The Iron Cage and the Gaze? Re-Interpreting Medical Control in the English Health System

Iron cage

The classic version of the iron cage thesis posits a system of hierarchical control which is marked by direct supervision of workers and formal rules. Progression and promotion (especially for professional staff) is based on socialisation into an occupational culture. Knowledge is standardised, enabling a clear division of labour and so, workers are trapped but also entrap themselves within a culture of control (Reed, 1999). I argue that the latter form of control has become significant given the medical colonisation of managerialism. Despite its “repressive and dysfunctional” effects (Kitchener, Caronna, & Shortell, 2005; McKinlay & Arches, 1985), the strength of this heuristic has been its integration of theoretical and empirical, ideological and normative, structural and operational features. However, there is a danger that this conceptualisation is sociological deterministic in that observation of these features necessarily implies an iron cage of control. It overlooks, therefore, the role of human agency in resisting such cage mentality and its associated practices.

The start of the demise the Weberian iron cage is not precise but the advent of neo-liberal policies, such as new public management (NPM) in the latter half of the twentieth century has been considered a precipitating factor (Ferlie, Lynn & Pollitt, 2005). However, this must be viewed alongside a raft of other contemporary developments such as the crisis in the welfare state (of Western democracies), the advent of new forms of information communication technologies (ICT), and shifting cultural values (including disillusionment with hierarchy and a growing acceptance of consumerism and market-based approaches) (Exworthy, Powell, & Mohan, 1999). Notably, ICTs were instrumental in delivering a transformational impact on hierarchical structures and processes, overcoming many of the rigidities of the iron cage (Lash & Urry, 1987). ICT enabled, for example, performance data to be compared, audited and costed (Power, 1997).

In terms of professionalised occupations, this period also saw the emergence of a narrative of professions whose influence was seen as increasingly malign; professionals were regarded as self-serving interest groups whose actions were not always coincident with those of service users or the wider public (Elston, 2009; Freidson, 1970). However, generally less well observed in applications of the iron cage thesis has been the applicability and relevance to the (medical) profession. Doctors (in most jurisdictions) have traditionally enjoyed significant autonomy over the technical and operational aspects of their work (Harrison, 1999). Until relatively recently, doctors have been able to avoid the institutional, organisational and managerial imperatives which have often been applied to other occupations, including non-medical ones. Doctors have also enjoyed a high degree of self-regulation (at a macro and micro level), rather than external regulation. Thus, there has been limited evidence of a “monolithic iron cage of physician subjugation” (Kitchener et al., 2005, p. 1320).

However, the cage thesis still retains relevance and arguably has become more significant. At first sight, medical autonomy might seem to contradict the notion of centralised authority (implied by the iron cage thesis). Yet, recently, central authority has been re-cast in three ways. First, NPM represented a “widespread shift in the legitimacy of management” (Exworthy & Halford, 1999, p. 5) which privileged managers (vis-à-vis professionals). NPM also entailed the creation of a cadre of local managers who were often seen as agents of the (hierarchical) centre, providing a stronger line of local implementation of national policies (Harrison, 1998). Market-based competition for health services may have challenged further medical power. Secondly, the emergence of evidence-based medicine (EBM) (in the 1980s and 1990s) represented a new (and countervailing) paradigm of central knowledge to which normative and cognitive behaviours of medical professionals are increasingly expected to conform (Armstrong, 2002; Harrison, 1998). Third, the decline in trust in professions as a result of a series of “scandals” (especially in medicine; Dixon-Woods, Yeung, & Bosk 2011) eroded the profession’s claims to maintain autonomy and be self-regulating; in doing so, they have arguably ceded (some) power to external interests. However, although different medical specialties have enjoyed contrasting experiences of managerialism, the medical profession (as a whole) remains reasonable well established and resilient to external threats (Timmermans & Oh, 2010).

Arguably, as a result of recent developments, the notion of the cage has not become redundant but rather, recent reforms have transformed and intensified it (Farrell & Morris, 2003). Although they do not advocate a return to a form of Weberian bureaucracy, they recognise that it has some merits which have often been decried in recent neo-liberal reforms (see also Pollitt, 2009).

