Twenty-four individuals consented to participate: 16 were from Canada, seven from the US and one from Australia (Table 1). One eligible participant declined consent due to a confidentiality agreement with their current employer. Of the eligible workshop participants contacted, 2 were not reached due to undeliverable email addresses and 33 did not respond to our email invitation. Participants were recruited until no new themes were identified; therefore, not all workshop participants were sent a study invitation. Participants worked in a variety of healthcare environments including hospitals, academic research centers, universities, government organizations, and regulatory organizations. Participants had a range of experience supporting implementation activities in healthcare environments and reported working in implementation for 1.5 to over 20 years. Of the 24 participants, 11 (46%) had completed a “Practicing Knowledge Translation” course developed by the Knowledge Translation Program at St. Michael’s Hospital, Unity Health Toronto, Canada (13). In terms of knowledge, 14 (58%) participants rated themselves as very or extremely knowledgeable or familiar with implementation theories, models and frameworks, and 13 (54%) as very or extremely confident in selecting and applying them to their work. Sixteen (67%) participants reported frequently or always selecting an implementation theory, model or framework and applying it to their work.
Barriers and facilitators to identifying and selecting implementation theories, models or frameworks
Four broad categories and 10 factors, generated from the data, influenced identification and selection of implementation theories, models and frameworks and were relevant to tool development (Figure). Illustrative interview excerpts are presented in Tables 2 and 3.
Category 1: characteristics of the individual or team conducting implementation
Factor 1: attitudes about the importance of selecting theories, models and frameworks
Participants reported having a general understanding of theories, models and frameworks and described several uses in implementation research and practice. For example, many participants found Nilsen’s 2015 taxonomy (3) was useful for defining a theory versus a model versus a framework and referred to the taxonomy when describing their similarities and differences. Some participants said their understanding was grounded in their learnings from the “Practicing Knowledge Translation” course. Others described their understanding of implementation theories, models and frameworks in terms of their clinical or health discipline, such as the Iowa Model for Evidence-based Practice to Promote Quality Care (30) which originates in the nursing field. In general, frameworks and models were described as being descriptive and useful for clarifying aspects of a complicated process. Theories were viewed as being more explicit about how certain phenomena are operating and how change might be occurring.
Participants mentioned using 28 different implementation theories, models and frameworks to inform their work (Table 4). Participants described the important role that theories, models and frameworks play in advancing implementation understanding, especially regarding planning, developing and sustaining effective interventions and implementation strategies. Some of there described uses of theories, models and frameworks included: informing the research question; justifying and organizing an implementation project; guiding the selection and tailoring of implementation strategies; helping to achieve intended outcomes; and analysing, interpreting, generalizing, or applying the findings of an implementation project. Other benefits to their use included providing a good starting point for implementation, providing a systematic or pragmatic approach for implementation, avoiding overlooking key categories or processes of implementation, and increasing methodological rigor. Participants commented on the importance of engaging in practices that are informed by theories, models and frameworks and evidence.
While all participants agreed on the utility of frameworks and models, such as the Knowledge-to-Action Cycle (20), a few were skeptical of the value of using theory to enhance knowledge of the complexity of implementation; they preferred to avoid selecting a formal theoretical approach. Others lacked experience with theory-driven implementation. A few believed that implementation practitioners may not feel the same level of “pressure” to use a theory, model or framework in their role compared to an implementation researcher.
Factor 2: knowledge of existing implementation theories, models and frameworks
Knowledge of existing implementation theories, models and frameworks and where to find them were perceived to be important. Some participants struggled to identify new theories, models or frameworks to inform their work, and identified their lack of knowledge of the breadth of options as an important barrier. Most participants favoured one or more implementation theories, models or frameworks and used them repeatedly, stating that it was easy to use what was familiar. Many did not follow an explicit process for identifying a new theory, model or framework. Access to a comprehensive repository or database of existing implementation theories, models and frameworks was perceived as helpful. Participants also suggested having at least one implementation team member with up-to-date knowledge of what implementation theories, models and frameworks exist, where to find them and their uses.
