Once policy has been developed and agreed upon, it requires implementation. This phase is rarely straightforward – indeed it has been termed the ‘Achilles’ heel’ of the UK political system (Harris & Rutter, 2014). The idea of simply moving from one stable state to another as a result of planned change is widely acknowledged to be at odds with work on complex systems (Williams, 2015). Explanations for how and why policy does, or does not, become implemented include everything from the initial policy design to the resistance of the local population (Spillane et al., 2006).
Rather than just let policies drift into full or partial failure, governments are beginning to take an interest in ways in which the implementation of policy can be strengthened and supported. This interest includes improved preparation of a policy, its prioritisation, better tracking of policy to assess its impact, and – the subject of this paper – implementation support. Given the relative novelty of these types of approaches, little exists by way of an established evidence base testifying to their success or otherwise.
To help tackle this deficit, an evaluation of the Implementation Support programme introduced when the Care Act 2014 was passed in England, ending in the first quarter of 2016, was undertaken. Adult social care is a devolved matter and there is separate legislation determined elsewhere by the Scottish Government and Parliament, the Welsh Government and Assembly, and the Northern Ireland Executive and Assembly. The Implementation Support (IS) programme was confined to England, therefore the arrangements elsewhere in the UK are outside the scope of this study.
Following a brief overview of the Care Act and its purpose, and a few reflections on why implementation support merits attention, we present the key findings from our study. The study was one of a suite of research projects examining various aspects of the Care Act’s implementation. It was conducted by a team based at the University of Kent and Newcastle University. The views expressed in this paper are those of the authors and not necessarily those of the National Institute for Health Research (NIHR), the Department of Health and Social Care (DHSC), or its arm’s length bodies or other government departments.
The term ‘adult social care’ is commonly used to refer to personal care and practical support for older people, adults with physical disabilities, learning disabilities, or mental health issues, as well as support for those caring for them informally. The use of the term has become established in most governmental, professional, and academic writing, largely replacing the former general terms such as ‘personal social services’ and ‘community care’ (Gray & Birrell, 2013). While earlier conceptions stressed the residual nature of support for those in greatest need, more recent formulations emphasise wider ideas of individual and collective wellbeing. Indeed, it was the very fact that this policy domain had been the subject of such change that new legislation to tidy, freshen, and overhaul the sector was thought to be required.
Within England, local authorities have responsibility for oversight of their local adult social care systems within a national framework of policy and legislation. Local care systems operate as mixed markets of provision.
The Care Act 2014 introduced the most significant and ambitious change in social care law in England for 60 years, fundamentally overhauling the entire care and support system for adults, older people, and their carers. It consisted of two phases, the first phase being introduced in April 2015 with phase 2 to be introduced in 2016. However, in July 2015 it was announced that Phase 2 (introducing a cap on care costs) was to be deferred until April 2020 and has since been abandoned. New proposals are expected in a Green Paper but this has been deferred in part due to the impact of the ongoing Brexit developments on government business and, more recently, of the COVID-19 pandemic. The delay is relevant insofar as it reflects a failure of policy implementation in an area of significant public concern.
The Care Act was seen as a significant part of a new approach to supporting adults with social care needs and the delivery of adult social care services. Its overarching objectives were to reduce reliance on formal care, promote people’s independence and wellbeing, and give people more control over their own care and support. The Act was generally welcomed and achieved a high degree of consensus aided by a collaborative approach among the key partners who were intent on ensuring that the policy was made to stick when it came to implementation. Central to this commitment to implementation was the development of the Implementation Support (IS) programme noted above. Whilst all changes in policy, especially those of a complex nature, might benefit from some form of implementation support, it should be easier to achieve where the key parties are in agreement over the direction and objectives of the policy. This ‘collaborative policy design’ (Ansell et al., 2017) is the central feature of the Care Act IS programme. Before turning to it, we briefly reflect more generally on the nature of implementation support within public policy.
