“About sixty per cent I want to do it”: Health researchers’ attitudes to, and experiences of, patient and public involvement (PPI)—A qualitative interview study

Abstract Background Funders, policy‐makers and research organizations increasingly expect health researchers in the UK to involve patients and members of the public in research. It has been stated that it makes research “more effective, more credible and often more cost efficient.” However, the evidence base for this assertion is evolving and can be limited. There has been little research into how health researchers feel about involving people, how they go about it, how they manage formal policy rhetoric, and what happens in practice. Objective To explore researchers’ experiences and perceptions of patient and public involvement (PPI). Methods Semi‐structured interview study of 36 health researchers (both clinical and non‐clinical), with data collection and thematic analysis informed by the theoretical domains framework. Results In the course of our analysis, we developed four themes that encapsulate the participants’ experiences and perceptions of PPI. Participants expressed ambivalence, cynicism and enthusiasm about PPI, an activity that creates emotional labour, which is both rewarding and burdensome and requires practical and social support. It is operationalized in an academic context influenced by power and incentives. Discussion and conclusions Researchers’ experiences and attitudes towards patient and public involvement are a key factor in the successful embedding of involvement within the wider research culture. We call for a culture change that supports the development of effective organizational approach to support involvement.


| BACKG ROU N D
Patient and Public Involvement (PPI) in research is increasingly expected by funders, policy-makers and research organizations. This is an international movement; in some countries, such as Canada and the USA, the term engagement is preferred to involvement, but the intention is the same, namely to do research "with" people rather than "to" or "for" them.
The UK's Chief Medical Officer has asserted that PPI can make "research more effective, more credible and often more cost efficient." 1 One systematic review 2 found evidence of increased recruitment and retention, as well as some evidence that patient involvement helped in securing funding, in designing study protocols and in selecting relevant outcomes. A systematic review specifically on recruitment and retention found that PPI interventions had a modest but significant effect. 3 In primary care, Blackburn et al 4  However, the practice of PPI and the evidence base is still emergent, and the quality of impact reporting is inconsistent. 5 Considerable debate remains as to what can and should be measured, and how we define PPI as an "intervention." 6,7 Fundamental differences in understanding of the goals and value of PPI result in very different approaches to understanding impact-is it a technical process to improve the effectiveness, acceptability and feasibility of medical research, or is it rather about democratic rights and empowerment, or even an essential challenge to what constitutes "evidence"? [8][9][10] Qualitative research can provide insights into how PPI is implemented (or not) and how it may affect research, 11,12 but it cannot give us a definitive account of impact. Staley, et al 13 have argued that we may be better considering impact on researchersover time, on their behaviours, emotions and ways of thinking-than on specific pieces of research. Thus, health and medical researchers are exhorted to implement PPI as a way to improve the quality and effectiveness of their work, based on a, as yet, somewhat limited evidence base. This may be felt to be in tension with a predominant emphasis on evidence-based medicine, and the expectation that interventions should be rigorously evaluated before widespread implementation. Researchers' beliefs and behaviours are of course also fundamental to whether, how and at what stages of the research cycle PPI takes place. The cycle of tokenism presents a risk: researchers who are unconvinced about the benefits involve people in a half-hearted or superficial way, reinforcing their belief that it makes little difference. 14 There have been few studies of how researchers have responded to the pressure to move towards a more participatory model of knowledge production. 15 The privileging of scientific expertise over lay experience has been remarkably resilient, and it has been noted that researchers are reluctant to cede control over research. 16,17 Despite some interest in, and support for, PPI in theory, researchers may feel apprehensive and reluctant to change their practice, given other professional and institutional priorities. 18 Boaz et al's 15 interview study with 19 biomedical research centre scientists found little evidence of any change in these attitudes and concludes that "science remains the preserve of scientists, with patients and the public invited to 'tinker at the edges'." (p.592).

A focus group and interview study of the perspectives of 24
Australian researchers 19 is one of a small number of studies to focus explicitly on researchers' experiences of PPI. The report identifies a mix of practical barriers to involvement (such funding time and contributor involvement, lack of skills and difficulty finding people) as well as more relational issues, including anxiety and defensiveness; fear of antagonistic or "difficult" patients; and power hierarchies both within academia, and between academics and patients. Thirty-six participants were recruited through a wide range of avenues from across England, Scotland and Wales. We aimed for a maximum variation sample, 21 seeking variation across types of research (clinical, medical scientists and social scientists); research design (eg researchers using trials, cohort studies, qualitative methods); levels of seniority (from early career researcher to principle investigator); degree of experience of PPI; and age, gender and ethnicity.
We also included some people working as research managers or involvement coordinators. Table 1. Participants were interviewed at a location of their choice, usually their academic workplace. Interviews were video-or audio-recorded, depending on consent, and transcribed verbatim.

