Introduction

The network governance narrative of public management reform sponsored by post New Public Management (NPM) authors (Pollitt and Bouckaert 2011; Christensen and Lsgreid 2007) promoted more integrated and systemic approaches to the delivery of public services designed to address the fragmenting effects of earlier NPM reforms (Ferlie et al. 2016). The different streams of governance have created what Jones (2018) terms a complex sedimented governance architecture. While the NPM approach emphasises managerial authority and top-down implementation, network governance works through collaboration and partnerships. In this binary theoretical field, the decentred theory of governance offers a different perspective that focuses on how ‘people see the world’ (Bevir and Rhodes 2001). In this paper we use a bottom-up decentred lens to understand how local actors perceive reforms and use their agency to make sense of the changes in the delivery of stroke services.

The decentred theory highlights the importance of key actors’ beliefs and traditions, crucial aspects that are rarely analysed. It does not seek to ‘tame an otherwise chaotic picture of multiple actors, creating a contingent pattern of rule through their conflicting actions’, instead it embraces this messy experience and uses it to explain how governance occurs in this disordered reality (Bevir and Richards 2009). Bevir and Richards (2009) see everyday practices arising from situated agents whose beliefs and actions are informed by traditions and that these traditions can be detected as habits and practices. The decentred approach sees ‘practices’ as the sets of actions pursued by actors, which are influenced by their beliefs. Within organisations, these practices come to be viewed as ‘traditions’; the accepted ways of doing things. It is from these traditions that situated agents select their broad beliefs concerning the appropriateness of certain actions. In turn, these beliefs may change when confronted by ‘dilemmas’. A dilemma can be defined as an idea which—if it stands in contradiction to another established belief, practice, or tradition—poses a problem for individuals or groups. According to this approach such dilemmas can only be resolved by either accommodating or discarding the new idea (Bevir and Rhodes 2007; Geddes 2014). We add to Bevir’s approach by exploring the roles of context, audit and leadership at the micro level.

We consider the decentred approach as particularly well suited to analysing and explaining the micro implementation of a health care reform. The aim of this chapter is to use this perspective to better understand how a major health care reform was implemented ‘on the ground’ in two contrasting National Health Service (NHS) Foundation Trusts. Foundation Trusts are semi-autonomous hospitals within the National Health Service (NHS) in England that have various managerial and financial freedoms and relative independence from the Government’s Department of Health and Social Care. Through two qualitative case studies we investigated the implementation of reforms to NHS stroke services that began in 2007 in England, which were the result of a culmination of a longer effort from within the sub-specialism of stroke. The stroke reforms can be characterised as a ‘disruptive innovation’ in that it challenges existing stakeholders (commissioners, providers, and other interested parties) to rethink their current practices through the introduction of new ideas. These potential challenges to the prevailing arrangements, may conflict with the existing beliefs and traditions of local actors. This requires us analytically to identify the dilemmas (both operational and clinical) and existing traditions (as faced by the different actors but principally, clinicians) and the situated agency of these individuals within their organisations respectively. A decentred approach to governance (Bevir and Rhodes 2007; Bevir and Richards 2009) allows us to highlight the importance of situated agency and local traditions in the face of such dilemmas (Fraser et al. 2019).

Through our study, we want to prompt new thinking around the decentred concept of ‘dilemmas’. Can dilemmas only result in intractable tensions between two sets of beliefs, practices and traditions that individuals hold and then must resolve? Our findings offer an empirical base on which to build on Geddes (2014) call for “a more comprehensive articulation of the concept of ‘dilemma’.” However, unlike Geddes whose focus is on the macro political level, we examine the micro level of policy implementation. Before we go on to explain our methods, we provide a contextual summary of how stroke care has been transformed in the past decade or so.

