Building knowledge for policy and practice based on service user and carer experiences: A case study of Scottish adult safeguarding research

Summary This article presents Scottish adult safeguarding as a case study to illuminate some challenges of building knowledge for policy and practice based on service user and carer voices. It draws on five of our own research projects that have evaluated implementation of Scottish adult safeguarding legislation and/or asked more exploratory questions about risk, safety and support. Findings We show how practical and ethical issues limited our more evaluative lines of inquiry. We then show how increasingly participative approaches led to studies that were more accessible and that connected more deeply with service users’ and carers' lives, but that also faced greater challenges in the translation of their findings back into the policy and/or practice environment. Applications We conclude with an argument for ongoing dialogue between policy-makers, professionals, service users and carers, researchers, educators and students about knowledge, its different forms and sources, its generation and its use.


Introduction
Adult safeguarding has been the subject of legislation for more than 10 years, and of policy for more than 20 years across the UK. Adult safeguarding refers to the protection of adults from mistreatment or harm. Internationally other countries are debating how to respond as they recognise adult abuse as a policy concern (Montgomery et al., 2016). Yet the knowledge base for adult safeguarding is limited; and service users' and carers' voices are under-represented in research undertaken so far (Mackay, 2017;Wallcraft, 2012). Attempts to begin to plug these gaps open up a range of questions: in particular about the nature and facilitation of participation, and about the intersection of knowledge of different kinds with the policy-making process. The exploration of these questions is itself a work in progress in academic, policy and professional circles across a range of national contexts (Gregory, 2000;Hammersley, 2016;Newman, 2011).
This article offers Scottish adult safeguarding as a case study to illuminate some challenges of building knowledge for policy and practice based on service user and carer voices. We draw on five of our research projects with different methodologies and methods. We compare and contrast research that focused wholly or partly on the evaluation of practice within a given legislative framework, with research that asked some broader questions. We address the framing of research aims; access to participants; types of methods adopted; and how these all affect the nature of the knowledge generated. We conclude by offering some tentative learning, with applicability across policy fields, as well as to adult safeguarding itself.
First though we wish to broach the term 'service users' and its associated problems. We acknowledge that people who use services are not a homogenous group, and that the population of people labelled in this way includes people who 'use' services only with reluctance and/or under duress (Beresford, 2005;Smith et al., 2012). More specifically in relation to adult safeguarding, the term fails to reflect that investigations and some interventions can take place without express consent of the person concerned. This may include some people who could not be described as using services in any conventional sense: people who are reclusive, who might neglect their own well-being and who tend to avoid contact with health and care services (Braye et al., 2017). Nevertheless, we have chosen to retain the term as the least problematic one for now (McLaughlin, 2009).

Case study site
Adult safeguarding has been a defined area of Scottish law and policy since the passage of the Adult Support and Protection (Scotland) Act 2007. This legislation is concerned with the abuse, neglect and harm of 'adults at risk', defined as being 'unable to safeguard their own well-being' and 'more vulnerable to being harmed' due to their 'disability, mental disorder, illness or physical or mental infirmity' (Scottish Government, 2014, pp. 12-14). This harm might be perpetrated by anyone, including professional caregivers, family, strangers or the adult themselves. The Act places duties on public bodies to refer adults at risk to local authorities, and on local authorities to make inquiries; local authorities also have powers to intervene in certain circumstances. Work under this legislation might supplement and interface with existing services, or it might open up support services to people who had not previously used them. It is multi-agency in nature but is generally led by social workers and is co-ordinated via case conferences in more complex circumstances. The administrative arrangements are thus similar to UK and other Anglo-American child protection models (Filinson et al., 2008), though most proposed interventions require the adult's consent, making the position with respect to autonomy somewhat different (MacIntyre et al., 2018).
