Study on Resisting Hybridization between Modes of Clinical Risk Management

Description
Study on Resisting Hybridization between Modes of Clinical Risk Management: Contradiction, Contest, and the Production of Intractable Conflict:- Risk management is the identification, assessment, and prioritization of risks (defined in ISO 31000 as the effect of uncertainty on objectives, whether positive or negative) followed by coordinated and economical application of resources to minimize, monitor, and control the probability and/or impact of unfortunate events

Study On Resisting Hybridization Between Modes Of Clinical Risk Management: Contradiction, Contest, And The Production Of Intractable Conflict Abstract

This article explores and explains escalating contradictions between two modes of clinical risk management which resisted hybridisation. Drawing on a Foucauldian perspective, these two modes - ethics-orientated and rules-based - are firstly characterised in an original heuristic we develop to analyse clinical risk management systems. Some recent sociologically orientated accounting literature is introduced, exploring interactions between accountability and risk management regimes in corporate and organizational settings; much of this literature suggests

1

these systems are complementary or may readily form hybrids. This theoretical literature is then moved into the related domain of clinical risk management systems, which has been underexplored from this analytic perspective. We note the rise of rules-based clinical risk

management in UK mental health services as a distinct logic from ethics-orientated clinical selfregulation. Longitudinal case study data is presented, showing contradiction and escalating contest between ethics-orientated and rules-based systems in a high-commitment mental health setting, triggering a crisis and organizational closure. We explore theoretically why perverse contradictions emerged, rather than complementarity and hybridisation suggested by existing literature. Interactions between local conditions of strong ideological loading, high emotional and personal involvement, and rising rules-based risk management are seen as producing this contest and its dynamics of escalating and intractable conflict. The article contributes to the general literature on interactions between different risk management regimes, and reveals specific aspects arising in clinically based forms of risk management. considering some strengths and weaknesses of this Foucauldian framing. It concludes by

Keywords:
clinical risk management; conflict; ethics; Foucault; hybrids; mental health; organizational change; paradigm incommensurability; regulation

Introduction: The interrelationship between two modes of clinical risk management - explaining non-hybridisation, escalating contradictions and organizational crisis.

Formal risk management systems now provide a dominant and pervasive logic for governing an uncertain social world (Power, 2004). Such systems have expanded and colonised terrains previously occupied by less formalised self-regulation, including self-regulation by professionals. They have proliferated in UK government, regulatory agencies and public 2

services (Power, 1997; 2004), as well as private firms. These systems promise a means of dealing with potential hazards as what begins as a mere possibility of danger is converted into calculable objects of surveillance, regulation and control (Castel, 1991; O'Malley, 2004; Power, 2007).

Yet risk management's claim to calculation and objectivity may overlook local values, emotions and practices concerning social transgressions, rule-breaking and deviancy. For instance, by putting individuals' selves and feelings 'at risk', formal risk systems may involve 'cleaning-up' accounts for presentation to external auditors; they may encourage individuals to hide malpractice, 'game' reporting systems and undermine corrective learning (Gabe, Exworthy, Jones & Smith, 2012; Iedema, Flabouris, Grant & Jorm, 2006; McGivern & Ferlie, 2007; McGivern & Fischer, 2012; Waring, 2009). Risk may function above all as a moral idea in which the selection, handling and elaboration of risk functions to protect authoritative moral orders and risk management regimes meant to uphold them (Douglas, 1992). According to this 'risk as moral government' perspective, risks are not ontological facts, but social constructions where omissions, wrong-doing and blame are attributable to persons held accountable (Douglas, 1992; Luhmann, 1993).

If formal risk management systems operate as a form of moral government, they may also interact with indigenous risk practices and mentalities as individuals orientate themselves towards authoritative, external evaluations of their conduct (Ericson & Doyle, 2003; Ewald, 1991; Power, 2004; 2009a). According to Foucault's (1979) original concept of

governmentality, such an orientation towards risk may lead to an internalisation and strengthening of its rationalities, whether through compliance, participation, or even resistance (Gordon, 1991).

But what are the empirical dynamics of interactions between 'indigenous' risk management practices and formal risk management systems? Within the sociologically orientated accounting literature, recent scholarship indicates manageable tensions (Gendron, 2002; Rahaman, Neu &Everett, 2010), complementarity (Roberts, 1991) and ready hybridisation (Miller, Kurunmäki & 3

O'Leary, 2008) between different accountability and risk management regimes. By contrast, Armstrong (1994) suggests some potential contradictions arising between different discursive systems. Yet overall, this literature does not suggest conflictual interactions between formal risk management and indigenous risk management systems.

In contrast to this literature, we argue that interactions between alternative risk management systems may exert perverse and intractable effects, not previously adequately considered. Drawing on Foucault's (2010, 2011) recently published final lectures at the College de France, we develop an original heuristic to explore interactions between a rules-based mode of regulation, advanced by formal risk management systems, and a contrasting ethics-orientated mode more embedded in indigenous clinical practices. Whereas Foucault (1992:25) defined morality as a 'systematic ensemble' of values and rules of conduct presecribed to individuals through authoritative institutions, he contrasted these rules-based 'moral codes' with the different ways in which individuals might interpret and relate to them. Individuals may not merely conform to rules, but seek to constitute themselves as 'ethical subjects' through practices intended to transform their thoughts, emotions, and ways of being.

We apply our heuristic in an empirical case of a high-commitment health care organization where perverse interactions between contrasting modes of risk regulation are exemplified. Through a longitudinal case study of a mental health care setting - a Democratic Therapeutic Community (DTC) - we explore dynamics between rising formal clinical risk management systems and pre-existing self-regulation, clinically embedded. Whereas interactions between these modes are likely to be important in a number of settings, we propose the DTC may be an 'extreme case' (Eisenhardt, 1989) human service organization, well-suited for studying these interactions that may be less apparent in other settings. Contrary to much literature, this case reveals strong tension between the two modes of regulation, leading to escalating morallycharged conflict which ultimately, we suggest, triggers a crisis and organizational closure.

Our article contributes to the sociological accounting literature, firstly, by elucidating perverse interactions between formal risk management systems and indigenous risk management 4

practices. Secondly, we develop a sociological perspective on a related field of clinical risk management systems. Some sociologically orientated accounting literature examines corporate and financial accountability or risk management systems, including some health care settings (Miller et al., 2008; Miller & Rose, 2008; Rahaman et al., 2010). We apply these perspectives to the particular domain of clinical risk management. As clinical risk management involves

significant first order risks (mainly to service users and clinicians), as well as second order, reputational risks (particularly to managers and organizations, see Power, 2007), we suggest this context reveals some perverse interactions, previously overlooked.

The argument proceeds as follows. Firstly, we introduce our Foucauldian heuristic, situating the discussion theoretically in the sociologically orientated accounting literature on interactions between regulatory regimes. The growth of formal clinical risk management systems in UK mental health services is outlined and we introduce the DTC as a distinctive clinical setting. We then describe our ethnographic research design and empirical case study, revealing escalating tensions between self-regulatory practices and a rising formal risk management system. We find contradictions, contest, and no easy hybridisation. These findings are discussed theoretically in relation to our heuristic. We conclude by considering the strengths and weaknesses of the Foucauldian framing adopted, and suggest ideas for further study of interactions between risk management regimes.

Literature review and theoretical emplacement
A Foucauldian heuristic: Two contrasting modes of clinical risk management
When discussing his core concept of 'governmentality', Foucault explores the developing capacity to govern populations indirectly, through novel knowledge bases (including psychiatry), segregated institutions (including the asylum), and associated micro practices, such as systems of registration and accounting (Foucault, Burchell & Gordon, 1991). A governmentality

perspective is thus promising in analysing regulatory regimes in mental health care. Yet we are interested in how Foucault's thought evolved, especially in his recently published final lectures 5

on The Government of Self and Others (Foucault 2010; 2011). Whereas early Foucauldian analysis focused on technologies of power-knowledge and its internalisation by docile subjects (Foucault, 1977), he became increasingly interested in how subjects form a relationship with themselves, whether as subjects of disciplinary power or potentially as 'intensely free', selfactualising subjects (Foucault, 1988; Veyne, 2010). Foucault's later libertarian ideas explored distinctive themes of personal ethics, desire and self development, realised through freelyembraced self discipline (Starkey & McKinlay, 1998), linked to social praxis (Flyvbjerg, 2001).

Through exploring shifts in notions of morality between Greco-Roman, early Christian and modern secular periods, Foucault (2005) distinguished between ethics-orientated practices of personal conduct, important in antiquity, from rules-based codes prescribed and mediated by authority, more dominant in later periods (Kosmala & McKernan, 2010).

These contrasting schemas involve distinct forms of subjectivity. In the ethics-orientated mode of Greco-Roman culture, subjects may seek to actively constitute themselves as ethical subjects through ascetic exercises to establish an ethical foundation for engaging with the social world (Kosmala & McKernan, 2010). As Foucault (2010; 2011) argues, this ethics orientation seeks ethical self-government in oneself and others, accomplished through relations of care. Omissions and errors are regarded not as breaches of moral codes, but intrinsic to formative learning. By contrast, the code-orientated mode, more dominant later, involves governing

oneself and others through truth obligations in the form of rules mediated by authorities. 'Moral conduct' means becoming a subject of external truth, involving self-renunciation, conversion and rituals of atonement (Foucault, 2005; Foucault & Blasius, 1993). Subjective and experiential knowledge is subjugated to ideals of objective truth, requiring self-examination, confession, and 'deciphering' of the subject by authorities.

Although these two modes are historically situated, Foucault (1996) considered them relevant to contemporary forms of government. Subjugation to external authority involves internalisation of a dominant moral discourse, where subjects put their confidence in credible and authoritative notions of truth and attempt to self-consciously reform themselves through adherence to its rules. 6

However, an ethics-orientated mode seeks to produce more autonomous individuals, selfconstituted through more contested and 'agonistic' relations to domination (Foucault, 2005; 2010). This ethics-orientated mode entails social practices that involve neither subjection to moral codes, nor retreat from the social world, but seek ethical government through intersubjective relations of care, along with freespoken, practical critique (Foucault 2011).

We suggest these two modes of Foucauldian subjectivity offer a heuristically useful lens to explore contrasting forms of clinical risk management. According to Foucault (1992:24), risk handling may involve not merely subjects' compliance with rules of conduct, but practices involving a 'whole mental endeavour' of vigilance, intention and attitude, designed to avert dangers and contribute to the security of social groups. Whereas formal risk management systems emphasise regulatory control, informal, indigenous forms entail active co-production with service users, involving self-regulation.

How might we operationalise such a heuristic? Firstly, an ethics-orientated mode suggests linked clinical and social practices may 'transform' the self through care of oneself and others, seeking to develop therapeutic intersubjective relations. Clinical risk is understood and managed interpersonally, emphasising intense engagement in ethical relations, accompanied by active reflection and personal responsibility. It involves higher tolerance of risk, regarded as providing opportunities for personal learning through interpersonal feedback and challenge. Clinical risk management is here inherently co-produced between participants who might genuinely seek personal transformation and development.

By contrast, a rules-based mode suggests adherence to explicit rules and calculations of probability, with recording and reporting to external authorities. Clinical risk is understood and managed as an expert technology, practised by clinicians and managers. Rules-based clinical risk management distrusts experiential, subjective knowledge, and service users would be regarded as unpredictable and potentially dangerous. They would not be expected to

authentically engage in therapy, so requiring expert supervision and management. This mode involves low tolerance of deviance and risk. Second order evidence of risk management is 7

emphasised (Power, 2009b), involving demonstrable self-examination, confession of 'deviance', and externally verified conversion. Clinical risk management here emphasises internalisation of rules-based codes by service users, clinicians and managers to increase self-conscious compliance with dominant ideals.

