Essay 2

1 Introduction `̀ Home'', he mocked gently. `̀ Yes, what else but home?'' `̀ It all depends on what you mean by home. Of course he's nothing to us, any more Than was the hound that came a stranger to us Out of the woods, worn out upon the trail.'' Home is the place where, when you have to go there, They have to take you in.'' `̀ I should have called it Something you somehow haven't to deserve.''

Robert Frost (`̀ The death of the hired man'', 1915)

Although at first glance home seems to be the natural situation where care takes place, it does not take too long to imagine instances where that situatedness grows compli- cated. As Frost's stanzas show us, home can be an especially dense place of meaning and emotion around questions of caring for terminal illness. As the hired man returns to his former employers' home to die, the farming couple debate their obligation to care for him; whether their home is actually his home, and where he should go to die if they turn him away. Frost wrote early in the 20th century when home death was becoming increasingly rarer and rarer (Palmer, 1993). By the 1950s most Americans (60.9%) died in hospitals (Brim et al, 1970); this left the question of where dying ought to take place, and who should do the work of care, largely answered by modernity (Seale, 1998).

Recent trends, however, make the farming couple's debate over home care once again salient. There has been a clear spatial shift in death away from hospital and towards the home both nationally (Field and Cassell, 1997) and locally in Washington State (Brown and Colton, 2001). The purpose of my paper is to examine one specific set of consequences for this new geography: what home is like as a place for terminal caregiving. Indeed not all home deaths are hospice, but hospice is increasingly

Hospice and the spatial paradoxes of terminal care

Michael Brown Department of Geography, University of Washington, Box 353550, Seattle, WA 98195-3550, USA; e-mail: [email protected] Received 26 April 2002; in revised form 27 August 2002

Environment and Planning A 2003, volume 35, pages 833 ^ 851

Abstract. The purpose of my paper is to offer an understanding of home hospice from a perspective of political geography. Informed by critical political theories of care, and recent work on the geographies of public and private spheres, I explore one set of consequences of the spatial shift towards home death in metropolitan Seattle, Washington. Terminal hospice care done in the home creates an especially paradoxical home space. By blurring public ^ private boundaries, hospice care produces a political geography of home interpretable through four spatial paradoxes: a normative paradox of home being a good and bad place to die, a territorial paradox of control itself changing the home, a constitutive paradox between heart and welfare politics, and a relational paradox between autonomy and dependency. The implications for political and health geography, as well as political theory and hospice work itself are discussed as a consequence of recognizing these spatial paradoxes.

DOI:10.1068/a35121

structuring home deaths in the United States generally; and specifically throughout metropolitan Seattle (Hoyer, 1998; NAHC, 2001; NHPCO, no date).

Recent work in feminist political theory and geography provides an intellectual framework within which hospice work in Seattle can be understood (Butler and Scott, 1992; Fraser and Gordon, 1997; Harrington Mayer, 2000; Pateman, 1989; Tronto, 1994; White, 2000). This work challenges key tenets of liberal democratic theory and con- servative political thought that prevent us from interpreting care as a fully political set of relations. As it focuses in on many blindspots in political theory, so too must it speak to political geography, which often remains fixated on only certain kinds of political relations fixed at particularly large (that is, nation-state) spatial scales. It asks us to consider the home as a decidedly political geography (Young, 1997). In this way, recent theoretical work on care nourishes and extends recent work in political geography that appreciates multiple forms and locations of politics. It also fuels recent developments in health geography that seek a social understanding of where caregiving takes place, especially in the home itself (Dyck, 1998) and in the context of welfare-state restructuring (Brown, 1997; Gleeson and Kearns, 2001; Milligan, 2000). If care is to be reconceptualized as a political concept, geographers must explore its spatiality. In order to develop an understanding of the spatial dimensions of hospice care in the home, I draw on a series of in-depth interviews with forty workers, volunteers, and family carers involved in the six local hospice organizations in Seattle, Washington between 2000 and 2001. Interviewees were asked to explain the practice and philoso- phy of hospice, to describe their work in hospice specifically, and to describe the settings of hospice care. Interviews were recorded, transcribed, and returned for edit- ing. Only edited transcripts are used for quotation. A grounded-theory approach was used to code the interviews on the geographies of home death and caregiving. The goal of the interviews was twofold: to gain an understanding of what home is like in a hospice situation, and to determine how the nature of terminal care work has changed with this spatial shift from hospital to home.

From these interviews, I develop a political geography of hospice by drawing out the consistent theme of home as a spatial paradox, which emerged from the inter- views as a recurrent theme in understanding the `nature' of home hospice care. Four dimensions of this spatial paradox were most salient in Seattle: a normative paradox of home being a good and bad place to die, a territorial paradox of patient control itself changing the home, a constitutive paradox between heart and welfare politics, and a relational paradox between the autonomy and the dependency of the dying. Though discussed separately, I do not mean to suggest that the paradoxes are empirically discreet. Indeed, these contradictions can be understood simultaneously as they all signify home hospice. But parsing them out thematically helps specify the particular meanings that help us understand the home as a political geography of death and terminal care.

2 Theoretical orientations 2.1 Political theory and care There is a growing awareness in contemporary political theory that its objects of interest historically have been quite partial and incomplete. One of the most remarkable gaps has been the question of care and caregiving. As a starting point in rethinking care more politically, Tronto (1994, page 103) argues that the concept suggests,

`̀ ... a species of activity that includes everything that we do to maintain, continue, and repair our `world' so that we can live in it as well as possible. That world includes our bodies, our selves, and our environment, all of which we seek to interweave in a complex, life-sustaining web.''

834 M Brown

Here care is an emotional and subjective orientation, but it is also an orientation often linked to action. Care can be both emotion and labor, typically culturally signified. Its emotive and material facets mark it as a form of `heart politics' in Peavy's (1996) eloquent words. This reconceptualization of care challenges liberal-democratic and patriarchal assumptions and arguments that fracture justice and care in a binary system of political action. Whereas justice, a key concept in political theory, is understood rationally and as a universal; care is not drawn on the political theorists' map, it is a particular idiographic phenomenon, more practical than principle. It is highly emotional and irrational. It is women's work: prepolitical and congruent to the private sphere. Without being cared for in the private sphere of the home, liberal man could not fully and rationally enter the world of the public sphereöwhere he could pursue freely his already-articulated notion of the good life (Pateman, 1989). Yet care is political precisely because it embodies issues typical of politics in a democracy; questions over the allocation of public resources as well as agonistic relations wherein equity, justice, obligation, and rights are lived.

Whether it has been deployed normatively or descriptively, this dualism is well recognized and persistently deconstructed by feminists (Butler and Scott, 1992). It is not simply that questions of care have been ignored, but rather that the entire demarcation of the nature of welfare politics has been achieved only because of this exclusion. Considering explicitly terminal care and death, Young (2000, page 487) puts it well:

`̀ The public realm of politics can be so rational, noble, and universal only because the messy content of the body, meeting its needs, providing for production, caretaking, and attending to birth and death, are taken care of elsewhere.''

