Volume 7, No. 1, Art. 31 – January 2006
Community Care and the Location and Governance of Risk in Mental Health
Abstract: The concept of risk is now central to all areas of health and social welfare in the UK, although its exact character in relation to different groups varies. It has been argued that risk in mental health has been characterised by a preoccupation with the perceived risk of violence to others posed by those experiencing mental distress, particularly since the implementation of community care policies in the 1990s. The present paper draws on qualitative materials from semi-structured interviews with thirty-nine mental health social workers to demonstrate the significance for policy and practice of identifying where professionals see risk as being located. In the present study, three key sites were identified: firstly, risk was located in dangerous individuals, where the concept "high-risk" was particularly closely identified with young Black men with a diagnosis of schizophrenia. Secondly, social workers located risk in within-subject entities such as active psychotic illness, when it was the symptom rather than the whole individual that was subject to surveillance and control. Thirdly, social workers located risk in social context and regarded risk in multidimensional ways compared to their psychiatric colleagues. The paper highlights how a theory of risk location can be a useful conceptual tool.
Key words: risk, surveillance, mental health, social work
Table of Contents
3. When People Become Hazards: The "High-Risk" Individual in Mental Health
4. The Segmented Individual and "Non-Human Agency" as a Locus of Risk
5. Social Work and the Power of Context in Locating Risk
Concerns about community care policies in mental health have centred on the belief that these policies have resulted in large numbers of people with mental health problems being discharged from asylums with insufficient resources to support them and with significant numbers of people consequently ending up on the streets (DEPARTMENT OF HEALTH 1998; ZITO 1999). Anxieties relating to "the mad" have therefore shifted as the space they are perceived to occupy—both literal and symbolic—has changed (LEFF 2001). It has further been argued that the shift from hospital-based mental health care to community care has involved a dispersal of disciplinary strategies in the Foucauldian sense (COHEN 1985), and that the same ideological positions that were witnessed in asylums, mainly in the form of medical dominance, have shifted to new and more dispersed sites, specifically the management of the "dangerous welfare 'other'" in the community (ELLIS & DAVIS 2001, p.138). Social workers, along with psychiatrists, are therefore "at the front line" in terms of public and political expectations about how the perceived risks associated with rapid social change in mental health care should be managed or "policed". 
For Foucauldian theorists, governmentality in its contemporary form is characterised by neo-liberalism, an integral aspect of which is the role played by expert knowledge to survey the population, or the "body" of society, and to ensure its productivity. Whilst this "discipline" can take the form of explicit means of control, it is more closely identified with forms of regulation which are directed at individuals who come to police themselves. LUPTON has provided a useful summary of how risk can be understood in the context of a governmental strategy:
"Risk is governed via a heterogeneous network of interactive actors, institutions, knowledges and practices. Information about diverse risks is collected and analysed …Through these never-ceasing efforts, risk is problematised, rendered calculable and governable. So, too, through these efforts, particular social groups or populations are identified as 'at risk' or 'high risk', requiring particular forms of knowledges and interventions." (1999, p.87) 
It has been argued that there has been a shift from "dangerousness" to "risk" in social administration which has profound implications for professional practice (CASTEL 1991). The chief characteristic of this shift is that intervention is now focused upon constellations of "risk factors" rather than upon concrete individuals or groups of individuals, with the consequence that "there is no longer a subject" (1991, p.288, emphasis in original). The shift to risk heralds a new approach to surveillance, which CASTEL terms "systematic predetection" (p.288), whereby risk assessment starts with a general definition of the dangers to be prevented rather than with the direct experience of some kind of threat based on contact with an individual. CASTEL identifies this shift, not only in professional practices within psychiatry but also in all of the social care professions, where increasing detachment from risky individuals is emphasised. PARTON (1996) has identified one such form of detachment in the virtual abandonment of "relationship" in social work in favour of care packages, where monitoring is the core element of practice. 