The development of the Adult Support and Protection (Scotland) Act 2007 raised questions around citizenship, choice and human rights (Mackay, 2011;Stewart, 2012). Service user and carer representatives were consulted, though arguably might have been involved more fully, and rights-based arguments were employed both to challenge and defend the strengthening of state protective powers (Stewart, 2012). On the one hand, the Act might be viewed as evidence of Scotland's more socio-democratic approach to welfare services in comparison to similar policy developments emanating from Westminster (Keating, 2009). From this perspective the increased statutory responsibilities to support and protect adults can be seen to acknowledge the causal links between harm and social exclusion. This legislation has also been argued to respond to the gap identified, within previous inquiries into deficiencies in safeguarding, between general welfare powers and mental health and mental capacity legislation that left people who were unable to protect themselves in harmful situations (Stewart, 2012). On the other hand, some disabled people and their organisations viewed the Act as paternalistic and discriminatory, arguing that it might lead to more controlling, risk-averse practices (Scottish Government, 2009). This was allied to its characterisation as individualising a social concern, one that places the focus, or intervention, on the harmed person rather than the harmer (Keywood, 2017).
Our professional and research interests span both these critical debates and also the operation and experience of the Adult Support and Protection (Scotland) Act 2007 on the ground. Adult safeguarding is the term we are using for a field of ideas, practice, policy and research that is not confined to Scotland nor to any given legislative criteria, then; the Adult Support and Protection (Scotland) Act 2007 is one particular policy and legislative approach that has been taken to these issues. This distinction is important because the research we discuss below asked some questions with relevance to adult safeguarding in the broader sense, and some narrower ones about the legislation.

Evidence, policy and practice
The above distinction also draws us into debates about evidence or knowledge, and its relationship to policy and professional practice. These questions have been discussed, researched and theorised extensively in the social policy and social work literatures. In relation to social policy in particular, the ways in which rational and linear models of policymaking have been critiqued since the 'scientific' turn in policymaking of the 1950s (Newman, 2011(Newman, , 2013 are too rich and extensive to do justice to here. Instead we wish to pick out just one manifestation and approach to these debates: the opposition between an 'instrumental rationality' in policymaking (Durose & Richardson, 2016;Sanderson, 2002, p. 62), and some alternative contemporary perspectives on knowledge and its uses.
Instrumental policymaking can be seen as a product of the adoption of evidenced-based policy such as the What Works agenda of the UK New Labour Government (1997Government ( -2010Sanderson, 2002). This approach arguably ignores the more intrinsic reasons why policy initiatives may be unsuccessful such as the impact of structural inequalities meaning some people have fewer resources than others to self-manage their lives in the ways envisaged by governments (Clarke, 2005;Walklate & Mythen, 2010). In the domain of social care, where adult safeguarding mainly lies, there is evidence of this instrumental rationality in the language of personalisation (Barnes, 2011) and where independent living has come to mean independence from the government and the welfare state (Morris, 2011). In this light the Adult Support and Protection (Scotland) Act 2007 and its allied Code of Practice (Scottish Government, 2014) can be seen as an instrumental approach to the policy concern of adult abuse and harm: one that focuses on individuals at risk as opposed to social causes.
One side-effect of instrumental policy making is that it may seek the opinions of those citizens affected through the lens of a given policy rationale rather than a more expansive exploration of the nature of the social concern itself. Policy evaluation can be conducted through the same narrow lens of inputs and outputs: what citizens wanted and their satisfaction with the outcomes (Beresford, 2002). Furthermore, there is a reported preference amongst those policy-makers who take this instrumental view for particular forms of knowledge: for instance quantitative research evidence, especially with some capacity to capture causality (Newman, 2011;Webb, 2001), and/or other evidence that can be 'readily packaged' (Newman, 2011, p. 478) for policy-makers' use. If we apply this to our case study, adult safeguarding in Scotland, research might only be funded to support the effective implementation of the Adult Support and Protection (Scotland) Act's prescribed roles and tasks. Conversely, alternative lines of inquiry might allow more scope for contestation about the nature of social problems and desirable outcomes, and/or might accord greater attention to the agency of professionals and service users in the enactment of policy (Newman, 2011). They might also open up the criteria for authoritative forms of knowledge, a point we return to below.
It is worth noting that parallel tensions between instrumental and alternative ways of knowing underpin debates about the nature of social work and related professional practice, which have played out over a number of decades and across a range of international settings (Hothersall, 2019). In particular, debates about 'evidence-based' or 'evidence-informed' practice have examined the extent to which professional interventions ought to be informed by 'evidence' in more or less direct ways, as well as how evidence ought to be appraised and what counts as evidence at all (Kelly, 2017;Munro & Hardie, 2019). In England, recent initiatives to train social workers through shorter, employment-based routes rather than university degrees might be characterised as a re-assessment of the necessary knowledge base for practice based on instrumental thinking. They were famously championed by Michael Gove, then UK Education Secretary, in a speech that criticised practitioners who 'acquiesce in or make excuses for the wrong choices' people may make and downplayed the importance to social work education of a structural analysis of inequality (Ramesh, 2013). Critics argue that such thinking underestimates the complexity of professional practice, the contexts within which it occurs and the types of expertise required to do it well (Cartney, 2018;Thoburn, 2017).

