Introduction/background
In implementation science, health policies are often conceptualized as outer setting factors that enable improved population health by directing funding towards the implementation of evidence-based interventions (EBIs) [
1‐
3]. The emerging research focused on
policy development strategies [
4,
5], however, suggests “evidence-based” is a less helpful construct than “evidence-informed” [
6]. In this policy development subfocus area of implementation science, scholars increasingly recognize the need for pull rather than push strategies to ethically and effectively integrate research evidence from health science into the complex realities of real-world policymaking [
7,
8].
This
evidence-informed perspective also aligns with political science models that explain policy and public administration failures. In one such model, New Public Governance [
9,
10], failures are attributed to misalignment among the sectors needed for successful policy implementation (e.g., policymaker, service delivery sector, consumer, community member). The relationship between poor sector alignment and poor EBI implementation is routinely documented in implementation science as well [
11]. Updated implementation science frameworks [
12,
13] identify the need for collaborative planning rather than top-down strategies when innovations or contexts present high complexity.
Policymaking is one such highly complex decision-making environment. Policymakers and those facilitating policy development have to navigate conflicting constituent values, short time frames, and difficulties accessing timely, relevant knowledge. Many of the questions addressed by policy have multiple causal factors and competing courses of possible action [
14]. When surveyed, policymakers note the difficulty of accessing and appropriately applying research evidence given these constraints [
15,
16]. As a result, research evidence tends to either be unused or applied in non-optimal ways, posing barriers to implementation successes and health equity.
Codesign is a framework for collaboration potentially well suited to addressing the complexity of multi-sector policymaking. The purpose of codesign is to develop a space for sense-making among individuals with different cultures, beliefs, and forms of knowledge [
17]. As articulated by thought-leaders in this field, effective bridging among sectors requires a mindset shift among policy developers from expert to facilitator. Operationally [
18], this moves the developer from being the one to gather knowledge and produce recommendations and policies, to one who convenes and creates opportunities for individuals from diverse professional and lived experience backgrounds to create shared understanding and agreement on policy direction. As noted by Evans and Terry [
19], common features of design-based models include iterative stages of divergence and convergence, with a series of phases starting with discovery or inspiration, leading to design or ideation, and followed by implementation.
The integrative approach fundamental to codesign provides a model for bringing research evidence together with system knowledge and service user knowledge in an “additive” approach to policymaking [
20]. The use of research evidence in this collaborative framework reflects
conceptual use as articulated by Carol Weiss [
21]. Sometimes referred to as enlightenment use, conceptual use describes how research influences the way policymakers and those in the policy arena think about a topic. This may result in a shift of mindset or mental model or may confirm the participants’ view of the problem in a way that enhances motivation to act [
22,
23]. This also aligns with the predictions of New Public Governance (NPG) which posits that focusing on the process of policy formation rather than a specific policy structure (e.g., funding an EBP) is expected to produce better downstream outcomes [
24,
25]. In NPG, successful policy formation can be evaluated by assessing the extent to which considered policies are scrutinized by multiple perspectives, encompass clearly defined goals and strategies, are descriptively innovative, articulate how the policy could navigate the trade-offs of complex policy problems (wicked problems), are supported by diverse policy stakeholders, and are sufficiently flexible to be adjusted over time [
9].
Codesign and similar cocreation approaches are rapidly growing methodologies in health services [
26], but the literature on the application of codesign within health policy is limited. Consequently, the claimed benefits of this approach to health-related policymaking are largely theoretical. The complex nature of policymaking suggests the application of any methodology in this area, especially one with the ambitious claims of codesign, requires thoughtful theoretical and empirical scrutiny. The dramatic rise in visibility and popularity of codesign has led to widespread adoption of the term. In a survey of public service workers (
n = 466), 90% self-reported the use of codesign [
27] but proponents of formally described codesign approaches argue that it is rare to find the skills and mindsets for codesign within the public sector [
28]. Politically, some are concerned that the use of term codesign in policy spaces is being coopted to provide participants with a false sense of ownership while policy decisions continue to be dictated by more powerful actors [
29,
30].
To advance our understanding of codesign as a policymaking strategy in health policy research, we undertook a scoping review of the health policy codesign literature. Our aim was to characterize the existing state of research in this area, analyze available approaches against the claims of guiding theory, and propose recommendations for developing and evaluating codesign as a strategy for policy formation within policy dissemination and implementation science. In doing so, we adopted a broad definition of policy, including Big p, little p, and policy-enabled service development [
31]. Big p policies are laws that regulate resources. Little p policies are institutional norms and regulations. Policy-enabled service development includes policymakers or policy levers in service innovation.
