Background
Internationally, clinical networks have emerged over the last ten years as an important clinician-driven innovation for attaining system-wide improvements in healthcare delivery and patient outcomes [
1]. Commonly, the remit of clinical networks is to identify how and where improvements in health service delivery and patient outcomes can be made in the particular specialty represented by the network and to implement changes in association with key government health organisations. Important features of clinical networks are that they can provide a structure for liaising across institutions, allowing greater clinical input into models of service delivery [
2,
3]; provide 'bottom up' views on the best ways of tackling complex healthcare problems[
4] and are usually multidisciplinary involving doctors, nurses, allied health professionals, scientists, managers, and consumers [
1]. Clinical networks can also provide an organisational vehicle for embedding evidence-based care into health systems and engaging clinicians to change practice in line with evidence [
3,
5,
6].
The term clinical network has been used to describe many variants of networks ranging from fully integrated service delivery systems to informal communities of practice [
7]. In this paper the term refers to networks of voluntary clinicians[
8] that aim to improve clinical care and service delivery using a collegial approach to agree and implement a range of strategies. Unlike managed networks, they do not have a formal service delivery function involving the organisation and co-ordination of all clinical services that the patient needs through primary, secondary and tertiary care [
9]. In Australia, in 2004, the Agency for Clinical Innovation (then the greater Metropolitan Clinical Taskforce, GMCT) led the development and establishment of voluntary collegial clinical networks in New South Wales. The purpose of these networks is to provide opportunities for clinicians from a range of healthcare professions, and consumers to participate in the design and implementation of models of care and clinical plans; to develop and implement a range of projects to foster evidence-based practice and to provide clinicians with a structure that facilitates research and service delivery across local health care boundaries. These networks generally meet after hours and strongly rely on voluntary involvement of clinicians and consumers and have support from a Network Manager to assist with the co-ordination of network projects and network administration. The ACI clinical networks are commonly chaired by clinicians or consumers and are multidisciplinary in focus, involving doctors, nurses, allied health professionals, managers, and consumers. The remit of ACI networks is to identify how and where improvements can be made in the particular specialty represented by the network and to implement changes in association with key government health organisations in order to improve health service delivery and patient outcomes. Drawing on these experiences, clinical networks with similar purposes are currently being implemented in other Australian states. Examples of network initiatives include provision of education for implementation and workforce initiatives; development of policies, clinical guidelines and consumer resources; implementing processes for monitoring safety and quality.
Studies from a number of countries have shown that while there is some empirical evidence of positive impact on processes of care,[
10‐
12] there is uncertainty around what constitutes 'successful' network outcomes and what the factors, or conditions, that are required to help networks achieve their outcomes [
13]. While there are studies that have investigated the factors that affect the establishment of voluntary clinical networks,[
7,
8,
14] there are no studies that have researched what those connected to networks think is required to establish a successful network. Because clinical networks operate within a complex political, cultural and organisational context, research to identify conditions for successful networks from the perspective of those involved in networks can provide useful data for organisations wishing to establish future networks and maximise success in achieving outcomes.
In addition, there is limited empirical evidence demonstrating the benefits of networks to healthcare consumers and health services [
13]. This field of research could benefit from stronger evaluative designs and knowledge of which outcomes common to a range of networks (ie other than disease-specific patient health outcomes) should be measured. Identifying outcomes that are regarded as important by those connected to networks and policy-makers, and that could inform the measurement of outcomes across networks with different clinical foci would benefit this field of research. Taking account of stakeholder perspectives helps to ensure that a future set of outcome measures has utility and content validity [
15]. There are no studies that have investigated the views of those who work in or who are connected with networks about desirable outcomes. Clearly articulating the desired outcomes of networks can provide insights into ways network effectiveness can be measured.
The aims of this study were to identify key stakeholders' views on the conditions required to establish successful and effective clinical networks and also their views on desirable outcomes of successful networks.
Methods
Research Design
To answer the research questions of a) what do stakeholders identify as necessary conditions for establishing successful clinical networks and b) what do stakeholders identify as outcomes of successful clinical networks, a qualitative research design was chosen in order to obtain a full range of authentic participant views and for distilling key themes in this under-researched area [
16,
17].
