Limitations to thinking about KT
Although the KT research community has recently begun to acknowledge complexity, Van de Ven et al. (1999) were one of the first research teams in the area of the management of innovations to demonstrate empirically that new ideas do not follow a logical flow from generation to implementation. They argued that the appealing logic of the traditional approach, characterized by Rogers (2003), has one major drawback: complex systems theory describes all ‘living systems’ as being in a state of ‘dis-equilibrium’ (Dooley 1997) and thus the linear approach does not reflect reality. Their research found no support for a stage-wise model of innovation development and no support for a linear model of adaptive trial-and-error learning, particularly during highly ambiguous and uncertain periods of the process.
If KT is not a logical, linear process, does this mean it is random? Would it be as effective to bombard every healthcare organization with guidelines, protocols, policies, procedures and evidence and wait to see what unfolds? Van de Ven et al. (1999) have described this state as the result of a non-linear dynamic system, arguing that the innovation journey is not stable, predictable or random.
This leads to a wider consideration of the ways in which the systems we work in are routinely and implicitly conceptualized (Miller & Rice 1967). Systems have been described as how resources and human effort are organized to achieve a set of tasks. Closed systems are self-contained, have an internal logic and predictability; open systems are involved in importing, converting and changing resources from one state to another (Miller 1993). Open systems accommodate relationships – indeed they are predicated upon a set of assumptions about the understanding of the relationship and the tension between individual and group and the individual and the organization.
Lewin (1947) argued that conventional models of scientific analysis could not uncover the ‘gestalt’ properties of complex human systems. Therefore there is a need for social systems to acknowledge that the structured properties of the dynamic whole are different from the structural properties of sub-parts. This ‘balancing’ between the individual and the group, the sub-system and the whole system, influenced the early social innovations movement (Rice et al. 1950) and has led the way to learning organization theory (Senge 1990) and the role of systems thinking in creating effective organizations (Plsek 2001).
Senge (1990) describes two sorts of complexity: detailed complexity – logical, specific and mechanical – and dynamic complexity. The characteristics of dynamic complexity are time (the same action may have dramatically different effects in the short and long term), location (an action can have one set of consequences locally and a very different impact on other parts of the system) and unintended, non-obvious consequences. The essence of systems thinking lies in a shift to seeing inter-relationships rather than linear cause-and-effect chains, seeing processes of change rather than snapshots. Senge therefore advocates active involvement in reflection-on-action and drawing understanding and meaning from feedback on action and activity. This feedback moves to learning and the recognition of patterns that start to emerge. When the deeper structural and process patterns of systems are recognized we begin to learn how to influence those patterns.
This leads to the second proposition.
Proposition 2: The system-as-machine metaphor is profoundly unhelpful to our understanding of KT and of how to get new knowledge into practice.
The healthcare system-as-machine metaphor might appear anachronistic, but the experience of many patients and staff is that the systems in which they find themselves feel as if they are driven more by technology and rationalism to the detriment of other necessary qualities such as an emotional response, being present and drawing meaning from events (Brown et al. 2007).
Consequences of viewing the healthcare system as an ‘organism’ rather than as a ‘machine’
Moving from our logico-deductive paradigm into a more interpretive world of relationship and multiple perspectives can be disconcerting. If we agree with the logic and consider the evidence for a more interactive way of exploring KT, how do we go about developing and supporting this kind of research?
Is the debate about theory development and testing, as advocated by Estabrooks (2007), the ICEBeRG Group (2006), Grol et al. (2007), a step in the right direction? It is indeed important and timely if it embraces the open system, relationship approach. One bridge created between the traditional scientific world and the world of systems as organic entities is that of Van de Ven et al. (1999, 2000) empirical work. This leads to the third proposition.
Proposition 3: The (healthcare) system is best viewed as a complex, interactive, organic entity where experimentation, experiential learning and reflection are central to creating a culture of innovation, improvement and consequently effectiveness.
The innovation journey is defined as a sequence of events in which new ideas are developed and implemented by people who engage in relationships with others and make the adjustments needed to achieve desired outcomes within an institutional and organizational context (Van de Ven et al. 1999). This contrasts with the general expectation of the innovation process (as defined by Rogers 2003 and adopted by Greenhalgh et al. 2004) as involving sequential invention, development, testing, adoption and diffusion [For example, Greenhalgh et al. (2004) p582 define innovation as ‘a novel set of behaviours, routines and ways of working that are directed at improving health outcomes, administrative efficiency, cost effectiveness or users’ experience and that are implemented by planned and co-ordinated action.’ Emphasis added.].