Questions of care and death, then, cannot simply be mapped onto existing liberal democratic maps of the political. They transform its very foundations. Thus Tronto (1994) and White (2000) (and also Harrington, 1999) offer a new starting point to recognize care in the world. Foremost is a rejection of the assumption of autonomy for the political subject. We must recognize the interdependence of people throughout their social worlds. In other words, dependency or heteronomy is more often the norm than autonomy. Akin to this point is the argument that, whether emotion or action, care is always already a social relation. It structures relations between carers and dependents that help mutually to constitute subject positions of gender, class, abilism and race inter alia. Rather than just being a discrete attribute of individuals, care relationally helps make them how they are who they are. In this way, caring practitioner-experts alone cannot define care. Caring cannot be just a discourse of experts from within the welfare state. Nor can the intellectual work on care only be to `assess' one's need for it.

2.2 Multiple ontologies of hospice One of the most fascinating attributes of `hospice' is that it has a three-fold ontology. It is a philosophy, a welfare policy, and a place. The philosophy of hospice stresses the comfort of the dying patient in order to encourage the best death possible. It stands in stark opposition to the typical curative hegemony in biomedical discourse that encour- ages intrepid and heroic measures to save the life of the patient. Here, death is the enemy. For doctors, it is a symbol of their failure: they did not care enough. By contrast, hospice asks us to admit the inevitability of death for everyone, but especially for those with an obviously terminal illness. Once we admit that a patient will not get better, hospice argues that the curative philosophy is no longer appropriate; indeed it can be quite harmful where extending life may diminish the quality of life (for example, Duda, 1987; Webb, 1997).

Hospice and the spatial paradoxes of terminal care 835

With the mounting attacks on biomedical authority, and the increasing rationalization of expensive medical procedures, the philosophy of hospice in the United States gained respect and audience throughout the 1970s. By 1983 Medicare (the federal health-insurance system for elderly Americans) introduced the hospice benefit nationwide. Soon after, Medicaid (the federal health-insurance system for disabled and poor Americans), private insurers, and Health Maintenance Organizations offered the same benefit. Thus over the past twenty years hospice has become a policy of the North US welfare state and its private shadow state (Hoyer, 1998). Although it is true that a relatively small number of North Americans die in hospice overall (about 29% of all deaths in the United States in 1999), the estimated enrollment in the benefit has increased from 540 000 in 1998, to over 600 000 in 1999 (NHPCO, no date). The total number of hospices has risen from 31 in 1984 to 2273 in 2000 (NAHC, 2001). The hospice benefit offers comfort care to terminally ill patients who have six months or less to live. There can be no rehabilitative or curative treatment once hospice has been started. The core of the benefit is pain control. A hospice benefit grants a dying patient, and their families, care from a team of five specialists: a nurse (who manages pain control and corporeal comfort intermittently within 24-hour on-call service), a social worker, a home-health aid (who deals with `hands-on' duties such as bathing), a chaplain (who provides nondenominational spiritual support), and volunteers (who provide respite to the family carers). The fifth member of the hospice team is the family itself, which provides the day-to-day care of the patient. The Medicare hospice benefit may also cover medical appliances and supplies related to the terminal illness: outpatient drugs for symptom management and pain relief, short-term acute inpatient care, including respite care, and where necessary palliative physical therapy, occupational therapy, and speech or language pathology services (see www.hospicenet.org).

For geographers, however, it is especially interesting that this philosophy policy is also a place. Moreover, in contrast to the British model (where hospice began), that place is the patient's home for most Americans. Haupt (1998) reports that in 1996, 78% of American hospice patients were in the home. This figure is likely to be higher in Seattle, as only two hospice organizations of the six in Seattle, have in-patient facilities (one of which is about to close, and the other which contains fewer than ten in-patient beds). Hospice rarely takes place in the hospital itself, though it certainly can occur in a nursing home. Hoyer (1998) notes that, because Medicare limited the number of days hospice beneficiaries could stay in hospital and still receive the benefit, the benefit ruled its geography as the home. The point is, it is unusual for hospice to be in a stand- alone building where people explicitly stay for comfort care while they die. Most Americans who subscribe to the hospice benefit, and are embodying the hospice philosophy, are doing so in a residence of some sort.

2.3 Geographies of public and private This spatial imbrication of hospice care means a political geography of hospice care is needed to appreciate certain consequences in the spatial shift towards home death. For just as political theorists ask us to be suspicious of what the public ^ private divide hides in actually existing democracy, so too have political geographers asked us to question the dichotomy spatially. Both Milligan's (2000) and Staeheli's (1994; 1996) work is especially helpful here. Examining the provision of elderly care in homes, Milligan draws our attention both to the macrolevel and to the microlevel restructurings that are placing public welfare caring in `homespace'. As she puts it: ``the provision of formal and informal care within the homespace is creating a new site of caring in which the boundaries between public and private space are becoming blurred'' (Milligan, 2000, page 56, emphasis mine). The specific lineaments of home hospice's paradoxes are part

836 M Brown

of what makes home `new', and extends her insights empirically. Staeheli also provides some guidance here, by asking us to be careful in specifying whether we are discussing public and private space versus public and private actions. The two dimensions of these spheres are not always coterminous. And so I would further argue that home hospice exemplifies why Staeheli's warnings are so important. Hospice works precisely on the blurring of space and action; publicity and privacy. It is a public policy of the welfare state, but there are a number of private providers who offer a hospice benefit. Medicare and Medicaid will reimburse private insurance companies for provision of hospice, and the benefit is largely the same whether there is reimbursement or not (that is, whether the patient or the government is paying the bill). Although it takes place in the private space of the domestic residence, much of its work is done in civil society: by professional strangers delivering a service. There is obviously a mixture of public and private actors on the hospice team. Most of the team are paid professionals, but there are also unpaid volunteers and the bulk of the care work is done privately by members of the family. The relations of care between these actors and the patient are often both public and private. And then the hospice death itself is a curious mixture of public and private spheres. Awareness and preparation of one's own death must surely be an intensely private experience, yet in hospice it is worked through relations with social workers, chaplains, and families. One dies in front of them very publicly as it were. Hospice, then, is a particular constellation of public and private actions and spaces, largely occurring in typically private home space.

In sum, a preliminary understanding of the geography of hospice suggests quite a paradoxical space of terminal care. This reading stands in direct opposition to liberal- democratic theory's geography of the home as prepolitical, people as fully autonomous, and care work as largely emotional and charitable. The blurred boundaries between public and private require sensitivity to how power relations are spatialized in the home through the experience of hospice. Clearly home hospice is a philosophy and policy that spatializes a politics of care. Neither simply private space nor private actions; they are part of the constitution of how and where we `take care of our world'. To interpret what home is like as a consequence of this point, we must pursue a political geography of home care. This interpretation draws on, and spatializes, politi- cal theories of care. This geography is important not simply because it emerges from a spatial shift in where death typically takes place in the city, but also because spaces of death themselves are typically terra incognita: it is a geography that remains largely hidden and unknown (Palmer, 1993). Next I interpret what home is like, not simply because it is a complicated, paradoxical space where care is politicizedöbut also because it is such an unknown place, because death has typically been placed in institutionalized settings. The paradoxical nature of home hospice is interpreted through four themes that emerged from interviews with hospice workers. If the fore- going discussion argued that home is a paradoxical space, what I deal with next tries to trace how such paradoxes are made meaningful at home.

