognition, quality of living, health, disability, falls efficacy, mood/ affect, activities of daily living, muscle strength, physical activity, static and dynamic balance, carer strain, and carer's health. **Receives physical exercises, functional activities (eg shopping), physical activity promotion; risk enablement; environmental assessment, community engagement and provision of information. # Includes qualitative interviews with participants with dementia and their carers at month six ad 12 of the intervention RCT (Appendix 2). The academic researcher developed a tentative version of the topic guide, which was then discussed and edited in a meeting with the lay researchers. This meeting ensured that the interview questions were relevant, meaningful, and jargon-free for participants with dementia. It was found that the terminology and structure of some of the questions was difficult to understand. For example, the concept of control was deemed to be too abstract. The question 'How much control do you feel you had in developing the programme of physical activity?' was therefore changed into 'How much were you able to decide what to do in PrAISED?'

| Preparing
As per the NIHR guidance on co-producing a research project, 28 some preparatory training sessions were undertaken to enable the lay researchers to undertake their role in co-research. 8 The needs of lay researchers were carefully evaluated and discussed. The lay researchers were asked to think about the role and reflect on which aspects they might find challenging and which specific skills they believed they should develop/boost. Having been PPI members in various research projects in the University of Nottingham, the lay researchers felt confident to fulfil most of the duties that their role included. The lay researchers had previously been trained in data analysis, through a two-day course delivered by the University of Nottingham research staff. The training consisted in an introductory session and workshop on qualitative data analysis.
The course participants were then asked to work independently on the coding of a sampled interview excerpt before the second session. The homework was discussed in a group format in the second workshop. The lay researchers felt that, given the vulnerability of the target population, training to carry out the qualitative interviews with participants was required. A plan was then set up, which comprised a half-day session held by the academic researcher and attended by both lay researchers.
In the first hour of the session, the academic researcher introduced the concept of process evaluation and the scope of the study in a jargon-free language and appropriate format, and then responded to the questions that the lay researchers might have.
The following 2 hours were focused on training. It was agreed that the aim of the training would be to make the lay researchers confidence enough to be involved in the interviewing process in such a way that it produced benefits for themselves, the participants, and data collection. It was agreed that the academic researcher, who had expertise in qualitatively interviewing vulnerable populations would provide refreshers around interviewing techniques, using the interview schedule, managing interactions, non-verbal communication, sensitive topics, anonymity, and confidentiality.
In the remainder of the training session, the co-research team arranged the practicalities of the interviews (eg transport, dates). In accordance with good practice in co-research, 8,28 ground rules, and specific research tasks were established, based on their suitability to the lay researchers' skills and aspirations and to research needs. The themes in the interview topic guide were shared between the academic and lay researchers. It was decided that, based on the concept of 'expert-by-training' and 'expert-by-experience', the academic researcher (with a psychology background) would ask questions about personal beliefs and motivation, while the lay researchers would investigate issues related to quality of life, such as emotional support and independence (Appendix 2).
This session was also an opportunity to build up rapport between team members. Maintenance of an open, honest, and trusting relationship throughout (and beyond) the collaboration was key to successful co-research. For example, the team created informal safe spaces out of the research environment (eg meetings in the lay researchers' homes) to further develop the quality relationships and make the meetings more accessible. 29 The team agreed that the training process would be iterative throughout the study, based on the emerging needs of the lay researchers. This proved effective after the lay researchers experienced a moral imperative to offer advice to participants (eg signpost activities available in the community) who exhibited difficulties (eg social isolation) during the interviews. In line with Corbin and Morse, 30 the co-research team felt a moral obligation to pass on information which might help alleviate participants' emotional burden. However, In order to avoid breaching boundaries between the scope of the interview and the intervention, and to avoid a negative impact on data integrity, information was provided after the interview session had ended, when appropriate (ie a decision was made by the academic team on the type of information that would not have an impact on the trial and sent to the participants by mail).