Participants
A purposive maximum variation sampling approach was used to recruit participants from each of four groups, that is who were either directly involved as a participating clinician or involved directly or indirectly at a policy or strategic level. More detail about the definition of the four stakeholder groups: network drivers, network participants, Senior health service managers and Senior policy-makers, is provided in the list below:
Network drivers: individuals with detailed knowledge about the activities of at least one established clinical network, specifically network managers who had been in position at a network for a minimum of two concurrent years and current and past Network Co-chairs and Governing Committee members.
Network participants: participants in network activities and meetings as defined by records of attendance (medical staff, allied health professionals, nurses, and consumers) involved in networks that had been established for a minimum of two years.
Senior health service managers: senior managers at executive management level who held a clinical operations or clinical governance role at a hospital or from one of the eight (as there were at the time the study was conducted) metropolitan Area Health Services (AHS) that had clinicians involved in networks. Individuals in this group included senior local health service executives such as the general manager, clinical director, chief executive and director of clinical operations.
Senior policy-makers: executive level policy makers that worked in organisations with a relationship to the Clinical Network Executive (that is, network administrative body). Examples of these organisations are those with a statewide strategic role in: health policy and planning; quality improvement, patient safety initiatives; developing clinical guidelines and protocols and professional development.
Those in the network driver and network participant groups had to have been associated with a network that had been established for a minimum of 2 years in order to capture informed perspectives from those with at least some experience of working in or establishing networks. In addition to the criteria outlined above, eligible individuals were those aged 18 years of age or over and able to give informed consent.
We recruited participants from each of the four groups by asking the Clinical Network Executive Office (the main administrative office of the then GMCT), to identify those who met the eligibility criteria. Once selected, all potential participants were sent an advanced letter that explained the research aims and informed them that they met the eligibility criteria and that a researcher may later contact them to see if they are agreeable to participate. A phone call was then made one week after to four individuals in each participant group who were selected from the list (not in alphabetical order) provided by the Clinical Network Executive Office and appointments made for the interview. The initial number of individuals was selected to ensure that a range of views was captured. Additional participants were selected one at a time to ensure maximum variation. All individuals contacted consented to participate. The final sample size of 27 was reached by saturation of themes, that is, no new insights were identified in the data. Nine participants were from the network driver group; six were network participants; four were senior health service managers and eight were senior policy-makers. Those from the network driver and network participants groups (called 'network groups' below) were associated with 10 of the 20 networks established at the time the research was conducted.
Data collection
Individual semi-structured face-to-face interviews using a topic guide were conducted at the interviewee's place of work. The interviews were of 30-60 minutes duration and were audio-recorded. Participants were asked for their views on what they thought were the most important things that needed to be in place (that is, conditions) for clinical networks to be successful and to achieve positive outcomes. To elicit views on 'conditions' participants were prompted to nominate facilitators, or 'things that need to be in place' for successful networks to be established. Participants were also asked for their views on what they thought were desirable outcomes resulting from network initiatives or projects. In this context, the term 'clinical networks' referred to networks of clinicians that aim to improve clinical care and service delivery using a collegial approach to agree on and implement a range of strategies. Participants were asked to give their general views rather than to limit their responses to their direct experiences with any particular clinical network(s).
Data analysis
Interviews were anonymised and transcribed verbatim to produce transcripts of narrative text for thematic analysis. The factors of interest, namely conditions and outcomes, provided a framework for the initial categorisation of text. Thematic analysis with open coding was performed whereby each segment of interview text that related to the factors of interest was coded as a provisional theme. These codes were descriptive and linked with representative examples from the original text and the same code was assigned to data that represented similar themes. The process of generating themes was inductive. Themes were identified, coded, recoded and classified by examining regularities, convergences and divergences in the data. The themes and sub-themes derived reflect the language used by the participants. Refinement of our analysis led to both themes and sub-themes, the latter being conceptually linked to the main themes. We also highlighted differences and convergences in responses from stakeholder groups. For 'conditions' the resulting codes were organised into five sub-themes and on the basis of these, an overarching theme was identified: Effective Network Structure, Organisation and Governance. For 'outcomes' the resulting codes were organised into four sub-themes and two overarching themes were identified: Connecting and Engaging and Changing the Landscape of Care. Participant narratives have been used to illustrate meaning in the themes and summaries.