3 Spatial paradoxes of hospice 3.1 The normative paradox Perhaps the simplest, yet most compelling paradox discussed by hospice workers and families was the capacity of home to be coded normatively as both a good and a bad place to die. The sense of comfort and control (see below) afforded the dying patient by having them actively die in a space that is familiar, and noninstitutional obviously aids in `the good death' (for example, Duda, 1987; Webb, 1997). Indeed, the claim was such a doxa that interviewees had a difficult time elaborating on themes of comfort and control when probed why home was such a good place to die. For long-time hospice

Hospice and the spatial paradoxes of terminal care 837

workers or volunteers, however, the clear benefit of hospice was discussed in contrast to the horrible conditions they had witnessed in hospitals before there was hospice. One former nurse described it in detail:

`̀ In '65 I became the evening shift supervisor in the hospital and we implemented the ability to code people, which is probably hard for you to even imagine that you wouldn't have that. ... And my eyes were opened wide because we indiscriminately coded people, it didn't matter if they had end-stage cancer or kidney disease, anyone who started to die in the hospital abruptly got coded. It's like, well, this isn't what we meant. I mean, wait a minute, we need to think about what we're doing here. So we're talking 1964 to 1968 and so it's like, okay, so now we have all this aggressive treatment and options, which you know how much more so it is today, so how are we making the choices about what we use? Should we use it just because we have it to use or do we need to create some criteria, standards, medically and ethically and whatever else? So I began to say there's got to be a better way to do this and started to look around, has anyone found a better way to do this?'' Another way home hospice was coded as `good' was through the claim that it gave

ordinary people a chance to exceed their own expectations to do the difficult work of terminal caregiving. Here we can recognize Tronto's insight that understanding caring requires an appreciation of caregiver needs, as well as the needs for the cared for. A bereavement coordinator described the c̀oolness' of self-development and emotional fulfillment that carers often realized through their hospice work:

`̀ It is so cool to know what you are capable of or your species is capable of... . So if we normally just live in this little narrow range of our emotions, this experience can open us up that much more. And I can be moreöI don't have to die as I live and I don't have to continue living as I did live, I can be different, I can grow and mature.''

Another reason the home was a good place to die was because it allowed for physical intimacy between the patient and their loved one. Eloquently describing what she called `̀ skin hunger'', a retired hospice nurse reminded me that dying does not necessarily supercede our need for physical love:

`̀ I just remember when I was a fairly new nurse we had a woman who was about my age at the time, maybe 24, who was dying of leukemia, and she would go home on the weekends just so she and her husband could sleep together. I think they were still having sex too now and then, but mostly it was just that intimacy, that holding onto each other that you want to do, with both your grief and all of that, and the hospital doesn't give you that privacy to do that or permission. .... But I just can't imagine dying and not having that physical contact. That's such a normal part of your life but in the hospitals there's no recognition of that. So that was another thing that hospice really tried to talk to families about so that they could talk with us about it. We had to normalize it so that it was okay to talk about it. Am I too weird? I want to have sex with my wife and here she is dying, I don't even want to talk to her about it. Then we'll go in the other room and talk to her, and she goes, oh, I miss my husband, I don't think he finds me attractive any more and da da da. So that is most of hospice's work is to get the two conversations together.'' But the goodness of home was not a ubiquitous theme. Even the most enthusiastic

hospice workers readily pointed out that home was not necessarily or categorically the best place for everyone to die. In some ways it could be worse than the alternatives of hospital or a nursing home. Challenging my own rather romantic and stylized notions of home death, three themes emerged that drew out the darker side of home death.

838 M Brown

Most problematic was the issue of unresolved family tensions and disputes that only get exacerbated when a family member is actively dying and kin are doing the bulk of the labor to care for them. I was widely disabused of the notion that an impending death in a hospice situation prompts family members to put aside their differences and work together to enable a good death. A social worker cautioned against romantic notions of hospice in her work, which often raised unrealistic expectations for the practice of hospicing generally, and home death specifically:

`̀ [People] die the way they've lived. So if they've had a very chaotic, crazy kind of lifestyle, that's not going to change just because they know they're going to die.''

This sentiment was common sense for most hospice nurses and social workers. And so they explained that placing hospice in a home where the family dynamic was already strained or destroyed was ill advised. For as much as they wanted to see more hospice deaths in the population, they readily admitted it was not for all families. The bereavement coordinator (also quoted above) explained this sentiment to me quite sharply:

`̀ It is rare that in a time of great stress, people decide to work on really healthy coping skills. They go back to the ones that have worked, which can be alcoholism, can be overeating, can be a reckless behavior, can be checking out, any form of what that means. ... And it is the rare person who says I'm going to get a brain tumor [in order] to work on those pesky family dynamic issues I've always meant to work on. And it just doesn't happen. So the last thing they want to do is dredge up why we're not getting along now. Thus, a really dark side is your poor coping skills might be kicked into effect again and you're off and running and then grief is really ugly and hard and people don't support you and you're stuck with it. So yeah, I am so proud of hospice and I would never want to do anything else, but when people don't choose it, I go, `Hey! Not a bad idea!''' Another way in which home is not necessarily the place of good death is when we

recognize that some people simply cannot die at home, no matter how much they want to. In terms of disease ecology, certain illnesses and unplanned deaths make a hospice death impossible or impractical for home. Most obvious, perhaps, is the fact, that if a terminally ill patient has no family, or `primary caregiver' in the parlance of the benefit, then home hospice is not possible because the family do the bulk of the care.

We must also recognize that, if the family is unwilling to perform the work of hospice, or is unable to, then a home death will not occur. These inabilities can take many forms. They may be physical or practical, but they can also be cultural and emotionalöas the next excerpt, from a longterm hospice activist in Seattle, demonstrates:

`̀ Sometimes the idea of providing direct care to anyone, regardless whether it's their spouse, daughter, uncle, son, grandfather, grandmotheröjust the idea of providing that kind of care to someone is not within their idea of possibility for themselves. Or it may be role based or relationship basedöthey couldn't do that for their spouse or they couldn't do that for their mother or their father. Michael Brown: Changing a diaper or ö Interviewee: Right, it challenges the role that they're in with that person, they just can't get beyond that ideaösometimes it's just a matter of experience, they've just never done it before. And just haven't crossed that barrier of I don't know whether I can do this or not, because they haven't had any practice at it. Michael Brown: Is it also the denial of death at work?

Hospice and the spatial paradoxes of terminal care 839

Interviewee: Sure, there are definitely issues about death and dying that challenge that too. You know, some people don't want to have someone die in their home. That's just not possible for them to do, it's not comfortable to have somebodyö it can also be a matter of protection too on both parts, the person who is sick and the caregiver, because the person or person who could be a caregiverö some people don't want family to take care of them. They're very clear that they don't want their family to do that for any number of reasonsöcould be economic reasons, I don't want to take them from their jobs, from their source of income. It could be role relationshipsöI don't want that person taking care of me, so the driver could be the caregiver, the patient, or a combination of them. Or again, it's just the rules in the familyöyou don't do that and I'm not going to break those rules. You got to go someplace.''

The normative paradox of home in hospice care makes an argument for the relational and political dimensions of care. As a culturally situated process, the nature of care is highly contingent: neither universally good nor `bad'. So in turn, there are often debates whether or notöor in what respect^ the home is the best place to die. There is agonism around home hospice care that defines the home itself.