| Gathering
The participants with dementia in the PrAISED process evaluation and their respective identified carer were purposively selected by the academic researcher. This ensured that the sample was representative of the PrAISED RCT participants in terms of gender, ethnicity, relationship status, geographical location, and adherence to the exercise programme (ie low and high adherence).
The academic researcher contacted the participants through phone to enquire on their availability to take part in the process evaluation interviews and to confirm whether the participants were happy to be interviewed through co-research. None refused.
Once they had accepted to participate, the participants received a letter with the details of the appointment and an information sheet. The participants and their carers were interviewed (as dyads) through qualitative semi-structured interviews in their private home. The academic researcher carried out half the number of the interviews alone (ie with no layresearcher). It was agreed with the lay researchers that gathering interview data through two different configurations (ie with and without the layresearcher) would ensure easier identification of the added benefits and challenges of co-research.
In regards to the co-research interviews, the academic researcher and the lay researcher (one per interview) interviewed travelled to the participants' homes together and used the time before and after the session to brief and debrief. The debriefing was an opportunity to discuss any aspects of the interview session that might have had a negative impact on the emotional wellbeing of the lay researchers.
Verbal information (eg access to counselling services) and advice (eg how to process negative accounts from participants) were provided, when needed.
On the day of the interview, before the session commenced, each of the co-research team members introduced themselves.
Alternatively, the academic and layresearcher answered any questions that the participants might have, and gathered consent. The interviews were audio-recorded and continued until data saturation was achieved. The interviews were carried out from April to August 2019. In total, the co-research team interviewed seven participants with dementia and their carers (14 people in total). Carer participants were five spouses, one sibling, and one child. The sessions lasted on average 1 hour.

| Analysing
A professional agency transcribed the interviews verbatim. The transcripts were not returned to participants for comments. The transcripts were transferred onto NVivo 12. 31 The lay researchers decided to which extent they felt comfortable being involved in data analysis and the academic team agreed on an analysis plan. The academic researcher analysed independently (ie without the lay researchers) the transcripts of the interview he had carried out alone. Again, it was agreed that comparing different configurations in data analysis (ie with and without lay researchers) would facilitate learning points.
All the transcripts were analysed through inductive content analysis. The academic researcher underlined relevant pieces of text and wrote coding labels/ideas for each on the margin of each co-research interview transcript. The co-research interview transcript files were also sent to lay researchers, who read them and annotated their comments next to the text, independently of the academic researcher's annotations. Although some of the themes identified in the transcripts by the academic and lay researchers were in common, the annotations of lay researchers were instrumental in identifying further aspects relevant to the experience of dementia. In the example showed in Figure 2, for example, the layresearcher identified the themes of independence and privacy as central to the experience of physical activity in participants with dementia.
Once merged, the academic's annotations (from both the co-research and non-co-research interviews) and lay researchers' annotations (from co-research interviews) were used by the academic researcher to generate a tentative code book, including themes and subthemes emerging from the transcripts and their operational definitions. The tentative code book was passed to the lay researchers, who gave their feedback, which was used to aid construct refinement. For example, the construct 'capability' was expanded to include MG's observation that 'chronic conditions, such as arthritis, heart problems and stroke, might compound upon and mitigate capability'. The codebook will be used by the lay researchers, who will code two of the interview transcripts each. Inter-rater reliability between the academic and lay researchers will be tested through Cohen's Kappa coefficient. 32

| Writing and sharing impact
The lay researchers were involved as co-authors in all study outputs, such as the present paper and the process evaluation protocol. 25 They will also be co-authors of the upcoming main process evaluation report and a motivation paper, reporting on aspects affecting participants' motivation to engage in the programme. The co-researchers will schedule a visit to the participants to present study findings, so that they can give feedback, once the PrAISED trial is over (ie after month 12). The lay researchers co-presented research outputs with the academic researcher in seminars and lectures at the University of Nottingham, and a poster on the experience of coresearch at the 2019 Alzheimer Europe Conference in The Hague, Netherlands.

| Data collection
All three co-research team members annotated reflections of their experience of co-researching in personal diaries, throughout the collaboration. Reflective writing has become established as a method in its own right, as a data source and a key element of qualitative research. 33 It enables novel perspectives and insight on experience to emerge, which contributes to understanding and learning about practice. 34,35 Scanlon et al (2002:137) 36 contend that 'reflection enables practitioners to tap into knowledge gained through experiences'.
The co-research team agreed not to use any pre-imposed template to record their reflections, as this would allow the co-researchers' subjective experience to emerge more easily. However, to maintain consistency between the sources, each contributor received guidance to reflect around three main areas of co-research: methodological issues, benefits, and challenges (as per study aim).
It was agreed within the co-research team, that the personal reflections would be annotated as soon as possible after each qualitative interview with participants, to ensure retentions of fresh memories from the session. A summary of the reflective pieces compiled is featured in this study (see results).