Method
Purpose
We conducted this scoping review to capture policy codesign strategies across health-related disciplines to examine disciplinary and theoretical origins, activities, and the relationships among strategies and outcomes as reported in existing studies.
Study design
Scoping reviews are a methodologically rigorous approach to describing scholarly literature on a topic of interest [
32]. We used the most recent guidance for conducting high-quality scoping reviews, drawing from foundational literature [
32] and updated methods [
33‐
35]. The review was conducted in accordance with the Joanna Briggs Institute methodology for scoping reviews [
36]. Article selection and synthesis were conducted based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist.
We systematically searched relevant published research articles in April 2021 and updated our search in November 2022. Our search involved the following databases: PubMed, Academic Search Complete, Web of Science, EBSCO (MEDLINE, PsychInfo, and CINAHL), and Google Scholar. We stopped reviewing returns from Google Scholar after the first 100 articles. Google sorts by relevance and after the first 30 articles, we did not identify any additional eligible documents. We included text words contained in the titles, abstracts, and keywords of articles. All keywords and index terms were adapted for each database and/or information source and consisted of three levels of included terms. The first level aimed to capture efforts that aspired to use cocreation methodologies using the terms “codesign, cocreation, participatory design, coproduction, and community-based participatory research.” The second level aimed to limit the search to health-related areas with the term “health.” The third level aimed to limit the search to papers regarding policy and complex systems with the terms “complex health, complex public health, policy, community, healthcare system, and system.” The reference section of articles selected for full-text review were also searched, and potentially relevant articles were added to the full-text review list as well.
Inclusion/exclusion
Included articles described an effort by non-citizen led entities (e.g., government, academia, nonprofit, advocacy) to initiate a process to improve a health-related need through policy or system-level processes. The articles needed to include sufficient information to code the steps and strategies used in the process. Articles had to describe a process that clearly focused on engaging system and policy level changes. Articles not focused on a health or public health need were excluded. We used a broad definition of health, including social determinants of health.
Procedures
Citations for all identified articles were entered in a Microsoft Excel data worksheet developed by the team for managing systematic reviews. Article titles were reviewed by two independent reviewers for inclusion or exclusion, based on their titles/abstracts until reviewers reached agreement. The remaining title/abstracts were reviewed separately.
Study selection process
All titles and abstracts were screened independently by pairs of reviewers (SW, KA, BQ) and discrepancies were resolved by the pair. For screening full text articles, one pilot was conducted with 20 articles in which each article was independently reviewed by a pair of reviewers (SW, KA, BQ, BB, MP). When the full group of reviewers reached agreement on screening criteria, full-text articles were independently reviewed by three reviewers (SW, KA, MP) and spot checked by an experienced reviewer (SW).
Data abstraction
Following Joana Briggs guidelines, we abstracted data on article characteristics (region, type of article) as well as categories for health focus, policy focus, codesign definition, referenced theory, engagement level of policy beneficiaries, multisector involvement, codesign structure (phase, description), perceived benefits, and perceived challenges. Coding within category for health focus and policy focus was developed inductively by each coder. Codes were then recoded into a smaller set of higher-level codes by a single author (SW) and approved by three reviewers (BQ, BB, MP). Coding for codesign definition, theory, engagement of policy beneficiaries, multisector involvement, codesign structure, perceived benefits, and perceived challenges was developed by the process described below.
Data abstraction was conducted using an Excel form developed a priori and pilot-tested on a sample of five papers after which codes were refined. Four reviewers then reviewed a single study (SW, BQ BB, MP), and discrepancies were resolved by consensus. Data were abstracted by one reviewer (KA, MP, SW) and verified by an experienced reviewer for consistency and accuracy (SW).
Risk of bias assessment
We did not conduct a risk of bias assessment, consistent with Joanna Briggs Institute Scoping Review Methods Manual and scoping reviews on health topics.
Synthesis of results
Results were synthesized using frequencies and thematic analysis. Thematic analysis was performed by one reviewer (SW) and verified by a second reviewer (BB). Synthesis drew from public policy health concepts [
37], community sector literature, citizen participation frameworks [
38], and open-coding methods [
39]. Consistent with the purpose of scoping reviews, our intent in conducting this study was to characterize the state of the science in the area of health policy codesign.
Discussion
We conducted this review to examine how policy codesign is being defined and operationalized in health policy scholarship. We identified a small, growing literature that is expansive in geographical reach and topic area. Health policy codesign is being implemented across continents with somewhat higher representation in the literature among Anglophone countries. The approach is being used to address diverse health topics, including social determinants (housing, economic development), public health, and health services. The maturity of the science across disciplines is at an early stage. Reported outcomes were qualitative and not consistently defined but pointed towards common areas of interest for measuring outcomes.