Addressing policy failure and making policy stick by paying greater attention to implementation are becoming critical concerns, especially in the face of deep-seated and stubborn policy challenges that are increasingly acknowledged to be more complex and not subject to simple, linear solutions (Holmes et al., 2017). Four broad contributors to policy failure can be identified: overly optimistic expectations; implementation in dispersed governance; inadequate collaborative policy-making; and the vagaries of the political cycle (McConnell, 2015). We have elaborated on each of these contributors elsewhere (Hudson et al., 2019). They may be regarded as the implementation challenges that would have to be met by an implementation support programme of the kind with which we are concerned here. Certainly, the Care Act required expectations to be managed; governance to be in place at multiple levels – macro (national), meso (regional and strategic local authority), and local (service provider) levels; the active engagement of many different stakeholders; and a sustained commitment over time to ensure that the changes sought were sufficiently embedded. Building on McConnell’s work, we developed an overarching framework to structure our analysis (see Figure 1).
A Framework for the assessment of implementation support programmes.
Implementation support can take a variety of forms, with the mechanisms selected varying widely. They may, or may not, be measurable with some taking a visible, tangible form (e.g., regional meetings), some being of an experiential nature (e.g., inspiring leadership and/or management), and others being theorised (e.g., participation in regional meetings will facilitate ‘x’ and ‘y’). Mechanisms are not ‘things’ (or mediators) but part of an account of causality which only works when explaining the context within which they operate, and the outcome to which they contribute (Emmel, 2013). The identification of mechanisms shaping policy implementation support is important as it aids explanations in regard to why interconnections should occur. In addition, mechanisms can be used to describe the causal relations within a system which generate uniformity (Pawson, 2008). By linking these levels of explanation, there is the opportunity to transcend the divide between top-down and bottom-up approaches towards policy implementation (Sabatier, 1986, cited in McEvoy & Richards, 2003).
Implementation support mechanisms can be identified as having one of three main purposes: managing and regulating; problem-solving; and capacity building (Gold, 2014). All three were evident in varying degrees in the IS programme which accompanied the Care Act 2014 and to which we now turn.
Given the complexity of the changes introduced by the Care Act, the DHSC – known as the Department of Health at the time of the passage of the Act – and its key partners decided that a comprehensive programme of implementation support should be put in place both to ensure legislative readiness and increase the likelihood of smooth implementation (Hughes & Caunt, 2013). Three principles were established which underpinned the support programme:
The arrangements put in place to deliver on these principles involved the establishment of three key organisational innovations: a Programme Board; a Delivery Board and Programme Management Office; and a regional infrastructure. While some aspects of these features of support had been present in other policy programmes, the main innovation was that stakeholders were partners, taking on responsibility and not just giving advice. Figure 2 provides an illustrative overview of the programme structure established for implementing the Care Act (National Audit Office, 2015).
Programme structure for implementing the Care Act.
The Programme Board was upwardly accountable to the DHSC Major Programmes Board and had beneath it a Programme Management Office, a Support Delivery Board, and a raft of work streams. It had three key functions: support, assurance, and delivery. The Delivery Board had a much more hands-on role being tasked with driving timely and effective delivery; ensuring risks and other issues were identified and mitigated; and assessing data to monitor impact and drive the delivery of anticipated programme benefits. The Programme Management Office was established to support the work of the Board and was seen as central to the fulfilment of all three Board functions. In recognition of the potentially wide gap between central government and a multiplicity of local authorities, the decision was taken to develop a regional dimension to act as a conduit between regions and the Programme Management Office. The regional tier was not another organisational layer but amounted to some modest funding being found to support the regional leads who were left free to determine their own ways of working through networks of local authorities. Regional level support was anticipated to facilitate rapid dissemination of the latest tools and advice; increase the pace of local implementation; and link into assurance mechanisms where the local pace was thought to be falling behind. Organisationally this level of support was intended to build on arrangements for existing models connected with other programmes, such as Health and Wellbeing Boards (which bring together local health and care leaders to collaboratively improve outcomes for their populations), Transforming Excellence in Adult Social Care, and NHS Vanguards (sites leading the development of new models of care, some of which are overlaps of health and adult social care).