Whilst theorists such as CASTEL (1991) and PARTON (1996) suggest that social work may have already lost the struggle to retain a focus on the subjectivity of service users rather than objective risk factors, there are other theorists whose work suggests something different may be happening. Although writing in the 1970s, PHILP's (1979) historical perspective on social work predicts considerable areas of continuity rather than change in the way social work might be expected to operate under community care. This is because the space within which social work operates has always been in between objectified individuals and their subjectivity, and its role has historically been one of mediation between the two. Social work has thus been defined as a "liminal profession" (CHRISTIE 2001, p.12; WARNER & GABE 2004). According to PHILP (1979), social work practice is defined by the following three operations: the creation of subjects; the integration of objective characteristics and the function of speaking for the subject. The operation of creating subjects is described as follows:
"The creation of subjects is essential to mediation. The social worker faced, on the one hand, with an objectified vandal and, on the other, with a legal discourse has to attempt to present the underlying subjectivity of the vandal … The social worker does not say that the vandal did what he [sic] wanted to do, for in so doing the role of the social worker would disappear. What he [sic] does, rather, is to allude to the underlying character, the hidden depths, the essential good, the authentic and the unalienated. In doing so he is producing a picture of the vandal as a subject who is not immediately visible but who exists as a possibility, a future social being. Even if he does this without hope or cynically, he does it because it is the major factor which differentiates him from the policeman, the lawyer, doctor or psychiatrist. The object that these discourses deal with is one which is constructed out of facts and objective utterances. With his place in and between these the social worker cannot help but try to create people, subjects, where everyone else is seeing cold, hard, objective facts." (PHILP 1979, p.99) 
PHILP points out that social work cannot defend the subjective status of all individuals, but is "allocated those whose objective status is not too threatening" (p.99). Significantly, the boundaries that define those who are "too threatening" are constantly under negotiation. This paper argues that a period of sustained and intense negotiation is underway in relation to the entire "heterogeneous network of interactive actors, institutions, knowledges and practices" (LUPTON 1999, p.87) that form contemporary mental health services. 
The present paper's focus on the processes involved in constructing "risk objects" (HILGARTNER 1992) and identifying where these are located serves to highlight the extent to which harmful outcomes in mental health services, such as interpersonal violence, have not only been conceptually attached to certain groups of people but also to a range of entities associated with "riskiness". The process of constructing a risk object is in two parts: first, an entity is defined as an object and second it is linked to harm, which means that any entity, including one that is wholly conceptual, can be a risk object (ibid.). HILGARTNER also emphasises the importance of networks in which risk objects can be emplaced or displaced and the dynamic and conflictual nature of the processes involved in defining risk objects:
"… changes in the definition of risk objects can redistribute responsibility for risks, change the locus of decision making, and determine who has the right—and who has the obligation—to "do something" about hazards. Efforts to construct new risk objects, or redefine old ones, thus often take the form of intense struggles" (p.47). 
The agenda in mental health policy is indeed marked by an increasingly polarised set of tensions which are invariably the source of conflict. These tensions have been characterised in the literature as paradoxical in that they have constructed mental health service users as deserving of care within the community on the one hand, yet also as threats to its wellbeing (BARTLETT & SANDLAND 2000). Policy-makers must show that they are able to devise policies which will protect those who are mentally distressed and secure the best possible care for them. On the other hand, they cannot easily eschew the deep-rooted cultural concerns which continue to provide the wider context for political decision-making about mental health. The fact that these concerns have found a particularly powerful expression in the media, especially tabloid newspapers, constitutes a significant political pressure. Conditions at the political level have reconfigured the way professionals are held accountable for their practice in relation to so-called "high-risk" service users, and the anxiety engendered by numerous homicide inquiry reports have been a powerful political tool in this respect (EASTMAN 1996; MUIJEN 1996; SZMUKLER 2000). As PARTON (1996) has argued, an analysis of risk can highlight how social work practice has been reconstituted in important ways according to this highly-charged cultural and political context and it is in this area that this paper makes a contribution. 
This paper argues that the processes involved in the social construction and governance of risk can be best observed in analysing the way risk is located in different sites by practitioners in their everyday engagement with mental health service users. Qualitative materials from empirical work carried out with social workers are used to demonstrate how risk was located by them in three different sites, and a description of each of these serves to demonstrate the general process. First, the concept "high-risk" was associated with mental health service users who represented a "close fit" with the archetypal risk figure of the young male with a diagnosis of schizophrenia or personality disorder, where the focus was on the potential for violence to other people rather than for self harm by these individuals. The concept "high-risk" was associated even more strongly with young black men from this category, where Christopher Clunis1) stood as the proxy for young black males in general. In this instance, the location of risk was therefore closely related to the social location of the individual in terms of their age, "race" and gender. 
