From what we know to what we do: Human resource management intervention to support mode 2 healthcare research

There is a huge volume of evidence about what is clinically effective and efficient, but this is slow to translate into front-line practice. To address the problem, we need to support clinical academics and practitioners to co-produce research and service improvement; necessitating HRM intervention. Our case study shows common purpose in mode 2 research across clinical academics and practitioners can be attained by focussing upon their professional identity, within which their status and jurisdictional autonomy are key dimensions. Our study shows how development workshops, through which control is ceded by managers to clinical academics and practitioners, are used to co-design HRM interventions to support mode 2 research. Relevant HRM interventions are first, performance management that is non-intrusive and aligns with criteria clinical academics and practitioners value. Second, job design that allows autonomy and status enhancement for clinical academics and practitioners engaging in mode 2 research.


| INTRODUCTION
Co-production of impactful research-informed intervention, or what might referred to as mode 2 research, sometimes referred to as academic-practitioner collaboration (APC) (Bartunek & Rynes, 2014), or intervention research (Radaelli et al., 2014), has been a feature of business and management research for some time and continues to inform our research practice (Bartunek, 2011). HRM researchers, however, appear slow to engage in academic-practitioner research collaborations (APCs) towards mode 2 research (Harley, 2015;Lawler & Benson, 2022;Pfeffer, 2007;Rynes, 2007). And, despite encouragement to do so, we still lack understanding of antecedents to, the process of, and practical prescriptions for, HR interventions to support mode-2 research (Guerci et al., 2019). In Hewett and Shantzs' (2021) terms, we require insight into the process of co-creation of HR interventions with stakeholders, including in particular, the influence of employees, as well as understanding of content of HR interventions to support mode-2 research. Related to our professionalised setting set out below, we set out three inter-linked research questions: What are antecedent conditions shaping prospects for academics and practitioners engaging in mode 2 research in a professionalised setting? How do such antecedents determine the content of HR interventions that prove effective in engaging academics and practitioners in mode 2 research in a professionalised setting? How might we engage academics and practitioners in co-creation of HR interventions to support mode 2 research in a professionalised setting?
Our empirical study seeks to illuminate the challenges and solutions to the co-creation of HR interventions to support mode 2 research, focussing upon a project-based organisation with fixed-term funding. Our focal project-based organisation is an APC that seeks to draw in efforts of two groups of core professional employees with different perspectives and identities (clinical academics and clinical practitioners) in translational health CURRIE and SPYRIdOnIdIS 2

Practitioner notes
What is currently known?
• Mode 2 research is encouraged in human resource management, but we know little about how it works in practice • Professionals can successfully resist managerial intervention, including that aimed to support mode 2 research What this paper adds?
• Identifies the antecedent dimensions of identity that need to be invoked to support mode 2 research amongst core professional employees • Identifies the specific HR interventions-performance management and job design-that engage core professional employees in mode 2 research • Insight into how to engage core professional employees in co-creation of HR interventions

The implications for practitioners
• Do not impose HR interventions upon core professional employees, but allow them to co-create these HR interventions to support mode 2 research • Performance management should be non-intrusive and align with criteria valued by core professional employees to support their engagement in mode 2 research in a professionalised setting • Job design should allow for autonomy and status enhancement for core professional employees to engage in mode 2 research in a professionalised setting • Training and development workshops can be used for core professional employees to co-create HR interventions to support their engagement in mode 2 research in a professionalised setting research (generating evidence about what works for patient benefit and then diffusing this into practice) (Currie et al., 2014). Our APC case represents an endeavour that itself can be characterised as invoking a mode 2 approach.
Our APC case represents a transdisciplinary endeavour with expected outcomes of research that is both sufficiently academically rigorous to give rise to high quality peer-reviewed publications and sufficiently relevant to improve frontline healthcare delivery for the benefit of patients (Guerci et al., 2019). Such demands for better care delivery underpinned by generation of scientific evidence prove challenging to enact for clinical academics and practitioners because commonly they enact more specialist roles and career paths towards care delivery or scientific research (Currie et al., 2014;Dunn & Jones, 2010;Goldstein & Brown, 1997;Yanos & Ziedonis, 2006). Following which, one can assume core professionals impact how HR interventions designed to support such mode 2 translational health research are received (Chen et al., 2022). Indeed this proved so with clinical academics and practitioners resisting the managerially imposed performance management system put in place to engage them in mode 2 research. At which point we were called in to undertake action research (Bradbury, 2008;Reason & Bradbury, 2008) to examine the problem and work with APC management to support the development of HR interventions to motivate powerful core professionals to engage in mode 2 translational health research. In essence, our case study is likely to render very visible antecedents to, the challenges of, and practical prescriptions for the process and content of HR interventions to support mode 2 research in professionalised settings in line with our research questions. Hewett and Shantz (2021) define co-production, or more accurately they use the label 'co-creation' as a "continuous process in which HR and stakeholders create value through collaborative efforts to problem-solve and innovate in the design and use of human resource interventions to help them better satisfy stakeholders' needs". They argue co-creation generates greater value than if HR managers solely designed interventions, more so in start-ups and project-based organisations, both of which are characteristic of our empirical case under study. They emphasise that HR interventions commonly evolve once introduced in a process that is dynamic and iterative by nature. Employee response to a HR intervention is likely to shape the practice itself (Kehoe & Han, 2020;Van Mierlo et al., 2018).

