We provide six recommendations to advance policy D&I research through EPIS optimization:

(1)

Specify dimensions of a policy’s function.

(2)

Specify dimensions of a policy’s form.

(3)

Identify and define the nonlinear phases of policy D&I.

(4)

Describe the temporal roles that stakeholders play in policy D&I over time.

(5)

Consider policy-relevant outer and inner context adaptations.

(6)

Identify and describe bridging factors necessary for policy D&I success.

Recommendations 1–2 optimize EPIS by defining key dimensions of a policy so that researchers can determine which domain/construct it should occupy and to understand where policy exists within a causal pathway. Recommendations 3–4 describe how researchers can use EPIS to conceptualize policy implementation activities over time and specify which policy-relevant stakeholders are represented in domains/constructs. Recommendations 5–6 acknowledge that existing domains/constructs may be underdeveloped for considering policy D&I factors and offer guidance for researchers to advance EPIS specification. Although recommendations are illustrated through EPIS application (Fig. 1) [32, 44], we provide examples of how they can be applied to other D&I TMF. We provide hypothetical research examples to illustrate the applicability of these recommendations to global settings, across different health topics, and roles of policy in D&I efforts.

Fig. 1

Policy optimized version of the Exploration, Preparation, Implementation, and Sustainment (EPIS) framework

Recommendation 1: Specify dimensions of a policy’s function

Few D&I studies specifically investigated policy as the evidence-based thing or as a strategy to be tested. Most alluded to policy as a factor in a vaguely described outer context but did not report on its purpose. Outer contexts were described generally as the “public and broader policy context,” “community,” and “outer system level of a broader environment.” Inner contexts were more clearly defined as specific state agencies, school districts, or healthcare provider organizations. Few articles defined the domain constructs (e.g., leadership, service environment agencies, funders, advocacy groups) responsible for creating and implementing policy or who might benefit from its passage.

The first recommendation is to assess the policy’s function describing the fundamental purpose of a policy [50, 54]. Function dimensions include the following: (1) policy goal(s), (2) policy type, (3) context, and (4) capital exchanged. Specifying these attributes will help researchers determine what role(s) a policy plays in D&I success and which domain/construct it occupies. Researchers should first ask, “what is the goal or intent of this policy?” This recommendation echoes early policy implementation research which argued that correctly identifying policy goals is critical to determining whether implementation was successful [31]. Policies may aim to affect a broad or narrow scope of change or to formalize something that is already being done in practice. Policies with ambiguous goals may promote confusion around implementation activities and have little impact [31]. Researchers should review legislative documents, government and organizational strategies, press releases and news articles, conduct legal mapping studies [55], or key informant interviews to specify policy goals. A single policy may have one or multiple goals; researchers should determine which goal(s) are critical to their D&I effort. Specifying the policy goal will help clarify if the policy is the evidence-based intervention, an implementation strategy to promote adoption of an EBP/program, a mechanism (series of events that promote the success of another implementation strategy), a precondition (i.e., factor necessary to activate the mechanism), determinant (i.e., barrier, facilitator), mediator (i.e., a variable that intervenes on the relationship between the implementation strategy and outcome), or moderator (i.e., a variable that alters the influence of another implementation strategy) [43]. Lewis et al. (2018) provide a comprehensive description of these causal pathway terms, which can aid researchers in further identifying a policy’s goal [43]. Specifying the policy goal can also reveal outcomes of interest (see “Recommendation 2: Specify dimensions of a policy’s form”) from the policy D&I effort — thereby advancing new policy-relevant implementation effects beyond traditional D&I outcomes (i.e., acceptability, adoption, appropriateness, feasibility, fidelity, implementation cost, penetration, and sustainability) [56].

Researchers can then determine if the policy represents a “big P” or ‘little p’ policy type. Researchers can observe in-person or broadcasted hearings and/or document review of public policy records, policymaker meeting notes, white papers, and governmental strategies to help specify the functions of “big P” policies, although there may be many “behind the scenes” nuances to consider. Qualitative interviews with key informants may be needed to describe the function dimensions of “little p” policies if organizational documents (e.g., organizational strategy plans) are not publicly available. Researchers can investigate if/how “big P” policies turn into “little p” policies or vice versa over time.

Correctly identifying the policy goals and type will aid researchers in describing the outer and inner contexts where the policy originates and/or is implemented and potential implementation outcomes. The complexity of policymaking processes means that outer and inner contexts can be multi-level. The similarity in EPIS domain names makes this recommendation applicable to other TMFs including PRISM’s external/internal context [57] and CFIR’s outer/inner setting [49, 58]. Researchers should define all relevant policy contexts and levels to understand environmental factors that influence D&I processes. Finally, researchers need to identify the resources or capital exchanged (e.g., money, knowledge, data, training, political will) through the policy (then determine if those resources constitute a bridging factor, see “Recommendation 6: Identify and describe bridging factors necessary for policy D&I success”). Identifying the capital exchanged will help researchers understand why and when a policy is successfully implemented across multi-level contexts (i.e., “policy transfer”) [5].

