Improving Research Inclusion: learning from NIHR and Research Council funded studies in England

Patient and Public Involvement

In this paper we summarise the learning from these four projects, rather than report fully the PPI arrangements of each project. However, we briefly describe the PPI approaches here, and we consider the relationship between PPI and RI (see Table 1). In ATTUNE young people with lived experience participated in the funding application and design, and in the delivery of each component, including deliberations on ethics, and in the interpretation and writing up of results. They were represented in the overall advisory board, local and national advisory groups, and within specific research gatherings and interpretation processes. In Co-Pics and the related PhD project (both still being written up) on multiple long term conditions including psychosis, people with lived experience reviewed and contributed to and were involved in the application for funding, in the data collection, and will be involved in interpretation and writing up. In Co-Pics the advisory board included lived experience and a co-investigator on the funding application was a lived experience expert. In the BRC themes on Preventing Multimorbidity and the Youth Development, we worked with lived experience experts with the relevant experiences and conditions to inform how we should undertake research; in these studies we devised the engagement and involvement plans after making contact with potential participants, centring on flexibility and local and group specific adjustments in our approach. Lived experience experts were and will continue to be involved in interpretation of the data. The projects mostly followed an experience-based co-design methodology, so were not experimental studies or trials. The co-designed interventions are being or will be evaluated in the future. Where we used outcome measures in ATTUNE, which were recommended by the funder to promote comparison across studies, they were also reviewed and critiqued by young people’s panels, to inform selection, modification and presentation of questions. For this paper, we did not establish a separate PPI group, although we shared a summary of the paper for input and our checklist was scrutinised by a panel of five young people, who ranked the items in order of importance.

Agreeing concepts, vocabularies and evidence

The conduct of research is often shrouded in technical terminology, either reflecting specific methodological paradigms or disciplinary framings of the problem that the researchers seek to solve. Esoteric or technical language is sometimes necessary, linked to theory to help us advance science, or for communication between people sharing the same theories, concepts and vocabulary. However, language can act as a barrier to inclusion.¹⁹ For example, on interrogating concepts like personality disorder, disease or illness, or ethnic group. Agreeing an accessible vocabulary and explanations of causation often reveals unclear ideas or inconsistent definitions amongst scholars.²⁰ Such questioning helps improve the precision of the language used and exposes traditions and conveniences that must be revised. These negotiations ensure partnerships are grounded in shared concepts and meaning from the start, and that there is agreement on how words that carry power are to be used. This enables influence over the research to be a reality from the beginning.

Linked to language, once research is completed, the findings should also be accessible to the public, in a range of appropriate visual, audio, and sensory formats, beyond academic publications. For example, arts installations and exhibitions are an excellent vehicle for sharing research findings and for influencing practice and policy. We have provided photovoice exhibitions and arts exhibitions for the public and policy makers for ATTUNE, sharing young people’s art, soundscapes, animations and film outputs. Accordingly, we propose, RI can be more impactful if it encompasses not only study design and participation approaches, but also intentional designs of how the findings will be shared for wider knowledge dissemination into relevant communities and stakeholders. This adds value to the participants’ trust and enhances agency and ownership of their creative outputs and data (e.g. in ATTUNE and Co-Pics, respectively). We found this also helps to improve policy engagement, given creative outputs are known to be emotionally activating, more easily progressing to perspective-taking and consideration of options for future traction; rather than the traditional approach of providing research findings and expecting them to be valued and actioned by others facing multiple implementation barriers. Personal sharing of findings by potential beneficiaries also carries more weight. As another example, many young people in ATTUNE presented to parliamentarians and produced briefings for parliamentary consultations, informing emerging policy in real time.

