Patient and public involvement cultures and the perceived impact in the vulnerable context of palliative care: A qualitative study

Abstract Background Cultural values are crucial to the practice and impact of patient and public involvement (PPI) in research. Objective To understand different PPI cultures among research teams and the impacts of PPI associated with each culture type. Design A participatory action research design. Setting and participants The setting was 10 palliative care research projects. Seventeen patients and members of the public and 31 researchers participated. Intervention A programme consisting of four components: (1) training and coaching of patients and the public to prepare them for participation in research, (2) tailored coaching of the 10 research teams over 12‐18 months, (3) a community of practice, and (4) a qualitative evaluation. Results We identified three cultures types: relationship cultures, task cultures, and control cultures. We identified four areas of impact: the project aim became more relevant to the target audience, methodological reliability increased, the research products were better able to reach the public, and the awareness increased, associated with behavioural changes, among researchers regarding PPI. Discussion A relationship culture appears to be long‐lasting due to impacting the behaviours of the researchers during future projects. Different cultural types require different types of patients and researcher participants, assigned to different tasks. Conclusions Further research remains necessary to investigate the support required by researchers to enable relationship‐ and task‐oriented PPI cultures. Patient or public contribution Patient advocates and representatives contributed to our research team throughout the entire research process, as well as within the 10 implementation projects.


| INTRODUC TI ON
Patient and public involvement (PPI) refers to the involvement of health services users, such as patients, caregivers or community members, in research and is increasingly practiced and studied.
Involvement is defined as an active partnership between the public and researchers in the research process. Active involvement may take the form of consultation, collaboration, or user control. Many people define public involvement in research as 'doing research 'with' or 'by' the public, rather than 'to', 'about' or 'for' the public'. This would include, for example, public involvement in advising on a research project, assisting in the design of a project, or in carrying out the research. 1 The Canadian Institute for Health Research's Strategy for Patient Outcome Research (SPOR) adds that PPI is about 'meaningful' collaboration, including governance, priority setting, conducting research, knowledge translation, and evaluation. 2 Shippee et al distinguished four essential components for meaningful PPI: 1) patients should be involved early on, with a clearly defined role; 2) researchers and patient representatives should be able to develop equal partnerships; 3) patient representatives should co-learn and be prepared for their tasks; and 4) PPI should be systematically evaluated to improve its impact. 3 Several systematic studies have provided insights regarding the facilitators and barriers to PPI implementation. [3][4][5] One factor that is often mentioned is culture. Several reviews, such as Shippee, 3 Manafo, 4 and Chambers, 5 have mentioned relationships, a culture of involvement, and values as being components of PPI culture; however, these reviews have not provided an in-depth investigation into the impacts of PPI culture. 5,6 Based on current literature regarding organizational culture, PPI culture can be defined as how power is organized, ranging from concentrated to shared power, and the types of social relationships that are formed between patients and researchers. 7 Studies focused on PPI impacts have mentioned the relevance and appropriateness of the project to the users, the quality of the research, and improved relationships between researchers and communities. 1 Manafo et al 4 distinguished between three impact levels: PPI impacts in the near-term (the individual level), the intermediate-term (organizational level) and the long-term (systemic level). Chambers et al 5 found that the effects of power, diversity, and emotions, which are all related to our understanding of culture, are especially magnified when PPI is implemented in palliative care research. They concluded that evidence supporting the impacts of involvement was limited; however, when implemented effectively, PPI can benefit all stakeholders by improving the relevance and quality of research. Because PPI impacts have been poorly reported in previous studies, 4,8,9 we aimed to identify different PPI cultures and to assess the impacts of each culture type in 10 PPI projects in palliative care research.

| Design
This qualitative study applied a participatory action research design 10 to create new knowledge and change practices. We implemented PPI in ten palliative care research projects. This study was performed as participatory research because we did not study the participants from 'outside' or 'above'; instead, we collaborated with patients, the public, and researchers. Participatory action research acknowledges the nature of a research study as a complex social process that yields knowledge within practical contexts.

