Networked innovation in the health sector: Comparative evaluation of the role of Clahrcs in translating research into practice

Overcoming boundaries to access distributed knowledge

As work in the innovation field has highlighted, the knowledge required to develop and implement innovation is increasingly distributed across different groups and organisations. 35 This finding is also echoed in studies within health care. McAneney et al. ,36 for example, note that ‘the knowledge which is needed to solve problems and bring about changes is likely to be distributed throughout organisations and to come from many different sources’ (p. 1498). Recognition of the need to span boundaries is a long-established theme in innovation studies, and research has focused to a great extent on organisational boundaries. 37 However, the importance of accessing distributed forms of knowledge underlines not only the need for translation between different organisations, but also the need to circumvent what recent work has termed the ‘knowledge boundaries’ that constrain the flow of knowledge between different epistemic groups and communities.

Knowledge boundaries naturally arise, according to Carlile, because of the embeddedness of knowledge in practice – knowledge sticks at the boundaries of practice and is shared where practices are shared. 20 Such boundaries can be analysed in terms of syntactic (shared or different language), semantic (shared or different meanings) and pragmatic (shared or different practices) dimensions. 20 The value of this analysis is that it highlights the multifaceted and relational character of KT. 38 It thus provides a broader view of the challenges of translation that highlights the extent to which particular groups, such as researchers, policy-makers and practitioners, are connected or divided by their respective contexts, language and (politically invested) practice. These different groups can be viewed as ‘epistemic communities’, which are characterised by shared language, values and world views. 39

A number of studies have sought to establish the types of mechanisms and processes that can be used to support KT across knowledge boundaries in health care. 27,40 These studies build on an understanding that knowledge cannot easily be translated into a comprehensible form from one community for utilisation by another dissimilar community:41,42 knowledge, in other words, is embedded or ‘sticky’,43 as the producers and users of knowledge inhabit different worlds. 44 Knowledge is also politically invested,20,45 so there is a pressure within networked innovation initiatives for more collaborative forms of knowledge production to revert to (or even be undermined by) long-established modes of producing knowledge based on professional demarcations. 46

Given that the challenges of translating knowledge across the boundaries of diverse groups and communities are especially stark within the health-care field, attention has been placed on developing, managing and evaluating interventions that can facilitate this process. Among the mechanisms highlighted in the existing literature is the creation of boundary-spanning or ‘knowledge broker’ roles for individuals to link discrete communities;47–49 organisational-level activities (such as using forums and meetings) as places for the exchange of ideas between groups;50,51 and integrated KT processes and end-of-grant activities in which findings are translated for other audiences. 52

Considerable attention has been placed on understanding the process of ‘boundary spanning’ across organisational and epistemic groups within health care. While some research has highlighted the importance of material objects in bridging diverse groups,39 boundary spanning can also be achieved by individuals who, either formally or informally, enact a ‘knowledge broker’ role. Such individual boundary spanners have been described as performing functions such as acting in leadership and network-building roles, fostering relationships, and contributing an innovative perspective,53,54 suggesting that boundary-spanning individuals are those with experience of and credibility in both ‘camps’, such that they can move back and forth and broker understandings between different thought worlds. Such knowledge-brokering strategies can be used to address the language and cultural barriers between the worlds of research and decision-making by translating research and other evidence into different vocabularies. 55 The knowledge-broker role is designed so that individuals can act as facilitators of collaboration and ‘translators’ of knowledge from one community to another. Indeed, as the use of interpersonal contacts and good communication skills in the context of partnerships and research collaborations is emphasised in knowledge brokering, it has been described as particularly suitable for linking upstream research with downstream practice. 48 Therefore, one important strand of our work is on roles enacted by individuals to support the translation of knowledge across groups.

While much attention has been given to individuals performing boundary-spanning activities as brokers between diverse groups, previous studies also indicate that knowledge brokering can be enabled by organisations and structures, such as a whole network of ties or collaborative strategies. 56 These types of boundary-spanning mechanisms are about structures, activities or processes that an initiative may develop in order to facilitate KT within a collaborative community. For example, explicit strategies that have been described include initiative-wide activities developed to provide a ‘space’ for face-to-face interaction and discussion, such as consultation sessions, interactive multidisciplinary workshops, steering committees and the formal creation of networks and communities of practice. 57 The CIHR’s influential global integrated KT approach represents an organisational-wide strategy to support activities that are conducive for KT across the boundaries of the producers and users of research. 58 This approach is about building a structure and interlinking activities at the level of a whole initiative that is able to facilitate sharing knowledge across the boundaries of a wide range of actors, including health professionals, researchers, the public, policy-makers and research funders. 59 It may also include the formal allocation of resource to individual knowledge brokers.

Developing social networks to enable the translation of knowledge

In the recent period, recognition of the importance of boundaries between groups has been complemented by greater awareness of the central importance of social networks in enabling innovation that spans communities or practice groups. 31 Studies focusing on the analysis of network ties have highlighted the importance of network structures in shaping the flow of knowledge and information within and between organisations. 34,60 Here, the focus is less on individuals and more on the nature of the relationships and information flows across individuals and groups: their interconnectedness, in other words. Network analysts use a range of well-developed concepts (e.g. ‘nodes’ to denote position or location, and ‘ties’ to denote relationships or links among these positions) and methods [in particular, social network analysis (SNA)] to link the patterns of relationships in the network to the behaviours that arise from it. 61 Such tools allow analysts to reveal and represent the informal social relationships within and across organisations that shape work-related outcomes. This now-extensive body of work links helps to underline the impact, for example, on innovation of brokering across ‘structural holes’62 between different groups, as well as the pivotal roles of boundary-spanning groups and gatekeepers. It also highlights the different roles played by different kinds of social ties in knowledge sharing, with weak ties being linked to the sourcing of information and strong ties being linked to the establishment of trust and the sharing of tacit knowledge. 63 Importantly, social networks provide a relational form of governance (the ‘informal structure’) whereby an individual or group’s position (centrality) in a network relative to others plays a key role in their ability to exercise power. This informal structure has been found to be especially important in networked innovation settings (such as the CLAHRCs) where the exercise of power requires extensive boundary spanning and where sources of power influence innovation outcomes. 64

The attention paid to social networks in studies of innovation more generally has not so far been matched by work in the health-care management field. Here, recent studies, with relatively few exceptions, have focused on networks as formal, quasi-organisational entities, frequently associated with policy interventions. 65 As McAneney et al. 36 note, relatively few studies have addressed the implications of network structure for KT. Only in more recent work do we find social network analytical techniques being applied to uncovering the latent structure of ties between groups and individuals. Thus, a study by Currie and White66 uses SNA to identify the interplay between knowledge brokering and professional hierarchy. The authors argue that such brokering both is influenced by, and helps to mitigate, the impact of such hierarchy on knowledge mobilisation. They also highlight the role of groups in such brokering as against the actions of specific individuals.

Bringing together different interpretations: the role of cognitive maps

The importance of cognition and the underlying belief structures that groups and individuals bring to their interpretation of new situations has been identified as central to networked innovation and collaborative work. 67,68 Furthermore, work has shown how underlying cognitive structures or belief systems – sometimes termed ‘cognitive maps’ – are implicated in sense-making and decision-making processes. 69 However, while a number of studies have addressed the role of such cognitive maps in the actions of managerial groups,70 less attention has been paid to their importance in health-care settings. One exception here is the work of Sutherland and Dawson,71 who adopt a ‘sense-making’ approach to behavioural change among clinicians. They use the notion of cognitive maps to highlight the way in which actors ‘apportion meaning to situations, relationships, roles, and objects, and store such interpretations and understanding in cognitive structures, often in the form of “taken-for-granted” knowledge’ (p. 53). 71 Such structures or maps in turn influence how novel events or ‘equivocal’ situations are interpreted. 72

The arguments coming from this strand of work effectively underline the relevance of cognitive structures to the outcomes of networked innovation initiatives such as the CLAHRCs. As we highlighted at the beginning of this report, the notion of KT, which informed the original funding initiative for the CLAHRCs, was broadly defined and ‘equivocal’ in the sense that it was open to multiple interpretations. Organisational models and mechanisms of KT and its effectiveness within different contexts remain highly debated. It follows from this that the cognitive maps which different CLAHRC groups bring to this KT endeavour – what it means, and how it can or should be operationalised – may vary significantly across CLAHRCs, and may have important consequences for the way in which policy-inspired models are interpreted and enacted. A crucial question here has to do with participants’ ‘causal maps’ – that is, the associations they make between particular phenomena and particular outcomes. This is because such causal maps and their characteristics (e.g. cognitive complexity) have been closely linked to strategic visions and intentions and decision-making. 70 We recognise, of course, that intentions and cause–effect beliefs do not necessarily translate directly into behaviours and actions. 73 However, drawing the role of cognition and cognitive maps into our conceptual framework provides a useful complement to the emphasis on social interaction seen in previous research on networked innovation and collaborative work.

Management, governance and organisation

Much of the literature on KT tends to focus on the process by which this occurs, thus emphasising the micro-dynamics of interactions between individuals and groups. Viewing this process in the abstract, however, tends to neglect the important, if often antecedent, effect of the way in which that process is designed, directed and resourced.

Previous work has shown that efforts to engineer a process of KT may fail because management and governance are inappropriate to the task at hand. Relevant here is a study of the management and organisation of the Genetics Knowledge Parks initiative – a policy intervention designed to speed the translation of genetics and genomics science into clinical practice through the establishment of regionally based ‘knowledge parks’, where academic researchers, health-care practitioners and industry would work collaboratively. 74,75 This study found that the ambiguity that was inherent in these novel forms of collaboration generated uncertainties about its evaluation and the extent to which it was delivering against broadly defined, but extremely ambitious, objectives. This uncertainty prompted governing bodies (funders and policy bodies) to try to more tightly monitor and evaluate the work, thus creating burdensome reporting requirements and constraints on innovation.

This example raises an important issue of governance – that is, management practices often assume a linear model of innovation (i.e. planning and monitoring against predefined objectives and targets) whereas networked innovation is inherently an emergent and iterative process. 76,77 Equally, this example may reflect inappropriate forms of accountability and organisation being applied to a policy intervention. 74 The importance of management and organisation was fully recognised in the CLAHRC call for proposals, however, which highlighted the need to specify a director, management arrangements and explicit strategies for managing the collaborative partnerships involved in the CLAHRC.

The impact of the institutional and policy context

The example of the Genetics Knowledge Parks, outlined above, shows how the translation work occurring at the micro level within networked innovation initiatives is shaped by the wider institutional and policy context. The impact of the institutional and policy context on innovations in health care is widely reflected in the existing literature. 40,65,78,79 One conclusion here, then, is that, while much can be learned from models of KT in other nations, we must also be sensitive to the need to adapt and appropriate such models for the UK health-care setting.

Complementing this recognition of the importance of context at a macro level, there is a growing body of work on the ways in which context may shape more localised efforts to translate knowledge into health-care practice. Studies have shown, for example, that the same evidence may lead to different outcomes in different decision-making contexts,80 or that some translational activities, such as knowledge-broker roles, may be adaptable to different contexts. 49 At the same time, there remains a gap within the existing literature on how contextual features influence the KT process. This has driven calls for further research to address the influence of context as a prerequisite for developing more effective approaches in this area. 81,82

Qualitative data collection

Semistructured interviews were conducted face to face with individual members from three CLAHRC organisations within the UK as well as from the three North American initiatives. This was guided by an interview schedule, which was framed in terms of discussion of the three broad topics captured by the theoretical framework depicted in Figure 1 – that is, management and governance, KT processes/tools, and the development of social networks and relationships. Questioning encouraged interviewees to discuss specific examples of collaborative working (e.g. prompts such as ‘tell us about a situation where you have worked together well/where you have found it challenging’). The length of interviews was approximately 45–60 minutes, and the audio recordings were subsequently transcribed, providing texts of data for analysis. The interviews were supplemented by observational data and the collation of key documents, for the purpose of providing greater insight into the work taking place within the initiatives. This included attending meetings, including core management meetings, knowledge exchange, dissemination and engagement events. In addition, with respect to the UK CLAHRCs, original bid documents and NIHR interim reports were also collected. At the project level, in addition to attending project meetings, where opportunity arose, we also observed other activities, such as advisory board meetings and engagement events. We also collected written information about the projects, such as project outlines, information for participants, publications and other outputs.

In terms of the UK CLAHRCS, the first interviews conducted were with core members/founders of each CLAHRC initiative, such as the director, programme managers and other core members, and these members were reinterviewed in phase 2. Interviews were also conducted with members of the sample ‘micro-case’ project teams, which focused on three or four of the projects from each CLAHRC. The interviewees selected for phase 1 purposively targeted data collection from those in a variety of different types of roles (team leaders, managers, research and clinical professionals). Additional interviews were conducted with members of ‘specialist support services’, which were areas where each CLAHRC demonstrated particular types of expertise, such as academic advice in the form of economics and statistics support, KT and implementation expertise, and practitioner insight into areas of clinical services, commissioning, policy-making and capacity building. Depending on the different model of each CLAHRC, some of these members belonged to the clinical project teams, with others positioned in other parts of the initiative, such as central support and ‘cross-cutting themes’. Because of this variation, interviews with individuals in these types of roles were particularly important for understanding the distinctive ‘flavour’ of the vision of each CLAHRC model. In phase 2, follow-up interviews were targeted to be conducted with, primarily, the same representatives as from phase 1. However, owing to changes in personnel, particularly at the researcher level, these interviews typically focused on those in leadership and managerial positions from both central CLAHRC and project level, and also clinical professionals involved in project work.

In the first phase, interviews focused on the start-up of each CLAHRC, such as its set-up, aspirations, plans and initial formation of contacts. Documentary evidence, such as the NIHR funding application forms, website information and project-specific documents (such as study protocols), was used to provide additional information on these aspects and to reduce potential problems of retrospective accounting. In the second phase, interviews focused on what types of outcomes had been achieved, and the discussion focused on reflection about the experience of conducting a work programme within the CLAHRC context. In addition, textual information, such as interim reports and examples of outputs and dissemination activity, were used to supplement examination of the types of outcomes being demonstrated by each CLAHRC. In total, 67 interviews were conducted in phase 1 (Bluetown, 24; Greentown, 21; Browntown, 22) and 42 in phase 2 (Bluetown, 16; Greentown, 12; Browntown, 14).