Second, "high-risk" service users were understood, and their behaviour interpreted, according to the way in which social workers interpreted and located risk within the individual. An important feature of this practice was the location of the "causal force" (BARRETT 1996) of a person's violent behaviour in either their mental illness or in their personality in order to decide how that risk should be managed, both in moral and in practical terms. Third, the same or similar risk objects were found to be interpreted by professionals from different occupational groups in different ways with different meanings attached to them. For social workers, risk was often located in entities that, for other professionals such as psychiatrists, were associated with reducing risk and social workers were also much more likely to locate risk in contextual factors. The next section describes the approach taken to data collection and analysis. 
Data used in this paper were gathered in the mid to late 1990s in an inner-city social services department in the South-east of England. The catchment area was socio-demographically diverse and the psychiatric morbidity of the population was high compared with the national average. Participants comprised all thirty-nine social workers working for the department who were qualified as approved social workers2) (ASWs) under the Mental Health Act 1983, including six team managers. The identities of participants and service users were protected by using a coding system to replace all names and by the use of pseudonyms for place names. These codes were applied throughout the process of transcribing and data entry. The research proposal was approved by the social services department and by the ethics committees of two mental health trusts. 
Every social worker completed two questionnaires prior to their interview. Questionnaire 1 provided biographical information about them, for example, the length of their experience as an approved social worker. Some details about each participant are given following each quotation in the paper. Questionnaire 2 asked the social workers to identify service users on their caseload whom they would define as "high-risk" and this yielded data concerning 219 service users. A considerable amount of time during interviews was spent discussing the reasoning behind the definition of particular service users as "high-risk". Statements in the paper concerning "high-risk" service users refer to this population. Semi-structured interviews were carried out by the author with all thirty-nine participants, each interview lasting approximately one and a half hours. With the consent of the participants, each interview was tape recorded. 
Prior to the transcription of tapes, an initial analysis of the interview data was undertaken so that general themes could be identified, in keeping with the need to become "familiar with the data" (MAY 2001, p.139). Once transcripts were available, the qualitative materials were coded by broadly following the three stages of open, axial and selective coding defined in NEUMAN's (1997) taxonomy (after STRAUSS 1987). The main objective in open coding was, as BERG puts it, to "open inquiry widely" (2001, p.251) with the aim of identifying themes in the shape of abstract concepts. The focus in axial coding was on organising key themes and ideas based on the initial set of "open" codes. Finally, these themes were extended and reorganised so that patterns in the data could be elaborated. MILES and HUBERMAN (1994) recommend the use of matrices and other cross-data display strategies in order to identify patterns, and these were also employed during the later stages of data analysis. 
3. When People Become Hazards: The "High-Risk" Individual in Mental Health
It has been argued that the major difference between risks in human services and those in technology is that the main source of the potential threat, or hazard, in human services, particularly mental health services, is people: "Recognising people as potential sources of hazard creates a number of tensions in human services that are difficult to resolve" (ALASZEWSKI 2002, p.185). This section explores some of these tensions in relation to mental health service users in more detail. In the present study, to be "high-risk" in mental health was found to be increasingly synonymous with being a danger to other people rather than to oneself (such as through suicide or self neglect). From the accounts of the six team managers, there was a consistent message about the political pressure to prioritise the management of risk of violence to others by service users over and above risks of self-harm, self-neglect or suicide. The following quote summarises the point:
Manager: "… I think it is very easy to just forget the people who are risks to themselves who aren't necessarily a risk to anyone else and it is easier to rate them at a lower risk generally … I suppose I feel the issue of risk is violence to others rather than to themselves."
Interviewer: "Do you have any ideas about why that is?"