3.2 The territorial paradox Territoriality is a ubiquitous theme in explaining why the home is the proper site for the care and control of the dying patient. Home is not simply the object you control: it is a means of giving you a sense of that control. The argument in hospice philosophy is that patients are simply more in control at home than they are in institutional settings such as hospitals or nursing homes. Paid carers and volunteers expressed an acute awareness that they were `on someone else's turf', even if they were acting with external authority and control (for instance, bathing someone or deciding on or administering medication). From a hospice nurse:

`̀ Well, I would say that one thing is that patients have more control. In the hospital like it or not you know you have got a lot more control. It's your setting, you know where everything is, and you walk into someone's home, and it puts you at a whole other level. And as far as I'm concerned, it's an improvement. You lose a lot of control in the homes, but I personally think that the control should always be the patient. You know, so that's why I was never an ICU [intensive care unit] or a surgery nurse, it was, it just didn't appeal to me. So I think that that is probably the biggest deal: that you are really on their turf. And they get to call the shots. You know if we got to walk into every home and truly be able to do it all away, everybody would have a more peaceful death. But that's not the way it is. It's their lives, it's their death, it's their turf, and people continue [to] make choices that we think, brother, this is crazy, but it's their choice to make, and all you can really do is give them the options and the consequences of the choices they are making, but it's their choice. So, and that happens in the hospital to some degree, but it may happen in five minute increments, but in the home, it's like bigger.'' Themes of control, comfort, and familiarity were woven together for several res-

pondents. Often the patient's territoriality took the form of arranging elements in the home: typically, photographs were brought into the room, and placed close to the bed and pets and children were often allowed onto the patient's bedöa situation typically forbidden in institutional settings. The power in the comfort of bedding and clothing seemed especially important. Patients could have their own sheets, towels, blankets, and pillows at home. More than one hospice nurse noted that the patient's bed was often rearranged toward a favorite view. For one woman it was her backyard garden, that she had labored in for years, which helped situate death as a natural process.

840 M Brown

For another, it was a spectacular living-room panorama of Puget Sound and the Olympic Mountains in the distance. For others, it was the television. Sounds and smells also expressed this territoriality to the extent that they provided sensory comfort through familiarity. Familiar sounds may just be the regular din of a house, pets, or children. In terms of smells carers often discussed the importance of things such as cooking the patient's favorite meals (though see below), or having their own soaps and shampoos instead of institutional products These microexercises of power, it was claimed, worked to muster some sense of control over an inevitably powerless situa- tion. Being at home spatialized a politics of control and familiarity through what would otherwise seem to be a mundane set of environmental factors.

But home must change when people are hospiced there. So the paradox emerges: is it still `home'? In other words, the placement of hospice itself in the home has the inevitable effect of changing the home itself. And the extent to which there is a loss of territoriality because of this paradox was noted in several ways.

To a geographer the most obvious loss of territoriality was the fact that home itself sometimes had to change locationally in order for hospice to work. The research revealed a wide variety of home situations for dying people which were beyond their own residence. If higher degrees of hands-on care were required, and the patient could afford it, nursing homes, skilled nursing facilities, and adult-family homes could become home. In the nursing-home situation, the hospice team treated both the family and the staff as primary caregivers. Skilled nursing facilities are increasingly common in Seattle and can be thought of as a sort of hybrid between a home and a nursing home. Residents have their own apartments in a single building, and they are largely self sufficient for their own care needs, but they also have a nurse present in the building. In Seattle, these `Sniffs' as they are called, are practically indistinguishable from high-rise apartment buildings in surrounding neighborhoods. Adult-family homes are even more home like. Typically situated in large single-family homes in suburban neighborhoods, they are owner-occupied, and the owner provides primary care herself or himself to a small number of residents. Precisely because of the changing politics of care in society, the nature of domestic residences is changing for elderly and dependent populations.

If the dying patient remains in his or her own original residence (and often the demand to stay is itself the form territoriality takes), interpersonal dynamics within the home can change because a primary carer moves into the patient's dwelling to provide round-the-clock care for the terminally ill patient. Typically this is the situation with older women, who have outlived their spouses. If the patient can afford it, a live- in professional carer is an option, though hospice workers noted this was extremely rare because of the cost. Much more common would be a family member leaving his or her own home and moving into the dying patient's residence to provide direct 24-hour care. Either way, this new person around the house changed the environment if for no other reason than he or she had not lived there before hospice. Nationally we know that only 40.7% of noninstitutionalized hospice patients who had primary care- givers had their spouse in that role. Adult children were primary carers for 32.6% of these patients (Haupt, 1998). Although we do not know exactly where these caring relations were situated, these data confirm the interviews' themes that the home changes because of caring relations. My point is that having these people in the home can disrupt the normalcy of home and can sometimes challenge the territoriality of the patient.

The location of home can even change for both the carer and the patient together. One remarkable daughter changed her own residence in order to bring her mother home to hospice:

Hospice and the spatial paradoxes of terminal care 841

`̀ You know what I did. Okay, I knew she was coming home and I had a great apartment but it was on the third floor and there was no elevator. So I started looking all around my neighborhood because at this point we didn't know how sick she was going to be. So I found an apartment down the street and I was going to put her in that apartment and I was going to be here, and then it turned out maybe one week, so I rented yet a third apartment for my mother, and it was this big hugeöI realized it just couldn't be, she had to be with me all the time. So it was like this 1900 square foot apartment that was like all hardwood floors, perfect for a wheelchair, and so it was two bedrooms and the bedroom I put her in, it was real neat, it had French doors and it overlooked Lake Union. Because this was my thought, I thought at least she'd have something to look at if she was in bed and then she could also get around in the wheelchair on the hardwood floors, so it was perfect. It was like a lot of light, very pleasant so it worked out.'' Within the home itself, space must change to accommodate hospice. Most commonly,

the central area of the homeöthe living room or parloröis where the dying patient is placed. Although hospice philosophically is `low-tech', medical equipment is introduced into this space: electric hospital beds, wheelchairs, commodes, intravenous drug carriers, and equipment regulating medication disbursement. Towels are dyed red and are always on hand to clean and conceal blood and other bodily fluids. These are surely not the normal accoutrements of a standard living room or bedroom. They are clearly out of place. Thus they spatially code the room as a death scape, despite all efforts to the contrary. The division of labor in hospice policy also means that five different peopleömost of whom at least start out as strangersöare coming and going in the home. A hospice nurse went so far as to argue these visitations made the home very hospital like:

`̀ that's the part where I think it's similar to a hospital. You have a lot oföpatients would complain that they didn't want another volunteer to come in the house because they didn't want to see one more provider because there were already so many people they couldn't keep them all straight. Or you'd sit down as a nurse and they would think you were the social worker, or they'd ask the social worker to change the IV bagöso they would get you confused as to your role because to them you were all just strangers. So I think that is an interesting point you bring up around how canöboy, the extent of having the same people every dayöbut how can you gatekeep, and I guess that's part of what the primary caregiver is, is the gatekeeper to say who can come when.'' The home is culturally signified as private control over private space. If home is a

territorial form of control for the hospice patient, the stability upon which that power is presumed is not necessarily stable or fixed. The theoretical point that home is not completely one's turf is exemplified throughout this paradox. Whereas home is meant to territorialize control for the hospice patient, it may have to change for care to be provided well. Furthermore it changes precisely because of the work of care and its trappings.

