Professionalism Redundant, Reshaped, or Reinvigorated? Realizing the “Third Logic” in Contemporary Health Care

Recent decades have seen the influence of the professions decline. Lately, commentators have suggested a revived role for a “new” professionalism in ensuring and enhancing high-quality health care in systems dominated by market and managerial logics. The form this new professionalism might take, however, remains obscure. This article uses data from an ethnographic study of three English health care improvement projects to analyze the place, potential, and limitations of professionalism as a means of engaging clinicians in efforts to improve service quality. We found that appeals to notions of professionalism had strong support among practitioners, but converting enthusiasm for the principle of professionalism into motivation to change practice was not straightforward. Some tactics used in pursuit of this deviated sharply from traditional models of collegial social control. In systems characterized by fissures between professional groups and powerful market and managerial influences, we suggest that professionalism must interact creatively but carefully with other logics.

"new professionalism" that might embody the best of the professional ethic and secure its place at the heart of health care delivery (see e.g., Brennan 2002;Cruess, Johnston, and Cruess 2002;Irvine 1999;Royal College of Physicians 2005). Our focus in this article is on the fate of efforts to promote this "new professionalism" in an institutional field governed by multiple, often competing forces. The empirical material we use to advance our analysis derives from a study of a program of health care improvement initiatives in England. What makes this program analytically significant is that it explicitly and purposefully sought to mobilize professionalism in the pursuit of quality. We animate our analysis by identifying professionalism as an institutional logic (Thornton and Ocasio 2008) that may have particular valence in influencing health care professionals alongside ascendant corporate, state, and market logics (Goodrick and Reay 2011). We focus particularly but not exclusively on medical professionalism, and we begin by briefly scoping its history and current position in a complex and heterogeneous institutional field.

BAcKGROUnD
The Rise, Fall, and Resurrection of Professionalism?
On both sides of the Atlantic, the professions, and particularly medicine, enjoyed a privileged place in the organization of health care through the early and mid-twentieth century-the period often characterized as the Golden Age of medicine (Starr 1982). Autonomy and self-regulation were underwritten by an overwhelmingly positive public image of the professions, broadly endorsed by a generation of social scientists ( for an overview, see Light 2010). Given the esoteric knowledge base, information asymmetries, and potential for malpractice, functionalist sociology identified in the ideal type of professionalism an apparent solution to the challenge of ensuring that vulnerable patients obtained appropriate care and that the practitioner's duties to wider society were upheld (e.g., Parsons 1939).
The turn from functionalism from the late 1950s onward prompted a different set of concerns within the sociology of professions. Sociologists began to argue that claims of special credentials were strategic maneuvers aimed at securing exclusive rights to particular titles and practices in pursuit of occupational enhancement (Larson 1977)-rights that were established by the state through legal provisions that offered professions protection not enjoyed by nonprofessional occupational groups (Salter 2004). Critiques from the likes of Freidson (1970) and Larson (1977) saw the professions less as noble defenders of the public good than as cabals that used their institutionalized protection from competitive forces to advance their own interests.
Beyond academia, in the 1980s, institutionalized protection for professions came under challenge in health policy. Increasingly, the autonomy of the medical profession in particular came to be seen as a vice rather than a virtue, one that had given rise to, inter alia, spiraling health care costs, clinically unjustified variations in care, and a dangerously cozy relationship with the pharmaceutical industry (Mechanic and McAlpine 2010). In the United States, the rise of managed care saw insurers and other health care purchasers assert their power over the professions (Light 2000;Starr 1982), while in England, a series of reforms saw increasing managerial power over professional decision making, marketization of the National Health Service (NHS) (Klein 2006), and the erosion of professional selfregulation (Dixon-Woods, Yeung, and Bosk 2011).
At the same time, the very nature of professional work-what physicians and other health care professionals actually do-itself has changed dramatically (Noordegraaf 2011). Complex conditions and new treatment modalities have brought with them new risks and interdependencies with which traditional professional approaches to managing risk and assuring quality are ill equipped to deal. Professionals also now work in a very different occupational environment from that of the midtwentieth century. New interdependencies have been created; work is increasingly done by interprofessional teams rather than the heroic individuals of mid-twentieth century imagining; it is made visible, monitored, and controllable in multiple ways. The rise of evidence-based medicine and new technologies have opened previously inconceivable possibilities for the surveillance of professional conduct by external actors (Martin et al. 2013), including the state, managers, insurers, and patients, such that these actors now engage with the professions on the basis of "justifiable" or "verifiable" trust rather than the unconditional trust that perhaps once prevailed (Kuhlmann 2006). This nexus of changes within health care and the contemporaneous mutations in the nature and cultural accounts of professionalism have not occurred in isolation. Wider, societal-level changes have further eroded the influence of medical professionalism. These changes are manifest in the rise of alternative, overlapping "institutional logics" that have to some extent displaced the rules and norms of the logic of medical professionalism in determining behavior, notably the logics of market, corporatism, and state (Goodrick and Reay 2011;Scott et al. 2000). Institutional logics are "the socially constructed, historical pattern of material practices, assumptions, values, beliefs, and rules by which individuals produce and reproduce their material subsistence, organize time and space, and provide meaning to their social reality" (Thornton and Ocasio 1998:804). With its roots in neoinstitutionalist sociology, the institutional logics approach (Friedland and Alford 1991;Thornton and Ocasio 2008) offers a useful conceptual starting point for understanding the fortunes and potential of professionalism in health care; three distinguishing features of the approach are especially relevant for our current purpose.
