Constraining co-creation? An ethnographic study of Healthwatch organizations in England

While the potential benefits of co-production and co-creation are widely vaunted, the degree to which they precipitate innovative change in systems varies, and influences on their impact demand greater attention. We present an ethnographic study of organizations intended to foster co-creation in English health and social care. Comparing five cases, we find demonstrable activity and some tangible impacts. At the same time, the positioning of these organizations as collaborative insiders in local governance systems constrained the publics and issues that were prioritized in co-creative activities. Our findings highlight the significance of discursive forms of meta-governance in delimiting the scope of co-creation.


Introduction
The last decade has seen an explosion of interest in co-production and co-creation as means of improving the management and governance of public services through the input of wider constituencies, including above all the publics served.Building on a long tradition of work that notes that public services are inevitably the joint product of the work of 'providers' and 'recipients', scholarship on co-production and particularly cocreation emphasizes the wider ways in which public service users can, could or should add value to the governance of public services (Brandsen, Trui, and Bram 2018;Osborne, Radnor, and Nasi 2013;Torfing et al. 2021).
For all their promise, however, the extent to which these new forms have resulted in meaningful change to the organization and delivery of public services is not fully clear.Consequently, there are calls for more research on the 'mainstreaming' of co-creation in public services (Ansell and Torfing 2021b, 261), including the role of contextual conditions and the degree to which its distinctive contributions impact on systems and services in practice (Ansell and Torfing 2021a;Ferlie 2021), and for studies that demonstrate the conditions under which co-creation fails to achieve these aims (Baptista, Alves, and Matos 2020).
In this article, we respond to these calls with a comparative study of ostensibly similar initiatives in co-creation in five locales in health and social care in England.This field of public service delivery has seen rising interest in co-production and cocreation since the mid-2010s, including a prominent role for organizations known as 'local Healthwatch', whose mission is to 'strengthen the collective voice of local people' (Department of Health 2012, 4).With varying resources and approaches to involving and representing their constituencies, however, Healthwatch face divergent background conditions as they seek to realize their missions.
Drawing on comparative analysis of ethnographic and interview data, we find much evidence of co-creative activity, and some evidence of impact on services.However, we also find that some areas of public service organization and delivery were rendered 'out of scope' of the dialogue and deliberation of co-creation.In consequence, the distinctive impacts vaunted for co-creation -creative, ingenious solutions that address entire service systems (Bentzen 2022;Torfing et al. 2021) -were not to be found.This arose not because of conflict between the interests of public and professional stakeholders (Jaspers and Steen 2019;Steen and Sanna 2018), but rather from a confluence of interests among those involved.The result was that co-creation resulted in relatively narrow changes, in terms of substantive focus, intended objectives, and the publics whose views were advanced.Our analysis shows that willingness, enthusiasm and collective endeavour towards co-creation are not sufficient to give rise to outputs that are 'innovative yet feasible' (Ansell and Torfing 2021a, 221): rather, a shared interest in demonstrating the impact of co-creative endeavours appeared to prioritize the feasible over the innovative.We highlight the importance of accounting for the role of meta-governance in delimiting co-creation's scope and focus -including not just the specific steering activities of meta-governing agents, but also the less tangible but highly important influence of more discursive forms of meta-governance, and their consequences for the dispositions and actions of those involved in co-creation.

Co-production, co-creation and new forms of public governance
Despite (or perhaps because of) the increasing attention devoted to them in the public management literature, both co-production and co-creation defy universally agreed definitions.Helpful accounts of the development of the terms have been provided by others (e.g.Ansell and Torfing 2021b; E. M. Eriksson 2019), so we rehearse only a limited history here.
The notion of co-production in the original sense-i.e. that those traditionally thought of as the 'consumers' of services also have a crucial, even inevitable, role in making those services happen -can be traced back 40 or 50 years (e.g.Ostrom et al. 1978).However, more expansive notions of co-production's role in not just delivery but also design and organization of public services have developed more recently.Building on insights into co-production in service delivery, and on ideas from other literatures such as the joint generation of value by providers and consumers in privatesector service industries (Prahalad and Ramaswamy 2004), various authors have made the case for a 'strategic orientation' (Osborne, Radnor, and Nasi 2013, 141) towards coproduction.Given that the value generated by public services is not reducible to the moment of contact between provider and consumer, public services might better improve their quality and responsiveness by adopting more thoroughgoing approaches to co-production, engaging with service users 'at all phases of a (public) service lifecycle' (Osborne, Radnor, and Nasi 2013, 142), and treating them as stakeholders, rather than mere consumers.Others have similarly called for a shift from interest in 'rather mundane forms of co-production' (as an immutable feature of service delivery) towards a focus on 'systematic involvement of relevant and affected actors in the cocreation of new public services and delivery systems, [. ..] innovative solutions to complex problems [. ..] and public policies and regulations' (Torfing et al. 2021, 190).
Accordingly, Brandsen et al. (2018, 3) define co-production and co-creation, together, as 'a joint effort of citizens and public sector professionals in the initiation, planning, design and implementation of public services'.While the two terms are indeed often used synonymously (W.H. Voorberg, Bekkers, and Tummers 2015), others prefer to distinguish between them, positing co-creation as the more expansive and encompassing concept.For Torfing et al. (2021, 191), for example, co-production is a process that takes place predominantly bilaterally at the point of service delivery, whereas co-creation involves 'a broader group of citizens and organised stakeholders in [. ..] new services, entire service systems, and public planning solutions'.Our empirical and conceptual interest lies in this broader notion of co-creation as a multilateral process covering design, oversight, planning and management as well as delivery.We therefore use this term by preference, though we note that the policy and practice literature in particular also use co-production in this more expansive way.