In order to ensure the quality of results, two researchers conferred on the analysis separately coding text and generating themes and sub-themes and then comparing output. In addition, the results were circulated to participants to check that the findings had fidelity with their perceptions and experiences. Following this, minor adjustments were made. There was good concordance between researchers in the analysis of jointly reviewed transcripts and validation by participants did not show disagreement with the analysis.
Ethics approval
Ethics approval was gained from the Human Research Ethics Committee of the Australian Catholic University. All participants were informed of the objective of the study and that they were free to participate or withdraw from the study at any point. Participants gave written consent to be interviewed and for the interviews to be audio-recorded. Recordings and transcripts were coded so that the origin of each one could not be identified.
Discussion
Through thematic analysis, this study has identified stakeholder views on the important conditions for establishing well-functioning and successful clinical networks within the broader health system and also captured what stakeholders think are desirable network outcomes. Our study provides insight into the views of those with strategic and health policy and planning responsibilities and those who work within or are involved in clinical networks. Although mainly broadly supportive of network initiatives, the views of those in the Senior policy-maker and health service manager groups were not as uniformly positive as those in the network groups. Concerns were mainly focused on the need for network workplans to be aligned with state health and local health services priorities and for there to be a formalisation of alliance between the networks and local health service managers in order for networks to improve clinical outcomes. Those in the network groups were mainly concerned about bureaucratic obstacles to progressing their initiatives and also with practical considerations around availability of resources and not being able to implement innovations.
However, there was a broadly shared vision across all stakeholder groups that effective structure, organisation and governance be in place for networks to be well-established and effective in achieving health system outcomes. Lack of formal governance, structure and organisational focus was widely thought to result in poorly organised and functioning networks with limited ability to realise their objectives and liaise strategically with within the broader strategic health arena. The necessary conditions to achieve this is represented by the sub-themes of building relationships; leadership; strategic evidence-based workplans; adequate resources; and ability to implement and evaluate network initiatives would help.
An important condition was the development of an inclusive and collegial network ethos with a strong focus on working for healthcare consumers. The importance for network development of involving a critical mass of clinicians in health care service planning and delivery and forming interdisciplinary and interorganisational collaborations for achieving health outcomes and practice change has been identified as important in other contexts [
8,
10,
12,
18]. Other critical ingredients for network success was strong multi-level leadership to engage clinicians, drive change and to implement clearly articulated evidence-based workplans. Other studies have highlighted the importance of leadership in order to effect positive change in health systems [
1,
12,
14].
There was concordance amongst stakeholders that links with state health agencies and local health services were essential for establishing successful networks. However, tensions in relationships between some networks and health regions and health agencies was felt to constrain some network activities, although different reasons for this were nominated by the different stakeholder groups. For participants from network groups, there was concern that area health service boundary issues were a substantial barrier to realising the benefits of network innovations across a broad geographical area. While for those in the Senior health service manager group, the stated issue was that health managers were not formally involved in networks. Participants from all groups felt that these relationships would develop positively over time as networks became more established and more embedded within the health system. The feedback about the desirability of aligning network workplans with state health priorities on the one hand and objectives that may engage clinicians on the other hand, suggest that a balance needs to be struck to ensure projects are of interest to clinicians but also that resources are directed to innovations that will have the most impact on patient outcomes. In the early stages of network development, some quick gains from simple projects may be motivating and serve to keep clinicians on board for more complex projects.
Our findings converge with those in previous studies suggesting that there are a core set of preconditons that may be required by networks in different settings to ensure that they can deliver successful projects. Touati et al (2006) found a shared philosophy and vision to be important factors for success[
12] and Nies et al. (2003) found that the main obstacles to network success were: competing interests and priorities; overlapping catchment areas and complexity due to the involvement of multiple levels of government [
19]. A cross-sectional multiple case study of six managed clinical networks found that the three major determinants of successful networks were: professional dedication of network staff; legitimacy of the network and confidence of the staff and organisations involved [
7].
The present study also reports what participants identified as desirable outcomes of successful networks. The two main themes, connecting and engaging and changing the landscape of care, reflected two broad categories of both process and 'hard' outcomes. The nominated outcomes of interprofessional collaboration and strategic external partnerships were frequently cited by those in the network groups as being as highly valued as achieving clinical outcomes. Interestingly, this reflected the perspective of both those who were associated with networks that had been relatively newly established, and those whose network had been operating for a number of years. The value of these 'relationship' outcomes is that they provide a foundation for achieving health service delivery and patient outcomes. That this type of outcome was regarded as highly desirable and was also nominated as a condition ('building relationships') indicates the strength of its importance to the formation and development of well-organised networks. However, this does not necessarily indicate the relative importance of this condition/outcome compared to others, as participants were asked to nominate those conditions and desirable outcomes that they felt were important, not to rank them.