First, theorists of institutional logics have helpfully articulated how logics can form and evolve at multiple social levels and in multiple fields of organization, interacting and mutually shaping as they do so (Thornton, Ocasio, and Lounsbury 2012). This helps to explain why medical professionalism shares many characteristics of the higher-order logic of professionalism but also characteristics of professionalism in other fields (law, accountancy, etc.) at the same time as being clearly distinctive. Second, the approach illuminates not only how competing logics may coexist, with dominant logics ebbing and flowing through time, but also how even institutional logics that are in decline may continue to affect field practices (e.g., Goodrick and Reay 2011;Hinings 2005, 2009). Third, recent expositions of the approach attend to how change may occur within logics themselves, in response to shifts in societal-level "meta-logics" (e.g., Scott 2008:232), the influence of other institutional logics in the same and neighboring fields (e.g., Thornton et al. 2012), and the "bottom-up" agency of individual and organizational actors within a particular field (e.g., Seo and Creed 2002).
An account of health care informed by these analytic constructs might then identify the fading dominance of a medical-professional logic, including its relegation to a bit part rather than starring role, subordinate to the rules, norms, and cognitivecultural frames presented by ascendant logics of market and management-but not entirely without influence. Thus, Scott et al.'s (2000:338) study of providers in the San Francisco Bay Area documents the decline of medical professionalism from 1965 onward but notes "a continuing cacophony of contending logics and divided regimes" to which the medical-professional logic contributes an ongoing, if quieter, melody.
For some, the retreat of professionalism into the background, leaving the foreground to other logics in health care, deserves a qualified welcome, insofar as it reflects a reining in of the excesses of professions and professionals (Light 2000). Others, though, identify the potential risks-and actual iniquities-created by health care systems in which managerial or market logics are dominant. Fears are expressed that the professional's fiduciary duty to the patient may be replaced by organizational interests (Mechanic and McAlpine 2010), and unbridled managerial control may distort professional behavior in harmful ways (Bevan and Hood 2006).
Consequently, calls are increasingly loudly made for the continued importance of professionalism. Yet, what exactly professionalism might mean is itself disputed. Among the rallying cries is Freidson's (2001) notion of professionalism as a "third logic." In a notable turn from his earlier position, Freidson, once the hammer of the professions, puts forward a defense of an ideal-type professionalism (as distinct from "actual existing professionalism"). He envisages a professionalism that might act as a countervailing power against managerial and commercial forces, working to the benefit of health care quality and patients' interests. In this ideal type, he finds room for a much greater role for medical professionalism than it is seen currently to play in much of the institutional logics literature (e.g., Reay and Hinings 2005;Scott et al. 2000).
Separately, commentators in health care policy and practice have similarly argued for salvaging the best characteristics of traditional professionalism and recasting them in terms of the clinical, social, and organizational realities of today's health care systems and societies. Thus, leading figures including Irvine (1999) and Brennan (2002) have called, respectively, for "new professionalism" and "civic professionalism," with the professional "leading the way, not being brought along by regulations" (Brennan 2002:978). The medical profession itself on both sides of the Atlantic increasingly promotes the value of a "new," reconfigured professionalism (American Board of Internal Medicine Foundation 2002;Royal College of Physicians 2005). In England, the call for new professionalism has even been taken up by the state, where it is cast as a means of fostering leadership for quality, freeing and empowering staff, and reinstating accountability to patients rather than administrators (Secretary of State for Health 2008).
These various constructions of a reconstituted professionalism have in common a number of features. All are agreed, for example, that a new professionalism must embrace features of modern health care systems-such as evidence-based practice, active rather than passive patients, and wider networks of accountability and regulation-rather than revert to a tradition that puts individual professional autonomy center stage (Starr 1982). Even Freidson-though sometimes (mis)characterized as advocating a "nonsensical" "pure professionalism" (Noordegraaf 2007:781)-distinguishes between the ideal-typical characteristics of professionalism and its empirical realization: "reality is and should be a variable mix of all three logics, the policy issue being the precise composition of that mix" (Freidson 2001:181). But while these constructions envision a model that is quite different from the medical hegemony of the Golden Age, they also call for professionalism to retain its distinctiveness and collective-level autonomy.
Some strains of work in the institutional logics approach give reason to be cautious about the fortunes of such efforts to rehabilitate the logic of professionalism-or indeed any logic in a field of competing logics. Thornton et al. (2012:164) identify several ways in which field-level logics can mutate: one may displace another, or interaction between logics may result in the characteristics of one being incorporated into another. This poses dangers to the integrity of a reinvigorated professionalism: Evetts (2009:248), for example, notes that "occupational professionalism"-a discourse that guides the conduct of individual professionals who subscribe to the norms, values, and expectations of their collegium-may be appropriated by "organizational professionalism"-"a discourse of control, used increasingly by managers." Where organizational mandates displace collegial obligations, Evetts argues that "professionalism" loses the very characteristics that proponents of new professionalism wish to restore.
