Probst and Borzillo propose ‘ten commandments’ for effective communities of practice and suggests five key reasons for failure [8]. The researchers identify six key themes important to the establishment and maintenance of successful communities of practice: Leadership, Sponsorship, Objectives, Boundary Spanning, Risk-free environment and Measurements. These themes are explained and expanded upon as each theme is explored in relation to the literature identified for this review.
It must be noted that most of these studies are qualitative and there is varied statistical analysis and methodology reporting (Table 1). These papers have been read extensively and, where comments or discussions or conclusions from data, or from the project being discussed, are made, then these items are matched against the themes in Probst and Borzillo’s framework (Table 2). This is not an assertion that these themes have been formally studied as outcomes for each study. The additional themes of ‘Technology” and “Community” barriers and enablers have been included to cover a number of similar themes in these studies.
Theme 1: Leadership
Probst: The organisation can designate leadership roles to motivate community members to collaborate
Almost every study in this review commented on leadership, facilitation or moderation [21, 23–38]. Previous studies have commented on the lack of clarity around these terms in virtual communities of practice [21]. In this review, it appears that these roles, whilst overlapping, are different.
Facilitators/Moderators
The most common role described in the studies was of the facilitator or moderator. This role may arise in several ways. The originator of the group may end up being the initial leader and facilitator [23]. The facilitator may be appointed by the originators of the group [24–26] or the facilitators of the group may arise spontaneously [24].
If they arise spontaneously, then these moderators or facilitators tend to be part of the ‘core group’ which also characterises these virtual communities [23]. The ‘core group’ consists of a minority of active users, whilst often the majority is passive [25, 26]. Despite this passivity, these users are still seen as benefiting from the network as ‘legitimate peripheral participants’. As one GP put it, I have not used CHAIN much but it is a security blanket!” [26].
The tasks of the facilitator and moderator are, as Probst described, to improve collaboration [27, 28], but can also include making sure the rules of engagement are clear, keeping discussions focussed and processing memberships [23, 26, 27, 29].
There is some controversy about ongoing facilitation. One researcher believed that these networks can be self-sustaining [23], one found that it was definitely not [30], however most simply used facilitators, or had facilitators emerge, throughout the projects.
Leadership
In one study without formal facilitators, ‘leaders’ emerged. This ‘emergence’ demonstrated the opportunity for horizontal leadership to occur in VCoPs, in which marginalised or junior members of staff have the chance to emerge into leadership roles, potentially taking forward actions that arise from discussions [27].
In the same online midwifery forum, more senior nurses used their postings to praise other contributors and to validate the use of the forum, successfully encouraging usage. However, praise online actually fits better with the role of a moderator and from the perspective of Probst’s thematic analysis, the ‘leadership’ shown in validating the use of the forum by the organisation may fit better under ‘sponsorship’ [24].
Probst tells us that the role of the leader is in promoting collaboration [8]. However the definition of leadership in the articles reviewed is controversial. Li’s systematic review highlights the fact that the role of leader and facilitator may be separated or performed by the same person [21]. In terms of roles, in the articles reviewed it appears that it is actually the facilitator and moderator who promote collaboration. Leadership, when implying validation by the organisation, can actually be seen as equivalent to Probst’s ‘Sponsorship’ or the display of executive approval for the network. The main importance of the leader found in this review is in the initiation of the community. In many of these studies that role was actually performed by the study organisers [30, 31, 39]. In studies in which the study organisers are not the leaders, then this concept of leadership and initiation merge with Probst’s concept of sponsorship.
Theme 2: Sponsorship
Senior executives need to provide sponsorship to help communities reach their full potential
In business, Probst’s finding was that effective CoPs had a sponsor, or senior executive, who sanctioned the CoP. There was then a leader that drove the community [8].
The findings in the current literature review were that, in fact, in health the agenda is usually driven by the organisation attempting to start the community and/or the researchers founding the community. It is then the moderators and active group that continue to stimulate and promote knowledge sharing.
Sponsorship, initiation, vision or leadership was evidenced in many of the studies, as the groups were collaborations between stakeholders that were forming a network to solve a problem. Ultimately, someone had to start the network, then continue to support its activities. For example, the CHAIN network of evidence in the UK is part of the NHS Research and Evaluation network, ICUConnect is part of the ICU Monitoring Unit and the proposed e-collaborative platform for the Montreal Stroke Network is formed from a number of state and national stakeholders [26, 29, 32].
Once created, ongoing organisational support was essential to the success of projects. This was demonstrated well in a group of gerontological nurses that needed ongoing support from high-level nurses to legitimise work-based learning, before the use of the online environment was accepted [39].
