JOURNAL OF PALLIATIVE MEDICINE

Volume 13, Number 2, 2010

Mary Ann Liebert, Inc.
DOI: 10.1089=jpm.2009.0269

What Should We Say When Discussing Code Status

and Life Support with a Patient? A Delphi Analysis
James Downar, M.D.C.M., M.H.Sc.(C)1 and Laura Hawryluck, M.D., M.Sc.2

Abstract

Background: Patients and clinicians often find it difficult to discuss wishes regarding cardiopulmonary resuscitation (CPR) or code status. Some authors have published effective communication styles, but there are currently
no published guidelines for the content of a discussion about resuscitation or goals of care.
Methods: We identified a group of physicians with expertise in end-of-life care and communication, and used the
Delphi method to develop a series of consensus statements about the ideal content of a discussion of CPR and goals
of care.
Results: Twelve physicians agreed to participate in the study, generating nine consensus statements. These
statements addressed the following topics: timing the discussion; framing the discussion in terms of goals of
care; distinguishing between life-sustaining therapy (LST) and CPR; describing a cardiac arrest, LST, CPR, and
palliative care; describing what happens after a cardiac arrest; how to modify the discussion to respect a patients
medical condition or beliefs; offering a prognosis; making a recommendation; and the importance of trust and
rapport. There was consensus for each statement after the second Delphi round.
Interpretation: Physicians with expertise in end-of-life care and communication were able to develop consensus
statements for the ideal content of a discussion of CPR and goals of care. These statements can serve as
guidelines for physicians who feel uncomfortable with these discussions, in order to facilitate effective, informed,
and ethically sound decision making.

Introduction

atients and clinicians often find it difficult to discuss

wishes regarding cardiopulmonary resuscitation (CPR)
or code status. Many patients do not want to discuss CPR
with their physician,1,2 and many physicians feel that they do
not have the time or training to discuss this issue properly.3,4
As a result, many seriously or terminally ill patients receive
CPR and advanced life-sustaining measures without prior
discussion and informed consent,57 even though a large
proportion may wish to forego CPR.2
In order to help with the process of communication, many
authors have published effective communication techniques
(i.e., the style of communication).812 However, there are
currently no published guidelines for the ideal content of a
discussion of resuscitation. This leaves each clinician to decide
(for example) how to describe CPR and life-sustaining therapy (LST), how to explain a prognosis, or whether or not to
make a recommendation about CPR and LST to the patient.
These are important issues, as many patients prefer a shared
model of decision-making,13 and change their preferences for

resuscitation based on their understanding of life support14

and their likelihood of surviving CPR.15
In the present study, we sought to develop content guidelines for discussions of CPR and LST using expert opinion
from clinicians who frequently participate in discussions of
CPR and LST. Although every physician has their own approach, we hypothesized that we could identify common
ideas and develop recommendations for clinicians who do not
feel comfortable with these conversations and for those who
seek to improve their skills. In order to reconcile differences of
opinion, we selected the Delphi technique of consensus generation, which has been used previously for controversial
topics in end-of-life care.16
Methods
We used the Delphi method to develop content guidelines
for discussions of CPR. The Delphi method is a validated
qualitative technique for measuring and achieving consensus
in an area for which there is no scientific evidence.17 In
the Delphi method, consensus is achieved using a series of

1
Division of Palliative Care, 2Division of Critical Care, Toronto General Hospital, Toronto, Ontario, Canada.
Accepted September 22, 2009.

