Vol. 22 No.

3 September 2001 Journal of Pain and Symptom Management 717

Evaluating the Quality of Dying and Death

Donald L. Patrick, PhD, MSPH, Ruth A. Engelberg, PhD,
and J. Randall Curtis, MD, MPH
Departments of Epidemiology (D.L.P.), Health Services (D.L.P., R.A.E., J.R.C.), Medicine (R.A.E.,
J.R.C.) and Sociology (D.L.P.), University of Washington, Seattle, Washington, USA

Abstract
We propose a model for evaluating the quality of dying and death based on concepts elicited
from literature review, qualitative interviews with persons with and without chronic and
terminal conditions, and consideration of desirable measurement properties. We define quality
of dying and death as the degree to which a person’s preferences for dying and the moment of
death agree with observations of how the person actually died, as reported by others. Expected
level of agreement is modified by circumstances surrounding death that may prevent following
patient’s prior preferences. Qualitative data analysis yielded six conceptual domains:
symptoms and personal care, preparation for death, moment of death, family, treatment
preferences, and whole person concerns. These domains encompass 31 aspects that can be
rated by patients and others as to their importance prior to death and assessed by significant
others or clinicians after death to assess the quality of the dying experience. The proposed model
uses personal preferences about the dying experience to inform evaluation of this experience by
others after death. This operational definition will guide validation of after-death reports of
the quality of dying experience and evaluation of interventions to improve quality of end-of-life
care. J Pain Symptom Manage 2001;22:717–726 © U.S. Cancer Pain Relief
Committee, 2001.

Key Words
Dying, death, palliative care, quality of life, outcomes, terminal diseases

Introduction Our lives can be prolonged by physicians not one

day; we live as long as God has decided. But there
Unser Leben kann sicherlich durch die Ärzte um is a great difference whether we live miserably, like
keinen Tag verlängert werden wir leben so lange es poor dogs, or well and healthy; a clever doctor can
Gott bestimmt hat, aber es ist ein grosser ob wir do a lot in that respect.
jämmerlich, wie arme Hunde, oder wohl und
frisch, und darauf vermag ein kluger Arzt viel. Improving the quality of end-of-life care has
Johann Wolfgang von Goethe, in a conversation become a major agenda for patients, families,
with F. von Müller, August 12, 1827.1 and the loved ones of persons near death as
well as health care professionals, researchers,
and policy makers who organize and provide
care. What happens at the end of life is receiv-
Address reprint requests to: Donald L. Patrick, PhD,
MSPH, Department of Health Services, Box 357660, ing the attention of policy makers and the pub-
1959 N.E. Pacific, University of Washington, Seattle, lic at large, influenced by debates on physician-
WA 98195-7660, USA. assisted suicide, by scientific and technological
Accepted for publication: April 27, 2001. advances that can prolong life, and by the chal-

© U.S. Cancer Pain Relief Committee, 2001 0885-3924/01/$–see front matter

Published by Elsevier, New York, New York PII S0885-3924(01)00333-5
718 Patrick et al. Vol. 22 No. 3 September 2001