In terms of the North American initiatives, we conducted a few preliminary interviews with the leaders of each initiative. Following a snowball approach, we were able to gain access to additional people involved in the various projects and to collect documents including presentations from steering and working committees and minutes of meetings. We were also able to attend (and record and transcribe) several meetings. The North American fieldwork was conducted in a single phase spanning September 2010 to March 2012. In total, we conducted 49 interviews and observations (27 with Canada-Coordination, 11 with Canada-Translation and 11 with US-Health). The criteria to select the sample followed the UK CLAHRCs (as one aim of the North American study is to compare the results with the UK CLAHRCs). Therefore, we adopted similar criteria and focused on people who were involved in the initiatives from the beginning, having a broad view of the overall initiative and, in most cases, decision-making responsibility.

Collecting network data

The first task was therefore to identify individual members of each KT initiative through discussion with the health-care initiatives. Administrators of each KT initiative provided a membership roster. These lists were checked to remove individuals for whom participation in the study was not relevant. This included individuals who were no longer involved or employed by the organisation or who were on leave (i.e. sick/maternity/sabbatical). Delineation of each network sample was therefore defined by the criteria of CLAHRC membership. We then surveyed all individuals who we believed to be members of each CLAHRC. The same method was applied to our Canadian case site, Canada–Coordination.

With knowledge of who was in the network as well as demographic information on these individuals, it was then possible to assess the structure of KT between individuals in each health-care setting. An online survey was developed and individuals were sent an e-mail containing a web-link to access. The roster of CLAHRC member names was built into the electronic survey so that respondents were able to see the names of other CLAHRC members via a drop-down menu. Respondents were also able to name external contacts so that we were able to study KT relations beyond the CLAHRC. To show the development and changes of CLAHRC knowledge networks over time, the SNA survey ran for two data collection waves, providing data snapshots, which we refer to as time 1 and time 2.

Piloting

Prior to the official online release date, the survey was piloted, including testing and retesting, by a subsample of CLAHRC members and also members of the research team not involved with the SNA component of the study.

Response rates

The social network survey was sent to a total of 325 individuals across three CLAHRC health-care initiatives in the UK and 39 members of the Canada-Coordination initiative. The average response rate for the CLAHRCs was 71% at time 1 and 63% at time 2. The response rate from Canada-Coordination was lower than for the CLAHRCs. A breakdown of the response rates is provided in Table 1 .

Non-response and partial completions

Reminder prompts were sent to individuals who had not responded or who had provided a partial completion of the survey. The use of personalised e-mail prompts, telephone calls and reminders by CLAHRC leadership helped to achieve an increased response rate from members.

Validity

Our interim feedback reports and presentations to the KT initiatives allowed us to cross-check that the network ‘story’ that was being built made sense and resonated with CLAHRC members. In all cases, we received positive responses and helpful critical questions from the CLAHRCs. We also triangulated SNA and qualitative data sources.

Visualisation and analysis

Social network analysis was conducted using UCINET (Analytic Technologies, Lexington, KY, USA), with descriptive statistics and graphs in SPSS (SPSS Inc., Chicago, IL, USA) and Microsoft Excel (Microsoft Corporation, Redmond, WA, USA). In network visualisations, each ‘node’ represents a survey respondent/KT initiative member.

Generalisability

We attempt to provide comparative analysis of the KT networks but, as per our research design, we do acknowledge that KT activities may be influenced by specific contexts.

Missing data and network size

Social network analysis is sensitive to missing data,88 and this inevitably impacts on our study (more so for time 2 data). Social network data are notoriously difficult to collect. The study was ambitious and we relied on building and maintaining strong, positive relationships with the CLAHRCs and frequent communication and interim feedback of findings. Time 2 data collection occurred during the period prior to CLAHRC refunding, which may help to explain the lower response rate. To help overcome the issue of missing data we assumed symmetry for relations where it made sense to do so in our analysis. 89 We acknowledge the effects of network size on other structural metrics such as density,90 particularly for the smallest Canada-Coordination network.

Other issues

Beyond general issues related to inaccurate survey response (i.e. respondent misunderstanding questions), we highlight that network studies can be affected by the accuracy of respondent recall. 91 However, omission of contacts seems to be an issue for studies on weak or distant ties92 and this study focuses on ‘most important’ ties for KT activities, relations which are likely to be more obvious.

The boundary specification problem is a common issue when studying social networks with unclear boundaries, that is, where it is not entirely clear who is in the network. For this study, the network boundary was defined by CLAHRC membership. Our survey also permitted the inclusion of external knowledge contacts if these people were deemed important to respondents in the context of their CLAHRC KT work. A fixed vertex degree research design limited the number of contacts respondents were able to name to up to five external contacts. This restriction was chosen to yield data on the most important ties in KT contexts both internal and external to the CLAHRC, and to prevent people from simply recalling contacts from memory, which would likely result in bias towards most frequent and recent ties.

We considered the trade-off between respondent fatigue and a survey of the whole organisation and concluded that asking for relations between all CLAHRC members would have been too burdensome on respondents. Our fixed vertex research design urged respondents to think about whom to include and exclude rather than freely name everyone, or name too few.

The vision of the Collaborations for Leadership in Applied Health Research and Care

As the three case sites of translational initiatives within the UK were all created through the same UK NIHR funding programme, they were all designed to meet the same aim and generic mission. However, there was significant flexibility in the way in which this mission was interpreted by the leaders of different CLAHRCs. We term these interpretive acts of leadership as different ‘visions’ of collaborative translational research. Within our study we have explored how the vision of each CLAHRC has emerged from and interacted with the structuring of the initiative, particularly in terms of management and governance. By studying the CLAHRCs’ development over time, then, the qualitative fieldwork has been able to explore how these distinctive features of each CLAHRC influence their approach to KT.

Management, governance and organisation

The CLAHRC was originally formed around a small central leadership team, with the vision of the director strongly influencing the focus and direction of the CLAHRC model. As the director had a historically strong reputation in the local area, this helped to legitimise the CLAHRC as something that was perceived as of value by those in senior positions within the partner organisations.

The vision of the core management team has been strongly influential on the approach that each programme of work uses. Each programme of work is expected to use a rigorous scientific design and methodology in order to produce robust evidence that is suitable for publication in high-quality academic journals. Therefore, the model builds on the approach to scientific work that was historically conducted by the lead organisations, with the CLAHRC emphasising that through these work programmes the teams should foster collaborative relationships with relevant service areas. This vision is emphasised through the role of the leaders within each of the project teams, who provide scientific and methodological direction to the programme of work. However, although there is recognition of the overarching objectives expected from each team by the central management, there is no CLAHRC-wide strategy for how each team should be structured or how collaborative relationships should be formed and developed. As a result, the lead of each project team has been provided with extensive flexibility on how their individual programme is organised. As a consequence, each team tended to foster relationships with particular groups and communities as relevant to the local services on which they were gathering evaluative evidence. The influence of the CLAHRC was important here in formalising and legitimising this collaboration between clinical academics and targeted groups in the NHS.

Without CLAHRC, we would have some of those connections but I think the momentum, thrust and energy that’s going into current programme really wouldn’t be there . . . associating with individuals from other fields, groups that we wouldn’t normally be part of. This has really allowed us to reflect more objectively on work, and the direction we’re going.

BLUETOWN022

Structural features of the CLAHRC were used to communicate the overarching objectives of central management to the clinical project teams. This involved regular interactions between the centre and projects, management representation at project team meetings, and programmed meetings for project leaders and project managers. The positions which project members held in other environments (i.e. outside of their own team environment) were typically construed as ‘honorary’ – that is, not part of the main role which project members perform within their clinical team or central management group ( Box 1 ).

Although the majority of clinical project team members share similar types of disciplinary expertise, with most having clinical–academic experience, the structural organisation of the initiative provides access to other types of expertise. The extent of the CLAHRC-wide resources means that individuals with expertise such as health economics, statistics, systematic reviewing, sociology and communication are easily available for project teams to access. With the sociology theme, for example, each project team allocated a small proportion of their own resources to support the employment of a select number of people with this type of expertise. Although these team members come from different working cultures from the majority of the CLAHRC members, it is clear to the clinical project teams that the director values and respects the expertise that these individuals can provide. This helps to legitimise their contribution within the teams, even in sociological territory, which such teams would not normally view as part of their remit. At the start of CLAHRC, the cross-cutting activity for sociology was an undefined programme of work, but this provided an opportunity for these members to liaise with the clinical project teams to identify how they could support their programme of work. As relationships were built up, they quickly identified certain project teams where they could add value to the other work that was already planned by the team lead.

We are officially termed as a cross-cutting theme but we’re also embedded in the individual research, as in our jobs are paid out of individual projects.

BLUETOWN003

This cross-cutting work has become an embedded part of a proportion of the clinical project teams. Although they provide a different type of expertise to the clinical projects, the members of this cross-cutting theme enact their role in such a way as to fit in with the overall work programme. While, overall, the cross-cutting theme constitutes only a small part of the CLAHRC, resources were deliberately allocated so that the members of this group would be highly skilled and experienced, and therefore able to achieve this. They have also been able to contribute guidance to more junior members of the project teams who are involved with areas that overlap with their area of expertise, such as qualitative components. Observation from in-depth studies of the four clinical themes indicates that this approach has facilitated the ‘embedding’ of the cross-cutting theme members within the project teams.

Networks and collaboration

The qualitative interviews demonstrated that from the early stages of the CLAHRC’s development, a clear objective was understood to focus on working with stakeholder groups, such as collaborating with NHS practitioners and managers within the clinical project work. There was also acknowledgement that this required some compromise with established academic work practices, with some effort being required to produce work that is suitable for practitioners.

Getting researchers to understand practitioners is a covert aim of CLAHRC. So that you don’t go away for five years and then tell them what they should have done in the first place because practitioners don’t want to hear that.

CMBLUETOWN007

The activity of each of the clinical project teams means that they create links to defined health-care organisations involved in the CLAHRC partnership. The interaction between members of the project team (e.g. the project lead) is integral to fostering the relationships among the official partner organisations of the CLAHRC. As many senior CLAHRC members had pre-existing collaborative relationships with NHS trusts, they were able to enact ‘senior’ boundary-spanning roles. Many of the theme leads were in clinical–academic dual activity roles and held honorary contract positions with NHS organisations. However, their leadership typically reflected the wider ‘epistemic’ community of a university setting, emphasising academic values rather than the practical concerns seen in the health service environment. For university-based teams, the collaborative interaction was framed by the values of the academic community, with high-level academic publication considered as important for demonstrating value to these groups. This was seen as consistent with the vision of the CLAHRC, in that the collaboration is seen as creating a culture within health-care settings which is more receptive to high-quality scientific evidence.

It’s nonsense to say . . . the PCT health, local authority or the voluntary sector don’t consider evidence. They do. They just consider evidence perhaps in a different way than you or I perhaps might consider evidence . . . The CLAHRC process is about the better, the optimal decision making that we can bring, the greater rigour, to set different parameters for making the decision. That is the value.

BLUETOWN014

Thus, the CLAHRC emphasises clinical–academic evaluation of service delivery in terms of producing high-quality evidence. As a result, the CLAHRC’s applied health research is conducted on the terms set by the clinical–academic community, to which other groups must be able to fit in. At the same time, the CLAHRC’s strong scientific reputation within certain fields helps it to build links with communities that value this type of evidence, such as national policy groups and certain local clinical groups. Certain groups of health-care managers and policy makers were perceived as valuing exactly this type of evidence.

Publishing all this information in top quality journals as absolutely, absolutely key. Not just to the academic credibility of CLAHRC but to the managerial credibility of CLAHRC . . . dealing with managers, when they want to know how good somebody is the first thing they do is see whether other academics respect that person. And one method of that is you publish in the top journals. They don’t want to deal with somebody who’s not up there at the cutting edge. They want to be with the front people.

BLUETOWN001

The work taking place within these themes tends to be dominated by the traditions of the theme lead and the culture where the theme is physically located. Non-health care-based teams do use ‘in-built boundary spanners’, pre-existing contacts and mechanisms such as advisory boards to draw on insights about the local health-care context ( Box 2 ). This approach enables the teams to focus on their own academic areas of expertise but perhaps supports depth rather than breadth of approach.

Bluetown CLAHRC also includes a number of themes based directly in health-care organisations. As the teams working on these themes are closer to the issues of practice, their work has the potential to more easily impact on health-care practice. In ‘speaking the same language’ as the practice and community groups they intend to impact, they are in a better position to integrate the perspectives of these groups. This should facilitate the implementation of findings and local impact that these groups will have. However, as these teams are further away from the core CLAHRC, greater attention has been required to ensure that these themes feel part of the Bluetown CLAHRC community.

Knowledge translation

Bluetown CLAHRC was not built on an explicit CLAHRC-wide strategy for KT. However, as each project team was provided with extensive flexibility in how they organised their clinical study, each individual theme has developed their own strategy for dissemination of outputs, with a flexibility to interpret the aims of the CLAHRC to fit their own clinical context. As a result, the themes have developed their own tailored approaches to relating their findings to local health-care practice and policy. These are not, however, related to an overall CLAHRC plan or strategy for this approach to the work. Additionally, these approaches do not appear to be shared between the themes, and their context-specific design may potentially limit wider application.

Overall, the Bluetown CLAHRC is formed from members who represent both academic disciplines and health-care professional groups. This means that the cumulative work of Bluetown CLAHRC has the potential to innovatively incorporate many different perspectives and draw from different types of expertise. However, currently, these tend to focus on project–team connections, which limit the potential of a co-ordinated Bluetown CLAHRC effect across the region. The lack of a co-ordinated strategy to link with the different types and levels of the Bluetown partners may have limited the capacity-building of the CLAHRC as a whole, notwithstanding the benefit of the strong relationships fostered by each individual team.

As the CLAHRC developed, a new theme of knowledge management was developed by the central CLAHRC management team. This was designed as a high-level cross-cutting theme. The activity was separate from the clinical project work.

But having said all that I am very keen to in addition add a specific knowledge management function to the CLAHRC but I am very clear this is not based on discoveries made in the CLAHRC, although I would not exclude those, but it’s not based on that, it’s based on the knowledge out there in the world, what can we make of that. So knowledge exchange if you like.

BLUETOWN001

However, it drew on the infrastructure that the teams had used and fostered during the early stages of the CLAHRC. In particular, it was felt that, as the clinical project teams had demonstrated activity within the NHS organisations at an early stage, NHS managers and executives valued the work of the CLAHRC, and that this facilitated buy-in for this new venture.