| Context
The goal of our two-year research project, entitled 'PPI in 10 palliative care research projects' was to strengthen sustainable PPI in palliative care, including research, education, and practice (see Table 1).
Our project consisted of four components. Component I centred on the training and coaching of patients and the public to prepare them for participation in research. In collaboration with an experienced patient research partner, we provided three regional training sessions to meet the needs of the patients and the public. Component II focused on the tailored coaching of the 10 research teams, over 12-18 months, to facilitate the implementation of PPI. In total, we held 35 coaching meetings, which each lasted 1.5 hours, in which we worked interactively through serious gaming, using the 'participation game' and 'participation matrix' developed previously. 11 Serious gaming supports the dialogue about PPI and can be used in all phases of a research project. The 'participation matrix' was used to visualize PPI and to document the design, planning, evaluation, changes, and contextual influences that both support and hinder PPI during the lifetime of a palliative research project. Component III involved the development of a community of practice. We organized four meetings, at which the involved patients and researchers shared the challenges and opportunities they encountered during their projects. Component IV was the qualitative evaluation of the perceived impacts of PPI implementation.

| Participants
The participants in the 10 projects included researchers (sometimes referred to as healthcare/educational professionals) and patients, caregivers, patient representatives, and patient advocates, which will collectively be referred to as 'patients' for the sake of readability. The only inclusion criterion applied in our study was involvement in one of the 10 projects. A total of 17 patients participated in this study, three of whom were male and 14 of whom Representatives of a regional elderly association played two roles: 1. Representatives participated in the project group and were involved in every phase and step as partners.
2. Representatives gave advice on the pilot testing and the pre-and post-measures and were consulted to validate the results.

Project 3
The goal was to integrate advanced care planning in the care for intellectually disabled people, in personal care plans and communications between clients, family members, and healthcare professionals. Patients not in the palliative phase were involved as participants in the interviews in Phase 3. Family members were involved as participants in the interviews in Phase 2 and 3. Three of them were also members of the advisory council and gave feedback on the research process, participated in co-creation and implementation (Phase 4-7). Members of institutional client councils were informed of the progress and (preliminary) results of the project.

Project 4
The goal was to implement pro-active care planning in elderly care in a regional consortium consisting of 18 cancer networks. Phase 1: pro-active care planning was explored.
Phase 2: training of healthcare professionals in proactive care planning. Phase 3: studying the patient files of patients who were treated by healthcare professionals who followed the training. Family members of deceased ICD patients participated in the survey (Phase 3). Members of the patient and family members/ caregivers advisory council of the regional consortium on palliative care and the representative of the national ICD patient organization were involved as members of the project group in all phases of the project. Their task was to advise the researchers. The representative of the national ICD patient organization also gave advice on the research design during preparation.  Representatives of non-Western patient organizations and non-Western migrant national networks played several roles: as equal partners in the project group and as advisees in the advisory committee (advice on the development of the educational meetings, including the information package and workshops, and commented on the final research report). As research participants, they participated in the interviews and focus group discussions (Phase 1). As implementation fellows, they were involved in the co-design of the educational meetings. They co-developed and implemented the train-the-trainer course. Those who received the train-the-trainer course organized and implemented the educational meetings as workshop leaders.