Participation
Stakeholders both in and outside of government have long aspired to the widening of participation in policy-making and public service design and delivery; this has sometimes been in explicit response to the felt disempowerment of particular groups. For instance, movements of UK disabled people have long campaigned to be in control of their support (Oliver, 2009), whilst there have been feminist analyses of the exclusion of women and minority groups from the places where decisions about policies and laws are made (Lister, 2003), indicating a need to open up assumptions within policy-making processes (Sevenhuijsen, 2004). In UK health and social care, co-production is currently promoted as an ideal model (Pieroudis et al., 2019), following decades of efforts to increase service user and carer involvement. However, caution is needed about the extent to which such approaches have genuinely reconfigured existing power structures (Newman, 2013;Paylor & McKevitt, 2019). For example, whilst the UK's National Institute for Health Research now recognises co-production as a means to improve patient and public participation, the wider managerial processes around health service delivery limit its adoption and impact (Paylor & McKevitt, 2019). Similarly, Pilgrim (2018) observes that whilst knowledge of compulsory mental health inpatient services has been co-produced, co-production has yet to be extended to the development and running of such services.
Various conceptual frames have been developed to express the rationale behind efforts to increase participation in public services. Smith et al. (2012) identify four inter-connected approaches: specifically, managerialist approaches, which view service user feedback as a resource to help improve service efficiency; consumerist approaches, which treat service users as consumers in a market-place; neo-liberal approaches wherein listening to service users helps to legitimate and smooth the implementation of government policy, and rights-based approaches, which aspire to the empowerment of service users. Much earlier, Beresford (2002) and Beresford and Croft (2001) drew a simpler binary between top-down 'consumerist' and bottom-up 'democratic' approaches and raised questions about the potential of the former to lead to real change. Beresford (2002) notes, in particular, that consumerist approaches focus on improving services, with little transfer of decisionmaking power away from professionals, whereas democratic approaches have a broader political aim of empowering service users and improving their lives.
Notwithstanding this, such typologies may blur at the edges, borrow each other's languages and intersect in complex ways (Beresford, 2002;Smith et al., 2012).
Participation is also an important subject for research, and an extensive literature now explores its nature, facilitation and purposes here (e.g. Beebeejaun et al., 2014;McLaughlin, 2010). Approaches that have emerged since the 1990s can be thought of as falling along a continuum between emancipatory research (Oliver, 2009), where disabled people, or other defined groups, fully control the means of research production, and traditional methodologies where social scientists are in control and participants are viewed as passive subjects to be studied (Cossar & Neil, 2015). Between these two poles lie various types of participatory methods that aim to bring together professional researchers and service users, carers or other groups, to explore a concern and seek change. Research towards the participatory pole may be more oriented towards questions of social justice than more traditional types of social research, and may seek to empower those involved, either explicitly or implicitly. There may be a focus on how co-researchers work together on a daily basis (Fleming et al., 2014), acknowledging the complex power dynamics, ethical issues and intersection of different interests and epistemologies at play (Banks et al., 2013). Clearly, there are important epistemological issues at stake too in respect of the dialogue between research and policy. Specifically, research that challenges established orthodoxies about who can produce knowledge and in what ways; indeed, the very possibility of value-free knowledge itself; takes up an uneasy position relative to the presuppositions identified above about hierarchies of evidence for policy use (Newman, 2011). Our own work and its dissemination helps illustrate these tensions; we introduce our work below.

