ARTICLE IN PRESS

Social Science & Medicine 58 (2004) 2431–2444

Place of death: preferences among cancer patients

and their carers
C. Thomas*, S.M. Morris, D. Clark
Institute for Health Research, Lancaster University, Alexandra Square, Lancaster LA1 4YI, UK

Abstract

The place of death of cancer patients has become an important theme in UK cancer and palliative care policy. This
paper examines the place of death preferences of 41 terminally ill cancer patients and 18 of their informal carers, living
in the Morecambe Bay area of north-west England. We interviewed cancer patients referred to the research team by 13
specialist palliative care professionals; patients had an estimated 3 months of life remaining. The study design involved
an in-depth qualitative interview with each patient soon after referral to the study, followed by an interview some 4
weeks later and subsequent tracking interviews by telephone at 2–4 week intervals until death occurred. Interviews were
also conducted with main coresident carers soon after patient referral to the study and again in the post-bereavement
period. Thirteen factors were identified as shaping the place of death preference of patients and carers. These are
organised into four thematic domains: the informal care resource, management of the body, experience of services, and
existential perspectives. In documenting these factors, this paper adds significantly to current knowledge on the factors
that shape place of death preference, a field of enquiry acknowledged to be underdeveloped (J. Palliative Med. 3 (2000)
287). More importantly, it uncovers some of the reasons that underpin these preferences. Our research revealed a much
stronger preference for deaths in a hospice than had been anticipated, leading us to take a qualified stance on the
current policy drive in favour of home deaths by those charged with delivering UK cancer and palliative care services.
r 2003 Elsevier Ltd. All rights reserved.

Keywords: Cancer; Palliative care; Place of death; Patient preferences; United Kingdom

Introduction terminal illness, we would be likely to pay closer attention

Where would you prefer to die—at home, in hospital,
in a hospice, or elsewhere? Posing such a question
presents us all with difficulties, however aware we might
be that death is close at hand. Cultural mores discourage
such a direct question, and the sequestration of death in
contemporary society such that it is routinely hidden
from view and medicalised (Giddens, 1991), means that
we are ill prepared to consider the when and where of
dying. If we have the luxury of good health and the
prospect of many years left to live, the probability is
that, if pressed, we would express a preference to die at
home, attended to by those who love us (Higginson &
Sen-Gupta, 2000). In the face of serious and especially
*Corresponding author. Tel.: +44-1524-594092; fax: +44-
1524-592475.
E-mail address: c.thomas@lancaster.ac.uk (C. Thomas).
to where we would like to end our days, but also to be
more cognisant of the circumstances that might play a role
in determining the outcome. In that context, would dying
at home as opposed to dying in hospital or in a hospice
continue to be our preference? If not, why not?
This paper seeks to address such matters by reporting
the findings of a recent study undertaken in the
Morecambe Bay (MB) area in north-west England on
the factors that shaped preference for place of death
among 41 terminally ill cancer patients and 18 of their
carers. In this research, matters of preference for place
of death were explored sensitively through interviews
with patients and their carers, involving discussion of
the illness, care arrangements, and preparations for the
future. All patients had an ‘open awareness’ (Seale,
Addington-Hall, & McCarthy, 1997) of their terminally
ill state. This qualitative engagement with people
struggling with end of life concerns has brought to light