lenges of facing death and providing comfort therefore be sought from the patient through-
in dying. out the dying process and from loved ones and
Professional groups have considered what con- caregivers after death. In this way, quality of life
stitutes good quality end-of-life care.2–4 Prior at the end of life is different from the quality of
qualitative research has also described patients’ dying and death by virtue of focus and emphasis.
views of the quality of care at the end of life.5,6 Quality of life emphasizes activities and experi-
Not surprisingly, many patients desire survival as ences while living with a life-threatening or ter-
long as life extension has acceptable quality.7 In minal illness, and the quality of dying empha-
light of these preferences, the goals of high qual- sizes experience of preparing for, facing, and
ity end-of-life care include adequate pain and experiencing death itself.
symptom management, achieving a sense of con- Viewing the quality of dying and death as a
trol for persons who are dying, and respecting measurable concept by assigning a number to
wishes of patients and their loved ones. These reflect quality can be seen as an ultimate form
processes of care are sometimes linked to desirable of reductionism. As one Scottish novelist re-
outcomes such as improved quality of life at the cently noted when contemplating death: “Is . . .
end of life, a notion that has currency for both lay death only a negative number in our combat
persons and professionals.6 readiness or population totals? Or is it a major
The focus on outcomes requires methods to intellectual and emotional loss? How will our
identify the determinants of high quality dying species prosper if we treat ourselves according
and to evaluate interventions purporting to in- to Numerical Facts, as no more than arith-
crease quality of the dying experience.8 Efforts metic? Humanity, its potential and inherent
to improve the quality of dying have included strengths as expressed in the human brain, are
randomized controlled trials of interventions being systematically erased.”14 As applied scien-
such as the promotion of advance directives,9,10 tists using positivist principles, it is not our in-
educational programs with patients or physi- tention to define all that dying means or might
cians,11 and the provision of prognostic infor- possibly mean as this can be known only to
mation to patients and physicians.11 Yet these each person himself or herself or to demean
interventions have not been shown to be any of the richness of life or the dying experi-
highly effective in improving end-of-life experi- ence. The overall objective is to apply humanis-
ences. This ineffectiveness may be attributed, tic thinking and measurement principles to
in part, to the lack of accurate and responsive end-of-life experiences to obtain a summary
means for identifying the most important is- measure that can be applied to populations.
sues and for developing outcome measures to In this article, we present a conceptual model
assess these issues.12 and operational definition for measuring the
A taxonomy of quality of life for dying per- quality of dying and death distinct from quality
sons defined quality of life for patients and for of life at the end of life. This definition guides fu-
loved ones at the end of life.13 This framework ture measurement of patient preferences prior
also included a separate domain labeled the to death and assessment of the degree to which
quality of dying, defined as a personal evaluation these preferences can be and are followed.
of the dying experience as a whole, including a
subjective evaluation of concepts according to
expectations and values. This approach is con- Methods
sistent with the notion of the quality of death The conceptual model and operational defi-
defined by the Committee on End-of-Life Care nition were developed in five steps.
of the Institute of Medicine as “a death that is
free from avoidable distress and suffering for 1. Previous Studies of the Dying Experience:
patients, families, and their caregivers; in gen- After-Death Reports
eral accord with the patients’ and families’ Innumerable authors have considered the
wishes; and reasonably consistent with clinical, dying experience using the methods of the hu-
cultural, and ethical standards.”3 Each dying manities and social sciences, and our review
person, family member or loved one in this was necessarily selective. We focused on major
conceptualization may have a different sense research reports describing the actual end-of-
of what a good death might be. Views could life experiences of dying individuals that in-
Vol. 22 No. 3 September 2001 Quality of Dying and Death 719

clude after-death reports or comparisons of contrast, other studies found congruence be-
pre- and post-death reports. In the United tween relatives and hospice patients’ accep-
Kingdom, extensive surveys were conducted in tance of death21 and between professional car-
1969 and 1987 amid rapid demographic and egivers and patients ratings of outcomes of
social change and the institutionalization of palliative care.22,23 A ground-breaking study
the hospice movement.15,16 Symptoms domi- comparing relatives’ retrospective reports of
nated after death reports, including pain, terminal illness with patients and relative re-
sleeplessness, lack of bladder and bowel con- ports prior to death indicated only moderate
trol, mental confusion, trouble with breathing, reproducibility on symptoms of vomiting, dysp-
and many other sensations.16 Despite the fre- nea, and immobility and poor congruence on
quent desire for people to spend their last days pain, anorexia, and depression.24 The author
at home and as independent as possible, inten- of this study concluded that “this study and
sive and expert nursing needs were evident other available evidence indicate that relatives’
that placed considerable burden on caregivers, retrospective reports of terminal illness, mea-
primarily relatives. sured against current ratings, are moderately
Many other studies followed these early land- reliable for some items but can vary or be po-
mark surveys from the United Kingdom and tentially misleading over other aspects, includ-
have been well-reviewed in previous reports.3,13 ing pain.”24
Of particular importance is a major study of se- This literature review provided support for a
riously ill hospitalized patients and the dying conceptual model that incorporated both pa-
experience that included interviews with the tient pre-death preferences and after-death re-
patient’s surrogate decision maker between 4 ports of the dying experience in understanding
and 10 weeks after a patient’s death.17 In this and improving the quality of dying and death.
study, where the majority of patients were con-
scious and many had significant pain, dyspnea, 2. Qualitative Interviews Defining Domains of
and fatigue, family members after death be- Dying and Death
lieved patients preferred comfort over life ex- In previous studies, patients with advanced
tension. Nevertheless, life-sustaining treat- AIDS and their primary care clinicians were re-
ments were often used. cruited into a prospective cohort study.25,26 We
Common to all these studies is an emphasis conducted qualitative one-on-one interviews
on life events, including dying, as matters of with 52 patients (5 patients died during the six-
personal choice or autonomy. Dying persons month period) and held focus groups with 47
have become participants in decision-making patients based on sex, ethnicity, and injection
as joint adventurers with professionals in the drug use.27 Three focus groups, including 16
social construction of the end-of-life experi- patients with end-stage COPD, were conducted
ence.18 This role as active participants requires to compare explicitly with AIDS patients. Dur-
meaningful communication among dying pa- ing these interviews and focus groups, an expe-
tients, their loved ones, and clinicians. The im- rienced interviewer asked patients to describe
portance of autonomy and communication was what they considered a good and bad death.
demonstrated in a recent study of seriously ill Scripted probes were used to encourage pa-
patients, unrelated family members after death, tients to continue talking about the features of
physicians, and other care providers. In a mail a “good and bad” death.
survey, respondents reported good consensus One-on-one interviews and focus group ses-
and strong endorsement for the importance of sions were audiotaped and transcribed verba-
symptom management, ability to prepare for tim. The transcripts were reviewed indepen-
the end of life, and strong relationships with dently by the investigators to identify themes of
health care providers.19 a good death and a bad death. All themes were
Only a few studies, however, have directly underlined independently by each investiga-
compared the account given by dying persons tor. A preliminary coding scheme was devel-
with their loved ones or caregivers. In a follow- oped from reviewing the transcripts and this
up study to the early surveys in the United King- scheme was used and augmented by each in-
dom, reports of pain were found incongruent vestigator. To address the trustworthiness of
between caregivers and actual patients.20 In the data, we compared coding across investiga-
720 Patrick et al. Vol. 22 No. 3 September 2001