Introducing our research
We have undertaken a range of research in the field of adult safeguarding with a particular focus on service user and carer views and participation, some of it in collaboration with other stakeholders, including government, local authority and voluntary sector groups. Table 1 summarises the five projects that form our focus here. Some of this work we find difficult to classify into the models noted above, although we can see elements of the various typologies at play. For example, studies 2 and 4 had an evaluative element in relation to the Adult Support and Protection (Scotland) Act 2007 and might therefore align more closely with managerialist, consumerist or instrumental approaches to the use of knowledge. However, we do not align ourselves with these approaches, and indeed, important messages from studies 2 and 4 consolidated our awareness of other ways of knowing. Younger disabled adults were involved as co-researchers (10) and participants (42), all of whom accessed Altrum support services. Forum theatre and associated drama and arts-based methods to discuss risk and protection with four groups of participants. Each group was based in a different area of Scotland, and each met four times to get to know each other and develop their ideas. Sessions were planned and delivered and data were analysed as a participative research group. Services users and carers were identified via three local authorities as potential interviewees. A participative study with social workers as co-researchers. Adult Support and Protection (Scotland) Act 2007 cases were identified via agency databases. One-off Interviews were conducted with 29 practitioners and seven service users or carers though more were invited to take part.
3 Improving ASP Participation Project Team (2013) (KM) Improving service user participation in Adult Support and Protection (Scotland) Act 2007 processes. The aim was for four local authorities to recruit service users as ongoing members alongside advocacy and social workers in small co-production teams. Actual service user involvement varied widely. Three co-production teams were supported by the researchers and attended three whole project workshops. Each choose their own ASP issue to work on. The most successful team was based within a learning disability service. The service user team member sought other service users' views; and a contact summary sheet for service users was piloted in Adult Support and Protection (Scotland) Act 2007 inquires. (continued) (FSJ) Independent advocacy providers' and users' views about the Adult Support and Protection (Scotland) Act 2007 and its impact on wider support and advocacy services.
The aim was for participating advocacy organisations to invite people who used advocacy to take part. No service user or carer was successfully recruited.
One-off interviews were conducted with 20 workers and managers across six advocacy organisations. Some interviews were one-to-one and some were conducted jointly with more than one participant from the same organisation.
(FSJ and KM) Older people's understandings of safety and vulnerability, and the importance of safety in their lives. Ways to generate knowledge about this, for research and practice.
The project was designed in collaboration with an independent advocacy project and some of their older members (n ¼ 10), who participated in two focus groups to refine the methodology for the second stage of the study. In the second stage, 5 focus group attendees also became participants, piloting the methods in collaboration with their allocated research team member and generating substantive information about risk and safety in their lives. A pilot project that set out to design and try out methods as well as to generate substantive data. Focus groups were used to refine methods. One-to-one data generation in the second stage used methods selected by each participant, principally two-way and three-way interviews and guided walks of the participant's local environment and daily life. Data analysis was shaped and guided by further workshops/focus groups.
The following sections exemplify our learning. We begin with the more evaluative approaches, discussing challenges that arose in building knowledge in this area of work, based on service users' and carers' voices. The latter challenges we discuss in this section operate as a bridge to the following section: they are challenges for an evaluative approach to researching the Adult Support and Protection (Scotland) Act 2007 because they point to other types of pressing research question. We then discuss our research in the field of adult safeguarding, in the broader sense outlined above. We term these studies and aspects of studies 'exploratory' because of their greater flexibility, with reference to the types of questions asked and the methods used.

Researching experience of the Scottish legislation: Evaluative approaches
One element of our work has been researching experiences of, and thoughts about the Adult Support and Protection (Scotland) Act 2007, from people who have been subject to it. First, KM led a collaborative project with practitioners, which investigated practitioners', service users' and carers' direct experiences of the legislation's processes and outcomes shortly after its implementation (study 2; Mackay et al., 2011). Second, FSJ's solo study incorporated exploration of independent advocates' and independent advocacy users' views of these legislative processes as one element of its aims (study 4; Sherwood-Johnson, 2015). Other initiatives have similarly sought the voice of service users and carers to investigate the implementation of this legislation, and they have also experienced the challenges we discuss below to various degrees (e.g. ASP National Priority Working Group on Service User and Carer Engagement, 2014; Burns, 2018;Miller, 2012).
First there was the ethical issue of avoiding causing further harm. People who have been abused and/or have been through protection processes might find it traumatic to revisit those experiences, even if the focus of the research is more on the service response than on the detail of the abuse. Positioning oneself, or having oneself positioned as a 'victim', a 'vulnerable adult' or a person who needs or needed 'protecting' might also be distressing or diminish self-esteem (Brookes et al., 2012). Even the request to participate might lead to one or both of these kinds of harm. Hence, we needed to design sampling protocols for studies 2 and 4 respectively, which risked excluding a range of voices, in order to manage as best we could these other types of risks. We did this largely by drawing on the advice of our gatekeeper agencies and speaking only with people able to access their support.