0277-9536/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.socscimed.2003.09.005
 ARTICLE IN PRESS
2432 C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
a range of factors that shape place of death preference.
In documenting these factors, this paper adds signifi-
cantly to understanding about not just the ‘what’ but
also the ‘why’ of preferences—something missing from
the literature on place of death.
The study was undertaken at a time when greater
impetus was given to a long-term UK health-care policy
goal of increasing the proportion of cancer deaths that
could occur at home (Clark & Seymour, 1999). This policy
is presented as a matter of meeting patient choice rather
than as anything to do with other considerations such as
reducing service costs. For example, the NHS Cancer Plan
(Department of Health, 2000), while focusing on cancer
treatment services, acknowledged that although most
patients with advanced cancer would prefer to die at
home only about a quarter are able to do so. The plan
pledged an extra d50 million investment a year by 2004 in
hospices and specialist palliative care, with an explicit
emphasis on improving the capacity of services to support
end of life care in the community. In 2001, the Department
of Health announced the investment of an extra d2 million
to educate and support district and community nurses in
providing care to enable more patients with advanced
cancer to remain at home, if that was their wish (DoH,
2001). Subsequent official announcements on the funding
of specialist palliative care services have underlined this
intention. These developments have occurred in the
context of a broader health service policy commitment
to listening to the views of service users, with repeated
political promises to be responsive to the expressed needs
of patients and their families (Department of Health, 1997,
1998; Small & Rhodes, 2000).
Against this background, the idea that ‘home is the
best place to die’ has taken a firm hold in the minds of
many service providers and practitioners (Seale et al.,
1997). From this perspective it follows that services
should be reconfigured to boost the provision of
specialist palliative care in community settings. Our
research project grew out of these concerns among
service commissioners and providers in the MB locality
in the late 1990s. A service-instigated review of place of
death patterns for cancer patients in the geographical
patch covered by the former Morecambe Bay Health
Authority had revealed that compared with averages for
England and Wales the locality had a high proportion of
deaths in hospice, a low proportion of deaths in
hospital, and a similarly low proportion of deaths at
home.1 The question was posed: should there be a shift
1 small-area basis within the district (Gatrell et al., 2003).
It should be noted that in the MB area, and thus in this
paper, the term ‘hospice’ refers to a physical building in which
people are cared for and may die, rather than a philosophy of
care than can be delivered in home and other settings. The MB
area has two hospices of contrasting sizes, but in combination
they represent a provision of specialist palliative care beds for
the district that is above the per capita national average.
of specialist palliative care resources from the hospice
sector to the community in the light of national research
evidence that most cancer patients would prefer to die at
home? To the credit of local health services, it was
decided not to take precipitate action but to await the
findings of our NHS funded research project on patient
and carer preferences for place of death (Thomas,
Morris, & Gatrell, 2003).
Where do cancer patients die, and what are their place of
death preferences?
Much more is known about the factors that correlate
with where cancer patients actually die than about the
factors associated with their preference for place of
death. The literature on predictors of the location of
cancer deaths in the UK and other advanced industrial
societies, and indeed of deaths from other causes, is now
quite extensive (Brown & Colton, 2001; Gatrell et al.,
2003; Grande, Addington-Hall, & Todd, 1998; Higgin-
son, Astin, & Dolan, 1998; Higginson, Jarman, Astin, &
Dolan, 1999; Hunt, 1997; Karlsen & Addington-Hall,
1998).
In the UK and USA, this interest in predictors of
place of death has emerged in the context of long-term
trends away from the home towards institutionally
located deaths. Between 1967 and 1987, the proportion
of cancer deaths in hospitals in England increased from
45% to 50%, and from 5% to 18% in hospices and
other institutions (Cartwright, 1991). The increase in
hospice-based deaths is testimony to the rapid expansion
and diversification of hospice services in the UK in the
years following the opening of the landmark St.
Christopher’s Hospice in 1967 (Clark & Seymour,
1999). In 2000, the proportion of cancer deaths in
hospital in the UK stood at 55.5%, with 16.5%
occurring in hospice, 23% at home, and 5% in other
settings, principally nursing and residential homes
(Ellershaw & Ward, 2003). Within the UK, large
variations in place of death are found between and
within regions. For example, in north-west England
(1990–1994), the proportion of deaths occurring at home
ranged from 33% in South Lancashire to 22% in
Morecambe Bay—our study area (Hospice Information
Service data; in Higginson, 1999). In Morecambe Bay,
of 6900 cancer deaths between 1993 and 2000, 35%
occurred in hospital, 33% in hospice, and 22% at home,
but there were marked variations in this pattern on a
The fact that when compared with national averages the
MB area has a higher and lower proportion of cancer
deaths in hospice and hospital, respectively, is a
reflection of the relatively high number of hospice beds
per capita in the district. It seems that, given the
similarity between the proportion of deaths at home in
 ARTICLE IN PRESS
C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
MB (22%) and in the UK as a whole (23%), the hospice
sector in the MB area redirects patients from hospital to
hospice as a place of death, rather than from home to
hospice (Thomas et al., 2003).
Factors identified as predicting place of death for
cancer patients include: having informal carer support,
the health status and emotional capacity of the main
carer, the availability and use of specialist community-
based palliative care services; the need for specialist
symptom-control, tumour type, distance to services,
gender, age, socio-economic status, attitude to dying,
and the strength and visibility of the patient and carer
preference for a particular place of death (Addington-
Hall & McCarthy, 1995; Cantwell et al., 2000; Grande
et al., 1998; Herd, 1990; Higginson et al., 1999; Karlsen
& Addington-Hall, 1998). In our own analysis in the MB
area, we found similar predictors of place of death
(although we had no information relating to carers),
with service proximity effects playing an important role
(Gatrell et al., 2003).
In reviewing the literature on what they call
the ‘microlocation of death’, Brown and Colton (2001,
p. 801) note in relation to deaths at home that:
The presence of an informal caregiver, and the
[patient] being male, younger, and of a higher
socio-economic class are the nonclinical factors that
correlate positively with home deathy The associa-
tion between home death and informal caregiving is
the most consistent theme in the literature.
Clinical factors associated with deaths at home are:
tumour type—for example, respiratory, gastrointestinal,
and genitourinary cancers; the availability of palliative
care beds in the patient’s locality; a low requirement for
symptom-control measures such as equipment, technol-
ogy, or intricate drug regimes; and length of survival
time from diagnosis (Brown & Colton, 2001; Gatrell
et al., 2003).
Knowledge of cancer patient preference for place of
death is scantier. Higginson and Sen-Gupta (2000)
conducted a systematic review of the available literature,
identifying 18 studies, half undertaken in the UK. These
included five surveys of the general public, with the
remaining 13 studies presenting data on the views of
patients and carers dealing with advanced cancer.
Although Higginson and Sen-Gupta indicate that half
of the studies were flawed in design or reporting, they
conclude that home is the location of preference for end
of life care and death for over 50% of patients, with
inpatient hospices the second most preferred site. The
study they regard most highly, the one by Townsend,
Frank, and Fermont (1990) in the UK, found that of the
59 patients (from a sample of 84) with terminal cancer
who stated a preference for place of death in their then
existing circumstances, 34 (58%) chose home, 12 (20%)
2433
hospice, and 12 (20%) hospital (with 2% for ‘other’). It
was reported that preference for home death decreased
significantly over time, from 58% to 49%, with slight
increases in preference for hospital (20–24%) and
hospice (20–24%) (Higginson & Sen-Gupta, 2000,
p. 295). Hinton (1994) similarly found that the accept-
ability of a death in hospital rises as time progresses for
those with terminal cancer, particularly as problems
with self-care increase and relatives’ fatigue worsens. In
a UK study of 160 terminally ill cancer patients,
Dunlop, Davies, and Hockley (1989) found that of the
90 who could express a preference, 53% chose home,
with 29% preferring hospice, 14% hospital, and 3% a
nursing home.
What is known of the factors that shape place of
death preference? Thus far, few factors have been
identified and they are rather poorly understood:
symptom management, change in disease status, gender,
having a religious faith, and previous personal experi-
ence of death and dying (Butow, Maclean, Tattersall, &
Boyer, 1997; Higginson & Sen-Gupta, 2000). Higginson
and Sen-Gupta (2000, p. 299) conclude that ‘No
consistent conceptualisation of the factors that deter-
mine preferences for place of terminal care of patients
with cancer emerged from the studies reviewed and this
should be the focus for future work’. Further (Higginson
& Sen-Gupta, 2000, p. 299):
A preference for home care or death may be an
empowered expression of wishes or an aversion
from the perceived disadvantages of hospital care.
A preference for hospital or hospice care may
indicate a resigned acceptance of the inevitability of
inpatient care, a desire to save relatives and close
friends the burden of caring at home, a belief in
better care being provided, or a refusal to admit that
a cure is not possible. Many other interpretations are
possibley .
This paper offers some enlightenment on these
matters.
Methods
Study design, patient referrals, and samples
This 2-year (2000–2002) longitudinal observational
study involved interviews with terminally ill cancer
patients who were estimated by clinicians to have up to 3
months of life remaining. The study design also involved
a spatial statistical analysis of 6900 cancer deaths in the
MB area in an 8-year period (1993–2000), and interviews
with 16 health professionals involved in the delivery of
palliative care services in the MB area; but these data are
presented in detail elsewhere (Gatrell et al., 2003;
Thomas et al., 2003). The plan was to conduct an initial
 ARTICLE IN PRESS
2434 C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
patient interview soon after their referral to the project,
to hold a follow-up interview some 4 weeks later, then to
undertake tracking interviews by telephone at 2–4 week
intervals until death occurred. At the first interview visit,
the presence or absence of a coresident carer was to be
established, and an agreement for interview sought with
a coresident carer. Once a relationship with the carer
had been established, a post-bereavement interview
request would be sensitively made at an appropriate
moment.
After obtaining the approval of relevant NHS
research ethics committees, patients were referred to
the project team by 13 palliative care professionals in the
MB area. The majority of referrals (71%) were made by
seven specialist palliative care nurses. These patients
gave informed consent for the referral to occur after
receiving a project information sheet and discussing
potential involvement with the referring professional. Of
patients approached by professionals, 87% ðn ¼ 69Þ
agreed to participate. Once a referral was made, the
researcher made contact with the patient’s household by
telephone to confirm willingness to participate and to
establish whether they were well enough to be visited for
a first interview.
Of the 69 patients contacted by the researcher, three
refused to participate, a further 24 had become too ill to
embark on the first interview, and one declined because
his cancer had gone into unexpected remission. Conse-
quently, 41 patients commenced the interview series.
There was a high rate of sample attrition due to
worsening illness or early death, with only 24 patients
(59%) able to take part in a second interview and 17
(41%) able to complete the interview series. Overall, 65
face-to-face patient interviews and 38 telephone tracking
interviews were conducted. Eighteen coresident carers
agreed to be interviewed, a response rate of 72%; of
these, nine were interviewed for a second time at around
12 weeks post-bereavement.
This high rate of sample attrition is testimony to the
uncertainties involved in undertaking research with
dying patients, and the difficulties that clinicians have
in accurately estimating the illness trajectory and
proximity to death—an inherent problem in palliative
care research, certainly not unique to this study (Vigano,
Dorgan, Buckingham, Bruera, & Suarez-Almazor, 2000;
Christakis & Lamont, 2000).
Patient and carer interviews
All interviews were conversational in style, and
loosely structured around a range of pre-defined themes:
the cancer journey, quality of life, levels of pain, and
other symptoms, as well as feelings about current care
arrangements, and plans and desires for care in the
future. Preference for place of death was elicited through
sensitive questioning on a range of related issues, and
could be addressed more or less directly depending on
the ease displayed by patients and carers in discussing
such matters. The intention was to conduct carer
interviews separately where possible, but we complied
with respondents’ wishes where joint interviews were
preferred. Face-to-face interviews with patients and
carers were tape-recorded and fully transcribed, with
respondent consent, and were usually around 40 min in
length; all but two patient interviews were conducted in
home environments. Most interviews were conducted by
the project researcher (SMM); four were conducted,
under supervision, by three specialist palliative care
nurses who were part of the research team.
Analysis
All face-to-face interviews were fully transcribed, with
the data entered into the Atlas.ti analysis programme
(Scientific Software Development, 1997) for coding for
content and topic themes as well as key demographic
and clinical variables. The identification of analytical
themes was initially undertaken through close reading
and rereading of all transcripts by SMM, and of a
selection of transcripts by CT, in order to support the
validity of interpretation. This process was subsequently
developed through an examination of selected text at the
intersection between topic categories and sample char-
acteristics. Attention was paid to both typical and
atypical cases. The 13 factors influencing patient and
carer preference for place of death reported here
emerged through this close engagement with the inter-
view data. These factors were organised into the four
analytical domains reported after careful reflection by
CT. The data analysis thus approximated a grounded
theory approach (Glaser & Strauss, 1967), where topic
categories and analytical themes were derived from and
constructed through a systematic exploration of the
data.
Generalisation
Our heavy reliance on community specialist palliative
care nurses for the recruitment of patients to the study
meant that there was an under-representation of patients
who (i) had no contact with the specialist palliative care
services, and (ii) had tumours like those of the lymphatic
system that require close contact with hospital services
and which might be associated with a hospital place of
death preference. A very high proportion of patients in
the sample (73%) had some contact with one of the two
hospices in MB—whether for day-care and/or inpatient
care. These biases had a bearing on the pattern of
preference for place of death uncovered in the study.
However, since the main purpose in this paper is to
consider the qualitatively elicited factors associated with
place of death preference and not the numerical pattern
of preferences per se, the findings do have much wider
relevance and resonance, and thus have qualities of
 ARTICLE IN PRESS
C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444 2435