On the issue of the cost of the IS programme, the two stakeholders outside of the Department of Health (DH) – LGA and ADASS – funded their own input. In addition, DH made available modest funding to support the regional links. It may be worth noting, too, that the NAO (2015) raised no concerns about value for money in relation to the monitoring and support arrangements.
This brief descriptive account of the Care Act’s IS programme is testimony to the seriousness with which the mission was undertaken at central level. It suggests a keen awareness of the potential danger of policy failure and a determination to avoid it in ways that could mark it out as different and distinctive. The need for implementation to be in the hands of a multiplicity of local agencies – statutory, voluntary, and independent – is a key feature of the context surrounding the Care Act. Although highly detailed statutory guidance (the epitome of a top-down approach) was produced, there was also an appreciation of the influence of local contexts and dispersed power bases and the need to take these into account.
Our research, conducted between early 2017 and mid-2018 and funded by the DHSC Policy Research Programme, which comes under the auspices of the NIHR, focused on implementation support at three levels: macro (national), meso (regional and strategic local authority), and micro (local service delivery). Data were explored in three key areas: analysis of relevant theoretical and conceptual literature, a review of the support programmes (if any) for a number of previous and current national policy programmes, and an empirical study of the Care Act IS programme itself. In order to understand better the reasons for establishing the Care Act IS programme at national level, three data sources were utilised: documentary analysis of Care Act Programme Board minutes; information on Care Act Programme Board actions; and an analysis of 10 semi-structured interviews conducted with members of the Programme Board. A number of documents were produced for the Programme Board setting out ‘visions and priorities’ for implementation support, one of which provided a succinct explanation of the need for an IS programme:
A traditional approach to providing implementation support is unlikely to be able to meet the needs of all organisations given their breadth, role in providing social care and support and particular local circumstances. Similarly, those charged with implementation also have challenging financial constraints, other related policy issues such as Integration Transformation Fund, corporate requirements and/or partnership arrangements to address (Hughes, 2013: p.1).
The document cites a number of advantages in having a distinctive implementation support programme, including collaboration among stakeholders, clarity in dialogue, and flexibility in the programme management tools. Additionally, capacity – in terms of resources and finance – is advanced as an issue that several organisations in the public sector face. There is also a recognition that ‘…no one single approach will be universally applicable to all involved and that a heavily directed approach would neither be well received nor taken up’ (ibid:p.2).
To explore the issues arising at the meso (regional) level, interviews were conducted with five regional leads. At the micro (locality) level, six local authority case studies were undertaken which entailed interviews with senior managers, operational staff, and focus groups made up of service users and carers. The local authority areas were chosen to reflect the diversity present within English local government (see Table 1). Research approval was sought from the Association of Directors of Adult Social Services and granted on 21st July 2017, and HRA Social Services Research ethical approval for data collection with service users was obtained on 23rd February 2018 (Ref: 17/IEC08/0050). A final research report was published in late 2019 (Peckham et al., 2019).
Table 1
Case studies.