| THEORE TI C AL LEN S
The implementation of PPI requires researchers to change their existing behaviours and practices and, in the case of new researchers, adapt to this relatively new practice in a culture that is also adapting. There are numerous frameworks for seeking to understand behaviour change and organizational implementation. We adopted Michie et al's 22 theoretical domains framework (TDF) to guide both data collection and thematic analysis. Its focus on professional behaviour, evidence-based medicine and the adoption of new practices seemed particularly relevant to this study. We drew on, but adapted, the suggested questions to develop prompts for our own interview guide. During analysis, we also adopted a micromeso-macro level lens, incorporating wider factors into our developing understanding of implementing PPI, that is we looked at individual, organizational and broader cultural factors (eg funders).
Interviews were semi-structured. Participants were initially invited to talk about their current role and how they had first started involving patients and the public in their research. Prompts included both invitations to give further detail about how and why they had involved people, and prompts derived from the theoretical domains framework. In practice, some prompts were dropped or merged (Box 1).
The data were analysed using an iterative thematic approach, 23 supported by data management software (NVivo). Data were coded and categorized using a framework that was developed deductively based on the TDF, but we also included inductive codes that arose from the interviews. Further inductive analysis of the TDF coding was conducted to ensure that we looked beyond it and into the views of the participants. This further analysis was used to generate themes that encapsulate participants' experiences. In Table 2 below, we present the themes and demonstrate what domains of the TDF were used in generating them. Table 2 presents an overview of the themes and the domains of the TDF that were used to construct them.

| Practical and social support
Patient and Public Involvement required significant administrative labour. Participants talked about the lack of practical support to do this work, and the time and effort diverted from other activities.
Embedding PPI in practice was described as a complex process.
It required a range of resources, including time, money, space to interact with contributors, skill development and refinement, and an infrastructure to support PPI (including PPI coordinators). Generally, no additional researcher time was allocated for the work associated with involvement. This raised questions about its value and importance. I'd love admin support just so that we had everything streamlined… not me searching around for the right tailored information that they would need and typing it up myself. So that sort of thing would be really, really good but I will drop things deliberately to deal with PPI because it is incredibly important.
Maria, Senior lecturer and researcher, age withheld, CMSR Participants talked about the value of PPI coordinators, staff whose role was to support contributors and facilitate the process of PPI:

| Rewards and burdens of emotional labour
In addition to the time costs outlined previously, participants dis- This emotional labour and its associated costs were seen as necessary, important but also rewarding. Participants described how it had a positive impact on them emotionally as well as on their work:

| Academia, power and incentives
The "publish or perish" culture of academia was highlighted fre- In addition to the association between PPI and being junior and on fixed-term contracts, participants observed a gender divide in terms of who bears the administrative and emotional labour of PPI.
Not only was there an observation that women were usually responsible for PPI, but the participants suggested that this might be due to a wider perception about the skills needed and the value of PPI.