Farrell and Morris note three developments which have prompted a revision of the bureaucratic mode of control: interaction between professionalism and managerialism, managerial resistance to managerialism, and professional heterogeneity. The first refers to the apparent tension between different values and accountabilities of these discourses but also to the accommodation between them (Exworthy & Halford, 1999). Second, it may be that local managers do not fully enrol in managerial projects (especially those instigated by the centre); instead, they find local arrangements to work with professions (Macfarlane, Exworthy, Wilmott, & Greenhalgh, 2011). Cultural self-management may, however, become normalised amongst doctors. Third, the internal professional heterogeneity reveals stratification. One such division is medical hybrids within institutions (McGivern, Currie, Ferlie, Fitzgerald, & Waring, 2015). Whilst the managerial/administrative elite of the profession has long been apparent (Freidson, 1994), their role and significance has been amplified by recent managerialism. This also signals internalisation of managerial discourses within segments of the profession. Thus, the bureaucratic controls may not have disappeared but rather have been transformed and even intensified (Farrell & Morris, 2003).

The gaze

Typically, the gaze consists of various managerial strategies whereby control is internalised within norms and behaviours of occupations, individually and collectively. The disciplines of (especially) market-based discourses thus become inculcated within occupational logics and thereby become unchallengeable. Drawing on earlier work by Bentham, Foucault’s (1979) characterised the gaze in terms of the panopticon—the remote and supervisory observation of many prisoners by guards (or even one guard). However, it has also come to refer to the applications of this surveillance such as the hierarchical and comparative ranking of subjects (as in “league tables”), total quality management (Townley, 1994), and the public reporting of CEO salaries, mortality rates associated with individual surgeons and school performance (Pawson, 2002).

The effect of the gaze, it is claimed, transforms “recalcitrant, disordered and potentially threatening individuals into docile and obedient subjects” (Reed, 1999, p. 29) through separation, remote observation, hierarchical ranking, and normalisation. Individual agents are thus transformed into “calculable selves”; individuals become subject to the demands of the external logic (here, managerial) and the objects of it (by becoming individually committed to evaluating their own practice). Miller (1992) uses accountancy practices within public sector managerialism to argue that: “The calculative technologies of accountancy [and arguably, other panoptic approaches] hold out the promise that one can transform all individuals, whether they be managers, workers, doctors, or teachers into calculating selves” (p. 70).

ICT has been instrumental in creating and shaping the ways in which the gaze has been enacted. ICT offers the potential for more flexible processes of control in which such (managerial) discipline (especially in relation to managing medical performance) can be internalised by individuals (Reed, 1999). “ICTs seem to possess the inherent capacity for delivering an organisational control regime in which self-regulation and discipline… is a realisable project for management” (Reed, 1999, p. 24).

This self-discipline overcomes the limitations of continuous surveillance by a superordinate as control is internalised. ICT thus has the “benefit” of providing remote and unobtrusive monitoring (of performance) and cultural self-management (Kitchener & Exworthy, 2008). No specific external managerial intervention need thus be imposed. One of the limitations of the iron cage was its inability to effectively monitor or control activities which are “invisible” to the managerial gaze (such as doctors) (Kitchener et al., 2005). This is the essence of Foucault’s interpretation of the panopticon: “The surveillance is permanent in its effects even if it is discontinuous in its action” (Foucault, 1979, p. 201).

Responses by public service professionals to post-modern, and neo-bureaucratic, modes of control have evolved from binary tensions to more contingent responses, mediated by the type of control, local contextual factors and professional trajectories. Such contingent responses by professionals have included adoption of managerial discourses and perception of their identity, career and professional norms into terms of such discourses (McGivern et al., 2015). This panoptic gaze has, for example, been applied to doctors through performance management such as pay-for-performance schemes (McDonald, Checkland, Campbell, & Roland, 2007) and public reporting of their mortality rates (Gabe, Exworthy, Jones & Smith, 2012). It has also helped reveal poor (and even criminal) clinical practice (Dixon-Woods et al., 2011; McGivern & Fischer, 2010). The notion of the gaze is also helpful because it is applicable to the growing self-reflexivity of the medical profession. In the context of quality and patient safety initiatives, such reflexivity of the gaze might also engender a “micro-sociology” rather than questioning the “over-arching governance and structuring” of health care organisations or institutions (Iedema et al., 2006).