A second key ethical issue was informed consent. Some adults who have been subject to the Adult Support and Protection (Scotland) Act 2007 experience cognitive impairments like dementia or learning difficulties. We wished to resist the tendency towards immediate exclusion of those whose diagnoses raise questions about capacity to consent to research participation. This is because our views align with more nuanced approaches that problematise the concept of 'capacity' (Wiles et al., 2007), consider the benefits of research participation alongside potential harms, and also consider the broader risks of certain groups' exclusion from knowledge-making exercises (Dye et al., 2004;Munro et al., 2005). Therefore, our sampling protocols allowed judgements about capacity to be made on an individual basis. However, service user and carer participation in studies 2 and 4 remained disappointingly low.
We think gatekeeping processes contributed to these low response rates. Specifically, our means of accessing participants relied on collaboration with service-providing agencies. There is a documented tendency towards greater protectionism amongst organisations operating in this 'gatekeeper' role (Montgomery et al., 2017;Munro et al., 2005). In study 2 there was a staged procedure whereby samples were identified by a senior social work contact, then key workers invited service users to participate where they felt this was appropriate. There were thus two stages at which more potential participants might have been excluded on grounds of potential harm and/or capacity to consent than we might have deemed appropriate ourselves. On the other hand, the same study raised questions about the nature of the consent negotiated between some key workers and service users, given some of the interviewees recruited seemed to lack recall of being subject to the legislation or seemed not to fully understand the questions being asked. This is in part an ethical concern but also raises questions about the types of knowledge sought and the methods used, to which we return later.
In collaborating with gatekeeping agencies, there is also the challenge of use of staff time and other practical implications for that agency. For instance, FSJ asked independent advocacy organisations to invite their clients to participate in interviews as part of study 4. However, these organisations had limited resources and often did not have ongoing contact with people they had previously supported through these legislative processes. Approaching existing clients, therefore, was problematic from a resource perspective, whilst approaching past clients was also problematic in relation to the broader functions of the agency. The pool of potential participants was therefore reduced.
In short, then, our findings with respect to experiences of the Adult Support and Protection (Scotland) Act 2007 need to be read with an awareness that both ethical considerations and gatekeeping processes will have screened some people out. This will include some people whose experiences of harm and/or these legislative interventions may well have distinctive features: for instance, people with particularly limited abilities to understand or express themselves and people with particularly difficult relationships with service agencies.
We are also mindful that many studies around adult safeguarding, including some of our own, have confined their focus to people who have experienced the whole legislative processes from initial inquiries through to case conference as a starting point for evaluating implementation. For instance, study 2 specified that participants should have experienced 'substantial intervention' under the Adult Support and Protection (Scotland) Act 2007 (Mackay et al., 2011, p. 20). This focus helped to work through the different stages of interventions with practitioners so was useful for policy and practice purposes. However, it did mean the experiences of adults with more limited contact could not be addressed. This is important because their experiences might also be qualitatively different. Indeed, some independent advocates in study 4 suggested that progression into full legislative proceedings opened doors to fuller and more flexible use of resources, and better communication amongst professionals and with the service user, than these service users had previously been able access. In some of these circumstances, positive experiences of good communication and flexibility dropped off again, once legislative processes were concluded (Sherwood-Johnson, 2015). This raises questions about equality of access to properly funded and co-ordinated support, about the potential implications of safeguarding proceedings functioning or being seen to function as a gateway to support, about the 'cracks and gaps' in the legal definition of an adult at risk of harm (Sherwood-Johnson, 2015, p. 32) and about 'vulnerability' potentially being generated by budget cuts and associated gaps in more general support services. These are not questions it is possible to interrogate by means of a study focused only on people who have been judged to meet this legal definition, and only on their experiences subsequent to that point.