Whereas these contrasting modes involve different kinds of subjectivity, interactions between them may be more nuanced than this heuristic suggests. As Kosmala and McKernan (2010) write, a self-constituting, ethics-orientated subject is never entirely independent of moral codes, as social frameworks are "proposed, suggested, imposed upon him by his culture, his society and his social group" (Foucault 1996: 41). Whereas we have so far outlined distinct modes, some scholars suggest these might come together and blend in interesting ways (see Kurunmäki, 2004; Miller et al., 2008). Townley (1994), for instance, argues that disciplinary practices in Human Resource Management (HRM) may take a rules-based form that objectifies individuals as examined and inscribed subjects, as well as an ethics-orientated form emphasising subjective experience, intersubjectivity and minimal domination. But HRM practices may also take a third approach, aligning organizational objectives with individuals by attempting to produce selfmanaging, 'productive subjects'. Such hybrid practices seek to increase and make visible selfconscious and reflexive knowledge, whilst experiential and relational self-regulation is subjugated in favour of (internalised) expert knowledge (Townley, 1994).

According to these authors' perspectives, ethics-orientated and rules-based modes of regulation might be predicted to combine to produce possible hybrid forms. We capture in Figure 1 the main differences between these modes of clinical risk management, highlighting four main distinctions, derived from our reading of Foucault's (2010; 2011) final lectures on The Government of Self and Others. These distinctions are (a) their contrasting truth discourses and (b) 'practices of the self', in which (c) subjects' reflexivity is sensitised towards alternative sources of self-knowledge, and (d) where intersubjective relations are variously regarded as either important or untrustworthy aspects of managing substantive risks. But what would

possible hybridisation look like in our case? For the purposes of developing our heuristic, we 8

speculatively outline in this figure a possible third mode, drawing on the on the hybrid regulation ideas of Miller and colleagues (2008; Kurunmäki, 2004).

Ethics-orientated Truth discourses
Truth of clinical risk is 'discovered' through subjective knowledge, coproduced with others. Risk management is through self-mastery and inducing ethical forms of government in self and others. Practices of risk management link tolerance of deviance with selfdevelopment. Social praxis is integrated with shared learning.

Rules-based
Ideals of objective truth and calculable knowledge are codified and distrust subjective knowledge. Risk management is through adhering to rules & ensuring compliance by others.

Possible Hybrid
Blended truth discourses assimilate code-based truth with subjective knowledge. Risk management is through internalising external rules, assimilating them into the care of oneself and others.

Practices

Risk management is seen as an expert technology, requiring recording and reporting of deviance to experts. Self-development involves conversion to external ideals.

Practices are readily combined and internalised, hybridising expert technologies with indigenous practices. Low tolerance of deviance produces a proliferation of techniques, with rules-based practices more dominant. Reflexive sensitivity to horizontal relations is mediated by selfconsciousness towards authorities' perspectives. Self-examination and selfmonitoring draw upon external rules. Authentic intersubjective relations are balanced by an orientation to authority relations. Responsibility for risk management is promoted through ensuring mutual adherence to external rules, and punishing deviance.

Reflexivity

Reflexivity is sensitised towards 'horizontal relations'. Social relations and relations with oneself are reflected on in the service of mutual learning and self-development.

Reflexive selfconsciousness is directed 'vertically', towards the perspective of external authority. Self-examination and 'confession' to authorities are mediated through internalised rules. Intersubjective relations are held as untrustworthy and subject to sanctioned authority relations. Personal accountability is stressed, involving transparent selfrevelation to authorities, attributing blame, and visibly punishing culpability

Intersubjective relations

Authentically engaged intersubjective relations are held as key to substantive clinical risk management. Mutual responsibility is promoted through active and outspoken care of oneself and others.

Fig. 1 A Foucauldian heuristic: three modes of clinical risk management

9

However, is hybridisation the only possible scenario? In the next section, we turn to some sociological accounting literature which suggests alternative forms of interaction, from hybridisation to complementarity, managed tensions, and possible contradictions.

Interactions between risk management systems: hybrids, complementarities, managed tension or contradiction?

Hybridisation
Hybrids may be defined as composite phenomena produced by elements usually found separately. In biology, for example, hybrids are produced by crossing different species. In organizational terms, hybrids similarly represent a composite of two distinct modes of organizing that achieve a degree of stability and longevity (Latour, 1993; Miller et al., 2008)

Within the risk management field, the notion of hybrids suggests that interactions between diverse elements may produce a ready combination and fusion of practices, processes and knowledges. Miller and colleagues (2008) find that heterogeneous and disparate elements

between contrasting regulation regimes can constantly mix up and link. These readily combine to produce new hybrid forms, in "a continually inventive process, in which proliferation and multiplication is the norm' (Miller et al., 2008: 961). They argue that such hybrids can produce stable states that are resilient and overcome internal contradictions.

Why might such risk management hybrids proliferate? One driver is the movement of dominant risk practices through interorganizational networks that cross boundaries and penetrate individual organizations. A second driver is likely to be the movement of knowledge and expertise across boundaries, as if ideas move through some inherent force: 'novel types of expertise emerge, too, as financial expertise comes increasingly to be mixed up with other types of expertise, which earlier were viewed as distinctive and bounded if not its antithesis' (Miller et al., 2008: 952). For example, in the field of health care, financial and medical expertise may readily hybridise. 10

Comparing health systems in Finland and the UK following managerial reforms, Kurunmäki (2004) finds that in Finland, doctors readily combined medical knowledge with new financial knowledge, forming new hybrid knowledges including budget setting, cost calculations and setting prices. Amongst doctors in the UK, the hybridisation of financial and clinical knowledge has also proceeded, although more slowly.

Such hybridisation may have unexpected effects, though, as it may be impossible to control hybrid technologies that can 'take on a life of their own':

"Technologies produce unexpected problems, are utilized for their own ends by those who are supposed to merely operate them, are hampered by underfunding, professional rivalries, and the impossibility of producing the technical conditions that would make them work...Unplanned outcomes emerge from the intersection of one technology with another, or from the unexpected consequence of putting a technique to work....The will to govern needs to be understood less in terms of its success than in terms of the difficulties of operationalising it" (Miller & Rose, 2008: 35).

Applied to our empirical case, a hybridisation perspective predicts a stable and enduring fusion of ethics-orientated (clinical) and rules-based (more managerial) modes, influenced by rising rules-based clinical risk management, rather than merely co-existence or indeed possible tensions between these modes.

Complementarity
Complementarity suggests that certain entities (in this case, risk management systems) may come together in ways that mutually complete and add value to each other. Interactions between contrasting forms of regulation may be complementary as they provide a broader spectrum of possible responses.

11

Roberts (1991; 2001) and colleagues (Roberts, Sanderson, Barker & Hendry, 2006) explore this possibility in corporate governance systems where 'individualising forms of accountability' create authoritative fields of visibility by offering a remote view of corporate conduct. Through technologies such as annual financial reports and briefings for analysts and institutional investors, there can be fuller corporate disclosure. However, such visibility can also produce anxious and defensive effects, "creating a narcissistic preoccupation with how the self and its activities will be seen and judged" (Roberts, 2001:1553). 'Socialising processes of

accountability', by contrast, emphasise sense-making, open communication and dialogue through face-to-face meetings such as in canteen chat or, more formally, in a board of directors. The balance between individualistic and socialising modes of accountability may vary, but a preferred form - if it can be created - is the 'complementary mode', with a slight dominance of the socialising mode, supported by extensive external disclosure. Roberts (2001) argues that in order to support coherence of internal processes and their effective leadership, the socialising mode would remain dominant within face-to-face meetings, while external disclosure acts as a 'failsafe' device if this socialising mode is compromised, by bringing in external discipline to address poor performance. Nonetheless, he concedes that such 'creative' complementarity

entails a particularly difficult balance between individualising and socialising modes.

Applied to our empirical study, this complementary perspective suggests that an ethics-orientated mode would remain the default mode within the DTC, supported and enhanced by the rulesbased mode as required, potentially remedying difficulties arising within community relationships (see Roberts, 2001:1566).

Managed tension
By contrast, Gendron (2002) and Rahaman et al (2010) find that disparate governance regimes produce tensions; yet they argue that these can be managed and may be functional. Gendron (2002) examined client acceptance decisions by auditors in large Canadian Professional Services Firms, where taking on lucrative but risky clients increases the probability of litigation. Competing commercial and professional logics of action are here radically distinct, producing 12

levels of tension that can produce significant and overt conflict. However, Gendron (2002) found formal organizations can reduce contradictions and manage tensions through influencing local actions. By deliberately conveying mixed signals through corporate systems, firms could support the dominant logic, while the subordinate logic remained legitimate. Managing tensions in this way was regarded as advantageous, through subjecting the dominant logic to review andchallenge.

Rahaman et al's (2010) slightly different perspective seeks to hold 'competing regimes of practice' in balance so that neither one dominates the other. In exploring interactions between accounting practices and delivering HIV/AIDS programmes in Ghanaian Non-Governmental Organizations (NGOs), Rahaman et al (2010) found requirements to produce annual work plans and financial records had disruptive effects, as health care professionals were diverted into tasks for which they had not been trained. In managing this tension they responded strategically to create visible 'easy wins' for external audiences by "changing their health and prevention activities to both conform to funding rules and to maximise the probability of receiving future funding" (Rahaman et al., 2010: 118). Here we see a mixture of face compliance and gaming; there is enduring tension between competing regimes, yet no open confrontation.

In applying managed tension perspectives to our case study, we would expect to find a subordinated ethics-orientated mode being managed in tension with a dominating rules-based mode. Formal organizations would manage local tensions, resulting in the ethics-orientated mode being maintained to review and challenge rules-based regulation (Gendron 2002), or being strategically managed, even if partially disrupted (Rahaman et al 2010).

Overt contest, incompatibility and contradiction
A final possibility is of incompatibility and contest between different clinical risk regulation regimes. Here we draw on Armstrong's (1994) review of Foucault's influence on accounting research, in which he argues that empirical examples of 'incompatibilities between regimes of truth' point to internal contradictions and contest in disciplinary systems, which appear under13

examined in much Foucauldian scholarship. Armstrong (1994) questions what he sees as an 'over-socialised' notion of subjectivity in Foucault's work on governmentality, in which dominant truths are readily internalised, yet which also recognises resistance and contest in the form of local insurrections (such as in prisons) and subjugated knowledges (such as deviant subcultures in prisons):

"If resistance and subjugated knowledges are fashioned from materials extrinsic to the prevailing regimes of truth, the manner and co-existence of these different 'truths' need to be explored" (Armstrong, 1994: 33).

Interestingly, it is to such questions of resistance, conflict and 'philosophical militancy' that Foucault (2011) eventually turns in his final course of lectures at the College de France, entitled The Courage of Truth, suggesting that overt contest between truth regimes may indeed be a possibility. As applied to our case study, this perspective suggests contradiction and contest between 'incommensurable' (Armstrong 1994) ethics-orientated and rules-based modes of regulation.

In summary, our review of the literature explores the interrelationship between two contrasting clinical risk management regimes relevant to a mental health care setting. Theoretically, we have drawn on a broadly Foucauldian perspective, and developed a heuristic to analyse clinical risk management systems in our empirical case. In then reviewing some sociological accounting literature, we have explored varying interactions between different accountability systems, including hybridisation, complementarity and managed tension - as well as contest, incompatibility and contradiction.

A Democratic Therapeutic Community - a high commitment mental health setting
14

We next consider patterns of interaction between a rising rules-based mode and an embedded ethics-orientated mode of clinical risk management found in an unusual mental health inpatient (residential) setting: a DTC. As Eisenhardt (1989) argues, such atypical 'extreme cases' may be useful for studying dynamics that are more difficult to observe in other settings. The objects of the rising risk management system in this case include not only individual residents regarded as presenting a significant risk of harm to themselves and others, but also clinical staff and managers, formally responsible for managing clinical risk.

Responding to high profile incidents involving people with severe personality disorder in the late 1990s, the UK Department of Health developed new services to treat this problematic and potentially dangerous group. Severe personality disorders are regarded as challenging to manage and treat. They are deemed to act impulsively without regard to consequences; they are

associated with self-harm, suicide and homicide, and they have poor clinical prognosis. Accordingly, they have been described as 'the most difficult people to be encountered in clinical practice... (their) emotional impact on staff?ranges from anxiety to sudden unexpected anger and exhaustion... Their potential to act in dangerous ways and disrupt hospital settings makes them unattractive' (Moran, 1999: 21).