| D ISCUSS I ON
This study explored the implementation of PPI from the perspective of health and medical researchers demonstrating that it is a highly complex undertaking that can be both beneficial and burdensome.
Its implementation is affected by a range of micro, meso and macro level factors; it is operationalized within a "publish or perish" culture that can result in it being deprioritized, causing tension for those who see its benefit. Attitudes towards PPI range from cynical to ambivalent to positive. It is often the responsibility of women and junior staff, and is particularly complicated when it lacks support from senior colleagues.
These findings are particularly relevant in light of Staley's 13 call for research on the impact of PPI to focus on the impacts on researchers. In contrast to previous research, 15,18 we found that researchers were willing to change their practice, but that this was complicated by the culture in which they operate and a myr- In debates about PPI in research, the dominant discourse is one in which researchers hold the power and contributors do not, or at least have to find ways to negotiate power for themselves in an unequal relationship. 24 The problem with this discourse is that it overlooks how hierarchical academic life can be, and how power may be unevenly distributed amongst researchers; between research disciplines; and between researchers and the organizations and processes which shape their lives, such as university employers, funding bodies, ethics committees and peer review. It also overlooks the possibility that researchers may feel that they have considerably less power vis-à-vis patients/public than is sometimes assumed. Equally, more junior researchers may feel that they do not have the power needed to ensure involvement activities are embedded in research, particularly if they are working with a more senior person who treats it as a tick-box exercise.
Our findings also indicate that responsibility for PPI is not equitably distributed, with female and junior researchers often being tasked with it. There was a sense that junior staff starting their careers were more open to the notion of PPI and felt it was part of the culture of academia. Drawing on the concept of "emotional labour," the fact that women appear to have more responsibility for it than men may be unsurprising. In 1983, Arlie Hochschild 25 coined the term to describe the work of managing emotions that was required by some professions, predominantly undertaken by women. It has previously been considered in higher education in the USA, 26,27 but not with regard to PPI, which arguably extends the emotional reach of academia. As our findings indicate, PPI involves a significant amount of emotional labour from taking care of contributors, to suppressing emotions, to operating in a context that deprioritizes involvement. Devolving responsibility to women may be linked to socially constructed notions of what it is to be a woman and the perception that they are more likely to be skilled in these tasks.
At a more structural level, however good the relationship with individual patients and members of the public may be, researchers may resent the implication that their expertise is devalued. This may be more threatening to professional identity in some fields than in others, reinforcing inter-disciplinary hierarchies. There was an erroneous perception reported here that patient involvement in qualitative interviewing and analysis is uncontroversial, whereas it is not generally assumed that patients' lived experience qualifies them to get involved in statistical analysis. The fact that such disciplinary hierarchies may also map (albeit imperfectly) onto gender compounds the challenge to professional identity.
Research leaders who acted as positive role models and created an environment where PPI was commonplace helped reduce dissonance between "doing PPI" and "being a good academic." In other contexts, researchers who chose to pursue PPI very actively might feel they were going against the tide and damaging their careers.
Others might feel that their career was being damaged for them by senior researchers giving signals that PPI was an unimportant task, and by extension, the people in the organization to whom they delegated it were also unimportant. Our findings suggest that this view has not been entirely overcome.
Pressure from funding bodies, and especially a strong steer from the NIHR, has undoubtedly fostered a climate in which PPI is seen as a requirement rather than an option. However, it is questionable whether this should be understood as truly "game-changing," or rather a stimulus to "gaming," in which ticking boxes are literally as well as metaphorically required. The truth, as ever, lies probably somewhere in between, whereby researchers are genuinely making efforts to do more and better PPI, but at the same time finding it irksome and not always expecting to live up to the promises made in grant application forms in order to get the funding.
There were clear examples in our data of people holding very critical views of PPI but not feeling empowered to voice these publicly. In reality, PPI may be seen by researchers and funders alike as unlikely to be a deciding factor in funding decisions, in contrast with scientific criteria. This in turn further affects how researchers are seen within their organization if they are perceived to be spending too much time on PPI.

| Strengths and limitations
This is one of the first papers to draw on the TDF and implementation literature to look at PPI in health and medical research. As such, it considers researchers' views on if or why they should involve patients and public, and on how they do it and the range of factors affecting adoption. The TDF is a comprehensive framework for understanding meso and micro factors in implementation. However, it excludes the macro level, which is problematic for the implementation of PPI as wider socio-political factors are key in this debate.
Nonetheless, we used a micro-meso-macro approach to analysis to ensure we explored these crucial factors.
The TDF proved to be a useful lens in exploring the implementation of PPI. In analysis, it provided helpful direction and we did not consider it unduly limiting, as we were guided by it rather than following it deductively. On a practical note, we found it difficult to implement as an interview guide and had to modify it, as detailed in Box 1.
The design of this study allowed participants to freely express their views on involvement, leading to new findings about researchers' experiences of implementing PPI. Our sample was limited to participants working in England, Scotland and Wales and, although we achieved a good range in terms of career stage and length of experience of PPI, we were unable to recruit equal numbers of female and male staff. In spite of these limitations, key findings from this work may be transferrable to other settings.

| Implications
From our analysis, it is clear that more support is needed for PPIpractical support, including funding for PPI input to develop bids (which is not always currently available); time, including longer deadlines for funding calls and more dedicated assistance (eg administrative support) with conducting PPI activities. Intragroup and intergroup support in the form of researchers sharing practice within and across organizations is also vital. We call for a cultural change to challenge the tick-box approach that may have resulted from making PPI an operational requirement. Much like the Athena Swan initiative that is attempting to improve academic culture for women in science, an initiative to embed PPI could reinforce its value and promote it to a more prominent and important role. Some evidence of attempts to challenge research culture has been made in recent years, including publications focused on how to involve contributors in clinical trials 28 and qualitative research, 29 and in co-producing their involvement. 30 However, what is needed is a dedicated initiative aimed at solidifying involvement as part of research culture.
Such an initiative could draw on existing training and should include shared learning offered by academics and PPI contributors; sharing practice and experiences is a valuable form of training that might make PPI more accessible. Such a scheme may have a positive effect on culturally embedding PPI.

ACK N OWLED G EM ENTS
We are grateful to the participants for sharing their experiences to inform this research. This research was supported by a panel comprising patients, members of the public, AND health, clinical and medical researchers. We are grateful for their support.

CO N FLI C T S O F I NTE R E S T
The authors have no conflict of interests to declare.