Meanwhile, our conversations with many service users and carers have also challenged a fairly basic premise of our policy evaluation aims: namely, that professional safeguarding activity is something discrete, which people can discern and reflect on separately from other types of interventions. For example, in study 2 one participant was able to talk in depth about how she felt during the safeguarding investigation and what was helpful about the support she later received; in contrast another participant struggled to distinguish between interventions that had occurred under safeguarding legislation and earlier interventions under mental health legislation and domestic abuse police responses. This might have rendered her account of limited value within an evaluation of the Adult Support and Protection (Scotland) Act 2007, yet such accounts were richly informative in other ways, and have influenced our thinking about other types of research question.
We think the above points illustrate a key problem with research into people's experiences of a particular legislative approach, at least when considered in isolation. Namely, whilst the core research questions are of pressing interest to practitioners, managers and policymakers as they go about their specific legislative work, they are likely to prove a less than perfect fit with people's key concerns as service users and/or carers (Beresford, 2002). Such research might highlight, without being able to pursue, more intrinsic and cross-cutting policy themes, such as people's understandings and experiences of services at the intersection of several policy agendas, and indeed people's understandings of the concepts of safety and vulnerability themselves. It is these types of broader issues to which we now turn.

Adult safeguarding in a broader sense: Exploratory approaches
We have explored a number of research questions which concern adult safeguarding in a broader sense. FSJ was involved in a participatory project with younger disabled adults, investigating their key messages for professionals when supporting people to manage risk (Altrum Risk Research Team, 2011). This project took place in the context of the Adult Support and Protection (Scotland) Act's recent implementation, overlapped in time with project 2, and led to some discussion between the respective research teams. Rather than ask how these particular processes and interventions have been received, KM's subsequent co-production research project 3, involving practitioners and local authority service users, asked how participation in such processes might be facilitated, with a focus on developing practical tools to achieve this (Improving ASP Participation Project Team, 2013). Project 4 investigated the relationship of independent advocacy services to adult safeguarding services, and how and if service user and independent advocate participants perceived support services to have changed, given the increasing attention in policy and law to statutory services' 'protective' responsibilities. Some discussion of how the new legislation impacted on their work formed one element of this project but was not intended as its primary focus (Sherwood-Johnson, 2015). Finally, our current joint project in collaboration with an independent advocacy organisation and some of their older members asks what safety and vulnerability mean to older people, how they keep safe, and how important safety is in their lives (Sherwood-Johnson et al., 2019).
None of the above projects required people to have been subject to safeguarding processes in order to join the research team and/or become participants. Rather, they involved groups with some shared experiences of particular support services, together with some experience of one of the legal criteria of an adult at risk of harm, namely 'disability, mental disorder, illness or physical or mental infirmity' (Scottish Government, 2014, pp. 12-14). One effect of this was to temper the acuteness of the ethical issues discussed in the last section, because no request was being made for people who had been subject to abuse and/or distressing experiences of protection to re-visit those experiences. Rather, these participants' and/or participant researchers' expertise lay elsewhere: for instance, in managing risk in their own lives and engaging with support services more generally.
We have found that working collaboratively with service users and carers at an earlier stage of research projects has helped to open up illuminating conversations with respect to research questions. In study 3, local co-production teams each negotiated their own focus and went on to design practice tools of particular relevance to their local settings. In this way, collaboration ensured that usefulness was designed into the project from the start. Study 1 had its inception many months before research funding was secured and began with an explicit negotiation over the nature and focus of the research questions. Specifically, university researchers approached a social care service user group to raise awareness about the Adult Support and Protection (Scotland) Act 2007 and to propose some joint research about the ways they would like to see it implemented. Group members were keen to include much fuller attention to the positive as well as the negative aspects of risk in people's lives and raised the university researchers' awareness of the politics of focusing on just one side of this coin. The final questions reflected and blended the types of expertise and insights that both participant and university researchers brought to this debate, foregrounding what service users thought that practitioners should know alongside what those versed in policy felt were the key questions.
Collaboration has also helped to grow more creative methodologies. We have found these to be more accessible to some participants than the traditional interview or focus group design. Study 1 used forum theatre and other drama-based activities with groups of participants who each met several times. Study 5 uses interviews but is ethnographic too, in that it treats spending time with people repeatedly, over a period of time in their local environments as a research method. It incorporated a first stage of focus groups specifically concerned with refining the methods of the next stage of the project. These iterative designs allow relationships to be developed, and the focus of the study to be understood, reinterpreted and reconsidered over time in the context of those relationships (Beebeejaun et al., 2014;Ward & Gahagan, 2011). In our experience this has supported the generation of rich findings that connect with individuals' lives and real concerns, more fully than studies that only interviewed the person once.