Developing Stroke as a Specialty

Globally, stroke (referred to as apoplexy until the middle of the twentieth century) has been a leading cause of death and disability for many years (Johnston et al. 2009). The disease predominantly affects the elderly and historically there has been a lack of effective treatments (Pound et al. 1997). Drawing on Foucault’s (1973) work on the ‘spatialisation’ of disease, Daneski et al. (2010) give a detailed history of the treatment of apoplexy and stroke, highlighting in particular, the different ways this disease has been conceptualized over time, the complexities around different types of stoke (ischemic and haemorrhagic) and its development as a condition dealt with by geriatricians through the twentieth century. It is only over the last 30 years, through the establishment of a distinctive body of knowledge and the accumulation of evidence of treatment and effectiveness that stroke medicine has become an important clinical specialism in its own right. Two significant evidence-based developments have been notable in the fashioning of modern stroke care from the late twentieth century. The first is essentially organisational—the development of specialised Stroke Units (SUs) and Hyper-Acute Stroke Units (HASUs), where stroke patients receive specialised multidisciplinary care (first in a HASU and then ongoig treatment and rehabilitation in an SU) (Stroke Unit Triallists’ Collaboration 2007) rather than on general medical or care of the elderly wards. The second, is diagnostic and linked to the proven effectiveness of thrombolysis—a drug that offers a radical improvement in outcomes for certain stroke types (NINDS 1995). This international best-practice was translated into NHS policy through the National Stroke Strategy launched in 2007 (Department of Health). This transformation around the medical discourses of stroke was part of an international trend over the last two or three decades (Baeza et al. 2016) leading to stroke being categorised as ‘a preventable and treatable disease’ (RCP 2008) with new pathways for stroke patients inscribed in government policy over the past decade in the English NHS (Department of Health 2007). These developments posed complex organisational challenges for individual hospitals, clinicians and strategic level managers across England and Wales. Essentially the stroke reforms focused on the maxim of ‘time is brain’, resulting in stroke being regarded as an emergency and not an elective condition (Department of Health 2007). People with a suspected stroke should be immediately transferred to a hospital providing hyperacute services throughout the day and night—this includes expert clinical assessment, rapid imaging and the ability to deliver thrombolysis. They should receive an early multidisciplinary assessment and have prompt access to a high-quality stroke unit (Department of Health 2007). Regional strategic stroke plans were rolled out in different locations from 2008, influenced by the Royal College of Physicians (RCP) and guidance from the National Institute for Health and Care Excellence (NICE). By 2008 the transformation of stroke care was such that the RCP and NICE stated that:

Over the last two decades, a growing body of evidence has overturned the traditional perception that stroke is simply a consequence of aging which inevitably results in death or severe disability. (NICE 2008, emphasis added)

It is also important to stress that the proposed stroke service changes were implemented in full, on time and faced very little opposition from the relevant stakeholder institutions, clinicians or the wider public, this is unusual, marking the case as a ‘positive outlier’ (Flyvbjerg 2006; Fraser et al. 2017). The reforms enjoyed broad political and popular support (Boseley 2014). A number of studies have explored local developments linked to the strategic drivers of the 2007 national directives (see Fulop et al. 2015; Turner et al. 2016; Fraser et al. 2017; Fraser et al. 2019). The external forces of organisational (SUs) and technological (thrombolysis) innovations have prompted large change in the treatment of stroke, which has overturned the traditional perception of stroke that had been the norm. These innovations have produced potential dilemmas, and it is these that we want to examine at the micro level through two case studies whose methods we describe below.

Our Study

As part of a European Commission seventh framework funding programme, a comparative study examined the implementation of evidence into practice in stroke services in England (2 hospital sites) Sweden (2 hospital sites) and Poland (1 hospital site). This multiple case study approach (Yin 2003) enabled us to develop credible case and cross case analysis to establish the internal consistency of the information gathered and use this analysis to develop theoretical constructs from the data (Eisenhardt 1989; Eisenhardt and Graebner 2007). Our aim was to collect rich descriptions about their perspectives of evidence-based practice and the stroke care reforms. This chapter only uses the data from the two English case studies in order to focus on divergence at the micro level. The two English sites differed in terms of urbanity; patient demography; influence and existence of competing hospitals; difficulties in attracting and retaining skilled staff; and differences in community care arrangements.

Informants were purposively sampled to represent the different managerial and professional groups involved in delivering stroke care. The purposive sample included a range of both clinical and managerial staff from the hospital based SU, emergency medicine, radiology, ambulance service, community rehabilitation services, including physiotherapists, occupational therapists, speech & language therapists, dieticians and psychologists, commissioners of services and GPs. A total of 45 interviews were carried out in these two case studies. Below we provide brief profiles of the two English case study sites.