tors and examined passages together for 4. Review of Existing Instruments

themes that were not identified by both investi- A large number of quality of life measures
gators. All themes identified from the tran- have been employed with people with terminal
script review were identified in at least two dif- illness covering domains that may bridge qual-
ferent focus groups or individual interviews, ity of life at the end of life and the quality of dy-
suggesting that additional interviews were un- ing and death.34 We searched the literature
likely to have identified more themes.28 and selected studies of dying patients that ad-
dressed the conceptual domains or used mea-
3. Findings from Authors’ Prior sures to assess quality of life or quality of dy-
Qualitative Research ing.5,20,35–49 Concepts and domains from these
We reviewed a previously published qualita- measures were identified and listed for possi-
tive study of issues pertaining to the health and ble inclusion into the conceptual model and
social circumstances that patients considered operational definition.
states worse than death.29 In this study, 56 pa-
tients were interviewed from six groups: mild
to moderate cognitive impairment, terminal 5. Desirable Measurement Properties
illness (including cancer), chronic illness, The overall goal of creating a conceptual
AIDS, nursing home residents, and well adults. model and operational definition of quality of
Participants were asked to discuss circum- dying and death was to develop an outcome
stances or situations that might make death measure for use in evaluating the process of
preferable to living and why this situation led end-of-life care. Thus we considered that our
them to consider these situations as worse than conceptualization should include measure-
death. Transcripts produced phrases that ment purpose and ideal properties for an oper-
could be used for defining domains of dying ational definition.50 We considered that the
and death, including the percentage of respon- conceptualization should include: 1) a clearly
dents who mentioned different considerations. identified measurement model and domain
Subsequent studies of preferences for health structure for instrument development; 2) con-
states considered worse than death30 and their cepts and language used by patients and fami-
relationship to preferences for life-sustaining lies themselves; and 3) the expression of indi-
treatment31 further defined domains and spe- vidualized concerns or those issues and concerns
cific items for the measure of the quality of dy- most important to individual patients at the end
ing and death. of life as well as a standardized measure that
In another study, patients with a wide variety could be used across persons. The conceptual
of conditions who survived intensive care, pa- model also had to take into account different
tients residing in nursing homes, and families dying trajectories3 as well as the views of differ-
of these patients were interviewed to identify ent stakeholders, most notably the patients
circumstances in which they did or did not themselves prior to death and family members,
want life-sustaining treatments.32,33 These stud- caregivers, and loved ones both prior to and
ies helped to identify issues of concern at the after death.
end of life associated with refusal or withdrawal For improving the quality of end-of-life care,
of treatments. Finally, we reviewed results of a we consider the ideal instrument to be a previ-
qualitative study of patients with one of three ously validated one that could be administered
terminal diseases: metastatic cancer, end-stage retrospectively (after death) to proxy respon-
COPD, and advanced AIDS. The study used fo- dents such as family members or health care
cus groups of patients as well as focus groups of workers. A valid, retrospective measure will be
family members who had lost a loved one to more feasible than paired pre- and post-death
one of these diseases, to elicit the domains and assessments, because of the inherent difficul-
components of physician skill at end-of-life ties in identifying dying persons8 and the rela-
care.5 This study provided information about tive intrusiveness of pre-death interviews.51 Em-
the components of care that is important to pa- pirical evidence is needed, however, to deter-
tients and families and helped to distinguish mine the extent to which after-death reports
quality of end-of-life care from quality of dying are valid without pre-death assessment of pa-
and death. tients’ preferences.
Vol. 22 No. 3 September 2001 Quality of Dying and Death 721