| LITERATURE REVIEW: MODE RESEARCH, PROFESSIONAL IDENTITY AND HRM
Thus HR interventions need to be inscribed into the interpretive schemes of employees and gain legitimacy amongst employees to be effectively implemented and add value (Cooke, 2018).
Managers remain important to diffuse HR interventions across the organisation (Beer, 1997), but the balance of agency and control across managers and employees is a crucial dimension of the process of co-creation, and around which we require insight (Hewett & Shantz, 2021), in particular more contextualised insight into HR co-creation (Mirfakhar et al., 2018). In outlining receptive conditions for co-creation of HR interventions, Hewett and Shantz (2021) emphasise the importance of collaborative use and co-design by the wide range of stakeholders subject to HR interventions, particularly employees, whose involvement and participation means they are then more committed to their implementation Budjanovcanin, 2018;Meijerink et al., 2016), otherwise an implementation gap ensues (Piening et al., 2014). Following which, Hewett and Shantz (2021) advocate managers work closely with employees, to jointly utilise their skills and experience to design HR interventions that better meet users' needs. Given goals of managers and employees are not always compatible (Guest & Bos-Nehles, 2013), and that complex organisations, such as the empirical one under study commonly tend towards bureaucracy, then relational ties across stakeholders need to be cultivated to create integrative outcomes (Gittell & Douglass, 2012).
Institutional conditions associated with context represent antecedents for prospects of HR co-creation (Hewett & Shantz, 2021). In our empirical case of translational health research, both clinical academics and practitioners are powerful professionals, and as such bring strong identities to bear upon any attempts to manage their practices. Professional identity represents "an individual's self-definition as a member of a profession" (Chreim et al., 2007(Chreim et al., , p. 1515, rather than, for example, identity derived from their organisational membership. At the core of their claim to be a professional, first, they focus on client interest, rather than pursuit of self-interest (Wright et al., 2017). Second, to be a professional is to be granted autonomy and control your own work (Freidson, 1988). This is derived from a specialised knowledge base, into which professionals have been trained and socialised, and around which they have a discrete jurisdiction (Abbott, 1988;Freidson, 1988). Professionals are thus likely to defend their jurisdiction and autonomy against changes that threaten their identity and autonomy (Abbott, 1988), in particular resist new practices they perceive as attempting to control them (Noordegraaf, 2011). As part of their efforts to counter intrusion upon their jurisdiction, professionals seek to stratify themselves from other professions and occupations to enhance their status (Abbott, 1988). Any success in co-creating HR interventions is likely to rely upon engendering a constructive relationship between managers and professionals that hold different sources of power (Ferrary, 2005). To emphasise our point, we require greater insight into the process of co-creation of HR practices in professionalised settings.
When applying this to the translational health research field, first HR intervention needs to take account of divergent career trajectories enacted within the medical field. In particular, the role of clinical academics, professionals that translate discoveries to the bedside, is a threatened one, with young doctors forced to choose between research or practising medicine in their increasingly compartmentalised careers. Then even where they do gravitate towards research, they choose basic science rather than clinical science, because it is less constrained, generates clear-cut results, and is more likely to generate peer-reviewed publications and grants (Goldstein & Brown, 1997). Where they attempt to combine clinical practice and clinical research, role holders may experience confusion about what their job really is and associated ways of being; that is, being a clinical academic represents an identity challenge for role holders (Yanos & Ziedonis, 2006). Yet, while representing a less attractive role, clinical academics represent 'bridgers' between practice and science worlds, with their clinical experience and direct exposure to clinical pheneomena or service systems they study allowing them to recognise many real world issues that inspire relevant and innovative research, create a virtuous dynamic of ideas between clinical and research domains, and facilitate dissemination and implementation of evidence-based care (Yanos & Ziedonis, 2006). Their training and development, and rewarding their mode 2 efforts, has thus increasingly been seen as crucial to ensuring translational health research for patient benefit (Goldstein & Brown, 1997). Meanwhile, for clinical practitioners, the challenge is one of operational pressures associated with constrained resource and increasing demand for clinical services, which eat into their opportunity to engage in any research, even that directly applied to improving patient care (Currie & Suhomlinova, 2006). Yet engaging in translational health research is likely attractive for clinical practitioners, if space can be carved out for them to engage, on the basis it will enhance their status amongst peers (Currie et al., 2020).
Second, linked to the above, HR intervention needs to take account of divergent performance measures applied by government agencies to higher education institutions (orientated towards generating research income and high impact peer-reviewed publication), and health care providers (orientated towards reducing waiting times and lists, and ensuring services are patient safe). Following which, narrowness of clinical academic and clinical practitioner roles are reinforced. It used to be difficult to distinguish clinical academics and practitioners, with each enacting a tri-partite mission to provide teaching, engage in research and deliver clinical care. However, in the face of divergent performance measures, workload allocation that underpins job design in their respective employing institutions means clinical practitioners are unlikely to have protected research time, and clinical academics less likely to deliver care in provider settings (Currie & Suhomlinova, 2006). Third, derived from the above, the perspectives of clinical practitioners and clinical academics upon what constitutes relevant research also diverges. Academics in healthcare value experimental methods that produce statistical generalisability, aligned with procedural assessment criteria of their peers that review grant submissions or manuscripts for prestigious journals. Meanwhile, clinical practitioners' immediate need for improving their services is more likely met by applied research considered less rigorous by academics (Currie et al., 2014).
Unsurprisingly, therefore, clinical academics' and practitioners' career paths are likely to vary once they exit their initial training together, as they become socialised into their different specialist identities (Dunn & Jones, 2010). The challenge of designing HR interventions to engage them in mode 2 research in translational healthcare research, is thus rendered more difficult because there are two groups of powerful professionals, clinical academics and clinical practitioners, whose incentives, values and work roles are increasingly divergent.
Following which, users (in our case clinical academics and practitioners) influence the implementation of HR interventions and whether intended outcomes are realised Budjanovcanin, 2018;Jansseens & Steyaert, 2009;Purcell & Hutchinson, 2007;Shipton et al., 2016;Ulrich et al., 1989). Where users 'actively' engage in co-creation of HR interventions, this generates better outcomes: it satisfies multiple users' needs simultaneously; users feel more committed to co-created HR interventions, which in turn increases their use; stronger relational ties across the HR function and its users develop; HR managers and users enhance understanding of each others' perspectives (Hewett & Shantz, 2021). Finally, the need for insight into process-related understanding of implementation of HR interventions (Guest & Bos-Nehles, 2013) demands longitudinal research, that is contextualised and acknowledges the multi-actor and complex nature of HR implementation (Budjanovcanin, 2018). We set out the dimensions of our own research design along such lines in the next section of our manuscript.