Organizations undertake the innovation journey each time they invent, develop and implement new products, programmes, services or administrative arrangements, according to Van de Ven et al. (2000). Using a grounded theory approach, they derived a clear set of concepts for selecting and describing the objects to be studied on the journey and then developed systematic methods for observing change in the objects over time.
Important for KT research is the focus of this team’s work on process analysis and examining inter-relationships among variables at several points over time (Nisbet 1970, Pettigrew 1985, Van de Ven & Poole 1988). Their framework examined the development of innovations in terms of five key concepts –ideas, outcomes, people, transactions (relationships) and contexts. These concepts have been used to explain how the assumptions and observations about the core concepts diverge from orthodox literature on the subject and on the traditional views of KT research (Table 2) and how the innovation process works (Table 3).
Table 2. Assumptions and observations about core innovative concepts
|Themes||Literature implicitly assumes||Van de Ven et al.’s (1999) interpretation||Traditional KT view|
|Ideas||One intervention operationalized||Reinvention, proliferation Re-implementation, discarding and terminating||Emerging from scientific research, refined and tested for rigour Synthesized and passed to practice world to implement|
|People||An entrepreneur with a fixed set of full time people over time||Many entrepreneurs distracted, fluidly engaging and disengaging over time in a variety of roles||Experts developing the knowledge, opinion leaders in receipt and adapt to situation. Rest in the system influenced by leaders|
|Transactions||Fixed network of people/firms working out details of new ideas ||Expanding/contracting networks of partisan stakeholders who converge and diverge on ideas||Hierarchy of key influencers who are incentivized to embrace new ideas and permeate them through the system. Little explicit acknowledgement of changes in roles, relationships, ways of working|
|Context||Environment provides opportunities and constraints on innovation process||Innovation process created and constrained by multiple enacted environments||Acknowledge impact of environment and attempts to control its effect. Expectation that innovations can be introduced in resource neutral way|
|Outcomes||Final results orientation as stable new order comes into being||Final results indeterminate – many in-process assessments and spin offs; integration of new order into old||Expectation that final results will demonstrate tangible, measurable improvements within a specified time frame|
|Process||Simple cumulative sequence of staging phases||From simple to many divergent, parallel and convergent paths some related others not||Models describe a simple cumulative sequences of staging phases|
Table 3. Twelve elements of the innovation journey
|1. Long gestation lasting several years in which seemingly random events occur that precede and set the stage for initiation of innovations|
|2. Coalescence of factors – internal and external – to create a ‘shock’|
|3. Frantic search for resources to turn innovations into practical realities for adoption and diffusion|
|4. Initial innovative idea proliferates into multiple divergent ideas and actions|
|5. Setbacks and mistakes frequently encountered, resource and development timelines diverge; unattended problems ‘snowball’ into vicious cycles|
|6. Criteria for success/failure change, differ between resource controllers and innovation managers, diverge over time, trigger power struggles between insiders and outsiders|
|7. Emotional roller-coaster for staff involved – euphoria, frustration, pain, closure also characterized by high part-time turnover of staff|
|8. Investors/top managers involved throughout the process. needed to solve problems|
|9. New interdependencies created as a result of the innovation which in turn start to impinge on the wider organization|
|10. Entrepreneurs often involved in creating the (new) infrastructure necessary to gain support and legitimacy for collective innovative efforts|
|11. Adoption/implementation efforts do not wait until the innovation is completed – they occur throughout the whole development period|
|12. Innovations stop when they are implemented or when the resources run out|
Van de Ven et al. (1999) concluded that the process of innovation was messy and complex; it could not be reduced to a linear model of stages or phases. Despite articulation of the 12 steps (see Table 3), they caution against reducing the process to a planned, co-ordinated set of actions (as implied in Greenhalgh et al.’s 2004 definition). Rather, they explain the innovation process as a non-linear cycle of divergent and convergent activities that may be repeated over time and at different organizational levels to renew the cycle, if resources are obtained.