The potential of co-creation has caused excitement in the public management literature.At a theoretical level, this reconceptualization of the service user as active stakeholder is seen as more satisfying than the tendency to reduce her to either citizenvoter (in the old public administration) or consumer (in the new public management).At a practical level, co-creation opens the possibility of much more fundamental and radical change in public service delivery than older forms of governance could achieve, thanks to the transformative potential offered by bringing together the diverse, even conflicting, forms of knowledge of different stakeholders for problem-solving (Ansell and Torfing 2021b;W. H. Voorberg, Bekkers, and Tummers 2015).By fostering creative synergies through multilateral relationships of all stakeholders, co-creation goes beyond the service delivery focus of co-production and instead promise 'new and innovative solutions to intractable problems' (Torfing, Sørensen, and Røiseland 2019, 804) and impact on the whole 'service ecosystem' (Engen et al. 2021, 889).
Yet, as more cautious commentators have noted, this promise remains, for the moment, largely unfulfilled.Much of the literature remains conceptually oriented rather than empirically informed (Baptista, Alves, and Matos 2020).The question remains whether co-creation will ultimately turn out to be 'merely a fad, a way to re-engineer service management in the public sector, or a whole new ethos for public organizations' (Sicilia et al. 2016, 24).A comprehensive review of the literature has found a greater focus on inputs and process than outputs or outcomes (Voorberg, Bekkers, and Tummers 2015).Since then, further studies have emerged, including examples of success (Baptista, Alves, and Matos 2020), but as Ansell and Torfing (2021a, 5) concede, 'few authors have sought to broaden the focus from service production to policy making, societal problem-solving, and democratic participation'.Rather, while there is evidence of its impact at the service level (Baptista, Alves, and Matos 2020; Voorberg, Bekkers, and Tummers 2015), co-creation's vaunted transformational potential has been realized only patchily (Bentzen 2022;Bentzen, Sørensen, and Torfing 2020;Leino and Puumala 2021;Ongaro et al. 2021).The reasons for this inconsistency can be summarized under three headings: institutional context; the degree to which co-creation is genuinely inclusive of the diversity of perspective on which it thrives; and the internal dynamics of co-creation processes.
First, prevailing institutional arrangements appear to have a determining influence on whether co-creation takes hold in the first place, what form it takes, and with what prospect for impact (Ansell and Torfing 2021b).Different approaches to governance, consultation and policy formulation give rise to divergent realizations of co-creation (Ferlie 2021; Lahat and Sher-Hadar 2020; Pestoff 2018; W. Voorberg et al. 2017).Administrative cultures that tend towards consultation rather than unilateral action by the state are more likely to engage in co-creation in a way that supports innovation and influence at levels beyond that of the individual service, for example by delegating control to co-creation networks or at least including mechanisms to ensure that the fruits of co-creation are not simply put aside at key decision-making points.The approach taken by bureaucrats is also important, with those who understand cocreation as integral to broader organizational strategy (Ongaro et al. 2021) and respond with appropriate resourcing (Van Gestel, Kuiper, and Pegan 2023) securing a stronger role for co-creation in decision-making at the system level.
Second, the reliance of co-creation on the voluntary input of self-selecting publics brings advantages and disadvantages.The contributions of those most affected by societal issues to their resolution might supplement and enhance existing democratic institutions (Ansell, Sørensen, and Torfing 2023), but co-creation also risks creating new inequalities in access to decision-making (Verschuere et al. 2018), or exacerbating existing inequalities of service provision (Meijer 2016).Ensuring some kind of representational relationship between those involved in co-creative activities and wider constituencies might offer some means of mitigating these risks (E.M. Eriksson 2019; Grubb and Frederiksen 2022), but if co-creation relies on the input of a highly selective group of stakeholders, or is not pursued in an inclusive manner, both its legitimacy (Jacobs and Kaufmann 2021;Meijer 2016;Steen, Taco, and Bram 2018) and its creativity (E.M. Eriksson 2019; Leino and Puumala 2021;Meijer 2016) will be harmed -with the concomitant risk that its influence will be undermined, or that it will fail to find 'new and bold solutions' (Torfing et al. 2021, 195) and will instead be confined to tinkering at the service level.
Third, for all the rhetoric of inclusiveness, equitable collaboration and consensusbuilding, co-creation may remain subject to inequalities of power -between different stakeholder groups, and between state and non-state actors (Sicilia et al. 2016;Steen and Sanna 2018).Different groups involved will have different interests (Jaspers and Steen 2019), and while co-creation offers an opportunity to overcome 'ideological polarization' and reach consensus through rational deliberation (Ansell and Torfing 2021b, 173), the conditions for an equitable process are not always readily achieved (Martin 2012;Turnhout et al. 2020).Again, this poses a fundamental risk to the distinctiveness of co-creation and hence its value proposition.If co-creation fails to capitalize on its potential to bring conflicting views into creative synergy, then its innovativeness and its transformative impact on public service systems will be lost (Bentzen, Sørensen, and Torfing 2020;Leino and Puumala 2021;Meijer 2016;Steen, Taco, and Bram 2018).
Together, these issues have a vital bearing on what comes of co-creation efforts.Does it result in 'innovative public value outcomes' at the level of the system that draw on 'otherwise untapped experience, knowledge, resources and perspectives' (Ansell and Torfing 2021a, 216, 219), or is its impact confined to more marginal gains at the level of services?Our qualitative case study work, examining five instantiations of a national policy in England that sought to involve the public in the co-creation of health and social care policy and practice, offers some answers to this question.

English health and care governance: the rise of co-creation
The history of healthcare governance in England has been documented extensively elsewhere (e.g.Klein 2013).It reflects swings in wider trends in public governance theory and practice, from the centralizing tendencies of old public administration, through the reforms foregrounding competition and the import of private-sector management principles of the new public management, to more recent shifts towards networked, collaborative approaches associated with new public governance.The United Kingdom (UK), and England in particular, has found itself at the vanguard of these successive waves of reform (Ferlie et al. 1996;Newman 2001).However, each wave has not entirely washed away the remnants of what preceded it, leaving what others have referred to as 'sedimented' governance (Jones 2017) or a 'palimpsest' (Dickinson 2016) -a fertile ground for studying the impact of these antecedent conditions on co-creation (Ansell and Torfing 2021b).