Manager: "I think that is part of a number of reasons, partly to do with, if you like, the kind of whole society type thing of the media, the inquiries and everything, which are very much focused on the injuries to other people, the murders of the public, injuries to the public, rather than the attention that is given to the numbers of mentally ill who commit suicide or commit serious self-harm. I think that is part of the human defensiveness part of our job, protecting society is probably higher than protecting the individual from themselves." (Interview 34, female manager with 7 years ASW experience) 
Such a statement reaffirms the argument that risk in mental health has been politicised in explicit ways since the implementation of community care policies, with risk now meaning "danger to others". Risk was also identified closely with specific characteristics relating to age, "race" and ethnicity and gender, such that young Black men with a diagnosis of schizophrenia were a particular focus. This finding is consistent with findings in the literature that Black men are more likely to be perceived as "dangerous individuals" than people from other groups (BROWNE 1995; DUTT & FERNS 1998; HOLLANDER 2001; KEATING 2002; LORING & POWELL 1988). Homicide inquiry reports, exemplified by the inquiry into the murder of Jonathan Zito by Christopher Clunis, together with the powerful media accounts which have accompanied them, appear to have provided a template for this construction of risk (NEAL 1998). This is illustrated by the following interview extract and the reference to the service user (who is an African Caribbean man in his thirties with a diagnosis of paranoid schizophrenia) as being a "Christopher Clunisey type figure":
"He has actually got a history of violence; he doesn't like women; he has allegedly, allegedly I say because it was never brought to court, assaulted a 14 year old girl; and he doesn't appear to consider that he has any problems at all so I consider him a high risk and in a way he is not going to have any support networks apart from us and the psychiatric team when he is outside of hospital because he has not done anything yet and I have a horrible feeling that he will, and all we can do is keep a very close eye on him. I do see him as a very Christopher Clunisey type figure; that he moves around London a lot and we have to continue to chase him, so I would say he was very, very high risk actually, when I think about it, because he won't have any support networks." (Interview 39, female manager with 8 years ASW experience) 
Similar themes about "race" and the risk of violence emerged in a number of other interviews. The excerpt from a transcript given below exemplifies PRINS' observation, following his investigation and report in 1993 into the death of Orville Blackwood in Broadmoor Hospital, when he commented that, "(t)here was also a tendency to see such black patients as 'big, black and dangerous' because of their size and ethnic origins, where there was no evidence for such an assumption" (1999, p.132). The following extract also has the additional feature of risks associated with sexualised behaviour and the man's ethnicity:
Interviewer: "Were the risks perceived to be to himself or to others?"
ASW: "Both. Part of this behaviour pattern was making inappropriate approaches to members of the opposite sex."
Interviewer: "Did that ever extend to actual contact physically as far as you know?"
ASW: "As far as I know it didn't go as far as rape or anything like that, but he is a big built Afro-Caribbean man who would present as immediately threatening in any kind of approach and it was done in a very unsophisticated way."
Interviewer: "Where would you rank him [on your high-risk list]?"
ASW: "We are struggling with this. At one end of the spectrum people will define him as a Christopher Clunis." (Interviewer 33, male with 4 years ASW experience) 
As SAYCE (1995) has argued, the assumption that Black men are more likely to commit rape or sexual assault has a longstanding history and she suggests it is this assumption, among others, which has prefigured the assumed link between violence, "race" and mental illness. The consequences of such racialised constructions of risk in terms of the lived experience of individual service users are profound: "To be young and black, particularly for males, is to be deemed a greater risk and in need of increased surveillance and greater control…we are seen as requiring control as opposed to care …" (BROWNE 1995, pp.67-8). There is therefore a strong sense in which "racial otherness" and embodiment, specifically in terms of the symbolic significance of "black" (SHILLING & MELLOR 2001), have become central to contemporary constructions of risk in relation to mental health, including in professional practice. In the next section, the location of risk in the symptoms of mental illness is explored. 
4. The Segmented Individual and "Non-Human Agency" as a Locus of Risk
Work which has addressed the significance of psychotic symptoms as risk factors for violence implies strongly that it is the symptom and not "the person" which is the source of the propensity for harm (LINK & STUEVE 1994). The idea that mental illness is seen by professionals as representing a discrete part, or segment, has been developed in depth by BARRETT (1996) in relation to schizophrenia. 
In his ethnographic study of multidisciplinary practice (including social workers, nurses and psychiatrists) in a psychiatric hospital, BARRETT employs a Foulcauldian analysis of the way in which power is exercised through professional knowledges and practices which are taken-for-granted. BARRETT analyses how the person who is admitted to hospital becomes the subject of an assessment in which written documentation renders them a "segmented case" rather than a person. The "case" is sub-divided—for example, into "personality" and "illness"—before finally being reinvested with subjectivity. BARRETT identifies the most salient division as being one which separates the person from their illness. The process of segmentation is accompanied by objectification, so that the focus is on factors and variables rather than "the person": "People who work in psychiatric hospitals exercise power to transform persons by effecting subtle shifts in the way they are defined as parts or wholes, objects or subjectivities" (p.301). 