generalisability consonant with a qualitative research Table 2

paradigm (Bowling, 1997). Carer sample characteristics

Key characteristics N

Findings Gender
Male 5
Sample characteristics and place of death preferences Female 13

Age
Tables 1 and 2 summarise key characteristics of the Mean 61
patient and carer samples. Range 33–82
The samples were mixed in terms of individuals’
occupational histories and socio-economic status, hous- Relationship to patient
ing tenure and urban-or-rural residential locations. Wife 10
Table 3 shows the pattern of patients’ place of death Husband 4
preference together with actual place of death for the 35 Sister 2
who had died by the end of the study period. Daughter 1
If the preference percentages in Table 3 are recalcu- Son 1
lated among those patients who expressed a definite
Lived with patient prior to cancer diagnosis
preference, the most common preference was for a home Yes 16
death (36%). This makes our findings consistent with No 2
Townsend et al. (1990) and those of other studies
reviewed by Higginson and Sen-Gupta (2000). However, Total 18
as a sole preference, this home preference of 36% was
lower than would be expected from other studies (for
example, it was 58% in Townsend et al., 1990), and at
29% the hospice preference was higher than that found
elsewhere (20% in the Townsend study). What differ- that an additional 32% of patients held an equal
entiates our study more markedly, however, and preference for a home or hospice death, and that no
partially explains our lower ‘home only’ percentage, is patient expressed a wish to die in hospital. Thus,
preferences in our sample were overwhelmingly in
favour of either a home or hospice death. Only four
patients in the sample altered (as opposed to completely
Table 1
Patient sample characteristics changing) their preference over time: two who had
favoured a home death became more oriented toward a
Key characteristics N hospice death, one leaned from ‘hospice if I’m bad’
Gender toward home, and one expanded his preference from
Male 17 hospice to ‘hospice or local hospital’.
Female 24 Although our sample is small, when preferences were
matched against actual place of death, a clear picture
Age emerged. Patients who wished to die at home were not
Mean 67 always able to do so; in contrast, patients who preferred
Range 41–88 to die in hospice were all able to do so. If a patient had
an equal preference for a home or hospice death, all
Tumour type deaths, with one exception, occurred in hospice rather
Lung 9
than at home. A fifth of patients ended their lives in one
Colorectal 8
Breast 6 of the three hospitals in the MB area.
Oesophagus/gastric 6
Ovarian/endometrial 4 The nature of ‘preference’
Prostate 2
Other 6 Before setting out our findings on the factors that
inform place of death preference, it is important to
Living arrangements acknowledge that the apparent certainty of preference
Lives with a carer 24 type conveyed by figures in the paragraphs above, and
Lives alone 17 indeed in the wider literature on this topic, is in large
measure an artefact of analytic processes involved in the
Total 41
construction and presentation of numerical categories.
 ARTICLE IN PRESS
2436 C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444

Table 3
Patients’ preferred and actual place of death

Preference for Actual place of death in the Morecambe Bay area

place of death
Home, n Hospice, n Hospital, n Other, n Alive (end Total (preference) Preference met (of those
of study), n who died) (%)
n %

Home 6 2 2 10 24 60
Hospice 4 4 8 20 100
Home or hospice 1 7 1a 9 22 100
Hospital 0
Other 1 1 3 0
Not decided 1 4 2 1 2 10 24
No preference 2 1 3 7
Total (place died) 8 17 7 2 7 41
(24.5%) (50%) (20.5%) (6%) (100%)
a
This person’s place of death is not certain, but was either home or hospice.