The origins of local Healthwatch organizations can be found in reforms introduced by the UK's Conservative-Liberal Democrat coalition government in the early 2010s.Following something of a shift towards more collaborative and networked forms of governance under Labour governments from 1997 to 2010, the 2012 Health and Social Care Act marked a pronounced resurgence of the values of new public management (Asthana 2011), including greater use of competitive tendering and a stronger role for Clinical Commissioning Groups (CCGs), the purchasers of care.It also introduced local Healthwatch organizations as the key means of securing patient and public input into the National Health Service (NHS) and social care.At this stage, far from adopting the language of co-production or co-creation, Healthwatch was variously presented as 'a new consumer champion' and a means of giving 'citizens a greater say in how the NHS is run' (Department of Health 2012, 4) -a tension in constructions of the public that did not go unnoticed by commentators at the time (Carter and Martin 2016).
Yet within a few years of the Health and Social Care Act, debates about the governance of healthcare had shifted radically -in government, in the NHS nationally, and in local health systems.Influential in heralding this change was the publication of the Five Year Forward View by the NHS's national leadership (NHS England 2014), which set out looming challenges around funding, demand and productivity for the NHS -and espoused not greater competition, but greater collaboration, as the solution.Language in and around the NHS began to invoke the promise of co-production, at the levels of the individual patient encounter (e.g.Department of Health 2018), and the transformation of whole systems (i.e.what we refer to as co-creation) (e.g.NHS England and Coalition for Personalised Care 2016).The broader logic of collaboration also quickly took hold.It soon became the organizing principle for the NHS (Hammond et al. 2019), which was restructured into 42 regional bodies ('integrated care systems') responsible for leading integration of care, co-ordinating collaboration and co-production across organizational boundaries and agreeing local distribution of budgets (Pickett 2017).These shifts towards collaboration and co-creation, however, took place in a legal and regulatory environment still framed by the provisions of a more competition-oriented act of parliament.Indeed, only in 2022-with the passing of a new Health and Care Act -were integrated care systems formally incorporated as organizations with executive decision-making powers, finally regularizing a situation that had existed de facto for several years.
For Healthwatch organizations, the rising rhetoric of co-production offered opportunities to fashion a central place in the heart of the local governance of health and social care.Moving away from the consumerist overtones that accompanied their introduction, Healthwatch were increasingly presented in terms of strengthening 'patient voice at place and system levels, not just as a commentary on services but as a source of genuine co-production' (Secretary of State for Health and Social Care 2021, 19).But they faced challenges too, in part associated with the way that co-production had arrived on the English health and care scene -not through primary legislation, but as an answer that the NHS had largely identified for itself.Similar to other contexts (e.g.Torfing, Sørensen, and Paulsen Breimo 2023), co-production here had some of the hallmarks of a 'magic concept': a cure for many ills, and one whose key features were rather under-specified and malleable.Consequently, Healthwatch were left with a strikingly open brief for their work, in terms of the publics they were to convene, the issues they were to advance, and the degree to which health and care organizations were required to engage with them (Carter and Martin 2016, Martin andCarter 2017).They were also constrained by their own place in the system.The 2012 Act had set Healthwatch up as something akin to a contracted-out 'service': a function to be commissioned by local authorities (local government agencies) on a competitive tendering basis.Moreover, local authorities were allocated monies by central government, determined in large part by capitation, to fund local Healthwatch, but this funding was not ringfenced; consequently, Healthwatch could vary substantially in size and function.Finally, the emergence of the new integrated care systems added further complication, since the 'footprints' of these 42 systems did not map neatly onto the geography of the country's 150 local authorities.
Local Healthwatch thus found themselves in complex and multi-layered settings in which the traces of past trends in governance remained important -a common feature of many contemporary public service settings (Dickinson 2016) -and where differences of scale, scope and focus were evident in local realizations of an ostensibly similar initiative.Given this background, they offered a promising site for examining the influence of antecedent conditions on the products of co-creation, and whether and how these resulted in the changes anticipated (Ansell and Torfing 2021b).

Methods
As part of a wider mixed-methods study, from August 2019 to January 2021 we undertook ethnographic fieldwork in five areas, examining the work of local Healthwatch organizations in their wider systems (Zoccatelli et al. 2022).We sought to gain insight into the actions, practices and relationships involved in realizing Healthwatch on a day-to-day basis, getting beyond official accounts and attending in particular to the interactions that produced, reproduced and represented Healthwatch as a locus of co-creation (Rhodes 2015).
We selected Healthwatch for inclusion using a theoretical sampling approach (Eisenhardt 1989), sampling for variation in characteristics identified as potentially consequential through existing literature, discussion with stakeholders and earlier parts of the study; these included organizational structure, type of local authority, demographic and system characteristics of area covered, and budget.Table 1 summarizes the characteristics of the selected case-study sites.Having approached the Healthwatch identified and gained organizational agreement to participate, we undertook extensive ethnographic data collection in each, comprising: in-person observation of day-to-day activities such as meetings, visits and public events, enriched by informal conversations with actors in the field; semi-structured interviews; and collection of documentary evidence such as public-facing reports and internal documents.Mid-way through our data collection, the COVID-19 pandemic, and associated government interventions, necessitated a shift in our approach to data collection.Some Healthwatch activity continued but largely took place online; accordingly, we moved towards 'virtual' ethnographic observation, and carried out remaining interviews by telephone or video call.
