INTRODUCTION
Knowledge translation
Emergency medicine is a rapidly evolving field with consistently emerging practice-changing evidence. A critical issue in our field is that advances in research are often slow in making their way into practice.Reference Gaddis, Greenwald and Huckson 1 Impacts demonstrated by science often do not result in changes in bedside care, or in improving patient outcomes. Implementation science is the study of methods to promote the uptake of research findings in routine health care in clinical, organizational, or policy contexts. 2 In Canada, this is often known as knowledge translation, or KT.
Choosing Wisely Canada
Emergency physicians are being asked to be more efficient and effective in the care that they provide. Campaigns such as Choosing Wisely CanadaReference Cheng, Campbell and Chartier 3 are dedicated to highlighting the numerous areas where there are opportunities to eliminate low-value care. Some would argue that the greatest challenge facing healthcare is not generating new knowledge through research but identifying optimal methods of physician and system-level uptake when evidence is ready for implementation.
Implementing new practice
Barriers to implementation and achieving enduring change in clinical practice exist at the patient, provider, departmental, and institutional levels. Implementation is contingent on three key sources of change, which come with their own unique challenges: knowledge, attitudes, and behaviour.Reference Tavender, Bosch and Fiander 4 Knowledge acquisition can be challenging because of the volume of information, the time needed to stay informed, and the barriers to accessing educational resources. Because implementation is ultimately a local activity, an emergency department (ED) must identify its own knowledge and behaviour gaps. Robust data collection and monitoring to demonstrate a practice gap are central to changing attitudes, but these are also labour intensive. Finally, changing behaviour is difficult to achieve, with inertia and external pressures promoting the status quo.
While implementation science has gained traction as a focus of research, implementation occurs within a local and political context that should take into account local barriers and enablers for change.Reference Tavender, Bosch and Fiander 4 Department or system-wide change requires champions and committed hospital resources. New protocols ideally incorporate patient preferences or expectations as well as cost-effectiveness, medicolegal, or regulatory considerations.Reference Lang, Wyer and Haynes 5
Aim
Evidence-based emergency medicine care requires evidence-based implementation strategies. Recognizing the challenging environment unique to emergency medicine, the working group’s aim was to develop recommendations for the robust and reliable implementation of evidence-based practices in Canadian EDs.
METHODS
The Canadian Association of Emergency Physicians (CAEP) working group consisted of five practicing emergency medicine physicians who are active implementation scientists, a clinical nurse implementation scientist, and a medical student.
There were three parts to this project. Firstly, we conducted a systematic review of implementation strategies to change healthcare provider behaviour in the ED setting. Secondly, we conducted a national survey of emergency physicians to gain insight into different strategies used to plan, implement, and evaluate new interventions within the ED. Thirdly, we summarized our findings into a list of draft recommendations that were presented at the national CAEP Conference 2017 and online for public comments through social media.
Systematic review
A systematic review of literature was conducted (by author JC) to identify, appraise, and synthesize the best available evidence on implementation strategies that have been evaluated in ED settings to change healthcare provider behaviour. The review will be published in full separately. We included studies that evaluated implementation strategies as described in the Cochrane Effective Practice and Organization of Care (EPOC) Taxonomy. 6 The outcomes of interest included change in provider behaviour or change in patient behaviour. The search strategy was developed in collaboration with a library scientist for the following databases: MEDLINE, CINAHL, Embase, and Cochrane CENTRAL for primary studies complemented by a manual search of the past 5 years of journals, Annals of Emergency Medicine, CJEM, and Implementation Science.
Online survey
An online questionnaire (online Supplement 1) was sent to all CAEP members on January 24, 2017, with two further reminders. An additional, personalized invite reminder was sent from the authors to each non-responding CAEP member who had registered himself or herself with CAEP as an emergency medicine leader, educator, or researcher. Only respondents who answered that they had ED implementation experience or had been a research study primary investigator were asked to complete the survey.
The questionnaire branched into focused lines of questioning depending on how the responders identified themselves. Educators and clinicians were asked about methods for communicating new strategies, changing physician behaviour, identifying the barriers to change, and measuring the impact of the change. Leaders answered questions about the implementation techniques that they used, which techniques they recommended, the importance of timely implementation, and the effectiveness of implementation strategies. Researchers were questioned about the ways that they communicated their study findings and how they changed local physician behavior based on the results of their research findings.
Categorical responses were summarized using proportions, and free text responses underwent thematic analysis by two group members (NK and KdW).
Development of CAEP KT Recommendations
The KT working group met weekly between April and June 2017 to discuss the key findings of the survey and systematic review, and to draft an initial 10 recommendations through an iterative process. The systematic review, survey results, and draft list of recommendations were presented to an audience of peers at the 2017 conference of the CAEP Academic Symposium in Whistler, British Columbia. They were simultaneously posted online (www.canadiem.org/knowledge-translation-recommendations), and comments from the public were encouraged through the Fluid SurveysTM interface. The working group used this feedback to derive the final list of recommendations.