The strengths we perceive in these studies, however, are linked to their own challenges. Specifically, such studies are time-and resource-intensive due to relationship building and shared decision-making (Fleming et al., 2014), and may therefore be more expensive than other types of study. Few if any funders make allowance for the time and resources that may be required to develop a funding bid in a participative way, as in study 1. Even having secured funding for this explicitly participative project, we found systems for research administration and governance to be a hindrance to participation at times. For instance, our funder could not release funds until ethical approval had been secured, even having granted funding to a project that planned to spend several resource-intensive sessions developing the study design, prior to application for ethics approval. Other funders may be put off altogether from a study that is not designed to deliver straightforward answers to prescribed policy questions. This is notwithstanding our experience that other types of findings can prove equally if not more foundational to good adult safeguarding practice. For instance, whilst participants in study 1 could not necessarily report how an investigation under the Adult Support and Protection (Scotland) Act 2007 feels, they could certainly speak about feeling unfairly restrained, or judged and found wanting, over years of contact with professionals and carers, some of whom they trusted more than others, and the effect that this background might have on encounters with new professionals purporting to 'help' them (Sherwood-Johnson et al., 2013). Similarly, participants in study 5 could not necessarily describe how best to respond to abuse, but they gave vivid accounts of the balancing act involved in managing their dayto-day risks, in the context of their wider lives and relationships, which could help sensitise social workers and others with safeguarding duties to the potential complexity and emotional significance of this terrain to older people (Sherwood-Johnson et al., 2019).
Not only at the point of applying for funding, but throughout the research process, we have encountered challenges in the translation of the subject and worth of our research back and forth between its various stakeholders. This brings us back to our earlier distinction between research that focuses only on the Adult Support and Protection (Scotland) Act 2007 and adult safeguarding research. FSJ, in particular, feels that her lack of sufficient attention to the extent that this distinction might be shared was a significant contributor to the difficulties she experienced in the recruitment of service users to project 4. That is, asking independent advocacy workers to recruit service users to a study about adult safeguarding was problematic because 'adult safeguarding' or 'adult support and protection' had already become a shorthand for Adult Support and Protection (Scotland) Act 2007 processes or meetings in the eyes of these gatekeepers, perhaps mirroring a broader trend in the services they worked with. Despite purposefully broad inclusion criteria and the explicit articulation of project aims in terms that did not emphasise procedural issues, it was difficult for these gatekeepers to appreciate that people not subject to these legislative processes might contribute meaningfully.
The flipside of the above challenge is the translation of findings back into the policy and practice environment. Specifically, there is the risk of significant loss of meaning in translation between the wide-ranging and at times tentative findings of participative research, and some policymakers' and practitioners' desires for clear measurable data about a narrowly specified policy stream (Newman, 2011). For instance, the outputs of study 1 included some videos of role-played 'moments' in an 'Adult Support and Protection journey' (Altrum Risk Research Team, 2011). One video extract, the imagining of part of an Adult Support and Protection (Scotland) Act 2007 case conference, speaks volumes about the emotional intensity of certain relationships, and the types of disempowerment some participants feared in their interactions with relatives and professionals in the context of concerns about risk. However, because the extract does not depict professional roles and processes in a wholly technically accurate way, it was received by some attendees at dissemination events as an invalid criticism of these roles and processes, and hence of limited practical use to them. As we reach the dissemination stage of study 5, we are similarly aware that the findings do not point to simple 'how-to' messages for practitioners, or the need for tweaking of administrative procedures within a given policy stream. They are about people's lives more holistically, rather than services and policies specifically, and their application will require more thought, and perhaps more resourcing than this. Nevertheless, in our experience this makes them more, not less, important.