In this light, exactly what form a reinvigorated professionalism should take and how it should interact with other institutional logics in the health care field is unclear. Recent studies have shown how receding logics can retain an important, if subordinate, role in an organizational field (McDonald et al. 2013;Reay and Hinings 2009). However, no study has examined the fortunes of a purposive attempt to rejuvenate a receding logic, its composition, and the degree to which it can remain distinctive and command legitimacy in a field now dominated by other logics. We seek to fill this void through a study of a program of health care improvement projects that were premised explicitly on the idea of harnessing professionalism. We focus on the degree to which this approach appeared to succeed in gaining legitimacy among individual clinicians, the tactics used by leads to turn legitimacy into action, and the way this was received by professionals themselves. In our discussion, we reflect on how far this form of professionalism constituted a distinctive, autonomous logic alongside those of the market, state, and corporate managerialism.

DATA AnD METHODS
The program that was the subject of our inquiry was known as "Closing the Gap through Clinical Communities." Funded by the Health Foundation (a British charity that funds quality and safety improvement initiatives), it included 11 projects, each charged with improving health care quality and safety in areas where there were known deficiencies in current clinical practice. Each project comprised a core team that led and managed the project and a number of participating teams that undertook the improvement activities in their own organizations. Significantly, the program was explicitly based on notions of professionalism as a means of driving improvement, based on the guiding principle that professionally led change would be more likely to mobilize clinical staff into action than alternative approaches such as top-down, managerially imposed mandates.
This article derives from an evaluation of this program, involving in-depth ethnographic study of three projects that were purposively sampled from the program. These were: Research UK, a charity, which sought to improve the care of people diagnosed with chronic kidney disease (CKD), a long-term condition managed in primary care.
Case study selection, both of these three projects and of participating teams in local NHS organizations, was guided by theoretical and empirical literature and aimed to include variation in characteristics considered likely to influence the success or otherwise of this approach to improvement. These included clinical setting, the "quality gap" to be tackled, host organization, professional leadership, and organizational context. Table 1 provides an overview of the projects selected and key features; it also serves as a glossary of acronyms and specialist terminology for readers unfamiliar with English health care.
We undertook 63.5 days' ethnographic observation across the three case studies (ILCOP: 25.5; AAA-QIP: 21.5; ENABLE: 16.5), focusing on core teams' activities (internal meetings and events convened by core teams to bring participating teams together) and participating teams' implementation work in clinical and managerial settings. Additionally, we undertook 5 days' observation of overarching program-level events. We conducted 126 in-depth interviews with members of core and participating teams (45 each in ILCOP and AAA-QIP; 36 in ENABLE) and 11 program-level actors. Interviewees included, among others, clinicians from all involved professions, health care administrators, data managers, and commissioners (payers).
Interviews were audio-recorded and transcribed verbatim. Fieldnotes from observations were "debriefed" within the team, with each member's fieldnotes discussed with other members to document key events and begin to identify areas of analytic interest; debriefs were audio-recorded and transcribed. Relevant project documents were also collected for analysis, including plans, reports, and training materials. Data analysis was based on the constant-comparative method (Charmaz 2007). Analysis involved intensive engagement with the data for each of the case studies to ensure that they were each understood in terms of their own context and meaning, followed by comparison across cases to generate higher-order themes, and then further interrogation to attempt to identify reasons for differences and similarities across cases. NVivo 8 software was used to assist in coding the data, locating recurrent themes, and grouping themes together.

RESULTS
We present our findings in three sections. First, we highlight how the program grounded its activities in the ideals of professionalism. Then, we discuss the tactics used by core teams to secure the commitment and actions of their peers. Finally, we explore the responses of professionals themselves.

New Professionalism's Promise
In interviews, those involved in directing and funding the Closing the Gap through Clinical Communities program explicitly proposed that an approach founded on handing over leadership and control to clinicians might help to make improvement happen where managerialism and markets might fail: If you think of health care in terms of tribes, it's your own tribe wanting you to do things, rather than some rival or alternative or non-tribal person. (Program manager 1) If clinicians aren't fully engaged and fully involved in improvement then it won't happen, it can't be something that's driven from outside the professions. (Program manager 2) Across all three case study projects, those involvedfrom core teams to participating teams-were similarly enthusiastic about the potential for the professionally led approach espoused by the program to achieve positive change. They identified the need for and importance of a new, reimagined professionalism, arguing that initiatives led by professional insiders could confer credibility and legitimacy that was lacking when the leadership came from outside the profession. Two of the three projects were led by professional associations, which were described by project participants as having evolved from club-like organizations of a previous era into institutional structures that could both define and promote standards of practice and conduct for their members and command authority and allegiance: [We] went through phases when colleges were supposed to be, you know, old boys' club and old physicians talking nonsense. That is changing. . . . I take a great influence from the Royal College of Physicians so if the RCP says something, I take it seriously, as compared to few other organizations within the NHS. (Respiratory physician, participating team, ILCOP) Professionally led efforts to improve care were seen by leads and participants alike to have particular value and salience in a health care context laden with other priorities (often driven by central Clinicians participating in the projects thus endorsed an approach to improving quality that aligned with their own motivations toward improving patient care. Further, because the projects were free to use methods of their own choosing and to work in areas that the participants themselves recognized as most important, participants saw them as much less prone to "gaming," tick-box compliance, or perverse incentivization than managerial diktat or financial incentives: I thought this was a really good and useful new initiative that would pick up where peer review left off. [National] peer review [see Table 1], I felt, had become very bureaucratic, very processdriven and had lost sight of the point of peer review, which was to identify, on a peer-to-peer basis, areas for improvement. And traditional peer review has now become little more than a tick-box exercise and if you've got your paperwork in order, that's fine. (Respiratory physician, participating team, ILCOP) These views were not confined to physicians; nurses and others offered very similar responses, affirming a "new" kind of professionalism that appealed across, rather than solely within, professional groups: We are in a very privileged position I feel up here, because it was being driven by the clinicians. They see the value in working together closely. I think they do it informally I think and this just formalizes the process. (Network coordinator, participating team, AAA-QIP) The projects thus enjoyed considerable legitimacy with the clinicians whose behavior they sought to influence. Next, we explore the approaches core teams used in seeking to turn this legitimacy into something that would motivate clinicians toward change.