RI seeks to democratise knowledge and show what research is being done for public accountability. Greater involvement of those who would not usually participate in research can bring benefits to them.²¹ In our work with marginalised groups and with professionals delivering health and social care, it is always surprising how little of the research evidence is known. Therefore, any prior assumptions of what is known or not across participants in co-design and participatory methods should be questioned. Providing evidence summaries early in the process, again in multiple modalities, can help balance power as participants can challenge the prevailing evidence, what it means, and what it is based on. This helps to optimise trust and communication, and promotes early discussion; such a process also sets out shared and non-shared basic assumptions of group process, including exploring communications and collaboration and complex interdisciplinary dynamics.²²

Trust, power and participatory co-design

The design and delivery of research, and the ways in which research findings are shared, when informed and shaped by the beneficiaries of research, ensure that the presented findings are relevant, ethical, and more likely to be adopted for practical impact.^23,24 Epistemic injustice refers to people’s views being dismissed because of a specific identity characteristic such as a diagnosis, their identity (for example, ethnicity), or both (testimonial injustice). Their views might also be dismissed through biases in knowledge production methods leading to a failure to produce fair and representative accounts (hermeneutic injustice).²⁵ Through dialogue, with participants and intended beneficiaries and co-designed collaborative processes, knowledge production can address epistemic injustice. This should ensure that the public and participants come to trust the research process and consequently the later findings.²⁶

Participatory RI can identify research questions of importance to the public and patients, alongside refining or challenging those proposed by funders and professionals. However, an equally important function is to critique or to deprioritise questions or procedures that might reinforce stigma and prejudice, or halt research interests that are too far removed from patient experience. Such critique can prevent wasting resources on unhelpful approaches.

A shift to find solutions to research problems (agree vocabularies, recruitment, representation, ethical dilemmas) during the research requires time to explore and progress and may be viewed as demanding and costly. Therefore additional resources (time and funds) should be anticipated in the research design and funding application.² Consequently, implementing RI can be experienced as disruptive of conventional research pipelines, but this is necessary to shift power away from research teams and standardised procedures of exclusion, towards the interests of beneficiaries. Given the risk of epistemic injustice for marginalised groups, championing their needs and advocating for better RI should fall to the research team. We found a re-think was often required in response to poor recruitment; hence, pausing at these moments offered valuable opportunities to review research designs and underlying procedural, practical, philosophical and theoretical confusions and assumptions. Cultural and linguistic adaptation, for example, may be revealed to require more intensive effort. Lived experience perspectives often helped uncover these assumptions and provided guidance on when the adaptations we pursued were adequate.

In our approach to inclusive research practice, we made use of participatory methods, emphasising co-design and collaboration with the intended beneficiaries from inception.²⁷ However, participatory methods are seldom used or fully reported in health research, and even when they are reported, this may not permit replication given there are no widely agreed reporting standards.²⁸ Co-designed and collaborative research contrasts with traditional approaches found in hierarchies of clinical evidence (trials and systematic reviews being considered the most valid approaches); funding priorities and study designs are usually determined by governments, think tanks, charities, and learned societies and researchers based on these traditional hierarchies. Such an approach, unhelpfully, locks in a particular design, rather than allowing researchers to respond to the ongoing and necessary re-design and newly discovered support needs of potential participants to ensure meaningful participation.

Another benefit of co-design and participatory methods is that these enable those undertaking research to reflect on and adjust research procedures and create a team culture in which lived experience is central to discussions and processes. For example, in ATTUNE, young people impacted by adverse childhood experiences were involved in the creation and interpretation of research findings with the investigators. In Co-Pics, peer researchers received skills training, undertook biographical narrative interviews and will be included in analysis and interpretation. In ATTUNE, we also adjusted the way advisory board meetings were run, including briefing young people before the meetings, providing accessible summary materials before and during meetings, and using fonts and colours that maximised accessibility to those with reading difficulties or neurodivergences. An additional benefit of participatory approaches is that those expected to implement findings and wider public stakeholders become more informed of what matters to the intended beneficiaries, and what really is at stake for them.²⁹ This prepares and motivates them for implementation. For example in ATTUNE, creative art was emotionally activating for both the research team and policy makers; and at the same time it presented young people’s experiences, centred on their agency, and in a positive manner, emphasising their creativity as a strength. Moreover, co-design reduces power imbalances as stakeholders become more familiar with working alongside each other and directly with the intended users of the research. This prepares them to champion the implementation of research and anticipate which local regulatory and governance procedures need activation.