Results The conceptual model for this definition

A Conceptual Definition and its determinants are shown in Figure 1 as
We define the quality of dying and death as an expansion of a previously published frame-
the degree to which a person’s preferences for work.34 The oval represents the overall concept
dying and the moment of death agree with ob- of the quality of the dying experience and indi-
servations of how the person actually died as cates that preferences are moderated by the
reported by others. Like other prospective stud- unavoidable circumstances surrounding death.
ies of advance directives,32 level of agreement Although the timing for measurement is not
must be modified by unavoidable circum- contained in the model shown in Figure 1,
stances surrounding death that may have pre- preferences of persons who are dying by defini-
vented the realization of patient prior prefer- tion must be assessed prior to death and the re-
ences. For example, patients may state that ports and ratings of the quality of dying and
they wish to die with their children present, death obtained from others after death. In elic-
but the child’s death precedes the parent. iting patient preferences for dying and death,
Thus, agreement would not be expected. Un- we propose a focus whenever possible on the
avoidable circumstances are those that might last week of life. This time period is arbitrary
change patients’ preferences or reasonably and the actual time period of “dying” depends
prevent agreement between preferences prior on the trajectory of dying. Picking one time pe-
to death and observations after death. riod, however, allows standardization across in-

Fig. 1. Conceptual model for measuring quality of dying and death (contained in the shaded oval) and its place in
the overall conceptual model of factors affecting quality and length of life of dying patients and their families.
722 Patrick et al. Vol. 22 No. 3 September 2001

dividuals. In assessing the quality of dying and Table 1

death after death, we propose that the raters Conceptual Domains and Items for Measuring the
Quality of Dying and Death (QODD)
(family members or health care workers) be
asked about the last week of life to facilitate a Symptoms and Personal Care
standardized reporting period. If the patient 1. Having pain under control
2. Having control over what is going on around you
was unable to communicate in the last week of 3. Being able to feed oneself
life, ratings of the quality of dying may be un- 4. Having control of bladder, bowels
obtainable and a one-month or longer recall 5. Being able to breathe comfortably
6. Having energy to do things one wants to do
period would be necessary. Although arbitrary Preparation for Death
and subject to empirical verification, these con- 1. Feeling at peace with dying
sistent time periods will allow summarizing re- 2. Feeling unafraid of dying
3. Avoiding strain on loved ones
ports and ratings across patients and raters. 4. Having health care costs covered
5. Having visits from a religious leader
Domains of Dying and Death 6. Having a spiritual service or ceremony before death
7. Having funeral arrangements in order
Table 1 lists 6 domains and 31 specific issues 8. Saying goodbye to loved ones
derived from the literature review, patient in- 9. Attending important events
terviews and focus groups, prior research, ex- 10. Clearing up bad feelings
Moment of Death
isting instruments, and desisrable measure- 1. Dying in the place of one’s choice
ment properties. Empirical validation using 2. Dying in the state of one’s choice (i.e., asleep, awake,
traditional psychometric techniques and mod- unconscious)
3. Having desired people present at the time of one’s
ern test theory will be required to address the death
conceptual groupings. Family
1. Spending time with spouse/partner
2. Spending time with children
1. Symptoms and Personal Care (6 items). These 3. Spending time with family, friends
items are drawn primarily from the literature, 4. Spending time alone
but also previous instruments and qualitative 5. Spending time with pets
Treatment Preferences
interviews. This domain includes the control of 1. Have discussed end-of-life wishes with your doctor
symptoms (e.g., pain and breathing) as well as 2. Avoid using a ventilator or dialysis
many of the activities of daily living that are im- 3. Have the means to end life, if desired
Whole Person Concerns
portant to dying patients (e.g., feeding, blad- 1. Being able to laugh and smile
der and bowel control, energy, overall con- 2. Being touched and hugged
trol). These items reflect the need for control, 3. Finding meaning and purpose
4. Keeping one’s dignity and self-respect
self-efficacy, and autonomy that was identified
in the literature reviews, interviews, and focus
group analyses.
3. Moment of Death (3 items). This domain in-
2. Preparation for Death (10 items). Based on the volves the place (home, hospital) and state
previously identified notions of autonomy, this (awake or asleep) of one’s choice and the peo-
domain includes items involving planning, ple one wishes to be present at the actual death.
controlling and completing events prior to
death, particularly as they enable the patient to 4. Family (5 items). This domain involves spend-
structure the dying experience as they would ing time with spouse or partner, children, fam-
like. These include: attending important ily or friends, pets, or time alone. It includes is-
events, taking care of health care costs, taking sues specific to some individuals (e.g., only
care of funeral arrangements, saying goodbye those with children) as well as issues that apply
to loved ones, clearing up bad feelings and to all persons (e.g., time with family or friends,
avoiding strain or worry on loved ones. This time alone). The wide range of family struc-
domain also captures important religious and tures and variability in potentially important
spiritual customs and beliefs, such as having interactions, such as those with pets, require
visits from a religious leader and having a spiri- the inclusion of issues that are not universally
tual service before death. Finally, it includes applicable. The quality of dying, like the qual-
items involving acceptance of death such as ity of life, may be individual, creating a de-
feeling at peace and feeling unafraid. mand for measures specific to individuals or at
Vol. 22 No. 3 September 2001 Quality of Dying and Death 723

