Involving the public (patients, carers and the public) in health care research has been widely discussed and endorsed in English policy [1–3]. However, reviews of Patient and Public Involvement (PPI) impact and effectiveness suggest that there is a need to understand if different approaches to PPI achieve different outcomes [4–8]. In end of life care research there is some evidence that "bottom up" approaches focusing on value to the user may provide a framework for designing programmes to raise awareness of issues related to death and change behaviours [9]. This paper describes the involvement of members of a PPI group in a study that explored the views, expectations and experiences of living and dying of older people living in care homes. It considers how their involvement as co-researchers supported recruitment to a study on a sensitive topic, provided peer support and safeguarding for vulnerable participants, and informed how findings were interpreted both with the research team and the staff and residents of the care homes. It also discusses what the process revealed about power relationships within the research team and how public involvement is negotiated.

The involvement of the public in research with older people builds upon the wider work of peer facilitation/peer education where people respond well and identify with those in a similar situation or of a similar age [10]. The rise in user involvement in research is due to the emergence of participatory research that seeks to work 'with' people rather than 'on' them as subjects of research, political drivers for greater user participation, policies that reflects broader consumerist and democratic movements within society [11, 12]. It can involve consultation, collaboration or be user-led/controlled [13, 14]. Studies that have used PPI to focus on how older people anticipate end of life and their priorities for care have involved older people acting as peer educators whose role is to share information with people of a similar age and life experience [15–17]. PPI involvement in research on sensitive subjects, such as death and dying, can ensure that questions are framed to reflect the interests of their peers, develop a shared understanding of issues raised, and moderate the often unequal power relationship between researcher and participant [18]. This paper describes the involvement of older people as co-researchers in a study of living and dying in care homes. It considers the advantages and disadvantages of user participation in end of life and care home research

The Experiences of Older People in Care Homes (EPOCH) study aimed to understand how living in a care home influences, over time, residents' views, experiences and expectations of end of life care and symptom relief. To capture key experiences, events and change over one year, of older people resident in the participating care homes, a longitudinal, mixed method design was employed. Residents were interviewed up to three times and their case notes were reviewed four times over the year. Interviews were semi structured, recorded and fully transcribed. Issues and themes identified in the first interview were discussed in subsequent interviews to see if key events, or the passage of time, had affected how older people thought and talked about living and dying in a care home. Information about the deaths that occurred during the study was reviewed and an economic analysis of their prior service utilisation was carried out. Care home and health care staff (General Practitioners, Community and palliative care nurses) were interviewed once about their experience and views of providing end of life care and discussion groups were held with residents to discuss emergent findings towards the end of the year's data collection ([19] provides a detailed account of the study method).

The research study recruited 121 (47%) of the residents from 6 care homes (with no on site nursing provision) in three geographically disparate areas in England. Thirty care home staff and 19 NHS staff were also interviewed. Sixty three older people participated in up to three interviews over the year's data collection of which 23 (5%) of the residents died during the year. A detailed account of the study findings is provided in the final report [24].

The three PIRg members spent between 2-6 hours in the care home during each visit, and a total of approximately 80 hours in care homes over the course of the study. They made between 6 and 18 individual visits to the care homes. The PIRg members visited the care homes mostly at the beginning and end of the study. The PIRg members and researchers spent similar amounts of time in the care homes during the consent and recruitment phase. The researchers carried out all the data collection (interviews, care note reviews) and towards the end the PIRg members (accompanied by researchers) discussed the emergent findings with residents.

There were three areas where PPI involvement directly influenced the outcomes of the study. These were recruitment, governance and safeguarding, and the discussion and interpretation of emergent findings with the older people living in the care homes.

It would be misleading to suggest the public involvement in end of life research brings unalloyed benefit. It was time consuming and resource intensive to liaise with members and careful planning was necessary to prepare for PIRg participation and ensure there was a common understanding of the study purpose, the contribution of different members of the team and the particular sensitivities of talking about dying.

For the research team, research deadlines, adherence to the protocol as funded, and responsibility for the research study as a whole meant that PIRg members were not always treated as equal partners. The project was situated within the participatory as opposed to the emancipatory paradigm and hence the overall control remained with the researchers [25, 26]. Many researchers feel more comfortable in retaining control and working within a professionally defined construction of public involvement [27]. Beresford [11, 28] argues that there is an innate fear on the part of many researchers that service user and public involvement may reduce the rigour and reliability of research findings. While the researchers involved in this project had a strong commitment to public involvement in research, there were preoccupation about the impact of potential co-researchers' values on the research outcome that were not applied to other members of the research team. This perceived/presumed dichotomy between co-researcher as value-laden and emotive whilst the researcher is objective and unbiased is not helpful, not least because the researchers bring their own set of values and emotions to a study [29, 30]. The PIRg member who was excluded from visiting the care homes, considered she was as capable of being objective and unbiased about the methodology and findings as the researchers.

Within participatory research negotiation between the researchers and service users or the public can be one of the most challenging issues to face but there is little research on this aspect [31]. One practical way of addressing this is through a formalised agreement about roles and expectations [32]. In the EPOCH study there was no formal job description for the PIRg members and there was a risk that despite the honorarium and benefits in kind that PIRg members received, the study exploited their good will and interest. It was difficult to know how much time to ask a member to give to the project. These issues highlights the value of project specific role specification for PIRg involvement, the importance of regular review, and the need to develop open working relationships prior to a study's commencement.

There were also logistical challenges. PIRg members were available for varying amounts of time due to personal and other commitments. The researchers sometimes found it difficult to accommodate their needs; because care home visits could not always be organised with sufficient notice or could be cancelled by the care home manager with minimal warning.

This was an emotionally demanding study. The researchers and PIRg members agreed to complete a reflective diary after each visit to a care home, although, not all members did this. Mechanisms had been put in place to support the research team and the PIRg members, there were debriefing meetings to discuss areas of concern and matters that the research team had found difficult to address and PIRg members could access extra support independent to the study if wanted. Participation triggered memories of family members who had died and whose experience of care homes had not always been good. For two PIRg members a desire to make a difference in this area of practice had been a key motivation for their involvement.

The study was essentially exploratory, and while recommendations for future practice were based on the findings, the study itself did not result in any immediate changes in patterns of care or service provision. It was a source of frustration, for some members of the PIRg and research team that research may not directly influence practice in care homes when it was apparent (to them) that change was needed. In contrast one PIRg member considered the research experience had given her hope and challenged previous bad experiences of care home life. The care homes in this study had staff who were interested in improving the end of life care they provided.

This paper examined some of the practical implications of having PIRg members involved in end of life research in care homes. Their involvement included but extended beyond involvement at a distance in the review of research questions, documents and findings.

The recruitment period required considerable time and input from the researchers and from PIRg members resulting in a successful recruitment process that exceeded response rates of other studies in care homes [33], while also safeguarding residents during the conduct of research on potentially sensitive topics. There were power differentials that, despite the research team's efforts in involving the PIRg throughout the research process, persisted. It has also illuminated how the evidence created in a project is as much shaped by social relationships in the research process as the actual methodologies [34].

Future studies would benefit from developing a job description for PIRg members and a more formal means of setting out respective expectations. Future research may wish to elicit the views of focal participants in such studies and other stakeholders about the mediation of research by PIRg members. In this study these groups would include residents but also their family members and the staff of the care homes.