3.3 The constitutive paradox A third series of paradoxical themes emerge around the ways that power relations are spatialized through the home as obligations to care. In hospice, both the family and the state powers are constituted spatially through home. On the one hand care is an emotional worköa labor of love. It is selfless work done by familyöusually womenö out of what is often characterized as a prepolitical, private, and essentially a natural obligation (Elshtain, 1983). This sense of natural, familial obligation was noted by discussions of what the loving and caring family members did. That theme emerged in discussions of how beautiful hospice was. A family-carer member described her brother's hospice situation:

842 M Brown

`̀ ... if you feel comfortable and integrated in your space where you feel you are powerful, and have some agency to make things the way they are, there's more likely a chance in a casual exchange of bathing or feeding or talking or just sitting with someone, that reminiscences happen, that forgiveness can happen, that closure of the relationshipöthose conversations that can't happen in a one hour hospital visit with all of the interruptionsöyou don't feel safe, you don't know who's going to be coming in, there's no privacy. So some of those really precious moments, and you talk to people, the surviving family members will often talk about how that was the best time they spent together. I was thinking back when my brother died, we hadöthere were jigsaw puzzles at the dining room table at his house and he had four children, and they would come in and visit with him for a while, and work on the puzzles and talk with each other. So he wasn't out interacting with them, he didn't have that kind of energy, but he would listen to their voices and so there's that integration. He had colon cancer... and he was diagnosed in May and died in August, so it was really a shortöso I think that opportunity to have those moments that you can't always have at the hospital, they're not as dramatic as the television or movies portrays, but they're very powerful for people for memo- ries afterwards. And you can say your dad's going to be gone in another eight hours, let's say, and if you need to say something, just make time to do that. So you could see them going over and sit by them, each one in turn in the last eight hours would do that. There's the fullness of the memory that he or she was here and that you did everything you could and that sort of pride at some level, or satisfaction that you made their last days what they needed. And the house has memories of all of that, which may be painful to people at first but later on is really quite positive.'' But this sense of home structuring family obligation was also sketched in more

stressful tones through the theme of guilt. A surprisingly common example of this sense of familial guilt took the form of food and feeding. Arguably the strongest synechdoche of family obligation and the home; the political ecology of food often took on quite mammoth proportions in hospice work. On the one hand, carers felt that continually feeding or offering favorite foods to the dying patient was a special and intimate way to show love through care work. On the other, dying bodies need less nutrition as the body shuts parts of itself down, or deals with particular illnesses and their final assaults on organs or systems. This mismatch between family obligation in the home and patient's biological needs takes on a special saliency of weight in the home when it is a family member doing the work of caring.

`̀ Oh, yes, the [food-guilt]öit happens a lot in the male ^ female, especially olderö but it just happens, where the only thing the family members feel they can do is make food that the patient likes to eatömilkshakes or treats, that the patient really likes to eat. And then as the patient gets ready to die, realizing that those body systems are shutting down, eating food, it just won't work for the bowels and they don't really feel like eating much and they're losing their taste. And the family member now is feeling helplessöwhat do I do? I can't comfort them. So they'll push, and all those things that used to just make me nutsöthey'll say oh, come on, I just made this special for you, you're not going to let all my good work go to waste, are you? And oh, how could you do this to meöyou're not eating your soup. It's like I just want to strangle the cook, but I really do understand all that. But it's like after you see it and hear it enough, it's like oh gosh. And so then you spend time with the family talking to them about how it isn't good for them to be eating and so they aren't going to die of starvation or die from thirst.''

Hospice and the spatial paradoxes of terminal care 843

If on the one hand hospice is a structuration of a familial love, obligation, and, allegedly private, prepolitical home relations; it is also structured by the polity's obligation to care. Hospice may be love, but it is also public welfare (see Gottschalk, 2000; Harrington, 1999; Harrington Mayer, 2000; Katz, 2001). In terms of revenues, Medicare made 72.4% of hospice payments, and an additional 5% came from Medicaid (US DoHHS, 2000). Private insurance and managed care corporations copy the Medicare policy for the sake of ease and consistency. Thus in addition to family dynamics we can also witness state power in the home. I develop this theme below by (a) noting the ways that hospice policy defined, limited, or excluded patients and the state's obligation to care; and (b) tracing the retrenchment of the welfare state generally through hospice.

As an example of how governmentality operates here, consider how the state structures the dying body. Hospice is only available to those patients who are certified by their doctor and the hospice medical director as having six months or less to live. Patients must sign a statement choosing hospice, rather than curative treatment and standard Medicare covered benefits, for that illness. The overall point here is that governmental structures shape the practice of death and care such that what Califano (1997) calls ``death management'' is evident. This scheduling also proves problematic because it insists that people formally and publicly admit they are, in fact, actively dying. This declaration means that all curative measures must stop. The benefit will not pay for any curative treatment, only palliative care. But as one hospice manager reminded me carefully, not all terminally ill people are ready to die. Such governmentality obviously has powerful effects of subject formation for the patient: people have to identify as abject. They must admit failure in conquering disease.

As a policy with rules and regulations, hospice is not always immediately available to everyone who wants it. People who typically fall outside the entitlement are those individuals who have no private health insurance, but are also (a) too young for Medicare and (b) are too `wealthy' to qualify for Medicaid. The typical example given was that of a self-employed tradesman such as a carpenter or contractor. Typically what must happen for these people is that they must `spend down' their assets to qualify for Medicaidöbut whether or not they have the time for this financial strategy depends on the progression of their terminal illness.

A further resonant theme of state power is that of `dying on schedule'. For all the freedom and autonomy hospice is meant to encourage, the hospice benefit must presume a certain schedule to dying in order for the benefit to be available to all eligible citizens equally and fairly. The hospice benefit is structured around a six-month time frame. Patients are not meant to have more than six months to live in order to qualify for hospice benefit. Should they exceed the six-month limit, they can reapply for up to two additional half-year terms. After that they must `re-up' for hospice in sixty- day increments. There is no limit on the number of terms, but the terms themselves create a schedule. Hospices are reimbursed on a per diem basis, along these six-month terms, and budgets are planned accordingly. If several patients do not die on schedule, a deficit can emerge that taxes the organization's ability to extend care to more patients.

Furthermore, hospice does not necessarily schedule well with all forms of dying. For example, cancer is typically related to home and hospice deaths (for example, Brown and Colton, 2001). Although the percentage of hospice patients dying of cancer decreased from 76% in 1992 to 51% in 2000, the majority of hospice patients still suffer from cancer (Hospice Association of America, 2002). Yet as a hospice administrator explained, part of that correlation is because the progression of the cancer works well in the hospice time frame:

844 M Brown

`̀ And [the high rate of cancer deaths at home] is partly driven by the federal government. There's other diagnoses that you die from and could die at home often, but you can't. Medicare has given really tough criteria for the diagnosis of congestive heart failure and chronic lung disease, [for instance]. They're terminal illnesses but they're harder to put in that 6 month box than cancer. [But other diseases] they're all over the map. You can have chronic obstructive pulmonary disease for 20 years and be dying by inches for 20 years and you're clearly not going to qualify for hospice.''