Attempting to manage this severe personality disorder group through control-orientated responses tends to exacerbate these problems because they attenuate service users' autonomy and clinical relationships tend to be become control-dominated and dysfunctional, leading to further deterioration (Adshead & Jacob, 2009; Bateman & Tyrer, 2004). Conversely, the Reed Report (1994:16)2 argued the distinctive clinical approach of DTCs "have shown the most promising results of any form of treatment...in terms of psychological and behavioural changes during

2

The Reed Committee was established by the UK Government to review services for mentally disordered offenders

and examine in detail the needs of offenders with psychopathic disorder.

15

treatment, reduction of violent incidents in treatment settings, significant improvements following treatment and, sometimes, in the maintenance of these changes following treatment"

The Reed Report (Department of Health & Home Office, 1992; Reed, 1994) recommended the development of DTCs rather than prisons or secure hospitals in managing these conditions. Taking as a model a well-established DTC with a coherent clinical model dating from the 1940s, the UK Government established two new DTCs (including the one we studied) in the early 2000s to provide a national service. We studied one of these new DTCs which had about 30 male and female residents.

The national DTC service was managed within the National Health Service (NHS) as an experimental policy innovation. Unlike most NHS services which are locally commissioned, and governed through local NHS Trusts, DTCs were commissioned by a specialist national commissioning group, linked to the Department of Health. The DTCs had formal accountability relationships with national NHS commissioners, responsible for monitoring performance against contracts. However, each DTC was hosted by a local NHS Trust which was operationally responsible for the units, including their clinical governance - a form of corporate governance, focused on the delivery of care. (Department of Health, 1999a)

The DTC therapeutic model is distinct from other mental health services (Campling & Haigh, 1999), its principles have been well documented (Lees, Manning, Menzies & Morant, 2004; Rapoport, 1960), and they were strongly replicated in this new site (Ormrod, Ferlie, Warren &Norton, 2007). Especially salient is the model's philosophy of 'community as doctor' (Rapoport, 1960) emphasising collective treatment through a resocialisation programme, addressing severe emotional and behavioural problems. DTCs are residential settings, often geographically

separated from other health care services. Residents live in the unit for an extended period of twelve months. There is a blurring of roles between staff and residents, and the term 'residents' is used to differentiate their active participation from the more passive role of patients in other mental health services. 16

The DTC has a strong group-based culture and therapeutic model that encourages interpersonal engagement. All therapy takes place in group settings; residents are expected to take responsibility for their own and others' treatment as 'co-therapists', and to participate in community decision-making and community tasks. To increase openness to emotions and

sensitivity to relational dynamics, residents are required to withdraw from psychotropic medication before joining the community, and to abstain from alcohol and drugs.

Residents join voluntarily after being assessed and interviewed by the community, testing their commitment to intensive therapy. The community votes democratically on whom to admit and discharge, or whether to 'evict' residents deemed untreatable or not authentically engaged. Such evictions might take place without prior post-discharge planning. Although a principle of

permissiveness tolerates interpersonal disturbances and 'acting out', this is constrained by counter principles of communalism, reality-confrontation and democratisation (Rapoport, 1960). These principles promote strong, personal engagement in community life by openly providing and receiving feedback, holding each other to account, and being confronted with the consequences of damaging community relations. All members are expected to clarify each other's positions, understand others' concerns, and explore solutions mutually.

In these ways, formal external accountability is balanced by strong indigenous orientation to community decision-making, while an elected hierarchy in the resident group provides senior residents with some local authority. Senior residents may be elected to the 'Top Three' resident positions, responsible for chairing daily community meetings, selection meetings for new residents, and deciding with staff how to respond to problems that arise. Through democratic decision-making, residents can over-rule staff decisions. Clinical risk is actively handled

through face-to-face meetings that attempt to cultivate authentic exchanges. Meetings of the entire community are called frequently to manage disturbances and address wider repercussions. The declared 'culture of enquiry' (Lees et al, 2004) involves reflecting and questioning to enable the community to develop collectively, as part of the treatment model.

17

The staff team is comprised of a senior group of three consultant psychiatrists (medical doctors) including a medical director, clinical psychologists, psychotherapists, specialist nurses and social workers, as well as 'social therapists' (usually psychology graduates) a business manager and administrative staff. Social therapists and administrative staff are asked to bring non-clinical, lay perspectives and social norms that are intended to be less professionalised, and thus closer to residents' own experiences. All staff are expected to participate in the day-to-day running of the community, along with residents, involving communal tasks such as preparing food, cleaning and gardening, as well as community decision-making. However, staff also have particular clinical responsibilities: some senior therapists conduct psychotherapy groups, art therapy and psychodrama, while the wider staff group conducts gardening and social activities, and facilitates discharge planning and aftercare meetings.

There is a flattened (although not flat) hierarchy within the staff team and staff are expected to speak authentically with each other, challenging each other's judgements and assumptions. This is intended to surface marginal perspectives (including those of visitors and students), balancing formal expertise with intuitive and felt experiences. The full staff team meets daily, and staff attend community meetings with residents several times each day, allowing for direct and continuous review of clinical and community affairs.

Although the clinical director (the most senior medical doctor) is formally responsible for the DTC, and reports to the NHS Trust chief executive (CEO), a clinical management team (a senior nurse, psychologist and consultant psychiatrists) limits the clinical director's formal authority through upholding the DTC's decision-making principles. All staff attend a weekly 'reflective staff group' intended to explore and resolve internal tensions. This is facilitated by an external consultant psychotherapist who comments on, interprets and challenges group dynamics. To increase adherence to the original DTC model, this DTC is part of a 'community of communities' through which mixed teams of residents and staff visit each other's communities several times a year to question, challenge and learn from each other's practices.

18

We suggest that the DTC's distinctive therapeutic model reflects a Foucauldian 'ethicsorientated' mode of clinical risk management. Through mutual engagement in a therapeutic milieu, residents seek to liberate themselves from 'disordered' patterns of relating to themselves and others. In particular, it may be a final chance to transform long-established dysfunctional patterns of engaging with authority, through shared learning, reflection, and developing authentic interpersonal relations (see Coid, 2003).

Yet the DTC's democratic approach may also present risks such as manipulation, negative group dynamics such as scapegoating and emotional 'contagion', leading to collective disturbances (Baron, 1987; Barsade, 2002). The model thus has important implications for clinical risk management. There is firstly, likely to be higher tolerance of clinical risk as crises are treated as opportunities for remedial exchanges. Secondly, clinical risk is managed interpersonally through community meetings which reinforce mutual engagement, but may increase 'relational turbulence' within and between groups (Fischer, 2012). Thirdly, decision-making is shared, democratically determined and supported rather than imposed through hierarchical power relations, but this might limit influences of professional judgement. Finally, small group settings such as this may create challenging group dynamics internally, presenting additional risks that may be difficult to address in a communal spirit.

Rising formal clinical risk management in NHS mental health services
Although the national DTC service was developed in the early 2000s, its democratic-therapeutic model sat awkwardly with concurrent, UK mental health policy changes. High-profile homicides committed by people with severe mental disorder led to legislation to detain the 'dangerous' mentally ill, based on presumed risk to the public (Department of Health, 1998; Maden, 2007).

Forensic psychiatry was influential in these shifts in policy and practice, as severe personality disorder was brought into mental health law and clinical risk management systems, especially through a new medico-legal category of 'Dangerous and Severe Personality Disorder' (DSPD). 19

While the Royal College of Psychiatrists and civil liberty groups argued against individuals seen as 'clinically untreatable' being detained in hospitals (Feeney, 2003; Royal College of Psychiatrists, 2008), the UK Government introduced policies to detain patients in secure settings, based on the notion of "serious risk such people present to the public" (Department of Health & Home Office, 1999:6). New high secure services were developed in the field of DSPD,

reflecting public and political concerns about dangerousness (Heyman, Shaw, Davies, Godin & Reynolds, 2004; Manning, 2003), while formal clinical risk management systems developed structured actuarial scales to calculate 'psychopathic risk', probabilities of violence and reoffending (Dolan & Doyle, 2000; Langan, 2010). Forensically-orientated risk assessment and risk management thus became a central focus in mental health policy and clinical practice (Brown, 2006; Manning, 2003).

In this context, a statutory Care Programme Approach (Department of Health, 1999b) was developed as an administrative framework for differentiating and managing mental health patients according to levels of risk, supervising those deemed high risk (Godin, 2004). This clinical risk management framework involves a designated care coordinator to assess and record: "the nature of any risk posed and the arrangements for the management of this risk to the service user and to others, carers and the wider public, including the circumstances in which defined contingency action should be taken" (Department of Health, 1999b:53). Its operation

places accountability on named professionals; indeed, homicide inquiries have tended to scapegoat such named practitioners (Ryan, 2004).

How might the DTCs be influenced by the rising rules-based clinical risk management system? The three DTCs were centrally commissioned by the Department of Health, and they might be expected to have come under increasing central surveillance. As a high profile and unusual policy initiative, their clinical and performance activity was externally evaluated, as was the new units' effectiveness in replicating the original DTC model. The DTC we studied had frequent contact with the national commissioners, as well as its local NHS Trust. While formal risk management methods were increasing externally, the DTC's emphasis on building strong relationships with residents was initially regarded as an effective model of clinical risk 20

management. Despite the DTCs' well-established methods, they adapted and interpreted outside influences such as the Care Programme Approach, using these to strengthen in-house methods. In our study, the DTC was initially described by its NHS Trust's CEO as the Trust's 'jewel in the crown', held to exemplify the Trust's stated principles of service user engagement. In the longer term, however, the DTC's distinctive system of clinical risk management became increasingly isolated from rising formal risk management systems externally.

A longitudinal and ethnographic organizational case study - methods and data
Our empirical data are drawn from a four-year ethnographic study of interorganizational relations between one new DTC and its external referral agencies in health, social care and criminal justice organizations, spanning public, private and voluntary sectors across three conurbations and a rural area. 3 Given well-documented difficulties in coordinating

interorganizational approaches for this challenging population (Coid, 2003), early insight into the DTC's atypical methods (see Hammersley & Atkinson, 1995) suggested an initial broad theme of interorganizational collaboration.

One author (MF) worked in the region as a clinical consultant and had pre-existing professional links with the NHS Trust, permitting access to the DTC as an 'insider'. However, his role in the DTC was solely that of PhD researcher; he had no other professional involvement with the DTC. The other author (EF) acted as his second PhD supervisor with particular responsibility for organizational and policy themes. These origins influenced study design. There were fewer difficulties in winning insider status than in other ethnographies (e.g. Schouten & Alexander, 1995). However, these links were with senior professional staff and not with service users. In an

3

The study was part of self-funded PhD research (Fischer, 2008)

21

early meeting to negotiate research access, residents requested the study should be extended from its original interorganizational focus, to include participant observation within the DTC and interviews with residents, to understand their 'real' experience. The research design evolved, therefore, in consultation with participants, moving towards the broad experience of the DTC as a setting, including its interorganizational relations. Gaining research ethics approval across multiple agencies was protracted (16 months), but this enabled wider interagency relationships to be developed. Consistent with this emergent design, our ethnographic methodology explored participants' activities, beliefs, meanings, values and motivations, seeking to interpret organizational and social worlds as members did themselves (Hammersley & Atkinson, 1995).

Over a period of four years, ethnographic observation, 76 formal interviews and numerous informal interviews were conducted across DTC-related external agencies as well as within the DTC. Formal in-depth interviews (1½ to 2 hours duration) were conducted with practitioners and managers across public, voluntary and private sector organizations, as well as with government representatives, NHS commissioners, service users, relatives and user groups. The researcher observed meetings, interorganizational (outreach) projects, and followed individual 'cases' as residents were referred to and discharged from the DTC. Following Spradley (1979), a looselystructured interview guide 'funnelled' interviews from initially descriptive questions to a more specified focus. As is characteristic of ethnographic interviewing, interviews had an informal, conversational style, giving priority to exploring participants' perspectives rather than seeking answers to specific questions (Charmaz, 2006). A second phase focused on the DTC following a critical incident that took place during fieldwork, exploring its effects on external and internal relations. Numerous informal interviews were conducted over the total period of this research (seven years in total), which informed our understanding of the field, but to comply with research ethics approval, these were not included in the formal dataset.