| Empirical context
We empirically focus upon an APC (EliteMed), externally funded £10mn from a national government agency for clinical academics and practitioners (primarily doctors) to engage in mode 2 research to improve services for the benefit of patients across a range of long-term conditions, such as mental health, respiratory health (COPD), diabetes (Cooksey, 2006). There were nine such APCs funded across the English National Health Service (NHS).
Our main research interaction, at least initially, was with EliteMed managers, whom directed our attention upon performance management. As set out in our empirical analysis however, subsequently, through our longitudinal study, we engaged with core professional employees, that is the clinical practitioners and clinical academics enacting translational health research, following the failure of the performance management intervention. Through our engagement in workshops with core professional employees, we drove EliteMed management and their national peers towards greater recognition that implementation of HR interventions was a multi-staged process that included and blended with co-creation of HR interventions (Hewett & Shantz, 2021;Mirfakhar et al., 2018). In our empirical analysis set out below, we provide insight into the conditions for co-creation of HR interventions that led to more effective implementation.
In our specific empirical case of EliteMed, COPD provides a benchmark for success of co-production. Regarding practical relevance, the new COPD service co-produced through the APC, showed reduced cost and improved patient care (reduction in mortality rates, in large part because patients were triaged more effectively and secondary care and primary care were better integrated), following which it diffused from its originating site to 15 other healthcare providers in the region covered by the APC. Regarding demand for academic rigour in the co-production of the new service, clinical academics successfully bid for additional large-scale funding to further evaluate the new service, associated with which they built research capacity through supervising post-doctoral and doctoral researchers, and published in high quality peer-reviewed journals.
There are three main stakeholders involved in the co-creation of HR interventions we examined: EliteMed's executive management team that developed and implemented HR interventions to support mode 2 research; clinical, mainly with a medical background, academics employed by a university medical school; clinical practitioners, mainly doctors employed within healthcare providers (hospitals, community care, primary care mental health care) geographically close to the university.