Prior to the pandemic, we had undertaken 75 days of fieldwork across the sites, which we supplemented with further ad hoc contacts such as observations of online meetings and informal chats with stakeholders (Table 1).In total, we undertook 84 interviews in-person or remotely, with Healthwatch staff and volunteers (39 interviews), members of their boards ( 14), individuals in local health and care systems (such as local authority commissioners and officers, elected councillors, and managers in health and social care organizations and emerging integrated care systems) (10), and other stakeholders such as staff of other third-sector organizations in the area (21).During fieldwork, data collection was enriched and focused through ongoing deliberations within the team, and by presenting and discussing findings with an involvement panel comprising 15 Healthwatch staff and volunteers from other areas, who provided reflections on emergent ideas and direction for further data collection and analysis.
Data from field notes, interviews and collected documents were brought together for analysis assisted by NVivo software and based on a logic of cross-case comparison (Eisenhardt 1989) and an approach to coding based on the constant comparison method (Charmaz 2006).GPM read and re-read all sources systematically and then coded excerpts of the texts on a case-by-case basis to high-level thematic codes relating, for example, to the influence of context (including resources, populations, and organizational histories), specific activities undertaken by Healthwatch (such as interactions with systems, organizations, and publics), and the impacts and forms of influence claimed by Healthwatch and evident in data collected.We also developed more abstract codes, relating for example to the conceptualization of publics, the nature of Healthwatch as an organization, and the evolution of its purpose.In consultation with the other authors, GPM undertook an iterative process of reviewing and refining these codes, breaking them down according to more specific themes, merging and reconfiguring them, and systematically comparing cases to identify cross-case patterns.He then constructed a narrative synthesis of the key features of cases, patterns of activity and influences on the parallels and divergences between cases that were shared, further developed and agreed among all authors, and which forms the structure for the presentation of findings that follows.

Findings
We present our findings in three subsections, focusing on the context for co-creation and Healthwatch's place in it, the activities pursued by each Healthwatch, and the outputs of this co-creative work.Across sites, the institutional context appeared promising, as did the inclusiveness of the approach to co-creation.Stakeholders agreed on the important contribution that Healthwatch had to offer to local system governance, and shared a stake in demonstrating the impacts of an inclusive approach.The internal dynamics of co-creation activities varied markedly, however, depending on the Healthwatch organizations' resources and capacity, which impacted their ability to demonstrate the value of their work to local decision-making.Across the five sites, then, the institutional context and inclusiveness of co-creation were both promising, but its internal dynamics were more varied.Yet we also note a convergence across sites in the outputs of co-creation: the kinds of contributions Healthwatch sought to offer, and the ones they discarded.To varying extents, the organizations converged on a rather instrumentalist and conservative notion of what co-creation should createone that fell short of the innovative and transformational impacts envisaged by its advocates.

The context for co-creation
Across cases, participants within Healthwatch and the wider integrated care systems noted the need for these small organizations to identify a distinctive contribution to health and care governance.Healthwatch officials and wider stakeholders alike landed on a role for Healthwatch that foregrounded its local connectedness and ability to represent the views of the public in the round, especially those at risk of marginality.By fostering co-creation in a way that engaged such groups, Healthwatch were seen as a means of enhancing the quality of decision-making by incorporating a diversity of views (Torfing et al. 2021).
The relationship between something like Healthwatch on a very local level means they know the geography, they know the people, they get rooted in the community.(Local authority commissioner, Ambria) What do I expect of Healthwatch in this is-and I don't know if any of us have got this right yet -but there's layers of understanding about patient need and patient view and some of that top level about we all know it, the who shouts loudest thing, and I think all of us really try to get under it.How do you hear the seldom-heard groups?(Healthcare commissioner, Ethershire) Many features of the institutional context for co-creation were promising.Support from a range of constituencies was in place, most notably from state agencies whose support for co-creation is vital (Ansell and Torfing 2021b;Sicilia et al. 2016;W. Voorberg et al. 2017).Top-down policy support was also in place (W.H. Voorberg, Bekkers, and Tummers 2015): Healthwatch's status as the 'official' conduit for the public's voice conferred legitimacy on its activities.Participants from organizations that commissioned and worked with them noted their statutory positioning and the breadth of their mission -working for the public 'at large' rather than any specific interest group -as an important virtue: I do think it's helpful that Healthwatch have a statutory role; I think it's helpful that they have a generic oversight because it means that they're not drawn into one agenda in particular, because a lot of those other patient groups, by their definition, have an agenda.[. ..]There's a role around public engagement and supporting co-production: that's with Healthwatch.(Integrated care system manager, Ambria) All in all, there was little challenge and plenty of support, both for Healthwatch specifically, and for the idea of more inclusive, co-creative approaches to health and care policy more broadly.The ability of Healthwatch to bring the views of marginalized and 'seldom-heard' groups to the table formed an important part of their appeal, helping to address concerns that co-creation might be dominated by relatively elite communities and neglect those in most need (Meijer 2016;Verschuere et al. 2018).
The result was that stakeholders across the whole system had a shared interest in demonstrating that co-creation was happening in a visible, tangible way.
[Local authority] were really pleased with everything we were doing but they felt we could improve on-what's the best way to put it?-publicizing the successes.[. ..]Flag stuff up and maybe publicize more widely when we think we've achieved something.(Healthwatch officer, Cramford) The context for co-creation thus appeared to be a promising one.Healthwatch had a policy-mandated position, and were able to craft a role for themselves that other system stakeholders valued, particularly Healthwatch's ability to bring in the views that they struggled to access themselves.Moreover, all stakeholders recognized the need to demonstrate that they were actively engaging in the ascendent co-creation agenda.As we see next, this performative need had a strong influence on the activities pursued by the five Healthwatch -but they were not all equally equipped to meet it.