Due to the financial model of the CLAHRC, no specific funding was allocated to this area. However, NIHR Flexibility and Sustainability Funding money was allocated to this activity. As part of this, the core management team was expanded to include support for the clinical teams and CLAHRC as a whole with external engagement and communication and dissemination activities. In addition, a knowledge-exchange forum was developed. This was targeted at high-level managers within NHS and local authority organisations. The meetings acted as a place where NHS organisations could discuss issues that they considered important areas for further work, and the CLAHRC team developed these ideas into discrete programmes of work in which they provided the resource for tailored evidence to be produced.

In this sense, as led by the director, this CLAHRC has evolved to incorporate new objectives as it has developed over time. However, the original vision of the CLAHRC model has been central to supporting the development of these new components. In particular, the knowledge-exchange forum has drawn on the connections fostered by the clinical teams, and the reputation established within partner organisations and beyond. This has allowed them to generate a membership of executive-level personnel who are in decision-making and management positions within their own organisations. Overall, the conceptual vision of the CLAHRC has helped to expand the CLAHRC programme of work beyond the initial set of clinical projects ( Box 3 ).

Reflections on the Bluetown Collaboration for Leadership in Applied Health Research and Care model and its development over time

The Bluetown CLAHRC’s vision and structure enables it to collaborate readily with those communities that are more aligned with the working practices of its members. The partner organisations and the members involved in the initiative, therefore, tend to support the principle that only rigorous evidence should be used to inform service developments. However, within this constraint, the Bluetown initiative does support new operational approaches, and in particular emphasises that project teams need to develop collaborative relationships with external communities, especially management and decision-makers.

With no overarching CLAHRC-wide strategy to support KT activity, each theme has developed approaches tailored to their own local communities and clinical context to support the mobilisation of the knowledge produced through their programme of work. However, as the CLAHRC has progressed over time, central management have augmented the original structure to develop new aspects to support overall Bluetown CLAHRC work. For example, they identified that CLAHRC-wide support was required to support the dissemination activities of the individual project teams. As a result, new central support was established to facilitate engagement of the project teams with end-users of their research, and to support the translation of research findings through a CLAHRC-wide communication strategy within the region.

Management, governance and organisation

Greentown CLAHRC has had several changes in leadership since the decision to apply for a CLAHRC was made, with each one bringing different types of professional expertise to the role, ranging from clinical–academic, through business school academic, to health-services executive management. These changes in leadership are reflected in the ways in which the vision and approach of the CLAHRC has evolved from its conception and funding application through to its latter stages. Each new director has contributed a particular characteristic to the CLAHRC through their leadership. As the first director of the CLAHRC observed, the novelty of the role itself created a need for it to be ‘interpreted’ and ‘enacted’:

Even if they’d had a visible leadership role before the structured things were set up the brokering, engagement, it can be quite intangible. These sort of things have come to the fore . . . I am doing a leadership role that’s much more strategic and autonomous and I think there will be a degree of variability and interpretation of what the role is and how it’s enacted. My first port of call was to get practice on board. Those chief execs are quite frightening people sometimes. So get practice on board. So I went round every trust board that was a partner in CLAHRC and presented the CLAHRC.

GREENTOWN001

As this suggests, the agency demonstrated by the leadership of the CLAHRC has been important in shaping its approach to KT. The original head of the CLAHRC bid was a clinical academic with a strong reputation in mental health research. A change in leadership was made to a specialist in organisational studies, with the aim of emphasising that the CLAHRC model was about a different way of working compared with traditional, standalone clinical–academic projects.

Overall, however, Greentown CLAHRC is structured around a set of clinical research programmes of work designed by clinical academics (i.e. typically a professor from one of the medical school clinical subdisciplines). These focus on applied health research issues to do with the delivery of services for chronic and mental health conditions. Although the basic form of the clinical themes persisted over time, they were eventually complemented by the creation of explicit structures for cross-cutting work. A considerable proportion of the finance was reallocated away from clinical research into implementation science to emphasise the integral role of this approach to their proposal. This process of evolving the planned model for the CLAHRC has generally been viewed as a positive process that enabled the CLAHRC strategy to be well defined by the time it started, supporting earlier progress.

And I mean, that letter and the feedback, you know, it was fairly explicit and it provided a platform for the reworking in the bid. You know, along implementation lines and gave me legitimacy to lead it . . . And I think that referral process was really useful for us. I think we were able to hit the ground running to a much greater extent than the other CLAHRCs.

GREENTOWN001

Initiative-wide structures connected each clinical project team to members who provided specialist academic expertise, such as KT or statistical support. These fitted in with CLAHRC-wide activities, and were incorporated into the Greentown CLAHRC organisational model as a way of facilitating the sharing of different types of expertise with members from varied disciplinary and professional backgrounds.

This incorporation of this greater emphasis on implementation work was included after the majority of the core individuals had already ‘signed up’ to be part of the bid, and at a point when much of the focus for the clinical research projects had already been decided. In effect, the new KT strategy was grafted onto a CLAHRC structure that had evolved around a more traditional clinical–academic research model. Therefore, this new emphasis on an explicit implementation strategy had to be adopted by CLAHRC members with established interpretations of their roles, and where much of the content of the CLAHRC’s work programme had already been specified.

The majority of CLAHRC members are employed by the university, and many are co-located. However, as the CLAHRC spans multiple university departments, bringing together academics from a clinical–academic background within community health sciences, and social sciences from business and sociology departments, it created significant challenges for members with different types of expertise trying to work together.

For the clinical scientists this is a completely new way for them to do any work . . . They’ve never thought about the wider implementation aspects and actually how do you physically get research into practice.

GREENTOWN002

As Greentown CLAHRC is built on a cross-disciplinary academic collaboration, this CLAHRC has faced additional challenges of working across the academic norms and practices of several different academic disciplines, including clinical sciences, nursing and allied health, and management and social sciences. In practice, it has sometimes been difficult to align these different spheres, especially in relating the conceptual remit of the cross-cutting themes. A further change to the leadership was made in the middle phase of the CLAHRC’s time frame, bringing in a director with a background in NHS executive management. This move helped to bring about a shift towards greater engagement with different types and levels of NHS organisations, in order to make the CLAHRC more NHS facing. At the same time, a key part of the restructuring from the mid-term review of Greentown was designed to combat uncertainty about the role of different programme components, and particularly the cross-cutting themes. In particular, as the CLAHRC’s vision emphasised new ways of working, it had been difficult for individuals to comprehend what their role should involve, or to fit this into traditional career trajectories.

As part of this restructuring, roles were divided into academic and non-academic parts, with the aim of allowing each member to focus on their own form of expertise, and to provide clear access to other types of skills. As a result, the CLAHRC was able to become more NHS facing, with a clearly defined team to assist with the delivery of research to practitioner and policy communities.

Networks and collaboration

A typical configuration for the arrangement of positions within project teams had been designed by central management and was used across the initiative. Each core team was established around the team leader, which generally involved other colleagues from within the same academic subdivisions, and the employment of other team members who had trained within similar academic areas to fill designated roles for research and management of the programme of work. Additionally, management created a team structure that aimed to create connections between the core members of each project team, and other team members who could contribute different forms of expertise. These included ‘knowledge broker’ roles aimed at facilitating externally directed ties within health-care or community organisations. These acted as an explicit mechanism to support KT between different communities. An account provided by one team member who had taken on one such externally focused ‘knowledge brokering’ position describes some of her responsibilities within the team, and highlights how working across boundaries was achieved. She describes how an important part of this aspiration is for the team members in the designated brokering positions to spend time with external groups who are relevant stakeholders for to the intervention they are studying.

I did a lot of meetings etc., doing a presentation, explaining about the project. You need to tell them what’s happening with the project and how it’s going to be implemented locally, what the practice can have on them. And they’re a key part of making it work. Whether you . . . take their ideas up or we’re bringing ideas from the [initiative] to them, I just think it works both ways. So that won’t work or this won’t work and what about these participants, how is that going to work? Is there any money for that and then have we got stuff for that. Stuff that people in the university don’t understand.

Members in defined boundary-spanning roles within academic-centred project teams describe how they use team meetings to, for example, bring in insights about public health issues to the academic core team members. We observed in one discussion how project meetings were used as a place where the implications of the information provided by external groups was incorporated into a discussion on developing a sustainable implementation of the intervention which the project team were researching. Senior academic members of the team integrated this information by aligning it with the scientific and methodological approach for the project work, such as considering how this related to the study design, and the implications changes to the protocol would have on demonstrating the academic rigour required for high-quality journal publications.

I think a CLAHRC is about doing things differently to what’s been done before. I think the only way to do that is to bring in people who have different backgrounds and different experiences, who understand the world in a different way. Because I think one of the problems that the CLAHRCs are set up to address is that academic research is done for academics and clinical research is done for clinicians and the twain don’t meet. And I think this CLAHRC has been sensible in bringing in people from different backgrounds.

GREENTOWN002

The structure of Greentown is centred on themes defined by related clinical areas. This supports the achievement of collaboration between CLAHRC members and relevant stakeholders in the local area. The commonality of many activities covering mental health, and related projects within the stroke rehabilitation theme, has provided an opportunity for CLAHRC members to develop communities around these clinical areas. In building on the formal roles to link project teams with external groups, efforts were made from the project team leadership to develop relationships with communities relevant to their research topic. With the senior leaders already having a reputation within their field, this provided a platform for connecting with stakeholders from within the local region, and in presenting a body of ‘CLAHRC-type’ work to national clinical groups.

As part of the CLAHRC-wide strategy, the building of ‘communities of practice’ is actively promoted as an approach which can ultimately support the implementation of evidence into local policy and practice. Four groups are perceived as particularly important: academics, clinical practitioners, health services decision-makers (managers and commissioners) and service user representatives. In particular, creating connections and building on links of existing networks of practitioners is viewed as a strategy to support this aim. In addition, the importance of developing relationships with opinion leaders within the local area has been described as an important part of the implementation process. The CLAHRC has developed an inclusive register of associates, where anyone who is interested in the CLAHRC work can sign up to receive updates about the CLAHRC work and are invited to events. Many of the activities and events are formed around the clinical topics that are of mutual interest, which forms an emergent community for people within the local area to connect to certain groups within the CLAHRC.

Knowledge translation

Greentown adopted an explicitly translational approach for informing external communities, such as commissioners, decision-makers and clinical practitioners, about the results of the clinical teams’ programme of work. This was based on a standard template and style of writing which emphasises the implications for policy and practice. These ‘bite-sized’ outputs operate as a tool to support KT of the work of the CLAHRC to the local practitioner and commissioning community. They require the academic teams (with assistance from specialist support members of the CLAHRC, and those with insight into practice) to tailor the way in which their work is presented for different types of audiences ( Box 4 ).

In addition to organisational-level mechanisms, roles performed by individuals acted as bridges to support KT between different communities. Members in defined boundary-spanning roles within academic-centred project teams describe how they use team meetings to bring insight to the academic core team members relating to issues about using the public health intervention in a community setting:

I add an ability to look at things from an outsider’s perspective and say, ‘why are you doing it that way?’ And I think there’s as much challenge to academia as there is to the NHS. I think this CLAHRC along with the others were set up with ‘we’re the academics, we’ve got the knowledge, we’re going to tell you what you need to know’. And actually one of the biggest shifts that’s happening here is that academics have to get a sort of understanding of a new process around the democratisation of research.

GREENTOWN009

Ultimately, these individuals were able to work with the academic team by enacting roles, such as ‘translating’ how information should be presented for different audiences. They also acted as ‘interpreters’ when providing insights into the practical implications of particular implementations of interventions, and helped to negotiate or legitimise the CLAHRC’s work in non-academic settings:

One of the goals was to try and encourage evidence based practice, and it’s using people like [the clinical practitioner who is a team member] to try and overcome some of the barriers. He’s very useful, as being a clinician he would command quite a lot of respect.

GREENTOWN018

However, this KT occurred very much at the periphery of the project team’s work, with little impact on academic work practices. The creation of explicit knowledge-broker roles provided a formalised link for the team to work with those with academic expertise, and we observed how the project team meetings were used by its members as a focused time and space in which the insight and knowledge of those from beyond the community of the core team, such as academic advisors, and those representing the vision of the central management of the initiative, could be considered.

Reflections on the Greentown Collaboration for Leadership in Applied Health Research and Care model and its development over time

As noted, this CLAHRC has explicitly aimed at developing a model in which its members work in new and different ways. However, the qualitative data suggest that, in practice, there have been issues with the effectiveness of the CLAHRC’s original model. Some members (including leads) have failed to fully engage with the work of the CLAHRC (a key objective of the CLAHRC model), as they have struggled to understand the purpose and remit of certain elements of this CLAHRC’s structure (e.g. what the purpose of the implementation theme is; what the role of knowledge brokers is), and how they can integrate the CLAHRC model into achieving clinical sciences programmes of work.

The changes that took place in the mid-term of the CLAHRC’s life cycle were designed to take these challenges into account. Overall, the vision of the CLAHRC was reaffirmed, including the organisational learning model and its structural features and roles. However, it was recognised that the original CLAHRC model had led to some confusion about the contribution of particular roles and themes. As a result, the restructuring involved redefining roles to emphasise the specialisms of particular members, while providing greater resources for individuals with the expertise to make the CLAHRC more ‘NHS facing’.

Management, governance and organisation

Browntown CLAHRC is built around a novel collaboration between the two main academic centres within the region: one which focuses on traditional forms of clinical science work, and another with expertise in allied health. This allied health ethos is reflected in the types of partnerships that are being built with the NHS, with strong links with senior and middle-level management in nursing and allied health within the partner trusts. These have been influential in informing the Browntown programmes of work.

This is not centric about the university. This is about an NHS collaboration that is distributed across a health environment, a health economy . . . It is about long-term conditions, and about knowledge mobilisation.

BROWNTOWN002

In practice, this CLAHRC model faced challenges in creating a coherent overall programme of work that integrated members from a wide range of professional and disciplinary backgrounds, including different types of academics, and different groups of health-care practitioners. Nevertheless, especially at the core management level, the overall emphasis of the CLAHRC is to develop integrated work programmes involving both academic and NHS groups, and for these connections to be across different levels (i.e. to create links at high- and medium-level health-care management, and also NHS practice-level participation), and also to build the work of the CLAHRC across the whole Browntown region. In particular, the capacity-building objective of the CLAHRC is about building up the skills and expertise in all partner organisations, and ultimately bringing about an evidence-based applied health research culture across the Browntown region.