A further useful feature of BARRETT's work in the context of the present paper is the identification of schizophrenia as a "causal force". It is in this sense that conduct can be attributed to a "non-human agency" rather than to the individual or their "self" (WEINBERG 1997). The moral implications of this are clearly profound, in that if conduct such as violence is ascribed to a non-human agency then the individual concerned is no longer morally accountable for it. This is somewhat different to CASTEL's notion that the focus is entirely on "risk objects" rather than "the subject", since it is the shifting relationship between the two which is the focus of professional attention. Processes of classification in the sense introduced by BAUMAN (1991) then become a predominant concern, something that was evident in interviews with social workers. In relation to a number of their "high-risk cases", many social workers seemed concerned to identify where the cause of risk was located. The following extract from one interview illustrates well the way in which risks are deemed to be rooted in either the person's mental illness or in their personality:
Interviewer: "If you were to compile a list like this again today, would it exclude any of the names you put on the original list?
Interviewer: "You seem clear about that?"
ASW: "Hold on …" [Pause]
Interviewer: "Who were you having thoughts about? What went through your mind?"
ASW: "[Service user's name]"
Interviewer: "I take it you were considering this question about excluding her?"
ASW: "As of today, as it were. She has got a pretty volatile personality and I suppose I kind of experienced it in a slightly different way this week, which perhaps in my mind reduced the absolute risks that that poses."
Interviewer: "How did you experience it?"
ASW: "It seemed less rooted in her mental illness. It seemed to be more personality-based and by virtue of that, what is wrong with being volatile? If that is part of your personality in terms of psychiatric risks?" (Interview 33, male with 4 years ASW experience) 
In the above interview extract, the social worker momentarily reconsiders his assessment because this service user's behaviour appears more "personality based" than "rooted in mental illness". As such, the moral evaluation of it is different. In crude terms, her volatility rooted in her personality is cast as relatively harmless and her "subjectivity" (PHILP 1979) comes to the fore: "What is wrong with being volatile?" Her volatility rooted in mental illness, however, is considered a risk object subject to control by this social worker. 
The following extract from another interview illustrates the relative sense of ease with which the work with some high-risk service users was approached by social workers in the present study:
"Two of them suffer from manic depressive illnesses and when they become unwell they are a definite risk to others as well, so their admission always follows the same pattern. They become very unwell very quickly and the police are always involved and there is always threats of assault towards others, etc. In between admissions, they are very easy to work with. It is very specifically towards their specific illness and how it presents itself, very clear, and they are very clear about it as well. And the other [service user], I also put in the same light. His insight is a little bit less than the other two, specifically who, strangely enough, two of my most easiest clients to work with, and the two who show most insight; very easy to work with. They have developed immensely, but with each admission there has been a risk towards others so we are aware of it. It is quite strange in that way I was going through my case list. It felt quite strange to include them [on the high-risk list] in a way, because in other ways…they have shown so much potential and insight." (Interview 11, female with 2 years ASW experience) 
The predictability of the behaviour of these service users when "unwell" is clearly a major factor in the way their risk is managed. The important role of "insight" is also emphasised by this social worker. Notice how insight on the part of the service users suggests an alliance between them and the social worker in relation to the illness, since: "… they are very clear about it as well". This example again highlights the shifting nature of the sites in which risk is located. 
The treatment of mental illness as an object which can be contained was a theme which recurred in a number of interviews. Although service users who fitted this model were still considered "high-risk", it was felt by many social workers that their riskiness was relatively easy to work with. This was primarily because it was identified so closely with their mental illness. In other words, the risks they posed were finite in their dimensions and it was relatively straightforward for social workers to "create a subject" (PHILP 1979). These were people who appeared to provoke less anxiety or fear amongst professionals and they tended to evoke positive reactions from all members of the multidisciplinary team. Their level of insight was closely linked with a high level of compliance with treatment regimes. They might thus be defined as the "deserving ill". This was in contrast with most people whose personality was regarded unambiguously as the source of their high-risk status; these individuals were more likely to be regarded as responsible for their conduct, and thereby as "undeserving". In the next section the paper moves on to consider the location of risk in a wider context. This is most readily demonstrated by focusing on the theme of disagreement within multidisciplinary teams, which was one focus in the interviews with social workers. 
5. Social Work and the Power of Context in Locating Risk
It has long been argued that one of the key issues in collaboration between psychiatrists and social workers is the important ideological difference between them in terms of the models of explanation applied to mental health/illness. In crude terms, psychiatry has been identified with medical models and social work with social models of explanation (MILLER, FREEMAN & ROSS 2001). These ideological differences are believed to be a major source of some of the divisions and conflicts which are said to occur within multidisciplinary teams and such differences were evident in some, although not all, of the interviews with social workers in the present study. 