Whilst sponsorship describes the process of the corporate world well, in the health context there are some differences. Mostly, the networks have an initial purpose of knowledge sharing that supports the organisation, or the researchers’ study, and thus are a collaboration of multiple stakeholders such as a health service, the researchers and clinicians, rather than the domain of a single company.
Theme 3: Objectives
Clear objectives provide members with responsibilities and motivate them to contribute more actively
Each VCoP studied had an objective, however these objectives ranged from clear and specific to broad. The success of networks with specific objectives initially appears to support this statement [24, 25, 31, 34, 39]. For example, the development of evidence-based ‘best practice’ statements for gerontological nurses in Scotland led to the better uptake of evidence-based practice, using a Virtual College and CoP. However, a number of networks had broad objectives within a specialised group of practitioners and were also successful [23, 24, 34]. For example, Nagy’s network for PACS online radiology systems had a broad objective to “facilitate and accelerate PACS through education and communication”. Within that framework, users developed their own goals and content through posted queries and responses. A similar pattern was found in Brooks’ midwifery forum [27].
However, when a busy psychologists’ network was reviewed for the outcome of ‘professional identity creation’, there was less success. The network had not been set up for this, and perhaps its broad goal of providing a ‘meeting place where ..professionals…can establish valuable relations; sharing experiences information and practices...’ contributed to the lack of specific identity formation [35]. Also, a network of nursing academics experienced some problems with lack of focus [30].
Probst describes clear objectives and sub-objectives for CoPs. For example, a car manufacturer may have a broad objective of improving engine performance, with sub-objectives around building and exchanging technical knowledge around each of the engine parts (valves or internal combustion for example). The findings from this review are that specific objectives are helpful although, particularly in a specialised area such as midwifery or radiology systems, some networks succeed without a high degree of clarity around their goals.
Theme 4: Boundary spanning
Boundary spanning enables members to engage in internal and external benchmarking practices
Most groups in this review benefited from a heterogeneous make-up, although there were some problems. In almost every study, there were either a variety of practitioner types, or a variety of organisations participating. Booth found that linking CoPs in different sites via the virtual college accelerated their guideline development process for nurses [31] and Curran’s rural emergency departments benefited from their city cousins sharing expert knowledge and from the use of knowledge experts [40]. The evidence-based CHAIN network in the UK described the effective knowledge sharing between groups as a demonstration of strong and weak tie theory [26]. In this instance, strong ties are between users that know each other best, but weak ties between users only distantly acquainted or introduced via the network led to the greatest knowledge sharing.
However, if the group is too heterogeneous, there can be problems, as there is either not enough overlap for effective communication or antagonistic viewpoints between competing groups [30, 35].
Probst describes members of CoPs either being fed with external expertise, or making use of other CoPs either within, or from without, the CoPs company. This view differs from the health experience in that often these networks do not originate within a single ‘company’ or stakeholder. The boundary spanning occurs through the interaction between either different professional groups or different organisations, or both, whilst some used external experts.
Theme 5: Risk-free environment
COPs should be used as an especially valuable opportunity to express and test ideas in an informal and risk-free environment, thus requiring a strong degree of safety and intimacy between members
A risk-free environment came through as important in this review. Moderators were encouraged to enforce rules of no offensive language and ‘model citizen behaviour’ [23, 27] and protocols were developed about how users are to behave online with expectations of themselves and each other [34].
In addition to lack of risk, positive reinforcement was also important, along with a non-hierarchical atmosphere. One nurse said “I think if you keep encouraging people they will think and be creative” [39], whilst another commented that “It’s (the online environment), you know, a free atmosphere; to be able to do it without any comeback” [36].
A demonstration of the risks that users fear was the fact that Penn’s Suicide Prevention network had still not progressed to its original goal of online psychiatry advice due to legal concerns [34]. In addition, in an online anaesthetic network reporting on critical incidents, it was felt that some of the lack of reporting was due to the general culture of low reporting of incidents. This network also commented that users requested anonymity as an option, likely for the same reason [41]. Probst’s review demonstrates that a risk free environment is important in business to encourage growth. In health, although an environment must be risk free, it should also be positive and encouraging. This type of environment builds trust and thus improved communication.
Theme 6: Measurements
Empirical evidence suggests the use of measurements to assess the value of communities of practice
There was very little formal measurement identified in this review. One study found that regular feedback provided to participants assisted them in decision-making [31]. However, several studies commented on the value of informal ‘benchmarking’ or ‘validation’ of their own practice against that of other users and organisations [27, 39, 40], while other participants generated their own ‘closing the loop’ of actions resulting from the online discussions [24].