| Data collection
EliteMed managers struggled to engage clinical academics and practitioners in the APC through a performance management system they had imposed, following which we were commissioned 12 months after the start of the APC, to evaluate their model of mode 2 research, with a focus upon how HR interventions best support their aims.
In the words of EliteMed's Director, "we were to work with them to adapt HR practices if deemed necessary". Our research was both retrospective, to examine the performance management intervention they put in place at inception of the APC, and 'real time', to examine prospects for HR interventions in performance management and "any other domains of HRM we deemed relevant to support our aims" (EliteMed Director); that is, our approach represented 'action research' (Bradbury, 2008;Reason & Bradbury, 2008).
Our action research approach is one suggested as a more feasible route to mode 2 research in management settings (Kiesser & Leiner, 2012). HRM research lacks such a tradition, nevertheless action research is not uncommon (Styhre, 2004;Vashdi et al., 2007;Zhang et al., 2015); indeed action research is particularly relevant for HRM focussed upon systemic interventions in complex organisational contexts, where the introduction of HR interventions may not develop as planned (Bleijenbergh et al., 2021). We fed back to the EliteMed executive managers in a 'safe' learning space for both parties (Guerci et al., 2019). Following which, we were invited to feed back our research to clinical academics and practitioners within workshops put in place by EliteMed managers, which encouraged clinical academics and practitioners to reflect upon HR interventions to support mode 2 research, and where necessary, co-create them (Molineux, 2013).
As detailed in our empirical analysis, whilst our engagement with EliteMed managers in meetings was participative, within the workshops we 'sat back' and fed back our research to promote reflection and action by clinical practitioners and academics. Following which, our role reverted to non-participant observation as the latter drew upon our research to engage in co-creation of performance management systems, and other HR interventions (specifically job design) they felt relevant to support the mode 2 research aims of the APC. We then followed up with workshops with the core professional employees to examine the process and effect of co-creation of HR interventions and understand why they worked (or not) to support the aims of the APC for mode 2 research in the translational health research domain.
In total, we conducted 134 semi-structured interviews. All managers of the EliteMed executive team (12) were interviewed four times (48 interviews). EliteMed managers were interviewed at inception of our study, and after 12-month of our study, prior to the researchers feeding back about HR intervention failure. EliteMed managers were interviewed again following researcher feedback, 18 months into our study, capturing their reflections upon failure of their original performance management intervention, and then a final interview took place following co-creation of HR interventions with clinical academics and practitioners after 30 months of our study. Twenty five clinical academics and 18 clinical practitioners (i.e doctors) involved in co-creation of HR interventions were interviewed twice, over the lifespan of the research, around 15-16 months (commenting on the original HR interventions), and then again around 32-33 months (commenting on modified HR interventions that were co-created) into our study (in total, 86 interviews with clinical researchers and practitioners). All interviews were fully transcribed.
As participant observers, fieldnotes were generated across 4 meetings with EliteMed managers (8 h) and, as non-participant observers we collated fieldnotes from a further 56 h of workshops in which clinical academics and practitioners co-created HR intervention. Periods of observation lasted between 1 and 8 h at any time. During observations, we actively immersed ourselves in EliteMed. We became 'groupies' on the EliteMed event scene, engaging in informal conversations with various research subjects before and after workshops, and at serendipitous encounters during coffee breaks and lunches, around workshops. Notes were taken during or immediately following such observations or conversations as appropriate. These fieldwork notes were encompassed within a case study database, alongside interview transcripts and documentation, and subjected to analysis as detailed below.

| Data analysis
Our starting point for research was a problem-centred one, directed by EliteMed managers to interrogate failure of the performance intervention they had implemented in a top-down way. Following their directive to focus upon performance management, we followed an abductive logic. Abductive reasoning is characterised by constant dialog between theory and empirical findings, which involves an analytical strategy based on continuous formulation and iteration of questions and answers from literature to both focus and explain emerging findings (Mantere & Kekokivi, 2013). So, our literature review informed analysis in more deductive way, for example, identifying how HR intervention might support mode-2 research, but the influence of professional identity upon this emerged from the data, following which we perused relevant literature, that of sociology of professions, retrospectively.
Our analysis was one carried out across temporal brackets aligned with our action research, within which the level of analytical generalisation was raised (Yin, 2003). Analysing data in the first temporal bracket, we confirmed failure of the performance management intervention to support mode 2 research, following which we raised level of analysis as we sought a theoretical explanation for 'failure'. At this stage, literature about professional identity provided theoretical insight, specifically understanding the failure of the performance management intervention as one where it compromised the status and jurisdictional autonomy of clinical academics and practitioners. Our second temporal bracket was particularly action research orientated as we sought to drive reflection and action by EliteMed managers. This buttressed our understanding the performance management intervention to support mode 2 translational health research should align with professional identity. In particular, we ascertained that collegial approaches to decision-making characteristic of professionalised settings should inform a more collaborative approach to intervention. In a third temporal bracket, we followed through to observe workshops within which HR interventions were co-created with clinical academics and practitioners to support mode 2 research, and drew upon relevant literature (e.g. Hewett & Shantz, 2021) to explain its effectiveness. Following which, in our final temporal bracket we ascertained the effect of co-created HR interventions upon mode 2 research in the translational health research domain.
In our final stage of analysis, we drew our insights across the four temporal brackets together, to understand the relationship between co-creation of HR intervention across managerial and professional stakeholders and professional identity to arrive at our final theoretical interpretation as presented in our next section, the empirical presentation.
Throughout the data collection and analysis process, we emphasise the need for researchers to enact mode 1 controls over quality as they enact mode 2 research. Action research can play into managerial agendas, where it is commissioned by managers in pursuit of organisational advantage. We took care with our data collection, ensuring we interviewed a range of core professionals, less and more receptive to mode 2 translational health research, and remained wary of individuals' attribution of actions and their effect through complementing interviews with observation. More crucially perhaps, our analysis outlined above was one we sought to continually authenticate with our research subjects, specifically the core professionals subject to HR interventions (Watson, 2000).