This cyclical pattern is not unique to innovation: Senge (1990) describes reinforcing (divergent) and balancing (convergent) behaviours and, within quality and safety literature processes such as PDSA (plan, do, study, act), there are implicit connections between periods of convergent and divergent behaviour (Thor et al. 2004). Equally, within action research and action science, spirals or cycles of learning, doing and reflection are well-recognized and embedded processes (Titchen & Manley 2006). However, how well this underlying dynamic is understood within KT is an important theoretical and practical question. Van de Ven and colleagues found that the divergent/convergent cycle was the underlying dynamic in developing a corporate culture for innovation, learning in innovation teams, leadership behaviours of top managers, building relationships with other organizations and developing (an appropriate) infrastructure for innovation (p213).
Building on the organizational learning theories of Senge and the innovation work of Van de Ven et al., there is a growing argument that KT should be viewed as another type of innovation process, which becomes subject to the vagaries of the process journey (see Table 4).
Table 4. The divergent–convergent model of behaviour within innovation journeys
|Divergent behaviour Branching and expanding process of exploring new directions Random, chaotic patterns Creating ideas and strategies Inspirations and negotiations||Theme launching||Convergent behaviour An integrating and narrowing process of exploiting a given direction – linear periodic pattern Implementing ideas and strategies Pushing ideas into currency|
|Learning by discovery Exploratory search||Learning||Learning by testing Trial and error|
|Pluralistic leadership Encouraging and balancing diverse views||Leading||Unitary leadership Encouraging unity and goal consensus|
|Building relations and porous networks||Relationship building||Executing relations in established networks|
|Creating infrastructures for collective advantage – running in packs||Infrastructures||Operating within infrastructures for competing advantage|
Do recent innovations in nursing reflect any of these elements?
Table 5. Nursing innovations in relation to the management of innovation
|Themes||Interpretation of the innovation process||Practice development movement||Magnet hospitals movement|
|Ideas||Reinvention, proliferation Re-implementation, discarding and terminating||Emerging from clinical nursing practice, patient experiences and traditional research routes Proliferation of ideas Re-interpretation Re-implementation||Emerged from research on high performing organizations, transposed to acute hospital sector; emphasis on systems/contextual issues|
|People||Many entrepreneurs distracted, fluidly engaging and disengaging over time in a variety of roles||Actors in system perceived as ‘stuck’, requiring support to innovate and creating own improvements – special internal and external support roles created to enable innovations to happen at local level||Innovation mediated through formal management executive structure with external organizations providing legitimacy|
|Transactions (relations)||Expanding/contracting networks of partisan stakeholders who converge and diverge on ideas||Creative problem solving and action supported by local facilitation coached by external experts||Fixed networks of people working out new ideas|
|Context||Innovation process creates and constrained by multiple enacted environments||Seen as key element within innovation process. Elements of context – culture, leadership part of change process||Central to transformation process itself|
|Outcomes||Final results indeterminate – many in-process assessments and spin-offs; integration of new order into old||Final results often seen as indeterminate – seen as a weakness in traditional KT endeavours- yet often reporting multiple in-process spin-offs, changes in teams, cultures, contexts, working patterns. Issue of time sensitivity to test for outcome changes||Final results relate to observable improvements in patient outcomes (mortality and failure to rescue) as a way of defending, justifying the shift in the power balance, relations at local level|
|Process||From simple to many divergent, parallel and convergent paths some related others not||Complex, multilevel, multifaceted, guided and supported through expert external and internal facilitation||Processes not viewed as important as outcome change viewed as staged, logical, sequential|
However, one issue not explicitly addressed by Van de Ven et al. (1999, 2000) but which has important implications for health care – and nursing in particular – is the question of what happens to innovations if they are introduced into contexts described as ‘underdeveloped’ (Towell & Harries 1993). Miller (1993) describes such contexts as having four key characteristics: lack of control over the immediate environment (victim rather than master); tightly integrated systems in operation (routines, attitudes, beliefs are all connected and fixed); relatively closed systems (automatic resistance to any sort of innovation and only certain sets of transactions allowed with other systems) and control over the external environment is nil (which often translates into feelings of apathy, worthlessness and impatience).
The term development in such situations would therefore imply a change in the direction of influencing and controlling the environment, thereby enabling the individual within the system to exert greater authority and control. Implicit in Van de Ven et al.’s work is this notion of intrinsic autonomy to act at local level. However, they do not deal with the matter explicitly – perhaps because the organizations studied all wanted to innovate and therefore were at a level of organizational readiness.