10
The goal was to obtain insights into the current practices of palliative emergency care provided by family physicians and the development of implementation strategies. Phase 1: exploration of current care by literature, interviews, and focus group discussions.
Phase 2: development of a national guideline and implementation plan.
Phase 3: implementation of training materials by the Dutch Association of Family Physicians.
Patients receiving palliative care and family caregivers were involved in data collection by participating in interviews. Patient representatives and patient advocates (Zorgbelang) were involved in the sounding board meeting of the regional consortium on palliative care during the preparation of the research proposal. They were also involved throughout the project, as members of the advisory council. They provided advice on data collection and the development of the guideline.
Note: Patients are people who are in need of palliative care services. (Family) caregivers are those who are or were involved in providing palliative care, such as spouses, children, friends, neighbours or volunteers. Patient representatives are linked to a local patient group or organization. They represent the stakes of specific patient groups or people in need of palliative care and/or have experience in palliative care, such as family caregivers/family members, etc. Patient advocates are staff members of an umbrella organization, operating at the provincial level. Members of institutional client councils are generally family members or residents of a specific healthcare institution. Self-nominated person advocates are based on personal interest, without being members of a patient organization or institution. Patients with interest in palliative care are chronically ill individuals who require health care, are linked to an umbrella organization operation at the provincial level, and advocate for people in need of palliative care. In the article, we refer to all of these groups as 'patients' and distinguish only where necessary.
TA B L E 1 (Continued) than 2 years research experience), and 21 were senior researchers.
Their research experience ranged from 0 to more than 20 years.
Nine projects were performed at five universities throughout the Netherlands, and one was performed at a knowledge institute with close university ties.
We used two sampling strategies: self-selection for projects and purposive sampling for patients involved in the projects. The senior researchers associated with each project recruited patients for this study.

| Ethics
This project was reviewed by the ethical review board at Zuyderland -Zuyd University (16-N-108). Participants signed informed consent forms. Participants received verbal and written information describing the goals of the study and how confidentiality and anonymity were assured. We asked the interviewees about the perceived impacts of PPI, and the barriers, facilitators, and strategies that strengthened PPI. We used the 'most significant change' interview method (IM, AvdH). 12 We used an open approach, allowing the interviewee the opportunity to decide their stories. 13 The interviews lasted from 30-90 minutes, and all interviews were audiotaped and transcribed verbatim.

| Data analysis
In the analysis, we applied the model of organizational culture. 7 Machado described a model with two axes: one axis ranging from

| Trustworthiness
Member checks were used to increase the credibility of the study.
We sent the participants summaries of our analyses for feedback.
We also used data triangulation (the literature, the experiences of patient representatives, advocates, and researchers), methodological triangulation (observations and interviews), researcher triangulation (analyses by different researchers), and peer debriefing (discussing the applied codes). To guarantee the external validity, we used thick description. We described the context of the project, the project characteristics, and the selection of participants. 15

| Patient and public involvement
Our project group consisted of three researchers (two junior and one senior), one patient advocate (a staff member of Zorgbelang Limburg, which strengthens patient and public involvement in the province of Limburg, the Netherlands), and two patient representatives, who were also affiliated with Zorgbelang Limburg. One was the chair and founding member of the regional Patient Sounding Board Palliative Care and has extensive experience in patient representation. The other has been involved for several years with a volunteer-managed hospice. He had no previous experience in PPI in research but brought experience as a member of various committees from his professional past. The project was based on coleadership and co-responsibility, and PPI was implemented at every phase.

| Participation cultures
We identified three culture types: relationship cultures, task cultures, and control cultures. We will first discuss the most participative PPI culture and finish with the most traditional PPI culture. We use pseudonyms to ensure the anonymity of participants. 'If someone shines, one's motivation will be biggest and one's participation most impactful' During Project 8, the task division between the researchers and Lisa, the patient representative, was also clear. Lisa was asked to provide a critical eye to review the work performed by the researchers.
She did not receive pre-defined tasks but experienced the freedom to take a pro-active role. For instance, she presented the idea of performing a pilot interview and arranged an interviewee for this purpose.
Her ideas were taken seriously by the researchers, who gave her the opportunity to contribute freely.
In contrast to the relationship culture type, the contributions that the researchers expected from Halima and Lisa were not to provide a lay or 'authentic' perspective, based on the actual experiences of people requiring palliative care but, instead, to provide their critical perspectives, based on their professional experience as advocates.
Halima was a professional trainer of migrant groups, whereas Lisa was a former social worker and a member of the regional patient and caregivers advising council.