In other words, not all terminal illnesses suit the policy model well. Besides issues of governmentality and bureaucratic rationality, we can see the

power of welfare-state restructuring through home hospice in what Tronto (1994) calls `̀ privileged irresponsibility''. By this she draws our attention to the fact that, because care relations are power laden, those with resources can abdicate their social respon- sibility to care (through the welfare state). With the spatial shift to home death, terminal caregiving shifts from the public sphere of the state apparatus to the private sphere of the family (Harrington Mayer and Kesterke Storbakken, 2000). It shifts (relatively) from (mainly female) nurses to (usually female) primary caregivers. The bulk of the labor shifts from skilled paid labor in the welfare state to unpaid familial love in the home. Indeed, one of the persistent myths about hospice, according to interviewees, is that it provides 24-hour professional care, seven days a week. Families, according to interviewees, are often shocked to find out that they will be doing the bulk of carework: in terms of both time and effort. In a very real sense we are privatizing the responsibility and the labor of terminal care by replacing it into the home (Knijn, 2000). A former hospice administrator explained:

`̀ Well, in a way if you've got a willing family member, they do all the hands on, the bathing, the turning if they want to be turned, the feeding and the cleaning up if there's incontinence, but the injections. But nowadays you don't have to give injections as much because there's so many other ways, to administer pain medica- tion now. And so because that kind of stuff is high touch, low tech, you can train someone to do it. They wouldn't know the dosing, but you could teach them to watch for signs of too much or too little. But again then it's that idea of well, anybody can do it.''

This theme is captured by research on the enormous cost savings of hospice. Kidder (1992), for example, reports that in the late 1980s Medicare saved $1.26 for each $1.00 it reimbursed on hospice care. According to the National Hospice and Palliative Care Organization (NHPCO, no date), that figure jumped to $1.50 in 1995.

My point here is that home is a paradoxical space where different exercises of power constitute the normalcy of care within it. The federal welfare state is imbricated with family love and obligation. These two forms of care mutually constitute the home. That mutual constitution is sometimes hard to see because home is coded as private, prepolitical space. Clearly caring complicates this simple political geography, and does it in quite complex ways. The peculiar hybridity of obligation to care in hospice, which is difficult to see in the home, begs the question of whether or not Humphrey's (1967, page 182) defense of the welfare stateö`̀... to do for ourselves together what we each cannot do alone''öis being heard.

3.4 The relational paradox Liberal theory has always had a hard time with dying individuals. They are similar to children, the mentally ill, and animals; they are bodies that refuse to conform to ideals of democratic-theory aims. So they undermine liberal institutions. Dying denotes an explicit heteronomy (the absence of autonomy). The dying are often not

Hospice and the spatial paradoxes of terminal care 845

the rational, autonomous, free individuals who are pursuing the aims of life, liberty, and property. The fourth spatial paradox of home deals explicitly with the issue of patient autonomy and dependency. Hospice is often expressed as a clear example of medical treatment that puts the patient and family at the center of care and decisionmaking. As hinted at in various ways above, hospice is seen as encouraging and maintaining patient autonomy so that the patient can have the sort of death she or he wishes for, under conditions of his or her own choosing. As demonstrated in the section above, hospice patients are encouraged toöand doöexercise their auton- omy through territoriality, and clearly have a wide range of choices over the extent and direction of their own care. But, as Messler (1995) points out, this portrayal can mask relations of dependency and heteronomy structured in the home. For instance, home can often be defined for them, and is a result of circumstances beyond their control. Home must also be a place where they are not alone, and in this way there is a built-in dependency to the geography of hospice.

The processes of nondecisionmaking and mobilization of bias are also at work here. Amongst the interviewees there was a recognition of the controlled set of choices for hospice patients. Most notably, perhaps, is the legal fact that hospice agencies cannot engage in assisted suicide or voluntary euthanasia; for example, various means were employed by hospice agencies to control overmedication. But the patient's heteronomy and dependency also worked in more subtle ways too. A former hospice nurse, one of the early advocates for a hospice in Seattle, relayed the following illustrative anecdote:

`̀ Let me give you a very practical example, very practical. [A hospice team] comes into a home and says, `Let's get a hospital bed in here'. And it's suggested to the family and the family go, well, these people know what they're saying, we probably should get a hospital bed in here. Well, let's just take that very tiny example about how different is that than sleeping in your own bed and if your spouse is still able to be with you even part of the time, let's just assume it's a spousal relationship, might you not prefer that? But how did that choice get made? Was it really a choice or was it a recommendation or was it we think this is best and they go okay. And things are moving very quickly at that time, there's a lot to be done, it's like, okay.''

Thus although hospice clearly exemplifies patient freedom and autonomy, it cannot escape the inherent dependency of a dying individual: who is a body that cannot care for itself.

I would also argue that the dying patient's autonomy pivots on the very ironic surrender of autonomy of the (usually female) family carer (who themselves must have no less of a claim on that virtue). Whereas I will take up the gendered division of labor in hospice work in future work, here I want to press the point that for family caregivers there is a clear relation of dependency here that is at odds with the whole project of autonomy in the liberal-democratic state (Fraser and Gordon, 1997). This point is exemplified in a number of ways. Most poignantly by the paradoxical emotions carers face as they hospice someone they love. A very eloquent hospice social worker explained the paradox thus:

`̀ So it's things like holding onto the patient while letting go, increasing attention to the patient during the illness while starting to detach from the patient in terms of his or her existence in the future, remaining involved with the patient while sepa- rating from the patient, planning for life after death while not wanting to betray the patient by considering life in his or her absence, communicating feelings to the patient while not wanting to make the patient feel guilty for dying or bound to this world when the patient needs to let go, balancing support for the patient's increased dependency while supporting the patient's continued need for autonomy,

846 M Brown

redistributing family roles and responsibilities while not wanting to do anything that would call attention to or cause more losses for the patient, taking care of the patient's needs while taking care of one's owns needs, experiencing the full intensity of the feelings involved in anticipatory grief while not trying to become over- whelmed, focusing on the patient as a living person while remembering that the patient is dying.''

The sheer number, complexity, and ubiquity of tasks family carers are asked to perform diminish caregiver's autonomy. A clear refrain of caregivers being overwhelmed because of the dying patient's dependency came through loud and clear. Here is just one example:

`̀ Well, I can think of different experiences like one person who the wife had a lot of pain and we were trying to manage it with oral medication and then when we talked with him about we would have to go to a subcutaneous pump infusion, he just freaked out. He said, I can't deal with that. I hate needles, I can't look at needles, there's no way I can have any part of that. And I tried to explain to him that he really wouldn't have to do that, all he had to do was learn how the pump worked and if there were any problems to call the nurse on call. She would get out there and help him. It just freaked him out too much, he couldn't deal with it, so she couldn't stay at home.

And I think that's the fascinating part of caregiver stress, this kind of ten- sion and struggle that they're going through while doing all the practical stuff. Well, somebody is dying in the living room and it's not just somebody. I keep saying somebody but it's their mom, their dad, their husband, their child, their sibling, their neighbor, college friend, whoever that relationship is.''