| Failure to impose HRM interventions: The importance of professional identity
EliteMed managers initially instituted performance management that threatened the professional identity of both clinical academics and practitioners, and thwarted their interaction: Performance management is one of the major barriers to our engagement with our practitioner colleagues around this Given clinical professionals value their identity, intrusive performance management intervention was unwanted: It is all about managing your own work and being given autonomy to do that. You always have the authority to shape things the way you personally envisage them, and that can never change That clinical academics and practitioners distanced themselves from the EliteMed projects in which they participated, was reflected in their dis-engagement: It's the (EliteMed) way or the highway … they set out a template for how we should behave in the training workshops … and frankly, I said fine, take your money away. I could do without the headache.

(M.R26-Clinical Academic)
In summary, HR interventions to support mode 2 research did not align with professional identities of both clinical academics and practitioners and so they failed to engage in collaborative processes. Considering this resistance, and informed by feedback by ourselves, EliteMed managers reflected upon this failure, which we detail in the following section.

| Management reflection upon failure of imposed HRM interventions
Our investigation provided the basis for EliteMed managers to act upon the failure of HR interventions imposed upon professionals. We were invited to participate in several strategic meetings with EliteMed managers where they emphasised their need "to reflect on recent experiences and foster learning" (meeting observation, M.R5-EliteMed Manager). During these meetings, EliteMed managers, especially those dealing with the implementation of HR interventions, talked about iterative learning processes: We need to spend a lot more time perhaps as a team working through problems to learn how we can approach and how we can change what we do.

(meeting observation, M.R12-EliteMed Manager)
Managerial reflections were forward looking in scope aiming to convert their reflections about "why they [referring to clinical practitioners and academics] resisted our practices for quite a long time [into explicit action regarding] … how are we going to respond to all this?" (meeting observation, M.R1-EliteMed Manager). EliteMed managers understood it was not just the content of HR interventions that stymied collaborative processes, but that the process through which they were designed and moreover, implemented, was viewed as intrusive by professionals: They never appreciate managers interfering with their work. It is their work, and they want to manage their own work.

(M.R3-EliteMed Manager)
They don't want people, particularly managers imposing targets on them.

(M.R12-EliteMed Manager)
EliteMed Managers reflected on professionals' expectations for a high degree of control over their academic or clinical practice to protect their professional identities. Following which, "we tried to give people more freedom to shape things" (meeting observation, M.R10-EliteMed Manager).
Belatedly, working with the research team, EliteMed managers recognised "our practices are not being perceived positively" (M.R10-EliteMed Manager), and so sought to refine "how we make people inspired and eager and motivated" (meeting observation, M.R11-EliteMed Manager). In the next section we summarise our findings about the subsequent process of co-creation of HR interventions with clinical academics and practitioners.

| Management of Co-creation of HR intervention with core professionals
Following failure of the managerially imposed performance management intervention to engage clinical academics and practitioners in the APC, our research team were tasked with organising workshops to co-create HR interventions. These were re-orientated: Such a managerially-led approach had generated a "lack of trust from us about their [referring to EliteMed managers] goals and motives" (M.R17-Clinical Practitioner). And core professional employees, "welcome direct interaction with the research team to address this, it's crucial we enage with you rather than managers" (M.R21-Clinical Academic).

Away from our [EliteMed managers] previous aim to equip them [clinical academics and practition-
It was not that clinical practitioners and academics were necessarily reluctant to engage with mode 2 efforts or more specifically the co-creation of HR interventions, but more they expected the latter to "take place in a constructive manner that proves rewarding for those that engage and adds value for EliteMed through our [clinical practitioners'] understanding of what motivates our colleagues to carry out translational health research" (M.R6-Clinical Practitioner). The constructive manner proposed by clinical practitioners and academics was one where "we should be left alone to design our own management, without interference from them [referring to EliteMed, managers]"

(M.R24-Clinical Practitioner):
Sometimes it can be frustrating having to do things for EliteMed, and you can not see its relevance.
We have to do what make sense to us and they [referring to EliteMed Managers] have to look and understand our interests and use our knowledge to inform more well rounded practices. And retrospectively it would've made much more sense for them [referring to managers] to ask to engage in this process a year ago when we started our journey with EliteMed, but at least they doing so now, and it's important to just let us get on with it. to engage clinical practitioners and academics in mode 2 research through EliteMed, following which the research team stepped back. "Informed by the research we asked them to explore challenges that have emerged and how to overcome these to influence their work" (M.R1-EliteMed Manager). Workshops were created keeping in mind the need to embed reflection and co-creation of HR interventions "in the context within which they [referring to clinical practitioners and academics] have to work, be collaborative in nature, invite contributions from both clinical academics and practitioners so that workshops goals were co-developed with those professionals" (M.R6-EliteMed Manager).
To aid the co-creation process, workshop participants had identified colleagues that represented "first amongst equals, those clinicians with sufficient influence and status to facilitate not just talk, but action" (M.R12-Clinical Practitioner). They represented 'knowledge brokers' (Currie & White, 2012); that mobilised the right knowledge (of their perspectives and practices) to the right people at the right time (from practitioners to their academic colleagues on the other side of the 'fence' or vice-versa); that were mandated by their academic or practitioner colleagues to engage in the workshops. (see 'Practitioner Note') The concerns of clinical academics and practitioners were thus brought to the fore in the workshops. They emphasised HR interventions, "needed to be dynamic and subject to context" (M.R31-Clinical Practitioner). Over time, initiated by their workshop interaction, knowledge brokers from both sides of the fence brought the colleagues they represented together in 'communities of practice' (Lave & Wenger, 1991) focussed upon their shared interest in a clinical service domain. Such communities were to prove crucial when wider professional engagement was required to scale up evidence-based clinical services derived from the Elite Med APC (see 'Practitioner Note').
The content of HR interventions co-created in such a manner, which represented the antecedent conditions for clinical practitioners and academics to engage in translational health research, is detailed below.