Local authority case study | Population size (2017) | Urban/rural |
---|---|---|
North East metropolitan district unitary authority | 200–300,000 percentage 65years+ – 17% |
Urban/rural split – 99%/1% |
Smaller northern unitary authority | 100–200,000 percentage 65years+ – 19.5% |
Urban/rural split – 68%/32% |
Northern metropolitan district council | 300–400,000 percentage 65years+ – 17% |
Urban/rural split – 82%/18% |
Large southern county council | over 1 million percentage 65years+ – 17.9% |
Urban/rural split – 72%/28% |
Rural eastern county council | 100–200,000 percentage 65years+ – 19.8 |
Urban/rural split – 61%/39% |
London Borough Council | 200–300,000 percentage 65years+ – 10.3% |
Urban/rural split – 100%/0% |
As part of the preparations for the field work undertaken to assess the value and impact of the Care Act IS programme we conducted a rapid mapping exercise of other English policies with not dissimilar aims to those of the Care Act in order to understand what, if any, policy support had been made available. The aim was to establish if there were general lessons regarding how support processes may best be developed to aid the local implementation of national policy. It is perhaps notable that no published review of the issue of implementation support exists. The search strategy was guided by the knowledge and experience of the research team, and by the external advisory group which guided and informed the research. Five criteria governed the selection of policies: scale (was it a national policy applicable to local areas across the country?); purpose (was the focus on implementation support?); reach (was support extended to every locality?); learning (was there an evaluation or other evidence base?); and significance (did the policy have a statutory underpinning and guidance?). The mapping exercise initially identified fifteen policies which, after reviewing them against our selection criteria, led to a focus on five of these: the Community Care Support Force, Sure Start, Health and Wellbeing Boards, the Troubled Families Programme, and the NHS Vanguards. Although many of the policies themselves had been subject to an independent evaluation, few of the studies detailed the approaches of implementation support (if any) that had been offered. Further information about the rapid review can be found in Chapter 3 of the final research report (Peckham et al., 2019).
In the light of our rapid review of other policies, the main conclusion to draw is that implementation support has tended to be regarded as somewhat marginal to successful policy implementation – at best a useful accessory, but not thought to be central to the success of a policy. Such a finding, although disappointing, serves to underscore the importance of our study of implementation support in regard to the Care Act.
Against this context of the availability of implementation support in other policy domains, the Care Act IS programme was both unusual and distinctive. Although the other policies we examined shared some of the same approaches to implementation support, the Care Act employed a more extensive range of support mechanisms in order to address every aspect of implementation (e.g., nationally produced guidance, regional level working groups, training, stocktake reporting mechanisms). It was also widely welcomed and not considered to be burdensome or an unhelpful imposition.
Finally, and perhaps not so surprising given the limited attention it has received, our mapping of other policies, and the implementation support provided, shows negligible learning from the type of IS programme we have described in respect of the Care Act. The implementation of each policy has proceeded in separate silos, seemingly oblivious to, and unaffected by, what has happened (or not happened) in other policy areas. However, there are some glimmers of hope which suggest that this compartmentalised approach to policy implementation may be shifting with evidence of a growing interest in learning from implementation support approaches. This was especially notable in the development of the respective support programmes for the Care Act 2014 and NHS Vanguards initiative (Billings et al., 2019; Checkland et al., 2019; Coleman et al., 2020). In the case of the latter, this research commenced upon completion of our study of the Care Act IS programme and was informed by our work. Findings from our study are echoed in those emerging from the Manchester-based Vanguards study.
The findings from our research are structured around the six criteria set out in the assessment framework based on McConnell’s work (see Figure 1):
From the outset it was acknowledged that the Care Act was a complex and ambitious piece of legislation and that implementation would not be easy, especially being dependent on diverse contexts and the involvement of multiple stakeholders each presenting specific challenges well identified in the policy literature (Davies et al., 2008, Russell et al., 2008). The approach adopted in the Care Act IS programme built on the collaborative nature of the development of the Act itself and involved key national stakeholders working in partnership to develop and support the implementation process. The range of support mechanisms employed within the implementation approach set the Care Act apart from previous policies reviewed in this study and briefly described earlier.