Measurement, benchmarking and feedback
The VCoPs in Probst’s review had more measurable goals, such as cost reduction or product improvement. However, he still notes that members posting online ‘stories’ of how their experiences have led to positive change motivates other members. In the health context, these measurements may be more likely to be member-generated, including benchmarking of practice or having feedback about organisational changes that have been triggered as a result of the discussion, rather than formal manufacturing targets.
Technology and community features
Whilst not specifically addressed by Probst and Borzillo, a number of other themes were found in this literature review, which have been grouped under the headings Technology and Community Features.
Technology
Making the technology easy was commonly cited as highly important. The concept of ‘easy’ included ease of use, ease of access and flexibility of options for communication [24, 27, 28, 30, 34, 37, 41].
Communication options in most studies included an asynchronous method, either by email or discussion boards [23, 24, 26, 28, 34, 37, 39, 42], while some studies used these with a mix of features including chat, content sharing and synchronous web-meetings [23, 34, 35, 39]. Email reminders were also suggested to be useful [26, 37, 41].
Whilst the previous features were more uniform, a number of areas were controversial. Some studies used passwords [28, 42] though lost passwords and online delivery created barriers for others [37, 39, 40]. The online environment was of real benefit to most [24, 27, 35], though one study found that the culture of face-to-face interaction amongst nurses was a barrier to use of online environments [30]. Lastly, training was mentioned as necessary by some [39] whilst others aimed to avoid training through simplicity of design [24].
Ease of use is paramount in any online community. Communities should offer asynchronous communication methods such as email and discussion boards and may consider other options such as chat and content repositories. When setting up a community, consideration needs to be given to the pros and cons of passwords, access, identification and training.
Community features
Effective communities of practice result in knowledge sharing [15]. This knowledge sharing can be encouraged by voluntary involvement, as self-selection appears to encourage users that are willing to share knowledge to participate [27, 28]. A particular feature of the CHAIN network of evidence in the UK is the reciprocity of members, that is the generosity of members when responding to queries from others [26]. However, whilst this active membership is essential, passive users can still be seen as Lave and Wenger’s ‘legitimate peripheral participants’, gaining support from watching the ‘expert’ users [25, 26]. The validation of each others’ practice and a desire to understand current knowledge are other factors that help sustain an online CoP [24, 27, 40].
Whilst online membership is helpful in overcoming barriers of geography and time [24, 27, 30], bonds can be strengthened through face-to-face meetings [31, 32]. In fact, one network started online, with physical chapters developing as a result [23].
Communities can help professionals overcome isolation through connecting with colleagues and sharing knowledge [27, 38]. One nurse said “I feel fairly isolated [because] I don’t have many peers (advanced practice nurses) in my organisation. The listserv helps give me ideas when I have no-one else to bounce ideas with in my hospital”.
In addition to the features mentioned by Probst and Borzillo, self selection, a desire to knowledge share and the blending of face-to-face and online involvement are desirable. It is worth noting that it is not just the active users that benefit from membership in such communities.
Implications
From this review it can be seen that there may be a role for VCoPs in general practice training, although a planned approach to research is needed. A VCoP for general practice training may decrease the social, structural and professional isolation aspects of training, thus improving trainees’ sense of connectedness and improve their knowledge sharing opportunities. The benefits of these outcomes could include higher general practitioner trainee satisfaction and knowledge, particularly whilst in rural placements, with implications for possibly helping to overcome workforce shortages and quality health care delivery in these areas.
Another potential benefit of a VCoPs for general practice training is that VCoPs can offer the potential to make invisible work visible. This might enable areas of practice that have traditionally occupied lower status in general practice to gain significance as members communicate their experiences. An example of a VCoP for general practice trainees could include online expert medical moderators facilitating case discussions, answering questions and helping to build a shared knowledge resource for trainees. During this process, under-represented or marginalised areas such as workers’ compensation related illness or youth mental health may be highlighted in discussion, thus raising their profile as well as providing practical tips for trainees with little exposure to these difficult areas.
Limitations
There are a number of limitations to this study. Firstly, the initial model is drawn from the business literature, with business outcomes in mind. In health, CoPs often involve several organisations, rather than one business. They may also be non-profit and the outcomes being measured may be more related to clinical care delivery or knowledge sharing and overcoming professional isolation. It was also unclear in the Probst and Borzillo model how many of the CoPs were in fact VCoPs and there was no subset analysis on this differentiator, which is noted in the Probst and Borzillo paper.
Secondly, the overall data quality of many of these papers is limited and in particular there is very little rigorous outcome data. Future studies must include an examination of efficacy in addition to qualitative review.
Finally, the themes that have been generated from each paper are not formal themes that have been evaluated in each paper. In many cases they are drawn from descriptions of the project or interpretations of the data by authors, but with variable data quality (see Table 2).