RESULTS
Systematic review results
The results of the systematic review can be found in the online Supplements 2 and 3 and included 33 studies.Reference Auleley, Ravaud and Giraudeau 7 – Reference Goyal, Fein and Badolato 39 Most included studies were randomized controlled trials (n=32) conducted in urban academic emergency practice settings (n=27) in the United States of America (n=21) and published between 2000 and 2017 (n=33).
A total of nine different implementation strategies were used; reminders (manual or computerized interventions that prompt healthcare workers to perform an action) (n = 22), educational meetings (courses, workshops, or other educational meetings) (n = 15), educational materials (distribution of educational materials to individuals or groups to support clinical care) (n = 11), and clinical practice guidelines (systematically developed statements to assist healthcare providers and patients to decide on appropriate healthcare for specific clinical circumstances) (n = 7) were the most common implementation strategies used. Most studies (20/33) included more than one implementation strategy, and some type of educational strategy was included in the majority of these multifaceted interventions (17/20).
Primary outcomes in the included studies incorporated a range of healthcare provider level outcomes (e.g., proportion of X-rays ordered, knowledge, use of clinical practice guideline), patient level outcomes (e.g., proportion of patients referred to the stroke unit, patient satisfaction), and health system level outcomes (e.g., level of stay, time to disposition decision).
Survey results
In all, 455 emergency physicians replied to the survey out of a total CAEP membership of 2,500 (response rate of 18%). Only 301 responders had experience in implementation or were study primary investigators; these included 119 clinicians, 76 leaders, 59 educators, and 47 researchers (see online Supplement 4 for details of responders).
Eighty percent (61/76) of ED leaders felt that timely practice change after the publication of new guidelines was very important, and 70% (53/76) thought that implementation efforts in their department were “somewhat effective.” Most respondents (148/240) felt that it was difficult to measure the impact of an implementation strategy.
Eighty-five percent (33/39) of researchers thought that journals had a role to play in the dissemination of research study results, over and above the publication of a manuscript. Thirty-one percent of researchers (12/39) reported a journal where they had published had aided in knowledge dissemination of the findings, for example, through social media, the press, or broadcasting of interviews.
The most common methods of communicating new guidelines or pathways were ED rounds (departmental educational meetings) (262/301) and emails (228/301). Other routes included presenting at multidisciplinary rounds (116/301) and meetings with hospital administration (85/301). A few reported publishing in local news media (34/301), using Facebook or Twitter (49/301), links to blogs (32/301), promoting a new app (41/301), or using infographics (65/301).
Free text analysis resulted in the identification of 13 themes, outlined in Table 1. The coding scheme that relates each code to a theme can be found in the online Supplement 5.
Table 1 Themes derived from the CAEP emergency physician survey
Feedback from the CAEP academic symposium
The draft recommendations, feedback, and edited final recommendations are outlined in Box 1 and Supplementary Material 6.
Box 1 CAEP Knowledge Translation Recommendations
∙ Before implementing change, ensure your physician group understands the evidence, acknowledges a gap in local practice and is willing to change. Agree on group-level targets before an implementation program starts.
∙ Consider following guidelines rather than implementing an intervention based on a body of work that has not been adequately validated. Both premature and delayed implementation can have negative consequences.
∙ Practice changing interventions should target clinical problems where care can be significantly improved.
∙ Develop implementation strategies that address local circumstances.
∙ Effective practice change interventions may include, but are not limited to, institutional support, local champions, standardized order sets, education, audit and feedback.
∙ Involve all relevant stakeholders - including physicians, nurses, patients, administrators, and allied health care providers. A team based approach to developing, integrating, and supporting an implementation strategy is the key to success.
∙ When possible, measure the impact of the implementation strategy by tracking compliance, following directly and indirectly related patient-oriented outcomes, and investigating unanticipated consequences.
∙ Make successful interventions sustainable by ensuring that the relevant business plan provides adequate resources and that key tasks are incorporated into relevant job descriptions, roles, and responsibilities.
∙ Formal ED implementation studies/trials should be rigorously designed and include assessments of individual provider, patient and system level barriers prior to designing and implementing the intervention.
∙ Protocols of ED implementation studies should be registered in an appropriate
SUMMARY OF RECOMMENDATIONS
Recommendation 1: Before implementing change, ensure that your physician group understands the evidence, acknowledges a gap in local practice, and is willing to change. Agree on group-level targets before an implementation program starts.
At a basic level, the group of clinicians involved in implementing a change of practice must be convinced by the evidence supporting it. This will involve an examination of local data and available published guidelines and/or best practice evidence. Physician buy-in at the start increases uptake of the new intervention. Group agreement on the aim of the implementation and the outcomes that they will work towards will help facilitate change. For example, a group may decide that an intervention to improve the diagnosis of pulmonary embolism should result in fewer missed cases of pulmonary embolism.
Recommendation 2: Consider following guidelines rather than implementing an intervention based on a body of work that has not been adequately validated. Both premature and delayed implementation can have negative consequences.
