Research utilization scholars have identified organizational context and culture as important factors influencing research use 2123233. Kitson et al. 34, Estabrooks 35, and Lomas et al. 36 have persuasively argued that changing practice is not just a matter of focusing on the behaviour of individual practitioners but also requires attention to the social, cultural, and material context within which practice occurs. Contextual factors that have been identified as promoting the successful implementation of evidence into practice can be grouped under the broad themes of culture 3738 and leadership 83739. Culture refers to the basic assumptions, values, and beliefs that are embedded in institutions and organizations 3740. Organizations most conducive to facilitating change are those that are described as 'learning organizations' 3741, which refers to an organization's capacity to recognize the value of new knowledge, assimilate it, and then implement it as the basis of decision-making. As part of such a context, decentralized decision-making, collaboration, teamwork, receptivity to change, and shared goals for improvement are typically valued 8373840. Reducing the uncertainty that results from inconsistencies in unit management practices, technology-driven routine work, and the complexity of teamwork have also been identified as necessary precursors to increasing research utilization 32. The chances of successful implementation are enhanced further in contexts where clinical decision-making is informed not only by evidence from systematic reviews and randomized controlled trials, but also by patient preferences and clinical experience 3742. It is argued that there is greater likelihood for successful implementation where evidence is consistent with patient narratives and experiences as well as the tacit knowledge of practitioners 42. Considering clinical experience and patient preferences as valid sources of evidence requires broader evaluative techniques than 'hard' outcome measures alone 8. Thus organizations most conducive to research use have the resources to incorporate multiple methods and sources of feedback into their evaluative framework 42.

McCormack et al. 8 have emphasized the importance of transformational leadership, also referred to as shared partnership and distributed leadership 39, for creating a culture of inclusion that values all levels and rank of staff. Balanced power, shared purposes and goals, shared responsibility for work, and mutual respect are requirements of shared leadership 39 to effectively alter the prevailing organizational culture and to create a context more conducive to the integration of evidence and practice.

Clearly, contexts can invite or reject innovation, complement or inhibit the activities required for success, and sustain or alter adherence to entrenched practices 39. While there has been much progress by way of identifying which aspects of context may influence innovation adoption in healthcare, much less progress has been made in terms of understanding how and why contextual contingencies interact the way they do, and how these interactions influence KT 32. This is troubling, given the importance of understanding context for facilitating successful implementation. Many conceptual models depict relationships between the various different aspects of context 93743 without recourse to theory to facilitate understanding of why relationships assume the form they do, and the underlying factors of the complex realities of practice. Even those models that are informed by theory are limited in their capacity to conceptualize causal mechanisms that link structure and agency. Either there exists such deep bias in favour of viewing structural properties as overconditioning the actions of agents (e.g., situated-change theory 44 and theories of culture 45), or quite the reverse, agency is said to be the primary characteristic or driving force of behaviour (e.g., planned behaviour 46 and community of practice 47 theories).

Critical realism, with its commitment to elucidating both the structures which constrain and enable activities, and how individual action reinforces, challenges, or transforms structural impingements, offers a promising way to remedy the tendency to either strip agency of structure or structure of agency. Critical realism has effectively been used to evaluate cardiac rehabilitation programmes 48 and diagnostic and treatment delays in breast cancer 49. It has informed an analysis of the effects of racism on occupational relationships between nurses and doctors, and how these are mediated by professional ideology 50, the sociopolitics of evidence-based medicine 51, as well as the fields of evaluation (realist evaluation) 14, organizations 52 (see the notion of engaged scholarship 53), and health promotion 54.

Critical realism furnishes a sophisticated understanding of context. In critical realism 2728, a distinction is made between the real (underlying nature and causal powers of objects/agents), the actual (what happens if/when those powers are activated), and the empirical (what is experienced/observed) 5556. This distinction is central to an ontological conviction that there exists a reality distinct from and greater than the domain of the empirical 2728, and that this reality is comprised of structures and mechanisms independent of our perceptions. Mechanisms can coincide under real world conditions to produce emergent properties contingent in time and space, properties which are irreducible to those of their constituents 55. The notion of contingency contrasts with positivist notions of universal logical necessity (natural laws, generalizable truths) by highlighting the uncertain nature of phenomena (i.e., that propositions may hold true only under certain circumstances). In the domain of the actual, there are many mechanisms concurrently active where some reinforce one another, and others frustrate the manifestations of each other. In this sense, it can only be said that a certain object tends to act or behave in a certain way 57.