Four key concerns guided the research:
These concerns need to be viewed through a complex adaptive systems lens since the context, or system, within which a policy is implemented is never static. Systems are viewed as self-organising and emergent from within complex structures and there is therefore a need to comprehend and interpret the relationships between the elements which make up the ‘system’ in order to understand ‘what works’ (Westhorp, 2012). Within a complex system, interactions are generally non-linear, that is, an action does not always have the same outcome as the result is dependent on the context within which the interactions occur. In addition, emergent behaviours are often unpredictable, due mainly to the influence of people who will react differently to the same situation (Williams, 2015). This may be because they are subject to differing pressures and expectations reflecting the particular power plays operating in any given situation.
Policy sets the context within which those with a remit for its delivery must make crucial decisions on the shape of implementation. In everyday parlance it is often said that things should not be ‘taken out of context’. This similarly applies to policy implementation, since there is now a growing body of evidence that an intervention that is successful in one location does not deliver the same results elsewhere (Health Foundation, 2014; Horton et al., 2018). As Dixon-Woods (2014: p.89) points out: ‘History is littered with examples of showpiece programmes that do not consistently manage to export their success beyond the home soil of early iterations’.
All of this connects with the long-standing literature on ‘receptive’ and ‘non-receptive’ contexts for change pioneered by Pettigrew et al. (1992). The quintessential task of implementation support could therefore be said to be to assist the organisational shift towards a ‘receptive’ implementation context. Weiner (2009) describes this as ‘organisational readiness’ for change – a state of being both psychologically and behaviourally able and willing to take action in a desired direction. Of relevance here is the health system transformation initiative launched by WHO Europe, which includes a self-assessment checklist to enable policy-makers to reflect upon, and assess, their readiness for change and whether or not the requisite capacities and capabilities are in place for successful implementation to occur (WHO, 2018).
It is therefore likely that the implementation support process will more easily flourish in some contexts than others – indeed a recurring theme throughout this account has been the receptive political and professional context within which the Care Act IS programme functioned. Not all policies can be expected to be characterised by such a high degree of political and professional agreement and engagement; in fact, most will almost certainly be the outcome of divisive and contentious disagreements.
A useful framework for understanding the role of context is Matland’s (1995) classic work on the impact of conflict and ambiguity on implementation. The premise is that the different characteristics of policies have varying implications for the way they are implemented – and, by extension, for the ways in which implementation support programmes might best be constructed. Matland uses a distinction between issues about the extent of policy ambiguity on the one hand, and issues about policy conflict on the other, to develop the matrix below (see Table 2).
Table 2
Matland’s Model of Conflict, Ambiguity and Implementation.
LOW CONFLICT | HIGH CONFLICT | |
---|---|---|
LOW AMBIGUITY |
ADMINISTRATIVE IMPLEMENTATION
|
POLITICAL IMPLEMENTATION
|
HIGH AMBIGUITY |
EXPERIMENTAL IMPLEMENTATION
|
SYMBOLIC IMPLEMENTATION
|
There are important implications arising from this analysis for ensuring the right model of policy implementation support is associated with each domain of the matrix. Broadly we can hypothesise that:
These categories are not mutually exclusive – policies could contain several elements – but the task of policy-makers and practitioners is nevertheless to determine which policies require what mix of support to give them the best chance of effective implementation.
In the case of the Care Act, the policy is probably best understood as ‘experimental implementation’. Although the passage of the legislation was characterised by relatively low conflict, it incorporated some new and largely untested ideas that were always likely to be open to interpretation – high ambiguity. In these circumstances a bottom-up approach showing sensitivity to local context alongside support for problem-solving was (in line with Matland’s hypothesis) the correct approach.
A limitation of our research is that it was undertaken after the IS programme had closed and while this had the advantage of allowing time for reflection, it also meant we were, to some extent, dependent upon participants’ recall of past events. A second possible limitation of our research was its restricted coverage to six local authorities that may have not been typical of developments elsewhere. However, we do not consider this to be a major drawback since, given the complexity of local government, no two authorities are likely to be the same in every respect.