Case Study 1 (CS1)

This is a district general hospital in a non-metropolitan region which serves a population of 230,000 (with pockets of deprivation) and a staff of 3300. The hospital was established in 1981, has had foundation trust (FT) status since 2006 and is affiliated with the local medical school. It is in an unfashionable location and is less prestigious than other specialist hospitals that are in ‘more attractive’ areas of the region, however, these are too distant to compete in terms of stroke patients. On a macro level, the 2007 National Stroke Strategy provided a push for improvements in the delivery of stroke services, and at a meso level, the host commissioner has invested heavily in improving the hospital’s stroke services.

Case Study 2 (CS2)

This is an outer London teaching hospital, serving a population of 500,000 with a staff of 4000. The hospital’s SU has been upgraded to a large combined hyper acute stroke unit and SU since 2008. Prior to this, the hospital had a comparably poorly staffed, non-exclusive ‘proto-stroke unit’ with 23 beds, often occupied by non-stroke patients, which meant that many stroke patients never made it to the SU until many days post-stroke. The hospital had to invest very heavily and rapidly in terms of medical and radiological resources as part of the London stroke plan to arrive at a point at which the minimum standards were achieved—more so than other, better established sites.

Findings

The stroke reforms and the subsequent reconfiguration of stroke services resulted in a series of dilemmas that we explored in our two case studies. We followed the decentred approach to reanalyse these data and three important themes emerged that we develop below.

The Importance of Context in Shaping Beliefs and Traditions

Contextual factors are important in shaping individual beliefs and traditions and so these need to be considered when we examine how a dilemma is dealt with by individuals. Here we consider the stroke reforms as a potential dilemma that could cause problems but also present opportunities that were dependent on contextual factors, which were very different in the two case studies.

The context for CS1 was of a hospital that was perceived as a poor performer in stroke care by the local service commissioner, whose medical director provided a damning assessment:

They weren’t doing the right things, and the things they were doing, they weren’t all doing right. (Medical Director of the local service commissioners; CS1)

This hospital’s location was also important, it was geographically isolated which impacted negatively upon its ability to recruit good staff, which was cited as a contributing factor to its historical poor performance:

We have huge difficulty recruiting, because we’re on the edge of the world, you see. […] half the catchment area that you would normally have people from, is sea here. So, unless there’s oil rig workers that want to come and help us, we’re stuck. (Nurse consultant, CS1)

A context of historical poor performance and geographical isolation therefore shaped the local actors’ beliefs and traditions and this has had an influence on how they responded to the perceived dilemmas to change their established way of working and improve the delivery of stroke care.

The context for actors in CS2 was quite different as the hospital saw itself in a strong strategic position as a result of the 2008 London Stroke Plan, which reconfigured stroke services in ways that favoured outer-London hospitals like this one. Winning the bid to have a HASU was perceived as vital for this hospital as it would enable it to gain more resources not only for stroke services but for the hospital as a whole. However, to achieve this, it would have to switch from its traditional management approach that was predominately operational to a new more strategic one. There was a recognition that changing the prevailing management traditions needed outside, specialist, strategic management expertise to develop a succesful HASU bid. The management consultant who was bought in explained the change in management that was needed:

[M]ost Trusts are heavily operational. They’re not - they don’t tend to be very strategic because it’s all about the crisis today and next week… And what I was trying to say was, ‘Look, you know, things are not going to be as they always were. They’re going to move – special stuff is going to move to tertiary centres… So that’s a key element to your strategy which they hadn’t really been thinking about… [A]nd you get a kind of critical mass of those kind of services [e.g. stroke]. And then you become the default [hospital for other services]. (Independent Management Consultant; CS2)