| Performance management
Clinical academics and practitioners agreed that EliteMed managers had to influence the "measuring of outcomes and improve our performance" (workshop observation, M.R19-Clinical Practitioner), but in a way that was not perceived "to be a burden" (workshop observation, M.R24-Clinical Academic) or "a threat" (workshop observation, M.R41-Clinical Practitioner). A critical element of monitoring that motivated those professionals was to "indicate where change is needed rather than to control what we do" (workshop observation, M.R43-Clinical Academic).
Core professionals explained that "performance management intended to improve efficiency usually challenges the performance of the project or fails to harvest the desired effect" (M.R12-Clinical Academic). Following which performance management criteria and approach were refined to align more with the professional identities that clinical researchers and practitioners expected: We're given that autonomy that aligns more with the type and nature of our work (M.R22-Clinical Academic) They have realised that we are professionals, and we value discretion and autonomy in what we do

(M.R19-Clinical Practitioner)
We witnessed several discussions between those professionals negotiating indicators of success. Clinical practitioners emphasised "impact upon service improvement" (M.R34-Clinical Practitioner) remained core to judgement of success of project investment. But we also note that there was equal emphasis upon quality of evidence produced by research as reflected in citation impact of peer-reviewed publications, alongside the gaining of external research income linked to projects, the latter recognised as important because service improvement required scaling up: Doing impact work requires time and additional resources. People assume that research translates into everyday practice and wonder why academics would not want to get involved, when of course the reality is that on a day-to-day basis we are extremely busy. I am now happy to engage in translational [driving academic evidence into practice] work (M.R12-Clinical Academic) The performance management system, when co-created with core professionals, was now perceived as less intrusive. It could be argued that the actual performance criteria to which clinical academics and practitioners had not shifted much; what was crucial was less the content in this case and more the process through which the performance management intervention was co-produced that engendered its acceptance. A formal project management structure was created to represent both professional groups, who met often to review project performance in regular workshops. These workshops were created to facilitate negotiation and shared understanding of performance requirements and associated metrics. The traditional understanding that knowledge was "produced by clinical academics and consumed by us (referring to practitioners)" (M.R19-Clinical Practitioner) was challenged and a more shared and iterative approach to knowledge development through collaborative processes was emphasised. Many alluded to the fact that the process of co-creation of HR interventions to support translational health research encompassed "working together and learning from each other" (workshop observation, M.R19-Clinical Practitioner) and developing "a common language, shared rituals and a shared purpose" (M.R-17-Clinical Academic). Thus they felt other HR interventions were also required to motivate those professionals to engage with each other in collaborative processes, in particular job design.

| Job roles and responsibilities
While the workshops were set up address the failure of the performance management intervention, the remit given to clinical academics and practitioners by EliteMed managers was wider. Discussion of professional identity and jurisdictional autonomy, led to discussions that higher status needed to be associated with any knowledge brokering role to be enacted by clinical academics and practitioners, for such roles to prove attractive. The more prosaic issue of 'time' to enact the role was also highlighted. Following which, workshop participants sought to more explicitly scope out duties and responsibilities of the knowledge brokering role, and ascertain the title that might be ascribed to the role that would be valued by their colleagues.
Status gain related to existing professional identities but derived from a leadership role they fulfiled as "quality improvement leaders" (M.R34-Clinical Practitioner) was highlighted by core professionals. For this leadership potential to be fulfiled, and status gain to ensue, their roles within EliteMed had to be carefully co-designed with professionals and according to what "we value" (M.R42-Clinical Academic), notably their professional identity and autonomy: And it's that combination of academic research with patient services in our job roles and clinical input and academic input that together hopefully will mean that change happens quicker (M.R2-Clinical Practitioner) To keep our research time we have to generate research income and for those of us who engage with clinical practice in any designed role it is much more feasible to do so.