| Data analysis and model generation
Once compiled, the individual reflections from each co-researcher were collated by the academic researcher and stored safely in a password-protected computer to ensure data confidentiality. At the end of the co-research experience, the co-research team joined in a three-hour session held in one of the lay researchers' home (MG) to analyse the data and co-produce a model for good practice in coresearch with carers of people with dementia.
In this session, the co-researchers individually identified and extrapolated from their own diaries aspects they deemed as relevant for good practice (Table 1). In a team effort, the co-researchers then aggregated similar aspects reported by the different co-researchers and generated overall principles on how researchers should address these aspects to pursue good practice in co-research.

F I G U R E 2 Excerpt of interview transcripts, with annotations from the academic (CDL) and lay researcher (MD)
TA B L E 1 Identification of relevant aspects in co-research and how these can be addressed in practice

Coresearcher Aspects identified as relevant through the co-researchers' reflective diaries
Overall principle (ie how to address the aspects). The academic researchers should…

MD (lay researcher)
The lay researchers feeling worthy/motivated Adopt a non-tokenistic approach, build rapport/ foster trusting relationships, give back to lay researchers, keep lay researchers in the loop

| Lay researchers MD reported
Being involved with the study through the whole process, contributing knowledge gained from lived experiences, was fulfilling and made me feel I had, in a small way, given back to the community. In the context of data collection, meeting participants and carers in their homes to discuss personal information required a great deal of experience in managing sensitive situations. A natural ability to relate and empathize with a wide sociocultural base was also required, to ensure participants and carers were treated with dignity

Coresearcher
Aspects identified as relevant through the co-researchers' reflective diaries Overall principle (ie how to address the aspects

| Lay researchers MG added
This has been the most affirming and positive piece of work that I have had the pleasure to be part of, in the 10 years that I have given to PPI, primarily in dementia studies.
Approaching ing of the study and committed to its success. When we met with the academic researcher, who embraced our roles and ensured we were fully confident to make this further contribution, we knew we could positively meet the challenge. All this enabled us to establish a connection with the participants, and this was clearly such a strong factor in ensuring the interviews were both structured as planned, but still allowed them to evolve organically, as our empathy and understanding helped widen and deepen the experience.
My introduction to the participants as a layresearcher was made easier, as a photograph of me and my husbands fronted the large PrAISED work folder that each participant had been given on their involvement in the study. I think this further helped to lower barriers, as my photograph had practically been with the participant and carer for months. This seemed to create an initial bond of trust, which became a significant part of the interview process. While the physical voices of the carers were heard more often, their inner thoughts and feelings were more often than not put to one side. The carers tended to prioritize reporting of the participants' needs and their experiences. Therefore, I felt that my responses as a carer with lived experience in a small way helped to draw out the reality of the impact of the dementia on their relationships and lives. Therefore, as a team, we gained a more comprehensive picture of their involvement in the study and the significant changes it brought in their daily living.
The opportunity to debrief with the academic researcher, who brought his professional capacity as a psychologist into every discus- to work through personally, also at the level of professional rapport with members of the academia.
In conclusion, the experience for me has been very powerful. To meet, at last, with those taking part in the study felt like truly being embedded in the research process. It gave me further insight into others' lived experiences, enabling me to prepare for a future, which could see me on a parallel journey.