Strategies for Securing Commitment: Seduction, Deliberation, Coercion, Enforcement
We identified four strategies used by core teams in various combinations over the course of the program to attempt to translate good will into real influence: seduction, deliberation, coercion, and enforcement. Some of these resembled traditional modes of influencing behavior within professional collegia (e.g., Freidson and Rhea 1963); others went well beyond it.
Enacting the Collegium: Seduction, Deliberation, and Coercion. What we term seductive tactics were an especially important feature of early phases of the projects. Clearly operating through collegial principles, they sought to appeal to professional values, sensibilities, and identities and were explicitly persuasive in character. They aimed to convince participants (and would-be participants) that changes in norms, practices, and behaviors were an important professional responsibility and that delivering on this duty would likely not only improve outcomes for their patients but also bolster occupational status, offering a defense against external attempts at control. Core teams offered demonstrations of inspiring leadership, provided evidence of apparently poor practice and of the benefits of change, and emphasized how, if clinicians did not seize the initiative themselves, it would be seized from them by those outside the professions. In these ways, they sought to reinforce both the identity and the responsibilities of the collegium of professionals, soothing misgivings by stressing how the changes sought had been selected and endorsed within rather than without the professional communities, with some effect on participants: There's always this thing, some managerial person comes along and says, "You've got to do X, Y, Z," and everyone just puts their hackles up. Whereas if it's coming from within, I think it's much better and . . . I don't think anyone can argue with the principle of it. (Vascular surgeon, participating team, AAA-QIP) But while seductive tactics often created the necessary receptivity among professionals, they did not always do much more than this. Whatever its appeal in principle, relying on clinicians' intrinsic motivations to translate legitimacy into motivation for change was prone to obstacles in practice. Feeble or perfunctory efforts that stopped well short of what the core teams believed was needed to secure improvement were reported, as were inertia and non-engagement. "No one's disagreeing with the concept of trying to improve things and trying to reduce mortality. The problem is when you get down to the fine detail and individuals, and no one really wants to change. . . . As in every aspect of medicine, there's an attitude, "Well I've done this for the last 10 years, why should I change now?" Well the answer to that is, "Mortality isn't low enough," but no one ever thinks it's their problem. (Vascular surgeon, participating team, AAA-QIP) A second major reason for the faltering of efforts to secure participation and action through an appeal to professionals was the force exerted by competing institutional logics. Mandatory expectations from other masters were often impossible for professionals to evade: Daily work has to be the priority; you can't tell the waiting room to go home. Thus, in a regime dominated by rather different logics, the seductive appeal of projects premised on professionalism did not translate automatically into motivation to engage in the changes proposed. Core teams were therefore obliged to look to other tactics. A second important tactic was that of creating opportunities for deliberation, which involved bringing participants together to talk through and take ownership of the changes. The AAA-QIP project, which sought to change the behavior of vascular surgeons, a traditionally highly autonomous group, used this tactic extensively. The project proposed a care pathway to standardize what would happen to patients across multiple sites, but this pathway was supported (of necessity) by an incomplete evidence base and provoked controversy and complaint. Deliberation was deliberately used to counter these challenges; the core team convened regional meetings at which affected stakeholders from participating teams worked together across occupational groups to adapt the care pathway to their region, on the assumption that once the community took ownership of the design, the norming effects would then take care of the laggards.
It's a really good starting place because they will dismantle it and then put it back together with what works for them and that's a really good place to be. Everybody has said they're quite happy, because I think they did that to some degree. The paperwork came from somewhere, we all had a look at it and thought, "We didn't like that; that's a good idea; have you seen this?" (Service administrator, participating team, AAA-QIP) The effects of these deliberative forums went beyond the processes of discussion and consensus making. They were also, crucially, concerned with the creation of a collectivity and with the performance of professional identity.