(M.R17-Clinical Academic)
Clinical academics and practitioners emphasised that to engage in any role in the collaborative processes, jurisdictional identities, relative position and power related to their academic and practice responsibilities should be aligned with these new roles: I am not anti-impact, don't get me wrong, I think it is very important for clinical academics to actually have a voice in improving clinical practices but it is more important to ensure that we are able to do our academic jobs just as well, and that means basically focus on producing new promising knowledge through rigorous research.

(M.R12-Clinical Academic)
I am a clinician and I want to progress in this career. It's always attractive to work with academic practice as it help with career progression, but the priority always remains with the clinical work. In a meeting about an interaction with clinical academics, one clinical practitioner explained "time is something that I don't have. To engage stronger with clinical research requires time to set up those links and that is not enough, you need time to cultivate the relationships with academics to be able to do good quality meaningful work" (M.R10-Clinical Practitioner).
In response, EliteMed managers worked with clinical academics and practitioners to institute what they labelled either an, 'Impact Champion' role for clinical academics and 'Improvement Science Fellow' for clinical practitioners, which not only offered them the possibility to enhance their professional identity over time, but also mediated work pressures highlighted above so collaborative processes could be realised: My role has changed a lot. I've been awarded an Impact Champion role, which builds on the work we've been doing in EliteMed but provides me with my own funding source and protected time to do a lot more improvement work. Those engaged as Impact Champions or Improvement Science Fellows opened opportunity for collaboration by allowing clinical academics to co-supervise PhDs with clinical practitioners focussed upon service improvement, which enhanced those professionals' legitimacy as an innovation leader: The theory practice gap, people work in research and people work in practice, and I'm interested in ways of bringing those two things together through my role as leading PhD supervision in the field of service improvement. We can learn from each other and for me, engaging with EliteMed has presented a unique opportunity to cultivate stronger contacts with the clinical frontline and something that creates a very strong reputation of myself as a legitimate leader of innovation.

(M.R12-Clinical Academic)
My role has changed significantly over the last six months. I am now an Improvement Science Fellow, and this means I have to work closely with various clinical teams to support clinical innovation and change. It seems a perfect opportunity to go up a step with my career.

(M.R12-Clinical Practitioner)
In summary, the efforts of clinical academics and practitioners towards co-creating and instituting the Impact Champion or Improvement Science Fellow roles appeared fruitful in motivating them to engage in collaborative work with each other: EliteMed has given us a bit of a structure and support to work beyond the norms of our research here, and it's given us a real opportunity to get on and develop something together and do something a bit more exciting, a bit more out there, part of it is of course helping them to improve their performance but also collecting data for our publications. A win-win situation as you say in business (M.R13-Clinical Academic)

| DISCUSSION
Our empirical analysis has focussed upon co-creation of HR interventions between managers and those they seek to manage in a professionalised setting, an APC, which aims to generate mode 2 research across clinical academics and practitioners. In addressing our interlinked research questions, our study highlights the importance of professional identity that underpins clinical academics' and practitioners' engagement in mode 2 translational health research, their response to any imposed HR interventions, the content of HR interventions, and the need to co-create the HR interventions, which, in our empirical case, acts as an antecedent to mode 2 research in the translation of scientific evidence into clinical practice. Regarding the latter, reflecting on our experience of mode 2 research, we suggest that managers should not be privileged as the user group for HR interventions in a professionalised setting for those academic researchers, like ourselves from business schools. The lesson here is to avoid capture by those that commission your research, in our case the organisational managers. We highlight our own pursuit of the type of collaborative research with core professionals, encouraged by Hewett and Shantz (2021), from which we derive the following insight.
Derived from their identity, the type of core professionals upon whom HR interventions were imposed in our empirical case, value their jurisdictionally-based autonomy, following which they expect to be free from managerial intrusion. Co-creation of HR interventions inevitably changes power and control dynamics (Ferrary, 2005;Hewett & Shantz, 2021). Core professionals must be allowed to lead, not follow, others in the process of co-creating HR interventions. The role of academic HR researchers, as well as managers, is thus to 'stand back', allowing those professionals that are 'first amongst equals' to facilitate discussion and decision-making amongst their professional peers.
As highlighted in our empirical analysis, a climate needs to be cultivated within which workshops are seen as free of managerial or researcher intrusion. Thus, control needs to be ceded to the core professionals so that HR interventions align with incentives they judge as relevant, for them to engage with mode 2 translational research. This is not to dismiss a need for interactions across the three stakeholder groups of core professionals, managers and academic HR researchers, such relational ties proved crucial to co-creation of HR intervention (Gittell & Douglass, 2012). Managers provide the mandate for core professionals to co-create HR interventions that support translational health research in our case, and then follow up decisions made by core professionals about their content by ensuring they are implemented across the organisation; for example, incentives and job design have resource implications for an organisation (Beer, 1997). Meanwhile, researchers act as neutral knowledge brokers, and in our empirical case study, their analysis of failure of HR interventions facilitated the efforts of core professionals to co-create antecedent conditions for translational health research with which core professionals engaged.
Our empirical analysis highlights what HR interventions may represent receptive antecedent conditions for the translational health mode 2 research. Initially, even though failure ensued, EliteMed managers appeared to identify a set of consistent cues in the HR domain that were likely antecedents for core professionals to enact mode 2 research on the basis they linked to professional identity held dear by our focal core employees, doctors. Indeed, the HRM literature confirms their relevance, but we highlight, at least initially, EliteMed managers misaligned the content of the HR interventions with identity of the core professionals they aimed to enage in mode 2 translational health research.
In our professionalised setting, training and development proved more effective when it brings clinical academics and practitioners together to enhance their shared interest in the development of evidence-based patient care in a collegial manner characteristic of professional organisation (Mintzberg, 1979). In essence, training and development should allow professionals to co-opt managerial practice to align with their identity, rather than co-opting professionals into managerial practice (Waring & Currie, 2009).
Second, we confirm the significance of performance management and the influence of professional identity upon what is acceptable for clinical academics and practitioners to engage them in mode 2 research (Foss et al., 2009;Gagné & Deci, 2005). Managerial intrusion in the realm of any HR intervention, not least externally imposed performance management systems to evaluate collaborative processes for mode 2 research, is particularly unwelcome since it challenges professional autonomy and status (Raelin, 1985). In our empirical case, over time, managers recognised they need to understand, and be prepared to learn about, professional practice, in this case made more complicated because there were two sets of professional practice to consider. For professionals, they showed appreciation of the managerial context within which they work, specifically the strategy and performance accountabilities (Hewett & Shantz, 2021).
Third, whilst the focus of the training and development workshops in our study was to address failure of the performance management intervention to support collaborative processes towards mode 2 research, job design emerged as an important issue for clinical academics and practitioners. Confirming such assertions in literature (Bach et al., 2008;Foss et al., 2009;Yan et al., 2013), job design had a particular effect upon professional status , but also mitigated operational work pressures, and consequently affected whether professionals are prepared to engage in knowledge creation through mode 2 research.