Limitations
This article discusses issues to do with participation (methodological, theoretical and ethical) drawing on five of our previous joint or respective research studies. There is a focus on methodological issues rather than the findings of each study, and the limitations of our methods with respect to participation in each case are integral to our discussions throughout. However, the specific issues we have raised are based on our own experiences and interpretations of those experiences, rather than on any process of independent or more systematic review. Readers are referred to Altrum Risk Research Team (

Conclusions
Our learning about participation in the context of building knowledge for policy and practice has implications both within and beyond adult safeguarding; we consider it in spheres of progressively greater transferability here. First, there are certain implications for increasing service user and carer feedback on the implementation of the Adult Support and Protection (Scotland) Act 2007, and similar types of processes across the UK and elsewhere, the importance of which we would not wish to underplay. We have noted that there are ethical and practical difficulties in recruiting participants into this type of study, and that these difficulties are particularly acute for researchers external to the agencies who undertake this legislative work. Hence the in-house and collaborative evaluation activities that have developed in recent years in various parts of Scotland are particularly to be welcomed (Burns, 2018;Miller, 2012). There is scope to develop these further and to share their findings more widely. There is particular scope to widen involvement in their design and conduct, following key innovations such as Burns (2018) and Montgomery et al. (2017).
An argument might also be mounted that outcomes-based models of adult safeguarding, such as the Making Safeguarding Personal approach adopted by English practice guidance (Department of Health, 2017), circumvent some of the challenges we raise. This is because this model of practice begins with capturing the wishes of the service user, and proceeds in a way that incorporates evaluation of progress against these outcomes on a formative and summative basis (Cooper et al., 2016;Department of Health, 2017). We would argue, however, that a number of problems remain. First, there are related practical challenges concerned with the resources required for practitioners and agencies to fully implement this approach, together with the danger that it becomes a service audit tool (Department of Health, 2017). Second, it is a key finding of ours and others that at a time of crisis following the experience of harm, poor physical and mental health may limit individuals' capacities to engage in the planning of their own support, and professionals may then legitimately need to guide or lead (Mackay, 2017;Mackay et al., 2011). This can be conceptualised as maximising individuals' capacities to act autonomously in the longer term (Bergeron, 2006;MacIntyre et al., 2018), but it also resonates strongly with ideas about 'care' and approaches to 'vulnerability' that do not rest on the kinds of instrumental assumptions we have sketched out above. Specifically, just as policy approaches that treat individuals as autonomous actors on a level social playing field misrepresent the contexts in which our needs for care and protection arise (Fineman, 2008;Fyson & Kitson, 2007), so approaches to evaluating safeguarding policies will inevitably paint skewed and partial pictures, if they rely too heavily on the testimonials of those able to provide them for traditional research and/or evaluation exercises, whether or not those exercises are built in as part of service provision. Third, we have shown above that where such exercises begin with the intention to evaluate a specified policy or service, they risk poorness of fit between their research questions and the substantive concerns of service users' lives. These may not translate fully or well into the policy's conceptual schema, and they are also likely to cross a number of policy streams; in adult safeguarding, for instance, they may cross into domestic violence policy (Williams & Zerk, 2018) or more general social care .
Whilst not claiming this as a panacea, we have shown that research questions and methods designed and implemented in more participative ways have potential to address this poorness of fit and also to engage a wider range of voices. In doing so, such activities unsettle some key presuppositions about what knowledge is, who makes it and how it might be used. The challenges for these approaches, then, are in the interfaces with funding agencies and potential users of the research findings. This is a key manifestation, indeed, of Newman's (2011) tension between research that is critical and research that makes a difference, in the sense of commensurability with established ways of influencing policy and practice. All this underlines the critical dialogue about knowledge itself that needs to be ongoing between centres of research and/or educational activity, centres of research funding, centres of policymaking and places where public services are provided or co-produced. Spaces where service users, carers, practitioners, policymakers, educators, students and others can collaborate without too rigid an agenda of service evaluation require to be maintained or developed, in short (Ward & Barnes, 2016), despite these straitened times.
In social work there is a long tradition of criticality about knowledge and its uses, with significant bodies of theory addressing the relationship of its different forms to the practical tasks of the social worker (Hothersall, 2019;Kelly, 2017). Nevertheless, there is an ongoing need to challenge neoliberal incursions into the education and governance of the profession, and to support practitioners to gain or re-gain research-mindedness, including participation in research themselves. This would assist them in the complex process of applying research findings, particularly from the types of more exploratory study outlined above. Like our findings about safeguarding themselves, none of these insights into participation and the building of knowledge for policy and practice are necessarily easy, or quick, or cheap to implement. However, it seems to us that they take us closer to knowledge exchange that truly involves exchange.