A bereavement coordinator shared the following anecdote as a way of helping me not only understand the sense of being overwhelmed that a family feels, but also how the family's dependency on professional staff can cascade into stress for hospice workers as well:

`̀ The charge nurse [was on the phone with a caregiver], you know, a zillion calls a day, this is happening, this is happening, what can you do, and she has to slowly talk them through [a specific care task]. Their nurse may not be there, the social worker is the one on the phone, so she talks them through this. And I was waiting for her to finish a call, and she hung up and she goes, `Oh, for God's sakes, these people!' Which was exactly how I felt a million times and every hospice worker feels this way as in,`Can't they get it?!' kind of frustration. And I turned to her and said,`Yeah, you know, they act like someone's dying in their living room, don't they?' And she just went, `Oh, I'm so sorry, thank you.' Because my God, somebody IS dying in their living room! And then I get a flat tire and it like throws me into a panicöbig deal!

But they are basic people and suddenly we are asking them to give supposi- tories and figure out to how get more calories in and change linen in a bed with the person in it, medication management, and holy cow! So her phrase is we're going in and asking people to prepare a Thanksgiving dinner and up to this point they've only made Jell-O. And that's really important for us to keep remembering because we do get skewed as workers.'' An especially spatial form that caregiver's heteronomy took was signified by the

role of hospice volunteers. In the division of hospice labor, they are not charged with caregiving tasks per se; rather their role is to provide respite for the family caregiver. Typically this means staying with the patient, thus giving the family member a chance to leave the house and take a break from the relentless labor of terminal care. The respite lasts usually one to two hours. It is notable here that the way in which caregiver autonomy is structured in hospice is through the spatial paradox of physically leaving

Hospice and the spatial paradoxes of terminal care 847

the home to gain freedom! Once again for caregivers, home is not the place of liberal autonomy. When the caregiver returns, the work of care is still there waiting for their selflessness.

Situated as it is in the home, hospice care on the one hand extends patient autonomy, but it can never do so completely. Here we must acknowledge the relational and power- laden nature of care. The patient's freedom and capacity to stand alone are always limited and curtailed, precisely because caring is always part of gendered and classed social relations. And networks-of-power web those social relations. Their own autonomy is only gained at the ironic expense of dependency. This seems to be an especially difficult theme to see given dependency is such a dirty word in the present context of neoliberal state restructuring (Fraser and Gordon, 1997). Furthermore, the heteronomy of the family careröusually womenöis what enables the patient's autonomy in the first place.

4 Conclusion Were the hired man to appear at the farmers' door today, the couple might ponder the same questions, but their saliency would be different, because death at home is not simply `natural' or uncomplicated. Parts of its meaningfulness change across time and space, whereas others stay constant. I have offered an interpretive political geography of home hospice based on critical theories of care. Overall that geography is best described as a paradoxical space: one that can be described as one way, as well as its opposite. For hospice care, home is a pharmakonöit is ironic. The paradoxes take many forms and are not mutually exclusive. There can be a normative paradox, where home is a good and bad place to die. There can be a territorial paradox, where the spatial exercise of power by the patient can change the nature of that place itself. There is a constitutive paradox, where caring relations, their efficacy, and their legitimacy are structured governmentally and through family. Reflecting care's inherent relationality, patients' control and autonomy often rests on their own, or others' surrender of autonomy or increased dependency. These may be relations of love and support in the family, but they are also part of welfare-state restructuring with all the attendant governmentality incumbent in state power. The overall point of tracing out these complexities is not to reject or trash hospice care. It is rather to offer a careful and considered interpretation of it. This interpretation draws on, and speaks to, current issues within both political theory and geography.

The stress on spatial paradox within this paper should not be taken to mean that the paradoxical nature of the home itself is new. As geographers and nongeographers have long noted, home has always been an especially dense point of meaning and contradiction (Busch, 2000; Mack, 1993; Pratt and Dowling, 1993), `̀ a product of ever-shifting social geography of social relations past and present,'' in Massey's (1993, page 172) potent words. My point here is that home hospice is a new and especially acute weight on that density. As welfare continues to be privatized in multiple ways, and changing demographics mean the burdens of care fall especially hard on the so-called `sandwich generation' (between obligations to their parents and their children), it becomes both theoretically and empirically important to move beyond simple constructions of domestic space.

There are several implications for current literature that intellectually situate this project. For political geography, I show the value of continuing the project of problem- atizing the public ^ private divide in liberal theory, especially when it is rethought spatially and in a context of welfare-state restructuring. To open up new forms of politics and the political means we must also open up the locations in which we search for them. The recent turn towards decentering the state in political geography signals the need to appreciate not just the diffuseness of state powers, but also other forms

848 M Brown

of power/knowledge that are no less political or geographic. These `̀ intra-household politics'' (Taylor and Flint, 2000, page 349) might seemingly problematize claims that political geography's fundamental unit of analysis is the nation-state and its public sphere (compare Staeheli, 2001). Thus political geographers must continue work that explores politics within the home, and between state and home.

For health geography there is the need to increase its engagement with political theory. The so-called postmedical or health geography has been nourished by social and cultural theory to be sure (for example, Gesler and Kearns, 2001), but questions of governance and politics remain undertheorized, though this seems to be shifting (for example, Gleeson and Kearns, 2001). This need to draw on, and speak to, political theory seems especially apposite for health geography given the retrenchment of the welfare state, but also in light of the longstanding interest of health and medical geography in health inequalities (see Gatrell, 2001).

As a spatial strategy of care, home hospice seems to reproduce many of the most problematic and vexing problems within liberalism. Home is a space of control, auton- omy, and freedom for someöbut not for others. The naturalism of familial obligation is uncritically favored over the communal obligation inherent in state welfare policy. In the will to limit and retrench state power to preserve freedom, women must bear the burden by surrendering some of their own freedom and autonomy. For as much as hospice is a progressive antidote to the disempowering ways American society has dealt with death, it also reproduces its own share of problems. And moreover, it does so in particularly spatial ways.

For political theory, this work suggests two points of attention. Foremost, it suggests that death and dying, as well as caregiving around it, is an important point of focus, as death has political meanings in multiple and changing ways (Harris, 1971). Although the problems death presents for political theory are often treated anecdotally or ad hoc, there are precious few extended considerations of death itself in contempo- rary political theory (Harris, 1971; Seery, 1996). Just as important, there is the pressing need to understand the spatialization of abstract political concepts, and the spatiality of social processes generally (Brown, 1997). Rather than seeing any sort of empirical inquiry as an empiricist move, an encounter with critical geography may liberate thinking, as White's (2000) careful work has demonstrated.

For hospice workers and advocates, I feel this paper has important, albeit modest implications, because much of the empirical material was drawn from their lifeworld. Certainly it is important to recognize and appreciate these broader social forces and processes at work, even if they are beyond your will to control or abate. Here the arguments by political theorists about the need to understand care as a social relation are especially salient. Although hospice workers must always assess and evaluate care, and have their eye on the explicit tangible aim of improving patient care and enabling `the good death', we must also interpret care and the relations that enable and con- strain it. We must see the larger social narratives at work in everyday events and practices. Hospice may well be a terrific way to improve terminal-patient care, but it also instanciates structures of patriarchy and welfare-state privatization. Within ordi- nary tasks such as controlling pain or providing comfort these are larger questions of rights and obligations and their `proper' forms. In Milligan's (2000) words, `̀ Who bears the burden of care?''. What is the appropriate relation between public and private spheres in a liberal democratic society? These questions are being asked and answered as patients are discharged home, as family members bathe and change them. They are a rather more hidden side of welfare reform alongside questions of employment (for example, Peck, 2001). As the same time, I would offer a plea to recognize the thoroughly spatial ways care and hospice are structured and practiced.