The notion of the gaze has limitations. The ideal-typical notion of “one” observing “the many” needs to be reversed in the case of the online panopticon; a limited number of agents (doctors) are being watched by (potentially) a vast number of “observers,” not simply health-care managers but patients and the public too (Lyon, 1993; Zuboff, 1988). Their ability to understand and interpret the data about medical performance is likely to be limited as will be their willingness to act upon it (Shekelle, Lim, Mattke, & Damberg, 2008). The continual generation of information creates a “tyranny of light” (Tsoukas, 1997) which may obscure as much as it enlightens; for example, the increasing volume of data about medical performance may obfuscate patient or managerial decision-making and create a sense of misplaced trust. Also, the gaze may be unable to reach into inter-personal aspects of professional performance. Some aspects of such practice are ineffable and not easily reducible to standard measures amenable to comparison (Exworthy et al., 2003). Crucially, the medical profession has played a key role in defining the creation and implementation of gaze-like performance measures, not least through the interpretation of esoteric knowledge.

Medical performance

Traditionally, the definition and measurement relating of medical performance remained solely the remit of the profession itself (Causer & Exworthy, 1999). Doctors were only deemed eligible to monitor their own performance, either individually or collectively; the essence of peer review. Patient, managerial and political interests were largely excluded from commenting on or taking action on cases of (poor/bad) medical performance. Performance data were not usually shared with colleagues and less likely, with external stakeholders. Although doctors themselves would often privately claim that they “knew” about such performance, there was a professional ethic which stifled open/public criticism of their peers; Causer and Exworthy (1999) term this a “notion of equality of competence.” Action to rectify “poor” performance was instigated and administered by the medical profession (Rosenthal, 1995).

Central to doctors’ ability to maintain control over their own performance was the extent to which their medical practice was not amenable to tight managerial control and/or was “hidden” from external view, or gaze. The operating theatre or the physician’s office were thus private spaces which helped sustain their autonomy. The advent of ICT (notably the internet) has provided the opportunity to store, accumulate and compare data about the performance of doctors (Waring & Currie, 2009). Consistent with the audit society thesis (Power, 1997), ICT—as new “softwares of control” (Fournier, 1999)—enables stricter application of performance measures and also greater panoptic scrutiny of doctors (Noordegraaf, 2011).

A particular expression of the audit society and the application of ICT has been public reporting of medical outcomes (Gabe et al., 2012; McGivern & Fischer, 2010). Public reporting is part of transparency which has become “a widespread normative doctrine for the conduct of governance” whose value is rarely questioned (Hood, 2007, p. 192). It involves the publication of the medical performance associated with named individual doctors. Cardiac surgery has been the pioneer of public reporting. Publication has previously been in paper-based reports but is increasingly disseminated via the internet (Bates & Gawande, 2000). Public reporting has tended to record aspects of performance, usually mortality rates (rather than patient experience although this is now being introduced in some specialties). Also, previous anonymised reporting systems have become transparent by identifying named doctors.

However, the notion of performance becomes problematic; the published performance data are not a factual account of a doctor’s actions but a discourse which conveys particular interests. For example, public reporting can mask the contribution of the wider clinical team for which the senior doctor is responsible. As health services have become subject to market competition, such comparison have become useful for patients as consumers. Public reporting facilitates control of surgeons since surgery might be seen as more easily measurable than some other clinical services (Katz, 1998; McDonald, Waring, Harrison, Walshe, & Boaden, 2005). However, the impact of control over doctors which is exercised by patients (acting as consumers) and managers (acting as regulators of medical profession) is not fully understood (Gabe et al., 2012; Shekelle et al., 2008). Public reporting can also provide patients (as consumers) with information upon which they might “choose” between competing providers; yet, much evidence (mainly in the US and UK) suggest that patients do not choose in this way (Shekelle et al., 2008). Public reporting may thus be as much a mechanism to convey and amplify peer reputation, than to control the profession (Bevan & Wilson, 2013).