Factor 3: training related to implementation theories, models and frameworks
Participants talked about the relationship between selecting implementation theories, models and frameworks in research or practice and their training experience. For example, most participants selected theories, models and frameworks for which they received specific training. Major barriers to selection included inadequate background or research training in implementation theories, models and frameworks, and lack of training or expertise in implementation research methods or practice. Some participants spoke about the challenge of getting others (e.g., senior administrators, healthcare providers) to buy into the use of a certain theory, model or framework, especially if they were not familiar with the application of theory. Facilitators to selection included gaining appropriate training through participation in capacity building activities, such as accessing implementation workshops, conferences, coaching, mentoring, train-the-trainer approaches or communities of practice. Examples included working with someone who was formally trained on the theory, model or framework, or receiving feedback from implementation experts who used it to inform their work.
Category 2: characteristics of the implementation theory, model or framework
Factor 4: language and terminology used to describe the theory, model or framework
Language and terminology were key factors for identification and selection. Participants described the language used in implementation theories, models and frameworks as “complex”, “abstract”, “complicated” and “confusing”. In particular, the use of jargon and lack of clear construct definitions were identified as major barriers. Further, several participants struggled with overlapping constructs, and the inconsistent terms used to describe them across theories, models and frameworks. For example, the same term or definition may be used for different constructs, or different terms or definitions may be used for the same constructs. A few participants commented on the inaccurate and inconsistent use of the term theory versus model versus framework, both in research and in practice settings. This appeared to be common with theories versus frameworks (e.g., calling something a theory but referring to a framework). Facilitators included the importance of clear and concise language, and clearly-defined constructs to help differentiate among the various theories, models and frameworks.
Factor 5: fit of the theory, model or framework to the implementation project
Another key factor for identification and selection was the level of fit or appropriateness of the theory, model or framework to the implementation project. Specifically, a poor fit between the context in which the theory, model or framework was developed or had been applied, and the context of the implementation project was identified as a major barrier. For example, many theories, models and frameworks were developed for a specific condition or health behaviour and had not yet been applied in different contexts. Important aspects of the context included the research question, purpose or goal; health problem; setting; population; and level of behaviour change. Evidence that the theory, model or framework had been applied in practice in a similar context (such as relevant examples of applications in the literature) facilitated appropriate selection. Participants stated that seeing a description of the contexts in which the theory, model or framework was previously used was helpful when determining fit. Being aware of a theory, model or framework’s underlying assumptions and its limitations also informed appropriateness and applicability. Other related challenges included the interchangeability, compatibility and adaptability of implementation theories, models and frameworks. For example, some participants struggled with the trade-offs of selecting one theory, model or framework over another. Participants perceived that guidance on comparing different options would facilitate appropriate selection. Some noted that theories, models and frameworks often overlap or are highly derivative of each other, which adds to the complexity of combining more than one within an implementation project. It was deemed helpful to highlight theories, models or frameworks that fit well together, such as the research by Michie and colleagues linking Capabilities Opportunities Motivation Behaviour with the Theoretical Domains Framework (31). For others, implementation theories, models or frameworks that allowed for some modification were appealing, but participants struggled with how to modify or change aspects to improve fit while maintaining fidelity to key elements.
Factor 6: ease of use of the implementation theory, model or framework
Ease of use in practice was perceived to influence selection of a theory, model or framework. Some participants described implementation theories, models and frameworks as “not intuitive to use” and difficult to operationalize in the context of their own implementation project, even when the theory, model or framework was viewed as a relevant option. Facilitators to selection and use included existing online tools and publicly available resources, such as websites dedicated to specific theories, models or frameworks (e.g., the Consolidated Framework for Implementation Research). In terms of measurement challenges, a few participants cited a lack of relevant measures for key variables across theories, models and frameworks, as well as variability in the extent to which measures were developed to assess constructs. Participants preferred theories, models or frameworks that were “highly actionable”, “pragmatic” and “easy to operationalize” in practice, with detailed processes for the measures themselves that were compatible with their setting.