Wider structural barriers that prevent use of health and social services can be replicated in research institutions. Innovative models can relocate research processes into community venues, ones that are more familiar to and trusted by participants; however, this means improving the infrastructure for research in the community in recognised safe spaces such as schools, charities, museums and art galleries.³⁰

Ethical issues

Research inclusion often brings structural and procedural barriers to the fore, generating ethical questions and dilemmas when designing and delivering research. Adopting RI means upholding an ethical responsibility to ensure justice and fairness in research, where all groups have an opportunity to benefit from advancements in science, rather than a privileged subset.

Globally, ethical dumping (the practice of researchers from high-income countries conducting unethical or exploitative practices in lower income countries) and helicopter research (research teams entering and leaving communities after the research is completed without considering the impact on the community) are being challenged and must be rigorously avoided and managed.³¹

One way to combat ethical violations is to ensure research processes to recruit participants and deliver research are developed with and subject to critique by intended beneficiaries.³² Oversight committees need to be attentive to potential unanticipated ethical dilemmas and see these as an opportunity to learn. For example, in ATTUNE, we designed participatory methods involving multiple modalities of creative arts and communications (verbal, audio, video, sensory) and different approaches to co-design; this was time consuming and required thoughtful and sensitive engagement with the intended beneficiaries throughout. This posed many ethical questions around wording and presentation that were addressed through participatory deliberative processes.³³ These were especially helpful if there were sensory challenges or difficulties with literacy or verbal ability.

Such approaches may be less familiar to funders, researchers and ethics committees.³⁴ Therefore, an important task is to explain and upskill such committees³⁵; this is additional work that researchers must undertake and so the timelines and costs should be considered for progress milestones to be realistic and achievable.

There are many other examples of how research needs to be adapted to respond to ethical questions. Adequate support is required for people to enter research, especially if potential participants are frail, vulnerable, or have experienced lifetime adversity and unfavourable socio-economic conditions; this requires more time to assess what kind of support is needed and to build trust. Participants will need more support than is usual during and after the research. It is relatively easy, in our experience, for such groups to be systematically excluded by tight inclusion criteria or when the research infrastructures and recruitment teams do not have capacity to provide the additional necessary support. Judgements about suitability may carry inherent prejudices about an individual’s desire and ability to participate, especially when relying on busy clinicians or others working in health systems who are not directly involved in the research but are assisting with recruitment. These matters are often compounded by poor communication of the research aims and what participation involves, especially with novel participatory designs which are open to misunderstanding and misinterpretation. Most research activity follows standardised operating procedures especially for trials, therefore novel co-designed and participatory designs may be questioned or mistrusted by potential participants and generic recruitment teams. These too may unevenly impact specific groups, such as women, and require resources and time.

People experiencing poverty, mobility issues, disabilities, as well as multiple health conditions may require more time in travelling, flexibility in entering and leaving research, and resources (for example, taxis may be required rather than public transport). There may also be child-care needs to permit participation. These too need resources and time. People with carer responsibilities need support during the research process, especially if they are managing multiple social challenges like paying fuel bills, living in rural or distant areas, and if they need to prioritise immediate health and welfare appointments over research (for the person they are caring for or themselves).

For young people with care experiences or those living in vulnerable, chaotic or precarious situations, there are additional ethical complexities that require careful consideration. They may struggle with lengthy consent procedures³⁶ and may not wish to discuss research participation with parents or carers. Whilst researchers and regulators insist on assurances of uniform informed consent processes and of confidentiality, young people living in such precarity will be excluded, risking neglect of their significant health, welfare, and social care needs. In addition, if they have troubled and chaotic lives, young people may not be able to participate in tightly scheduled research, or in a consistent manner. Research should also not assume levels of literacy or ability to express experiences verbally; for example, many struggle to complete questionnaires that require significant periods of concentration, and ones which ask about deficits or traumatic exposures such as maltreatment that may be distressing. Depending on the timing or nature of the traumatic event, it may be difficult to recall. Again, the ethical dilemma is that such young people may be excluded, therefore, procedures and practices and ethical conventions may need to be revised. We engaged with social care organisation and local authorities who are usually responsible for safeguarding protocols and protections. The participation of social care leaders in research is a priority in the UK currently, yet this requires a larger infrastructure than is currently available.