Although the measure of performance in cardiac surgery (30 day post-operative mortality rate assigned to the lead surgeon) is still defined and measured by the profession, there has been some resistance by doctors to the involvement of external agents in the surveillance (Gabe et al., 2012; McGivern & Ferlie, 2007). However, the shift towards mandated participation in public reporting in ten medical specialties in England (since 2013) points more towards bureaucratisaton (iron cage) combined with a post-bureaucratic (gaze) modes of control. However, the presence of doctors (especially cardiac surgeons) in key policy-making positions in the government ministry (Department of Health) and a regulatory agency (Care Quality Commission; CQC) has moderated the extent of the cage. Furthermore, the CQC’s collaboration with the cardiac surgeons’ professional body in presenting these mortality data on the internet has also affected the salience of the gaze.

At the same time that public reporting is becoming normalised within the (surgical) profession, questions have been raised about the value of mortality rates as a measure of performance. Whilst apparently precise measures (such as mortality) have enabled the spread of public reporting due to their comparability, a standardised, comparable measure of mortality inevitably neglects ineffable practices (such as consultation skills) or unmeasured aspects of care (such as patient experience). Hence, public reporting may not act as any sort of cage and may present a “narrow gaze” (which has been defined largely by the profession).

Public reporting does not rely simply on bureaucratic techniques of direct control (associated with the cage) or a narrow gaze (described above) but more on transforming doctors as “calculable selves” (associated with the gaze). While scholarly and professional attention focuses on “micro-circuits” of surveillance (Iedema et al., 2006; Reed, 1999), there is a danger that counter-vailing forces of the professional power are neglected. Using a narrow gaze upon their performance, surgeons have been able to present a reassuring assessment of their own performance. This might signify that public reporting is being used by the profession to restore trust and promote a new professionalism.

Financial incentives relating to medical pay

Although pay is the largest single item in financing health systems, the issue of doctors’ pay has received relatively little attention (compared with other English health policy issues). The example of financial incentives relating to the pay senior doctors (“consultants”) in the English NHS provides a pertinent application of models of medical control. Consultants are mostly hospital-based doctors who are responsible for the delivery of health-care to assigned patents. By contrast, English primary care doctors have different contractual arrangements (McDonald et al., 2007).

In 2011, there were over 47,000 consultants in England (Review Body on Doctors' and Dentists' Remuneration, 2013). This was equivalent to 4% of the NHS workforce but 13% of the NHS budget. Consultants are not evenly distributed with current shortages in emergency medicine and in poorer, deprived areas. Three-quarter of consultants’ pay consists of “basic pay” (salary) which was worth £88,732 (€111,590) (mean annual basic pay) in 2013-14 (Health and Social Care Information Centre, 2014). Additional payments for specific activities (such as weekend working) are also paid. They are also eligible to apply for Clinical Excellence Awards (CEAs).

CEAs were established in 1948 at the inception of the NHS. Their ostensible purpose was to ensure that the medical profession participated in the newly-established government-run institution. Otherwise, doctors might be tempted to practise their medicine privately. The awards are currently given to consultants who meet criteria across five domains, indicative of a quality of work “over and above” contractual obligations, as determined by a system of local and national committees. Awards range in value from £3,000 (approx. €3,785) to £75,000 (€94,640) per annum across 12 levels, which are paid in addition to basic salary. As a result of CEAs, the median annual total earnings of a consultant are £109,000 (€137,545) (2011-12) (NAO, 2013). The overall cost of CEAs is over £500million (€630,939million) per annum.

The five CEA domains of medical performance include: delivering a high quality service, developing a high quality service, leadership and managing a high quality service, research and innovation, and teaching and training. The criteria for allocating awards have changed over time but since 2001, greater emphasis has been placed on “delivering and improving local and health services,” which includes participation in local managerial and organisational decision-making processes (among other things). This emphasis has been complemented by rising managerial and lay representation on award committees. Although chaired by a lay member and including managerial representation, doctors still account for 50% or more of the representation of these committees.