185

186
rounds. In round 1, we identified a group of critical care and
palliative care physician experts and asked each to provide
written answers to a series of questions (Appendix A). The
answers were analyzed using open coding and grouped using
axial coding. The codes were then organized into subthemes,
which were organized into themes that were used to generate
nine separate statements. One of the authors ( J.D.) reviewed
and coded the responses and generated the statements, while
the other author (L.H.) reviewed the coding scheme and the
statements.
The statements were generated by one of the authors ( J.D.)
based on dominant themes from round 1. A dominant theme
would be either an idea expressed by more than one participant, or a well-justified idea expressed by one participant. If
two dominant themes appeared contradictory, J.D. attempted
to generate a compromise that would allow for both views.
In round 2, each expert was then asked to rate his=her
agreement (on a 9-point Likert scale; 1 least agreement,
5 neutral,9 fullagreement)witheachoftheninestatements.
These ratings were then collected and analyzed for consensus.
Consensus was defined as 70% of respondents giving a rating
of 7 (agreement) or better for a given statement.
Participants
We aimed to enroll clinicians with expertise and experience
in end-of-life care and communication. We reviewed the recent literature to identify investigators who have published
articles relevant to end-of-life care and decision making. In
addition, we used snowball sampling by asking local experts
at the University of Toronto to suggest individuals nationwide with an interest in this area, in order to include experts
who might not have a research focus. We focused on clinicians
with a practice that would involve frequent CPR discussions:
palliative care and critical care physicians. We solicited participation using e-mail invitations.
Ethics and confidentiality
This study was approved by the research ethics board of
the University Health Network, Toronto, Canada. All participants provided voluntary and informed consent prior to
enrolment, and all participants could opt not to provide any
personal demographic information. Responses were sent directly from participants to the authors via e-mail or facsimile,
but all identifying information was removed from the responses prior to coding.
Results
We began enrolment and round 1 in June 2008, and completed round 2 in March 2009. We offered participation to 25
clinicians, of whom 12 agreed to participate and completed
round 2. Demographic data for the participants are provided
in Table 1. Participants were equally split between critical care
and palliative care specialties. Not all participants agreed to
provide all the information requested.
Responses from round 1 were analyzed using open coding
and grouped using axial coding to form answers to the 10
questions asked in the question guide. The coding scheme is
provided in Appendix B. The dominant or salient themes
were organized to form 9 statements to be evaluated in round
2 (Appendix C). Participant agreement scores from round 2

DOWNAR AND HAWRYLUCK

Table 1. Participant Demographics
Participant demographic
Age (mean [SD])
Female gender
Religion
Protestant
Catholic
Christian Orthodox
Buddhist
Sikh
Agnostic
Decline=none
Year of graduation from
medical school
1970s
1980s
1990s
2000s
Country of residency training
Canada
Canada and United States
Current practice location
Ontario
Quebec
Alberta
Years in practice (mean [SD])
Specialty certification
Internal medicine
Family medicine
Critical care medicine
Palliative care medicine
Respirology
Anaesthesiology
Emergency medicine
Practice type
Academic hospital
Community hospital
Community practice
Primary care
Referral-based
Inpatient
Outpatient

Value
44.3 (7.9)
58%

Number
responding (n)
9
12
12

1
1
1
1
1
1
6
10
1
6
1
2
10
9
1
10
7
1
2
14.8 (8.3)

10
10

5
3
6
4
2
1
1
10
8
1
0
3
10
9
2

SD, standard deviation.

are provided in Table 2. More than 70% of experts rated their

level of agreement at 7 (agreement) or greater for all 9 statements in round 2, and so consensus was achieved. One respondent felt unable to rate his=her agreement with 5
statements.
Participants offered a broad range of responses about the
ideal timing of a discussion of code status, with most feeling
that it should be done early, or following any significant
change in diagnosis, level of supportive care, or goals of care
(Statement 1). Participants recommended that resuscitation
and code status should be discussed in the context of the
goals of care, rather than as therapies themselves (Statement
2). Participants expressed little agreement over the distinction
between LST and CPR, with some feeling that they should be
discussed as separate concepts, and others feeling that they
were best discussed together (Statement 3). Participants provided brief and clear explanations of a cardiac arrest, life-

HOW SHOULD WE ASK ABOUT CODE STATUS?

187

Table 2. Participant Agreement Scores for Each Statement

Level of Agreement
Statements
1
2
3
4
5
6
7
8
9

n=a

2
3
1
4
3
2
3
2
1

3
3
2
3
3
4
2
3
3

6
5
9
3
3
3
4
4
7

1
1
1
1
1
1
1

1
1

1
1
1
1

1
2

Frequency table of agreement scores for each statement. 1 least agreement; 5 neutral; 7 agreement; 9 most agreement.