Me coming along and saying, "These are the standards, get on with it": you'd say, "On whose authority, mate?" These are all intelligent people, they're busy, they need to have a reason to get involved, and the reason to get involved is that their peer group, that they subscribe to, has said it's important. (Core team member, AAA-QIP) It's about us being able to check what's happened, so as a clinician in a busy team myself, I know if I've got to go to a meeting in three months' time and present what we've done, I'm going to get something done to present: I don't want to stand up and look a fool. And I think that will prompt people to get stuff done. . . . They don't want to turn up at meetings or be highlighted as people who had this chance and didn't use it. (Core team member, ILCOP) Seduction and deliberation together appeared to go some way toward engendering commitment from participating sites but again, did not always appear to translate into personal motivation or collective will among clinicians. Increasingly, therefore, core teams resorted to more coercive tactics for instigating change, using collective-level influence and edict. Again, however, these tactics retained the source of the imperative within the collegium itself. Two projects began to augment the scope for peer pressure by altering the way they published figures from audit databases; they moved away from confidential feedback to individual teams toward more open publication to allow teams to compare their own performance with that of others. The core teams recognized that this was an approach that needed care, so as not to "single people out too much" (Core team member, ILCOP). It thus involved "not so much naming and shaming; it's more about openness, it's allowing people be to be subjected to peer pressure" (Core team member, AAA-QIP). In AAA-QIP in particular, there seemed to be an acceptance of the inevitability of this more open approach to audit since surgical outcome measures were already in the public domain in neighboring disciplines such as cardiac surgery. Core teams promoted-and broadly gained acceptance for-the idea that open comparison of data and management of performance within the profession was better than the alternative of an externally imposed managerialist regime (Meyer and Rowan 1977) or "trial by media": [Participating teams] are very mindful of the fact that our mortality is high and it needs to be driven down. And I don't think they'll want to see themselves having been outliers on any graph that's published, because they don't want a visit from The Guardian [a national newspaper], they don't want The Guardian coming saying, "Look, why are you an outlier?" (Vascular surgeon, participating team, AAA-QIP) It allows us to compare ourselves with other centers as well and I think it's important nationally, to have some standards and some way of recording actually that what we're doing is right. (Vascular surgeon, participating team, AAA-QIP) ILCOP and AAA-QIP thus increasingly embraced "harder tactics" (Aveling et al. 2012) to change behavior, explicitly seeking to place limits on individual-level autonomy to fulfill wider societal accountabilities. Insofar as they maintained professional autonomy at the collective level, within the collegium, and operated through modes such as peer influence and top-down pressure, such tactics were by and large accepted by those subjected to them. Here, then, was evidence of a "borrowing" from other field-level logics (including managerialist tactics of measurement as a means of control) but a retention of the imperative for change within the collegium itself.
Underwriting the Collegium: From Coercion to Enforcement. At times, however, the tactics went further still: they went outside the collegium and sought to build synergies with external mechanisms of change. In particular, the projects came to utilize alignments with wider aspects of health care governance, including forces associated with other institutional logics. In ENABLE, this alignment with external forces existed from the start: the project sought to consolidate and build on a set of qualityrelated requirements that had been implemented by the state through financial incentivization (QOF; see Table 1). In ILCOP, the core team decided over time to invoke the power of nonclinical service managers where the professional model alone seemed insufficient to secure clinically led change: These [hospitals] can't just sign up for this project and not do anything about it, so, e-mailing the trusts to tell them that we're going to be speaking to their CEOs, I think it was a way of, you know when you're at school and they're like, "We're going to talk to your mum if you don't improve!" (Core team member, ILCOP) The improvements targeted by AAA-QIP, meanwhile, were aligned with wider pressures on hospitals and practitioners: a move toward rationalization of vascular surgery prompted by evidence of associations between volume and outcome, greater transparency and publication of surgical outcome data, and the introduction of a population screening program that demanded similar quality to those espoused by AAA-QIP. To a large extent, therefore, the project was operating in a wider environment that already underwrote its aims: They all know the screening program is going to come in their area in the next six months and therefore they're all keen to provide the same service of care, and [AAA-QIP's] care pathway bundle will ensure the consistency of the service that's provided and it will enable them to audit their practice as well. (Service manager, participating team, AAA-QIP) Handled skillfully, alignments with broader shifts could motivate participants toward change while avoiding the charge of infidelity to the professional ethic: If I can get the commissioners [healthcare purchasers] to commission it, then half my work's been done for me, because then you've got people who will ring up chief executives and make their lives miserable, and so the whole thing rolls without me being Dr. Strangelove in the corner. (Core team member, AAA-QIP) The Response: Professionalism Revitalized, Powerless, or Co-opted? Thus, by design, by modification, or by accident, the projects came to rely not just on the mixture of coercive and cooperative strategies internal to professions but on alignments with key features of the wider systems in which they operated. They evinced varying responses from the professionals they sought to target, including some successes (Table 2).
However, efforts to reanimate professionalism as the organizing principle of improvement efforts were heavily mediated by those wider systems. All three projects faced difficulties in aligning professional with market, state, and corporate logics in a way that did not subordinate the professional logic. This proved challenging. For one thing, any incongruence between the aims of the project and the managerial and market logics manifest in the priorities of managers and commissioners (payers) tended to undermine, in the view of participating teams, both the project objectives and the sense that a professional ethic might prevail. Where managerial priorities were out of kilter with the plans developed by ILCOP's participating teams, for example, the professional ethic could be derailed: There's been no manager involvement: the cancer manager, for example, hasn't been involved with this at all. I think it is partly because it is a clinician-led, clinician-driven initiative, and that's right and proper, but it's also partly because cancer managers in particular tend to focus an enormous amount on targets, meeting the targets, organizing things they have to do like peer-review targets. (Respiratory physician, participating team, ILCOP) Even in AAA-QIP, which as we noted previously benefited from an apparent synergy with wider moves toward rationalization of provision, the and whose medical significance is questioned by some Note: See Table 1 for full explanation of acronyms and terminology used in Table 2.