Our qualitative engagement with patients revealed that preference. In headline form, but not in order of
eliciting their preferences was not a matter of uncovering importance, these are as follows:
something that exists as a ‘pure’ type, a clearly definable
entity, something that can always be straightforwardly Factors shaping patient and carer preference for place of
categorised and counted. Preferences for place of final death
care and death were not simple choices, and were not
necessarily positive choices. 1. Patient’s social network and living arrangements.
We found that preference was rarely stated categori- 2. Patient’s assessment of the carer’s capacity to care.
cally, but took the form of a stronger or weaker leaning 3. Patient concern for the welfare of the carer/family.
in one direction, qualified by speculation about how 4. Carer’s attitudes and willingness to care.
things might change with events. Patient and carer 5. Symptom management.
statements on these matters were always made with 6. Patient’s fears of loss of dignity.
reference to important contextual factors—relationships 7. Patient and carer perceptions of the reliability of
with significant others, symptoms, physical limitations, services, and the degree of ‘safety’ they offer.
interactions with health professionals and health ser- 8. Patient’s attitude to a hospice.
vices, and so forth. They expressed much uncertainty 9. Patient’s experience of hospitals.
about ‘how things will go’. Thus, preference could not 10. Patient’s knowledge and experience of community
be considered in the abstract. People tended to take a services.
pragmatic view, acknowledging the contingency at work 11. Patient’s attitude to nursing homes.
in their lives. This partly explains the fact that, as Table 12. Patient’s attitude to, and outlook on, death and
3 shows, almost a quarter of the sample (24%) were dying, including religious faith.
‘undecided’ about where they would prefer to die. 13. Previous personal experience of death and dying.
On a deconstructionist epistemological note, discus-
sions with patients and carers about preferences always These factors can be organised into the following
already rested on the unspoken assumption that the thematic domains: the informal care resource (factors
default preference for place of death is the home. That 1–4); management of the body (factors 5 and 6);
is, exchanges between interviewee and interviewer were experience of services (factors 7–11); and existential
generally premised on a mutual, culturally normative, perspectives (factors 12 and 13). While these domains
assumption that the most ‘natural’ place to prefer to die and factors are presented as analytically distinct in the
would be the home. Thus, talk about place of death discussion below, they interact in complex ways in
preference tended to take the form of explanations as to individuals’ lives.
why the death may or may not be able to occur at home.
The informal care resource
Factors that shape place of death preference This domain concerns a rich mixture of life features
bound up with immediate living circumstances, relation-
Analyses of the patient and carer interview data ships, attitudes, and family histories. Of critical im-
yielded 13 factors that influenced place of death portance is the patient’s social-relational context,
 ARTICLE IN PRESS
C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
hinging on answers to questions like: does the patient
live alone or with one or more others?; which family
members, friends and neighbours make up a patient’s
immediate social network?; what are the other demands
in the lives of family and friends (childcare, paid work,
other caring responsibilities)?; which members of the
social network live close by?; what is the quality of the
inter-personal relationships involved—between a patient
and his or her spouse, adult child/ren, and wider family
and friends?
The answers to such questions describe the informal
care resource that patients have access to—the resource
that offers both practical assistance and emotional
sustenance. Patients’ understanding of the nature of
this resource, and of the security it offered, played a
central role in their formulation of place of death
preferences. At one end of the spectrum were patients
whose social networks and living arrangements were
such that they felt extremely well supported by a spouse
and/or wider family and could contemplate ending their
days at home. Take, for example, this patient with
prostate cancer who lived with his wife and expressed a
preference to die at home:
Male patient: I would say I have unlimited resources.
I have a treasure chest of friends and family. B8
(patient identifier), age 63.
At the other end of the spectrum, and less common,
were those who felt relatively unsupported and isolated,
for example, a woman who lived alone and had no
family or close friends nearby:
Female patient: Obviously, because I live on my own,
there will be a time when I won’t be able to manage.
Interviewer: And what do you imagine happening?
Patient: Well I don’t know really. I mean, I think at
the end of my life I would like to go to the hospice. I
don’t want to struggle around managing on my own.
I’m not that interested in staying at home if it’s
terribly, terribly inconvenient. I don’t really know if
that will be possible.
B1, age 50, with lung cancer.
The majority of patients in the sample reported
having something of an informal care and support
resource, whether or not they lived alone, and living
alone was not necessarily associated with a preference to
die in an institutional setting.
Whatever the scale of their social networks, patients’
place of death preferences were also strongly influenced
by their assessment of carers’ capacity to care, both
physically and emotionally. This was irrespective of a
carer’s expressed desire to engage in the labour of
caring. If the caring capacity was deemed to be limited in
important qualitative ways, patients often looked
2437
towards the hospice as their place of death preference.
In making this assessment, patients had in mind factors
like: the health status of a carer; the other demands on
the carer’s time—childcare, jobs, and other caring
responsibilities; the ability of the carer to undertake
tasks like cooking and bodily care; and the carer’s
emotional response. These assessments were often age,
generation and gender related. The following extracts
illustrate some of the inter-personal complexities in play:
Interviewer: But there wouldn’t be anyone like that to
help you, if you were more poorly?
Male patient: I don’t think so, no, I don’t think so.
Oh, I’d have to be in [hospice] I think.
Interviewer: Your son, or daughter or-?
Patient: No, there wouldn’t be nobody here. I
mean, I don’t expect my son to spend all his time
here, he’ll have enough on his plate with [his wife’s]
MS [multiple sclerosis]. L9, age 88, with lung cancer,
lives alone.
The presence of a spouse was no guarantee that care
would be forthcoming, as this woman explained:
Female patient: You see my husband has never been
to the hospital with me to see a specialist. He’s never
been to the doctors to see him. He’ll take me down in
the car but he won’t go in [y], so I’m relying on
other people to go in with me as backup [y]. Well,
what are we going to do at the end? So I just think I
hope I end up in the hospice, at least they look after
you, I mean you don’t want to, but I think, well, I
won’t get any support here. B6, age 61.
Another female patient expressed a preference for a
hospice death, despite living with her son, in part bound
up with his perceived limited caring capacity:
Female patient: We get ready-made mealsy because
he doesn’t cook. I’ve tried, but it’s a disaster. I used
to like cooking and baking and now I don’t any
more, well, I can’t. So we just have something that’s
8 minutes in microwave and that’s it.
Interviewer: And you’ve had to move your bed
downstairs?
Patient: Well, he brought it down, because he
didn’t know what I was going to be like when I came
home. So he brought it down and I was mad. Because
his dad finished up in a bed downstairs and I had him
to nurse. But, then was I glad, because I just can’t get
up stairs, I struggle.
K6, age 75, with lung cancer, lives with son.