22

Table 1 Summary of Fieldwork Phase I Participant-observation Hours DTC Interorganizational 'outreach' projects Former residents External agencies Phase II DTC NHS officials TOTAL - Phases I & II 195 35 95 10 3 8 76 42 51 7 Days/meetings observed 9 15 1 15 14 4 32 Formal Interviews

Observations and interviews were triangulated against texts collected throughout the study. These texts consisted of emails and correspondence, minutes of meetings, policy documents and proposals, copies of confidential inquiries, newspaper clippings, organizational performance figures and committee reports. This material highlighted overlooked areas of investigation and shed light on conflicting accounts.

Aiming to be immersed in the field, yet retain freedom of movement and thought, the researcher adopted an observation-orientated approach. 'Observing in order to write' supported close attention to dialogue during meetings (Emerson, 2001), yet permitted wider participation such as informal exchanges during coffee breaks or over lunch, participating in social activities and sharing car journeys. Exchanges at the periphery of meetings deepened field relations, allowing insight into backstage behaviours and unofficial perspectives in which personal material and 23

mutual interests were shared (as described by Schouten & Alexander, 1995). These informal exchanges led to important personal involvement (Gans, 1999), permitting ethnographic insight through 'subjecting oneself' to the same situation and experiences of other participants (Van Maanen, 2011). Detailed observational notes and personal reflections were recorded later each day after on-site work.

While this design opened access to backstage regions, it presented additional demands on managing the fieldwork role. Although participants generally accepted the observer role in meetings, there were often 'inclusive overtures'4 (Emerson & Pollner, 2001) outside them. This entailed navigating between invitations for further involvement, whilst declining participants' requests to join local advisory groups, undertake research for members' organizations and contribute to other local purposes. The researcher treated such proximity not as providinginsight per se, but as significant material requiring further, reflexive questioning of his reactions to the setting (Hinshelwood & Skogstad, 2000). Negotiating levels of involvement in members' worlds provided useful access and supported a 'snowballing' sampling technique, sometimes giving access to backstage areas. For instance, residents invited the researcher to stay overnight (democratically outvoting the objections of DTC leaders), to allow participant-observation of informal aspects of residents' lives in the community, hosted by the Top Three residents rather than duty staff, (this is an interesting example of role-blurring).

During informal activities the researcher role was more participative: preparing food, shopping and doing household chores, developing friendships with residents and staff in the DTC, chatting outside with smokers, and securing the building against intruders at night time. Such informal involvement revealed behaviours and attitudes usually concealed to staff. For instance, whilst helping a junior staff member fix a bicycle, the staff member privately revealed his ambivalence

4

Emerson and Pollner (2001) describe 'inclusive overtures' as part of the dynamics of inclusion and distance in

fieldwork relations, as indigenous members draw fieldworkers towards certain activities and forms of participation .

24

about the DTC model. Residents' curiosity about the research prompted personal disclosures about their experiences. Gaining access to the backstage areas of community life surfaced insider perspectives on staff-resident tensions. The researcher recorded such observations later, waiting for a natural break before finding somewhere quiet to write.

Balancing involvement and critical distance is a common dilemma in ethnography (Schouten & Alexander, 1995). Critical distance was sought, firstly, by exploring diverse perspectives from external agencies, service user (patients) and pressure groups as well as DTC members, NHS Trust managers and government officials. Secondly, in-depth interviews cultivated researcher sensitivity to alternative interpretive schemes, particularly through studying fieldwork incidents and accounts from different perspectives (Emerson 2001). Finally, an orientation to critical reflexivity (Alvesson & Sköldberg, 2000) was developed through fieldwork diaries, writing reflexive memos, and regular discussion with PhD supervisors of emerging empirical and analytic themes, including the researcher's personal responses to the material (see Arnaud, 2012).

Fieldnotes and interviews were transcribed and NVivo software was used to assist data management and analysis. A modified grounded theory method was used at this stage in analysing ethnographic fieldnotes, using key incidents and memo-writing to move from open to focused coding (see Charmaz & Mitchell, 2001; Emerson, Fretz & Shaw, 1995). This analysis was originally developed in a PhD thesis (Fischer, 2008) within a broadly group-psychoanalytic framing. A clear empirical focus within the PhD was the organizational crisis and sudden closure, precipitated by a critical incident (a homicide).

Theory was developed from this analytic focus through comparing deviant cases (Katz, 2001), and revising analytic concepts and memos, using a loose initial framing. In Gendron and Bédard's (2006) social constructionist account of the constitution of audit committees within Canadian firms, they firstly used a loose framing, later focusing their analysis by bringing in theoretical literature on actor reflectivity. Similarly, we drew upon theoretical literature to frame 25

our analysis of contrasting modes of clinical risk management. We firstly explored Foucauldian scholarship on governmentality, focused on mundane techniques of audit and risk management (McKinlay & Starkey, 1998; Miller & Rose, 2008; Power, 1997; 2007), then later broadened our focus to include Foucault's late scholarship on 'technologies of the self'. Through our reading of this literature, we developed a guiding theoretical framework as a heuristic to act as a mid-level sense-making device (see Figure 1). This allowed us to move iteratively between data and theory to produce our empirically-grounded framework (see Eisenhardt, 1989).

The DTC's ethics-orientated mode of clinical risk management
In its early ethics-orientated phase, the DTC presents as a high commitment setting in which residents and staff are interpersonally and emotionally engaged. The DTC's clinical model emphasises sincere relational engagement and shared commitment to democratically-held rules as its essential method of managing clinical risk. Despite its 'high risk' population, the DTC is governed through residents' and staff members' commitment to a complex set of rules and structured programme of groups, developed over the 60-year history of the original DTC. These rules are strongly maintained by the community and they are seen by residents and staff as integral in building trust and collaboration within the DTC, and in sustaining the community as a 'safe place'. Such personal commitment is regarded as central to handling the relational

problems of people with severe personality disorder, such as strong negative reactions to hierarchical authority and a tendency to readily disengage from clinical treatment (Lees et al., 2004; Manning, 1989). Its democratic-therapeutic approach contrasts with how formal systems seek to manage severe personality disorder through increasing physical security, actuarial risk calculations by professionals, and close integration with the criminal justice system (Exworthy &Gunn, 2003; Seddon, 2008; Tyrer, 2007).

For residents accustomed to other health care settings, the DTC's participative methods can offer an idealised 'place of hope' where reciprocal involvement is central: "After so much rejection, I had a complex about nobody wanting me. But when I got there, it seemed like a place of so much hope that I just burst into tears. I have always 26

had problems with professionals, but the groups really appealed to me. It was really empowering. (One resident) asked me about my attempted suicide...and then he walked out really upset. And it just really affected me that - oh my God - I can have so much effect on people" Former resident Residents are expected not simply to comply with DTC rules, but to learn how to understand and interact with them adaptively. Greater behavioural disturbance is tolerated than in other mental health settings, as clinical crises and rule-breaking are regarded as opportunities for reflective sense-making and learning. All members are encouraged to grasp others' concerns and explore alternative perspectives and solutions. "It was a central task for all those coming to the community that they would learn to recognise and understand their own feelings and subsequent risk - and through relating to others within a structure specifically designed to slow incidents and issues down so that they can be understood and thought about before acting." Senior therapist 1 Clinical risk is managed by community members drawing upon, interpreting and sometimes amending DTC rules, generally supported through the resident hierarchy. Much relational engagement takes place in a mutually constructed liminal area that is neither formally part of the organization, nor entirely private (Vaivio, 2006; Warner & Gabe, 2004). Through the DTC's model, members are encouraged to explore these interactions within the community as a therapeutic milieu. For example, in an informal discussion with a new resident, Simon5, his upset response to feeling misunderstood by other members prompted the following exchange, in which fellow residents linked Simon's emotional reaction to the intended use of DTC rules. Resident A: "After the way you felt last night, you're looking for rejection and finding it when it isn't there."

5

All personal names have been disguised

27

Resident B: "You know, you haven't let us know what is going on for you. We aren't mind readers...you didn't approach Top Three and so you created distance for yourself and for others... Resident A: "You just need time to hear that what has been said to you throughout your life, well that really isn't what is being said to you in here" Informal discussion between residents Thus, maintaining shared emotional investment in community rules is held to be central to maintaining the community as a "safe and therapeutic space" (senior therapist B). Anxiety or concerns held by any resident or staff member is expected to be taken to one of the elected Top Three residents who would then liaise with duty staff. They would together judge whether the issue could be contained until the next scheduled meeting, or if an emergency meeting of the whole community should be called at any time of the day or night, to provide support and safety. While this might involve minor issues such as an untidy kitchen (which often provoked strong feelings), meetings would also handle incidents such as self-harm, address disputes, and arrange support for distressed residents.

In the case of Miranda, a resident who had cut herself through self-harm, 'Top Three' alerted duty staff to provide first aid and assess whether further medical attention was needed. An emergency community meeting was called to inform the community and agree which residents should accompany Miranda to hospital, together with a staff member. When she returned, another meeting was called to organise peer support during the night. The entire community is obliged to attend such meetings in order to understand events from multiple perspectives. Solutions to handling the incidents and members reactions (such as anxiety) were then extensively debated until a majority decision was reached. Such decisions are intended to be actively supported by the whole community. "Behaviour gets challenged - even any slight worry they've got about you acting out or anything. At first I was, like, I don't really get what you're worried about. I never had

28

the insight before to realise where all my problems came from. But it really makes you aware of yourself; it was like I had a life again." Resident C Residents are often "sensitive to changes in atmosphere" (senior therapist), and may call meetings about themselves or others perceived to be experiencing distress, or who are seen at risk of harm. Meetings may be used to organise support, averting possible escalation. For instance, a practice of 'floors and doors' involves a rota of volunteers who would sleep nearby a resident feeling vulnerable, or remain awake through the night to provide active support. Community meetings are chaired and conducted by Top Three residents and are ritualistic and formalised in tone. There are frequent votes for decision-making, taking the form of five minutes' discussion, a call for objections, two 'tellers' stand to count votes for, against, and abstentions (voting takes place by show of hands). These meetings are minuted in detail by an elected resident acting as secretary, who also reads out the minutes in the following morning's community meeting. "The meeting is very formal (ritualised), starting with a name-round 'for the visitor'. There was a reading of the previous day's very detailed minutes, and notes of discussion, which seemed verbatim...This was listened to in silence, with an almost religious respect...it felt the reading was being received like a sacred text in a monastery." Fieldnotes The staff team fosters the therapeutic functioning of the community. Many staff members are visibly invested in the 'community as doctor' model; their authority is gained not through formal hierarchy, but their perceived emotional commitment to the community, along with their ability to demonstrate authentic interpersonal exchange with residents, as well as each other. Upholding DTC methods, though, can be filled with ambiguity, prompting the exercise of what some therapists describe as 'concertina-like' authority, whereby staff seek to promote clinical perspectives, but without undermining the DTC's democratic methods. "If the culture of enquiry is not carried by residents, it becomes something that the staff are left to do. And when questions come from staff rather than residents, we are accused 29

of being too psychotherapeutic (and) making residents feel vulnerable and abused. The longer (this) goes on, the less communication takes place, and momentum builds for things to take place behind the scenes." Senior therapist 2 However, this clinical model may amplify personal and organizational reactivity. The DTC's emphasis on therapeutic (rather than managerial) responses means that members' emotional and behavioural reactions can have a cumulative, disruptive, effect. Reactions associated with