least dynamic measures with different but com- Our conceptualization permits a focus on the
parable items.52–55 features of a good or bad death that are dis-
tinct from the important features of quality of
4. Treatment Preferences (3 items). Treatment life at the end of life, such as functional status
preferences include items about having dis- or satisfaction with one’s functional status. The
cussed treatment preferences with the pa- domains of dying and death are focused on the
tient’s physician, wishes to avoid the use of ven- wide diversity of individuals’ experiences with
tilator or dialysis, and having the means to end dying including social and spiritual compo-
one’s life if desired. These items reflect physi- nents and are, for the most part, distinct from
cian–patient communication, as well as issues the medical care received at the end of life. We
of control and autonomy. believe that differentiating these concepts–
quality of life at the end of life, quality of medi-
5. Whole Person Concerns (4 items). These items cal care at the end of life, quality of dying and
draw on both interview and research data5,19 death–and measuring each independently are
suggesting that dying patients feel the need to the major steps in identifying the determinants
be understood as a whole person. They include: of a good dying experience and identifying
being able to laugh and smile, being touched and evaluating interventions that will improve
and hugged, finding meaning and purpose, and the dying process.
keeping one’s dignity and self-respect. The prospective design suggested by the
The concepts and domains listed in Table 1 model allows information obtained directly
apply both to preferences prior to death and from patients to be compared to that reported
ratings of quality of dying and death obtained by proxy respondents. We hypothesize that the
from others after death. Unavoidable circum- most valid assessment of the quality of dying
stances surrounding death might include un- and death will use patients’ preferences about
avoidable medical events, such as small bowel ob- dying and death (assessed as ratings of the im-
struction or a sudden worsening of respiratory portance of specific items) to inform the after-
status; family or support events, such as the declin- death assessment of the quality of this aspect of
ing health status of a spouse or unavoidable ab- the dying experience. Empirical research will
sence of a family caregiver; structure and processes be necessary to determine how important these
of care, such as reimbursement practices; and pre-death preferences are to the validity of the
personal and social environmental changes, such as measure of the quality of dying and death.
changes in values or changes in residence and Assessing patient preferences about dying
caregivers. These circumstances will not always and death is an important part of the concep-
be easy to identify or classify given differences tual model for evaluating the quality of dying
of perspectives among everyone involved in a and death, but the practical aspects and limita-
person’s dying and death, including the person tions of this endeavor require careful attention.
himself or herself. Patient preferences, if possible to obtain, may
also change prior to death and thus the stability
of these preferences is important to ascertain.
Discussion Dying patients are hard to identify in advance,
This conceptual model provides concrete and the accuracy of prognoses is limited, mak-
guidance to the spectrum of domains for mea- ing reliance on patient preferences difficult.3
suring the quality of dying and death and sug- Furthermore, it is estimated that approximately
gests a method by which one could develop one in three persons are unable to be inter-
and provide evidence of validity to proxy re- viewed in the last week of life.11 Preferences for
ports obtained after death. No doubt individ- dying and death are implicit in advance care
ual disagreement exists on exactly the concepts planning informing treatment decisions at the
and measures on which to place the most im- end of life and these outcomes may need to be
portance, as the end-of-life field involves po- elicited as part of advance care planning.56 Pa-
tentially divergent stakeholders. The domains tients who can be identified prior to death, in-
elicited, however, suggest that the quality of cluding patients with terminal diseases and
the dying experience can be distinguished those in hospice care, provide a means for vali-
from the quality of life, even at the end of life. dating proxy reports used more universally.
724 Patrick et al. Vol. 22 No. 3 September 2001