These extracts reveal that for some patients a hospice
might seem to be the only realistic place of death option.
Thus, we record ‘hospice’ as their place of death
preference, but in such circumstances preference can
hardly be thought of as entirely positive.
 ARTICLE IN PRESS
2438 C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
A related but separate theme is the association
between patients’ place of death preference and their
concern for the welfare of the carer and wider family.
This concern was most often expressed as ‘not wanting
to become a burden’ upon loved ones, as not wanting ‘to
put them through it’. To an important degree, this
stance appeared to represent a means by which patients
could sustain their own roles as family stalwarts and
carers. This was an especially strong concern among
those patients whose informal carers were their adult
children, but also for two male patients who lived with
their sisters. This concern meant that some patients had
a preference to die in hospice because they believed it
would relieve the emotional and physical burden on
carers in the final stages. For example, one woman who
had nursed her husband with cancer, and knew what his
death had been like for her children, wanted to protect
them as her own death approached:
Female patient: If I was really sick, I mean sick sick, I
wouldn’t want my kids to look after me. I don’t think
I’d want that, I think I’d sooner go somewhere to be
looked after. Yes I think I would sooner thaty
because they were so good with their Dad, and they
went through so much, I wouldn’t want that again.
L16, age 70, with lung cancer, lives alone.
Female patient: If things got sort of distressing to
the family then I want away, yes. K5, age 67, with
colorectal cancer, lives with husband.
Female patient: I think I’d rather go in the hospice,
I think it would be easier for people.
Interviewer: It would be easier for your husband
and—?
Patient: And family, yes, yes. B6 age 61, with breast
cancer.
Turning to carers’ attitudes and expressed willingness
to care, in both pre- and post-bereavement interviews,
most carers, especially spouses, were willing to shoulder
the ‘care burden’ and indeed wanted to do so. To these
carers, a death at home was the preferred option, with
hospice or hospital care being a last resort if all else
failed:
Female carer: We’re in our 70 s. I mean we’re not
young, are we? But no, I feel it’s my duty to look
after him. It isn’t a task. It isn’t a task at all. Not—
whatever I do, I’m pleased to do it—because I love
him so much, and it isn’t a problem. I will look after
him as long as I can breathe. And if that sounds
soppy I don’t care. That’s how I feel [y]. Patient
L14, age 72, with gastric cancer.
In post-bereavement interviews, some carers ex-
pressed regret and sorrow that a home death proved
not to be possible in the face of worsening and
unmanageable symptoms.
However, a small minority of carers did find caring
burdensome, and had a preference for the death to occur
in an institutional setting. Here, for example, is the sister
of a man with pancreatic cancer who was in residence in
her house:
Female carer: Now, this blip [incontinence] that he
had last week happened twice and he was beside
himself, and I just made light of it, I said, ‘Oh forget
it, it’s nothing, don’t worry about it,’ but I thought I
can’t cope with this. I’m not a nurse, it’s not your
child that’s been sick. This is a man and he’s full of
cancer. And you somehow get this smell in your
house that you—. Now that’s what hospitals are for.
Patient L4, age 58, with pancreatic cancer, cared for
by sister.
Management of the body
Place of death preferences were much influenced by a
range of corporeal matters, factors bound up with
symptom management and the maintenance of personal
dignity and identity in the face of an increasingly
unreliable body. However, uncertainty about what
might happen to the body was also associated with
indecision about place of death (24% of patients were
‘undecided’, see Table 3), as the following account
demonstrates:
Female patient: I don’t know what I’ll need in the
future, do I? Because I don’t know how it’s going to
progress. Slow–fast–suddenly? I don’t know. I’ll just
deal with it when it comes I suppose, that’s me,
though I do like to know or have an idea so I can be
half prepared. K5, age 67, with bowel cancer.
Despite their contact with specialist palliative care
professionals, most patients and carers felt they lacked a
detailed understanding of likely symptom progression
and of how they would recognise the signs that death
was imminent. While this was very problematic for
some, leading to indecision about place of death and
carer arrangements, patients and carers reported that it
was difficult or impossible to broach these topics with
professionals because this was perceived to be ‘morbid’
and sensitive territory.
Patients’ concerns about the management of symp-
toms had a bearing on their place of death preference,
particularly culturally stigmatised symptoms such as
incontinence associated as it is with ideas of unaccep-
table mess and disgust. For example:
Male patient: I had said that when I got close to
dying that I wanted to go into the hospice and die,
because I didn’t wish to burden my family with the
 ARTICLE IN PRESS
C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
unpleasantnesses, whatever they might be. And
I still hold to that, but whereas before I was
quite certain that I would do it, I would now only
do it if the unpleasantnesses were there. So, I have
altered my tack on that. K1, age 67, with rectal
cancer.
Thus for some patients, maintaining and maximising
personal dignity given the prospect of loss of control
over bodily functions, of extreme pain, or of ‘going
mad’, was bound up with a preference to place their
final care in the hands of professionals in a hospice
setting, should the need arise. In one exceptional case,
these concerns were associated with an expressed
intention to commit suicide, although this patient
actually died in hospital. End of life care by profes-
sionals in an institutional setting was certainly a
preference for those patients who thought it ‘not right’
that a son, daughter or sibling should be required to
engage in intimate care tasks such as toileting, bodily
washing, or clearing up faeces and vomit. Exceptions
might be made if a relative was a nurse. Carers
sometimes concurred with these gendered and genera-
tional care boundaries:
Female carer (sister): Well it was difficult you see,
because with it being my brother, I was limited as to
what I could do as regards his hygiene, so this is why
I got him going to the hospice once a week, but he
managed his stoma all right. But then again, once he
started having the ‘accidents’ he hadn’t the energy to
put his pack on right—put—what I call the towel on.
And of course, we had all this waterproof bedding
which the social services had put on the bed, so I
couldn’t do it. Fortunately my daughter works at the
hospital, so she would come round and see to him on
a personal hygiene level. This was the awkward part
with him—he didn’t have a wife, and if you have a
wife I suppose you have children, you might have a
son that can help, but he didn’t have anyone, so I
found it that hard. That was the difficult thing for
me—I suppose—but some people take it different
than others. Well personally, I don’t like it, I don’t
like that part of it—I don’t mind passing his laces
when he could put his shoes on, I don’t mind
combing his hair and I don’t mind—but I says, when
it comes to anything else, I can’t. Patient L15, age 66,
with bladder cancer.
Carers tended to be particularly worried, as were
some patients, about symptoms associated with mental
impairment. Encountering such symptoms was terrify-
ing and could override a ‘home’ place of death
preference for carers. For example, in a post-bereave-
ment interview one carer recalled the admission of her
husband to hospital because of his changed and
2439
extremely distressed state of mind, where he died shortly
afterwards:
Female carer: [My husband, B] was roaring, it was
like animal noises, and he was punching out and
flaying with his arms, so [the ambulance crew called
to the scene at home] decided to bring a chair to put