| The academic researcher's personal reflections
Involving the lay researchers in the PrAISED process evaluation A mixed-gender research team was also a valuable asset with the couples who adopted strict gender roles. It was noticed that female participants would more often keep eye contact with the female researcher, while the male participants would find it easier to relate with the male researcher. It was also observed that a mix of personalities in the research team was helpful to enhance data collection, as the participants were more likely to find a type of personality they were better matched with and open up more easily about their experience.
Other than personal characteristics, the lay researchers brought added value to the interview process, given their lived experience with dementia. They displayed great 'situational sensitivity', which is 'consideration of the interest and vulnerability of the particular participant, rather than application of general rational principles'. 37 For example, knowing first-hand the sensitivity of terminology, though this was not established as a ground rule during the preparatory work leading to data gathering, the lay researchers avoided using the term 'dementia', unless it was first mentioned by the carer or the participant during the interview.
In terms of data analysis, the differences between the two different configurations (ie with and without the input of lay researchers) were more marked. Having different background and experiences, the lay researchers' transcript annotations grasped different nuances of the interviews. For example, MD tended to focus on the use of language and its underlying meaning, while MG on aspects related to quality of life and social networking. The inclusion of the perspective of lay researchers in the generation of themes and subthemes was therefore instrumental to contain academic researcher's bias and allowed validation of process evaluation data with people with lived experience with dementia.
There were also some challenges experienced in co-research.
Undertaking co-research might potentially delay data collection, as extra time is needed to agree on appointment dates, which are suitable to all those involved in the process. While co-delivering the interview with lay researchers aims to make the process more equitable for all research participants, it must be recognised that when lay researchers are carers, they might more easily empathize and relate with their own peers (ie the carer-participants). In such scenario, the participant with dementia might still perceive the interview as disempowering. The empathic bond between the lay researcher and the carer participant might also have repercussions on data collection. It was noticed that some carers tended to confide in the lay researcher out of the formal interview session, revealing information that were relevant for the study in confidence to the lay researcher.
This suggests the importance of rigorous research protocols to be agreed within the co-research team, prior to contact with the research participants.

| Generation of a model for good practice
The model for good practice in research with carers of people with dementia ( Figure 3) features research-stage-specific and overall principles. While the former apply specifically to co-researching with carers of people with dementia, we feel that the latter represent transferable information which can be used in co-research with other members of the general public (ie without experience of dementia).
In relation to the research-stage-specific guidelines, the 'think and plan' stage requires the set-up of plans for long-term collaboration with lay researchers (during and after the research cycle) early on in the research 38 and an appropriate selection of the lay researchers. Ideally, the layresearcher role should be co-designed with PPI members who wish to be involved. In order to facilitate effective and meaningful contribution in the research process, attention should be dedicated to both the lay researchers' aspirations, preferences, and goals to be obtained from involvement and to the skills they need to fulfil their research role.
It is also ideal to select lay researchers with different skillsets, background, experience from each other, so that each of them can contribute uniquely to the process. Although separate ethical approval was not a requirement for co-research in this study, ethical and governance demands may vary across different institutions, as reported in the literature. 15,39 Academic researchers should therefore work proactively to ensure that approvals (eg letters of access for lay researchers to have contact with research participants recruited through the NHS) are in F I G U R E 3 A model for good practice in co-research with carers of people with dementia place when involving lay researchers. In the 'prepare' stage, the model emphasizes the importance of training offered to lay researchers, which needs to be on-going throughout the project, based on emerging needs.
At this stage, it is also essential for good working relationships to discuss, negotiate, and establish clear research roles.
In the 'gather' stage, the co-research team members should share commuting to the location of the data collection, which will promote bonding and allow pre-session preparation and post-session discussion. This may also resolve transportation issues that the lay researchers may have to reach the interview locations. To ensure effective data collection, all those involved in the session should be made comfortable. In order to make the interview process equitable for the participant with dementia, the team should send photos of the researchers in advance, so that the participant can get familiar before the session. The photos could also help the participants with memory impairment to remember the co-researchers in the follow-up visit. In the context of a co-research interview, four people might be present.
Given the 'expertise' of co-researchers, the participants with dementia and their carers may be in a disadvantaged position, as they might be less articulate and feel less confident to voice their views. This potential power differential may require great skills of interaction management on the part of all researchers, to ensure equitability in the interview process. 15 Again, this highlights the importance of training.
Training is also crucial to ensure data integrity. An important methodological issue we encountered is the moral imperative to give support and advice that lay researchers may experience when hearing the participants' difficulties. This may breach boundaries between research interview and intervention and have an impact on data integrity. In the lack of appropriate training, there is a risk of actively influencing responses, as opposed to merely elicit them.
During the interview process, it is also crucial to ensure the physical and emotional wellbeing of the layresearcher, who should always be supported by the academic researcher. As found in previous studies, when hearing the difficulties of the participants, the lay researchers might reflect on their own situations and become emotionally stressed. 15 Therefore, strategies should be in place for lay researchers to discuss any concerns/ emotional issues. Ideally, a specific distress protocol should be in place to provide verbal and written support to lay researchers.
In the 'analyse' stage, the model emphasizes the importance of lay researchers interpreting/analysing data independently of the academic researcher, before results of the analysis process are merged.
This will ensure different perspectives and their own voices to be reflected in the research outputs. In line with previous studies, 39 in the 'write and share impact' stage, lay researchers should contribute as equals in writing reports and dissemination materials (as active co-authors, by writing their own sections) and in presenting research outputs at conferences, seminars, and talks.
In relation to the overarching action points (ie applicable to all stages of the research cycle), the academic researchers should strive for non-tokenistic and meaningful collaboration in all research stages. In order to engage effectively in co-research, they can seek help and guidance from experienced researchers in the field and/or university resources. The academic team should allocate extra resources in place to pursue good practice in co-research, including extra funds, time, and staff. 15 In the context of a large RCT, financial and staff resources were such that co-research could be adequately funded, even though it had not been considered/costed in the original project proposal/ protocol. However, given its resource requirements, academics willing to undertake co-research should calculate costings from the initial stages of the project application.
Throughout the co-research experience, it is crucial to cultivate good rapport within the team, and foster trusting relationships. This could be facilitated by members spending time away from work together. The whole experience of co-research should ideally be recorded in reflective diaries, which can be used not only as key data sources, to advance good practice, but also to promote personal and professional development. Finally, as also reported in previous studies, 29  The study is characterized by certain strengths and limitations. and resources and the sustained commitment of university institutions.