Danermark et al. 57 use the example of a match to illustrate the notion of tendency. The object (match) has within it the causal power for fire, but ignition requires this power to be triggered by agential mechanisms through the act of striking, as well as by mechanisms of nature including sufficient oxygen, dry conditions, etc. Irrespective of an agent's intent, numerous combinations of mechanisms may influence whether the causal power (fire) will manifest itself in the realm of the empirical. Thus generative mechanisms are real in the sense that they provide the conditions that serve to constrain or enable an individual's action. For critical realists, explanatory power derives not from counting the co-presence of observable phenomena and inferring causation on the basis of empirical co-occurrence, but from 'identifying causal mechanisms and how they work, and discovering if they have been activated and under what conditions' 58. Consequently, context becomes redefined as the interrelationship between real and emergent or possible properties of structures and agents:

The (local) mix of conditions and events (social agents, objects, and interactions) that characterize open systems, and whose unique confluence in time and space selectively activates (triggers, blocks, or modifies) causal powers (mechanisms) in a chain of reactions that may result in very different outcomes depending on the dynamic interplay of conditions and mechanisms over time and space 54.

In illuminating these aspects of context, critical realism is a perspective that is equally pertinent to program evaluation. Proponents of critical realist evaluation 5960 have argued that the central question is not so much whether certain interventions work in a generalizable way, but what will work with these stakeholders/actors in this setting at this time. This shifts the focus of evaluation of interventions from a programme-based view of what works to causal pathways 51. Opening the 'black box' 61 of implementation is necessary to better understand the relations between the innovation and structural and agential properties 62 that inform uptake, the need for refinement, and the factors important for replication 63.

For critical realists, agential capacity is not innate or static, but relational. It is activated in the mobilization of various forms of capital: social, cultural, and material/economic 64. Power is exercised in relation to others who are likely to mobilize stocks of capital and resources in order to promote their own interests. Human action is enabled and/or constrained by power inequities, but this action, in turn, reproduces and/or transforms those structures of power 49. For example, those perceived as having specialized knowledge, and social and economic authority often prevail. These stocks of capital are not randomly assigned but tend to follow time-honoured cleavages of race, gender, and social class, suggesting that social structures (including institutional practices, policies and regulations, cultural norms) play a role in the production and often (re)production of inequalities amongst social groups. This 'indebtedness of agency to structure', as Scambler 65 terms it, underscores the dialectical relationship that exists between human action and structures of power 49.

Power relations may be ubiquitous, but they are expressed in different ways in different settings, in part because other mechanisms are also at play which may be local manifestations of much broader processes (e.g., gender and race relations, management-labour relations). Contemporary neoliberal 'logics' of management practice (concerned primarily with profitability, cost reduction, cost-per case efficiency, and standardization) must figure prominently in any such discussion. This kind of managerialism 66 seeks to parse healthcare into discrete tasks that can be easily measured against written standards pertaining to how much time can be spent on a given task and how it should be done 67. The measure of care lies with the physical task rather than the quality of human interaction and, as a consequence, the relationship between the care provider and recipient is not always quintessentially one of caring unless those most closely involved make a point 68 and have the requisite capital 49, to make it so.

Thus interventions aimed at 'humanizing' care must acknowledge that such interventions intersect powerfully with other dynamics (decision latitude, service delivery trends, atomization of the nuclear family leading to loss of proximal family members, etc.) in ways that, by virtue of the underlying causal powers at play, have the ability to either enhance or undermine change initiatives. Critical realism proffers a view of evidence-based practice that concentrates on an elaboration of mechanisms and the logic of causation rather than a programme-based view of what works in terms of research-manipulated interventions and independent outcome measures. It is an approach to implementation evaluation (also referred to as formative evaluation) 63 that, when combined with outcomes evaluation, creates a powerful 'hybrid style approach for implementation research' 63 which provides a clearer direction for action because the decision maker not only has knowledge of the outcomes but also of what produced the occurrence or absence of targeted outcomes.