| Control cultures
We defined control cultures as based on control exercised by researchers, who acted as the traditional powerholders and the responsible individuals for the project. They invited patients to participate in roles defined by the researchers.
The third PPI culture type, which is also the most traditional type,   In task cultures, the patient representatives also impacted the project relevance because they were professionally close to people who required palliative care and were able to advocate for their stakes. In control cultures, the patient representatives might not have a pro-active say in the project aim; however, the researchers do need to be able to legitimize their project to the patient representatives. Therefore, the researchers remain obligated to consider the patient perspective, at least to a certain extent.

| Increased methodological reliability
In all three PPI cultures, the projects have become methodologically more reliable because the involved patients provided a lay perspective on the materials used, such as recruitment letter and survey questions. They speak the same language as the intended audience.
Guy, a researcher working on Project 5, added that a personal relationship between the patients and the researchers was essential to allow the researchers to feel comfortable inviting the patients to discuss methodological issues.
Halima, working on Project 9, previously trained many patients with migrant backgrounds. She mentioned that she warned the researchers to avoid including discussion topics that are considered 'taboo' in certain cultures. During Project 2, Willy proposed to perform a pre-test for the questionnaire and arranged test-persons for this purpose, which improved reliability. Dorry (Project 7) was able to recruit many people due to her social network with the patient organization. Patients were more willing to participate when someone 'from inside' asked them to do so. Hanna, a researcher working on Project 7, noticed that interviewee recruitment was much easier when working with patient representatives who have close ties to the target audience, resulting in a more diverse sample.
In control cultures, the researchers mentioned an impact on the methodological reliability, although the patients themselves often state they are unsure whether they have made any impact because they are not informed by the researchers how their feedback is used.
Sometimes, the researchers in this project culture can neglect opportunities. For instance, Jet, the involved caregiver in Project 3, recalled an instance where the researchers showed her film recordings of interviews that had been performed with clients presenting mental disorders, regarding the ends of their lives.
'The clients were interviewed by their own healthcare professionals. On the one hand, that is, of course, a relationship of trust. On the other hand, that is also an inherently unequal relationship. The language used was also too complicated, and questions were suggestive. But they showed the film only after they had conducted the interviews. I was not consulted before, so I could not have any impact'.

| Research products were better able to reach the public
In the relationship and task cultures, the participants also perceived that the final products were better able to reach the public. Patients feeling invited to take pro-active roles were key to this impact.
During Project 5, Maria, the patient representative, was involved in the implementation phase and wrote the press release. Dorry   Arguably, the gap between researchers and patients is larger during the palliative phase than during other healthcare stages, which increases the potential impacts that PPI might have in the palliative context.

| D ISCUSS I ON AND CON CLUS I ON
These findings provide us with a better understanding of the relationship between culture and PPI impact. Although Shippee et al, 3 Manafo et al, 4 Chambers et al, 5

ACK N OWLED G EM ENTS
We would like to thank Manon Beckers-Houx, who actively participated as part of our project team as a patient representative but was unable to be listed as a co-author for this article. She has had a tremendous impact on the project. We would also like to thank all of the participants in our project for providing us with insight into the patient and public involvement and for sharing their experiences.

CO N FLI C T O F I NTE R E S T
No conflicts of interest.

AUTH O R S' CO NTR I B UTI O N
Inge Melchior has substantially contributed to the data collection, analysis and interpretation, and to the conception, design and writing of this manuscript. Anouk van der Heijden has substantially contributed to the data collection, analysis and interpretation, and to critically revising this manuscript. Esther Stoffers has substantially contributed to the conception and design of this project, its data collection, analysis and interpretation, and to critically revising this manuscript. Frits Suntjens has substantially contributed to the coaching of the 10 projects (our intervention), the data analysis and interpretation during our coaching period, and to critically revising this manuscript. Albine Moser has substantially contributed to the conception and design of this project, its data collection, analysis and interpretation, and to the conception, design and revising of this manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.