sustaining therapies and palliative care (Statement 4). Participants felt that it was important to explain the short- and
long-term sequelae of a cardiac arrest and successful resuscitation (Statement 5). Participants generally recommended a
patient=family-centered approach to this discussion, which
would usually not need to be modified depending on the
clinical condition or cultural context (Statement 6). Most
agreed that physicians should provide a prognosis (Statement
7) and a recommendation (Statement 8) about whether or not
a patient should receive CPR=LST, but that these should always be made in the context of the patients goals of care.
Finally, Statement 9 combines various recommendations
made by one or more participants that do not specifically
address any question in the guide.
Discussion
In this study, we were able to generate nine consensus
statements to act as guidelines for the content of discussions of
CPR or code status using the Delphi process. This is the first
study to our knowledge to develop such guidelines. These
statements represent the opinions of Canadian clinicians who
have these discussions on a frequent basis, and are intended
as guidelines for clinicians who are uncomfortable or unsure
of how to approach a discussion of resuscitation and life
support. These statements cover a range of issues, predominantly content-related but also stylistic. They are intended to
facilitate effective, informed and ethically-sound decision
making. There was a high degree of consensus for each of the
statements among the experts.
The Delphi format has a number of strengths when used to
develop consensus statements.17 Anonymity avoids dominance by senior or more powerful individuals, while iteration
and feedback allow individuals to reflect and change their
opinions when appropriate. Finally, the measures of agreement can be used to provide more information than just the
final statement. In the present study, we approached specialists from two fields (50% palliative care medicine, 50%
critical care medicine), so we chose a high threshold of
agreement for consensus (70% of participants) so that each
statement would need to be supported by participants from
both specialties. Ultimately, consensus was achieved in the
second round, which may be a reflection of similar approaches
among the participants, or the fact that the statements aimed
to be too inclusive. Subsequent efforts, with more diverse

participants and more specific recommendations might produce greater disagreement, requiring a revision of the statements.
The purpose of this study was to address primarily the
content of a discussion of CPR and life support, but many
statements addressed points of style (e.g., the timing of the
discussion, framing the discussion around goals of care). This
likely reflects the fact that content and style are often difficult
to separate, and both affect the impact of this discussion on
patients and family members.9,1820 Many statements include
suggested phrases similar to those provided by the End-oflife=Palliative Education Resource Center (EPERC).21
These statements are specifically not intended to establish a
gold standard for CPR discussions, with all the medico-legal
implications of such a standard. For example, these statements advocate for a model of shared decision-making, which
includes a recommendation (Statement #5) from the physician
about whether or not the patient should receive CPR or LST.
This model is widely advocated22,23 and often preferred by
the patient,13 but some surrogate decision makers find such
recommendations undesirable.24 There are many acceptable
models of clinicianpatient interaction that can be applied to
discussions of goals of care,25 and clinicians should feel free to
use any approach that is consistent with the legal and ethical
principles that apply to their practice.
Some of these recommendations are inconsistent with
published evidence. For example, Statement 7 advises clinicians to offer a prognosis for survival in simple terms, and
gives an example of a qualitative estimate of high risk (e.g.,
very few people are able to survive it . . . ). In contrast, one
study has suggested that quantitative expressions of risk (e.g.,
90% of people do not survive) lead to better comprehension
among surrogate decision-makers.26 Another recent study
suggested that physicians prefer to prognosticate in terms of
quality of life rather than mortality.27 Future revision of these
recommendations, with broader inclusion of participants
from other specialties may change the focus of this recommendation.
These statements are not intended to establish or modify
accepted definitions in the literature. For example, in Statement 4, the explanation of palliative care does not precisely
match that of the World Health Organization,28 and the distinction between basic and advanced LSTs is neither clear
nor universally accepted. These explanations were suggested
by the participants for a non-medical audience. Participants

188
recommended aroutine explanation of palliativecare.Thismay
be very important, as palliative care is often overlooked as an
alternative to LST, even for patients with a very poor prognosis.5
The clinicians who developed these statements have considerable expertise and experience in this area, but their
opinions are subject to bias. For example, Statement 6 recommends a fundamentally patient-centered approach, which
may not be acceptable to patients who prefer a family-centered
or group-centered approach.29,30 This likely reflects the cultural
background of the experts, who are mostly North Americanborn and trained.
The participants consisted entirely of Canadian specialists
in critical care or palliative care medicine. We sought to enroll
Canadian participants solely because legal, ethical, and cultural norms are often particular to each country, and so we
chose to focus on the Canadian context. It would likely take a
much larger study to identify commonalities between practitioners from different nationalities, and statements may
need to be less specific in order to achieve consensus. Clearly,
recommendations developed by experts with one set of legal,
ethical, and cultural norms may not be relevant to clinicians
with another set of norms.
The fact that we enrolled only critical care and palliative care
specialists is a reflection of the process we used to identify
experts (i.e., publications and snowball sampling). There are
many clinicians in a variety of specialties (e.g., geriatrics, oncology) with expertise and experience discussing goals of care who
would not have been approached for enrolment in the present
study. In addition, we do not have demographic data for the
experts who refused participation in this study, so we cannot
determine if any selection bias may have influenced the results.
In conclusion, we feel that these recommendations would
be very helpful to any clinician who wishes to discuss goals of
care with their patients. We hope that this study is part of an
ongoing process in which recommendations can be developed, revised and improved by the input of a broader range of
clinicians with expertise in this area.