| CONCLUSION
Summarising the contribution of our study to HRM literature, it is three-fold. Regarding our first question focussed upon antecedents for mode 2 research, our analysis responds to calls for insight into implementation of HR intervention that is contextualised and takes account of institutional conditions, in our case that of a professionalised setting (Mirfakhar et al., 2018). We show common purpose can be attained by focussing upon professional identity of academics and practitioners, which means those seeking to manage mode 2 collaborative processes need to ensure academics' and practitioners' status and jurisdictional autonomy (Abbott, 1988;Freidson, 1988) are recognised in HR intervention.
Regarding our third research question, we provide insight into the process of co-creation of HR interventions that act as antecedents to mode 2 research (Hewett & Shantz, 2021). Within our institutional context, an effective stance towards co-creation of HR interventions and the translational mode 2 research that follows is one whereby, both managers and our team researching HR interventions allow core professional employees to control discussion and decision-making free from intrusion. This aligns with professional expectations about their jurisdictional autonomy (Abbott, 1988;Freidson, 1988). Further, in the process, any academic team researching HR interventions needs to take care not to be captured by the agenda of senior management, and to recognise the importance of engaging core professional employees in the HR implementation process, otherwise HR interventions will fail to add the value senior managers expect (Mirfakhar et al., 2018).
Regarding practical implications, responding to wider concern of other directors for the nine translational health research APCs, of which EliteMed was one, our study outlines relevant performance management, job design and training and development interventions, to support co-production.
Regarding its limitations, we suggest our analysis is transferable to other professionalised settings, nevertheless we encourage further research in non-clinical professionalised settings examining prospects for bridging identies through co-creation of HRM interventions across academic and practitioner communities to support mode 2 research (Hewett & Shantz, 2021).
Finally, we provide a reflective note regarding our learning as researchers about mode 2 research approaches. We did not feel our own identity as scholars was eroded in the course of our study, indeed engaging in mode 2 research provided insight that might otherwise have been unavailable to us (Butler et al., 2015;Empson, 2013). We engaged with EliteMed managers and core professionals on their (practitioners') turf and produced actionable knowledge to address a pressing organisational problem they had identified (Bleijenbergh et al., 2021). The needs of EliteMed managers were time critical, and if we were to engage them in the research for which they commissioned us and make impact upon practice, the generation of peer-reviewed publications should not be of immediate concern. Nevertheless, management practitioners were interested in our theoretical interpretation as a means to understand why such prescriptions were likely to work as means of abstracting lessons that were actionable beyond the immediate problem. Further, over time, we successfully bid for research grants, which included managers and core professionals involved in mode 2 translational health research, and published in high quality peer-reviewed academic journals about the management of translational health research initiatives. Through our mode 2 research we thus created a 'win-win' outcome for both us and managers and core professionals with whom we engaged (Werr & Greiner, 2008).
In conclusion, we thus argue mode 2 research can prove both academically and practically legitimate, and should be seen as complementary to, not competing with, traditional modes of research (Guerci et al., 2019).

ACKNOWLEDGEMENT
The research study was funded through National Institute for Health Research and Care (NIHR) Applied Research Centre West Midlands.