Ideally, preference elicitation for patients caring for dying patients, researchers inter-
should take place near death. Prior research ested in defining and improving the quality of
suggests that treatment preferences are rela- dying, and policymakers interested in improv-
tively stable over time30 but further research will ing care for the dying. Features of the quality
need to be done to determine whether prefer- of dying are likely to differ across different
ences about dying and death are also stable populations and care settings. Furthermore,
over time. the dying experience is likely to differ across
Studies on proxy reports have addressed who health care systems and types of health care
is the best proxy respondent and how proxy re- providers as well as across cultures and social
ports change with bereavement.21,51,57 These groups. Identifying variability and identifying
studies suggest that proxy reports of observable systems, providers, and processes that provide
behaviors and characteristics are more consis- the highest quality dying experience for differ-
tent with patient reports than reports of unob- ent persons could improve the dying experi-
servable feelings or preferences. This finding is ence. Measures of dying and death could also
similar to other fields in which proxy reports of be used to evaluate interventions in random-
functioning and activities have been compared ized controlled trials or demonstration projects
to patient reports, observations, or performance on continuous quality improvement. Research-
measures.58,59 Comparison of the person’s prior ers and providers interested in testing such in-
preferences about dying and death and the terventions will need reliable and valid mea-
proxy’s assessment of these preferences is one sures of the quality of the dying experience
way to examine validity and identify optimal that are sensitive to small but important
proxy respondents. Proxy reports after death are changes in order to accurately assess the value
being obtained at varying time intervals post of such interventions. These uses await re-
death. This is an important topic of research and peated experience in research and clinical
should include examination of circumstances practice of various models for defining and
surrounding death, the context of care-giving, measuring the quality of dying and death.
and the priorities of caregivers and providers
themselves that have been examined in prior
studies and are the subject of continuing inquir- Acknowledgments
ies.22,42,60,61 The reliability of reports obtained Support for this paper was provided by the
many months after death have not been exam- Agency for Health Care Policy and Research
ined, and evaluation of the time interval is an im- (R03 HS09540) and a grant from the Robert
portant part of the research agenda for measur- Wood Johnson Foundation. Dr. Curtis was sup-
ing and improving the dying experience.8 ported by the Open Society Institute Project on
Finally, a cautionary note is in order that pa- Death in America Faculty Scholars Program.
tient preferences about dying and death may
conflict subtly or strongly with family members,
loved ones, and professional caregivers.62 Au-
References
tonomy may be advocated but then denied, par-
ticularly in circumstances where patients are de- 1. Goethe JWv. Goethe’s Gespräche. Vol. 23. Zur-
ich: Artemis-Verlag, 1949.
pendent or a professional judgement labels the
patient “incompetent.” Vulnerable populations 2. Lynn J. Measuring quality of care at the end of
life: a statement of principles [see comments]. J Am
require special attention. Effective communica- Geriatr Soc 1997;45:526–527.
tion is necessary and little is known as yet about
3. Field MJ, Cassell CK, eds. Approaching death:
how well these preferences are communicated, improving care at the end of life. Washington, DC:
if ever, and throughout the dying trajectory. National Academy Press, 1997.
Each stakeholder may hold a determined posi- 4. Emanuel EJ, Emanuel LL. The promise of a
tion, but the social context may force consensus good death. Lancet 1998;251:21–29.
at the surface where conflict is not easy to ob- 5. Curtis JR, Wenrich MD, Carline JD, et al. Under-
serve or mediate. Further research will be neces- standing physicians’ skills at end-of-life care: per-
sary to determine the effect of these factors. spectives of patients, families, and health care work-
Measuring the quality of dying and death ers. J Gen Intern Med 2001;16:41–49.
could be an important advance for clinicians 6. Singer PA, Martin DK, Kelner M. Quality of end-
Vol. 22 No. 3 September 2001 Quality of Dying and Death 725

