1a. Creation of new knowledge

New knowledge can come from a variety of sources, including researchers at academic medical centers, delivery systems, medical groups, health plans, and other stakeholders. Foundations, government agencies such as AHRQ, and other organizations can serve as vital facilitators of such knowledge creation. For example, within its patient safety portfolio alone, AHRQ supports the creation of new knowledge through more than 100 research and demonstration grants and contracts, which will generate a broad array of research findings and products within the next year.

The knowledge produced will vary greatly in terms of its readiness for utilization within health care delivery systems. Some of the new knowledge will emerge from basic research in its early stages. These findings will likely require further research before they can be considered for adoption and implementation in medical practice. Other projects (e.g., the evaluation of educational approaches and curricula, identification of best practices regarding organizational and systems-based processes, studies on the effectiveness of medical informatics tools) may create knowledge and products that can be adopted and implemented more quickly in a receptive environment. New knowledge will also vary greatly in other respects, such as the importance of its contribution to science, the strength and appropriateness of its supporting evidence, and the extent to which it supports or deviates from existing bodies of knowledge in the field.

1b. Knowledge distillation

Because the knowledge generated in the patient safety portfolio is extensive and variable, it needs to be inventoried, synthesized, evaluated, and prioritized for dissemination. The knowledge distillation process aims to identify results that are valid for generating action. These results are then forged into “actionable” messages or products ¹² relevant to particular audiences and end users. Achieving agreement on what is actionable is potentially a big challenge; the distillation process can help to forge consensus. Knowledge distillation helps to reduce the signal-to-noise ratio for potential users and dissemination agents by emphasizing findings that are both important and actionable, thus increasing the likelihood that research evidence will find its way into practice.

The distillation process involves an assessment of the new knowledge produced by the patient safety research portfolio in the context of the entire relevant body of research literature and the practical tacit knowledge of health care professionals. The process may take a variety of forms, including development of evidence summaries and synthesis papers, and the use of expert panels and structured consensus processes for identifying priorities and actionable findings and products.

It is important that the knowledge distillation process be guided by the ultimate goal facilitating dissemination and promoting implementation of research findings and products relevant to specific target audiences. This goal should be reflected in the composition of the groups conducting distillation activities. While researchers or knowledge creators can and should be important participants in this process, the imperative of moving research into practice means that other organizations that have a broader perspective should be represented as well. For example, for AHRQ's patient safety portfolio, distillation activities might include some research grantees and various external groups well versed in the practice of developing evidence syntheses (e.g., Institute of Medicine, the National Quality Forum, and the Cochrane Collaborative). Additionally, perspectives of relevant professional and policy associations (e.g., the American Physicians and American Nurses Associations, the American Hospital Association, and the Joint Commission on Accreditation of Healthcare Organizations), who may not have been traditionally involved in the knowledge distillation function, should be included.

It follows that the criteria used in knowledge distillation should reflect the concerns of potential knowledge users (e.g., transportability to real life conditions, feasibility, and the evidence needs of health care delivery organizations and clinicians), as well as traditional scientific considerations (e.g., strength of evidence, consistency with other knowledge, and generalizability across populations and settings). ^{11, 13, 14} Balancing the need to move promising findings and tools more quickly into practice settings with scientific requirements for rigor and replication is a critical task for the knowledge distillation process.

2a. Creation of dissemination partnerships and knowledge transfer teams

The framework calls for the development of dissemination partnerships that can help garner scientific and professional support for key research results and products to facilitate their adoption and implementation in everyday practice. Dissemination partnerships link knowledge sources (researchers and knowledge distillers) with potential intermediaries (individuals and organizations that can function as knowledge brokers and connectors to the practitioners and delivery organizations in the health care system).

Partnerships that involve knowledge broker organizations (e.g., Institute for Healthcare Improvement, National Patient Safety Foundation, academic publishers, media, and educational institutions) can amplify awareness of the research findings among interested professionals and the public. Partnerships with membership organizations serving specific constituents (e.g., American Medical Association, American Nurses Association, and the Institute for Safe Medication Practices) can also be particularly effective. Acting as intermediaries between the knowledge and user communities, these connector organizations can help ensure that research-based recommendations actually influence the way medicine is practiced. ⁴ They can do so by identifying priorities for dissemination, shaping key messages and broad guidelines for specific target audiences, and developing specific interventions for adoption by potential user organizations.