For nursing practice, the notion of an ‘underdeveloped community’ would seem to have some currency (Titchen & McGinley 2003, Down 2004, Manley 2004) and indeed the magnet initiative reflects organizations’ deliberate attempts to create environments that actively support innovation and enable local autonomy (Aiken et al. 1994). The genesis of the practice development movement (Garbett & McCormack 2002) was widespread dissatisfaction with the quality of care patients received and a growing sense that technological solutions promoted by quality improvement techniques did not sufficiently address the wider organizational learning and cultural issues. Whilst Van de Ven et al. explored change agent roles within the wider construct of leadership, they did not explore the concept of facilitation or the role of an external change agent who comes into an organization and attempts to support a change process (see Table 5).
In addition, Miller (1993) discussed the dilemma for the consultant involved in (rural) development work. This dilemma was that the change agent role could be perceived as both ‘top down’, i.e. the change agent being recruited to execute the objectives of the organization, as opposed to ‘bottom up’, where the change agent would negotiate with local groups what they wanted to achieve and would work with them to achieve it. Miller’s experience was that these two roles were not mutually exclusive and that an expert change agent knew how and when to enable both processes.
Thus, we are left with a strong argument, with supporting evidence, that making things better in systems is a complex, multilevel process. This includes agreement that context plays a major part in enabling or frustrating the success of the venture, and acceptance of the premise that much of the adoption of new knowledge is contingent upon learning styles and culture, levels of autonomy and support in terms of resource injection, and new ways of defining boundaries around power, groups, communication and action.
This tension between the naturalistic enquiry of innovation researchers such as Van de Ven et al. and the more involved action researchers (e.g. Titchen & McGinley 2003) and, more recently, intervention scientists (Brown et al. 2007) raises the central conceptual issue. The KT scientific community knows that single intervention studies are of limited value (Greenhalgh et al. 2004), and exhortations to undertake more theory-based and complex multilevel intervention studies are increasing (Estabrooks 2007, Graham & Tetroe 2007, Grol et al. 2007). However, what does not seem to have been agreed yet is what the conceptual framework should be, whether process (evaluation) is as important as outcome (evaluation), and how to investigate both the inner world and an understanding of the individual and how individuals within teams and systems create change. The biggest question hangs over whether orchestration of the above elements can be facilitated by individuals especially equipped to manage such complex, multilevel processes over time.
The community and practice development literature offer models to describe this facilitation role and, indeed, attempts have been made to describe similar ‘enabling’ roles within KT (see, for example, Harvey et al.’s 2002 concept analysis of the term facilitation and Stetler et al. 2006). However, the science around KT needs to explore these complex processes more systematically in the light of viewing the system as a complex entity in its own right. Accepting the core elements identified by Van de Ven et al. (1999) and arguing that the supporting or enabling processes need to be more explicitly articulated (particularly in ‘underdeveloped’ contexts) lead us to acknowledge that we are dealing with a complex set of phenomena. The underlying world views (ontological position) need to reflect that complexity and interrelatedness. Having set out the co-ordinates for the ontological positioning, it is then possible to select (and devise) appropriate methods (or epistemologies) that will help to refine the science of KT, thus speeding up the spread of new ideas into practice (Kitson & Bisby 2008).
The conditionality of the argument at this stage leads to the last two tentative propositions. On the basis of preliminary comparisons between the Van de Ven et al. framework and high level elements of nursing practice innovations, the following propositions are posed.
Proposition 4: Successful innovation (knowledge translation) into any system is a function of the local autonomy experienced by individuals, teams, the unit involved in the change process and their ability to translate this into purposeful and planned action.
Proposition 5: Innovation (knowledge translation) is most effective when it involves key stakeholders in personal development; control of immediate physical resources; the context and increased autonomy and control of the external environment.
The following hypothesis is put forward as a logical consequence of the five propositions.
Successful translation of knowledge into (healthcare) practice is a function of:
The way in which participants (individuals) in the system understand the nature and characteristics of the new piece of knowledge and accept it;
The level to which they can make informed, autonomous decisions about how they can use the new knowledge to improve outcomes;
How they negotiate and renegotiate relations with others (individuals, teams, internal, external relations) in their system and
How they attract necessary resources to sustain the changes/improvements in practice.
This process is enabled by expert facilitation where trained individuals (change agents, facilitators, knowledge brokers, consultants) simultaneously work with individuals, teams and the wider system to manipulate contextual factors and support the experiential learning of individuals and teams in managing the new knowledge.
In this emerging hypothesis it is argued that the innovation process is inherently so complex that it requires trained experts to enable it to happen effectively. Because it is a process, and because it involves many actors over time, the theoretical and methodological challenges posed for KT research are significant.