of-life care: patient’s perspectives. JAMA 1999;281: flect those made by patients or their families? Palliat
163–168. Med 1993;7:219–228.
7. Patrick DL, Danis ML, Southerland LI, Hong G. 24. Hinton J. How reliable are relatives’ retrospec-
Quality of life following intensive care. J Gen Intern tive reports of terminal illness? Patients and rela-
Med 1988;3:218–223. tives’ accounts compared. Soc Sci Med 1996;43:
1229–1236.
8. Teno JM, Byock I, Field MJ. Research agenda
for developing measures to examine quality of care 25. . Curtis JR, Patrick DL, Caldwell E, et al. The
and quality of life of patients diagnosed with life-lim- quality of patient-doctor communication about end-
iting illness. J Pain Symptom Manage 1999;17:75–82. of-life care: a study of patients with advanced AIDS
and their primary care clinicians. AIDS 1999;13:
9. Schneiderman LJ, Kronick R, Kaplan RM, et al.
1123–1131.
Effects of offering advance directives on medical
treatments and costs [see comments]. Ann Intern 26. Curtis JR, Patrick DL, Caldwell E, et al. Why
Med 1992;117:599–606. don’t patients with AIDS and their clinicians talk
about end-of-life care? Barriers to communication
10. Danis M, Southerland LI, Garrett JM, et al. A
for patients with AIDS and their primary care clini-
prospective study of advance directives for life-sus-
cians. Arch Intern Med 2000;160(11):1690–1696.
taining care. N Engl J Med 1991;324:882–888.
27. Curtis JR, Patrick DL. Barriers to communica-
11. Investigators TSP. A controlled trial to improve tion about end-of-life care in AIDS patients. J Gen
care for seriously ill hospitalized patients: the study Intern Med 1997;12:736–741.
to understand prognoses and preferences for out-
comes and risks of treatments (SUPPORT). JAMA 28. Lincoln YS, Guba EB. Naturalistic inquiry. Bev-
1996;274:1591–1598. erly Hills, CA: Sage, 1985.
12. Lo B. Imporving care near the end of life. Why 29. Pearlman RA, Cain KC, Patrick DL, et al. In-
is it so hard? JAMA 1995;274:1634–1636. sights pertaining to patient assessments of states
worse than death. J Clin Ethics 1993;4:33–41.
13. Stewart AL, Teno J, Patrick DL, Lynn J. The con-
cept of quality of life of dying persons in the context of 30. Patrick DL, Starks HE, Cain KC, et al. Measur-
health care. J Pain Symptom Manage 1999;17:93–108. ing preferences for health states worse than death.
Med Decis Making 1994;14:9–18.
14. Kennedy A. Original bliss. New York: Alfred A
Knopf, 1999. 31. Patrick DL, Pearlman RA, Starks HE, et al. Vali-
dation of preferences for life-sustaining treatment:
15. Cartwright A, Seale C. The natural history of a implications for advance care planning. Ann Intern
survey: an account of the methodological issues en- Med 1997;127:509–517.
countered in a study of life before death. London:
King Edward’s Hospital Fund for London, 1990. 32. Danis M, Southerland LI, Garrett JM, et al. A
prospective study of advance directives for life-sus-
16. Cartwright A, Hockey L, Anderson JL. Life be- taining care [see comments]. N Engl J Med 1991;
fore death. London: Routledge & Kegan Paul, 1973. 324:882–888.
17. Lynn J, Teno JM, Phillips RS, et al. Perceptions 33. Danis M, Patrick DL, Southerland LI, Green
by family members of the dying experience of older ML. Patients’ and families’ preferences for medical
and seriously ill patients. Ann Intern Med 1997;126: intensive care. JAMA 1988;260:797–802.
97–106.
34. Stewart AL, Teno J, Patrick DL, Lynn J. The con-
18. Seale C. Constructing death: the sociology of dy- cept of quality of life of dying persons in the context of
ing and bereavement. Cambridge: Cambridge Uni- health care. J Pain Symptom Manage 1999;17:93–108.
versity Press, 1998.
35. Cohen SR, Mount BM. Quality of life in termi-
19. Steinhauser KE, Christakis NA, Clipp EC, et al. nal illness: defining and measuring subjective well-
Factors considered important at the end of life by being in the dying. J Palliat Care 1992;8:40–45.
patients, family, physicians, and other care provid-
36. Cohen SR, Mount BM, Strobel MG, Bui F. The
ers. JAMA 2000;284:2476–2482.
McGill Quality of Life Questionnaire: a measure of
20. Higginson I, Priest P, McCarthy M. Are be- quality of life appropriate for people with advanced
reaved family members a valid proxy for a patient’s disease. A preliminary study of validity and accept-
assessment of dying? Soc Sci Med 1994;38:553–557. ability. Palliat Med 1995;9:207–219.
21. Hinton J. The progress of awareness and accep- 37. Axelsson B, Sjoden PO. Quality of life of cancer
tance of dying assessed in cancer patients and their patients and their spouses in palliative home care.
caring relatives. Palliat Med 1999;13:19–35. Palliat Med 1998;12:29–39.
22. Higginson IJ, Wade A, McCarthy M. Palliative 38. Byock IR, Merriman MP. Measuring quality of life
care: views of patients and their families. J Pain for patients with terminal illness: the Missoula-VITAS
Symptom Manage 1997;24:227–281. quality of life index. Palliat Med 1998;12:231–244.
23. Higginson IJ, McCarthy M. Validity of the sup- 39. Greer DS, Mor V. An overview of National Hos-
port team assessment schedule: do staffs’ ratings re- pice Study findings. J Chron Dis 1986;39:5–7.
726 Patrick et al. Vol. 22 No. 3 September 2001