Novel research with dramatic results or in particularly interesting areas of practice can often be implemented prematurely. This can be problematic because findings from these studies may eventually turn out to be an overestimation of the reported effect or even incorrect. We recommend following expert guidelines on the topic rather than implementing an intervention that has not yet been validated in more than one trial. Further, a delay in implementing evidence-based guidelines may deprive patients of beneficial diagnostic or therapeutic options. Therefore, a balance is needed so that new evidence is implemented soon after validation has occurred or the evidence has been adopted by expert guidelines.
Recommendation 3: Practice changing interventions should target clinical problems where care can be significantly improved.
Interventions require effort and investment and should be targeted to have the most impact. It is necessary to review current local practice (for example, by chart review) to identify where practice diversifies from the best evidence. Any intervention should be aimed at reducing these specific gaps between the evidence and clinical practice.
Recommendation 4: Develop implementation strategies that address local circumstances.
Local culture and circumstance can play a significant role in the success or failure of an implementation initiative. Although it may be tempting to replicate implementation strategies that have been shown to be successful at other sites, consideration must be given to the environment in which the implementation is occurring with modifications to address the local barriers to change.
Recommendation 5: Effective practice changing interventions in the ED setting may include, but are not limited to, institutional support, local champions, standardized order sets, education, audit, and feedback.
Implementation strategies composed of single interventions can be easily ignored and are rarely successful. To maximize the likelihood of adoption, consideration should be given to strategies for communicating, facilitating, and tracking the desired change at multiple levels. ED workflow should not be disrupted by the new procedures and should be provided with adequate support to facilitate the change. Constructive feedback on both compliance and clinical outcomes is critical to ensure continued buy-in.
Recommendation 6: Involve all relevant stakeholders, including physicians, nurses, patients, administrators, and allied healthcare providers. A team-based approach to developing, integrating, and supporting an implementation strategy is the key to success.
Implementation strategies have the greatest potential for adoption and sustainability if all of the relevant stakeholders are involved from the beginning of the project. The degree of involvement and the number of participants will depend on availability and resources but, at the bare minimum, should include the stakeholders whom are most directly impacted by the project. This engagement achieves multiple objectives, including buy-in from the stakeholders, the identification of local barriers and facilitators, the incorporation of differing values/perspectives, the development of novel local solutions, and the identification of local champions.
Recommendation 7: When possible, measure the impact of the implementation strategy by tracking compliance, following directly and indirectly related patient-oriented outcomes, and investigating unanticipated consequences.
To determine the impact of the implementation strategy, it is critical to measure both process and outcomes measures. Process measures, such as compliance with the intervention, are important to ensure that improved outcomes are a function of a change in process and identifies when the implementation strategy is achieving adequate penetration. Compliance is best measured by mapping out the process of care step by step and a priori defining a simple measure of compliance at each step. Measurement of patient-oriented outcomes (rather than surrogate outcomes) ensures the intervention results in improved clinical outcomes. Defining patient-oriented outcomes is best achieved by rank ordering the outcomes based on importance and ease of capture or feasibility. In other words, at the top of the list would be items that are critically important and simple to capture. Measuring negative or unanticipated consequences is important to ensure that deleterious consequences are not occurring.
Recommendation 8: Make successful interventions sustainable by ensuring that the relevant business plan provides adequate resources and that key tasks are incorporated into relevant job descriptions, roles, and responsibilities.
One of the challenges with successful interventions can be sustaining the implemented change. Without a plan, the new intervention will slowly stop being used. To enable continual and persistent change, the project should have a business plan where longitudinal roles are clearly defined. To ensure sustainability, different aspects of the intervention are incorporated into appropriate personnel roles (frequently clinicians, registered nurses, administrators, and business clerks). This makes continued behaviour change an attractive option because the workload is dispersed and the process is automated.
Recommendation 9: Formal ED implementation studies/trials should be rigorously designed and include assessments of individual provider, patient, and system level barriers prior to designing and implementing the intervention.
There is a clear need for high quality, rigorously designed implementation science trials that specifically address the unique challenges in the ED. These studies should use an evidence-based approach to developing their implementation strategy, for example, by evaluating pre-existing barriers in multiple EDs, before developing an implementation strategy that targets these barriers.
Recommendation 10: Protocols of ED implementation studies should be registered in an appropriate database. Registration information should outline the rationale for the intervention components and assumptions about how the implementation strategy will work.
Registering of implementation science clinical trials (i.e., clinical trials.gov) is an important step to minimize the risk of publication bias, selective reporting of results and to ensure the transparency of the research protocols.
CONCLUSION
We have made recommendations for future ED implementation strategies, based on a systematic review of existing literature and the experience of Canadian emergency physicians who have lead ED practice change. The recommendations cover the identification of the evidence-practice gap, planning of the implementation strategy, measurement of outcomes, and planning sustainability of the intervention. Future high quality ED implementation research is needed to clearly identify optimal implementation strategies.
Acknowledgements: We thank Kelly Wyatt, Jennifer Artz, and Shanna Scarrow (CAEP) for their help in administering the survey and scheduling our group meetings.
Competing interests: None declared.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/cem.2017.432