It is important to highlight the role of the regional stroke networks as a contextual factor, as a dominant theme for CS2 much more than in CS1. These provide a regional network for supporting regional innovation and improvement through supporting high-level strategic planning and sharing of best practice. It is also the case that the extent of the local reforms was much more radical in CS2 (because of its London location) than in CS1. This may be linked to the differences in demographic and geographical challenges faced in the two sites. In CS2 the regional stroke networks were integral, for example, in working with senior strategic managers around decisions related to not only the upgrade of existing services to deliver better stroke care through HASU provision—but also the downgrading and closure of some existing units (Fulop et al. 2015; Turner et al. 2016; Fraser et al. 2017). Deciding how many HASUs would be needed in CS2 and where they should be located provoked significant dilemmas for members of both the formal networks alongside other interested actors. This ‘disruptive innovation’ was skilfully managed in CS2 and provided an opportunity for a radical redesign, not only of how best to deliver stroke care to the regional population, but also a significant break from past models of stroke care as the disease was effectively redefined as a regional concern as opposed to an institutional concern (Fraser et al. 2017). Winning the HASU bid would positively change things for this hospital, thereby creating a new situation that can produce opportunities and problems. This might not fit exactly with the decentred concept of a dilemma but nonetheless it creates an interesting new reality that impacts on local actors, which produces positive and/or negative outcomes depending on the context. Another significant contextual aspect of the stroke reforms was the heavy use of audit, which we explore in the next section.

The Role of Audit in Shaping Attitudes and Behaviour

Checking and inspection in the form of audit is often resisted by health professionals who tend to desire freedom of practice (Foster and Wilding 2000). However, our study suggests the audit regime ushered in by the stroke reforms was mainly welcomed by senior clinicians. Despite the pressure placed upon her, the senior stroke nurse informant in CS2 suggested that the new audit culture around stroke services had productive value for her as a manager because without it, ‘things would slide and slip’. This perspective chimes with Idedema and Rhodes’ (2010) idea of surveillance as an “ethic of care for self and other”. Whilst the senior stroke nurse clearly disliked some elements of audit and the pressure which it placed upon her, at the same time she was very willing to embrace it in order to achieve the mandated goals of the new model of stroke care.

It’s, first and foremost I think it’s the patient care and safety. And also, they are standards that are set within the stroke network which have to be met. So according to these standards we are also actually being assessed according to the evidence that has been given that – so if you don’t meet, if you don’t practice evidence-based treatments, you’re bound to fail on your standards, so it brings down the whole, the Trust as a whole. So, I think those are the motivations. (HASU coordinator and nurse; CS2)

Networks were central in CS2 to the establishment and embedding of new audit practices and the pursuit of standardisation in stroke care. Networks encouraged competition through transparency alongside broader collaboration and emphasised the productive potential of disciplinary power (Foucault 1973; Martin et al. 2013). The networks encouraged an organisational focus of audit (as opposed to a focus on individuals or clinical teams). Because the networks were collaborations between professionals and managers, they drew on both viewpoints, and also, they were considered to be locally rather than externally constituted and hence had high levels of legitimacy.

A senior nursing colleague in CS2 agreed that despite being critical of elements of the ‘bloated’ audit regime, it is necessary because it ‘pushes’ them to improve care. The disciplined and pressurised nature of increased surveillance can be channelled in productive ways by senior clinicians in order to improve clinical practice. This is particularly so with respect to nursing, which was seen as a problem in this trust, a problem that a strict surveillance apparatus can address. These data consistently showed how respondents in this hospital highlighted the ways in which the external audit demands allowed them to monitor and govern the competence of the nursing staff locally, illustrating the strength and enduring nature of hierarchical governance, as well as the perhaps unexpected welcoming of an audit regime by health professionals.

For CS1 the new strengthened audit regime was also welcomed but for different reasons, it bought the prospect of more resources to tackle their historical under achievement in stroke care:

Yes, I mean we, our Sentinel audit [audit of stroke service performance] is never good and it hasn’t been historically, because we’ve not had the right, you know, we’ve not had the right resources in place, we’ve not had the ward in the right area, enough beds. So, I’d be really interested to see what happens. (Matron; CS1)

The role of national guidelines was cited as significant in CS1, and the support of the local service commissioner was also essential in terms of funding. There seemed to be enough capacity within the stroke service to meet the government targets for patients to be treated on an SU, indicating that new ways of doing things (ideas) can become traditions if there is a supportive environment (resources). When it came to why evidence-based practice and audit performance might be important to institutions, economic reasons were frequently cited:

Well obviously there will be quality issues, which are similar to what I would look at on a personal level. But also obviously there are financial drivers as well, […], obviously, PCT [local service commissioners] who are providing a lot of our funding and stuff, you know, they will want to be seeing that we are giving best to our patients […] But that obviously has an impact on the Trust at how much revenue that is bringing in, as to whether we don’t get full payment if we’re dropping below a level and then there are certain sort of carrots dangled if you meet certain extra criteria, that there is extra monies available, so obviously that’s very important to the Trust as an organisation. (Senior sister SU; CS1)

There are a number of carrots and sticks that add credibility to the new governance regimes brought in by the stroke reforms and finally, the implicit idea that failure could ‘bring down the Trust’ in some way, adds a further pressure to conform or more accurately a pressure on senior managers to achieve conformity for the sake of the hospital as a whole. Somewhat paradoxically, these new contextual pressures can also be seen as enabling the clinical managers to use the new ‘dilemma’, in the shape of the stroke reforms, to motivate and/or demand staff to change their beliefs and practices.

The new context of digitally delivered and data heavy requirements of audit led to a performative disciplinary framework (Spicer et al. 2009) centred on professional reflexive action as this nurse manager explains:

It’s our professionalism and we also need to feed back to the Department of Health the things we’re actually doing. So, we’re actually achieving governance within our own unit. We’re actually monitoring ourselves, we’re actually achieving targets, and we’re quite open if we’re not achieving, and what we actually do to change. You can have eight very good weeks or something and then very bad week. So, what happened in that bad week? How do we tackle that week? How we are doing and prove what we’re doing and it’s about developing ourselves. (Ward manager and nurse; CS2)

The stroke reforms, with their ‘audit heavy’ aspects of practice also managed to disrupt and challenge the previous beliefs and traditions of professional autonomy. Professional autonomy is now reframed as the ‘freedom’ to provide excellent treatment:

I think people will talk to you about, you know, are you allowed to have autonomy to do things badly? And that’s ultimately the bottom line, is do you allow people to give suboptimal treatment? And I think what we’ve established is, no you’re not allowed to give suboptimal treatment. (Doctor; CS2)

This is an interesting quote because it presents the view of a senior consultant highlighting the problems with previous traditional beliefs of clinical autonomy. Essentially, if autonomy means freedom to deviate from the new agreed way of delivering stroke care, then such autonomy has become unacceptable and so this belief needs to be changed and recast in the new reality.

The issue of this new autonomy was illustrated by another doctor in the hospital who felt that the power of the stroke reforms was rooted in the fact that they chimed with his longstanding clinical beliefs and traditions, as opposed to other regulations that challenged them, again the stroke reform dilemmas were producing positive organisational and individual outcomes:

[I]n some ways [I feel my autonomy is] augmented [by the stroke reforms], because I have the authority of the Trust to basically drive through things that I feel are urgent, urgently. So, I suppose if you didn’t agree with it, I suppose my autonomy is threatened by the four hour [A and E target] wait, but it’s probably not threatened by [the stroke reforms], but that’s because this aligns with our clinical priorities, whereas the other thing doesn’t. (Doctor; CS2) (emphasis added)

This was also the case in CS1, with the Early Stroke Discharge (ESD) initiative that this therapist describes as a positive innovation:

It’s [ESD] coming from on high, I guess, inasmuch as it’s part of the accelerated stroke improvement programme from the Department of Health downwards, and so I guess it’s come from on high. But obviously the clinicians on the ground have always known that the patients would benefit from rehab [rehabilitation] in their own home, rather than in the hospital. (Senior lead of therapies; CS1)

These data illustrate how traditions and beliefs can be successfully challenged and reshaped if the proposed modification fits with actors’ individual and professional longstanding beliefs. They also show the enabling and productive possibilities of organisational surveillance (Iedema and Rhodes 2010; Martin et al. 2013).

The Importance of Leadership in Implementation

Finally, we used the decentred approach to explore the role of leadership, an aspect that wasn’t found in the existing decentred literature. Our data show that leadership was another key factor in shaping how the stroke reforms were perceived and implemented by actors within the two trusts. Leadership was significant at different levels, firstly, in CS1 it was seen as important by the local leaders that the national stroke reforms were accompanied with resources that would enable them to address their historical deficiencies in stroke care.