Researchers need to specify a policy’s function to determine if their framework should include the policy of interest as an outer/inner context factor, bridging factor, implementation strategy, or as the innovation factor. This is critical because it will help guide researchers to hypothesize about potential contextual constructs and relationships that influence D&I processes and outcomes. An example scenario for applying Recommendation 1 is presented in Table 1.

Table 1 Example for specifying dimensions of a policy’s function (Recommendation 1)Recommendation 2: Specify dimensions of a policy’s form

Few D&I studies have investigated policy as the evidence-based intervention to be implemented or as the implementation strategy. As a result, policy developers, their decision-making processes, and policy components are infrequently defined in D&I articles. To better conceptualize policy, researchers should clearly define the policy’s form: (1) its origin and creators, (2) structural components, (3) dynamism, and (4) (un)intended outcomes. Specifying a policy’s form will reveal the structures and processes that influenced how the policy was developed and can guide empirical research measuring how specific policy characteristics influence D&I outcomes [50]. Knowledge about policy structure (i.e., what it specifically enforces) can help researchers investigate which role policy plays in a causal pathway for D&I efforts and where it should be placed in the TMF (e.g., outer/inner context, innovation factor).

Policy origin refers to how the policy was developed and the stakeholders involved in its creation. For example, was the policy developed by agency staff, an expert workgroup, via a collaborative process with the public or advocacy groups? Understanding the origin story creates transparency in the policymaking process [11] to reveal the nature of “evidence” (e.g., research vs. personal beliefs) used to inform decisions and the types of interests represented during policy development. Social network analysis can aid in identifying actors involved in the policy’s creation.

If policy is the evidence-based “thing” to be implemented, the EPIS innovation factors domain can be specified. In other TMFs, researchers can specify policy within the innovation [49, 58], evidence [59], or intervention domain [57]. EPIS’ “innovation developers” construct can help define the policy’s origin. But policies might serve another role (e.g., as a determinant), and specifying where the policy developers reside (i.e., in outer or inner contexts and whether partisanship is part of the policies’ impetus) and their networks of influence can be useful to understanding which stakeholders need to be strategically engaged in the D&I effort or be the target of D&I strategies. For example, Purtle et al. identified US state legislators as a target group involved in policy decisions that impact children’s exposure to adverse childhood events (ACEs) [60]. They found that democratic policymakers were more likely to engage with dissemination strategies that included projected lifetime costs to the public system associated with every nonfatal ACE case, while economic data did not alter republican’s engagement on this policy issue [60].

Specifying the policy structure requires asking whether the policy is enforceable or effective enough to impact implementation. Researchers should determine if the policy represents a funded or unfunded mandate, suggested guidelines, or some other structure that will impact the urgency and compliance of stakeholders. Document review of the policy itself should clarify structural components. Informational interviews with policy developers can also yield insights on policy structures.

Dynamism describes the policy’s intent and potential for permanence. Researchers should investigate if the policy has an expected lifetime (e.g., 5-year demonstration project). Time-limited policies may have temporary political/public support that diminishes over time, ultimately leading to the policy’s dissolution. For example, COVID-19 mask mandates were commonly implemented as time-limited policies that increasingly generated public backlash mounting political pressure on politicians and public health agencies to prevent mandate renewal [61]. Policies without time limitations can face competing or supporting policies over time, political pressure, or advocacy from the outer and inner contexts that influence policy longevity. Researchers can investigate a policy’s dynamism by using legal mapping methods [55] or document review including white papers, government or organizational reports, legal, news, and social media sources. The prevalence of siloed health agencies [62, 63] suggests that competing or complementary health policy implementation efforts and political support exist, and qualitative interviews can help explain how these factors impact dynamism of the focal policy. Longitudinal media analyses and public opinion survey data can reveal how support for a policy changes over time and influences its permanence.