Table 2. (continued)
confluence between managerial and professional agenda could sometimes undermine, rather than support, the project's efforts to engage clinicians, for example, where external targets were already being met or exceeded: [Hospital management] will only help to drive it if it becomes a hard target, that we're not doing this and we've got to be. . . . At the moment, you know, [the hospital managers are] saying, "OK, the target is 3.5 [percent elective mortality]; our overall mortality is 2.7," so we can't even say, "Actually, we're not performing and you need to do it." I can't see management being too keen to invest. (Vascular surgeon, participating team, AAA-QIP) Across cases there was a sense that while sometimes necessary, interactions with other logics could constitute a dangerously double-edged sword. External, hierarchically imposed targets could bolster projects' own efforts at engaging clinicians but could also provoke goal displacement and even perverse incentives. Moreover, efforts to actively generate confluences between professionalism and other logics sometimes had counterproductive consequences. In ENABLE, for example, project leads sought to ensure alignment between the project's objectives around CKD management and wider, state-driven managerial mandates. In practice, however, they found that this apparent synergy did not always work as anticipated: [GPs have] probably got all the money from QOF [for] CKD that they can get. Most of them have-if you look at the register for example, you just have to make a register and you get all the money [available through the QOF system]. It doesn't matter how many people [are on the register]. (Core team member, ENABLE) In consequence, family physicians (GPs) seemed reluctant to engage fully with ENABLE. English primary care is a system that, as others have noted (McDonald et al. 2007), is particularly "crowded" with the extrinsic incentive mechanisms of market and state logics. The powerful motivator of financial incentives, in which "adequate" performance according to imperfect metrics was rewarded, might be understood as crowding out the intrinsic motivation to excel in service quality that ENABLE sought to encourage. But it would be a mistake to construct the problem quite so simplistically: it was also true that some GPs were skeptical about the extent to which it was appropriate to identify and manage CKD as a disease (rather than as a normal part of human aging). For them, professionalism meant engaging critically with the evidence, not simply accepting what they were being asked to do: There's an awful lot of skeptics, some doctorsthere's a huge argument . . . about CKD, and they didn't believe it at all. (Core team member, ENABLE) Similarly, in the other two projects, there was resistance to the more hard-edged approaches taken by the core teams. Some clinicians challenged the desirability of the changes being proposed, suggesting for example that the evidence base for certain interventions was inconclusive. For others, collective, clinical ownership of the improvements was not enough for them to endorse managerialist-style monitoring and enforcement of changes at the expense of individual professional autonomy. Again, the potential adverse consequences for patients were invoked as part of these arguments: One should always be striving to improve but one of my concerns is that we're now putting things into place [that are . . . ] taking the effort out of thinking for yourself. I think that guidelines are just that, they're guidelines, and any individual patient's treatment should be tailored for that individual patient and should be based on evidence and best practice, but not dictated by evidence and best practice. (Vascular surgeon, participating team, AAA-QIP) Core teams were thus not always successful in reassuring participating clinicians that their tactics remained true to the professional ethic: for some, a professionalism based on enforcement seemed little more than a fig leaf for the managerial logic (cf.

Evetts 2009).
Context was also important. AAA-QIP and ILCOP needed to secure the attention and commitment of practitioners outside their traditional spheres of influence (defined largely by their respective professional associations, the RCP and the Vascular Society), and here they could not be confident of legitimacy and sway. The regional meetings convened by AAA-QIP, for example, were better attended by surgeons (who were members of the Vascular Society) than by radiologists, anesthetists (anesthesiologists), and nurses-who were not members yet were essential to the project's multidisciplinary vision.
A vascular surgeon will be a surgeon who works only in vascular surgery. Anesthetists tend to work in a number of areas: certainly in theatre, I can't think of anybody in the country that just does vascular, just provides anesthetics for vascular-surgical procedures. (Anesthetist, participating team, AAA-QIP) ILCOP, similarly, found it easier to engage physicians, clinical nurse specialists, and multidisciplinary team coordinators than radiologists, surgeons, oncologists, and pathologists: [Pathologists] tend to not site specialize, which I think is why we in lung cancer are behind, because of pathology. . . . If you don't have expertise, you don't have any ownership, and if you don't have any ownership you won't be involved, you don't feel involved in service development and improvement because you're just doing general service-level work. (Clinical oncologist, participating team, ILCOP) Besides their positioning outside the scope of the intradisciplinary influence of professional societies and collegial peer pressure, groups such as pathologists in ILCOP and anesthetists in AAA-QIP also lacked a crucial sense of ownership of the issues around quality. Consequently, their motivation to engage seemed considerably weaker: these were not "their" problems to fix.
In the primary care context in which ENABLE operated, meanwhile, GPs' generalism militated against ownership of a problem that affected only a small proportion of their patients and made it challenging to define CKD management as a legitimate problem deserving attention: We are constantly being asked to do more, especially in terms of audits and QOF work. We're also facing very significant cuts in our budget and going to commissioning meetings the whole time, . . . that's taken up a huge amount of spare time. So this important clinical project has regrettably gone under some of these more pressing initiatives. (GP, participating team, ENABLE) All three projects, then, faced challenges in motivating participants through the range of strategies they adopted in environments characterized by dependencies on multiple occupational groups whose interests and motives were not always aligned and by strong influences on practitioner behavior deriving from other logics. While the use of hard tactics based on coercion and enforcement seemed a necessary corrective to the softness of seductive and deliberative approaches, it also gave rise to resistance from clinicians who conceived of professionalism differently. Clinicians' receptivity to the program thus varied according to the constraints of different clinical contexts, the perceived legitimacy or otherwise of the tactics adopted within those contexts, and the varying sway held with clinical groups whose engagement was also influenced by wider interests, pressures, and structures (see Table 2).

