We used a pragmatic combination of processes for framework modification during our evaluation activities in the CAPE programme [22]. The CAPE programme included a range of interventions: seed funding for academic and policy professional collaboration in policy-focused projects, fellowships for academic placements in policy settings, or for policy professionals with HEI staff, training for policy professionals, and a range of knowledge exchange events for HEI staff and policy professionals. We modified the SPIRIT framework through iterative processes shown in Table 1, including reviews of literature; consultations with HEI staff and policy professionals across a range of policy contexts and geographic settings in England, through the CAPE programme; and piloting, refining and seeking feedback from stakeholders in academic-policy engagement.

Table 1 Processes to modify the SPIRIT Action Framework for academic-policy engagement interventionsFindings

A number of characteristics of the original SPIRIT framework could be applied to academic-policy engagement. While keeping the core domains, we modified the framework to capture dynamics of engagement at multiple academic and policy levels (individuals, organisations and system), extending beyond the original unidirectional focus on policy agencies’ use of research. Components of the original framework, the need for modifications, and their corresponding action-oriented implications are shown in Table 2. We added a new domain, ‘Impacts and Sustainability’, to consider transforming and enduring aspects at the engagement interface. The modified action framework is shown in Fig. 1.

Table 2 Components of the original and modified SPIRIT action framework with corresponding action-oriented implications of the modificationsFig. 1

SPIRIT Action Framework Modified for Academic-Policy Engagement Interventions (SPIRIT-ME), adapted with permission from the Sax Institute. Legend: The framework acknowledges that elements in each domain may influence other elements through mechanisms of action and that these do not necessarily flow through the framework in a ‘pipeline’ sequence. Mechanisms of action are processes through which engagement strategies operate to achieve desired outcomes. They might rely on influencing factors, catalysts, an aspect of an intervention action, or a combination of elements

Identifying relevant theories or models for missing elementsCatalysts and capacity

Within our evaluation of academic-policy interventions, we identified a need to develop the original domain of catalysts beyond ‘policy/programme need for research’ and ‘new research with potential policy relevance’. Redman et al. characterised a catalyst as “a need for information to answer a particular problem in policy or program design, or to assist in supporting a case for funding” in the original framework (p. 149). We expanded this “need for information” to a perceived need for engagement, by either HEI staff or policy professionals, linking to the potential value they perceived in engaging. Specifically, there was a need to consider catalysts at the level of individual engagement, for example HEI staff wanting research to have real-world impact, or policy professionals’ desires to improve decision-making in policy, where productive interactions between academic and policy stakeholders are “necessary interim steps in the process that lead to societal impact” ( [24], p. 214). The catalyst domain expands the original emphasis on a need for research, to take account of challenges to be overcome by both the academic and policy communities in knowing how, and with whom, to engage and collaborate with [25].

We used a model proposing that there are three components for any behaviour: capability, opportunity and motivation, which is known as the COM-B model [26]. Informed by CAPE evaluation activities and our discussions with stakeholders, we mapped the opportunity and motivation constructs into the ‘catalysts’ domain of the original framework. Opportunity is an attribute of the system that can facilitate engagement. It may be a tangible factor such as the availability of seed funding, or a perceived social opportunity such as institutional support for engagement activities. Opportunity can act at the macro level of systems and organisational structures. Motivation acts at the micro level, deriving from an individual’s mental processes that stimulate and direct their behaviours; in this case, taking part in academic-policy engagement actions. The COM-B model distinguishes between reflective motivation through conscious planning and automatic motivation that may be instinctive or affective [26].

We presented an early application of the COM-B model to catalysts for engagement at an academic conference, enabling an informal exploration of attendees’ subjective views on the clarity and appropriateness, when developing the framework. This application introduces possibilities for intervention development and support by highlighting ‘opportunities’ and ‘motivations’ as key catalysts in the modified framework.

Within the ‘capacity’ domain, we retained the original levels of individuals, organisations and systems. We introduced individual capability as a construct from the COM-B model, describing knowledge, skills and abilities to generate behaviour change as a precursor of academic-policy engagement. This reframing extends the applicability to HEI staff as well as policy professionals. It brings attention to different starting conditions for individuals, such as capabilities developed through previous experience, which can link with social opportunity (for example, through training or support) as a catalyst.

Engagement actions

We identified a need to modify the original domain ‘engagement actions’ to extend the focus beyond the use of research. We added three categories of engagement actions described by Best and Holmes [27]: linear, relational, and systems. These categories were further specified through a systematic mapping of international organisations’ academic-policy engagement activities [5]. This framework modification expands the domain to encompass: (i) linear ‘push’ of evidence from academia or ‘pull’ of evidence into policy agencies; (ii) relational approaches focused on academic-policy-maker collaboration; and (iii) systems’ strategies to facilitate engagement for example through strategic leadership, rewards or incentives [5].