CO N FLI C T O F I NTE R E S T
None to declare. * These items are not applicable, as this is not the study reporting findings from the process evaluation.

A PPE N D I X 1 (Continued)
A PPEN D I X 2

TO PI C G U I D E FO R Q UA LITATI V E I NTE RV I E WS
NB -the following questions are suggestions and prompts -some answers may be anticipated earlier in the discussion and others turn out to be not relevant. The interviewee may also raise additional topics and issues which they feel are particularly relevant and these should be followed up in the discussion.

Pre-interview
• Researcher introduces himself and engages in small talk to break the ice with participant (eg give thanks for being invited over, gives compliments about the home, and asks how the person is doing on the day).
• Researcher explains his professional role and the purpose of the visit • Researcher goes through the Information Sheet with the participant. The following will be clearly explained: 1 The interview will be audio-recorded to have an accurate record of what was said 2 Anything mentioned during the interview is confidential and no one except members of the PrAISED research team will know what was said 3 In using any information in a report, the participant's anonymity will be maintained

Motivation
• Why did you decide to take part in the programme?
• Were you encouraged by anyone to take part or was it your own choice?
• What helps you keep going with the programme?
• On a scale from 1 to 10, how much do you feel you want to continue with the Activities, once the programme has finished?

Autonomy and control
• Is it important for you to decide what you do or do you prefer to leave it to others?
• (If yes to previous question), how much have you been able to make those decisions?
• How could we make you feel more involved?

Intervention characteristics
• Does the programme of physical activities suit your needs and preferences?
• What part of the programme of physical activities do you like the most?
• What part do you like the least and how could this be improved?

Self-efficacy and emotional support
• Do you feel you are able to do the activities as well as you would expect?
• Do you have any concerns or anxiety about taking part in the programme/doing the activities?
• Did you receive encouragement and support from your therapist(s) and carer(s)?
• Is there anything that would help you feel more confident to do the activities?

Support (Practical)
• Do the therapists give you practical support? For example, do they show you how to do the activities, when to do them and where to do them?
• Does your (carer role) give you practical support? For example, does he/she remind you how to do the activities, when to do them and where to do them?
• What could be done to better support you?

Independence
• How has the study programme affected you? (eg on your health and activity) • Has it given you greater independence?
• Have you noticed a change in your quality of life?
• Are there any activities you would like to be able to do that are not part of the programme?

Expectations
• Have you any personal goals you would like to achieve from the study?
• If yes, what goals are you looking to gain?
• Do you think you can achieve these goals and do you need support to do this?

FI N A L R EM A R K S
• Any final thoughts and feedback on the programme?
• Would you be happy to meet up again in 3 months' time to see how you are doing?
A PPEN D I X 3