Dissemination partnerships are vital to developing the network of support needed to facilitate the transfer of research into practice because they provide an authoritative “seal of approval” for new knowledge and help identify influential individuals, informal networks, and organizations to champion these research-based recommendations. These partnerships actively contribute to the knowledge transfer process, helping to create demand, or “pull,” from the user community by leveraging off existing relationships among health care professionals and organizations. In addition to facilitating transfer by using well-established and credible messengers, dissemination partnerships bring in relevant contextual knowledge about end-user environments, motivations, and constraints that smooth the transition from research to practice.

Focused dissemination partnerships can take the form of multidisciplinary knowledge transfer teams, such as those that have been effective in disseminating research-based cancer prevention programs. ¹⁵ These teams—comprising researchers, representatives from intermediary and end-user organizations, and representatives of AHRQ dissemination offices and programs—create tools to promote adoption and minimize barriers among relevant health care organizations and clinicians. To translate knowledge into more usable formats, they may develop prototype or model intervention packages that capture the essence of the research-based change(s) and offer incentive structures that address potential barriers and facilitate adoption and implementation of the intervention. Functioning as boundary spanners, these knowledge transfer teams would identify what decisionmakers in user organizations need to know; explore the policy and organizational implications of implementing new research findings within particular settings; determine the relative benefits and costs of the change for particular organizations; and stimulate support among important stakeholder groups.

2b and 2c. Mass diffusion and targeted dissemination

In the framework, actionable, evidence-based research findings can reach audiences in two ways: mass diffusion and targeted dissemination or marketing. Using both strategies in parallel will help the research findings and products reach “the tipping point”— the stage where the pace of adoption of a new idea, product, or process accelerates markedly, largely because there are enough early users to convince the remaining holdouts to adopt the innovation. ¹⁶

In broad terms, mass diffusion involves broadcasting information to broadly defined audiences, with the objective of sharing knowledge and creating awareness. The typical mechanisms for mass diffusion of research results and innovations are journal publications and professional conferences. Other one-way information transmission approaches (e.g., fact sheets, newsletters, and Web sites) are used as well. In addition, the national and local media can be used to share information with policymakers, professional audiences, and the public at large.

Targeted dissemination involves planning for the delivery of specific messages to particular audiences using appropriate transmission mechanisms and messengers. While the objective of mass diffusion is providing information and raising awareness, targeted dissemination intends to persuade, influence, and motivate action. Targeted dissemination has taken the form of one-way “information-push” mechanisms, such as continuing medical education (e.g., workshops and conferences) and clinical guidelines. In general, these methods have been found wanting in their ability to create lasting change. ^{17, 18}

Our framework extends beyond the information-push models. We suggest that targeted dissemination efforts be conducted by partnerships involving knowledge sources working alongside knowledge broker and connector organizations, which can use the means at their disposal to communicate with and influence their particular constituencies. Many studies ^{8, 19} show that these partnerships could develop multi-faceted dissemination strategies for patient safety research, with an emphasis on the channels and media that are most effective for particular user segments. Information-push methods should be supplemented by more interactive and personalized opportunities, including workshops and working sessions with representatives of knowledge broker and connector organizations.

The effectiveness of these efforts depends on having comprehensive strategic plans for targeted dissemination. A transfer strategy for any particular knowledge or product cluster should specify a coherent and mutually reinforcing set of activities aimed at communicating key messages and enhancing the capacity of the target professionals and organizations to comprehend and act in response to those messages. Specific audiences and stakeholders must be identified, and their missions, needs, and contexts understood, in order to develop initiatives that will be adopted by the user organizations and health care professionals. It is particularly important to understand how potential innovators and early adopters and their professional networks within end-user organizations can be mobilized to extend the transfer of knowledge within their organization.