him in. Well, he fought and he fought [y]y. [In
hospital] they gave him something to sedate him
every 10 minutes, but then he was back again
fighting, with his arms and what have you. Patient
B8, age 60, with prostate cancer.
Experience of services
This thematic domain brings together the health and
social service experiences that played a role in shaping
patient and carer place of death preference. During their
cancer journey, patients were likely to have encountered
a mix of hospital- and community-based services, and in
this sample a high proportion (73%) had some contact
with a hospice on a day-care or inpatient basis. In some
ways, the relationship between service experiences and
place of death preference is simple and direct: if the
service was perceived to be ‘good’ then that service site
(hospice, community, hospital) could be considered as a
place where one might end one’s days; if aspects of a
service were perceived to be ‘bad’ or indifferent, then
that service site was discounted as a place of death
option.
The qualities that constituted ‘good’ services and
professionals included: reliability, trustworthiness, dis-
playing a ‘caring’ and ‘understanding’ attitude, indivi-
dualised care, an appreciation of the needs of carers,
availability, accessibility, and responsiveness to calls for
help. Another important quality was that a service
engendered a sense of personal safety and security: the
belief that a safety net would be in place should
symptoms become uncontrolled. For example, although
holding a strong preference for a home death, this carer
talks of the hospice as follows:
Female carer: Well [the hospice] offers you a lifeline.
They’re only at the end of a phone. We know that we
can ring at any time with a problem. And I feel that
we could ring up and something would be done to
help us. We feel confident with them, don’t we, B.?
Male patient: Oh yes.
Carer: We feel confident that they’re there for us if
we—I mean we wouldn’t call on anybody unless it
was desperate, would we? We’re a bit independent.
But—um—who knows what we may need. L14, age
72, with gastric cancer, lives with wife, age 70.
‘Bad’ qualities in services and their staff were the
opposites of the ‘good’ qualities described.
 ARTICLE IN PRESS
2440 C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
Against these criteria, hospices, most community-
based nursing services (specialist palliative care nurses,
district nurses, Marie Curie nurses), and some GPs were
generally evaluated highly. Positive experiences of both
specialist community-based palliative care services and
primary care services bolstered many patient and carer
preference for a home death. Encounters with hospice
services often resulted in a sole, or equal (with home)
preference to die in hospice. In contrast, while some
features of hospital services were reported to be
excellent, many aspects of hospital care were viewed as
problematic and distressing. As already noted, none of
the patients in our sample expressed a preference to die
in hospital; in fact, a few expressed a strong preference
not to die in hospital. The following extracts give a
flavour of these service evaluations and their relation-
ship to place of death preferences:
Female patient: [In hospice] You’re well looked after,
you’re well taken care of, nothing is too much trouble
for them. I mean they came down, it was late at
nighty .
Male carer (son): I was going to say, it was a
quarter to 9 at night when we’dy got down there
and thought, ‘Well, they won’t let us in.’
Patient: And they did, didn’t they?
K6, age 75, with lung cancer, lives with son—hospice
was place of death preference.
Female patient: y if it did get to the stage where J.
[my husband] couldn’t look after me, I wouldn’t
mind going in the hospice at all, now I’ve seen how
good they are and how capable they are. And they
look after you so well. I wouldn’t mind going in
there. But not the hospital.
Interviewer: Not the hospital?
Patient: Oh no, no, not if I can avoid it I can tell
you. B10, age 63, unknown tumour type.
Female patient: Last year I didn’t go into hospital.
After my first experience I decided it wasn’t a good
idea. [y] I’d had to go in for a week, but it’s very
difficult with the medical ward up here, because
sometimes it can sort of turn into a geriatric ward
which is all right when you’re very ill but then as you
come around a bit it’s quite depressing really. So this
time I just thought ‘no way I’m going into hospital’,
so I had a few friends come up, but then I just learnt
to manage really. B1, age 50, with lung cancer, lives
alone—strong preference not to die in hospital.
Female carer (wife, post-bereavement interview):
Probably just over two weeks before he died the
district nurse came, and again I can’t praise
her enough, she was absolutely excellent and she
spotted that B. was getting very low, in himself,
which was not his nature. He was very, very
outgoing, very positive about things. And he, I think
it was the first time I ever heard B. admit to her
that he was fed-up. And she sent for our GP. Patient
B8, age 60, with prostate cancer—both patient and
carer preferred a home death, but this was not
achieved.
While hospices were evaluated in overwhelmingly
positive terms, many patients noted that prior to visiting
one they had harboured negative pre-conceptions
about these institutions, presupposing them to be grim
places that one entered to suffer and die. On personal
acquaintance, patients and carers, with few exceptions,
revised their views, expressing surprise at the welcoming,
caring, and even ‘cheerful’ hospice ambience and
at the range of services offered. For many, day-care
or inpatient hospice experiences contributed to the
formation of a hospice place of death preference—
either as a main preference or an equal preference with
‘home’.
In sharp contrast, patients expressed negative views
about nursing homes, whether or not they were
personally acquainted with either a nursing or residen-
tial home, and some saw themselves excluded from this
care sector on grounds of age. Only one patient,
someone who was already resident in a home, expressed
a preference to die in such a setting. Questions about
nursing homes in interview evoked an ‘only as a last
resort’ response, even in cases where standards of care
were known to be good:
Interviewer: The next thing [to ask you about] is
nursing homes?
Female patient: Oh, shut up—for goodness sake
(Laughter) [y].. Oh God, I hope I don’t need that,
ever [y]. With all the old dears sitting with white
boots. [y]y Oh dear God—God preserve me [y].
Was it yesterday we were talking? and I said to J. [my
husband], ‘If I hadn’t had J., I doubt I would
have been able to stay here on my own.’ I would
have needed a lot of care coming in, do you
understand me. I mean I’ve got family, but you
can’t expect family to be driving up here every day
and staying for—I mean they’ve got their own
families. So I suppose one would have to consider
that sort of thing. I mean you obviously wouldn’t
want to. L5, age 76, with bowel cancer, lives with
husband.
Female Patient: If I had to, and I was saving
somebody work, then I would go, but it wouldn’t
be with a smile (laughs). Well, I don’t think it
would—oh gosh, no, and you can get some horrible
people—not running them—the people that are in
them can be horrible too. L17, age 81, with stomach
cancer, lives alone.
 ARTICLE IN PRESS
C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
Existential perspectives
This final thematic domain is concerned with patients’
and carers’ attitudes to death and dying, their previous
exposure to the death of others, and their own earlier
encounters with serious illness. Having or not having
religious faith is also of relevance. Such concerns and
existential orientations tended to lead to either a strong
place of death preference or, in contrast, to adopting a
position that such matters are of little or no conse-
quence.
In their narratives these attitudes and experiences,
accumulated over a lifetime, were woven into patients’
accounts of their current illness, present day-to-day life,
expectations of what lay ahead, and preference, or
absence of preference, for place of death. Some narrative
extracts have already hinted at this—for example, the