Identifying or measuring the intended and unintended outcomes of policy implementation represents the final form dimension. Policy outcome measurement can be the primary research aim or contribute to understanding the policy D&I process. For example, Crable et al. investigated implementation strategies used by Medicaid policymakers’ to encourage substance use treatment providers to adopt EBPs during each EPIS phase [7]. Citing policy reach and fidelity outcomes from state evaluation projects helped contextualize the impact of implementation strategies used in Preparation and Implementation phases [7]. Public testimony from constituents, advocacy groups, and lobbying firms can reveal potential unintended outcomes of policy implementation for researchers to investigate. Researchers should consider whether a policy is generating upstream and downstream outcomes and across which contexts. Upstream outcomes include the use of research evidence in policymaking and the overall fit of a policy with contextual factors. Downstream outcomes include how the evidence-based policy impacts quality, access, equity, and costs — which can be measured using large population health surveys or claims data. Qualitative descriptions and quantitative measures can be used to examine policy outcomes, and this methodological area is ripe for advancement [64, 65].

In EPIS, the innovation factors domain is commonly used to examine the developers, characteristics, and fit of an EBPs but can easily be adapted to investigate policy forms. Researchers should use “innovation developers” to describe the policy’s origin story, “innovation characteristics” to reveal its structure and its dynamism, while “innovation fit” describes the (un)intended consequences of a policy and its overall fit with contextual factors (Table 2). Policy forms can similarly be specified in RE-AIM/PRISM fit considerations regarding intervention/policy components or the overarching issues domain where policy representativeness, reasons, costs, benefits, and value can be defined [57]. In CFIR, researchers can adapt the innovation domain to specify policy forms including its source (i.e., origin). Trialability, adaptability, and complexity can reveal the potential dynamism, and cost informs one outcome [58]. Regardless of TMF used, researchers should specify if policy outcomes occur in outer and/or inner contexts.

Table 2 Example for specifying dimensions of a policy’s form (Recommendation 2)Recommendation 3: Identify and define the nonlinear phases of policy D&I across contexts

Like policymaking, D&I processes are not linear [32, 59, 66, 67]. Our scoping review revealed few studies that examined D&I efforts across multiple EPIS phases. Most research focused on Implementation phase activities with little to no attention to how policy initially influenced or later modified implementation activities. Studying the nonlinear nature of policymaking and implementation processes is critical to understanding how and why evidence-based policies are adopted [21, 26].

Researchers should identify and define the nonlinear phases of policy D&I (Table 3). This process may require drawing different construct operationalizations within EPIS phases since contextual factors can yield different levels of influence and interaction over time. Researchers should identify the activities and stakeholders that characterize each D&I phase. Researchers can use EPIS phases or generic pre-, mid-, and post-implementation language to benchmark policy D&I activities. EPIS is particularly well-suited for achieving this recommendation given its temporal exploration, preparation, implementation, and sustainment phases and their dynamic relationship with other framework constructs. Researchers could integrate the use of group model building methods like causal loop diagrams to describe the role of policy over time, where reinforcing loops to indicate D&I momentum and balancing loops indicate stagnation [68]. Causal loops might vary depending on the EPIS phase in which they are proposed to occur. Mixed methods can further illuminate the stories behind causal loop diagrams to reveal contextual factors that motivated each phase.

Table 3 Example for identifying and defining the nonlinear phases of policy D&I across contexts (Recommendation 3)Recommendation 4: Describe the temporal roles that stakeholders play in policy D&I over time

Recommendation 3 highlights the need to understand how outer and inner contexts change over time, while Recommendation 4 advises researchers to specifically investigate how stakeholder roles and responsibilities in these contexts change over time. While some articles included in this review mentioned policy as a determinant of D&I efforts, they seldom described specific outer context “leadership” such as government officials charged with shaping or enforcing policy. “Interorganizational networks” of stakeholders were more frequently identified as having some distal influence over D&I processes, but their roles as implementation partners or intermediaries facilitating implementation efforts were not discussed. Several articles focused on the role that inner context “leadership” played in prioritizing and directing implementation efforts. Fewer articles addressed the role of stakeholders’ “individual characteristics” influencing implementation efforts.

Stakeholders involved in policy D&I efforts can enter, exit, and change positions over time. Researchers should document these positions, responsibilities, and movements in their framework to understand who is making decisions about policy development, dissemination, and implementation. Researchers can start by identifying the outer or inner context “leadership.” In addition, the role of outer context “interorganizational networks,” “advocacy groups,” “clients/patients,” and inner context frontline implementers as well as “intermediaries” who support the implementation of policy across contexts should also be considered. Some stakeholders will be involved throughout the entire policy lifetime (e.g., exploration, preparation, implementation, sustainment) or during time-limited phases where they make strategic contributions. Researchers should optimize their TMF to conceptualize the influence of all relevant stakeholders across outer and inner contexts and to determine if they serve in a bridging factor role, such as an “intermediary” aiming to align outer and inner contexts to promote policy implementation. EPIS includes multiple constructs describing stakeholders across domains, enabling researchers to capture how these roles change over time. If using other TMF, we recommend detailing individuals involved [58], specifically who is facilitating [59] policy D&I processes and the representativeness of stakeholders [57].