The activities of co-creation
Demonstrating influence and impact was, as might be expected in light of the context described above, a frequent focus of Healthwatch events we observed, both internal and for public audiences.'Getting your seat around the table' (Healthwatch officer, Ambria) and having something worthwhile to say once there could address the needs of Healthwatch themselves to demonstrate influence, and of integrated care systems to show that they were embracing co-creation in the way demanded by national leaders (NHS England and Coalition for Personalised Care 2016; Department of Health 2018).
They discuss who would be good to have on the panel at the [Annual General Meeting] and how they should approach it-'What effect has [Healthwatch Brierley] had on your work?' [. . .Healthwatch officer] says that the best way to evidence our achievements [is to show] who uses us.Public Health and the CCG.If these organizations come to us, then we're having an impact.The important thing is showing how you are valued, how you are involved.It's not about the reports being used.'It's simply that we're invited to do these things'.(Fieldnotes, internal meeting, Brierley) Exactly what that activity involved, however, varied markedly between Healthwatchand this depended on the history, resources and relationships of each organization (cf.Table 1).While all five organizations' work aligned with the ideas of co-production and co-creation to some extent, some were able to realize the ambition of 'systematic involvement of relevant and affected actors' (Torfing et al. 2021, 190) in system governance more effectively than others.
In Ambria, a particular focus was collating insights about the lived experiences of different populations in the area and presenting them to commissioning and provider organizations.One example was its work as a convenor of voluntary and community sector groups around the county.The local authority commissioned Healthwatch Ambria to bring together a large forum, covering a wide range of interest groups who could identify the population's needs and provide input into health and care system planning.The forum was based on extensive background research by Healthwatch 'to establish what we would need for engagement, lived experience, coproduction' (Healthwatch officer, Ambria).It served as a means of 'getting that flavour, getting a broad sense of what's gone on' (Forum member, Ambria) and ensuring it influenced decision-making: 'a good channel to communicate information backwards and forwards and [. . .] an open door into the council' (Healthwatch officer, Ambria).
Cramford and Brierley, both covering inner suburbs of big cities, shared some characteristics.Both benefitted from senior officers with long histories of involvement in local politics and public service governance, allowing them privileged access to official forums of the kind that was not so readily available in Ambria and elsewhere.
[Healthwatch chair] says that he knows half the councillors, he helped them get elected -he wants them to know that, i.e. that he's an active member of the local party.Almost to prove his point, one of the councillors walks past and says, 'Oh, you're here for your meeting with [official]?' (Fieldnotes, meeting between Healthwatch and local authority, Brierley) Beyond that, their approaches differed.Its connections meant that Healthwatch Cramford enjoyed an established place on joint boards with other organizations in the system, allowing it to contribute as an active partner in key decision-making processes.It supplemented this work with projects focused on the healthcare needs of key marginalized communities, which it could then draw on in co-creating local decisions around planning, policy and practice.
Healthwatch Brierley convened its own forums focused on co-creating healthcare in specific areas such as young people's services and mental health, but the most notable part of its work was the development of what it called its 'community intelligence hub', which collated information from a wide range of sources, including its own surveys and forums, patient-related data from system records, and neighbourhood-level demographic and socio-economic data.This proved to be an attractive resource for local health and care organizations, which drew on it in decision-making and appreciated the way its use was readily recordable and auditable -simultaneously allowing them to demonstrate value and opening doors to influence.
[Healthwatch manager notes that] we are starting to get commissioners and providers using Healthwatch data in their own reports-they use it in their dashboards and they reference Healthwatch [Brierley].So the data from Healthwatch has become part of their regular reporting.(Fieldnotes, internal meeting, Brierley) Darnside, the smallest Healthwatch in our sample and one of the smallest in the country (Table 1), could not match this scope of integration into local decision-making.Its minimal employed staff (between them amounting to less than one full-time-equivalent post) covered official meetings to which they were invited, but much of the rest of its activity was driven by a sizable cohort of unpaid volunteers, with varying levels of time commitment.The most involved volunteers spent much of their time linking with local healthcare organizations, attending meetings and scrutinizing reports.This activity served effectively to engage the system and demonstrate activity, since these inputs in turn demanded further responses from healthcare organizations.However, its focus remained at the level of individual organizations rather than at the level of the system; moreover, for some, there were doubts about its impact beyond the performative: We're not representative of anybody, but still, when you turn up, [. ..] at least it's somebody from the outside and these things are not necessarily being well scrutinized.[. ..]We were just turning up and sitting at meetings and I think there's nothing wrong with that.(Volunteer, Darnside) Finally, Healthwatch Ethershire, though much larger than Healthwatch Darnside, also struggled to demonstrate its value to its local system.It covered a large, rural county, which made it challenging to find areas of common interest across communities that might offer fertile ground for co-creation.The area was dominated by a major healthcare provider organization with a strong reputation and multiple alternative routes for engagement with the public.It was hard, therefore, both to demonstrate a meaningful set of co-creation activities, and to command engagement from the organizations involved: [Healthwatch Ethershire chair] asks how [healthcare organization] use the reports that they provide.The response is essentially, 'We don't think that they're that important'.[Director of communications] answers, saying 'They are part of a suite of reports that we rely on.We don't want to rely on just one thing'.She's talking about triangulating the Healthwatch report with the social media with the patient experience data.[. . .Chief executive] says that 'we're used to huge volumes of patient feedback and Healthwatch [Ethershire] will struggle to get that sort of volume'.[Healthwatch chair] agrees that what they can do is not as big [. ..] though he does add that for the recent consultation about priorities they got as many as 700 responses.The others don't seem overly impressed by this.(Fieldnotes, meeting between Healthwatch and healthcare organization, Ethershire) Consequently, rather than the efforts to convene and engage in co-creation activities witnessed in Cramford and Brierley, Ethershire's focus tended like Darnside's towards simpler, single-organization issues, for example access to general practices, where it had the opportunity to show some impact while ensuring that its efforts covered the length of the county.