A review of the qualitative material from the present study showed that professionals working in multidisciplinary teams often interpreted the same information about risk differently, assessing the level of risk to be higher or lower than colleagues according to the different meanings they gave to "risk objects". The following extended extract serves to illustrate the complex ways in which these differences are related to ideological differences and the way risk is accordingly constructed by professionals. It is noteworthy that the respondents quoted below did not always use the language of risk in their accounts. However, the process through which entities such as a day hospital are conceptually linked to harm can clearly be seen. It is this conceptual linkage which warrants the use of the term "risk objects":
Interviewer: "Do you think that the psychiatrist normally shares your views about the level of risk faced or posed by this client?"
ASW: "I think [the consultant psychiatrist] sees him as being higher risk than I would. I would definitely say that he has potential for doing serious damage to someone and [the consultant] has written a report that says she sees him as potentially committing homicide, which I would agree with certainly. However, the majority of the time I wouldn't particularly see [the service user] as being a risk, although I think [the consultant's] feelings are that he is potentially more of a risk more of the time." […]
Interviewer: "Have there ever been any decisions taken in relation to the care of this client with which you have disagreed?"
ASW: "Not really. I am saying that because I can see both sides. For example, I suppose [the service user] attending a day hospital, I can understand why both [the community psychiatric nurse and the consultant] think he should do that because he is obviously going to be in closer contact with services and he is going to be monitored more closely than he is doing so now on discharge. Whereas I can also see his point of view, which is that he doesn't actually see himself as fitting neatly into the mental systems … a lot of people he knows have attended there for quite a long time, are actually quite different in their level of functioning. He has had a very high level of independence and he wants to retain that. I think he is going quite a long way along the lines of how we want to work with him in terms of keeping in contact with us and I think expecting him to go to a day hospital 5 days a week would probably be counter productive […] [the service user] does want to retain as much independence as he can but he recognises that he has had a couple of very severe breakdowns and he recognises that he needs to remain on medication if he is going to be reasonably sure of being stable … at least, I think he does … […] OK, I am sure that he will have problems that he will have to face when in the community and I am fairly confident that he will be able to deal with a lot of these problems. I just feel that, I can completely understand why he wants as normal a life as possible and to me, I feel we should be supporting that more with clients rather than getting them to fit with our medical model." (Interview 7, male with 1 years ASW experience) 
It is important to note that the social worker quoted above constructs the day hospital as "conceptually linked to harm" (HILGARTNER 1992) by virtue of seeing this service user's attendance against his wishes as "potentially counter productive". From the social worker's account, the medical team view the attendance of this person at the day hospital solely as an effective means through which he can be monitored and the risk he poses thereby reduced. They have located risk in this service user, identifying him as "a dangerous individual" in terms of a propensity for violence, which is regarded as a permanent feature of his character. The alternative, as highlighted by CHISWICK (1995), is to see dangerousness as arising out of certain circumstances or conditions. 
A similar pattern can also be seen in the following extract from another social worker from the same team, again with reference to a psychiatrist. In this case, the focus is on medication as being linked with harm, but also "hospital" which is constructed bi-dimensionally and therefore becomes a risk object because of the disruption caused by frequent admissions:
Interviewer: "In your experience, are there any specific issues over which there are more likely to be disagreements with other professionals?"
ASW: "I think there is always the issue of medicine and how people … what it means for someone to be taking medicine and, something else that is also very, very, underplayed is the impact of someone's home. [Service user's name] last year was admitted to hospital three times, which [the consultant psychiatrist] interpreted as being a case that was poorly managed, inadequate medication, it was quite blaming what she was doing and she missed out that [the service user] was decamped from her flat because the flat was going to be renovated, moved to another place and then had to move back within that year as well and that completely shook her world and she broke down because she was so angry that these changes were being imposed on her and it wasn't helpful for [the consultant psychiatrist] to say: 'look at you, you just keep breaking down', blaming her, kind of thing, without understanding what was happening in her life and I think that is where I end up doing the most kind of advocacy work, by saying to medical people that, when you see someone who is two stone heavier than they were three months ago: 'look, what is their motivation for taking medicine? We are not offering them something that is helpful.'" (Interview 6, female with 6 years ASW experience) 
The medical team are portrayed as viewing the hospital and medication one-dimensionally as part of a risk management strategy, whilst the social worker views them as bi- or even multi-dimensional, since they each may also serve to increase the level of risk. It is not being argued here that the operation of creating subjects precludes social work from also engaging in processes of surveillance and control. The point is that social work can be understood as doing both by virtue of its role in speaking for others and by "engaging in a process where the client is encouraged to see within himself his possibilities for social adjustment" (PHILP 1979, p.103). The paper now explores the wider implications of these findings. 