Overall, it can be concluded that the Care Act IS programme significantly helped ensure the implementation readiness of local agencies. Key successes identified for the IS programme relate to its securing policy legitimacy, the successful navigation of complex issues through stakeholder engagement, and ensuring the readiness of local implementation agencies.
Stakeholder engagement at a macro (national) level was regarded as the key feature of the IS programme. The close relationships secured between the key national stakeholders (DHSC, LGA, ADASS) were unique with no comparable example found in the other policy domains we reviewed. The approach demonstrated engagement, drawing on existing relationships, brought in external expertise as required, facilitated the sharing of ideas and avoided a traditional mechanistic and top-down way of working by seeking a more flexible and adaptive approach influenced by the needs of regional leads and local service providers.
Given the relative novelty of the Care Act IS programme with its particular features that are not replicated in other policy domains, there is correspondingly little empirical evidence to draw upon, making the evidence base for implementation support programmes thin. There are three key messages arising from the Care Act IS programme we studied:
Underpinning everything the research investigated and its findings is the irrefutable evidence that the Care Act was in large part a popular piece of legislation amongst sector organisations that generated a great deal of stakeholder consensus. This cannot be said of all policy and therefore inevitably limits what can be learned from this study alone about the wider potential of policy support programmes. But if nothing else, the findings reported here demonstrate the potential value of implementation support if making policy stick is a desired outcome. It is clearly a topic that merits further investigation, especially at a time when government’s ability to find lasting solutions to complex ‘wicked’ problems has perhaps never been so tested and under such scrutiny.
Research funded by National Institute for Health Research Policy Research Programme Project PR-R14-1215-21006.
The views expressed in this paper are those of the authors and not necessarily those of the National Institute for Health Research (NIHR), the Department of Health and Social Care (DHSC), or its arm’s length bodies or other government departments.
The authors have no competing interests to declare.
Ansell, C, Sorenson, E and Torfing, J. 2017. Improving policy implementation through collaborative policy-making. Policy and Politics, 45(3): 467–86. DOI: https://doi.org/10.1332/030557317X14972799760260
Billings, J, Mikelyte, R, Coleman, A, Macinnes, J, Allen, P, Croke, S and Checkland, K. 2019. Supporting integrated care in practice: Perceptions of a national support programme for the development of new care models in England. Journal of Integrated Care, DOI: https://doi.org/10.1108/JICA-05-2019-0018
Checkland, K, Coleman, A, Billings, J, MacInnes, J, Mikelyte, R, Laverty, L and Allen, P. 2019. National Evaluation of the Vanguards New Care Models Programme Interim Report: Understanding the National Support Programme. Manchester, University of Manchester.
Coleman, A, Billings, J, Allen, P, Mikelyte, R, Croke, S, Macinnes, J and Checkland, K. 2020. Ambiguity and conflict in policy implementation: The case of the New Care Models (Vanguard) programme in England. Journal of Social Policy, 1–20. Available from: DOI: https://doi.org/10.1017/S0047279420000082
Davies, H, Nutley, S and Walter, I. 2008. Why ‘knowledge transfer’ is misconceived for applied social research. Journal of Health Services Research & Policy, 18(3): 188–190. DOI: https://doi.org/10.1258/jhsrp.2008.008055
Dixon-Woods, M. 2014. The Problem of Context in Quality Improvement. London, The Health Foundation.
Emmel, N. 2013. Sampling and Choosing Cases in Qualitative Research: A Realist Approach. London: Sage Publications. DOI: https://doi.org/10.4135/9781473913882
Gold, J. 2014. International Delivery: Centres of Government and the Drive for Better Policy Implementation. London: Mowat Centre/Institute for Government.
Gray, AM and Birrell, D. 2013. Transforming Adult Social Care: Contemporary Policy and Practice. Bristol, Policy Press. DOI: https://doi.org/10.2307/j.ctt1t88zr5
Harris, J and Rutter, J. 2014. Centre Forward: Effective Support for the Prime Minister at the Centre of Government. London: Institute for Government.