While enjoying ostensibly similar starting conditions, then, the five Healthwatch's activities differed in their focus, scale and influence, and in the extent to which they embodied the ambitions of co-creation.Healthwatch Cramford and Brierley seemed well embedded in the key decision-making activities of their local systems; Darnside and Ethershire were somewhat detached; Ambria was somewhere in between.Reflecting the findings of others (Ansell and Torfing 2021b;Bentzen, Sørensen, and Torfing 2020;Meijer 2016;Van Gestel, Kuiper, and Pegan 2023;Vedeld 2022;W. H. Voorberg, Bekkers, and Tummers 2015), the resources available to them, their networks, and the responsiveness of other players in the health and care system influenced the internal dynamics of co-creation.

The outputs of co-creation
Yet for all this divergence, commonalities were apparent too.
Almost universally, Healthwatch participants took care in how they positioned their organizations in relation to other agencies in the local system, conscious of the need to demonstrate activity and impact.This objective could be realized effectively through a collaborative set of relationships in which Healthwatch positioned itself as a system partner, rather than as a critical pressure group.This was a matter of deliberate choices on the part of senior Healthwatch decision-makers faced with the question of how best to secure influence in their local systems.They contrasted the role for Healthwatch as they saw it with that of one of their predecessor organizations, Community Health Councils (CHCs), which were seen as rather more confrontational in their disposition, and with other 'pressure groups': There are two camps: those who want to work within the system [. ..] and there are people [. ..] who are 'old-school CHC'.[She] likes that there are tensions but people expect her as chief executive to manage those tensions.It's up to her to explain the terms of reference of [Healthwatch Brierley] and its different roles (to others and to the board).So for example, explain what 'critical friend' means.(Fieldnotes, internal meeting, Brierley) There's an organization, "Defend Our NHS", which is anti-[private-sector provider] and it has presented its work to [local authority].But Healthwatch [Darnside] keeps its distance from them-our executive board says we need to develop a relationship with'.[private-sector provider] (Fieldnotes, internal meeting, Darnside) Healthwatch participants justified this stance in terms of its effectiveness: cocreation depended on trusting relationships, they argued, which in turn required that they be taken seriously as part of the system rather than be seen as outsiders.Ultimately, such an approach was more likely to achieve improvements of the kind that mattered to local publics: As we began to grow we realized that it was much, much better to form a strategic working relationship with the providers, with the commissioners and to become a critical friend.We've been far more effective by doing that.And now we can go in at very high level to adjust things that might be working wrongly or badly at a low level.(Healthwatch officer, Cramford) We are seen as being pretty central to the discourse of what's going on and not just a fringe group.[. ..]I don't feel co-opted, I feel much more interested in influencing-we're not powerful enough to go on opposition and win, that's the simple long and short of it.(Healthwatch officer, Brierley) Participants from Healthwatch thus articulated an implicit understanding of what was, and what was not, likely to gain traction, and thus influence of some kind, in their local systems.In doing so, they referred not to any explicit terms of reference for cocreation, or any rules regarding what was inside and outside their scope of influence: rather, their experience taught them which issues were worth pursuing and which were out of bounds.Likewise, participants from other organizations in local integrated care systems welcomed Healthwatch's 'participating in a constructive way' (Healthcare manager, Cramford): 'if they've built good relationships and adult relationships, when they are bringing critique to us about what we're doing in the system then it'll be easier to land it' (Integrated care system manager, Ambria).In all, this approach seemed to pave the way for constructive dialogue between Healthwatch and others, and secure influence on at least some decisions.
However, this disposition also meant that some activities were shunned.Healthwatch tended to focus on opportunities where their officers felt they could most readily make a difference, or where their priorities otherwise matched those of others in the system: for example, by representing the views of the public at large, or of specifically identified groups, typically those at risk of marginalization.Healthwatch maintained a cautious distance from other voices of the public that challenged system organization in a more fundamental way.An officer in Healthwatch Brierley put it frankly: Every couple of board meetings somebody almost invariably brings some leaflet along and says, 'There is a campaign [. ..] and you should do something'.Three or four years ago it used to be discussed; now it's dismissed.That's not what we do: we're not a campaigning organization.[. ..]I would rather put more energy into getting the hospital to finally do what they've said they've wanted to do for 18 months now, which is get us in to help them design clinical pathway studies.
Participants acknowledged that some important issues, including those that might be most transformative, were eschewed, but took a pragmatic perspective: 'You have to step back from it and say, "Well what are we really trying to achieve here?Is the greater good served by this approach?"'(Healthwatch board member, Cramford).
A partial exception was Healthwatch Ethershire.In the past, it had periodically sought to assert its presence and command greater attention through an approach that other organizations characterized as confrontational, including engaging in politicized debates where other Healthwatch feared to tread.Ultimately, the strategy had had limited success: the rest of the system hardened its stance towards Healthwatch, and the current leadership now pursued an approach that sought to reduce the risk of conflict with the dominant organizations in the system, for example by focusing on simpler, service-level issues of the kind described above.
[Healthwatch] made a deliberate decision to be much more independent and challenging, which is fine, it's in their role, [. . .but] our relationship did become difficult.[. ..]We had to feed that back to the local authority when it came to a review period.I think it was taken on board and I think we see less of that now.(Healthcare commissioner, Ethershire) Thus publics and issues that appeared to threaten systems' orientation towards consensus were ruled out of co-creation activities.Healthwatch deliberately constrained the scope of their contributions according to their perceptions of acceptability.The full richness of insights, ideas and critique from the breadth of the public that co-creation may offer was carefully filtered before it even reached discussion and decision-making forums: 'feasible' solutions took precedence over 'innovative' ones (Ansell and Torfing 2021a, 221).And when Healthwatch did not police these boundaries themselves, as in Ethershire, the rest of the system stepped in.