To conceptualize stakeholders’ roles over time, researchers can draw multiple time-bound versions of their EPIS framework. For example, researchers can specify stakeholder roles and responsibilities in outer and inner contexts, or as bridging factors during the exploration phase, and then re-specify those roles for the preparation phase to see which elements changed over time. Researchers can use multiple data collection methods to identify stakeholders including policy and meeting document review, social network analysis, ethnographic observation, stakeholder surveys, and qualitative interviews. Snowball sampling techniques [53] can reveal unexpected stakeholders across phases. An example scenario for applying Recommendation 4 is provided in Table 4.

Table 4 Example for describing the temporal roles that stakeholders play in policy D&I over time (Recommendation 4)Recommendation 5: Consider policy-relevant outer and inner context adaptions

TMF should guide the translation of research into policy and practice and elucidate and explain the relationship between contextual determinants, D&I strategies, and outcomes [33, 36]. Existing TMF present an incomplete organization of factors that impact policy D&I. Very few studies in the EPIS scoping review examined how specific policymakers (i.e., not just “leadership”), political institutions (i.e., polity structures), and politics played a role in D&I efforts. Most articles in the scoping review used a fraction of the EPIS constructs within outer and inner contexts, bridging factors, and innovation factors domains. Some articles did make adaptations to outer and inner contexts (Additional File 4). We argue that researchers should incorporate and define new policy-conscious constructs, as needed, to better understand the studied context or test new hypotheses about policy D&I processes, relationships, and causal pathways (Table 5). However, researchers should be careful not to include an unwieldly number of constructs that hinders meaningful investigation of the relationships between each.

Table 5 Example for considering policy-relevant outer and inner context adaptions (Recommendation 5)

Researchers should conduct literature reviews and speak with stakeholders in the study setting to identify relevant TMF adaptations that are necessary to conceptualize policy and guide empirical research. Potential adaptations to EPIS’ outer context include adding constructs like “political support” (to address partisanship), “societal stigma” (toward an issue or population targeted by the policy), “workforce capacity” (if implementing a policy that impacts provider responsibilities), and “news and social media attention” (which can sway societal and political support for a policy). Researchers should consider inner context adaptations which can include defining an organization’s “local service environment” (to describe how the existing service array might change due to policy D&I efforts). Adapting EPIS and other TMF to include relevant contextual influences helps to reveal new relationships between D&I strategies, mediating and moderating factors, and mechanisms that produce both desired and unintended outcomes [43].

Recommendation 6: Identify and describe bridging factors necessary for policy D&I success

Bridging factors represent structures, relationships, intermediaries, and processes that support outer-inner context alignment, policy transfer, and D&I success [44, 50, 69]. Like stakeholders, bridging factors may be omnipresent throughout all phases of dissemination or implementation or have a time-limited role [7, 50]. Although “bridging factors” language is specific to EPIS, these alignment enhancing factors can be conceptualized as domain-spanning linkages in other D&I TMF (e.g., boundary spanners that work across contexts to promote implementation outcomes). Recent research describes how contracts [69] and renegotiated reimbursement rates [7] between government agencies and clinical service providers are formal structures that function as bridging factors. Relational ties, like partnerships between government agencies and provider organizations, can also represent bridging factors. Stakeholders (e.g., lobbyists, consultants, advocates) who support the passage of a policy in the outer context and its implementation in the inner context serve in bridging factor roles [11, 50]. Researchers should investigate personal (e.g., financial) and professional (e.g., influence) gains individuals receive from serving as a bridging factor. Data and information sharing processes between outer and inner context entities (e.g., measurement-based care reporting) can also serve as bridging factors to promote cross-context alignment [69, 70] or policy transfer. Despite the important role bridging factors serve in achieving D&I success, their functions and forms are significantly understudied, and few studies in the scoping review enhanced our knowledge of their capacity to activate change.

Researchers should investigate and describe the presence or absence of necessary bridging factors for policy D&I success (Table 6). Such research would augment knowledge about how “big P” policies transfer from the outer to the inner context, how inner context “little p” policies are spread to the outer context, and how policies can diffuse across contextual levels [71,72,73]. Researchers can use qualitative methods to ask key informants about the nature and utility of structures, relationships, intermediaries, and processes supporting outer-inner context alignment and policy transfer processes. Snowball sampling techniques and social network analyses can help identify intermediaries and relational ties critical to policy implementation. Asking questions about how evidence is used to inform policymaking or how a policy is implemented can reveal when formal structures or processes serve as bridging factors.

Table 6 Example for identifying and describing bridging factors necessary for policy D&I success (Recommendation 6)