Outcomes

We retained the elements in the original framework’s ‘outcomes’ domain (instrumental, tactical, conceptual and imposed), which we found could apply to outcomes of engagement as well as research use. For example, discussions between a policy professional and a range of academics could lead to a conceptual outcome by considering an issue through different disciplinary lenses. We expanded these elements by drawing on literature on engagement outcomes [28] and through sense-checking with stakeholders in CAPE. We added capacity-building (changes to skills and expertise), connectivity (changes to the number and quality of relationships), and changes in organisational culture or attitude change towards engagement.

Impacts and sustainability

The original framework contained endpoints described as: ‘Better health system and health outcomes’ and ‘Research-informed health policy and policy documents’. For modification beyond health contexts and to encompass broader intentions of academic-policy engagement, we replaced these elements with a new domain of ‘Impacts and sustainability’. This domain captures the continued activities required in achievement of desirable outcomes [29]. The modification allows consideration of sustainability in relation to previous stages of engagement interventions, through the identification of beneficial effects that are sustained (or not), in which ways, and for whom. Following Borst [30], we propose a shift from the expectation that ‘sustainability’ will be a fixed endpoint. Instead, we emphasise the maintenance work needed over time, to sustain productive engagement.

Influences and facilitators

We modified the overarching ‘Policy influences’ (such as public opinion and media) in the original framework, to align with factors influencing academic-policy engagement beyond policy agencies’ use of research. We included influences at the level of the individual (for example, individual moral discretion [31]), the organisation (for example, managerial practices [31]) and the system (for example, career incentives [32]). Each of these processes takes place in the broader context of social, policy and financial environments (that is, potential sources of funding for engagement actions) [29].

We modified the domain ‘Reservoir of relevant and reliable research’ underpinning the original framework, replacing it with ‘Reservoir of people skills’, to emphasise intangible facilitatory work at the engagement interface, in place of concrete research outputs. We used the ‘Promoting Action on Research Implementation in Health Services’ (PARiHS) framework [33, 34], which gives explicit consideration to facilitation mechanisms for researchers and policy-makers [13]. Here, facilitation expertise includes mechanisms that focus on particular goals (task-oriented facilitation) or enable changes in ways of working (holistic-oriented facilitation). Task-orientated facilitation skills might include, for example, the provision of contacts, practical help or project management skills, while holistic-oriented facilitation involves building and sustaining partnerships or support skills’ development across a range of capabilities. These conceptualisations aligned with our consultations with facilitators of engagement in CAPE. We further extended these to include aspects identified in our evaluation activities: strategic planning, contextual awareness and entrepreneurial orientation.

Piloting and refining the modified framework through stakeholder engagement

We piloted an early version of the modified framework to develop a survey for all CAPE programme participants. During this pilot stage, we sought feedback from the CAPE delivery team members across HEI and policy contexts in England. CAPE delivery team members are based at five collaborating universities with partners in the Parliamentary Office for Science and Technology (POST) and Government Office for Science (GO-Science), and Nesta (a British foundation that supports innovation). The HEI members include academics and professional services knowledge mobilisation staff, responsible for leading and coordinating CAPE activities. The delivery team comprised approximately 15–20 individuals (with some fluctuations according to individual availabilities).

We assessed appropriateness and utility, refined terminology, added domain elements and explored nuances. For example, stakeholders considered the multi-layered possibilities within the domain ‘capacity’, where some HEI or policy departments may demonstrate a belief that it is important to use research in policy, but this might not be the perception of the organisation as a whole. We also sought stakeholders’ views on the utility of the new domains, for example, the identification of facilitator expertise such as acting as a knowledge broker or intermediary; providing training, advice or guidance; facilitating engagement opportunities; creating engagement programmes; and sustainability of engagement that could be conceptualised at multiple levels: personally, in processes or through systems.

Testing against criteria for useful action framework

The modified framework fulfils the properties of a useful action framework [20]:

(1)

It has a clearly articulated purpose: development and evaluation of academic-policy engagement interventions through linear, relational and/or system approaches. It has identified loci for change, at the level of the individual, the organisation or system.

(2)

It has been informed by existing understandings, including conceptual work of the original SPIRIT framework, conceptual models identified from the literature, published empirical findings, understandings from consultation with stakeholders, and evaluation activities in CAPE.

(3)

It can be applied to the development, implementation and evaluation of targeted academic-policy engagement actions, the selection of points for intervention and identification of potential outcomes, including the work of sustaining them and unanticipated consequences.

(4)

It provides a structure to build knowledge by guiding the generation of hypotheses about mechanisms of action in academic-policy engagement interventions, or by adapting the framework further through application in practice.