interpersonal conflicts, suicide threats, self-harm and intimidation can provoke 'ripple effects' of emotional contagion (Barsade, 2002), undermining therapy. "You live and breathe the (personality disorder) experience here...every pore is fully immersed... The dynamics...of this patient population...seep everywhere. You do enter a similar (personality disorder) experience to that of the residents...the staff room (mirrors) what's happening with residents and vice versa... All these things are polluted by the dynamics." Therapist 3 At times of frustration and exhaustion, democratic decision-making can increase community tensions rather than containing them therapeutically. Despite a tendency to idealise the DTC model, an important shadow side is of 'groupthink' (Janis, 1982), along with scapegoating of presumed troublemakers. "The first two months I didn't know where the hell I was... I was totally overwhelmed and exhausted and I was fighting with them...I took a lot of the community's anger because of my behaviour. A lot of people were intimidated by it and I ended up getting scapegoated. I remember sitting there crying every day, knowing I just needed to stop hurting myself, I needed to be different." Resident D There can be strong group pressure to evict members seen as untreatable, without considering adverse clinical consequences or post-discharge planning. According to the DTC's 'treatability' rule, members who break community rules, such as the prohibition on drugs and alcohol are deemed to have automatically discharged themselves, prompting an emergency community meeting. The community can elect to 'readmit' residents for a temporary 24 hour period in 30

which those on 'treatability' should show motivation to authentically re-engage with the community and its rules. But readmission also depends upon the wider community's appetite for reconciliation. "People had simply had enough. The community kicked him out in the middle of the night and then later realised that they may have been unnecessarily angry... Sometimes we end up making crap decisions." Therapist 4 Despite such disruptive potential, the model provides a self-regulating mechanism as participants tend to develop a personal interest in others' emotional reactions. Experienced members, for instance, advocate managing community reactions, tactically "slowing things down" (resident) rather than provoking crises. "We really don't want to increase anxiety; people will just go 'pop'. The community has learned to contain stuff, otherwise we end up having (emergency community meetings) all night long." Therapist 5 The DTC's ethics-orientated model attempts not to minimise clinical risk, but maintain it at subcritical levels, making it amenable for remedial work. Incidents and crises are seen as providing a means of reflecting upon and challenging DTC decision-making and learning from events. Accordingly, notions of 'connecting', 'relating' and allowing others to 'understand what's really going on' are emphasised above formal control.

In managing clinical risk, then, the DTC's approach hinges upon strongly participative and democratic engagement. Although accountability relationships and reporting were active between the DTC, the host NHS Trust and the national commissioners, the DTC was commissioned to treat a challenging clinical population according to well-established DTC methods; indeed it was required to demonstrate the methods' clinical and economic effectiveness through government-commissioned evaluations. This early phase of the DTC was generally regarded by NHS Trust managers as successful: there were no significant incidents, independent evaluations had been positive, and the NHS Trust sought to extend the model in some local

31

services. However, its unusual commissioning arrangement led to some tensions, particularly when new commissioners who were unfamiliar with the DTC's approach came into post. "We go to get beaten up by commissioners, but then we go home again - there's nothing much we can do about (how we work). This is what we were commissioned to do". Clinical management team Nonetheless, given strong local support by the NHS Trust and a promising evaluation (Fiander, Burns, Langham & Normand, 2004) the DTC's indigenous methods of clinical risk management remained dominant. In summary, although the ethics-orientated mode tolerates disruption, it seeks a restorative approachof self-regulation in which clinical risk can be contained and handled therapeutically by the community, notably through community meetings.

Critical incident: The DTC model of clinical risk management is questioned
The continued operation of this model was threatened by a critical incident which took place during fieldwork, involving two former residents, Mark and John, who had recently completed a twelve-month course of therapy. The two men formed an intimate relationship whilst in the DTC and, unknown to DTC staff, moved into a shared apartment after leaving the unit. Shortly afterwards, Mark stabbed John to death during a drunken row. Although both men had been regarded as successfully 'treated', Mark was detained in a secure hospital, charged with John's homicide, just weeks after leaving the DTC.

John's death led to intense shock and grief within the community, but its psychological impact was accentuated by feelings of self-doubt, guilt, and torn loyalties to the two men. Moreover, acute anxiety about potential consequences of the incident and likely official inquiries eroded members' confidence in the DTC as a 'safe and therapeutic' liminal space. These internal reactions were compounded by the homicide's repercussions within the wider health care system. Officials from the local NHS Trust (with line managerial responsibility for the DTC), NHS commissioners (who funded the DTC) and the Department of Health (responsible for 32

mental health policy) each reacted with anxiety as a high-profile innovation threatened to become a policy embarrassment. "Of course the homicide caused a great furore. The chair of (the commissioners) began to get anxious he was going to end up with a homicide inquiry... (and) panicked into commissioning a risk report of the entire national service... Did it make the Trust anxious? Oh my God, yes...it already had two homicide inquiries going on (unrelated to the DTC), both of which will severely criticise (it)." Senior official Officially, the homicide was not the DTC's responsibility. The incident occurred after the men's discharge, and accountability for aftercare formally lay with local agencies. Nevertheless, NHS commissioners regarded residents' democratic participation in each other's discharge plans and the DTC's emphasis on community decision-making rather than staff interventions as 'clinical laxity'. Health)". "The homicide is telling: discharge planning was done by other punters rather than by clinicians. Staff keep saying that these are some of the most dangerous and manipulative patients that there are. Yet they don't take clinical responsibility for them...it's literally the case of 'the lunatics running the asylum'. It's scandalous...the project is unsafe." Senior official Some senior officials were concerned with "not embarrass(ing) the Minister (of

A rising rules-based mode of clinical risk management produces escalating conflict
Under strong external pressure to bring the DTC in line with 'normal' clinical risk management procedures, the NHS Trust's board imposed its standard risk management procedures and governance arrangements. The Trust sought to enforce strict control through weekly risk reports from the DTC, official inspections, and greater involvement of Trust managers in clinical decision-making.

33

This imposed clinical risk management system interacted with the DTC's ability to function therapeutically, disrupting its indigenous model for managing substantive clinical risks. The circumstances of this critical incident are complex, as indeed are possible internal and external influences on parties' various reactions to the event. Nevertheless, our analysis suggests that the following four-stage process transformed the DTC from its self-regulating ethics-orientated mode to a dysfunctional unit, riven by escalating conflict between contrasting modes of clinical risk management.

Imposed formal risk assessment
Firstly, what had been an essentially personal, therapeutic space was increasingly formalised through transparent reporting of calculable risks. Initially, the NHS Trust conducted an immediate internal homicide inquiry, bringing in senior executives and clinical experts from the field of forensic psychiatry, whose knowledge base and practices were centred on actuarial risk calculations by professionals, and risk management enforced through high levels of physical security. The inquiry investigated the DTC's documentation and recorded interviews with DTC residents and staff, as well as external professionals involved in the case. It produced a highly critical report of DTC's risk assessment practices, concluding that 'no-one was flying the aircraft; and no-one was looking out of the window' (Trust internal inquiry).

The national commissioners appointed a senior director to conduct a 'root and branch' comprehensive risk assessment of the DTC, along with the two other DTC units, involving independent scrutiny of all policies, procedures and practices, organizational as well as clinical, and inspections of the physical environment. The local NHS Trust directed that DTC staff conduct comparable local risk assessments and report plans for managing identified risks to the Trust board. This shift from clinical risk management based on substantive first order risks to assessing and reporting calculable second order risks was initially resisted by the DTC. "Bringing to light some of the risks that there are over has not been an easy process. It has meant clashes in management styles and expectations. And I guess the director and 34

the DTC team feel they are getting reined in and being unfairly questioned and scrutinised." Trust executive i Nevertheless, the Trust board insisted that DTC risks be formally identified, recorded and reported, in line with its other mental health services. These included regular audits of potential ligature points (physical features that could support a noose for strangulation), formally recording clinical discussions associated with discharge planning, and recording what managers constructed as 'advice on clinical risks' given by staff to residents. Trust managers challenged the DTC's strongly participative interpretation of the Care Programme Approach, arguing that formal clinical responsibility should override residents' decision-making. "You are in conflict straight away between the trust's quite directive approaches versus that of the DTC which says we have to get agreement from (residents). To some degree we do need their cooperation, but it is a question of who has the last say...(about) their willingness to identify risks and to have (an approach) for dealing with them" Trust executive ii Given perceived difficulties in conducting such formal risk assessment, the Trust's director responsible for risk management set up a weekly risk management meeting with the DTC clinical director and clinical management team to assess and manage identified risks, tying these to reports of ongoing clinical incidents. "The nature of how the service runs is risky. It's really brought into focus the risks in the service... It's been a wake-up call for the execs that jeez, we should have been on top of this earlier and made sure systems were in place. We're carrying a lot of risks that we weren't aware of." Trust executive iii

External steering of clinical risk management
Secondly, these formal risk assessments enabled Trust managers and commissioners to steer clinical practices. The DTC had traditionally invited professional visitors to the community to observe some of the democratic-therapeutic model (partially to stimulate interest amongst 35

potential referrers). After the homicide, these visits became a means for Trust managers and commissioners to directly assess the community, compare interactions against standard clinical risk management practices, and direct alternative responses.

For instance, the CEO and Chair of the NHS Trust conducted a joint visit, during which they suspected two residents were in a sexual relationship. Sexual relations were discouraged, but not prohibited according to DTC rules, as they were considered potentially useful for therapeutic learning. Fearing a repeat of the recent circumstances of the homicide, they demanded the clinical director stop this relationship (suggesting the residents be threatened with discharge if they did not comply), and formally record this as a clinical risk management intervention. "We said to the clinical director: look, you need to do something to stop it. These people should be concentrating on their therapy, not on having this relationship... A lot of work had to go on from here to say have you counselled those individuals, have you recorded that you have counselled (them) and have you advised the different agencies. community should have been firmer about what was acceptable." Trust executive iv Such external steering was experienced by DTC as interfering in the community's democratic decision-making, and some staff felt it undermined the DTC's well-established model of clinical risk management. "There is something about the unknown that unnerves them so the trust goes to what they think is solid ground... But residents should not be made to feel like subjects of risk, debris of pathology...whose behaviour needs to be tagged and monitored. Are residents allowed to show (feelings), or are they slapped down to keep their feelings inside?" Senior therapist 2 External officials' strong reactions to the community's 'inner life' accentuated internal anxiety about repercussions following the homicide, and what some feared might be a 'witch hunt'. The

36

"The pressure has been immense...enormously traumatic... We have all this anxiety around...a chain reaction...that the finger of blame needs to be pointed somewhere for (officials) to be satisfied..." Senior therapist 6

Some adoption of rules-based regulation, but growing contradictions
Thirdly, despite proclaimed opposition, there was some adoption of externally imposed risk management practices by the DTC clinical management team, now determined to assert stronger clinical authority on the community.

"The question of the role of staff is important. Do we lead or are we for only part of the community? Maybe staff will need to take a stronger position, like the Director has said. Is it that the residents rule the community instead of staff? So it is on their terms? Because then we lose our respect and authority." Therapist 4 Senior staff adopted a subtly stronger position, overruling community opposition. community was becoming "more authoritarian, but it's not explicit... We act less as one organization; a sense of being in it together" (Therapist 7). The

Anxious about their professional careers and livelihoods, some staff weakened adherence to DTC principles, covertly conveying confidential clinical information to external agencies. Despite overtly resisting the CEO's demands to stop sexual relationships between residents, a senior therapist later pointed out to a resident that he should inform his probation officer about his relationship with another resident: "you have to understand that if your probation officer isn't seen to be completely on the ball, her career is on the line - you need to find a way of managing that by revealing more about what's going on" (senior therapist 8). In a subsequent staff team discussion, therapists were preoccupied about what clinical information they should convey to external agencies, "covering ourselves, in case the shit hits the fan" (senior therapist 1).

37

Staff seem anxious - what is our stance in relation to probation? (A DTC leader) says, 'no I didn't tell them about (the residents' relationship)...well actually I did, (but) said to (the probation officer) I didn't say that did I?' (The DTC leader) suddenly looks

mischievous. A senior nurse tries to clarify: 'if he puts himself at risk, what will that do to us? What are the implications for us?' This time more reassuring, (the DTC leader) says... 'we have to work with (probation) - and we have to get (the resident) to work with them.' Fieldnotes

Residents were highly sensitive to this sensed shift in staff's emotional investment, particularly from senior clinical staff who were perceived to have adopted more formal risk management perspectives. "There is not one community here - there are two. I really don't trust staff. You can't call it a community when you can't talk with them about anything. They have far too much control - you can't call it democratic." Resident E This division provoked strong negative reactions from residents, articulated in a growing distance between staff and resident subcultures, which eroded residents' adherence to DTC rules.