Efforts to enact little p policy (organizational adoption of practices) was slightly more common than big p policy or program development efforts. Studies were twice as likely to engage policy users as representatives in decision-making than as full owners or as informants. Full ownership approaches tended to be complex, requiring significant resources devoted to community mobilization [
41,
50], dialogue, and voting [
56]. Representative approaches tended to be more time-limited, using discrete engagement events such as workshops, to engage participants from multiple sectors [
19,
43,
63]. Informant approaches were focused on improving the robustness of information available to policymakers in their decision-making.
Relying on the author-reported benefits of codesign, we identified a trend association between the use of full ownership approaches and the reported level of community mobilization and knowledge of community needs. We also identified higher reported benefits in novel ideas within representative ownership approaches. Because the reported outcomes were purely descriptive, we identify this as a productive area for future research. Potential trade-offs between community mobilization and novel ideas suggest different policy codesign structures may yield different benefits and should be used to solve different types of policy problems. This aligns with the general guidance of research and policy participation frameworks [
66]. In Gupta’s split ladder of participation framework, for example, the model suggests only using complex, policy user intensive efforts when the health topic is controversial, with little agreement among sectors in values or beliefs about the relevant research evidence.
We did not attempt to examine associations between the linearity or phases of policy codesign efforts and outcomes because of the variation in approaches. We observed general trends; however, all approaches identified a phase in the process in which information had to be “put together” in some way to propose a policy solution, with these solutions coming middle to late in the policy formation process. Accordingly, a common feature noted across the articles reviewed and that spanned disciplinary approaches, was the greater amount of time required to facilitate a policy codesign effort when compared to traditional timelines. Studying the trade-offs between timeframe and the development of trusting relationships is an important area of future research. It is likely that long timeframes could hurt the acceptability or scalability of policy codesign and researchers will be motivated to find ways to shorten timeframes to improve efficiencies. Findings from the articles reviewed here suggests facilitation expertise might be a factor in reducing timeframes; on the other hand, studies cautioned that imposing urgency when forming partnerships with individuals marginalized by race, poverty, and/or disability will further harm those individuals. The intended benefits of a proposed policy should not outweigh the potential for harm among individual participants in the codesign process [
67,
68]. Distilling insights from experienced community/multi-sector facilitators and examining time-limited activities and methods from participatory design could yield a set of guiding principles for making these tradeoff decisions. Researchers studying policy codesign can assist by clearly documenting activities within phases, time devoted to phases, and participant perceptions of the codesign process using participatory process measures [
69‐
71].
Knowledge gathering and integration, possibly the area of greatest interest for implementation science, was infrequently described in meaningful detail among the reviewed studies. Traditionally, the field of participatory design and codesign has not considered the use of research evidence as a core element of knowledge synthesis. Although as formally trained designers are moving more into health services research, the use of research evidence within design engagements is becoming more prevalent [
72]. In our review, research evidence was formally presented in the policy codesign efforts of facilitator teams that came from health services research organizations [
45,
59,
65], whereas research use was less formally introduced or not introduced by facilitator teams with participatory design or public policy backgrounds. The use of research evidence within anti-racist and decolonizing movements within health services research is currently contentious; a key question for the field is how to make appropriate use of research evidence without imposing this use on communities [
73,
74]. Policy codesign, and codesign broadly, provides a promising framework and set of methods for resolving this tension.
In codesign, participants can request information to round out their view of a topic, including the research evidence, without this needing to be pre-selected or imposed by a facilitator team. In Owens [
59], for example, the facilitator team used an established systems-design model, Theory U, which does not require a formal “research evidence” component. In the project, the codesign participants requested a review of best practices in jail-based reentry for opioids and a research team conducted a rapid evidence review (RER) on the topic. The RER was translated into a 10-min video and research brief and sent back to the team to review and discuss. Interestingly, the paper notes that when asked if they “learned something new,” no one on the team noted the RER findings. However, the cross-service model developed strongly resembled existing evidence-informed recommendations while also including innovative components (e.g., peer navigation support). This and other literature suggest research evidence can be a valuable information source but should not dominate and sometimes may not be necessary (or could be harmful) when designing policies and systems to solve intractable health problems [
75,
76]. The promise of policy codesign, albeit in an infant stage, is having a framework for selecting and synthesizing information and engaging partners to create the most transformative or effective policy solution possible for that moment.
Limitations
Because we limited our review to policy efforts, we did not review a much larger literature reviewing the use of codesign in health services research and in health service program development. Review of these efforts already exists [
26,
77,
78]. We also excluded citizen-led mobilization and advocacy efforts even though these actions form an important part of the health services improvement landscape.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.