The main focus of this paper has been on the way contemporary social work practice is constituted in relation to risk, specifically the location of risk in different sites and according to different contexts. The paper argues that to focus on the question "what is the locus of risk?" is useful in a field such as mental health in at least one significant respect, which is that it throws into sharp relief the complex, multidimensional and often contradictory nature of constructions of risk. The implications of locating risk in "dangerous individuals" compared with locating it in specific symptoms of mental illness or in the wider social context in which individuals may find themselves are major. In simple terms, the "dangerous individual" will be subject to quite different forms of constraint and surveillance—such as via the criminal justice system—compared with the one who is identified as "safe" but for the risks located in their active psychotic symptoms when unwell. Under these latter circumstances, the paper has shown how it is the symptoms that may become the focus of risk management, and this management may be undertaken by an alliance which is forged between practitioners and the "insightful" individual service user concerned. 
There are clearly profound moral implications in terms of locating risk in different sites. In effect, the "story" about risk which can be told, both by and about an individual service user, will be very different according to what the "root cause" of their propensity for harm is presupposed to be at any given time. Focusing on the sites in which practitioners (and others) locate risk can therefore be a helpful conceptual tool in terms of understanding the fluid and unstable nature of constructions of risk in general, and may serve to clarify sources of conflict and ambiguity in relation to multidisciplinary working in particular. The findings in this paper also suggest that a theory of risk location may usefully be applied in a wide range of other contexts, both in health and social care and beyond. 
The author would like to thank the following: Professors Jonathan GABE and Mike BURY for their valuable contributions to the ideas in this paper; the Economic and Social Research Council, which funded part of the research; and everyone from the social services department where the research took place.
1) Christopher Clunis is a young African Caribbean man with a history of mental illness, who killed Jonathan Zito at Finsbury Park underground station in London in December 1992. <back>
2) Approved social workers are mental health specialists who, alongside psychiatrists, have legal powers to compulsorily detain people in hospital. <back>
Alaszewski, Andy (2002). Risk and dangerousness. In Bill Bytheway, Vivien Bacigalupo, Joanna Bornat, Julia Johnson & Susan Spurr (Eds.), Understanding care, welfare and community: A reader (pp.183-191). London: RKP.
Barrett, Rob (1996). The psychiatric team and the social definition of schizophrenia: An anthropological study of person and illness. Cambridge: Cambridge University Press.
Bartlett, Peter & Sandland, Ralph (2000). Mental health law: Policy and practice. London: Blackstone Press.
Bauman, Zygmunt (1991). Modernity and ambivalence. Cambridge: Polity.
Berg, Bruce L. (2001). Qualitative research methods for the social sciences. Boston: Allyn and Bacon.
Browne, Deryck (1995). Sectioning: the black experience. In Suman Fernando (Ed.), Mental health in a multi-ethnic society (pp.62-72). London: Routledge.
Castel, Robert (1991). From dangerousness to risk. In Graham Burchell, Colin Gordon & Peter Miller (Eds.), The Foucault effect: Studies in governmentality (pp.281-298). London: Harvester Wheatsheaf.
Chiswick, Derek (1995). Dangerousness. In Derek Chiswick & Rosemarie Cope (Eds.), Seminars in practical forensic psychiatry (pp.210-242). London: Gaskell.
Christie, Alastair (2001). Gendered discourses of welfare, men and social work. In Alastair Christie (Ed.), Men and social work: Theories and practices (pp.7-34). Hampshire: Palgrave.
Cohen, Stanley (1985). Visions of social control. Cambridge: Polity.
Department of Health (1998). Modernising mental health services: Safe, sound and supportive. London: Department of Health.
Dutt, Ratna & Ferns, Peter (1998). Letting through light: A training pack on black people and mental health. London: Department of Health.
Eastman, Nigel (1996). Towards an audit of inquiries. In Jill Peay (Ed.), Inquiries after homicide (pp.147-172). London: Duckworth.