Drug-taking came to light overnight when a resident started throwing furniture around. Police 'legged it' into the resident's bedroom and made two arrests. Next morning, a senior resident hesitantly acknowledges her involvement - she has taken amphetamines, cocaine, ecstasy, ketamine, 'dope', the list seems endless. She's had previous drug convictions and is terrified of being arrested... Eight other residents were involved, but alcohol is far more widespread - unofficial partying has been going on for weeks. Staff look visibly shocked, and insist the full picture of others' involvement is disclosed. Tom (another resident) storms out of the community meeting shouting, 'I'm addicted to 'speed', I need help, not this!' Top Three (had) colluded by protecting intoxicated residents, keeping them away from staff, and distracting staff by accusing them of bullying. Fieldnotes 38

The staff team were shaken by this incident and its repercussions triggering further involvement from the Trust. The DTC clinical management team immediately discharged residents held directly responsible and demanded the community evict a further six. They next directed all remaining residents to consent to police tests for drug and alcohol use, insisting that those unwilling to sign a consent form would be evicted: "the Trust was more disturbed by this outbreak of drug taking and the (outside) gossip around this. I insisted on random drug testing not very DTC!" (senior therapist 8).

In the following community meeting, forms were being passed between residents in silence, consenting to police taking random samples of residents, of saliva, urine, blood and hair, any time of day or night: I'm struck by the seemingly draconian and legalistic consent form... the clinical director is taking advice from the drugs liaison police. Afterwards, junior staff disagree about the new arrangements: 'we are far too reactive, we really undermine the residents. Surely, there's another way? They really have a sense that we are constantly checking up on them and it's really not helping us or them'... But DTC leaders refuse to compromise, 'we need to work with (this decision) - it's not going to be reversed'. Fieldnotes These measures further increased division and eroded community morale as staff struggled to maintain shared purpose. In place of ethics-orientated self-regulation, a procedural mindset was becoming more dominant. Despite staff insisting they remained committed to therapeutic

participation, residents resented what they experienced as a 'betrayal' by therapists whose belief in democratic practices appeared compromised. In turn, therapists noted a major change in the community's democratic-therapeutic ethos, resulting in a loss of their moral authority: "We tried to hang onto some semblance of authority. But the power the trust and commissioners were exercising over us diminished our ability to manage the attacks from residents, paradoxically...while the level of attacks was going up. I felt I no longer had the moral authority to provide guidance." Senior therapist 2 39

Politicisation of community space
Finally, the DTC's capacity to contain clinical risk was disrupted by members' reactions to the rising mode of formal risk management. "It's like a prison stand-off - like people are trying to psyche each other out. Who's going to break first? Who's going to be able to stay silent the longest? The most powerful people are the ones who say nothing. Because you can't work them out; they give nothing away." Resident F Within the staff team, there was increasing uncertainty and division about how to manage deteriorating relations and escalating clinical incidents, producing a sense of paralysis. There is an air of urgency - every member of the DTC management team is present. Community tensions have reached an intensity staff have never encountered before, it seems the community is on the verge of breakdown. A senior therapist, warns, 'if staff and residents don't work together, we'll have to close, it's just too dangerous'... A senior therapist says it's a problem of emotional exhaustion: 'It's critical that we deal with the sense of staff deficit, we're as stuck as they are - we're like a broken record." Fieldnotes of staff meeting Fuelled by intense anger and recrimination, deteriorating relations between staff and residents fostered a dissident and secretive resident subculture. "The past months have been hell. There's been a complete lack of trust. The community has been in chaos... There was no protected time, no retreat? it's like a year in Beirut." Resident C DTC members' sense of a mutually-constructed liminal space collapsed as informal leaders emerged, resisted authority, steered resident decision-making and orchestrated conflicts with staff. Instead of upholding the DTC's democratically established rules, residents protected each

40

other from staff scrutiny as they devised drug deals, organised illicit parties, and established a new norm of sexual relations between residents. "There was this mob mentality - there was absolutely no functioning aspect of the group I could appeal to. The staff picked up the pieces every morning after a bloody bomb had gone off over night... The community became disembowelled - every day was like a nuclear reactor without the container. There was nothing around to hold the explosion." Senior therapist 2 Staff-resident exchanges became embattled, undermining the DTC's capacity to manage clinical risks and 'ordinary' disturbances. Despite daily risk reports being sent for the personal attention of the Trust CEO, managers' efforts to instil order in the DTC acted perversely, exacerbating disturbances and increasing substantive clinical risk. "We had nothing to hold onto because everything kept changing. The trust's risk

management group met with us week after week and couldn't understand why we were still open; they kept saying...that the level of risk was just too high." Senior therapist 6

Collapse of the DTC
Given escalating incidents, tensions between the DTC and the Trust and commissioners became openly confrontational. As external NHS officials engaged with the heightened disturbance within the DTC, their personal reactions became part of the story. Trust managers and

commissioners were emotionally "pulled into an all-consuming (engagement)...you give your whole life to that unit." (Trust executive iii) "I didn't go in with body armour, but at one point I thought I might have to. One of the staff was crying and I thought, here we go?! My ears were already burning before I got there. And by the time I did, I think that burning effigies would probably be next on the agenda." Senior official 41

As NHS officials were drawn into charged conflicts, political pressure increased on the host Trust, producing a downward spiral of reactions. "In a situation in which the commissioners didn't understand the model and acted highly emotionally, this upped the ante even more... It makes the board anxious, it really does." Trust executive iv Unaccustomed to such antagonism, one official described his experience of meeting with DTC leaders and staff as "poisonous...the atmosphere is just so intense that people just get fried up? I have never in all my experience faced that degree of hostility. It is the only organization that (the national commissioners) agreed never to meet single-handed" (Senior official)

Although the commissioned risk report concluded all three DTCs were 'basically safe', warning that greater clinical risk would arise with rapid closure, the NHS commissioners and local Trust decided to quickly close the unit, discharging all residents within a matter of weeks. Despite positive independent evaluations (Fiander et al, 2004), officials had become emotionally drawn into confrontation and lost confidence in the unit. As one described: "(They decided)...to wrap up quite a lot at one go. I'm astounded at the failure of the Trust to support the place. We end up with the service collapsing because it did was what it was asked to do. But politically, there were some 'shenanigans' went on and the thing collapsed. There is a serious underestimation of the dynamics of these

(interagency) relationships and how they work." Senior official In summary, the homicide was a critical incident with two key effects. Firstly, the DTC's indigenous ethics-orientated method of clinical risk management was undermined by the external imposition of formal risk management, triggered by the homicide. Secondly, tensions between these two incommensurable modes of clinical risk management contributed to greater clinical disturbance and escalating conflict, both internally and in interaction with external agencies. The

42

story ends with what had been seen as a promising organizational experiment unexpectedly being closed6.

Concluding discussion: why contradictions and intractable conflict rather than hybridisation?

Our finding of overt contest between conflicting modes of clinical risk management contrasts with much literature reviewed earlier. Returning to our heuristic (see Figure 1), our case elucidates interlinked dimensions along which the two contrasting modes of clinical risk management interacted perversely.

Conflicting truth discourses: From background tension to overt contest and conflict
We suggest the homicide can be seen as a critical incident that precipitated a collision between ethics-orientated and rules-based modes of clinical risk management. Interagency tensions over accountability and blame for this incident shifted rules-based clinical risk management from a background resource to an alternative truth regime, proposed, suggested, and imposed (Foucault, 1996) by worried managers and commissioners.

As Kosmala and McKernan (2010) argue, ethics-orientated regulation is never entirely separate from rules-based forms; the balance between them is dynamic and tends to vary between different epochs. Nevertheless, rules-based regulation may remain a background social

framework. In our empirical case, the pre-incident DTC had drawn upon, re-interpreted, and

6

Both the original DTC, dating from the 1940s, and the other 'sister' DTC were later closed, as NHS policy

subsequently shifted away from DTCs towards high secure, forensic settings.

43

assimilated into its ethics-orientated clinical risk management certain external codes, notably adapting some principles of the Care Programme Approach (as described below). authorities' increased involvement with the community and its 'inner life' brought into play contradictory truth discourses about how risks should be constructed, identified and handled. Clashes arose not just over substantive clinical risks (notably, sexual relations, drugs and alcohol use), but second order risks of reputation and performance (Power, 2007), important in NHS regulation systems. But

So how did these competing 'truth regimes' interact? In contrast to much of the literature reviewed earlier, we find the development of escalating contest and contradictions with two key effects: firstly, erosion of the ethical basis of self-regulation and secondly, the developmemt of intractable conflict. Imposing rules-based ideals undermined valued aspects of therapeutic

relations and eroded staff and residents' shared commitment to an ethics-orientated mode of selfregulation. This did not, however, lead to rules-based clinical risk management becoming dominant. As parties' various efforts to steer risk management practices and decision making became mutually mistrusted, the community's carefully 'negotiated order' (Strauss, Schatzman, Bucher, Ehrlich & Sabshin, 1964) was disrupted. In its place, overt (and covert) contest between conflicting notions of clinical risk management interacted perversely, eroding moral authority and undermining members' trust in the DTC as a 'safe and therapeutic space'.

Contradictory practices: Non-hybridising, but conflicting modes
The DTC's indigenous clinical risk management practices had developed gradually through a participative method involving deep relational engagement, reflecting on shared experience, and collective decision-making. These practices were relatively well defined, tended to be internally coherent, and they were strongly embedded within the community. Furthermore, they were tied to a set of ethics-orientated principles that promoted therapeutic meaning and interpretation. In the charged and risk averse policy environment following the homicide, certain DTC practices such as democratisation (residents' influence in democratic decision-making), permissiveness (privileging experiential learning over strict adherence to rules) and communalism (care plans 44

based on relationship building, rather than administration) came to be regarded as inherently risky and in need of urgent reform. Indications of inadequate discharge planning triggered interventions from the host NHS Trust and national commissioners who imposed statutory NHS clinical risk management practices, minimising scope for local interpretation.

These introduced rules-based practices were experienced, though, not as neutral 'technologies' (Miller & Rose, 2008), but morally laden and designed to attribute blame and punishment. Demands that DTC staff should monitor, record and report resident activities - and that residents should 'confess' intimate relations, report illicit drug taking, and allow examination of bodies through substance testing - were experienced as authorities' efforts to control and dominate 'subjects of risk...debris of pathology' (senior therapist). Clinical risk management practices thus became aligned with deeply held and contested positions.

Whereas heterogeneous practices certainly mixed in the DTC's post-incident phase, instead of the complementarity (Roberts, 2001), managed tension, (Gendron, 2002; Rahaman et al., 2010) or hybridisation and proliferation of practices (Miller et al., 2008) noted by other scholars, we find clinical risk management techniques were used as devices by parties seeking to control and outmanoeuvre each other. Through intensifying their use of rules-based codes, for example, DTC leaders attempted to weaken residents' - and some staff members' - engagement in the embedded ethics-orientated regime (through which they could democratically outvote DTC leadership). Conversely, residents' anger at what they experienced as 'untrustworthy' attempts by DTC staff to take control stimulated a strongly politicised response amongst residents who became increasingly engaged in challenging, subverting and overtly refusing hierarchical authority. In relation to our Foucauldian framing, the DTC's original ethics-orientation may be seen as later developing an 'agonistic exteriority' to domination (Foucault 2011), in which residents' outspokenness, self-determination, and opposition to authority became central.

45

Altered forms of reflexivity: From intersubjective relations to externally orientated defensiveness
DTC members' reactions towards imposed formal clinical risk management after the homicide shifted patterns of reflexivity that had been sensitised to relations within the community. As emerging second order risk representations recast DTC practices from 'exemplary' forms of service user involvement to risky 'clinical laxity', members' sensitivities to external judgement and criticism increased. We suggest this shift altered intersubjective reflexivity to produce more selfconscious intra-subjective patterns, orientated towards external judgement.