Overall, Browntown CLAHRC is formed from members of a wide mixture of academic disciplines and different groups of health-care professionals. This heterogeneous composition means that the work of Browntown CLAHRC has the potential to innovatively incorporate many different perspectives and draw from different types of expertise. This provides this CLAHRC with a strong foundation to develop a novel approach for an inclusive and collaborative model of applied research – one better able to span the boundaries between the ‘producers’ and ‘users’ and research. The clinical project work is structured into a small number of clusters which each constitute a programme of work within a common clinical topic. This deployment of members into relatively large subcommunities helps to foster collaboration between members who come from different departments and organisations. On being involved in a project that was one of a set within a larger programme of work, one respondent comments about the benefits of belonging to the translational initiative, as opposed to doing standalone project work:

It was a group of like-minded people where we could work together, because often we would all be doing independent things and not working together. So it was an opportunity for people to work together in some sort of funded activity which would have more power and influence than an individual academic working on their own.

BROWNTOWN010

It was felt that having an interlinked set of projects could potentially have a greater impact. The overall programmes of work were often designed to inform different levels or aspects of health services for the same clinical areas. This more ‘holistic’ package of activity was perceived to have the potential to have a greater impact. In belonging to a larger team, there could be a co-ordinated presentation of activities to policy groups, with team members able to draw on the established connections with other groups to legitimise their own work through the high reputation of other colleagues. However, as members remain part of their ‘home’ organisation, this presents some operational challenges, as limited day-to-day interaction can mean that some members cannot easily access tacit information, or do not have a day-to-day reinforcement of the CLAHRC ‘approach’.

Many of the team members have several roles across different parts of the initiative, such as being members of clinical project teams while holding positions within the core management of the initiative. The central management team comprises a large group of CLAHRC members, meaning that internally the CLAHRC adheres to a distributed leadership style. This helps to support the diffusion of the values and aspirations of the central management team within the project work, and helps to reinforce the vision of the CLAHRC across the whole work programme. Many senior CLAHRC members, both from within the core management and the theme leads, are implicit boundary spanners – sometimes by dint of having ‘dual contracts’ with both NHS and university bodies – who contribute hybrid expertise and ‘belong’ to more than one community. They help to support the aim that the various programmes of work should incorporate different perspectives. The overlap between a large core management group and those in positions of leadership within the project teams helps to spread this vision. A smaller core group co-ordinates the overall CLAHRC-wide organisation, and presents the external-facing view of this CLAHRC model, and is influential in driving the vision throughout the CLAHRC. Overall, the onus is on project leadership to co-ordinate the different types of knowledge into one coherent programme. The leaders of the CLAHRC itself view their role as one of facilitating a new form of collaboration, rather than providing specific scientific or methodological expertise ( Box 5 ).

The work taking place within these themes or project work is not dominated by the vision or traditions of one individual (e.g. the theme lead), but a culture has been created where individuals can contribute from their own perspective. As there is no one dominant culture influencing the approach of the work programmes, members with ‘specialist’ types of expertise are integrated members of project teams, allowing different types of knowledge to routinely inform the programme of work.

In [this translational initiative] you’re going into situations all the time where everyone in the room has got lots of different roles. That can be a bit of a challenge at times with people having to approach things from lots of different perspectives. It’s very much going in and out of roles sometimes.

BROWNTOWN013

Members within the Browntown CLAHRC typically evolved more flexible and overlapping roles, reflecting the need for expertise to support different aspects of the work programme across the initiative.

Networks and collaboration

Although the overall theme of each programme of work remained largely unchanged during the study period, the CLAHRC model and leadership in Browntown has allowed different groups to shape the focus and direction of the work packages. By collaboratively working with other stakeholder groups, the work packages were shaped by the values and insight of different communities, in what one respondent describes as an ‘organic process’.

It’s a tool for facilitating research, applied healthcare research. To enable patients and clinicians and commissioners to make sense of decisions about what to, about what types of treatment to provide . . . The overall structure of the research design didn’t change but it was such an organic process really, what we set out to do is what we’re doing, but their support and interest and feedback was important.

CMBROWNTOWN002

In describing the work programme within the clinical theme, the participant highlights the flexibility of the plans for the project work, and how this allowed the integration of insights from different groups to inform and shape the direction of the work ( Box 6 ).

The project teams across the CLAHRC were composed of a mix of academics from different disciplines and clinical practitioners. The senior management group actively encouraged teams to continue developing their original proposals based on discussion and dialogue with stakeholder groups. This more emergent approach was enabled by certain features of the CLAHRC model and membership. Within the project team, for example, several of the team members were what we will term ‘hybrid’ individuals, that is, they were affiliated to both academic and practitioner communities, and they helped to foster an environment where no one group dominated the direction of work programmes. Instead, all team members were encouraged to actively interact with other groups, and to be flexible in doing their project work. Thus, the vision of the CLAHRC fostered an environment where members were encouraged to develop new work practices and build relationships, rather than to conform to particular disciplinary approaches.

Across the Browntown CLAHRC, many of the members were able to act as boundary spanners precisely by virtue of their ‘hybrid’ academic and health practitioner background. This supported a more fluid integration between research and practice and the building of sustainable relationships, as the overlapping roles conferred membership of both the CLAHRC community and of external groups of managers and commissioners. At the same time, those in leadership positions helped to create an environment where knowledge from different perspectives was routinely shared across teams. Boundary-spanning mechanisms, such as project meetings, were used to support the fluid integration of different perspectives into the various components of work involved.

You see everyone has got a different perspective. Whether you’re a commissioner, you know, perhaps coming from a public health or social services background, but you’re commissioning. Or a manager in the NHS, perhaps social services seconded to NHS. Or a doctor or a nurse or a psychologist or a GP or a service user. You’ve all got a different understanding of what the care pathway is and what needs to be done to improve it. And so very much we deliberately wanted to incorporate a collaborative project between all those different groups.

BROWNTOWN008

Knowledge translation

This CLAHRC’s KT approach drew from an established implementation model, the Canadian ‘knowledge into action’ cycle. In particular, many of the members of Browntown CLAHRC come from the NHS partner organisations, and many of the academics involved with this CLAHRC also hold NHS contracts and have been practically involved with roles at the local NHS organisations. In this sense, the CLAHRC organisation itself acts as a key mechanism to support the translation of knowledge between different communities.

I think CLAHRC is a boundary spanner. That’s its job, that’s what it is. Because the organisations, the NHS organisations, I mean, they do talk to each other because of, you know, Department of Health policy and stuff but they all have different ways of implementing policy. And so they don’t necessarily talk to each other but through us there’s work happening that can be, you know, translated across the different organisations. So we do act as kind of a, it’s almost like a phone exchange.

BROWNTOWN017

The Browntown CLAHRC model included a number of work programmes which were designed from the outset to focus on explicit late-stage implementation activity that would produce tangible impact at an early stage in the CLAHRC time frame. These themes were not designed to produce new research evidence, but were intended to align with the NIHR aim of learning more about KT by conducting implementation activity.

We need to demonstrate progress with implementation from the start of CLAHRC. They are implementation projects using research methods, as well as making a difference to practice, and we’re very much committed to adding to the body of knowledge about knowledge translation . . . Implementation work is really a cross between action research and participatory research, and you need to shape the projects as you work with the key stakeholders.

BROWNTOWN001

The implementation programmes of work were designed to produce results and impact on practice at an early stage. Thus, although they drew on established research expertise, they also emphasised collaboration with the local partners, including at all levels (executives, middle-management and ground-level staff) with whom the activities were taking place. These strands of work thus helped to reinforce and promote the expectation that all programmes of work, including the research themes, would generate outputs that could be applied to inform health-care practice.

The KT themes, in particular, focus on developing evidence that will help to make a practical difference to local health-care services. The clinical priorities for implementation are identified collaboratively, and the project work is seen as socially embedded in the organisations where change is happening. This facilitates the effective translation of knowledge into action by individuals and teams.

There’s a sense in which often researchers come to the NHS with their research ideas and then you try and get sign-up from them to take a project forward. This is turning the coin over completely and it’s saying, ‘we’re a resource; we want to work with you. What are your priorities?’ It’s been getting that ownership that I think has meant that we work in a different way, but also we get a different response back from NHS managers and clinicians. But also seeing, I guess it’s the role of brokers in all of this.

BROWNTOWN001

The main focus of the work of these themes is on piloting and evaluating innovative strategies for implementation, which are then planned to be rolled out across the NIHR CLAHRC for Browntown partnership (with further evaluation of their impact). One key mechanism in this effort involves focusing resources on individuals who are employed by the health-care trusts. A role was designed for these members to act as ‘facilitators’ to build up research activity within their organisations based on the priorities and general ethos of Browntown CLAHRC-type work.

We have what are called ‘research and development facilitators’. The model which was developed was to actively engage each partner healthcare organisation by having a person working within them. So we would bring CLAHRC to these organisations. My role is to bring research evidence to my organisation, and encourage people to use the research evidence much more determining the way that work is completed. And also to help people here to articulate some of their, some of the issues which they have which could have a research solution. They could find a solution through research or at least some preliminary studies.

BROWNTOWN022

It is clear from our analysis that members from throughout the CLAHRC recognise that there is an underlying strategy for this initiative to facilitate KT, and that this is conceptualised as involving an integrated, collaborative approach between academics, health-care practitioners and managers in order to facilitate implementation work. Thus, although the CLAHRC does designate certain work packages as ‘implementation’ projects, the vision enacted throughout the CLAHRC emphasises the need to bridge the second translational gap, and not limit KT activity to discrete programmes of work.

It’s not that simple as just having an implementation arm in CLAHRC . . . I think that a lot of people have been practitioners and have done research, there are very few pure academics who have never really linked in and have been in practice or not in the NHS . . . if you look at nearly every lead they’ve either been a practitioner or had a role in the NHS before as well as doing research. They’re all boundary spanners.

BROWNTOWN021

However, while this priority given to KT is very apparent at the core management level of the CLAHRC, it is perhaps more difficult for this type of approach to radically influence all areas of the CLAHRC work, and in particular for innovation approaches to inform the approach of the research-focused clinical theme work. In addition, the challenges of conducting implementation research are also highlighted, as the inductive co-production approach here differs from the design of conventional academic projects. Thus, while achieving KT through the seamless integration of research and practice is an important part of the Browntown CLAHRC vision, it depends heavily on the contribution of those occupying ‘hybrid roles’. Crucially, the individuals within these roles are not deemed peripheral to the project teams, but are well positioned to incorporate insights from different community perspectives into the work of the teams.

I work between a number of different organisations, so principally the NHS and academia. It was useful that I am actually from an academic background myself. It makes it a little bit easier in terms of understanding what academic opportunities there might be which the NHS might be able to tap into . . . I work with the initiative really because I work for the NHS and ensure that the CLAHRC work is embedded within this NHS organisation. So it’s very much that boundary-spanning role, I have two identities.

BROWNTOWN022

Reflections on the Browntown Collaboration for Leadership in Applied Health Research and Care model and its development over time

The vision of the Browntown CLAHRC collaboration aims to foster a change in the culture of members across the partner organisations by building the capacity of its members to engage with applied health activities. The fact that the core of the CLAHRC is not university-centric, but is instead based on the traditions of allied health academics and practitioners, has helped to privilege health services’ concerns in shaping the programme of work. Although discrete KT activities were established at its inception, the vision of the Browntown CLAHRC is for all programmes of work to engage with diverse perspectives and traditions. In this way, the CLAHRC builds capacity for innovation through a work environment where work practices are able to draw on a diverse range of perspectives.

The Browntown CLAHRC has evolved incrementally over time as new activities have been developed in response to partner need. From its inception, the CLAHRC has emphasised the scope for such incremental growth both through grafting on new programmes of work and through the expansion of the CLAHRC community to include new types of partners. Although not emphasised in the original bid, collaboration with industry has also emerged as an important strand of this evolution. One result of this approach has been the ability of the CLAHRC Browntown model to adapt to a changing policy landscape, as the ‘organisational memory’ of relationships with the primary care trusts has been used to forge relationships with new commissioning organisations.

Management, governance and organisation

The pilot project involves four main players (organisations): (1) the Regional Community Care Access Centre (RCAC), which is a community health provider that organises home, school, and hospital care, developing customised ‘care plans’ and providing support from health-care professionals, nurses, physiotherapists, social workers, registered dieticians, occupational therapists, speech therapists, and personal support workers to provide a range of care and support services; (2) the Ottawa Association to Support Children (OASC), which is another community health provider that provides specialised care for children and youth in Ontario with multiple physical, developmental, and associated behavioural needs; (3) the social services (SS), which is an agency that develops case resolution mechanisms to provide recommendations and referrals for families with children with complex care needs who are experiencing difficulties accessing support and services in the community; and (4) the White Hospital, which is the ‘hub’ of the project in that the children with complex care needs are patients of the hospital and, therefore, the main treatments are provided at the White Hospital.

The pilot project funds three key people to manage the project: (1) a project manager, (2) the most responsible physician and (3) the nurse co-ordinator. The project manager supervises the pilot project: she ensures that processes, communication pathways and flow maps have been developed; conducts staff training sessions; co-ordinates the meetings and presentations for the steering and advisory committees; and prepares reports. The most responsible physician reviews the overall complex medical needs of each child and co-ordinates communication with all the specialists at the White Hospital, other tertiary paediatric centres specialists, and the community physicians. The nurse co-ordinator works very closely with the most responsible physician, interfacing between the doctors, nurses, and managers at the White Hospital, and the other agencies; the nurse co-ordinator is the link person for all of the families of the children in the project. Interestingly, along with a solid governance structure, informal relationships played a central role in the pilot project, as is outlined in the example below ( Box 7 ).

From the case example in Box 7 , it is clear that informal relationships facilitated tight collaborations and promoted trust among the players involved in the pilot project. The development of trust, as we will show below, was also a relevant element in supporting the health-care network of the Ottawa community.

Networks and collaboration

The pilot project exploits existing networks between the players involved including RCAC, OASC, SS, independent paediatricians, and the White Hospital personnel (doctors, nurses and staff). In fact, while the project involved hiring some key people who could facilitate collaboration among the players, many of the relationships between, for example, social services (RCAC) and the nurses and doctors in the hospital (the White Hospital) were already tight. This aspect (prior networks) has positively affected the overall project because while formal networks are relatively easy to establish, it takes time for people to begin working together productively. Interestingly, the pilot project network developed collaborations aimed at improving the quality of health-care delivery that were not limited to the players that were involved in the initial project. In fact, the managers (of the four agencies) soon realised that in order to clearly identify the needs of the children involved in the pilot project, it was important to involve the patients directly. Therefore, the project manager, in accordance with the directors of RCAC, OASC, and SS, decided to involve the parents of two children involved in the project. This involvement included having the parents sitting in the steering committee of the project that meets monthly and makes decisions regarding how the co-ordination of care at the White Hospital and across its network can be improved.