One explanation for the continued implementation of CEAs (and their financial benefits to doctors) for over 60 years is social closure (Weedon, 2002). The profession and state are in an inter-dependent relationship relating to the content of work, training, performance appraisal and pay, for example (Klein, 1990). Doctors have secured and maintained a monopoly of medical labour through state (legislative) protection. However, the state (in the form of the NHS) has arguably depressed doctors’ incomes by being the main employer of medical labour (Klein, 1990).

The case of CEAs illustrates weakly the application of the cage thesis. Whilst the introduction of (the forerunner of) CEAs was a bargain struck between the state and medical profession, the original application of autonomous peer review remains largely intact. Recent media interest in doctors’ pay has occasionally mentioned CEAs but, despite medical scandals, there has been no concerted attempt to link CEAs to the more toxic issue of bonuses in banking or high salaries in the public sector. Also, information about CEAs is available on the internet but the absence of debate perhaps denotes minimal external gaze; CEAs thus remain “hidden in plain sight.” Arguably, over time, CEAs have become more aligned with the wider (regulatory and economic) interests of the state. Changing criteria for awards and committee membership both denote a shift towards wider organisational/corporate objectives and discourses. That said, three aspects lessen bureaucratic control over doctors in this example. First, the infrastructure of health policy-making is riven with medical staff, not least award committees. This arguably nullifies the extent of external threats to their position. Moreover, the comprehensive application of rules by which CEAs are administered might be considered a bureaucratic hallmark of this mode. Yet, financial support for CEAs has been maintained despite challenging fiscal climate elsewhere in the NHS. Second, the criteria for awards remain largely medically determined and applied, with relatively little influence from managerial or lay interests. Data about individual medical performance data (pertaining to the five domains) are passed to hospital chief executives who are required to offer support for applications to higher level CEAs. However, such information only provides a moderating influence on committee decisions and can hardly be interpreted as significant extension of control over the profession. Third, the coverage of CEAs to (currently) over three-fifths of all consultants creates (or perhaps reinforces) a sense of entitlement to these financial rewards among doctors, rather than simply a reward for their “excellent” medical practice. This might have a socialising effect on junior doctors. CEAs act, therefore, as an incentive to maintain medical collegiality and adherence to professional values, rather than controlling and managing their performance.

CEAs could also be interpreted in terms of the gaze. Certainly, some of the domains (which doctors must address to be awarded a CEA) have, over time, been re-defined to reflect a more managerial orientation. This might support the notion that a managerial gaze is being normalised within the profession. The third CEA domain, for example, refers to demonstrating a sustained commitment to the values and goals of the NHS. This might be evident in terms participating in the annual process of agreeing doctors’ “job plans” with the hospital, observing the guidelines which regulate the amount of private medicine which they (as NHS employees) can undertake, and showing a commitment to achieving agreed service objectives. Such evidence denotes the cultural self-management associated with the managerial gaze. Doctors’ behaviours might thus become aligned with the managerial definitions in certain domains. However, doctors do not necessarily need to perform well on this domain alone. Therefore, the specific interpretation of these behaviors (by CEA committees, dominated by doctors) might simply reinforce a medical perspective, rather than challenge it.

Re-interpreting the cage

The ideal-type of the iron cage applies weakly to the entirety of public services and health care. Given the professional autonomy of doctors, it never was a “monolithic” approach to control of this occupation. This underlines the need to re-assess it and apply it more appropriately. In health-care, however, the demise of the cage has been over-stated (Farrell & Morris, 2003) but it would be misleading to claim that the cage is no longer evident or relevant. Aspects of the cage have been and remain evident in health care institutions; the salience of each aspect has been contingent upon the institutional and professional setting. Indeed, certain aspects of the cage have become intensified in terms of implementation and salience. Such neo-bureaucracy has not removed cage-like apparatus but rather extended its application. In doing so, it has revealed many of the dysfunctional consequences and contradictions of the forms of NPM which have been implemented in the UK and elsewhere (Farrell & Morris, 2003).