However, the local leaders also pointed to the fact that the drive for a change in local practice came from outside and not from them internally. The push factors were characterised as emanating from national guidelines and regional initiatives that they needed to implement:

It’s an external, very external driver, so because we didn’t do it, the research proved that, the research trial that had been done proved that patients who had the thrombolysis did very well. (Stroke nurse specialist; CS1)

This evidence presented these nurse leaders with a dilemma, the dilemma being that their current stroke care was not producing good outcomes for their patients, the patients were being failed and these nurses were motivated to lead a change in practice to confront this dilemma. It is usually powerful hospital consultants as clinical leaders that can enact change but there was a perceived lack of this type of medical leadership in CS1, which was a crucial problem that needed attention for improvements to occur. Another stroke nurse specialist stressed the need to improve the internal leadership capacity in order to enact change:

Leadership within the stroke team has been perceived as a problem historically. And it was one of the things that we really tried to tighten up on in the last six months. (Stroke nurse specialist; CS1)

The problem that the nurse cited was that there was a lack of leadership from the senior consultants in the trust and this was a causing a blockage to change. Using the decentred approach, we can explain how the dilemma that the stroke specialist nurses were experiencing (delivering a poor level of care) led them to use their situated agency to lead a change in practice, to modify traditions in response to a dilemma.

A consultant explained how the fact that the leaders in this change process were nurses and had been recruited into the trust was an important factor in ‘winning others over’, particularly other nurses, in the hospital:

most of the leadership has come from elsewhere. And it’s come from a very motivated stroke nurse specialist that we brought in from elsewhere and who had been doing this before, who is, you know, just one of those people that knows their subject, is passionate about their subject and communicates well. And that just sort of sells it to half the Trust. And I think, to some extent, I think the fact that it’s been a nurse specialist that’s not quite been, let’s say, not actually been leading, but to a large extent has been driving the agenda, I think has made it a lot easier to sell, certainly to the A & E nursing staff. I think if it was a stroke consultant coming down to this department trying to tell them what to do, they’d be saying, ‘Oh I don’t understand that, that’s all above my head, you clever doctors, you know, off you go and speak to our consultants. (A&E Consultant; CS1)

The data from case study site 1 illustrates that the traditional power of the medical profession can be sidestepped by motivated nursing professionals in order to achieve change. However, this internal hospital view contrasted dramatically with that of the medical director of the local service commissioner who felt that it was only strong managerial and medical leadership that would be able to drive the necessary changes in the hospital’s stroke services, nursing leadership on its own, in his opinion, was not enough. Here, he sums up what he saw as the hospital’s continued failure and explains what he sees as the reasons and possible solutions:

They [the hospital] didn’t set up an acute [stroke] service for some time. They managed to set up an in hours thrombolysis service. But not at weekends […] The business case was to deliver 24/7 thrombolysis. […] My principal focus is service transformation, and one of the things I’ve observed is that getting service transformation without a clearly definable clinical leader is very, very difficult. […] [T]here is a vacuum of clinical leadership in stroke at the hospital. […] Now, I think if they had had a clinical champion, it would have gone faster, but there was nobody with a voice at middle or senior management team level… (Medical director of the local service commissioner; CS1)

This view from this medical director was that change was difficult without senior medical leadership and that nursing leadership on its own was unlikely to be as effective. Which, contrasts with the hospital’s senior consultant’s view that the nursing leadership was why thrombolysis was implemented at all and would not have succeeded if it had been led by doctors. In terms of the decentred approach the key question is, what type of leadership is the most effective way of reshaping beliefs and traditions, and these data suggest that the answer is very dependent on the micro context. However, the medical director also recognised the limits of simple hierarchical governance at the micro level:

I guess the other aspect of that is the issue of being held to account, so it’s clear that it’s a regional priority, it’s clear that it’s a national priority, the DH [central government] kicks the SHA [regional health planner], the SHA kicks the PCT [local service commissioner], the PCT kicks the provider [hospital]. The boot that’s on the DH’s foot is big and spiky and it hurts the bottom of the SHA when it makes contact. The boot of the SHA is big and spiky, and it hurts the bottom of the chief executive when it makes contact, because both of these spikes are P45 [a notice of dismissal]. There is no P45 issue that one can reason with the [hospital]. It’s contractual, it’s difficult, and it’s highly complex. (Medical Director of the local service commissioner; CS1)

This quote vividly illustrates the perceived limits of simple top-down implementation. From the perspective of informants within the hospital, influence from respected leaders (specialised nurses in this case) seemed to be successful, while the local service commissioner informant feels that hierarchical authority is the only way to achieve radical, lasting change.