As the project manager highlighted regarding the evaluation process of the project: ‘So the two parents from family forum that also sit on our steering committee for this Pilot Project reviewed our questionnaires and helped us get to the questions we wanted so that it’s more a participatory evaluation approach’. This quote underlines a collaborative climate where actors belonging to different networks are willing to bring their contributions and are supported by a common aim: to identify ways to improve the quality of health-care delivery services for children with complex care needs. One of the most relevant issues that emerged from the feedback with the (two) families who sit in the steering committee was the difficulty of having all of the different community services up to date with the most recent changes regarding the children’s condition – this problem was made very challenging by the fact that each child is frequently seen by a number of specialists.

Knowledge translation

Knowledge translation within Canada-Coordination was focused on relations between clinicians and with family members of the children involved. It was facilitated not only by direct interactions among these groups, but also by the development of artefacts that could help span the boundaries between them. In this context, the introduction of the Single Point of Care (SPOC) document is an important example because it radically improved the relationship between the White Hospital and the health-care networks. The SPOC is a paper-based medical sheet including all basic information about a child’s health, such as current health status, current treatments (medications), and any other detail that can be helpful at school (e.g. current allergies) or in emergency situations (particular drugs that the child needs to take if his/her condition suddenly becomes severe or life-threatening). The SPOC is issued by the White Hospital and is carried by the families, who now no longer have the difficult and at times confusing task of collating themselves all the medical information related to their children produced by different specialists. The SPOC is also shared with a number of organisations in the Ottawa community, such as police and schools. In particular, it is very relevant that, for example, if the child develops a new allergy or changes one medication, the school nurse is aware of the changes. While the creation of the SPOC will be broadly discussed in the next section – the SPOC being a KT tool – the following case is meaningful in highlighting how SPOC contributed to promoting networks and collaboration across different players in the Ottawa community ( Box 8 ).

The SPOC also helped to reduce redundant examinations because the parents show the medical sheet to each specialist who would edit it as appropriate and gather information about forthcoming tests. In sum, the introduction of the SPOC improved efficiency (being a co-ordination mechanism for different specialists) and provided the families with more awareness about the conditions of their children. In fact, while on the one hand the SPOC is acknowledged by any doctor as an official document because it is issued by a hospital, on the other hand, the (simple) way the SPOC is structured allows the family to interpret (in general terms) the health status of their child.

In sum, according to the project founders (the leaders of RCAC, OASC and SS), the involvement of the parents had produced significant benefits. This point was also confirmed by a number of interviews that we conducted with the families of the children involved in the pilot project because a number of issues of the families in terms of co-ordination emerged that could be brought up for discussion in the project meetings (steering committee).

Reflections on the pilot project model and its development over time

The pilot project is an example of how informal networks are effective in promoting KT and collaboration across community partners, including health-care partners such as a hospital and the SS, yet also including also non-health-care partners such as the police department and the schools. The management of the project includes periodic controls undertaken by an independent entity that is the White Hospital Research Institute. The White Hospital Research Institute surveys doctors and families of the children involved in the project and makes sure that co-ordination and health-care delivery services are continuously improved. This is evidenced by feedback from both clinicians and patient families.

Management, governance and organisation

The Canada-Translation Centre started in 2010 with the aim to promote and co-ordinate clinical research carried out by clinicians who work at Com-Hospital and are also appointed by Uni-Canada, a university in Quebec. According to the regulatory framework of the Canada-Translation Centre, it carries out clinical, epidemiological, and health services research studies, provides consultation to other researchers in the hospital, and provides educational services, including seminars and workshops, related to research.

The Canada-Translation Centre is led by Johanna, Associate Professor at Uni-Canada and the vice president (VP) of Academic Affairs at Com-Hospital. The Canada-Translation Centre organises monthly meetings (last Thursday of every month) that are held in the hospital. As at the time of our research the Canada-Translation Centre was in its initial stage of development, the main objective of the meetings that we observed involved the establishment of guidelines and systems.

From a governance perspective, the three main bodies that collaborate with the Canada-Translation Centre are (a) the REC, which is responsible for reviewing the scientific and ethical aspects of all research projects involving human subjects; (b) the research review office, which is co-ordinated by the research administrative secretary and provides administrative support to both the EC and the REC; and (c) the Quality Assessment (QA) Unit, which provides assistance to hospital staff on various aspects of QA analysis and evaluation, that is, projects that assess current performance or practice.

The governance aspects of the Canada-Translation Centre are extremely formalised, reflecting the interest of the academics at Uni-Canada to secure control of the development of the project, especially in its early stages. The board of directors supervises all research activities, while the VP of Academic Affairs (who is also the Canada-Translation Centre’s Director) supervises all main research committees. The Quality and Risk Management Committee is an independent body chaired by the VP of Professional Services. According to an interview with Johanna, the rationale of this design lies in the desire to have objective control of research activities by an independent body (the Quality Committee).

Networks and collaboration

The network of the Canada-Translation Centre is formed by people who work in the hospital who also have at least a teaching appointment or, more often, a professorial appointment, at Uni-Canada. The design of the three key bodies of the centre sought to promote collaboration and cross-fertilisation between the hospital and Uni-Canada. In particular, within Family Medicine at Uni-Canada there are four departments that have been always very independent and loosely coupled; however, with the start of the Canada-Translation Centre, members of these four departments started meeting twice a month to try and pursue common research objectives. In sum, the main objectives of the Canada-Translation Centre are to (1) tighten the collaboration between Uni-Canada and Com-Hospital by promoting networks and (2) bring in expertise from external collaborations.

In terms of improving the network within the Canada-Translation Centre (Com-Hospital and Uni-Canada) one of the initial challenges was to try to build relationships between key actors who had no experience of working together. For instance, quality and risk management staff rarely interacted with academics prior to the start of the project. This group quickly became committed to collaborating with the academic groups, as highlighted by the director of the quality and risk management at Com-Hospital:

And I convinced Johanna that the uniqueness of the relationship between Quality and Research here has to live somehow and has to be formalised in whatever plan she does. And I think she saw that as an opportunity. If it were someone else here than me I don’t know if it would work the same. It’s just because I started out working in Research and I, you know, I did that for ten, twelve years and I know the language and I know their challenges and I know, you know, how they think and how they . . . operate.

While it is important that people within Com-Hospital are able to work together (e.g. the research team and the quality and risk management department, as per what was highlighted above), it is also relevant that the (more practice-oriented) researchers at Com-Hospital collaborate and develop networks with the professors at Uni-Canada. The associate dean of Inter-Hospital Affairs plays a role in doing this by trying to promote fruitful collaborations between the hospital and the university. As he highlighted to us:

It was a natural thing to develop research along these lines. Again my role has been to facilitate communication between the hospital and the university, both with the Dean of Medicine and with other research leaders within the university. I’m not a researcher myself. My role is mainly to put people together. And to use the links that we have with the Ministries to help support what the hospitals do. The other way in which I interplay with the hospital is I’m on the board of directors. So being on the board of directors of course some of these initiatives to fund this research infrastructure comes to the board and at times I speak to it in a supportive fashion in order to help promote this venture for the hospital.

Knowledge translation

Included in the Canada-Translation Centre mission for the period 2010–15 there are five main domains of research to be pursued:

• to support and conduct high quality clinical and health services research relevant to the patients and services provided by Com-Hospital

• to promote KT and exchange activities to support evidence-informed decision-making in practice, management and policy

• to provide decision support for clinical and management leaders through high-quality rapid systematic reviews of scientific evidence

• to support the application of evidence through quality improvement and other implementation initiatives

• to provide a high-quality training programme and environment for students interested in research careers.

From the above it is clear that the Canada-Translation Centre has a specific focus on KT. In fact, each research member (i.e. a professor) must also be a clinician (i.e. he/she needs to spend a minimum number of hours per week in the hospital), a rule which is aimed at promoting applied research. As outlined by the case example below ( Box 9 ), one of the strategies that the Canada-Translation Centre adopts for promoting implementation research is to work on small projects whose short-term tangible results are directly testable in the hospital environment.

Reflections on the Canada-Translation Center model and its development over time

In contrast with the pilot project in Ottawa, this initiative is much more complex and involves more than 50 people including Com-Hospital physicians, Uni-Canada’s academics, and administrative personnel. Staff involved are fully aware of the barriers between academics and practitioners and between people who work in different departments (in the case of Com-Hospital) and in different institutions (at the hospital and at Uni-Canada). Therefore, specific boundary spanners – such as the associate dean of Inter-Hospital affairs – were identified to connect people with different background, professional and personal interests.

Management, governance and organisation

The governance structure of US-Health is relatively flat. Federal-Health is the government organisation that manages the tendering process for contracts, and US-Health (the network) is one of the recipients of these contracts. Informal relationships between Federal-Health and its collaborations over time, however, help to ensure that contracts are tailored to the strengths of members of the US-Health network. For example, Federal-Health frequently discusses potential research projects with the contractors of previous bids to elicit their interest.

Once a contract is signed, a project officer is assigned to each contractor; there are few project officers who manage multiple contracts nationwide. The main task of the project officer is to make sure that all milestones are met during the very short period of the contract. Moreover, the project officer can review ongoing documents, can come to visit a contractor to see how the research project progresses, and can provide suggestions and indications. The project officers need to write periodic reports for Federal-Health; therefore, Federal-Health can closely monitor each contractor (and each research project). The way a contractor (e.g. a member of the US-Health network) manages the research project is very subjective. While a proposal on how they will reach the objectives needs to be provided to Federal-Health, no specific guidelines on how to undertake the research are given. However, the contracts (and the research projects) are generally very specific in their objectives and concrete results. Very often, the research output is a ‘toolkit’, that is, a document with specific recommendations on how to manage a health-care problem in practice.

Many of the people whom we interviewed and who were involved in one or more US-Health research projects (i.e. contracts) highlighted that the success of the project depended to a large extent on the project officer. The effects of the project officer role can be both positive and negative, as shown by the example below ( Box 10 ).

Networks and collaboration

US-Health has a very complex network including, as we previously noted, a number of large partnerships. In this section we provide some insights that focus on two main partnerships where we were able to conduct interviews: Health-partnership and Health-Consulting. We chose these two networks because they are very different (one is a partnership whose participants are academics while the other is a consultant company with expertise in health care). Although collaboration on bid writing between partners was limited by a lack of funding, some collaboration did take place across US-Health contractors.

Health-partnership (primary contractor) is a partnership led by a US university (name not disclosed) which managed some 10 contracts with Federal-Health within US-Health. The contracts of Health-partnership involve themes such as improving hepatitis C virus-screening practices and testing uptake in select primary care providers, preventing pressure ulcers in hospitals, co-ordinating care across primary care practices, using innovative communication technology to improve the health of minorities, avoiding readmissions in hospitals using technology, and reducing infections caused by particular bacteria.

Health-Consulting is a much smaller entity, being a private consultancy company with some 15 senior consultants. As a result, it is not in a position to develop clinical research involving patients. Health-Consulting’s strategy within US-Health was, therefore, to find collaborators (subcontractors) who could undertake implementation research in health-care structures. To do this, they drew on a wide network of subcontractors. The contracts managed by Health-Consulting focused on the relationship between patient information management in hospitals and risks of complications and mortality while patients are hospitalised (this contract investigates both electronic and paper-based medical records) and on the development of performance measures for injurious falls in nursing homes and rehospitalisation of patients discharged from hospitals to home care. Although collaboration on bid writing between partners was limited by a lack of funding, some collaboration did take place across US-Health contractors.

Knowledge translation

With the overall goal of translating research into practice, US-Health links many of the largest health-care systems in the USA with top health services researchers who are identified through a tendering process and are managed through short- to medium-term contracts.

It provides a network of delivery-affiliated researchers and sites with a means of testing the application and uptake of research knowledge. US-Health is the successor to another large-scale initiative which was completed in 2005. All of the large partnerships (i.e. the prime contractors who work with a specified range of other organisations) have a demonstrated capacity to turn research into practice for proven interventions, targeting those who manage, deliver or receive health-care services. As per the above, both Health-partnership and Health-Consulting develop research that is focused on a wide variety of demand-driven, practical, applied topics which are of interest to the partnerships’ own operational leaders as well as to the project funders. The overall programme (US-Health) emphasises projects that are broadly responsive to user needs and operational interests and which are expected to be generalisable across a number of settings. An example of implementation research conducted by Health-Consulting is outlined in the case example below ( Box 11 ).

Reflections on the US-Health model and its development over time

The US-Health case, if compared with the two other North American cases, involves a number of partnerships, each partnership involving a number of organisations and players. Moreover, the funding model is quite unique, depending on bids for contracts (very often agreed between parties – e.g. Federal-Health and, in our case, US-Health). Even though there was little interaction across the US-Health network as a whole, collaborations proliferated within partnerships. Moreover, most projects have been successful and some projects were adopted nationwide.

Ambidextrous networks: brokerage and closure

Research indicates that developing and maintaining social network ties can be costly in terms of both time and energy. 100 Therefore, a structure whereby everyone is connected to each other is not viable. Instead, it is preferable to have small clusters of individuals with high levels of network connectivity (closure) that promote implementation and embedding knowledge into practice, for example in work teams or communities of practice. 101 Clusters working in isolation, however, limit the development and transfer of novel information. 62 Thus, it is also important for there to be ‘brokering’ or ‘boundary-spanning’ ties; networks that connect between work team clusters as well as ties that connect to other people outside of an organisation or entity. 102,103 Closure and brokerage, therefore, offer different benefits for KT. 104,105 We describe this ability to reconcile structures of brokerage and closure as ‘network ambidexterity’. In our study, we assess the extent to which the CLAHRC KT networks feature such ambidexterity by incorporating elements of brokerage and closure ( Box 12 ).

Structures for generating new ideas (brokerage)

Brokerage structures have been associated with the opportunity to ‘bridge’ between distinct knowledge camps. This offers potential access to new, in other words ‘non-redundant’, information, as well as to support the fusion of new ideas from different information sources.

Structural holes are ‘gaps’ in a network created where there is a lack of connectivity between actors. Structural holes present opportunities for knowledge brokerage because brokers are positioned at these network gaps. The network gaps presented by structural holes encourage the formation of separate pockets of information and ideas, and so opportunities for boundary spanning between distinct knowledge camps are enhanced. Bridging offers similar opportunities for innovative knowledge transfer. Here, there are two individuals who are not directly connected but, unlike structural holes where the gap remains, the gap has already been bridged by a broker who acts as an intermediary between parties.