DOWNAR AND HAWRYLUCK

4.

5.

6.

7.

8.
9.

10.
11.
12.
13.

14.

Acknowledgment
This research was supported by funding from Associated
Medical Services, Incorporated.
Author Disclosure Statement

15.

16.

No competing financial interests exist.

References
1. Heyland DK, Frank C, Groll D, Pichora D, Dodek P, Rocker
G, Gafni A: Understanding cardiopulmonary resuscitation
decision making: perspectives of seriously ill hospitalized
patients and family members. Chest 2006;130:419428.
2. Hofmann JC, Wenger NS, Davis RB, Teno J, Connors AF Jr,
Desbiens N, Lynn J, Phillips RS: Patient preferences for
communication with physicians about end-of-life decisions.
SUPPORT Investigators. Study to Understand Prognoses
and Preference for Outcomes and Risks of Treatment. Ann
Intern Med 1997;127:112.
3. Nelson JE, Angus DC, Weissfeld LA, Puntillo KA, Danis M,
Deal D, Levy MM, Cook DJ: Critical Care Peer Workgroup
of the Promoting Excellence in End-of-Life Care Project:

17.
18.

19.

20.

End-of-life care for the critically ill: a national intensive care

unit survey. Crit Care Med 2006;34:25472553.
Sibbald R, Downar J, Hawryluck L: Perceptions of futile
care among caregivers in intensive care units. CMAJ 2007;
177:12011208.
Rady MY, Johnson DJ: Admission to intensive care unit at
the end-of-life: Is it an informed decision? Palliat Med 2004;
18:705711.
Mirza A, Kad R, Ellison NM: Cardiopulmonary resuscitation
is not addressed in the admitting medical records for the
majority of patients who undergo CPR in the hospital. Am J
Hosp Palliat Care2005;22:2025.
Teno J, Lynn J, Wenger N, Phillips RS, Murphy DP, Connors
AF Jr, Desbiens N, Fulkerson W, Bellamy P, Knaus WA:
Advance directives for seriously ill hospitalized patients:
Effectiveness with the patient self-determination act and the
SUPPORT intervention. SUPPORT Investigators. Study to
Understand Prognoses and Preferences for Outcomes and
Risks of Treatment. J Am Geriatr Soc 1997;45:500507.
Back AL, Arnold RM, Quill TE: Hope for the best, and
prepare for the worst. Ann Intern Med 2003;138:439443.
Lautrette A, Darmon M, Megarbane B, Joly LM, Chevret S,
Adrie C, Barnoud D, Bleichner G, Bruel C, Choukroun G,
Curtis JR, Fieux F, Galliot R, Garrouste-Orgeas M, Georges
H, Goldgran-Toledano D, Jourdain M, Loubert G, Reignier J,
Saidi F, Souweine B, Vincent F, Barnes NK, Pochard F,
Schlemmer B, Azoulay E: A communication strategy and
brochure for relatives of patients dying in the ICU. N Engl J
Med 2007;356:469478.
Campbell ML: Breaking bad news to patients. JAMA 1994;
271:1052.
Buckman R. How to Break Bad News. Baltimore, MD: Johns
Hopkins University Press, 1992.
Quill TE, Townsend P: Bad news: Delivery, dialogue, and
dilemmas. Arch Intern Med 1991;151:463468.
Heyland DK, Cook DJ, Rocker GM, Dodek PM, Kutsogiannis
DJ, Peters S, Tranmer JE, OCallaghan CJ: Decision-making
in the ICU: Perspectives of the substitute decision-maker.
Intensive Care Med 2003;29:7582.
Fried TR, Bradley EH, Towle VR, Allore H: Understanding
the treatment preferences of seriously ill patients. N Engl J
Med 2002;346:10611066.
Murphy DJ, Burrows D, Santilli S, Kemp AW, Tenner S,
Kreling B, Teno J: The influence of the probability of survival
on patients preferences regarding cardiopulmonary resuscitation. N Engl J Med 1994;330:545549.
Hawryluck LA, Harvey WR, Lemieux-Charles L, Singer PA:
Consensus guidelines on analgesia and sedation in dying
intensive care unit patients. BMC Med Ethics 2002;3:3.
Jones J, Hunter D: Qualitative research: Consensus methods for
medical and health service research. BMJ 1995;311:376380.
Glavan BJ, Engelberg RA, Downey L, Curtis JR: Using the
medical record to evaluate the quality of end-of-life care in
the intensive care unit. Crit Care Med 2008;4:11381146.
White DB, Engelberg RA, Bereknyei S, Curtis JR: Toward
shared decision making at the end of life in intensive care
units: Opportunities for improvement. Arch Intern Med
2007;167:461467.
McDonagh JR, Elliott TB, Engelberg RA, Treece PD, Shannon
SE, Rubenfeld GD, Patrick DL, Curtis JR: Family satisfaction
with family conferences about end-of-life care in the intensive care unit: increased proportion of family speech is associated with increased satisfaction. Crit Care Med 2004;
32:1484-8.