Ellis, Kathryn & Davis, Ann (2001). Managing the body: Competing approaches to risk assessment in community care. In Rosalind Edwards & Judith Glover (Eds.), Risk and citizenship (pp.126-141). London: Routledge.
Hilgartner, Stephen (1992). The social construction of risk objects: Or, how to pry open networks of risk. In James F Short & Lee Clarke (Eds.), Organizations, uncertainties and risk (pp.39-53). Boulder, Colorado: Westview Press.
Hollander, Jocelyn A. (2001). Vulnerability and dangerousness: The construction of gender through conversation about violence. Gender and Society, 15, 83-109.
Keating, Frank, Robertson, David, McCulloch, Andrew & Francis, Errol (2002). Breaking the circles of fear: A review of the relationship between mental health services and African and Caribbean communities. London: The Sainsbury Centre for Mental Health.
Leff, Julian (2001). Why is care in the community perceived as a failure? British Journal of Psychiatry, 179, 381-383.
Link, Bruce G. & Steuve, Ann (1994). Psychotic symptoms and the violent/illegal behaviour of mental patients compared to community controls. In John Monahan & Henry J Steadman (Eds.), Violence and mental disorder: Developments in risk assessment (pp.137-159). Chicago: University of Chicago Press.
Loring, Marti & Powell, Brian (1988). Gender, race, and DSM-III: A study of the objectivity of psychiatric diagnostic behaviour. Journal of Health and Social Behaviour, 29, 1-22.
Lupton, Deborah (1999). Risk. London: Routledge.
May, Tim (2001). Social research: Issues, methods and process. Buckingham: Open University Press.
Miles, Matthew B. & Huberman, A. Michael (1994). Qualitative data analysis: An expanded sourcebook. California: Sage.
Miller, Carolyn, Freeman, Marnie & Ross, Nick (2001). Interprofessional practice in health and social care: Challenging the shared learning agenda. London: Arnold.
Muijen, Matt (1996). Scare in the community: Britain in moral panic. In Tom Heller, Jill Reynolds, Roger Gomm, Rosemary Muston & Stephen Pattison (Eds.), Mental health matters: A reader (pp.143-155). London: Macmillan.
Neal, Sarah (1998). Embodying black madness, embodying white femininity: populist (re)presentations and public policy—the case of Christopher Clunis & Jayne Zito. Sociological Research Online, 3(4). Available at: http://www.socresonline.org.uk/3/4/2.html [Date of access: December 15, 2005].
Neuman, William L. (1997). Social research methods: Qualitative and quantitative approaches. Needham Heights: Allyn and Bacon.
Parton, Nigel (1996). Social work, risk and "the blaming system". In Nigel Parton (Ed.), Social theory, social change and social work (pp.98-114). London: Routledge.
Philp, Mark (1979). Notes on the form of knowledge in social work. Sociological Review, 27, 83-111.
Prins, Herschel (1999). Will they do it again? Risk assessment and management in criminal justice psychiatry. London: Routledge.
Sayce, Liz (1995). Response to violence: a framework for fair treatment. In John Crichton (Ed.), Psychiatric patient violence: Risk and response (pp.127-150). London: Duckworth.
Shilling, Chris & Mellor, Philip (2001). The sociological ambition. London: Sage.
Szmukler, George (2000). Homicide inquiries: what sense do they make? British Journal of Psychiatry, 24, 6-10.
Taylor, Rod (1994/1995). Alienation and integration in mental health policy. Critical Social Policy, 42, 81-90.
Warner, Joanne & Gabe, Jon (2004). Risk and liminality in mental health social work. Health, Risk and Society, 6, 387-399.
Weinberg, Darin (1997). The social construction of non-human agency: The case of mental disorder. Social Problems, 44, 217-234.
Zito, Jayne (1999). Community care: Past failures, future directions. Probation Journal, 46, 101-105.
Joanne WARNER is Senior Lecturer in Social Work. Her main research interests are: sociocultural approaches to risk and mental health and documentary analysis.
Dr Joanne Warner
School of Social Policy, Sociology and Social Research
University of Kent
Bridge Wardens' College
Chatham Historic Dockyard
Kent ME4 4TE.
Warner, Joanne (2006). Community Care and the Location and Governance of Risk in Mental Health [30 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 7(1), Art. 31, http://nbn-resolving.de/urn:nbn:de:0114-fqs0601310.