The decentred approach is useful as it sees beliefs and traditions as resilient to simple top-down governance, instead the different actors’ beliefs and traditions need to be considered and understood by local leaders. In the first case study the hospital consultant felt that this was best done by a stroke specialist nurse, rather than by medical leadership. The medical consultant’s views of nurses from CS1 are in total contrast to those in CS2. There was a notable tone of disdain around the competencies and abilities of nurses from internal senior figures at the CS2 hospital. Senior doctors expressed a desire for their nurses to get back to basics—‘not touchy-feely stuff’—based on ‘hard facts’ and ‘testing’ and increased discipline through the use of audit. This type of language was also used by other respondents and perpetuated a narrative and belief in which nursing in this hospital was seen as failing, which was noted as being a historical belief that was resistant to change:

The [general] nursing has got a lot better, but they’re still our weakest link and we have, I mean it’s really difficult to know what to do about that… Again, I think there’s a difference between the professions. The therapists are very self-motivated. They will ask questions, they will go and look up answers, they will set up teaching groups for themselves, they will make themselves aware of what’s going on. The nurses don’t seem to do that. Again, I may be being very nasty to the nurses, it may well be that they spend so much time wiping bottoms and cleaning up vomit and that sort of thing that actually they don’t have the energy and that is fine, because when push comes to shove nobody else does that work and that’s their ultimate goal. (Doctor; CS2)

The above quote from a senior consultant illustrates some very clear, generalised and entrenched beliefs of the perceived competences, skills and goals of different professional groups. These beliefs have developed over several years and so are difficult to alter which led to general nurses in this hospital being closely regulated and controlled by the clinical leaders, who were primarily doctors. When compared to the situation in CS1, it shows that power is contextually situated, and the decentred approach encourages us to examine these contextual micro factors. The internal hospital leadership of the stroke services in CS2 was by contrast very aspirational and eager to develop their services as this manager made clear:

[I]t was very clear from Darzi’s [a leading surgeon who was appointed by the Labour government in 2007 to undertake a review of the NHS] work that he wanted these specialist centres and then little sort of polyclinics in a little part. So, what we did not want to be was [unimportant]; we really wanted to become one of the players. (Clinical Services Manager for stroke; CS2)

Driving the change process here was a desire from a senior manager to become ‘one of the players’ and the default option to colonise further specialisms on the back of the success of securing stroke specialist status. These new competitive beliefs clashed with the previous traditions that were more collaborative and network based. Post-reconfiguration stroke services would have an explicit and quantifiable value to the management of this hospital because it would need multiple ancillary services in order to function at the requisite ‘gold’ standard designated by national and local guidelines and the funds for these services would be guaranteed by the commissioners. This would make stroke service delivery central to the successful clinical, economic and organisational functioning of the hospital as a whole. The importance and significance of the stroke services and their workers would all increase. Although these are positive changes for the stroke services and for the whole hospital, they are also disruptive and challenging. For example, this was evident with respect to the collegial nature of consultant decision making and workforce planning within the hospital. Notwithstanding the increased importance of stroke to the organisation, the lead stroke consultant had no direct power to force other hospital services such as radiology to change their traditional practice in line with his own desire that radiology offer consultant on-call coverage for thrombolysis decisions which is linked to the explicit requirements of the London stroke care model. There was a need to change the consultant radiologists’ longstanding traditions which could not simply be achieved through hierarchical governance, but instead would need careful negotiation and understanding.

Again, the regional stroke networks in CS2 were also important as they combined clinical and managerial knowledge, power and legitimacy, and thereby provided a focal point for leadership and strategic direction. The stroke networks here existed in a unique space between the senior strategic management at a regional level, the local commissioners and the local hospitals. Simultaneously however, they drew on elements of all three through their diverse membership. Whilst none of the network actors derived hierarchical jurisdiction, or line management authority over other members of the network, they were able to harness their ‘soft’ powers of pursuation and collegiality to encourage change.