Health Foundation. 2014. Perspectives on Context. London: The Health Foundation.
Holmes, B, Best, A, Davies, H, Hunter, D, Kelly, MO, Marshall, M and Rycroft-Malone, J. 2017. Mobilising knowledge in complex health systems: A call to action. Evidence & Policy, 13(3): 539–560. DOI: https://doi.org/10.1332/174426416X14712553750311
Horton, TJ, Illingworth, JH and Warburton, WHP. 2018. Overcoming challenges in codifying and replicating complex health care interventions. Health Affairs. 37(2): 191–197. DOI: https://doi.org/10.1377/hlthaff.2017.1161
Hudson, B, Hunter, DJ and Peckham, S. 2019. Policy failure and the policy-implementation gap: Can policy support programmes help? Policy Design and Practice, 2(1): 1–14. DOI: https://doi.org/10.1080/25741292.2018.1540378
Hughes, A. 2013. Implementation Support: Vision and Priorities. London: Department of Health.
Hughes, A and Caunt, M. 2013. Care and Support Reform Programme Board: Programme Scope: Work-streams, Projects and Enablers. London: Department of Health.
Matland, R. 1995. Synthesising the implementation literature: The ambiguity-conflict model of policy implementation. Journal of Public Administration Research and Theory, 5(2): 145–74.
McConnell, A. 2015. What is policy failure? A primer to help navigate the maze. Public Policy and Administration, 30(3–4): 221–242. DOI: https://doi.org/10.1177/0952076714565416
McEvoy, P and Richards, D. 2003. Critical realism: A way forward for evaluation research in nursing? Journal of Advanced Nursing, 43(4). DOI: https://doi.org/10.1046/j.1365-2648.2003.02730.x
National Audit Office. 2015. Care Act First Phase Reforms. HC 82, Session 2015–16. London: The Stationery Office.
Pawson, R. 2008. Invisible mechanisms. Evaluation Journal of Australasia, 8: 3–13. DOI: https://doi.org/10.1177/1035719X0800800202
Peckham, S, Hudson, B, Hunter, DJ, Redgate, S and White, G. 2019. Improving Choices for Care: A Strategic Research Initiative on the Implementation of the Care Act 2014. Southampton, National Institute for Health Research. Available from: https://www.researchgate.net/publication/339289592_IMPROVING_CHOICES_FOR_CARE_A_STRATEGIC_RESEARCH_INITIATIVE_ON_THE_IMPLEMENTATION_OF_THE_CARE_ACT_2014?channel=doi&linkld=5e47f090458515072d9f38ef&showFulltext=true.
Pettigrew, A, Ferlie, E and McKee, L. 1992. Shaping Strategic Change. London, Sage.
Russell, J, Greenhalgh, T, Byrne, E and McDonnell, I. 2008. Recognising rhetoric in health care policy analysis. Journal of Health Services Research & Policy, 13(1): 40–46. DOI: https://doi.org/10.1258/jhsrp.2007.006029
Spillane, J, Reiser, B and Gomez, L. 2006. Policy implementation and cognition: The role of human, social, and distributed cognition in framing policy implementation. In: Honig, M (ed.), New directions in education policy implementation: Confronting complexity. Albany: State University of New York Press.
Weiner, B. 2009. But will it work? Issues in Government Studies, 32: 1–17.
Westhorp, G. 2012. Using complexity-consistent theory for evaluating complex systems. Evaluation, 18: 405–420. DOI: https://doi.org/10.1177/1356389012460963
Williams, B. 2015. Prosaic or profound? The adoption of systems ideas by impact evaluation. IDS Bulletin, 46(1). DOI: https://doi.org/10.1111/1759-5436.12117
World Health Organisation. 2018. Leading Health System Transformation to the Next Level. In: WHO Expert Meeting, 12–13 July 2017, Durham, UK. Copenhagen: WHO.