DIScUSSIOn
Our analysis sheds empirical light on the scope of the promise that writers within and beyond the health professions have claimed for a reinvigorated professionalism as a means of enthusing and motivating clinicians to engage with health care improvement initiatives. Evident from our findings is that faith in the potential of "new professionalism" is not merely a preoccupation of academic commentators or the elite: it is shared by frontline practitioners, many of whom saw in clinically led and owned projects a moral authority and potential for influence that was not to be found in other institutional logics that pervade modern health care systems. In principle, practitioners on the ground welcomed the way in which these projects sought to re-empower them as professionals. In practice, ensuring that good intentions translated into concrete action was rarely straightforward. Romanticized appeals to professionalism were not enough; instead, enrolling clinicians required the use of multiple tactics from seduction and deliberation to coercion, sometimes aligning with other logics, sometimes bumping against them.
We show that professionalism retains a legitimacy and a particular influence in the current health care field. As others have found (Goodrick and Reay 2011;Hinings 2005, 2009;Scott et al. 2000), it is perhaps less potent than it once was, but it nevertheless holds influence. The peer pressure that derives from the "company of equals" (Freidson and Rhea 1963) of the professional collegium remains an important means of securing commitment to making improvement. To the extent that they dealt with socially cohesive occupational groups, the core teams in our study found that they could supplement their efforts to lead change with the more diffuse influence of peers, and in combination, this gave rise to legitimacy and motivation for change among participants. But such influence varied with different professions and even among specialties within the medical profession. The notion of a new professionalism held much more allure for hospital physicians than those in primary care, due to the latter's more diffuse clinical work and their increased subordination to the logics of market and the state, epitomized in the system of incentives to which primary care physicians were exposed. Unlike English hospital doctors, GPs work in small businesses, largely physician owned. In hospitals, the influence of logics appeared to be largely "segmented" (Goodrick and Reay 2011)-that is, logics that might otherwise be in tension with one another could coexist, albeit somewhat restlessly. In primary care, we found what Harris and Holt (2013;cf. McDonald et al. 2013) term interweaving, with a single group affected by multiple logics and thus less able to resist their dominance. The threads of some logics in this weave, however, were more evident than others. Our findings thus suggest that where multiple institutional logics coexist, segmentation and interweaving may have different consequences. While segmentation may mean that the influence of subordinate logics is tangible (in the accounts of clinicians at least and perhaps in their behavior), the influence of subordinate logics may be much less easy to identify where logics are interwoven into the practice of a single group.
More than this, however, our findings suggest something of a shift in the institutional logic of professionalism in health care itself. Institutional logics are not static (Seo and Creed 2002); they may have a protean character, mutating as they evolve. We see some evidence of this in the range of tactics-from seduction to enforcement-adopted by the core teams. The danger, of course, is that if the health care professional logic absorbs so much of the content of other logics that it is no longer distinctive, then it can hardly offer much of a countervailing power. Here, Thornton et al.'s (2012:165) tentative distinction between "assimilation" and "blending" of institutional logics is helpful. Assimilation involves the incorporation of some of the components of one logic into another while "the core elements of the original logic prevail." Blending is a more fundamental change in which "institutional logics are transformed by combining dimensions of diverse logics." Thornton et al. (2012:166) acknowledge that "the difference between blending and assimilation requires further theoretical elaboration"; our findings suggest that the distinction is conceptually crucial but empirically slippery.
Conceptually, there is no necessary contradiction in assimilation: as Kuhlmann (2006) argues, for example, the medical profession's growing adoption of managerial technologies such as performance management through data collection and comparison may be seen as reasserting professional power but also recasting professions as progressive rather than conservative forces. Numerato, Salvatore, and Fattore (2012) develop this line of argument further, proposing that professionals' interactions with managerial technologies should be understood less in terms of necessary opposition between conflicting modes of organizing and more in terms of novel articulations and hybrids. Our data provide support for the suggestion that rather than spelling the end of professionalism, a shift toward greater accountability and management of performance might be better understood as a professionalizing strategy (Green et al. 2011)-a means of ensuring that professionalism retains its moral stature despite the challenges that face it-or as "professionalism finally realized" (Light 2010:283), presenting a means for professions to fulfill their side of their compact with society. But the legitimacy of such a shift with professionals themselves is not, as we have seen, universal.
Viewed in this light, nevertheless, the ingenious complex of approaches to achieving influence adopted in the three cases might be viewed as a sign of the potential vitality of professionalism in the health care system: not so much a rebirth of professionalism despite the power of state, corporate, and market logics but rather its rebirth through constructive interaction with those countervailing powers (Light 2010). Yet, our findings suggest a need for caution in such optimistic pronouncements and highlight the overlap in practice between assimilation and blending. While productive synergies could sometimes be achieved, interaction with other logics could also have more ambiguous consequences for professionalism. In particular, the logics of management and market could easily overwhelm the intrinsic motivation and sense of professional pride and identity that the three projects sought to channel. Even seemingly complementary managerial targets sometimes displaced professional goals or soured good feeling toward those goals so that professionalism became tainted by association for some participating clinicians. Building on Thornton et al.'s (2012) concepts, we suggest on this basis that while it may be possible to incorporate components of other logics while retaining the autonomy of an institutional logic (i.e., assimilation rather than blending), such changes may also have consequences for the balance of competing logics in the wider field: a professional logic that assimilates aspects of the corporate logic offers a weaker counterweight against other logics.