The use of incentives in research is often questioned, but for participants from poorer sections of society with multiple responsibilities and few assets, remuneration should reflect the time burden, and the emotional and cognitive demands involved in participation (e.g. for travel costs and time). Providing cash incentives may help, but can also compromise welfare and benefits, if the rules of welfare receipt specify limits to additional sources of income. Remuneration should also take account of time away from work, carer roles, and childcare costs.

There are other examples of ethical challenges: young people may choose social media communications, or digital methods, which are often not deemed to be adequately secure and so these are not favourably viewed by data security, ethics or sponsor committees. Yet, if this is their preferred form of communication, it may improve representation, in part through building trust offering a more convenient form of easy communication. One of our studies used in NIHR OH BRC DIME used targeted social media adverts to reach a broad UK geographic region as well to recruit 300 participants in a few months (NIHR BRC), whereas for other studies (ATTUNE) this was deemed too risky and not supported. Consequentially, greater flexibility is required in governance procedures and protocols; greater (time) investment is needed for research teams to understand participants’ lives and how to adapt the pre-specified research infrastructure as a study proceeds. A phase of testing out different methods might helpfully be included in future research designs.

Governance and assurance policies can become constraining and prohibitive, reinforcing structures of disadvantage. Therefore, we advocate for all research organisations and employers to review their policies and practices, on an ongoing basis, and perhaps even for each study, whilst evolving new procedures. Watchful attention to potential for improvement should be welcomed, rather than seen to challenge or violate protocols and standard operating procedures. Research Ethics Committees, Advisory Boards and Data Monitoring and Ethics Committees (DMEC) are part of the current expected governance infrastructure in the UK and along with sponsors determine what is considered acceptable; we found that raising the challenges and issues outlined at these committees led to productive opportunities to re-imagine inclusive research practice as studies proceeded. Practically, revising procedures requires ethical re-review which can delay studies if the governance processes are themselves not sufficiently well resourced to reflect the demands of inclusive research. Research teams will need to investigate these processes and consider the parameters of what is possible within resources (financial, research capacity and infrastructure).

Integrating research inclusion and PPI

Patient and Public Involvement (PPI) is a widely accepted practice in health research, usually enacted through appropriate representation on advisory boards, in the development of new research ideas, sometimes with a presence in research teams, and through the design and refinement of recruitment strategies and materials to be put before participants. Resourcing PPI well, commensurate with objectives, can protect against tokenistic, extractive and exploitative forms of inclusion. Building trust takes time and researchers need to guard against potential for exploiting knowledge-rich communities. Despite the advancements in PPI frameworks, there remain structural and procedural barriers, as well as areas for improvement especially for populations who remain excluded from formalised approaches. PPI efforts can become professionalised and bureaucratised. For example, many PPI panels consist of members with similar health conditions and from similar social contexts without much consideration of intersecting vulnerabilities (e.g., race/ethnicity, sex, gender, neurodivergence, sexual orientation, socioeconomic status, disability) that shape health outcomes.^37–40 PPI needs to be further adapted to facilitate inclusion of those located at the intersections of multiple disadvantages related to identities, health challenges, and socio-economic position. This is where RI approaches are complementary, and distinct from PPI as a form of representation based on health condition or care experiences. Indeed, RI seeks to identify those who are likely to not be recruited to existing PPI procedures or into research, or who will fall out of research. RI remains open to discovery of more complex identities and situations which practically affect participation in PPI activities and research. Yet intensive PPI and RI processes that takes account of people who are overlooked usually end up being undertaken with insufficient resources. Furthermore, funder agencies may resist paying the additional expenses, even if the makes the proposed research fairer and more ethical, and even if it appropriately challenges historically ineffective or exclusionary processes.