These altered patterns of reflexivity following the homicide - shifting from an interpersonal to an externally facing focus - disrupted members' sensitivity to intersubjective relations within the community. Whereas the DTC's original therapeutic model promoted a sophisticated 'craft' of intersubjective reflexivity to manage first order risks, members' increasing orientation towards second order scrutiny produced reactive defensiveness. Whereas increased awareness of external perspectives might have served as an important learning resource (Kosmala & McKernan, 2010; Luhmann, 1993), we rather see increased self-consciousness and defensiveness amongst DTC residents and staff, fearing "the finger of blame" (senior therapist), disciplinary action and, in the case of DTC staff, potential damage to their professional careers.

Emotionally charged intersubjective relations: The emotional organization
Finally, an important aspect of this case was the development of intersubjective relations characterised by strong personal and group engagement, which were experienced as emotionally intense. This analysis fits with much of the approach of one school of organizational studies which emphasises the role of emotions in organizations, as a corrective to rationalist, functionalist or institutionalist accounts (Fineman, 2008; Gabriel, 1999). In our case study, high levels of emotion and personal values invested in contested modes of clinical risk management

46

produced intense and intractable 'relational turbulence' (Fischer, 2012) within and between organizations, to which commissioners, managers, clinicians and residents actively contributed.

Much sociological accounting literature does not consider emotional loading in risk management and accountability regimes, tending instead towards an assumption of 'cool' climates, characterised by pervasive rational techniques. An important exception is Roberts et al's (2006) study of meetings between executives and investment advisers, involving major 'points of anxiety'. Executives' anxieties about face-to-face meetings with advisers, 'seeing the whites of their eyes', led them to intensive preparations. These meetings produced strong self-disciplining effects, accompanied by a strong desire to avoid conflict as executives conscientiously rehearsed presentations to investment advisers.

In our case, by contrast, external authorities were drawn into emotionally charged face-to-face meetings with DTC members, whilst this closer involvement with the community only increased officials' anxiety and reactivity to its "intense atmosphere" (senior official). This high

emotional loading produced 'heated' intersubjective exchanges and mobilised underlying tensions, leading to escalating and intractable conflict. Instead of rehearsed performances

(Roberts et al 2006), we see competing beliefs and values, polemical resistance and, in the case of some residents, attempts to undermine formal clinical risk management.

Incommensurable clinical risk management regimes: moving from contradictions and contest to intractable conflict
Returning to our Foucauldian heuristic, our case study reveals how contrasting ethics-orientated and rules-based modes of clinical risk management interact across the four dimensions described above. Empirically, we find neither hybridisation (2008), complementarity (Roberts, 2001) nor managed tension (Grendron, 2002; Rahaman et al, 2010), but rather heightened contradictions and contest between modes, in which differences are exaggerated and reinforced, leading to intractable conflict. 47

These tensions may partly be explained as incommensurability between contrasting 'truth regimes' or paradigms which interact but 'talk past each other' as they lack commonality. In Kuhn's (1996) study of scientific revolutions, he argues that there are occasional but radical shifts between incommensurable scientific paradigms (such as the transition from Newtonian to Einsteinian physics). His notion of paradigm includes specific theories, rules and methods, as well as internally coherent constructions of problems and possible solutions. Burrell and Morgan (1979) used the concept of paradigm incommensurability in the field of organizational studies to examine different schools of organizational analysis, reflecting differing epistemologies, theories and values. Indeed radical paradigmatic disputes are evident over substantive and theoretical questions about organizational change and leadership (Learmonth, 2006) and strategic management, where schools of thought have proliferated (Scherer, 1998). According to these perspectives, incommensurability involves radically different orientation systems that compete over issues of shared importance, where an acceptable means of adjudicating these tensions and conflicting values is lacking (Scherer, 1998; Tadajewski, 2009).

We question, though, whether this paradigmatic argument is sufficiently sensitive to micro-level interactions that may be important empirically. Whereas the notion of paradigm emphasises differences operating at the highest cognitive level (Kuhn, 1996), this perspective may neglect micro-level processes, beliefs and practices that proliferate in the 'undecided space' between rules-based risk management systems and the situated practices and orientations of active subjects (Iedema, Jorm, Braithwaite, Travaglia & Lum, 2006; Iedema & Rhodes, 2010; McGivern & Fischer, 2010). For instance, we found some evidence of early hybridisation as the DTC appropriated certain external rules-based 'resources' (Kosmala & McKernan, 2010; McGivern & Fischer, 2010), modifying certain elements of the Care Programme Approach from an administrative technique to a relationship-based practice. DTC staff insisted that statutory meetings with other agencies should include residents as a means of influencing statutory decision-making. Such re-modelling was initially regarded by the local NHS Trust and outside agencies as exemplary (if sometimes challenging) user-orientated practice. 48

Whereas the pre-incident DTC adopted certain external influences, in the events following the homicide, interactions between contrasting clinical risk management modes disrupted internally coherent practices and mentalities, such as DTC clinical leaders' commitment to democratictherapeutic principles. Instead of ready hybridisation, we found contradictions in which

differences between positions were reinforced, building tensions and overt conflict. In the homicide's aftermath, interactions between parties destabilised the internal coherence of ethicsorientated clinical risk management, partially disrupting its established practices, processes and expertises. Then, as the rules-based mode impinged upon indigenous practices, ideological and emotional tensions between these two modes escalated. Thus, interactions between clinical risk management regimes mixed distinctive techniques with underlying values. For instance,

authorities' focus on formal risk management dealt less with 'ontological facts'7 than with second order constructions of deviance, through which culpability might be externally attributed and sanctions applied (Douglas, 1992). This focus created moral outrage within the DTC, prompting vociferous 'truth telling' about what some perceived to be authorities' self-interested and defensive motives. These two interacting 'truth regimes' thus contained and mixed emotions and values as well as rational elements.

One theoretical suggestion emerging from the case is, therefore, the possible linkage between strong emotions and the possibility of hybridisation 8. Strong and sustained surges of organizational emotion may produce 'intractability' and reduce the scope for ready hybridisation to occur in the field. Conversely, we suggest hybridisation may be more likely to occur in settings with weaker emotional engagement.

7

In the case of the DTC, the government-commissioned risk report found the DTC to be 'essentially safe', while the

external homicide inquiry focused on outside agencies formally held responsible for John's and Mark's post- discharge care.
8

We are grateful to a referee for suggesting this point

49

We suggest these findings may helpfully rebalance the current scholarship on hybridisation, particularly its current emphasis on micro-technologies of control through routines and scripts. In Power's (2011) recent review of Foucault's impact on sociology and accounting scholarship, he stresses themes of power-knowledge, involving pervasive micro-technologies of disciplinary control. Within this perspective, there is a redirection of analytic attention away from ideology and values to an emphasis on classifications, routines and scripts, involving: ' a shift in ontological commitment from the cognitive basis of social order to a conception of order grounded in surface habits and practical action' (Power, 2011: 50). In our analysis, by contrast, we find the pervasive presence of competing values-based practices and risk management ideologies, accompanied by intense organizational emotions, and face-to-face confrontations. This is a 'hot' rather than 'cool' organizational climate, where values trump routines and scripts. Indeed, shared commitment to indigenous beliefs within a mode of ethics-orientated care was an important basis of internal resistance whereby some members sought to actively 'deface the currency' (Foucault, 2011) of imposed control technologies.

Foucault's (2010; 2011) later research into morally-charged refusal and courageous resistance in relation to normative codes helpfully informed the development of our heuristic. This heuristic enabled us to analyse our empirical case study in a structured and theoretically informed manner, sensitive both to rules-based technologies of 'governmentality' (Miller & Rose, 2008) as well as ethics-orientated care of oneself and others. Yet our findings suggest important limitations to Foucault's description of ethics-orientated care of the self. Whereas ethics-orientated social practice may indeed involve polemical resistance and 'the courage of truth', under some circumstances it can also produce perverse and ultimately self-damaging dynamics.

Does our case study have wider implications? Although we acknowledge this is a single case study drawn from an unusual setting, extreme cases such as this may illuminate important dynamics that are less visible in other settings (Eisenhardt, 1989). Similar contradictions may emerge where formal risk management practices are imported into other value-laden and emotionally charged settings, particularly where strong interpersonal engagement has historically 50

supported locally-generated modes of self-regulation. Such settings include human service organizations with clinical or educational orientations, social movement organizations and even religious settings, whose 'inner logics' may resist externally imposed risk management practices.

Future research should study further examples of contradictions as well as hybridisation between different modes of risk management. In the broader field of risk management, future work should specify conditions which limit ready hybridisation, exploring the possible utility of the concepts advanced here of values-based practices, organizational emotions and ethics-orientated resistance. Finally, future work which continues to move sociological accounting literature traditionally developed within corporate risk management and accountability systems into distinctive and (from this perspective) novel fields including clinical risk management may helpfully illuminate and problematise the broad contemporary phenomenon of rising formal risk management systems apparent across diverse organizations in much of society.

Acknowledgements
We would like to thank our research participants for providing generous fieldwork access during the 4 years of this empirical study. We acknowledge and thank the Editor (Christopher Chapman) and two anonymous reviewers for their comments and most useful guidance in developing this paper. Our particular thanks to Yiannis Gabriel as MF's first supervisor during this PhD research.

Bibliography
Adshead, G., & Jacob, C. eds. (2009). Personality disorder: The definitive reader. London: Jessica Kingsley. Alvesson, M., & Sköldberg, K. (2000). Reflexive methodology: New vistas for qualitative research. London: Sage. Armstrong, P. (1994). The influence of Michel Foucault on accounting research. Critical Perspectives on Accounting 5, 25-55.

51

Arnaud, G. (2012). The contribution of psychoanalysis to organization studies and management: An overview. Organization Studies 33, 1121-1135. Baron, C. (1987). Asylum to anarchy. London: Free Association Books. Barsade, S.G. (2002). The ripple effect: Emotional contagion and its influence on group behavior. Administrative Science Quarterly 47, 644-675. Bateman, A.W., & Tyrer, P. (2004). Services for personality disorder: Organisation for inclusion. Advances in psychiatric treatment 10, 425-433. Brown, P. (2006). Risk versus need in revising the 1983 Mental Health Act: Conflicting claims, muddled policy. Health, Risk & Society 8, 343-358. Burrell, G., & Morgan, G. (1979). Sociological paradigms and organisational analysis: Elements of the sociology of corporate life. Ashgate: Heinemann Campling, P., & Haigh, R. eds. (1999). Therapeutic communities: Past, present and future. London: Jessica Kingsley. Castel, R. (1991). From dangerousness to risk. In G. Burchell, C. Gordon, & P. Miller, The Foucault effect: Studies in governmentality. Chicago: University of Chicago Press, pp. 281-298. Charmaz, K. (2006). Constructing grounded theory: A practical guide through qualitative analysis. London: Sage. Charmaz, K., & Mitchell, R.G. (2001). Grounded theory in ethnography. In P. Atkinson, A. Coffey, S. Delamont, J. Lofland, & L. Lofland, Handbook of ethnography. London: Sage, pp. 160-174. Coid, J. (2003). Epidemiology, public health and the problem of personality disorder. British Journal of Psychiatry 182, 3-10. Department of Health. (1998). Modernising mental health services: Safe, sound and supportive. London: Department of Health. Department of Health. (1999a). Clinical governance: Quality in the new NHS. HSC 1999/065. London: NHSE. Department of Health. (1999b). National service framework for mental health: Modern standards and service models. Department of Health, & Home Office. (1992). Review of health and social services for mentally disordered offenders and others requiring similar services: Final report. Cm 2088. London: HMSO. 52