More is at stake here than an exhortation to be mindful of context as a kind of general backdrop for interventions. In seeking to understand how mechanisms play out in a particular setting, with particular agents at a specific time, we must also take account of how reflexive agents perceive, negotiate, unwittingly reinforce or selectively resist the effects of these broader trends and influences in the context of their own life biographies, socialization, and the micro-social context of peer relations in the workplace. Critical realism is a perspective that deems the creative tactics of individuals to deal with impingements in the social and material contexts of everyday life to be of equal importance to the social structures that furnish such impingements 49. Deep underlying generative mechanisms do form the basis for structural impingements on human activity, but structural relations of gender, class, and race can for example be actively resisted or reproduced during encounters with the healthcare system by practitioners (or patients) mobilizing their own stocks of capital in particular settings and contexts 49.

Diverse disciplines, practices, and literatures have identified the problematic nature of engagement as a central issue for a myriad of professional practices. Taking agency seriously means finding ways to work with practitioners to help them understand their situation, examine their values, identify barriers and opportunities for change, implement solutions, and evaluate the results while never losing sight of the ways in which generative mechanisms operate to constrain and/or enable change in particular settings. This requires a more sophisticated approach to engagement and dialogue that draws in and works with the whole person in his or her 'multiple literacies' 69. This is where the arts, as a medium for reaching and engaging care providers, can be particularly powerful.

There can be considerable leeway for evidence to be (mis)interpreted, resisted, adapted, and even dismissed by potential adopters 19. It is therefore imperative that when bringing evidence-based innovations to practice and encouraging their adoption, use is made of approaches that view potential adopters as beings capable of reflecting critically on their own assumptions, and on the relationship between their practice and its context 15. KT strategies have ranged, for the most part, from passive unplanned efforts (diffusion; e.g., publication of research findings), to targeting and tailoring the evidence and the message for a particular audience (dissemination; e.g., direct mailing), to systematic efforts to encourage adoption of the evidence (implementation, e.g., use of incentives and sanctions) 70. There is evidence to suggest that interactive educational interventions such as workshops can result in significant changes in professional practice 17. The arts, however, have been neglected as a KT strategy despite their enormous interactive, educational, and emancipatory potential; an omission that our model, CRARUM, specifically remedies.

Literature and theatrical performance are increasingly being used as a means to humanize medical education 71. Shapiro and Hunt 72 contend that live theatrical performances contribute significantly to medical education because they have 'a uniquely compelling emotional quality, making it difficult to avoid or intellectualize the struggles and suffering portrayed.' A growing number of health researchers are turning to theatrical performance as an innovative approach to extending research findings beyond the discipline in which they were generated, and thereby making research more accessible and relevant in health care settings 30737475. Examples of research-based productions include schizophrenia 76, substance abuse 77, breast cancer 73, prostate cancer 78, ovarian cancer 72, AIDS 72, Alzheimer's disease 30, and traumatic brain injury 79. Dramatic performance is particularly effective in engaging imagination and fostering sympathy because it privileges the phenomenological complexity of life. It draws the observer into a particular social and cultural world with all its textures, sounds, gestures, and movements 30 in contrast to textualism, which flattens out 'the flux of human relationships, the ways meanings are created intersubjectively as well as intertextually, embodied in gestures as well as in words...' 80.

Dramatic performances have been successful in helping practitioners and medical trainees reflect on the care they provide and increase their understanding of patient care issues 727881. For example, in post-performance evaluations of the drama No Big Deal? 78, based on a study of prostate cancer survivors and their spouses, oncology physicians, nurses, and allied health professionals indicated that attending the performance resulted in a new level of awareness or understanding of how patients are affected by cancer diagnosis and treatment. Post-performance evaluations of a research-based drama about personhood in Alzheimer's disease 31 found experienced nursing and allied health professionals acquired a new level of understanding of the expressiveness of persons with cognitive impairment. Deloney and Graham 82 have similarly validated the use of drama as an effective method to provide training about end-of-life issues and doctor-patient communication. These evaluations support the effectiveness of research-based drama as a KT strategy with the potential to positively impact practice.