two women who talked about how they had, in the past,
nursed their husbands at home until their deaths, and
who themselves now preferred to die in hospice because
they did not want to impose their own dying upon their
adult children (patients L12 and K6).
Place of death preferences were thus bound up with
individuals’ biographies as well as with current circum-
stances. In a few cases, previous near-death experiences
in individuals’ own lives meant that they held no place of
death preference, for example:
Male patient: I’ve been through the business of not
knowing what’s going to happen to me before. In
1977 [my heart nearly failed] and since then things
arey—it’s not something that concerns me particu-
larly. Well, it’s nice to live for another year. L10, age
56, with kidney cancer.
A stoical stance towards life’s closure could also mean
that place of death matters appeared to be of no great
concern. Here, for example, are accounts from two
individuals, one of whom was a man of the Church; both
were recorded as ‘undecided’ about place of death
preference:
Male patient [Vicar]: I mean life does seem unfair at
times, and I’m not complaining. Why should I? And
who said life was fair anyway? Some of us have, I
think I’ve had, a fairly good life. I’ve enjoyed most of
it. One’s not going to be here forever anyway, and
ought to have the wit to recognise this, one hopes—
and I hope there’s—I believe there’s something better
eventually beyond, so the whole process of dying
isn’t—isn’t too painful or difficult, the actual
moment of it, you know—I mean, I will see. B2,
age 65, with leukaemia, lives alone.
Female patient: [The consultant] just said, ‘Now I
can’t give you a time E.,’ he said, ‘how long it will be.’
So, I said, ‘Well that’s something, that’s a quiz, isn’t
2441
it?’ And he said, ‘Yes.’ ‘Well,’ I said, ‘Never mind
about it, whatever happens, happens, and I’ve had a
good life. I’ve got to 81. I said, ‘there’s many a little
kiddie that hasn’t had a life, and they’re just dying of
[cancer], what the hell have I got to grumble about?’
He said, ‘That’s a nice way of looking at it.’ [y]. He
said, ‘Right, now you’re all right? Do you want
anybody to come and talk to you?’ I said, ‘No thank
you. You’ve talked to me and told me what I want to
know. I’ve got it there and I know what I am going to
do.’ He said, ‘What are you going to do?’ I said, ‘Be
happyy, I’ll be happy when I die, I hope. I hope I’m
giggling from here to kingdom come.’ So I think this
is why I keep giggling, you know. I am happy, it’s not
a put on. L17, age 81, with stomach cancer, lives
alone.
For a few patients, their religious belief was a guide to
where they should die, as in the case of the only patient
who wanted to die in a nursing home, and was
interviewed there:
Female patient: I couldn’t cope when I got home, so
there was only one thing to do was to stop and see
what was happening, and I think, being quite a
religious sort, I think that the push, falling onto the
floor, was God saying, ‘Right, we’ve got your now.
Push—this is it, where you go now is where you’re
stopping.’
Interviewer: So you imagine now you’re going to
stay here until the end of your days?
Patient: I’ve told them all. I’ve sat here about 1
o’clock in the morning, this is the Monday night—at
1 o’clock, I’m sitting here and I look up and round
there there’s like a squiggle—now there isn’t, but
there was. It went round and round, a whirly thing,
and then it went out through the door, and a voice
said, ‘You are now here. We’re leaving you all our
good will and good wishes and spiritual help for the
future.’ B9, age 85, cancer of the oesophagus.
Concluding remarks
Singly and in combination, the 13 factors outlined
above shaped the place of death preferences of 41
terminally ill patients and 18 of their carers. Our
organisation of these factors into four thematic domains
has added coherence to the conceptualisation of place of
death preference: the informal care resource; manage-
ment of the body; experience of services; and existential
perspectives. It is testimony to the strength of qualitative
research methods that the narratives of relatively small
numbers of patients and carers can extend significantly
the limited range of factors previously identified
(Higginson & Sen-Gupta, 2000). More important, such
 ARTICLE IN PRESS
2442 C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
research has been able to throw light on what lies behind
these factors. In the narrative extracts we have glimpsed,
for example, the unsupportive husband, the sister who
finds it difficult to have her sick brother in her house, the
devoted spouse, the son who cannot cook, the nurse who
is reliably responsive in times of need, and the desire to
be in the hands of hospice professionals when symptoms
are ‘unpleasant’.
In response to Higginson and Sen-Gupta’s (2000)
speculations, cited earlier, that ‘A preference for home
care or death may be an empowered expression of
wishes or an aversion from the perceived disadvantages
of hospital care’ (Higginson & Sen-Gupta, 2000, p. 299),
we can say that it can be both simultaneously. Similarly,
we can confirm that all of the following may operate as
factors, both singly and in consort: ‘A preference for
hospital or hospice care may indicate a resigned
acceptance of the inevitability of inpatient care, a desire
to save relatives and close friends the burden of caring at
home, a belief in better care being providedy‘
(Higginson & Sen-Gupta, 2000, p. 299). However, none
of the patients and carers in this study preferred
inpatient care because of ‘y a refusal to admit that
a cure is not possible’ (Higginson & Sen-Gupta, 2000,
p. 299).
The data leave us with two other strong impressions.
One is that preference is a socially contingent leaning in
a particular direction, rather than an abstractly arrived
at certainty. This leads us to be wary of the apparently
resolute character of preference as reported in more
quantitative research literature. Second, despite their
very serious ill-health, most patients wished to exercise
agency until the last possible moment. An example of
this is preferring to die in a hospice because this is
perceived to relieve the burden on family members: it is a
means by which people can sustain their own roles as
guardians and carers in familial and other intimate
relationships. Another example is that patients may opt
to place their final care in the hands of hospice
professionals so that personal dignity and identity is
protected in the face of loss of bodily control, or as a
way to avoid culturally prohibited boundary crossing: ‘it
is not right that my son should have to clean me up’. In
this way, patients expressed their understanding that
hospices were places that could deal with what Lawton
(2000) has termed the ‘unbounded’ dying body. Of
course, social scientists who have made a study of death
and dying in contemporary society would not be
surprised by this second observation. They have argued
that the individualism characteristic of Western society
extends into our dying days such that choice, control,
and self-determination are the expected hallmarks of the
‘good death’ (Seale, 1995; Field, 1996; Clark &
Seymour, 1999; Seymour, 1999).
One reading of our findings is that there are many
reasons why cancer patients would prefer not to die at
home.2 We have seen that the reasons given for dying in
a hospice make perfect sense in many circumstances,
whether formulated positively or in the spirit of resigned
acceptance. These include: limitations imposed by the
informal care resource or living circumstances; the drive
to protect loved ones and relieve them of the burden of
care; the desire to sustain personal dignity once bodily
control is lost; the attraction of ‘safe’ professional care
in the face of pain and other distressing symptoms; and
witnessing the exceptionally ‘caring’ qualities of hospice
care compared with most hospital care. It should also be
noted that this preference for hospice deaths is likely to
spread among cancer patients as a consequence of recent
changes in hospice services in the UK. That is, since
greater proportions of cancer patients are becoming
familiar with hospice settings—following hospices’ move
towards the provision of day-care services and short-
stay inpatient care followed by discharge to community