40. Greisinger AJ, Lorimor RJ, Aday LA, et al. Ter- past accomplishments and a measurement para-
minally ill cancer patients: their most important digm for the 21st century. Ann Intern Med 1997;
concerns. Can Pract 1997;5:147–154. 127:743–750.
41. Higginson IJ, McCarthy M. A comparison of two 53. Joyce CRB, O’Boyle CA, McGee H. Individual
measures of quality of life: their sensitivity and valid- quality of life: approaches to conceptualisation and
ity for patients with advanced cancer. Palliat Med assessment. Amsterdam: Harwood Academic Pub-
1994;8:282–290. lishers, 1999.
42. Hearn J, Higginson IJ. Outcome measures in 54. McHorney CA. Health status assessment meth-
palliative care for advanced cancer patients: a re- ods for adults: past accomplishments and future
view. J Public Health Med 1997;19:193–199. challenges. Annu Rev Public Health 1999;20:309–
43. Morris JN, Suissa S, Sherwood S, et al. Last days: 335.
a study of the quality of life of terminally ill cancer 55. Patrick DL, Chiang YP. Measurement of health
patients. J Chron Dis 1986;39:47–62. outcomes in treatment effectiveness evaluations:
44. Morris JN, Sherwood S. Quality of life of cancer conceptual and methodological challenges. Med
patients at different stages of the disease trajectory. J Care 2000 Sep;38(9 Suppl):II14–25.
Chron Dis 1987;40:545–556. 56. Post SG, Pulchalski CM, Larson DB. Physicians and
45. McMillan SC, Mahon M. A study of quality of patient spirituality: professional boundaries, compe-
life of hospice patients on admission and at week 3. tency, and ethics. Ann Intern Med 2000;132:578–583.
Cancer Nurs 1994;17:52–60. 57. Teno JM, Lynn J. Putting advance-care planning
46. McMillan SC, Weitzner M. Quality of life in can- into action. Clin Ethics 1996;7:205–213.
cer patients: use of a revised hospice index. Can
58. Lawton MP. Assessing quality of life in Alzhei-
Pract 1998;6:282–288.
mer disease research. Alzheimer Dis Assoc Disord
47. Payne SA, Langley-Evans A, Hillier R. Percep- 1997;11(Suppl 6):91–99.
tions of a ‘good’ death: a comparative study of the
views of hospice staff and patients. Palliat Med 1996; 59. Magaziner J, Bassett SS, Hebel JR, Gruber-Bal-
10:307–312. dini A. Use of proxies to measure health and func-
tional status in epidemiologic studies of community-
48. Turner K, Chyee R, Aggarwal G, Phillip J, et al. dwelling women aged 65 years and older. Am J Epi-
Dignity in dying: a preliminary study of patients in demiol 1996;143:283–292.
the last three days of life. J Palliat Care 1996;12:7–13.
60. Spiller JA. Domiciliary care: a comparison of the
49. Ventafridda V, DeConno F, Ripamonti C, et al. views of terminally ill patients and their family care-
Quality-of-life assessment during a palliative care givers. Palliat Med 1993;7:109–115.
programme. Ann Oncology 1990;1:415–420.
61. Fakhoury WKH, McCarthy M, Addington-Hall
50. Patrick DL, Erickson P. Health status and health JM. The effects of clinical characteristics of dying
policy. New York: Oxford University Press, 1993. cancer patients on informal caregivers’ satisfaction
51. Fowler FJJ, Coppola KM, Teno JM. Methodologi- with palliative care. Palliat Med 1997;11:107–115.
cal challenges for measuring quality of care at the end 62. Loew F, Rapin CH. The paradoxes of quality of
of life. J Pain Symptom Manage 1999;17:114–119. life and its phenomenological approach. J Palliat
52. McHorney CA. Generic health measurement: Care 1994; 10:37-41.