If professionalism's legitimacy varied within occupational groups, then its influence on professional motivation was even patchier across them. The multidisciplinary nature of contemporary health care delivery fractures the affinities clinicians feel not just into professional groups (physicians, surgeons, nurses, etc.) but into specialties and subspecialties. Thus, while ILCOP and AAA-QIP were able to capitalize on their associations with professional societies, they could not always exert influence, pressure, and opprobrium beyond their own boundaries: the influence of institutional logics was segmented (Goodrick and Reay 2011) by subspecialty, specialty, and profession. In consequence, motivation and coordination across professional groups-an important prerequisite for quality improvement-was hard to achieve. Segmentation as well as interweaving can thus limit the scope of influence of an institutional logic. Those seeking to draw on a logic of professionalism should attend not only to making change "clinically led" but also to the question of which clinicians are leading what.
Taken in the round, our findings suggest that a reinvigorated professionalism does hold legitimacy that other logics do not but that features of the logic of professionalism and the composition of the wider institutional field are both crucial in mediating its influence. As Gray (1997:47;cf. Martin et al. 2004) points out, professionalism is not an immanent property of professions themselves so much as a product of "institutional settings that allow the fiduciary ethic of health professionals to exist and flourish." Interaction with other logics may have a capricious impact on the ability of appeals to professionalism to motivate, but so too can the dynamics of professionalism itself; consensus on what it means to whom remains elusive and consequential. What this perhaps suggests most of all is that those seeking to motivate clinicians toward a given end should cherish professionalism but recognize first, that it must be nurtured skillfully if its advantages are not to be undermined and second, that relying solely on exhortations to live up to a vaguely defined professionalism-particularly where other logics are dominant-is perilous.
cOncLUSIOn Redundant, Reinvigorated, or Reshaped Beyond Recognition? Finally, we return to the theorists who have considered the fate of professionalism in contemporary environments where it appears increasingly marginalized. Here we offer three contributions.
First, our findings suggest that in some fields at least, professionalism remains much more than a disciplining discourse for bending professionals to managerial priorities. In Evetts's (2009) terms, occupational professionalism retains its distinctiveness from organizational professionalism. The program studied here was funded and run by nongovernmental organizations and professional societies, and though their influence was variable, they were able to promulgate a notion of professionalism that was clearly not synonymous with unquestioning subservience to managerial priorities. Evetts (2006:137) asks: "why do states allow professions to flourish?" One answer is that there remain things that professions can do better than states-including making judgments about quality of care and encouraging their members to act on these appropriately. This is the line taken by recent policy in England (Secretary of State for Health 2008), and our findings here suggest that it is to some extent realized in practice.
Second, our findings highlight the care needed in finding accommodations with other logics if professions are to avoid, in Freidson's (2001) dystopian vision, losing their souls. Alignments with managerialist and market logics may be productive (cf. Waring and Currie 2009), and as Numerato et al. (2012) suggest, professional and managerial logics should not be understood as necessarily opposed; "hybrid" forms may combine the merits of both. To remain a worthwhile influence, however, even hybridized forms must retain something of the soul of professionalism; care must be taken to ensure that assimilation of components of other logics does not slip into blending.
This brings us to our third and final theoretical contribution. Some argue that the logic of professionalism is changing. Our findings confirm this: conflicts over the content of professionalism and over what it was to "behave professionally" were evident throughout our data. However, also clear from our findings was that to be influential, professionalism must be underwritten by collective, institutionalized arrangements. Such an analysis challenges those (e.g., Noordegraaf 2007:774, 781) who argue that professionalism is a set of embodied characteristics of an individual practitioner-"reflexive practice" or "artistic, intuitive processes which some practitioners do bring to situations of uncertainty"-rather than something determined, negotiated, and operationalized collectively, through professional collegia. Our findings highlight the importance of an institutionalized means of translating professionalism in the abstract into something meaningful to professionals. If anything, the need for professional institutions is greater now than ever, in a context in which individual professional behavior is subject to so many competing influences. The authority of the professional collegium remains, we suggest, central in distinguishing professionalism from other logics.

AcKnOWLEDGMEnTS
We would like to thank participants in the study and three anonymous reviewers for their constructive criticism.

FUnDInG
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the Health Foundation. Mary Dixon-Woods' contribution to the writing of this paper supported by a Wellcome Trust Senior Investigator award (reference WT097899MA) and by University of Leicester study leave at the Dartmouth Institute for Health Policy and Clinical Practice, New Hampshire.
Georgia Herbert is a research assistant in nutrition in the NIHR Biomedical Research Unit in Nutrition, Diet and Lifestyle at the University Hospitals Bristol NHS Foundation Trust and the University of Bristol; she was previously a research associate in the SAPPHIRE Group, University of Leicester. Her research interests cover dietary behavior change, organizational change, and the delivery of health and social care. Her work has been published in journals such as Social Science & Medicine and Health Expectations.
Mary Dixon-Woods is professor of medical sociology, SAPPHIRE Group, University of Leicester. Her research covers health care quality and patient safety, bioethics, and methodological innovation, particularly in relation to qualitative methods and evidence synthesis. A current Wellcome Trust senior investigator, her research has been published widely in social scientific and clinical journals, including The Milbank Quarterly, The Lancet, and Social Science & Medicine.