services—the number who wish to die in hospice is likely
to grow. The anti-hospital stance of our respondents,
more prevalent in this study than in others (Higginson &
Sen-Gupta, 2000), has been partly explained by the
biases in the sample. However, the generally low levels
of preference that are found for a hospital death
continue to pose a serious challenge to the hospital
sector. The same can be said of the nursing and
residential care sector.
In making recommendations to palliative care service
commissioners and providers in the Morecambe Bay
area, our findings pointed to the need to maintain
hospice provision at its current level in the locality, and
indicated that additional resources should be invested in
community-based palliative care services to support a
greater proportion of deaths at home. We suggest that
this message has more general relevance: there should be
a greater questioning of the ‘home is best’ mantra, and a
fuller appreciation of the factors that shape preference
for place of death. In our view, there should certainly be
additional investment in specialist and non-specialist
community-based palliative care services to increase
opportunities for home deaths—especially 24-h district
nursing services, out-of-hours on-call specialist palliative
care services, all-week Marie Curie night sitting services,
and more comprehensive social services support. But
policy makers, service commissioners, and providers
should also understand that many patients and carers
have sound reasons for desiring other end of life options,
particularly the opportunity to die in a hospice. We
2
Other researchers in areas where the per capita availability
of hospice beds is below the national average may wish to
examine whether the high level of preference for a death in
hospice found in our study could be partially explained by the
large proportion of terminally ill cancer patients in contact with
a hospice in the MB area (consequent upon relatively large bed
numbers).
 ARTICLE IN PRESS
C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444
would argue that inpatient hospice care in the UK
should also be in receipt of investment and support—to
serve the needs of cancer patients everywhere, and, of
course, so that patients with non-malignant conditions
can also have the privilege of accessing hospice care.
Acknowledgements
We are grateful to the patients and carers who agreed
to participate in the research at such a difficult time in
their lives. The health professionals who have assisted
the project in a variety of ways are also owed many
thanks. We are grateful to the Research and Develop-
ment Department, NHS Directorate of Health and
Social Care North, for generously funding this research
(Grant Ref: RDO/28/3/10).
References
Addington-Hall, J. M., & McCarthy, M. (1995). Dying from
cancer: Results of a national population-based investiga-
tion. Palliative Medicine, 9, 295–305.
Bowling, A. (1997). Research methods in health: Investigating
health and health services. Buckingham: Open University
Press.
Brown, M., & Colton, T. (2001). Dying epistemologies: An
analysis of home death and its critique. Environment and
Planning, 33, 799–821.
Butow, P. N., Maclean, M., Tattersall, M. H., & Boyer, M. J.
(1997). The dynamics of change: Cancer patients’ prefer-
ences for information, involvement and support. Annals of
Oncology, 8(9), 857–863.
Cantwell, P., Turco, S., Benneis, C., Hanson, J., Neumann, C.
M., & Bruera, E. (2000). Predictors of home death in
palliative care patients. Journal of Palliative Care, 16(1),
23–28.
Cartwright, A. (1991). Changes in life and care in the year
before death 1969–1987. Journal of Public Health Medicine,
13(2), 81–87.
Christakis, N. A., & Lamont, E. B. (2000). Extent and
determinants of error in doctors’ prognoses in terminally
ill patients: Prospective cohort study. British Medical
Journal, 320, 469–473.
Clark, D., & Seymour, J. (1999). Reflections on palliative care.
Buckingham: Open University Press.
Department of Health. (1997). The new NHS: Modern,
dependable. Cm 3807, London: HMSO.
Department of Health. (1998). In the public interest: Developing
a strategy for public participation in the NHS. London:
Department of Health.
Department of Health. (2000) The NHS cancer plan: A plan for
investment. A plan for reform. London: Department of
Health.
Department of Health (DOH). (2001). Caring for cancer
patients at home, 25 July 2001. London http://www.doh.
gov.uk/newsdesk/latest/4-naa-25072001.html.
2443
Dunlop, R. J., Davies, R. J., & Hockley, J. M. (1989). Preferred
versus actual place of death: A hospice palliative care team
experience. Palliative Medicine, 3, 197–210.
Ellershaw, J., & Ward, C. (2003). Care of the dying patient: The
last hours or days of life. British Medical Journal, 326,
30–34.
Field, D. (1996). Awareness and modern dying. Mortality, 1,
255–266.
Gatrell, A. C., Harman, J. C., Francis, B., Thomas, C., Morris,
S. M., & McIllmurray, M. B. (2003). Place of death:
Analysis of cancer deaths in part of north-west England.
Journal of Public Health Medicine, 25(1), 53–58.
Giddens, A. (1991). Modernity and self-identity: Self and society
in the late modern age. Cambridge: Polity Press.
Glaser, B. G., & Strauss, A. L. (1967). The discovery of
grounded theory: Strategies for qualitative research. New
York: Aldine De Gruyter.
Grande, G. E., Addington-Hall, J. M., & Todd, C. J. (1998).
Place of death and access to home care services: Are certain
patient groups at a disadvantage? Social Science and
Medicine, 47, 565–579.
Herd, E. B. (1990). Terminal care in a semi-rural area. British
Journal of General Practice, 40, 248–251.
Higginson, I.J. (1999). Which cancer patients die at home?
District data for North West Region, Fact sheet no. 31.
London: Hospice Information Service.
Higginson, I. J., Astin, P., & Dolan, S. (1998). Where do cancer
patients die? Ten-year trends in the place of death of cancer
patients in England. Palliative Medicine, 12, 353–363.
Higginson, I. J., Jarman, B., Astin, P., & Dolan, S. (1999). Do
social factors affect where patients die: An analysis of 10
years of cancer deaths in England. Journal of Public Health
Medicine, 21, 22–28.
Higginson, I. J., & Sen-Gupta, G. J. A. (2000). Place of care in
advanced cancer: A qualitative systematic literature review
of patient preferences. Journal of Palliative Medicine, 3,
287–300.
Hinton, J. (1994). Can home care maintain an acceptable
quality of life for patients with terminal cancer and their
relatives? Palliative Medicine, 8, 183–196.
Hunt, R. W. (1997). Place of death of cancer patients: Choice
versus constraint. Progress in Palliative Care, 5, 238–242.
Karlsen, S., & Addington-Hall, J. (1998). How do cancer
patients who die at home differ from those who die
elsewhere? Palliative Medicine, 12, 279–286.
Lawton, J. (2000). The dying process: Patients’ experiences of
palliative care. London: Routledge.
Scientific Software Development. (1997). ATLAS.ti visual
qualitative data analysis, management and theory building.
Version WIN 4.2. Berlin: Scientific Software Development.
Seale, C. (1995). Heroic death. Sociology, 29, 597–613.
Seale, C., Addington-Hall, J., & McCarthy, M. (1997).
Awareness of dying: Prevalence, causes and consequences.
Social Science and Medicine, 45(3), 477–484.
Seymour, J. E. (1999). Revisiting medicalisation and ‘natural’
death. Social Science and Medicine, 49, 691–704.
Small, N., & Rhodes, P. (2000). Too ill to talk? User involvement
and palliative care. London: Routledge.
Thomas, C., Morris, S. M., & Gatrell, A. C. (2003). Place of
death in the Morecambe Bay area: Patterns and preferences
for place of final care and death among terminally ill cancer
 ARTICLE IN PRESS
2444 C. Thomas et al. / Social Science & Medicine 58 (2004) 2431–2444

patients and their carers. Project report, Institute for Health
Research, Lancaster University, UK.
Townsend, J., Frank, A., & Fermont, D., et al. (1990). Terminal
cancer care and patients’ preferences for place of death:
Prospective study. British Medical Journal, 301, 415–417.
Vigano, A., Dorgan, M., Buckingham, J., Bruera, E., & Suarez-
Almazor, M. E. (2000). Survival prediction in terminal
cancer patients: A systematic review of the medical
literature. Palliative Medicine, 14, 363–374.