Three main SNA metrics tap into this notion of brokerage: Gould and Fernandez brokerage,106 actor-level betweenness107 and structural holes. 108 Table 3 presents average brokerage, betweenness and structural holes scores for each CLAHRC at two time points.

We analysed the extent to which the CLAHRC networks provided brokerage opportunity, with the results presented in Table 3 . The greatest difference between CLAHRCs is at time 1. Greentown had the highest betweenness and Gould and Fernandez brokerage scores, meaning that there were more individuals in this CLAHRC acting as ‘bridges’ and more disconnected pairs of contacts than in CLAHRCs Bluetown and Browntown. At both time points, efficiency scores show that CLAHRC members had a relatively high proportion of non-redundant ties that provided access to structural holes. However, the constraint scores suggest that a significant proportion of these relational investments were focused around a single contact rather than many different pockets of knowledge. By time 2, brokerage, betweenness and structural hole scores became more similar and there was less variance between CLAHRCs.

Structures for embedding knowledge (closure)

Closure in tight-knit (dense) networks has been associated with the ability to build trust and stability, which in turn helps to create an evironment that supports embedding and implementing knowledge. 101,109

The simplest measure of network structure with which network connectivity or closure can be assessed is network density. Density is the actual number of ties divided by the possible number of ties. As outlined in Table 4 , in the Greentown CLAHRC, density was 8.6% at time 1, decreasing to 7.8% at time 2, the most connected of the three CLAHRCs. Browntown was the least connected of the CLAHRCs at both time points (5.3% at time 1, 6.7% at time 2). Greentown’s density decreased over time whereas Bluetown and Browntown densities increased. The results show that there was not a high level or connectivity or closure across any of the CLAHRCs. This is positive for knowledge innovation at the level of the CLAHRC entity level because the too much connectivity would have a negative effect on diversity of knowledge (more redundant information).

While density is a frequently used measure of overall connectivity, it is possible to have a densely connected network that is actually fragmented into two or more subgroups. As a check on fragmentation we also calculated geodesic distance measures for each network. Geodesic distance is the average number of links between one person and every other person in the network. This network measure has been popularised by the notion of six degrees of separation that examines how closely connected any two people are. 110,111 At time 1, we find that the Greentown CLAHRC is the most connected as it has the lowest average geodesic distance score (2.6, compared with 3.0 for Browntown). Therefore, there are, on average, 2.6 connections between every person in the Greentown network. Our findings for geodesic distance are in line with those for density, with Greentown being the most connected and Browntown the least connected. At time 2, the CLAHRCs have established equal geodesic distances between members. Everyone in each of the three CLAHRCs can reach each other within an average of 2.7 links, and this indicates that the KT networks are not fragmented despite the low global density scores.

The third structural measure that we report is reciprocity ( Figure 3 ). This measures the extent to which relations are two-way, so that, where a nominates b as a knowledge contact, b also names a in return. Reciprocity of ties is often used as a proxy for trust, which has been shown to be associated with the sharing of knowledge within a network. 112 Figure 3 provides a visual illustration of reciprocal KT ties in each of the CLAHRC networks at time 1. The data suggest that Greentown members at both time points have a very slightly higher level of trust in each other than in the other two CLAHRCs. At time 2, the level of reciprocity has fallen slightly in each of the CLAHRCs. This is partly attributable to the fact that not everyone responded to our survey.

FIGURE 3.

Reciprocal KT ties in the CLAHRC networks (reciprocal ties in red). (a) Greentown; (b) Bluetown; and (c) Browntown.

Implications of brokerage and closure for knowledge translation policy and practice

Network size is most often deemed to be a positive factor in exchange networks. The more knowledge contacts a person has relationships with, the greater the chance that one of them has the resource he or she needs. Density, that is the overall cohesiveness of a network, will have a curvilinear effect on KT. 113 By definition, some connectivity is required for a network to be a network rather than a set of isolated people. Dense networks are more ‘cohesive’ and ‘tight-knit’ than they are fragmented, and, for KT purposes, such a structure is functional in a way that ensures that network members will be held together (be it through common purpose, shared goals or mutual trust). The speed of KT will be positively impacted by network density. Indeed, research in health-care settings has shown that knowledge will diffuse more effectively in a community of practice. 6,114 In a highly connected KT network, short distances between contacts will provide faster access to knowledge and accurate transmission of information, while long distances (where individuals are not directly connected or even disconnected) can delay and distort the information. Given the limits on the number of ties that can be maintained, it is of course negative if all individuals are directly connected to each other in a KT network because knowledge obtained in such circles or cliques will likely be redundant rather than novel information. 115 In these instances of closure, the network becomes closed off from external influences. It is important for KT networks to feature brokerage via structural holes and bridging because the converse, a knowledge network where all individuals are directly connected, will more likely encourage individuals to conform to accepted ideas and practice rather than develop diverse and innovative ideas. Thus, there is a ‘trade-off’ or curvilinear effect to the amount of closure and brokerage optimally required for a KT network to be effective. 104,105 Table 5 summarises the implications of brokerage and closure for KT while Box 13 gives an example of ambidexterity.

Overall, we find no indication of fragmentation into separate clusters within the CLAHRCs. The CLAHRC knowledge-sharing networks were, however, not particularly closed or ‘tight-knit’, either (in the sense where all CLAHRC members are connected to each other), nor did they display a high degree of reciprocity. We found some notable changes in levels of closure over time. Overall, density, geodesic distances and reciprocity became more similar over time. The Greentown CLAHRC became slightly less connected, whereas the Bluetown and Browntown CLAHRCs saw an increase in their connectedness.

Taken together, our SNA metrics suggest that there had been some organisational restructuring of KT ties over time to moderately improve the capability for CLAHRC knowledge brokering. In doing so, the CLAHRC KT networks also became more similar. However, constraint did not decrease greatly over time which is likely to be because many structural holes would inevitably have been bridged as the CLAHRC developed. In particular, Bluetown had fewer structural hole opportunities at time 2 (efficiency decreased and constraint increased). We explore the actual changes in knowledge ties in the next section.

Information search strategies: translating knowledge across boundaries

Our social relationships with others can act as banks of resources, or ‘social capital’. 116 The portfolio of social capital resources we have access to will depend very much on the quality of our social networks. Arguably, networks that span organisational and professional boundaries will support ‘knowledge brokering’ because they bridge across information camps. 117

The CLAHRCs were organised into working teams referred to as ‘themes’. Each theme had a specific health-care focus and members of each theme therefore shared common reference points in terms of research/implementation goals, projects, working practices and sometimes co-located office space. Each CLAHRC had a different number and set of themes. We analysed the extent to which KT occurred within and between these thematic groups. We measure connectivity between themes as evidence of boundary spanning within the CLAHRC (where individuals seek knowledge from other themes more often than from colleagues of the same theme). Table 6 provides proportions of knowledge ties within and between CLAHRC themes at two time points (see Appendix 1 for a detailed breakdown).

When we consider the total percentage of ties within and between themes, we find that each of the CLAHRC networks has a relatively balanced proportion of ties. At time 1, 60% of Bluetown’s ties are within theme, which is slightly higher than the proportion for Greentown (59%) and Browntown (56%). By time 2, Greentown and Bluetown had increased the proportion of cross-theme ties, while Browntown’s proportions were unchanged. It is important to note the reorganisation of theme groupings in CLAHRCs Bluetown and Greentown between time points, whereas Browntown theme groups stayed constant over time.

Rather than focus on the total of within- and between-theme ties per CLAHRC, we felt it more necessary to highlight the networks of individual themes and ask CLAHRC members to consider the appropriateness of this activity given the remit of each theme. Proportions of cross-theme ties were reported back to the CLAHRCs as part of our stage 1 feedback. We detail the pattern of ties in the figures below, which capture cross-theme ties by theme group at two time points. Themes that were disbanded between time points have been omitted from the figures below.

In Greentown ( Figure 4 ), the programme management theme had the highest level of cross-theme ties over time (time 1: 90.0%; time 2: 94.4%). The most internally connected themes in CLAHRC Greentown at time 1 were stroke, children and young people (CYP), and primary care (with 20.6%, 32.5% and 33.8% of cross-theme ties, respectively). At time 2, the Greentown network saw an increase in the proportion of cross-theme ties for all themes (especially stroke and primary care, which became less internally focused).

FIGURE 4.

Greentown CLAHRC proportion of cross-theme knowledge ties per theme at two time frames. CYP, children and young people; RDSU, Research Delivery and Support Unit.

In Bluetown ( Figure 5 ), core team/programme management and cross-cutting themes had the highest level of cross-theme knowledge ties over time. The most internally connected themes in the CLAHRC Bluetown network at time 1 were stroke and transient ischaemic attack (TIA), pyschosis and maternity support (percentage of cross-theme ties 13.0%, 18.2% and 31.0%, respectively). At time 2, the Bluetown network saw an increase in the proportion of cross-theme ties. Psychosis, maternity support, and stroke became less internally focused. The cross-cutting themes maintained a high proportion of cross-theme brokering ties at both time frames (82.4% at time 1, 73.5% at time 2). The core/programme management team became more externally facing over time increasing its ties to other themes from 60.4% to 73.9%.

FIGURE 5.

Bluetown CLAHRC proportion of cross-theme knowledge ties per theme at two time frames. TIA, transient ischaemic attack.

The Browntown KT network ( Figure 6 ) had a similar proportion of cross-theme ties in each time period (44.4% at time 1 and 44.7% at time 2). At time 1, the themes with the lowest proportion of in-theme ties were diabetes (22.9%), depression (26.1%) and obesity (32.1%). At time 2, genetics, chronic obstructive pulmonary disease and obesity became much less internally focused. Core team, TaCT (Telehealth & Care Technologies) and IntelComm became more internally focused.

FIGURE 6.

Browntown CLAHRC proportion of cross-theme knowledge ties per theme at two time frames. UCHD, user-centred health-care design. TaCT, Telehealth & Care Technologies.

Drawing on internal and external knowledge ties

Table 7 compares ties with people inside and outside of the CLAHRC. At time 1, all of the CLAHRCs used an expansive information search strategy – networks were diverse and knowledge came in a variety of forms from multiple sources. In general, external ties were important for accessing new contacts and obtaining practical advice. CLAHRC members had previously collaborated with most of their external contacts and, therefore, brought these network resources into the CLAHRC. At time 1, members of the Greentown CLAHRC had the highest number of internal and external ties (average number of knowledge contacts per respondent).

By time 2, external knowledge ties became less important. Instead, individuals prefered to draw on knowledge from CLAHRC colleagues as the CLAHRC matured. There was a reduction in Greentown members’ number of internal and external KT ties. While the numbers of external KT ties also dropped for the Bluetown and Browntown CLAHRCs, these reductions were compensated by increases in the average number of internal KT ties as the CLAHRCs matured. At time 2, the Bluetown and Browntown CLAHRCs had adopted a more targeted information search strategy – internal ties with CLAHRC colleagues became most important. External ties that were maintained became closer at time 2.

As part of our survey of CLAHRC members, we asked respondents to indicate the institutional affiliation of their knowledge contacts. At time 1, based on number of ties, the NHS was the dominant institution for Browntown and Bluetown, whereas academia was the dominant institution for Greentown. At time 2, academia was the dominant institution for all CLAHRCs. At time 1, all CLAHRC knowledge networks included ties beyond the NHS and academia – to local authorities, central government, the private industry, third sector and service users. By time 2, Browntown had increased its number of ties to non-NHS/academic contacts, whereas Greentown and Bluetown had reduced theirs, which illustrates a greater focus on NHS–academic collaborations within these two CLAHRCs.

Types of knowledge resources provided by internal and external ties

We analysed how the KT networks provided different forms of knowledge resources to the CLAHRCs. The types of information sought out from internal ties within the CLAHRC at time 2 are relatively similar ( Figure 7 ) (for a more detailed analysis see Appendix 1 ). One exception is that of the Bluetown CLAHRC, which has a higher ratio of technical advice sought out internally. There are considerable differences in the external ties ( Figure 8 ). There is a much higher desire for advice on organisational backing and access to groups in Greentown than in the other CLAHRCs. In contrast, there is less desire for practical advice. The other notable difference is a greater search for management advice from external ties by the Bluetown CLAHRC compared with the other two.

FIGURE 7.

Knowledge resources provided by internal ties (percentage of CLAHRC total at time 2).

FIGURE 8.

Knowledge resources provided by external ties (percentage of CLAHRC total at time 2).

In Greentown, within-theme ties were used to obtain scientific advice, inside-CLAHRC ties mostly provided scientific advice, and outside-CLAHRC ties provided access to other groups, practical advice, and management advice. For Bluetown, within-theme ties were frequently used to obtain practical and technical advice, inside-CLAHRC ties were most often used to obtain management and technical advice, and outside-CLAHRC ties provided access to other groups. For CLAHRC Browntown at time 2, we found variance in the types of resources that were important to which themes. There is also variance in terms of where people within each theme go for resources. For example, those in Core Team, Obesity and Stroke sought practical advice from outside of the CLAHRC, whereas those in Depression and Knowledge into Action looked within their own theme for the same resource.

Epistemic boundary spanners

In this section we analyse instances where individuals act as brokers because they connect across professional disciplines in a KT network. This definition takes into account the attributes of brokers as well as their network position. (Non-participant actors have been removed from the original network because this analysis requires answers to survey questions about CLAHRC members’ discipline.) We use the attribute ‘professional discipline expertise’ to code CLAHRC actors into three epistemic groups: clinical practitioner (medicine, nursing, allied health), academic researcher (social sciences, pure sciences, medicine, health sciences) and business functions (management, commissioning and administration). Table 8 gives a breakdown of the composition of each group. The sociograms in Figure 9 represent KT networks between CLAHRC individuals, colour-coded by discipline.

FIGURE 9.

CLAHRC knowledge translation networks at time 2 (nodes coded by discipline). (Cases with missing attribute data were removed from the sociograms and excluded from the External-Internal Index analysis.) (a) Bluetown CLAHRC (102 nodes, 716 ties); (b) Greentown CLAHRC (101 nodes, 603 ties); and (c) Browntown CLAHRC (139 nodes, 1034 ties).

Using the External-Internal (E-I) Index routine118 in the UCINET software package, we measured the extent to which people with expertise in multiple disciplines have more heterophilous KT networks than individuals with single-discipline expertise. We found clear evidence of epistemic differences influencing KT networks. Scores are reported in Table 9 .