Discussion

The implementation literature notes that policy and innovation implementation are not rational, top-down and linear processes (Nutley et al. 2007; Hill and Hupe 2002). This chapter highlights this assertion and sheds light on the reality of policy implementation at the micro level. Firstly, we found that several contextual factors of each hospital had a significant influence on how the stroke reforms were received and implemented by key actors in the two instituitions. Although macro policy is important, the data suggest that top-down implementation is limited, as governance is locally reproduced through context-based narratives and frames. This was illustrated by the different types of leadership in the two case studies, in CS1 it was the nursing leadership that led to changes and improvements in stroke care, something that would have been impossible in the CS2 context where nursing was considered weak by the key actors. Secondly, it was surprising to find that the introduction of greater surveillance in the form of audit was welcomed by staff in general, and senior staff in particular who used it to govern junior staff and highlight the need for more resources in stroke care. Lastly, in terms of leadership, our findings begin to question the traditional view of medical supremacy in local leadership as already mentioned. Below, we expand on these three findings and discuss the strengths the decentred approach offers to health policy analysis.

We used the ‘decentred’ governance approach (Bevir and Rhodes 2007; Bevir and Richards 2009) to analyse whether, and if so, how, and to what extent, specific micro traditions were challenged by new ideas and beliefs that the NHS stroke care reforms introduced in two English hospitals. The decentred approach offered an illuminating way to analyse health reform by highlighting the importance of dilemmas, beliefs, traditions and the local political contests that arise from these, which allowed us to unpack the context of policy implementation at the micro level. Our micro level data show the importance of local context in both shaping and then adapting local actors’ traditions and beliefs and thereby influencing their responses to the resulting dilemmas. These contextual influences are then important in understanding how a macro level policy directive is locally interpreted and implemented, as Bevir and Rhodes (2009: 31) state:

The workings of any policy or institution depend on the ways in which all sorts of actors interpret and respond to the relevant directives.

Our data suggest that dilemmas are individually constructed, and the decentred approach indicates the importance and influence of micro level actors in how reform is governed and emphasises the need for a bottom-up approach to implementation. A top-down, centrally driven approach to reform is unlikely to consider local context within which individual actors will interpret and respond to the relevant directives. Furthermore, these individual responses will be collectively reinforced at the micro level and thus influenced by the local context, where the leadership will be important.

Our analysis contributes to how we might view and define the decentred concept of a dilemma. For example, the heavy use of audit in the stroke reforms could have posed a dilemma for the stroke clinicians, as it potentially challenges their clinical autonomy to direct their practice (Timmermans and Berg 2003). However, both doctors and specialist nurses in CS2 used audit to govern both the work of junior staff and other colleagues. They used a potential dilemma (audit) to challenge and shape other local actors’ existing beliefs of professional freedom. In CS1, audit helped highlight the deficiencies of stroke care, which could have caused a dilemma for the stroke care clinicians, but instead they used the audit data to successfully argue for extra resources, an example of disciplinary power being used as a productive force (Martin et al. 2013).

Elsewhere, we have discussed how professional jurisdictional power does not simply rest with the medical profession but is more nuanced (Baeza et al. 2016) and this chapter illustrates the importance of leadership in local implementation. Our data demonstrate how stroke specialist nurses can be effective leaders and can be more influential than some senior doctors in implementing change at the local level. In CS1 we found that the specialist nurse successfully took on a leadership role that was left by a medical leadership vacuum and this enabled her to make significant changes to improve local stroke care. This local leadership was crucial in shaping and reshaping the beliefs and traditions at the micro level.

Our paper shows, through empirical data, the promise of the decentred approach to health policy analysis. This approach could be used to analyse large scale, complex health reforms, for example, NHS England is encouraging health and social care leaders across local systems to come together to produce sustainability and transformation plans (see: https://www.england.nhs.uk/integratedcare/stps/view-stps/) to cope with the twin challenges of intense productivity pressures and reductions in funding. These new partnerships involve a wide array of local actors who have diverse perspectives of how their local health services should respond to the productivity and funding dilemmas. By analysing the various actors’ beliefs and traditions within the different local contexts the decentred approach can offer a rich examination of these large-scale service reforms.