3a and 3b. Intervention development, adoption, and implementation

One must distinguish between the scientific rationale for implementing new processes or treatments and the decision to implement those innovations in a particular organization. Implementation of innovative evidence-based practices in operational practice settings involves complex interrelationships among the innovation itself, the organizational structures and values, the external environment, and the individual clinicians. ^{11, 19} For new evidence-based practices or processes to be adopted within particular settings, a “change champion” must often overcome a variety of barriers, e.g., the lack of clear structures for evaluating and adopting evidence-based practices, lack of incentives and training for individuals and teams to adopt such innovations, the complexity of the required change, existing regulatory and accreditation requirements, and the staff and capital costs involved. ^{10, 20 – 22}

Initial implementation of innovation is often conducted as pilot or demonstration program within particular sections of the adopting organization. This approach is often used to ensure the fit of the innovation to the particular organizational context and to judge whether the innovation is meeting its objectives. In addition to a change champion, it is useful to activate multidisciplinary implementation teams ^{23, 24} to assist in the practical aspects of embedding innovations into ongoing organizational processes. These teams can help in creating new processes, assessing intended and unintended consequences, monitoring feedback, modifying elements of the innovation as required, and carefully orchestrating the implementation rollout to obtain the support of key individuals or groups. During the initial implementation stages, “reinvention” often occurs. Reinvention is an important process in which the adopting organization essentially begins to look at the innovation as being its own. Berwick ²⁵ suggests that reinvention is necessary for organizations to achieve successful, lasting implementation of change.

Changing practice takes a considerable amount of individual and organizational effort, and that effort must be sustained over time. Most end users will need assistance in taking abstract knowledge, and even prototype interventions, and applying them in their particular contexts. Many users must initiate organizational and process changes before research findings can be consistently adopted and implemented throughout their organizations. For example, in order for change to occur, the organization's practice culture needs to support action and facilitate appropriate change processes. ²⁶ Supporting action and change requires coordination, communication, and dissemination of information. The benefits of being open to change must be made evident throughout the organization.

External support for change may come from the knowledge transfer teams described earlier. These teams can have important continuing roles during the adoption and implementation stages of the transfer process. Knowledge transfer teams can offer expert assistance to the implementation teams in the adopting organizations, helping them to identify and address specific institutional and professional barriers to adopting the innovation. They can assist the implementation team in translating theoretical research findings or previously developed prototype interventions into organization-specific structures, operating procedures, training, and information tools.

3c and 3d. Confirmation, adaptation, and institutionalization

When sufficiently visible, and with support from the necessary authorities, pilots can become an internal dissemination agent for the larger organization. Early pilot successes—for example, when the innovation alleviates the problem addressed—can create their own momentum, inspiring other units within the organization to follow suit. This “internal” spread depends on key personnel in other units being aware of the issues addressed, the innovation being tested, the results of the pilot, and the implications for their units if the innovation is successful.

If an innovation is successful and it is adopted within the relevant units of an organization, the decision to sustain the innovation usually goes through confirmation and routinization stages. The confirmation process involves key personnel affirming the decision to adopt the innovation, given all the information available and the effort required to actually implement and sustain the innovation. Efforts to sustain an innovation include dealing with changes in personnel responsibilities and practices, processes, and, perhaps, power structures. Discontinuance of an innovation, due to lack of attention to these matters, is not uncommon. ²⁷ Backsliding and slippage are commonly suffered by innovations that have been initially adopted, but not continually supported from within the organization, as well as by its external environment. The length and complexity of the confirmation process depends on the nature and scope of the innovation. Generally, it lasts until the innovation enters the routinization or institutional stage.

Once key organizational personnel commit to the innovation, the innovation is incorporated into the structure of the organization. This is the routinization stage. ²⁸ Personnel and practice responsibilities are assigned, and policies, procedures, rules, regulations, and job descriptions are developed. The change is no longer an innovation; it is the standard for the organization. At this point, tracking the experience of the organization relative to the original problem addressed provides confirmation that improvement has occurred and it can also provide the organization with an example that its effort to change has been rewarded, improvements have been made, and are being sustained. The monitoring could lead to the question of how further improvements could be made in a continuous improvement environment.

As institutionalization proceeds, adopting institutions can serve as agents for further dissemination by sharing their experiences and lessons learned. Continuing knowledge transfer teams may provide a venue for collecting feedback from early adopters that can serve to improve the prototype model intervention packages, and reflect practical experiences in practice settings.

In the movement toward institutionalization and spread of change to other organizations, external forces, such as regulatory and credentialing organizations, can play important roles, thus necessitating attention in later stages of dissemination efforts. When there is a sufficient base of success in operational practice settings, the imprimatur of credentialing organizations, such as Joint Commission on Accreditation of Healthcare Organizations (JCAHO) and State licensure agencies for health care systems, can play a powerful role in furthering implementation. Mandates from regulatory agencies provide an important incentive needed by later adopters.