These patterns regarding the outputs of co-creation and the influence of Healthwatch seemed to harden over the course of the COVID-19 pandemic, which began mid-way through our fieldwork.Much of the activity of Healthwatch moved online, and some organizations took on very different roles in local systems (Zoccatelli et al. 2021).Among our sample, the changing roles and relationships engendered by the pandemic served to sharpen the contrast between those organizations that appeared most successful in making themselves central to their local systems -Cramford and Brierley -and those that were more marginal -Darnside and Ethershire.
Healthwatch Cramford and Healthwatch Brierley both advanced their existing respected positions in local governance.Healthwatch Cramford continued its work at the nexus of local government, political and civil-society activity; the work of the boards in which it participated intensified and its contributions appeared valued, and it secured additional funding from the local authority to lead community-based responses to the pandemic.Healthwatch Brierley was able to pivot its work rapidly towards Covid-19-oriented activities, and its intelligence hub offered a versatile resource that remained highly coveted by the system.
In contrast, Darnside's small Healthwatch's activities remained limited, as its volunteers turned their hands to supporting the community at a time of great need.Healthwatch Ethershire's more distanced relationship with other organizations in the system, meanwhile, meant it was left out in the cold during the pandemic.
The legislation for [public] meetings has been changed so that local authority meetings can happen remotely.But that's not happening in [Ethershire].This was just announced and we weren't even asked about our thoughts.[Chair] asked about it and was effectively told, 'We're too busy'.(Fieldnotes, internal meeting, Ethershire) The pandemic, then, served to solidify patterns of activity that were apparent in our cases.Those that had successfully positioned themselves as collaboratively oriented partners central to their systems' co-creative endeavours found their status enhanced.Those that were more marginal were further sidelined.As a Healthwatch officer in Cramford, reflecting on a conference call involving Healthwatch from around the country, put it: We're used to being a partner, not someone throwing mud.Other Healthwatch do the same as we do and they've got shedloads of work.What was clear on that call was that the people who work in partnership and in collaboration, they were just inundated because they were involved with the community response , this, that and the other.But the people who just [. ..] ask difficult questions and shame people, basically, they were screwed because they were like, 'Well these council meetings aren't happening anymore, or if they are, they're not public meetings that we can attend'.[. ..]They're a bit redundant really.

Discussion
Co-creation is seen as having great promise in the governance of public services and systems, but realizing the distinctive, innovative contributions it promises can be impeded by the institutional context, the degree to which it achieves inclusiveness, and the internal dynamics of co-creation processes, such as the distribution of power among stakeholders.We respond to calls for use of comparative methods to 'broaden the focus from service production' and examine the conditions that support or prevent co-creation and its impact (Ansell and Torfing 2021b, 5;Baptista, Alves, and Matos 2020) with a study of five Healthwatch organizations asked to play a leading role in fostering co-creation in their local health and care systems.We find that supportive top-down policy, and willingness among stakeholders (particularly the state agencies involved) to support co-creation, resulted in a context where Healthwatch's contributions were welcomed, and indeed where all parties had a stake in demonstrating their worth.Our findings align with those of Voorberg et al. (2015Voorberg et al. ( , 1344) ) on the importance of '(top-down) policy that supports co-creation' and confirm the importance of institutional support for co-creation (Torfing and Siebers 2018).We also find that Healthwatch's particular claim to be able to engage and speak for the local community, especially its more marginalized elements, was important in enhancing the legitimacy of the organizations themselves and of co-creation as a whole (Meijer 2016;Verschuere et al. 2018).In terms of institutional context and inclusiveness, all five Healthwatch organizations appeared to be well positioned to realize the benefits of co-creation.
In practice, some Healthwatch struggled more than others to deliver activities and associated outputs that achieved the aspirations of co-creation, in terms of 'a systematic involvement of relevant and affected actors [. ..] in fostering innovative solutions to complex problems' (Torfing et al. 2021, 190-91).This largely corresponded to their level of resources and the degree to which they had managed to establish firm collaborative relationships with other parts of the system.Where resources were scarce (Darnside) or where the system operated at arm's length from the work of Healthwatch (Ambria), Healthwatch's activities tended to be more consultative, and/or focused on bilateral relationships with individual organizations rather than co-creation at the system level.But perhaps more importantly, in four of the five cases, Healthwatch leaders alighted on an approach that was selective about which issues to champion and which ideas to bring into deliberation.This was not, as the literature might suggest (Jaspers and Steen 2019;Meijer 2016;Steen and Sanna 2018;Turnhout et al. 2020), a matter of inequalities of power within the co-creation process resulting in selectivity about which of its products to take forward.Rather, it came about because Healthwatch saw advantage in positioning themselves as trusted partners inside the system.Consequently, the solutions typically prescribed in the co-creation literature on how to manage the internal dynamics of co-creation and their impact on what cocreation creates are unlikely to be effective: this was not something that could be addressed by better mediation of interests on the parts of co-ordinators of co-creation (E.Eriksson et al. 2020), skills development (Steen and Sanna 2018), or a greater tolerance of the open-endedness of co-creation processes among public servants (Leino and Puumala 2021).