Over the past 25 years, various forms of neo-bureaucracy have been implemented in the English NHS. Notably, the introduction of market-like relations and competition in 1991 (and the variations thereof since) have disaggregated health care institutions into semi-autonomous units. Whilst this might appear to have decentralised decision-making in the NHS, it has equally established more centralised apparatus in the form of regulation of the market (Greener, Exworthy, Peckham, & Powell, 2009). This paradox of simultaneous decentralisation/centralisation has not reduced hierarchy but increased bureaucracy (Farrell & Morris, 2003; Peckham, Exworthy, Greener, & Powell, 2005).

An illustration of the decentralisation/centralisation paradox is the use of financial incentives to control the performance of medical professionals. Whilst the use of financial incentives has continued apace in the last decade, Clinical Excellence Awards have remained largely untouched since 1948. The initial resistance to financial incentives has, for the most part, moved to acceptance as primary care doctors, for example, have benefitted significantly (McDonald et al., 2007). However, it doing so, medical legitimacy to set and monitor standards and crucially reward their members, has been challenged by NPM strategies (of the cage). Medical authority may not have been usurped but their ability to control their profession has been significantly eroded. Farrell and Morris (2003) may thus only been partly correct in referring to “persecuted professionals” (p. 135).

Re-interpreting the gaze

As transparency has become a doctrine for good governance in many health systems (Hood, 2007), there has been a growing surveillance of medical practice, with the widespread application of strategies associated with the gaze. In particular, the connection between surveillance (gaze) and the “performance paradigm” of NPM has been pervasive (Bevan & Hood, 2006; Exworthy, 2010).

Often, professionals have “de-coupled” from surveillance strategies linked to performance measurement. Professionals have often adopted a “tick box mentality” to fulfilling the requirements of the technique, without changing their own work practices (Levay & Waks, 2009; McGivern & Ferlie, 2007). Over time, there has been a growing normalisation of NPM strategies through the professional colonization by its logics. This tends to involve greater self-monitoring and normalisation. Whether this represents a loss of medical autonomy or rather a re-definition of it depends much on the context into which these strategies are implemented. For example, more junior doctors have tended to inculcate the logic of self-monitoring more readily than their senior colleagues (Gabe et al., 2012).

Whatever the consequence for medical autonomy, it is apparent that two consequences have arisen: self-fulfilling prophecy and commensuration (Espeland & Sauder, 2007). The former refers to the changes in behaviour in response to predictions whilst the latter refers to the transformation of (often disparate) properties which are quantified and measured.

Doctors’ responses to financial incentives might illustrate the former while the development of hierarchical rankings of performance (which are then publicly reported) illustrates the latter. Both of these consequences reflect the sense in which market-based relations have pervaded medical performance and medical pay. As a result, individuals and institutions increasingly need to be self-aware and re-position themselves in relation to their competitors. Managerial and professional interest may thus become elided.

Cage and gaze re-visited

Conceptualisations of the confluence of control (through cage and/or gaze) and professional autonomy have become key areas of inquiry recently. Techniques (such as public reporting of medical performance or the use of financial incentives for health-care professional) need to be vague enough to be adaptable to local (professional) circumstances but sufficiently testing to capture the attention of professionals. This is the essence of Courpasson’s (2000) soft bureaucracy. This implies a more flexible, “plastic” cage rather than a rigid, iron one. Equally, it reflects a two-way surveillance whereby doctors are coupled to be managerial strategies but continue to enjoy some latitude in defining and implementing them. By and large, doctors continue to set the terms by which their performance is defined and rewarded in the CEA scheme.

The significance of the “soft bureaucracy” in institutional change is especially relevant given the enduring and malleable power of the medical profession. Equally, Scott’s (2008) institutional pillars indicate the consequences in medical performance and financial incentives. For example, Scott’s “cognitive pillar” denotes the “symbolic systems” of professions. Here, this might include the reputation effect of public reporting (Shekelle et al., 2008) and the use of CEAs as honorific symbols of the medical profession (Frey, 1997). Both are consistent with Frey and Neckermann’s (2009) analysis of non-financial recognition. Collegial control (cf. managerial control) thus becomes significant as the professions seeks to define and sustain its “own” definition and prescription of reward for its members.