Improvisational theatre is an important form of drama that is influencing the way social and health scientists are incorporating drama into their research 83. Developed out of the political-theatrical agenda of Augusto Boal, a Brazilian theater director, writer, and theorist, 'forum theatre' is a method of teaching lay non-actors how to recognize and transform the conditions of oppression in their lives. The theatrical goal is to engage the disempowered and to create ways to liberate the disenfranchised 84. A short play is performed, followed by an identical presentation in which audience members are encouraged to rise and physically replace the main character when they feel inspired to enact an alternative approach that might result in a more favourable outcome 85. Highly interactive and imaginative, forum theatre fosters critical thinking about the lived reality of the participants, the root causes and solutions to social problems, and change. The collaborative process is intended to address the need for participants to step outside 'the apparently solid 'matrix' of 'this time in this place' and collectively de-codify the 'myth of fixed reality' – engendering hope for transformation' 85. Attitudes, beliefs, conflicts, failures, successes, and aspirations are shared, and emerging from this process is a vision of how things could be different 83. Mienczakowski, for example, has used elements of Boal's forum theatre techniques in ethnographic performance projects about schizophrenia and alcoholism 83. His use of these techniques was intended to provide emancipatory opportunities and insights for both health professionals and health consumers 83.

By offering the potential to foster critical awareness, to facilitate understanding, and nurture sympathy, arts-based approaches are well positioned to strengthen initiatives that seek to transform health care. In a recent review of the literature on the use of research-based drama for KT 86, a number of areas for further exploration were identified. First, little is known about the extent to which drama impacts health audiences, why it has the impact that it does, and whether and how this impact leads to real world application. Second, distinguishing the aesthetic qualities of the performance from its content has yet to be done, and this too would lead to a better understanding of the particularities of drama that work as a KT strategy. Finally, because the most common methods employed in evaluation studies have been unstructured feedback (e.g., reflective journals from students and informal discussions), and self-report questionnaires, qualitative methods are recommended to generate a rich data set for understanding how research-based drama operates as a KT strategy.

How would CRARUM help to guide users in successfully implementing evidence into health care settings? Following the logic of critical realism, as a necessary first step, qualitative and quantitative methods of data collection can serve to identify causal generative mechanisms of existing care. These data reveal contradictions between espoused and enacted practice, and existing barriers to best practice. Elucidating the social, cultural, and material conditions under which practice occurs enables the intervention to be meaningfully individualized to the care setting. For example, understanding the context of long-term care would be critical when implementing an educational intervention for front-line dementia care practitioners about a patient-centred approach to care. Understanding how administrators and practitioners negotiate the potential paradox of front-line staff being mandated to provide patient-centred care, despite healthcare rationalization policies that constrain their ability to do so, would further inform the development of adoption strategies for facilitating the use of the new approach to practice. Moreover, because organizational hierarchies can limit practice change 2087, engaging administrators in the development of adoption strategies can increase the likelihood of successful implementation. Theorizing the dynamic interrelationship between individual agency, organizational rules and regulations, and the larger health care restructuring agenda can facilitate the tailoring of the intervention such that its relevance, feasibility, and success are maximized.

Yet, tailoring interventions to better fit local settings alone is insufficient to achieve optimal care settings. The arts provide an innovative approach to the challenge of engaging practitioners to imagine new possibilities for more humanistic caregiving practices by helping practitioners to meaningfully connect with their care recipients 31. The use of drama, for example, can raise critical awareness of taken-for-granted assumptions about standard care practices, and affect change through reflection on the nexus of personal assumptions, staff behaviour, and organizational policy 88. In so doing, it can facilitate the development and implementation of an agenda for change that derives from the critical awareness of stakeholders themselves 15.

Critical realist evaluation of an intervention would take into account both the process and context of change. This entails an exploration of outcomes (e.g., non-pharmacological approach to behavioural management in dementia care), but also the conditions that were present to enable those outcomes (e.g., administrator support of practitioners' adaptation of care to meet patient preferences, such as having an evening bath rather than a morning bath, instead of rigid enforcement of institutional routines). Qualitative and quantitative data collection can inform understandings of what did/did not occur within the setting relative to the intervention, and which structural and agential factors influenced adoption of, or resistance to the intervention. Thus, in addition to answering whether the intervention works, critical realist evaluation facilitates understanding of why it worked, for whom, and in what circumstances.