The results in the table indicate a tendency towards heterophily for individuals with multiple-discipline expertise, with more homophily found for individuals with single-discipline expertise. People with single-discipline expertise tend to translate knowledge to others with single-discipline expertise more than they do to those with multiple-discipline expertise. Conversely, people with more than one discipline specialism were found to translate knowledge to individuals with both single- and multiple-discipline expertise. We therefore suggest that individuals with multiple-discipline expertise are ‘epistemic boundary spanners’ by nature of their KT practices.

It makes prima facie sense that epistemic boundary spanners would have more diverse interactions than single-discipline specialists. In fact, these interaction preferences were related to social network structure. Epistemic boundary spanning was negatively correlated with ego-network density (r = –0.229, p = 0.00) and positively correlated with normalised brokerage (r = 0.295, p = 0.00). This meant that, in comparison with those with single-discipline expertise, CLAHRC members with multiple-discipline expertise had looser knowledge-sharing networks and often found themselves acting as brokers between people who were not themselves connected. Moreover, individuals with multiple-discipline expertise were also more highly engaged with the CLAHRC initiative than those with single-discipline expertise (r = 0.250, p = 0.00).

Implications of boundary spanning for knowledge translation policy and practice

Our findings highlight the different information search strategies undertaken by members of the KT initiatives and the social capital generated in support of boundary-spanning activities. Firstly, the knowledge networks at CLAHRC-theme level varied. Some themes were internally facing where team members mostly relied on theme colleagues for the translation of knowledge. Other themes were externally facing, with team members preferring to span theme boundaries in doing CLAHRC translation work. Team density will be intrinsically related to the nature of the work involved, the specific remit of each group and where the theme’s priorities lie on the research–practice spectrum. In our feedback to the CLAHRCs, we highlighted the network activities of themes and invited members to reflect on the network structure of each work group. Secondly, we also highlighted the change in information search from ‘expansive’ to ‘targeted’ strategies. At time 1, the CLAHRCs employed expansive information search strategies using diverse networks to draw on multiple sources of knowledge. By time 2, internal ties became more important and the search for information in the KT initiatives became much more targeted. The CLAHRC itself becomes a networked resource bank for KT as it matured over time.

Thirdly, there were also variances between CLAHRCs in terms of the use of internal and external ties, and where different types of knowledge resources were drawn from. KT networks were adapted as the CLAHRCs’ need for knowledge resources changed over time. Fourthly, the epistemic make-up of CLAHRC members was important for boundary spanning, particularly for the translation of knowledge across professional boundaries. We found that individuals with more than one disciplinary expertise had more diverse knowledge-sharing networks – translating to/from individuals with both single- and multiple-discipline expertise and, therefore, acting as ‘epistemic boundary spanners’. The results show that epistemic attributes and network position are related characteristics and that these factors influence preferences for in-/out-group KT across professional disciplines.

Governance co-ordination and control

In our analysis, we explored the extent to which the CLAHRC KT networks were centralised/decentralised and if a dominant ‘core’ set of actor could be identified. These factors would have direct implications for co-ordination and control of the networks and, thus, implications for the translation of knowledge.

Centralised knowledge networks feature hierarchy due to an asymmetric, or unequal, distribution of ties. Decentralised networks feature a more equitable distribution of knowledge ties, and so are more egalitarian in nature ( Box 14 ). Core actors are a set of dominant individuals in a network who frequently translate knowledge to each other, in contrast to peripheral actors who translate only to the core (on an unreciprocated basis) and not to each other ( Box 15 ). 119

The distribution of knowledge ties varied between CLAHRCs. As outlined in Table 10 , at time 1 Bluetown and Browntown had comparably more centralised (hierarchical) knowledge networks, in that many ties went to a small number of people in the centre of the network while there were fewer ties between peripheral individuals. In contrast, Greentown had the most decentralised network (where members had relatively equal access to and control over knowledge) at time 1. Refering back to Figure 3 , it is interesting to note that the structural location of reciprocal ties differs between CLAHRCs. Greentown’s decentralised network has reciprocal ties spread across the whole network, while reciprocal ties in CLAHRCs Browntown and Bluetown cluster at the centre of the network.

The network structures did not remain constant, and by time 2 the Greentown and Bluetown CLAHRCs had adjusted their levels of centralisation up/down and so became similar to each other. With the lowering of the Bluetown CLAHRC’s level of centralisation, Browntown became the most centralised CLAHRC at time 2.

At both time points, a set of individuals leading KT could be identified. Using a core–periphery algorithm,119 it was possible to ascertain how focused each CLAHRC was around a ‘core’ of individuals and who the core individuals in each network were. The composition of core actors in terms of their organisational role is important in identifying the professional types leading the knowledge network.

The results revealed similar core–periphery structures between CLAHRCs across the two time frames (a more detailed summary of the changes in network structures over time is provided in Appendix 1 ). For the Bluetown and Browntown CLAHRCs, senior CLAHRC members (directors or programme/theme leads) were found to be the most dominant core actors, suggesting some vertical control over KT relating to organisational hierarchy. Conversely, we found that the Greentown knowledge network ‘core’ was the least dominated by senior managers, reflecting its decentralised network structure. Here, core actors occupied a mix of organisational roles (predominantly theme leads, project managers and support staff) and this network diversity was suggestive of an organisational ethos of inclusiveness in a horizontally governed KT model. By time 2, the structure of Greentown’s network had transformed to be more aligned with that of the other two CLAHRCs, both structurally (more centralised) and compositionally (with more senior level CLAHRC members in the core).

Implications of network co-ordination and control for knowledge translation policy and practice

Although neither organisational form is inherently superior, there are strengths and weaknesses to each and accompanying impacts for the governance of KT models. Centralisation can enhance internal efficiency, co-ordination and dissemination of information and persuasive messages. 120–122 Networks that are centralised and/or have a core–periphery structure are easier to co-ordinate because the hierarchical structure supports the communication of directives, standardisation of practice and operational clarity. 123 Network centralisation, therefore, creates a backbone for the provision of certain elements of knowledge governance. However, the limitation is that some individuals will likely dominate the knowledge network and the skills and experiences of more peripheral actors may be relegated as a consequence, which will have serious implications for the translation of knowledge ( Box 16 ).

In the absence of control benefits, decentralised networks are actually more robust than centralised networks. This is because knowledge ties are more equitably shared across the initiative so the network is not as dependent on key (central) actors. This is arguably one of the reasons that the Greentown CLAHRC was able to avoid factioning (or even collapse) during several changes in leadership. The drawback is that decentralised networks, lacking a strong structure for internal co-ordination, may be susceptible to mixed messages that may result in role confusion124 [as our per our qualitative findings for the Greentown CLAHRC (see Chapter 4 )].

Maps analysis

Figure 15 provides an overview of the causal maps of the five initiatives. The arrangement of the constructs is by thematic area and the arrows represent a selection of the links shown in the collective maps generated with Cognizer^®. In fact, while the software provides all links that have emerged from the causal mapping exercise, for clarity reasons only the three top rated links per each initiative are reported (considering both the number of times the link was chosen and the ‘strength’ of each link).

FIGURE 15.

Comparison of collective maps.

Figure 15 highlights the following. Firstly, two-way arrows occurred in two out of three CLAHRCs while none occurred in either the Canada-Coordination or Canada-Translation. Although this result is not statistically significant due to the small sample, it does suggest that in the UK CLAHRCs there was more appreciation of the inter-relations between factors that influence health-care innovation initiatives, rather than seeing simple cause–effect relationships. Given the size and publicity associated with the overall CLAHRC initiative, it is possible that this has led those involved in health-care innovation projects in the UK to better appreciate the complexity associated with developing, mobilising and using knowledge in ways that improve practice. In the more isolated contexts of the two Canadian initiatives that we consider here, this complexity may be masked by the focus on getting the research done and applied in practice. It is also possible to relate one- versus two-way relationships between themes with the particular composition of health-care teams. For instance, the CLAHRCs teams were mainly formed by clinicians, that is, some doctors are team leaders; they act as managers and co-ordinators yet they have a clinician background. In contrast, in Canada, while a few of the managers are also doctors of medicine, others have a business background (this occurred in both the Canada-Coordination and Canada-Translation).

There are several implications from the comparative analysis based on the causal mapping exercise. Firstly, the pool of (28) constructs can be used in similar health-care initiatives to map the beliefs of those involved. We contend that the constructs identified here have relevance across settings, given the diversity included in our research.

Secondly, from a methodological perspective, the focus on aggregated indexes in collective maps provides information on whether or not a team/project’s beliefs are aligned. For instance, from the links density and their variance it emerges that beliefs among participants in some projects, for instance the Greentown CLAHRC and Canada-Translation, are more aligned than in others.

Finally, we showed that causal mapping techniques are able to identify relevant constructs from document and interview analysis (see Table 2 ); they can reveal perceptions of relevant characteristics of projects at the collective level (see Figure 15 ); and they can be effectively applied to health-care settings to provide relevant insights for practitioners.

Boundary spanning and knowledge translation

Our study suggests that it may be helpful to distinguish between different forms of boundary spanning in order to provide a greater understanding of how KT can take place across the boundaries of different groups within translational research initiatives. In considering the different types of mechanisms that have been used by project teams within our case study sites, we have identified that these can be categorised into two different types of boundary spanning – ‘bridging’ and ‘blurring’. With ‘bridging’ mechanisms to span boundaries, an intermediary (a person, event or object) acts as a facilitator for the translation of knowledge between one setting and another. In particular, there is frequent use of mechanisms explicitly designed to facilitate knowledge flow between different groups.

The ‘bridging’ approach was exemplified by Bluetown CLAHRC, where, as noted, themes were led by university-based clinical-academics, who typically held honorary roles within local health-care organisations. Although team members such as these generally had prior experience as health-care professionals, within the context of the CLAHRC, these individuals constructed their identity as belonging to a clinical–academic community. The team leader was influential in directing the subprogrammes of work, resulting in the traditions of a clinical–academic environment being the dominant ideology within the clinical project teams. However, at CLAHRC level, collaboration between clinical–academics and those from other academic disciplines, and between the project teams and local health-care providers and commissioners, was also strongly encouraged as part of the process of making research evidence more useful for health-care policy and practice.

Structural features within this model had been designed with the explicit purpose of linking the clinical project teams with specialist types of academic expertise (such as economics, statistical support and KT and dissemination skills). It was clear from our study that these individuals were required to be flexible in the approaches that they followed for doing their work. One participant from a specialist support service team described how she felt that the clinical team that she had been allocated to work with did not initially understand the type of perspective her expertise could provide, or how her role could usefully contribute to the programme of work that had already been planned within their theme. In working through this challenge, it was important that the individual was willing to be flexible in the type of working traditions that were followed. Although the initiative-wide strategy had initially aspired to the greater ‘blurring’ of knowledge from different academic perspectives, the challenges experienced meant that a ‘bridging’ approach for KT was more widely applied in practice.

In contrast to the Bluetown CLAHRC, our study within Browntown CLAHRC highlighted how boundaries were ‘blurred’. In this case, project teams were drawn from a more diverse set of groups, representing different disciplines and professional groups. However, rather than the more polarised differences seen in Bluetown, there was a greater degree of overlap or ‘redundancy’ between the types of disciplines and professional expertise represented in teams. For example, many of the academics involved were from professional sciences rather than disciplinary sciences, and so were closer to the epistemic concerns of team members who were health-care practitioners. 129 Our data showed that that there was a willingness to incorporate insight from team members from quite different perspectives, and that these ideas could be more fluidly translated into shared practices within the teams. As a result, the outputs from the work programme could be more easily understood within a health-care policy and practice context, without a reliance on a ‘bridging’ type mechanism to facilitate translation.

The contrast between the bridging and blurring approaches identified in our study is summarised in Table 12 .

Bridging versus blurring as approaches to boundary spanning

As Table 12 suggests, ‘bridging’ emerges from and sustains an environment in which little compromise or negotiation is required by collaborating groups around their epistemic differences. Instead, the emphasis is on the mechanism of the bridge as a translational process, in order for knowledge to be able to be understood in a different context. In ‘bridging’ approaches, the emphasis is on an intermediary enacting the translation process themselves. They perform roles, such as ‘translator’ and ‘interpreter’ of knowledge, between the environment of the clinical project team and their ‘home’ context. In this case, the emphasis is on the boundary-spanning mechanisms to possess the requisite capability, such as an individual adapting their own comprehension of different situations, or a meeting space which has an appropriate environment to stimulate two-way discussion.

With ‘bridging’ mechanisms, the project team themselves are not required to radically alter how they work. As a result, they can develop knowledge based on the traditions and norms of their own community. As has been previously highlighted,130 an advantage of mechanisms that focus on trusted intermediaries enacting knowledge-brokerage roles is that there is less pressure on researchers to develop a new set of skills for KT, as instead they rely on the broker to carry out this work. However, as power balances and legitimacy battles between epistemic groups influence KT and mobilisation,28 facilitation by the senior management of the CLAHRC can be important in creating an environment of openness to engage and interact with the ‘bridging’ mechanism.

Viewed positively, ‘bridging’ approaches may have the advantage that university-based teams are able to produce high-quality publications, while still meeting CLAHRC objectives on informing health-care policy and practice. In this sense, ‘bridging’ mechanisms can be useful as they provide a supportive structure for groups and communities with markedly different practices and forms of expertise. However, the extent to which this knowledge can be transformed into novel forms can be limited, as it is reliant on what can be achieved by the bridging mechanisms.

Our analysis of ‘blurring’ forms of boundary spanning illustrated an environment where knowledge could be more readily mobilised across organisational, discipline and professional boundaries. Where the epistemic differences across these groups are not great, knowledge can be co-produced through subtle transformation, to incorporate ideas and expertise from different perspectives. In this sense, ‘blurring’ forms of boundary spanning have the potential to develop ‘novel’ types of knowledge that are not embedded in traditionally established specialist domains. Although explicit mechanisms were not used to support this type of boundary spanning, implicit organisational features and processes were important in supporting KT through this boundary-spanning approach. Particularly important here were ‘hybrid’ individuals with dual positions who could act as boundary spanners, and also organisational processes that enabled members to take on multiple positions within the initiative. By creating teams that span a variety of diverse groups, this approach has the potential to develop an innovative programme of work, creating capacity for innovation through the co-production of knowledge at the interstices of many different perspectives. 34 For ‘blurring’ mechanisms, ‘figureheads’ did not have the same importance as has been shown for ‘bridging’ mechanisms,79 as there was a greater onus on all members of the initiative to be flexible in their work practices and epistemic commitments. At the same time, support from CLAHRC senior management was crucial to creating an environment where the perspectives of diverse groups were respected and were enabled to inform the direction and focus of the work programme.