Hospice and the spatial paradoxes of terminal care 849

Acknowledgements. I would like to thank the editor and the anonymous referees for the constructive criticisms. This paper has also benefited from helpful comments and suggestions from Kim England, Chris Fowler, Jamie Goodwin-White, Steve Herbert, Jonathan Mayer, Dick Morrill, Claire Rasmussen, Matt Sparke, and Lynn Staeheli.

References Brim O, Friedman H, Levine S, Scotch N, 1970 The Dying Patient (Russell Sage Foundation,

New York) Brown M,1997 RePlacing Citizenship: AIDSActivism and Radical Democracy (Guilford, NewYork) Brown M, Colton T, 2001 `̀ Dying epistemologies: an analysis of home death and its critique'',

Environment and Planning A 33 799 ^ 821 Busch A, 2000 Geography of Home: Writings on Where We Live (Princeton University Press,

Princeton, NJ) Butler J, Scott J (Eds), 1992 Feminists Theorize the Political (Routledge, London) Califano J,1997,`̀ Death management'', inThe Grim Reader:Writings of Death, Dying and Living On

Eds M Spiegel, R Tristman (Anchor Books, New York) pp 370 ^ 376 Dowling R, Pratt G, 1993, `̀ Home truths'' Urban Geography 14 464 ^ 475 Duda D, 1987 Coming Home: A Guide to Dying at Home with Dignity (Aurora Press, New York) Dyck I, 1998, `̀ Women with disabilities and everyday geographies: home space and the contested

body'', in Putting Health into Place Eds R Kearns,W Gesler (Syracuse University Press, Syracuse, NY) pp 102 ^ 119

Field M, Cassell C, 1997 Approaching Death (American Academic Press,Washington, DC) Fraser N, Gordon L, 1997, `̀A geneology of `dependency': tracing a keyword of the US welfare

state'', in Justice Interruptus: Critical Reflections on the Post-socialist Condition Ed. N Fraser (Routledge, London) pp 121 ^ 149

Gatrell A, 2001 Geographies of Health: An Introduction (Blackwell, Oxford) Gesler W, Kearns R, 2001 Culture, Health, and Place (Routledge, London) Gleeson B, Kearns R, 2001, `̀ Remoralising landscapes of care'' Environment and Planning D:

Society and Space 19 61 ^ 80 Gottschalk M, 2000 The Shadow Welfare State: Labor, Business, and the Politics of Health Care

in the United States (Cornell University Press, Ithaca, NY) Harrington M, 1999 Care and Equality: Inventing a New Family Politics (Knopf, New York) Harrington Mayer M (Ed.), 2000 Care Work: Gender, Labor, and the Welfare State (Routledge,

London) Harrington Mayer M, Kesterke Storbakken M, 2000,`̀ Shifting the burden back to families'', in Care

Work: Gender, Labor, and the Welfare State Ed. M Harrington Mayer (Routledge, London) pp 217 ^ 228

Harris K, 1971, ``The political meaning of death'' Omega 2 227 ^ 239 Haupt B, 1998, `̀ Characteristics of hospice care users: data from the 1996 National Home and

Hospice Care Survey'', Advance Data #299, National Home and Hospice Care Survey, National Center for Health Statistics, Hyattsville, MD

Hospice Association of America, 2002,`̀ Hospice facts and statistics'', unpublished report; available from Hospice Association of America, 228 Seventh Street SE,Washington, DC 20003

Hoyer T, 1998, `À history of the medicare hospice benefit'' Hospice Journal 13 61 ^ 69 Humphrey H, 1967 ``The welfare stateöa state of general welfare'', in The Welfare State

Ed. C Schottland (Harper Torchbooks, New York) pp 179 ^ 183 Katz M, 2001 The Price of Citizenship: Redefining the American Welfare State (Metropolitan

Books, New York) Kidder D, 1992, `̀ The effects of hospice coverage on Medicare expenditures'' Health Services

Research 117 pp 599 ^ 606 Knijn T, 2000, `̀ Marketization and the struggling logics of (home) care in the Netherlands'', in Care

Work: Gender, Labor, and the Welfare State Ed. M Harrington Mayer (Routledge, London) pp 167 ^ 187

Mack A (Ed.), 1993 Home: A Place in the World (New York University Press, New York) Massey D, 1993 Space, Place, and Gender (University of Minnesota Press, Minneapolis, MN) Messler M, 1995, `̀ The philosophy and practice of patient control in hospice: the dynamics of

autonomy versus paternalism'' Omega 30 173 ^ 189 Milligan C, 2000, `̀ Bearing the burden: toward a restructured geography of caring''Area 32

49 ^ 58

850 M Brown

NAHC, 2001,`̀ Basic statistics about hospice'', National Association of Hospice Care,1700 Diagonal Road, Suite 300, Alexandria,VA 22314, http://www.nahc.org/consumer/hpcstats.html (July)

NHPCO (no date) ``Facts and figures about hospice'', National Hospice and Palliative Care Organization, http://www.nhpco.org

Palmer G, 1993 Death: The Trip of a Lifetime (HarperCollins, New York) Pateman C, 1989 The Disorder of Women: Democracy, Feminism, and Political Theory (Stanford

University Press, Stanford, CA) Peavy F, 1996 Heart Politics (Black Rose Books, Montreal) Peck J, 2001 Workfare States (Guilford, New York) Seale C, 1998 Constructing Death:The Sociology of Dying and Bereavement (Cambridge University

Press, Cambridge) Seery J, 1996 Political Theory for Mortals (Cornell University Press, Ithaca, NY) Staeheli L, 1994, `̀ Restructuring citizenship in Pueblo, Colorado'' Environment and Planning A

26 849 ^ 871 Staeheli L, 1996, `̀ Publicity, privacy, and women's political action'' Environment and Planning D:

Society and Space 14 601 ^ 619 Staeheli L, 2001,`̀ Of possibilities, probabilities, and political geography'' Space and Polity 5177 ^ 189 Taylor P, Flint C, 2000 Political Geography 4th edition (Prentice-Hall, London) Tronto J, 1994 Moral Boundaries: A Political Argument for an Ethic of Care (Routledge, London) US DoHHS, 2000, ``National Home and Hospice Care Survey'', CD-Rom Series 13, (27 July),

National Center for Health Statistics, United States Department of Health and Human Services; available from National Center for Health Statistics, Division of Data Services, 6525 Hyattsville Road, Hyattsville, MD 20782

Webb M, 1997 The Good Death: The New American Search to Reshape the End of Life (Bantam, New York)

White J, 2000 Democracy, Justice, and the Welfare State: Reconstructing Public Care (Penn State Press, College Park, PA)

Young I, 1997, `̀ House and home: feminist variations on a theme'', in Intersecting Voices: Dilemmas of Gender, Political Philosophy, and Policy (Princeton University Press, Princeton, NJ) pp 134 ^ 164

Young I,2000,`̀ Political theory: an overview'', in ANew HandbookofPoliticalScience Eds RGoodin, H Klingemann (Oxford University Press, Oxford) pp 479 ^ 502

Hospice and the spatial paradoxes of terminal care 851