Factor 7: evidence supporting the implementation theory, model or framework
Empirical evidence of effectiveness, including strength of evidence supporting the theory, model or framework, influenced selection. Implementation theories were described as “fairly loose” and “without solid evidence” compared to theories in other scientific fields (e.g., physical sciences). Further, within a theory, model or framework, the level of evidence was perceived to be uneven across domains or specific processes. A summary of the evidence supporting a theory, model or framework, including the evidence used to create it and evidence of its effectiveness, was deemed to be an important facilitator. Participants also felt it was important that the theory, model or framework constructs and concepts had face validity and made sense in terms of the implementation research question or goal.
Categories 3 and 4
Other important barriers and facilitators mentioned by participants were related to characteristics of the healthcare environment (Category 4) and, to a lesser extent, characteristics of the implementation intervention or project (Category 3).
Availability of resources (Factor 10) within complex healthcare environments (Category 4), such as time, staffing and capacity, funding and access to data were identified as both barriers and facilitators to selection. Many participants also described a “tension” between time and robustness of implementation. For example, a lack of time to invest in the understanding and use of a theory, model or framework (e.g., competing demands or pressure to fix the problem right away) was a major barrier, while taking the time to create an implementation plan that included consideration of theories, models or frameworks at implementation onset was a facilitator. Theory, model or framework selection was also influenced by staff and stakeholder support, such as having an inadequate number of project staff available or being the sole implementation practitioner within an organization. It was deemed important to “assemble the right people at the right table” to avoid siloed practice and redundancy.
Finally, a few participants mentioned factors related to the implementation project (Category 3), such as consideration of the purpose, problem or goal and intended outcome (Factor 8). For instance, it may be inappropriate to select a theory when part of the research question or outcome of an implementation project was to further develop theory. Another relevant factor that presented a challenge to selection was the level of intervention complexity (Factor 9), including the type of intended behaviour change (e.g., individual, program, practice, policy), and the implementation stage (e.g., planning, evaluation, sustainability) for the project.
Features and functions of a decision support tool
Participants were enthusiastic and receptive to the idea of a decision support tool targeted to implementation practitioners. The following key features and functions were suggested to inform tool development. Illustrative interview excerpts are presented in Table 5.
Features or content items
Most importantly, the tool should include a comprehensive list of existing implementation theories, models and frameworks to choose from. Suggested content items included characteristics of the theories, models and frameworks matched with characteristics of the end-user’s implementation project (e.g., aim, scope and level of change). Participants suggested organizing the theories, models and frameworks according to their purpose (including their intended aim, scope and level of change) to align them with end-users’ needs. Alternatively, one participant (ID1) suggested starting with the project end goal or outcome, and reviewing theories, models and frameworks that include that outcome as a relevant construct. Many participants also suggested including the context in which the theories, models and frameworks have been applied, along with links to seminal articles and examples of real-world use. Linking the tool with seminal articles would allow end-users to see examples of what has been done, and perhaps gauge ease of use, as well as where the literature may or may not be saturated. Some participants suggested summarizing the evidence supporting each theory, model and framework to highlight those that have been validated. A few participants suggested content items related to the availability of implementation resources, such as the project timelines, number of stakeholders, guidance and team expertise, and financial support.
Participants suggested that the tool be simple and easy to use by the target end-user (i.e., implementation practitioners). They identified that it should provide the user with a modest set of key questions or prompts that start off broad and become more specific. For example, the tool could respond to the user’s input by guiding them toward more specific theories, models and/or frameworks. The tool should also be practical in that the level of content detail fits the intended tool audience and purpose. Being highly accessible through an open access web-based platform was also important. Further, accommodating a team-based approach (e.g., permitting access and use of the tool by an entire multi-disciplinary implementation team) would foster collaboration. Other suggested features included: interactive viewing or search capabilities (e.g., clicking on an interactive theory, model or framework diagram or figure for more information, or searching by key word or construct name); webinars or instructional videos led by experts on when (and how) to use the theory, model or framework; the use of “storytelling” (e.g., case studies) to increase personal connection; and built-in chat room capabilities to connect or collaborate with and receive feedback from others in the field who have experience selecting and using the implementation theory, model or framework. Finally, a few participants suggested an embedded evaluation component whereby users may consent to complete a survey to provide feedback on the tool.