The importance of ambidextrous networks

In advancing the networked perspective on innovation, scholars have drawn on theorists of social networks and social capital, notably Burt and Coleman, to better understand the ways in which knowledge is created and shared through social networks. 131,132 This work has emphasised the related structural features of ‘brokerage’ and ‘closure’ as playing vital roles in generating different forms of social capital which contribute to the innovation process.

The extent to which individuals are connected to others who are not themselves connected is known as brokerage within the SNA literature. Brokerage thus describes the potential for actors located at structural holes (in disconnected groups) to connect and exchange non-redundant (new) information. 62,131 From the standpoint of innovation, it represents an important mechanism through which new knowledge can be created via novel combinations of existing knowledge.

If measures of individual brokerage are aggregated, it is possible to find the extent to which a governance network contains these phenomena. In our study, we find that the CLAHRCs had somewhat similar levels of brokerage at time 1. The exception was the Greentown CLAHRC, which had slightly more people acting as knowledge brokers between disconnected others in the network. By time 2, we see very similar brokerage scores for each of the CLAHRCs, indicating convergence in network structure. The level of brokerage within each of the CLAHRCs was conducive to the fusing together of novel information sources through individuals brokering across the loosely connected structure of the network between the themes.

By way of contrast, closure occurs in cohesive groups and describes situations where subsets of actors in a network are structurally positioned so that they have a tendency to be closed off from others. In an organisational setting, this may be akin to an established work team of colleagues with shared interests, motivations and goals or a community of practice. Closure is linked to density, reciprocity and trust,101 structural conditions that facilitate action, and ‘getting things done’ in practice. It thus helps to account for the role of networks in ‘embedding’ new knowledge.

The CLAHRC networks were loosely knit structures, especially between the themes. Within the themes, however, there was a much higher level of closure. We argue that this has a prima facie logic because social processes occurring at the local group level are likely to develop closure over brokerage. Each CLAHRC thus acted as a space to bring together various individuals and clusters around different themes under a common KT vision. We observed slightly different levels of closure between CLAHRCs at time 1. In the first time period, the Greentown KT network had a higher level of closure than that of the other CLAHRCs. In network analyses terms, we found that the Greentown CLAHRC had higher density and reciprocity and lower average distance between contacts. At time 2, the CLAHRCs became more similar in terms of their levels of closure.

It is important to recognise, however, that brokerage and closure only represent local network conditions; that is, phenomena which emerge within one part of a wider network. Existing studies have tended to focus on these local conditions as discrete phenomena133 – focusing, in particular, on knowledge creation via brokerage – and have not addressed their implications for the innovation process as a whole. A process view of innovation defines it as a temporally and episodically structured, highly iterative design and decision process involving the creation, diffusion, blending and implementation of new ideas and knowledge at different stages. 134 From this standpoint, innovation is seen as involving not only brokerage across structural holes to facilitate idea generation, but closure within cohesive groups to implement and embed these ideas. It follows that addressing the role of networks in the innovation process involves developing a better theoretical understanding of how they might simultaneously support brokerage and closure conditions.

In simple terms, the need for both brokerage and closure can be understood in terms of their relative costs and benefits. Previous research indicates that there are costs to having social network ties in both time and energy. Therefore, a structure whereby everyone is connected to each other is not viable. Instead, it is preferable to have small clusters of individuals with high levels of network connectivity that promotes embedding knowledge into practice. Clusters working in isolation, however, limit the transfer of novel information. Therefore, it is also important for there to be brokering ties that connect clusters as well as there being ties to other people outside of an organisation or entity. We found that the CLAHRCs were successful in building ambidextrous networks encompassing areas within the network that had high levels of closure (well-connected KT network structures suited to embedding knowledge into practice) and loose network structures containing ‘structural holes’ that supported the fusing together of novel information sources through individuals brokering across the structural holes. This network ambidexterity was maintained at both time points.

The importance of this ‘ambidextrous’ character in social networks is particularly relevant to KT networks. In this respect, the discussion here helps to draw out the implications of our findings on the CLAHRCs for wider theorising and practical intervention in the health-care domain. Thus, an important finding of our study was that each of the CLAHRCs had built KT networks with the capacity for both brokerage (ties conducive to generating new, innovative knowledge) and closure (tie structures suited to capturing knowledge to implement or embed into practice). The CLAHRC networks were able to achieve such ambidexterity, moreover, despite manifesting significant differences in their structural forms in terms of their degrees of hierarchy and centralisation.

Development of networks over time

One important contribution of our study to our theorising of the role of networks in KT is the extent to which our study was able to identify changes in the CLAHRCs’ networks over time. Research to date has tended to view networks primarily in structural terms, as channels, conduits or ‘pipelines’ through which knowledge is transferred. 135 The more limited research available, focusing on networking processes, crucially establishes that networks are not static, but are dynamic and evolving as circumstances unfold. 136 Yet relatively little attention has been paid to understanding and explaining the social and political processes underpinning the formation and effectiveness of network relationships specifically as they relate to innovation. Nor is there sufficient understanding of the ways in which processes of networking relate to the formation, and transformation, of new network structures across institutional, scientific and professional boundaries – structures which may, in turn, enable or constrain further processes of innovation. 137

In this respect, our findings are significant in demonstrating the scope and implications of change in networks over time. In particular, we observed a significant pattern of convergence – that the CLAHRC KT networks became more similar to each other over time. We also found that the CLAHRCs became more self-reliant and self-referential as social networks. Thus, in our initial study, we found that the CLAHRCs had adopted ‘expansive’ information search strategies. CLAHRC members had previously collaborated with most of their external contacts and external ties were important for access to new contacts and practical advice. In our second-phase work, we found that all three CLAHRCs had more ‘targeted’ information search strategies. External knowledge ties had become less important, as members drew on knowledge from CLAHRC colleagues.

Institutional factors

There appeared to be a number of institutional factors that affected the management and governance of the initiatives and that helped to account for the differences between the North American initiatives and the CLAHRCs in terms of their development of a more emergent (vs. directive) relational capability and in terms of their development of a more blurred, co-production (vs. research- or practitioner-led) approach to the development of integrative capabilities. First, the longer-term and more open-ended funding for the CLAHRCs meant that they could be more receptive to emergent forms of networking (encouraging networking between organisations with no clear idea as to what this would lead to) rather than being narrowly focused on networking to secure resources to meet a pre-defined problem. Second, because CLAHRCs were required to provide matched funding, they had a vested interest in building some long-term relationships between the organisations involved. Finally, the CLAHRCs had a more specific remit to promote co-production with stakeholder groups.

Greentown Collaboration for Leadership in Applied Health Research and Care

The most internally connected themes in CLAHRC Greentown at time 1 were stroke, CYP and primary care (79.4%, 67.5% and 66.2% of within-theme ties, repectively). At time 2, the Greentown network saw an increase in the proportion of cross-theme ties, with stroke, CYP and primary care becoming slightly less internally focused.

Bluetown Collaboration for Leadership in Applied Health Research and Care

At time 1, the most internally connected themes in the CLAHRC Bluetown network were stroke and TIA, pyschosis and maternity support (percentage of ties within themes: 87.0%, 81.8% and 69.0%, respectively). At time 2, the Bluetown network saw an increase in the proportion of cross-theme ties. Psychosis, maternity support, and stroke became less internally focused. The cross-cutting themes maintained a high proportion of cross-theme brokering ties at both time frames (82.4% at time 1, 73.5% at time 2). The Core team became less internally focused at time 2.

Browntown Collaboration for Leadership in Applied Health Research and Care

The Browntown network has a similar proportion of cross-theme ties in each time period (44.4% at time 1 and 44.7% at time 2). At time 1, the themes with the highest proportion of in-theme ties were diabetes (77.1%), depression (73.9%) and obesity (67.9%). At time 2, genetics and inequalities became less internally focused. Core team and TaCT became more internally focused. (Note: some theme data are missing.)

Browntown Collaboration for Leadership in Applied Health Research and Care

At time 1, 32% of ties are outside the CLAHRC; this decreases to 23% in time 2 ( Tables 16 and 17 ). People tend to go outside the CLAHRC to gain access to groups and individuals. This is especially the case at time 1. In time 2, we see an increase in the number of ties outside the CLAHRC providing practical advice. At time 1 people have a relatively high percentage of ties outside the CLAHRC in which they had previously collaborated. This drops in time 2. This highlights the switch from external to internal ties as the CLAHRC matures.

At time 1, ties outside of the CLAHRC were, on average, not very close, whereas by time 2 the outside-CLAHRC ties that remained were considered much closer. Not surprisingly, ties to people within the CLAHRC were more frequent than those to people outside the CLAHRC.

Greentown Collaboration for Leadership in Applied Health Research and Care

At time 1, 27% of ties are outside the CLAHRC; this decreases very slightly to 26% in time 2. At time 1, people tended to go outside the CLAHRC to gain access to groups and individuals and for practical advice. By time 2, there was much less searching for practical advice outside the CLAHRC.

At time 1, people have a relatively high percentage of ties outside the CLAHRC in which they had previously collaborated. This drops in time 2. This highlights the switch from external to internal ties as the CLAHRC matures.

At time 1, ties outside of the CLAHRC were, on average, not very close, whereas by time 2 the outside-CLAHRC ties that remained were considered much closer. Not surprisingly, ties to people within the CLAHRC were more frequent than those to people outside the CLAHRC.

Bluetown Collaboration for Leadership in Applied Health Research and Care

At time 1, 27% of ties are outside the CLAHRC; this decreases to 19% at time 2. At time 1, people tended to go outside the CLAHRC to gain access to groups and individuals and for practical advice. In time 2, this remained the same.

At time 1, people have a relatively high percentage of ties outside the CLAHRC in which they had previously collaborated. This drops at time 2. This highlights the switch from external to internal ties as the CLAHRC matures.

At time 1, ties outside of the CLAHRC were, on average, not very close, whereas by time 2 there was an increase in the ratio of outside-CLAHRC ties that were especially close. Not surprisingly, ties to people within the CLAHRC were more frequent than those to people outside the CLAHRC.

The Canada-Coordination pilot project, Ottawa

This map comprises 10 participants; the most relevant input constructs that were selected by the participants during the initial survey (see Chapter 3 , Research methods) are:

• collaboration between partner organisations in the project

• identification of barriers to service change and redesign

• involvement of health-care and user representatives

• involvement of experts from multidisciplinary backgrounds

• conduct research that is focused on patient needs.

The most selected input constructs – that is, the constructs that are the most important drivers to achieve the objective of the pilot project – are:

• conduct research that is focused on patient needs

• involvement of experts from multidisciplinary background

• conducting ongoing review and evaluation of the project

• identify barriers to service change and redesign

• collaboration between partner organisations in the project.

In the pilot project map, again, there is (partial) overlap between the constructs that have been chosen as relevant and the constructs that were most associated with the achievement of the outcomes of the initiative. In particular, the pilot project collective map indicates that:

• Conducting research that focuses on patient needs leads to quality improvement.

• Involving experts from multidisciplinary backgrounds leads to building capacity.

• Collaboration within the agencies involved in the initiative leads to building capacity and improving efficiency.

• There are two-way relationships between building capacity and identifying barriers to service change and redesign and between improving efficiency and identifying barriers to service change and redesign.

• Conducting an ongoing review of the project helps improving the quality of health-care service delivery.

Figure 23 illustrates the relationships above described.

FIGURE 23.

Canada-Coordination collective map (10 participants).

In terms of the quantitative approach, Table 27 illustrates the main statistics of the pilot project and compares this with the mean of all initiatives.

Table 27 indicates that the participants of the pilot project have selected a relatively low number of links (and strengths). This is shown by the first three indexes, which are all below the mean of all initiatives. Moreover, the link density (and the link strength density) is low. The overall map density is ‘average’ if compared with the other initiatives.

The statistical analysis can be interpreted as follows: (1) the equivalence between the mean links strength and the mean ABS links strength shows that almost no negative relationships were selected; (2) the first two indexes of the table show also that there are relatively weak links between the constructs (low mean link strength and low mean ABS); (3) relatively low standard deviation (if compared with the mean of the standard deviation for initiatives) indicates that the range of the values (1, 2 or 3) for the relationships of the construct is less wide than the mean of all initiatives (more 1 and 2 rather than 3); (4) link density, link strength density and map density are lower than average (compared with the other maps). This suggests that few associations were chosen and most of the associations that were chosen had low values (i.e. weak relationship between inputs and outputs).

Canada-Translation

This map comprises 11 participants; the most relevant input constructs that were selected by the users during the initial survey (see Chapter 3, Research methods) are:

• conduct research that is focused on patient needs

• involvement of experts from multidisciplinary background

• collaboration between researchers and practitioners

• conduct research that considers local priorities

• involvement of health-care and user representatives.

The most selected input constructs – that is, these construct that are the most important drivers to achieve the objective of the pilot project – are:

• conduct research that is focused on patient needs

• involvement of experts from multidisciplinary backgrounds

• collaboration between researchers and practitioners.

In the Canada-Translation map, there is complete overlap between the constructs that have been chosen as relevant and the constructs that were most associated with the achievement of the outcomes of the initiative. In particular, the Canada-Translation collective map indicates that:

• Conducting research that focuses on patient needs is a central input; in fact, it leads to reducing inequalities, improving the quality, improving efficiency, and building capacity.

• Involving experts promotes collaboration between researchers and practitioners and collaboration between researchers and practitioners helps to improve efficiency.

• There is a two-way relationship between building capacity and enabling collaboration between researchers and practitioners.

Figure 24 illustrates the relationships described above.

FIGURE 24.

The Canada-Translation collective map (11 participants).

In terms of the quantitative approach, Table 28 illustrates the main statistics of Canada-Translation and compares this with the mean of all initiatives.

Table 28 indicates that the participants of Canada-Translation have selected a relatively low number of links (and strengths). This is shown by the first three indexes which are all below the mean of all initiatives. Moreover, the link density (and the link strength density) is low. The overall map density is the lowest if compared with the other initiatives.

The statistical analysis can be interpreted as follows: (1) the similarity between the mean links strength and the mean ABS links strength shows that only few negative relationships were selected; (2) the first two indexes of the table show also that there are relatively weak links between the constructs (low mean link strength, and low mean ABS); (3) low standard deviation (if compared with the mean of the standard deviation for initiatives) indicates that the range of the values (1, 2 or 3) for the relationships of the construct is less wide than the mean of all initiatives (more 1 and 2 rather than 3); (4) with respect to the other maps, the link density, the link strength density and map density are very low. This suggests that fewer than average associations were chosen and most of the associations that were chosen had low values (weak and medium relationships between inputs and outputs).