Rather, even before co-creative interactions began, they appeared to be regulated by an implicit sense of what was acceptable and what was not.Most Healthwatch were PUBLIC MANAGEMENT REVIEW careful to avoid issues that would place them outside the consensus, such as service reorganization or closure, or the question of the governance of the NHS itself, including its financing and the role of for-profit providers (despite evidence of wide public concern about contentious issues of this kind: see, e.g.Martin, Carter, and Dent 2018;Stewart, Dodworthand Angelo 2022).Working with the grain concentrated the labour of Healthwatch on those issues where there was common interest in change: broadening accessibility; improving quality; informing service redesign.Less tractable, but perhaps more fundamental, issues were put aside.What we see is perhaps akin to what Turnhout et al. (2020, 42-43) call the 'depoliticization dynamics' of co-creation, in which certain features of the matter under deliberation are taken as axiomatic, and those involved cannot easily stray outside this 'scientifically sanctioned rationality'.These dynamics are not universal -see, for example, E. Stewart's (2021) account of 'fugitive' co-production that takes place outside the invited spaces of healthcare governance -but they are perhaps more likely to emerge in settings where cocreation has become formalized and officially endorsed (E.Stewart 2021).Here, in consequence, co-creation became not a matter of the value derived from generative conflicts (Rossi and Tuurnas 2021) but a process in which consensus on the biggest issues had, by-and-large, already been reached.Steen et al. (2018) identify some of the potential 'dark sides' of co-production and co-creation, each of which represents an unintended consequence of its virtues.To their 'seven evils' we might therefore add an eighth: a failure to tackle issues that are of importance and interest to at least some of those involved arising from a premature orientation towards consensus.
How might we explain this tendency, common to most of our cases, towards a rather partial and conservative realization of co-creation, and a tendency to avoid those most innovative and transformative issues where its potential seems greatest?One might point towards the wider policy environment in which these ideas were being vaunted.We note above that policy support for co-production was palpable, but that its introduction as a 'big new idea' and solution for the woes of the English health and care system took place in a governance context in which the sediments of previous approaches still influenced behaviour (Dickinson 2016; Jones 2017).Even as the NHS's central leadership body discovered and extolled the language of collaboration and coproduction, it did so in a way that made clear that there were certain 'ground rules' that were not up for debate.For example, NHS England (NHS England 2016, 6, 13) made clear that 'the involvement of people, communities and stakeholders' should help to 'build ownership and support for proposals to transform health and care' and to avoid 'criticism '. As Hammond et al. (2019) suggest, NHS England thus embraced the role of a 'meta-governor', guiding the forms that co-creation came to take and making clear the boundaries of its territory.
Again, though, this explanation of the way co-creation was realized in our five sites seems incomplete.As a means by which states and other hierarchical authorities can 'design, frame, support and intervene in governance networks' (Torfing, Sørensen, and Røiseland 2019, 814), the role of meta-governance is acknowledged as important.It also holds important legitimacy, in ensuring that the participatory expression of democracy achieved through collaborative ventures such as co-creation are leavened by accountability to liberal and electoral democratic institutions (Sørensen and Torfing 2005).Yet in most sites there was no need for meta-governors to adjudicate what was permitted and what was not, or even for explicit terms of reference governing the boundaries of co-creative activities.We need therefore to look beyond approaches to meta-governance, such as framing, facilitation and participation (Sørensen 2006), that have been the mainstay of analyses in this field, and instead understand metagovernance in more discursive, less agential, terms (Jessop 2020).While studies have increasingly cast light on the agents of meta-governance (Ek Österberg and Qvist 2023), a more decentred analysis also has benefits.Meta-governance is not always traceable to a single authority (Gjaltema, Biesbroek, and Termeer 2020;Larsson 2020).An understanding founded in theories of governmentality would also locate metagovernance in its subjects, and the way they internalize an implicit understanding of their roles that is reproduced in their behaviours and their interactions with others (Martin and Waring 2018;Jessop 2020).Healthwatch officers themselves quickly gathered what was asked of them and learned to play by the rules of co-creation without explicit instruction; in the one case where they did not, other players left them at the margins.Drawing on a governmentality framework, Larsson (2020) argues that new collaborative modes of governance such as co-creation bring with them their own, 'post-neoliberal', form of governmentality, in which values of competition and selfsufficiency are replaced by values of co-operation and interdependence.Based on our analysis, we would suggest that just as they remain evident in the palimpsest of formal policy, these neoliberal values continue to infuse the subjectivities of those involved in co-creation, guiding them implicitly as to what might be within scope (for example, optimizing the quality and patient-centredness of service pathways by incorporating greater attention to the needs and preferences of service users) and what is not (for example, a questioning of the role of for-profit providers in improving the quality and efficiency of health and social care systems).This, then, was less a matter of metagovernance as the legitimate intervention of the state to ensure that co-creation served democratically selected objectives (Haveri et al. 2009), and more the implicit, almost imperceptible, influence of foundational values in marking out the limits of what cocreation could do.And here, the radical and innovative impacts that have been claimed to make co-creation distinctive fell beyond those limits.

Conclusion
Our analysis suggests that the increasing promotion of ideas of co-production in English health and social care by leaders in the system has indeed resulted in local contexts that are fertile grounds for co-creation.Healthwatch were readily recognized as key organizations in leading co-creative activities across systems, and were welcomed by other stakeholders in the system, who acknowledged that all those involved in health and care governance needed to adapt to more collaborative approaches to decision-making.The areas on which Healthwatch focused were ones where other agencies in the system recognized their limitations, and where they knew they needed help to avoid socially undesirable outcomes: ensuring patients' contributions to service redesign, for example, and avoiding exclusion of marginalized or 'seldom-heard' groups.Yet the forms taken by co-creation in practice were largely conservative and constrained.Even though they were not explicitly ruled out-of-bounds, Healthwatch officers knew that to be considered legitimate and serious players in the governance of health and social care, they needed to be selective about which issues they brought to the table.
These findings suggest perhaps the need for some moderation of the transformative prospects for co-creation set out by its more optimistic proponents, at least in settings characterized by sedimented modes of governance and still haunted by the spectres of previous forms.While co-creation might indeed come to represent a 'core principle of public governance' (Ansell and Torfing 2021b, 4) in many systems, the degree to which this will result in radical and innovative solutions that represent real departures from the status quo and incorporate the breadth of knowledge of public service stakeholders seems more doubtful.

Table 1 .
Key features of the case studies.