Department of Health, & Home Office. (1999). Managing dangerous people with severe personality disorder: Proposals for policy development. London: HMSO. Dolan, M., & Doyle, M. (2000). Violence risk prediction: Clinical and actuarial measures and the role of the Psychopathy Checklist. The British Journal of Psychiatry 177, 303-311. Douglas, M. (1992). Risk and blame: Essays in cultural theory. London: Routledge. Eisenhardt, K.M. (1989). Building theories from case-study research. Academy of ManagementReview 14, 532-550. Emerson, R.M. ed. (2001). Contemporary field research: Perspectives and formulations. Long Grove: Waveland. Emerson, R.M., Fretz, R.I., & Shaw, L.L. (1995). Writing ethnographic fieldnotes. Chicago: The University of Chicago Press. Emerson, R.M., & Pollner, M. (2001). Constructing participant/observation relations. In R.M. Emerson, Contemporary field research: Perspectives and formulations. Long Grove: Waveland, pp. 239-259. Ericson, R.V., & Doyle, A. eds. (2003). Risk and morality. Toronto: University of Toronto Press. Ewald, F. (1991). Insurance and risk. In G. Burchell, C. Gordon, & P. Miller, The Foucault effect: Studies in governmentality. Chicago: University of Chicago Press, pp. 197-210. Exworthy, T., & Gunn, J. (2003). Taking another tilt at high secure hospitals. The British Journal of Psychiatry 182, 469-471. Feeney, A. (2003). Dangerous severe personality disorder. Advances in psychiatric treatment 9, 349-358. Fiander, M., Burns, T., Langham, S., & Normand, C. (2004). [DTC] replication study - clinical progress & health economic strands integrated final report. London: St George's Medical School, University of London. Fineman, S. ed. (2008). The emotional organization: Passions and power. Malden: Blackwell. Fischer, M.D. (2008). An ethnographic study of turbulence in the management of personality disorders: An interorganisational perspective, Ph.D Thesis, Imperial College London, University of London: UK. Fischer, M.D. (2012). Organizational turbulence, trouble and trauma: Theorizing the collapse of a mental health setting. Organization Studies 33, 1153-1173. Flyvbjerg, B. (2001). Making social science matter. Cambridge: Cambridge University Press. 53

Foucault, M. (1977). Discipline and punish: The birth of the prison. London: Penguin. Foucault, M. (1979). On Governmentality. Ideology and Consciousness 6, 5-22. Foucault, M. (1988). Technologies of the self. In L.H. Martin, H. Gutman, & P.H. Hutton, Technologies of the self: A Seminar with Michel Foucault. London: Tavistock, pp. 16-49. Foucault, M. (1992). The use of pleasure: The history of sexuality volume 2. London: Penguin. Foucault, M. (1996). The ethics of the concern for self as a practice of freedom. In S. Lotringer, Foucault Live: Collected Interviews, 1961-1984. New York: Semiotext(e), pp. 432-449. Foucault, M. (2005). The hermeneutics of the subject: Lectures at the College de France 19811982. New York: Picador. Foucault, M. (2010). The government of self and others: Lectures at the College de France 19821983. Basingstoke: Palgrave Macmillan. Foucault, M. (2011). The courage of truth (the government of self and others II): Lectures at the College de France 1983-1984. London: Palgrave Macmillan. Foucault, M., & Blasius, M. (1993). About the beginning of the hermeneutics of the self: Two lectures at Dartmouth. Political Theory 21, 198-227. Foucault, M., Burchell, G., & Gordon, C. (1991). The Foucault effect: studies in governmentality: with two lectures by and an interview with Michel Foucault: University of Chicago Press. Gabe, J., Exworthy, M., Jones, I.R., & Smith, G. (2012). Towards a sociology of disclosure: The case of surgical performance. Sociology Compass 6, 908-922. Gabriel, Y. (1999). Organizations in depth: The psychoanalysis of organizations. London: Sage. Gans, H.J. (1999). The participant-observer as a human being: Observations on the personal aspects of field work. In A. Bryman, & R.G. Burgess, Qualitative research, vol. 2. London: Sage, pp. 39-54. Gendron, Y. (2002). On the role of the organization in auditors' client-acceptance decisions. Accounting, Organizations and Society 27, 659-684. Gendron, Y., & Bédard, J. (2006). On the constitution of audit committee effectiveness. Accounting, Organizations and Society 31, 211-239. Godin, P. (2004). 'You don't tick boxes on a form': A study of how community mental health nurses assess and manage risk. Health, Risk & Society 6, 347-360. 54

Gordon, C. (1991). Governmental rationality: An introduction. In G. Burchell, C. Gordon, & P. Miller, The Foucault effect: Studies in governmentality. Chicago: University of Chicago Press. Hammersley, M., & Atkinson, P. (1995). Ethnography: Principles in practice. London: Routledge. Heyman, B., Shaw, M., Davies, J., Godin, P., & Reynolds, L. (2004). Forensic mental health services as a risk escalator: A case study of ideals and practice. Health, Risk & Society 6, 307325. Hinshelwood, R.D., & Skogstad, W. eds. (2000). Observing organisations: Anxiety, defence and culture in health care. London: Routledge. Iedema, R., Flabouris, A., Grant, S., & Jorm, C. (2006). Narrativizing errors of care: Critical incident reporting in clinical practice. Social Science & Medicine 62, 134-144. Iedema, R., Jorm, C., Braithwaite, J., Travaglia, J., & Lum, M. (2006). A root cause analysis of clinical error: Confronting the disjunction between formal rules and situated clinical activity. Social Science & Medicine 63, 1201-1212. Iedema, R., & Rhodes, C. (2010). The undecided space of ethics in organizational surveillance. Organization Studies 31, 199-217. Janis, I.L. (1982). Groupthink: Psychological studies of policy decisions and fiascoes. Boston: Houghton Mifflin. Katz, J. (2001). Analytic induction. In N.J. Smelser, & P.B. Baltes, International encyclopedia of the social and behavioral sciences. Oxford: Elsevier, pp. 480-484. Kosmala, K., & McKernan, J.F. (2010). From care of the self to care for the other: Neglected aspects of Foucault's late work. Accounting, Auditing & Accountability Journal 24, 377-402. Kuhn, T.S. (1996). The structure of scientific revolutions: University of Chicago Press. Kurunmäki, L. (2004). A hybrid profession: The acquisition of management accounting knowledge by medical professionals. Accounting Organizations and Society 29, 327-347. Langan, J. (2010). Challenging assumptions about risk factors and the role of screening for violence risk in the field of mental health. Health, Risk & Society 12, 85-100. Latour, B. (1993). We have never been modern. Cambridge: Harvard University Press. Learmonth, M. (2006). Is there such a thing as "Evidence-Based Management"? A commentary on Rousseau's 2005 presidential address. Academy of Management Review 31(4): 1089-1091.

55

Lees, J., Manning, N., Menzies, D., & Morant, N. eds. (2004). A culture of enquiry: Research evidence and the therapeutic community. London: Jessica Kingsley. Luhmann, N. (1993). Risk: A sociological theory. Berlin: Walter de Gruyter. Maden, T. (2007). Treating violence: A guide to risk management in mental health. Oxford: Oxford University Press. Manning, N. (1989). The therapeutic community movement: Charisma and routinization . London: Routledge. Manning, N. (2003). Actor networks, policy networks and personality disorder. In S. Timmermans, & J. Gabe, Partners in health, partners in crime: Exploring the boundaries of criminology and sociology of health and illness. Oxford: Blackwell, pp. 140-162. McGivern, G., & Ferlie, E. (2007). Playing tick-box games: Interrelating defences in professional appraisal. Human Relations 60, 1361-1385. McGivern, G. & Fischer, M. D. (2010) Medical regulation, spectacular transparency and the blame business. Journal of Health Organization and Management, 24(6): 597-610. McGivern, G., & Fischer, M.D. (2012). Reactivity and reactions to regulatory transparency in medicine, psychotherapy and counselling. Social Science and Medicine 74, 289-296. McKinlay, A., & Starkey, K. (1998). Foucault, management and organization theory: From panopticon to technologies of self. London: Sage. Miller, P., Kurunmäki, L., & O'Leary, T. (2008). Accounting, hybrids and the management of risk. Accounting, Organizations and Society 33, 942-967. Miller, P., & Rose, N. (2008). Governing the present. Cambridge: Polity. Moran, P. (1999). Should psychiatrists treat personality disorders? Maudsley discussion paper no. 7. Institute of Psychiatry, Kings College London. O'Malley, P. (2004). Risk, uncertainty and government. London: Glasshouse Press. Ormrod, S., Ferlie, E., Warren, F., & Norton, K. (2007). The appropriation of new organizational forms within networks of practice: Founder and founder-related ideological power. Human Relations 60, 745-767. Power, M. (1997). The audit society: Rituals of verification. Oxford: Oxford University Press. Power, M. (2004). The risk management of everything: Rethinking the politics of uncertainty. London: Demos. 56

Power, M. (2007). Organized uncertainty: Designing a world of risk management . Oxford: Oxford University Press. Power, M. (2009a). Performance and the logic of the audit trail. Hurly-Burly: The International Lacanian Journal of Psychoanalysis 1, 193-201. Power, M. (2009b). The risk management of nothing. Accounting, Organizations and Society 34, 849-855. Power, M. (2011). Foucault and sociology. Annual Review of Sociology 37, 35-56. Rahaman, A.B.U., Neu, D., & Everett, J. (2010). Accounting for social-purpose alliances: Confronting the HIV/AIDS pandemic in Africa*. Contemporary Accounting Research 27, 10931129. Rapoport, R.N. (1960). Community as doctor: New perspectives on a therapeutic community. London: Tavistock. Reed, J. (1994). Report of the Department of Health and Home Office working group on psychopathic disorder. London: Department of Health and Home Office. Roberts, J. (1991). The possibilities of accountability. Accounting, Organizations and Society 16, 355-368. Roberts, J. (2001). Trust and control in Anglo-American systems of corporate governance: The individualizing and socializing effects of processes of accountability. Human Relations 54, 15471572. Roberts, J., Sanderson, P., Barker, R., & Hendry, J. (2006). In the mirror of the market: The disciplinary effects of company/fund manager meetings. Accounting Organizations and Society 31, 277-294. Royal College of Psychiatrists. (2008). Rethinking risk to others in mental health services: Final report of a scoping group. London: RCP. Ryan, T. (2004). Risk management and system failure analysis in mental health investigations. Health, Risk & Society 6, 303 - 305. Scherer, A.G. (1998). Pluralism and Incommensurability in Strategic Management and Organization Theory: A Problem in Search of a Solution. Organization 5, 147-168. Schouten, J.W., & Alexander, J.H. (1995). Subcultures of consumption: An ethnography of the new bikers. Journal of Consumer Research 22, 43-61. Seddon, T. (2008). Dangerous liaisons: Personality disorder and the politics of risk. Punishment Society 10, 301-317. 57

Spradley, J.P. (1979). The ethnographic interview. New York: Holt, Rinehart and Winston. Starkey, K., & McKinlay, A. (1998). Afterword: Deconstructing organization - discipline and desire. In A. McKinlay, & K. Starkey, Foucault, management and organization theory: From panopticon to technologies of self. London: Sage, pp. 230-241. Strauss, A., Schatzman, L., Bucher, R., Ehrlich, D., & Sabshin, M. (1964). Psychiatric ideologies and institutions. London: The Free Press of Glencoe. Tadajewski, M. (2009). The debate that won't die? Values incommensurability, antagonism and theory choice. Organization 16, 467-485. Townley, B. (1994). Reframing human resource management: Power, ethics and the subject at work. London: Sage. Tyrer, P. (2007). An agitation of contrary opinions. The British Journal of Psychiatry 190, 1-2. Vaivio, J. (2006). The accounting of "the meeting": Examining calculability within a "fluid" local space. Accounting Organizations and Society 31, 735-762. Van Maanen, J. (2011). Ethnography as work: Some rules of engagement. Journal of Management Studies 48, 218-234. Veyne, P. (2010). Foucault: His thought, his character. Cambridge: Polity Press. Waring, J. (2009). Constructing and re-constructing narratives of patient safety. Social Science &Medicine 69, 1722-1731. Warner, J., & Gabe, J. (2004). Risk and liminality in mental health social work. Health, Risk &Society 6, 387-399.

58



doc_951147022.docx