We propose that RI is a specific extension of PPI to address the needs of ultra marginalised groups that may not even be known, and therefore not involved, at the start of research. RI promotes ongoing reflexivity and curiosity to improve procedures, so it strengthens PPI which may rely on and promote inclusion only through fixed mandated procedures.

We found that researchers should avoid inflexible ways of working in PPI. Well-intentioned institutional plans seeking efficiency can so easily become inflexible and bureaucratised.⁴¹ The research team should build in additional thinking time and opportunities to consider modifications adjusted for the most marginalised from specific localities. This includes self-awareness and agility in grasping and overcoming barriers and knowledge gaps as a core task.

Sometimes the pressure of preparing and launching a new research project against a strict timeline and working through all the contractual, ethical, workforce, and administrative stages can generate strain in the research team even if they have good intentions to make an impact. This needs care of the research team, especially if lived experience is prominent as planning, setting up and delivering research can be emotionally and cognitively demanding. This placed additional stress on teams; consequently, the PPI and RI elements can be compromised and neglected.

Members of the research team often carry many, potentially undisclosed, experiences of poor mental and physical health, and so the content of the research may be distressing and emotionally challenging for them too. Building in reflective debriefs, and careful use of holidays and health days to overcome fatigue at different phases of the research, alongside flexible working and tailored support can be helpful. A good research culture should be able to accommodate discussion, to guide relationships within and between participants and researchers. Each research programme, we argue, should have an explicit values framework guiding expected behaviours to respectfully resolve disagreements and demonstrate accountability when resolving dilemmas. External facilitation can help team dynamics also. We recommend all members of the research team, irrespective of discipline or experience, undertake research inclusion training relevant to the specific project on which they are embarking, using and testing the participatory methods; for example, we have made good use of the approach recommended in living and learning labs (practice-based workshops where people learn by doing). These ensure embodied and experiential learning rather than intellectual knowledge.

To improve RI elements in PPI, we drew on our collective experience of lessons learnt over several projects, and what we found helpful. From these we compiled a checklist as a set of standards derived from our work on various research projects. Our desire is for these standards to be adopted, but different types of research and specific tasks may require tailored refinement (e.g. cultural adaptation of psychological interventions or developing randomised controlled trials in low resource settings).

Lived experience perspectives

The strategies we developed evolved in partnership with project specific PPI panels, peer researchers, and community partners. As part of testing the value of our learning, we shared a summary of this paper and checklist with five young people through a survey. We did this in this manner to reduce burden to them in terms of soliciting their views, recognising the paper is written for researchers rather than lived experience experts.⁴² We asked this group to rank our recommendations. We only present percentage endorsement for 1^st and 2^nd ranked statements; see Checklist).

Limitations

While this paper draws on our experiences and perspectives delivering a range of UK-based research studies, we acknowledge that our co-authorship team may not fully reflect the diversity of research teams working in this space, nor the breadth of intersectional identities that are marginalised or seldom heard in inclusion research. Our collective standpoint inevitably shapes what we have noticed, prioritised, and articulated as best practice, and there will be experiences, tensions, and insights that may not be fully reflected. The findings might be considered relevant to UK research studies, although our populations varied by age, sex, gender, ethnicity, sexuality, neurodivergence, and by place (urban, rural, coastal). Each study sought out a specific population with commensurate inclusion criteria. Hence, our findings will need testing against study designs and governance and research infrastructures in different countries. Feedback from five young participants helped identify the top checklist items, yet similar processes will be needed for different research populations. We hope readers will test and iterate the checklist, whilst we do the same, in order to produce a standardised reporting framework. We offer this work not as a definitive account but as a contribution to an ongoing conversation, and we actively encourage dialogue, challenge, and extension from researchers, communities, and participants whose voices and lived experiences are not adequately represented here.⁴³