The theme 'changing the landscape of care' reflected strong views about the value of networks in improving health services, patient outcomes and also the role of networks in developing and retaining an engaged clinical workforce through fellowships; post-graduate education, scholarships and conferences. However, some Senior health service managers and policy-makers expressed concerns about whether networks are well-placed to achieve some of the more ambitious outcomes, such as addressing clinical variation. However, the reasons for gaps between evidence and practice are complex requiring complex solutions [
20] so the type of project required to achieve outcomes related to addressing clinical variation may be overly ambitious for networks to address in the early stages of the development of a network and/or without support from experienced researchers and other experts.
Related to this, it is noteworthy that in the course of the interviews, some participants, mainly from the network groups, expressed doubts about potential difficulties in applying a standardised formula of measuring success and outcomes across all networks. This suggests that while it may be possible to develop a core set of outcomes to which networks could aspire to, this suggests that network-specific outcomes which reflect networks different clinical foci and priorities, differing time periods over which networks have been established and available staffing and other resources are also needed. This is supported by literature that recommends that comparative studies of health service change strategies, such as clinical networks, should use a range of measures of success and impact to enable a more complete assessment of effectiveness [
21,
22].
Accessing a range of views on a relatively new initiative is useful for those with responsibilities for establishing and working with networks as it can identify both issues and possible solutions. The findings from this study point to the importance of ensuring that favourable conditions are in place to maximize the effectiveness of networks in achieving outcomes. Participants views indicated that being able to succeed in achieving network objectives and therefore achieving both process and hard outcomes is partly conditional on the right factors being in place in order to establish a solid foundation for networks to design evidence-based projects. Future research is warranted on whether networks with clearly defined and formal structure, organisation and governance compared to more informal network arrangements are more likely to achieve their outcomes.
Strengths and limitations
The strengths of our study include the maximum variation sampling strategy that ensured that multiple perspectives were captured through in-depth interviews of informants from four groups related to clinical networks. The sample included both those directly involved in networks and those who were knowledegable about, but not directly involved in the day to day work of networks. Stakeholders were asked to give their general views; however many chose to contextualise their responses with reference to networks with which they had direct experience. The semi-structured interview technique allowed issues to be explored in a flexible manner. In keeping with the research method, interviewees were free to raise any issue that they felt were germane to the topic under investigation. As a result, it is believed that the information gathered was reflective of genuine concerns and views. The main limitation is that the sampled participants (network participants and drivers) came from 10 out of 20 networks. However, within these 10 there was good representation in terms of range of years that they had been operating, that is from two years upwards and also in terms of the variability of clinical area represented. In addition, it may be that the interviewees expressed publicly acceptable viewpoints. However, the interviews were anonymised and confidential in line with ethics requirements. In terms of transferability of results, clinical networks in NSW, indeed each network, may have their own unique culture and political environment that influence the responses of participants' directly involved in networks. These findings may therefore be most relevant for networks that have a similar model to that described here.
Conclusions
This study has provided new knowledge on what key stakeholders believe are important conditions for successful networks and valuable outcomes of networks. The findings suggest that a systematic evaluation of barriers and facilitators prior to the establishment of networks should be undertaken to ensure that favourable conditions are in place to maximize the effectiveness of networks. The findings also suggest that tools assessing outcomes of clinical networks in future evaluative studies should be multi-dimensional, covering health care outcomes and process outcomes such as building relationships and interprofessional collaboration.
Importantly, stakeholders held the view that effective clinical networks could realise significant benefits for healthcare systems and patient care as long as the right conditions were in place. This makes networks well-placed to work collaboratively alongside health authorities to deliver population health goals. These findings are likely to be generally applicable to other similarly organised networks and studies of those networks.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors contributed to the study's conceptualisation. EM, SM, GG and MH finalised the study design. EM, SM, GG and MH prepared the framework of this manuscript. EM analysed the data and led revisions of the paper. MHaertsch, CP and PC advised on protocol details including recruitment and provided input into the content of the manuscript. All authors read and approved the final manuscript.