HOW SHOULD WE ASK ABOUT CODE STATUS?

21. Fast Facts and Concept #23. End-of-life=Palliative Education
Resource Centre. www.eperc.mcw.edu=fastFact=ff_023.htm
(Last accessed August 30, 2009).
22. Truog RD, Campbell ML, Curtis JR, Haas CE, Luce JM,
Rubenfeld GD, Rushton CH, Kaufman DC: American
Academy of Critical Care Medicine: Recommendations for
end-of-life care in the intensive care unit: A consensus
statement by the American College of Critical Care Medicine. Crit Care Med 2008;36:953963.
23. Curtis JR, White DB: Practical guidance for evidence-based
ICU family conferences. Chest 2008;134:835843.
24. White DB, Evans LR, Bautista CA, Luce JM, Lo B: Are
physicians recommendations to limit life support beneficial
or burdensome? Am J Resp Crit Care Med 2009;180:320325.
25. Emanuel EJ, Emanuel LL: Four models of the physicianpatient relationship. JAMA 1992;267:22212226.
26. Paling J: Strategies to help patients understand risks. BMJ
2003;327:745748.
27. White DB, Engelberg RA, Wenrich MD, Lo B, Curtis JR:
Prognostication during physician-family discussions about

189
limiting life support in intensive care units. Crit Care Med
2007;35:442448.
28. World Health Organization Contributors: Definition of Palliative Care. www.who.int=cancer=palliative=definition=en=
(Last accessed March 10, 2009).
29. Bowman KW, Hui EC: Bioethics for clinicians: 20. Chinese
bioethics. Can Med Assoc J 2000;163:14811485.
30. Coward H, Sidhu T: Bioethics for clinicians: 19. Hinduism
and Sikhism. Can Med Assoc J 2000;163:11671170.

Address correspondence to:

James Downar, M.D.C.M., M.H.Sc.(C)
Division of Palliative Care
Toronto General Hospital
200 Elizabeth Street
9N-926
Toronto, Ontario M5G 2C4
Canada
E-mail: james.downar@utoronto.ca

(Appendices follow)

190

DOWNAR AND HAWRYLUCK

Appendix A. Consensus Statements

Statement 1: Timing the discussion

Clinicians may initiate a discussion of CPR=LST at any point during the course of an illness, or even when a patient is in good health.
However, some patients find this conversation upsetting, particularly if they have not developed a relationship of trust with the clinician.
Clinicians meeting a patient for the first time should exercise their judgment to decide whether this conversation is urgent or would be more
appropriate on a later occasion. Otherwise, the following scenarios should routinely trigger a discussion of CPR=LST:
1.
2.
3.
4.
5.
6.
7.
8.

Any nonelective admission to hospital.

Any elective admission for a high-risk surgical procedure.
The diagnosis of a serious, incurable or end-stage illness.
Any clinical deterioration that signals a change in prognosis.
Any admission to a critical care unit.
A patient who considers stopping therapy for a chronic or end-stage condition.
A patient feels that his=her quality of life is unacceptable.
The physician feels that further non-palliative treatment would be futile.

Statement 2: Framing the discussion

Any discussion of CPR=LST should be framed as a discussion of goals of care rather than code status. However, the clinician cannot simply
ask a direct question such as: What are your goals of care? or Which is more importantlong life or pain control? This discussion involves
an active exploration of the patient=familys values, current perception of their illness and their hopes for the future. Possible goals of care often
fall under one of three categoriescure, prolongation of life, and symptom control. Clinicians should not discuss CPR=LST until they
understand the relative priority of these goals.

Statement 3: Distinguishing between LST and CPR

CPR and LST are distinct therapies, and LST can potentially include many therapies (e.g., mechanical ventilation, vasopressors, hemodialysis, artificial nutrition, antibiotics, etc.). Depending on the goals of care, it may be necessary for the clinician to explain the distinction between
CPR and LST, or which types of LST may be appropriate.

Statement 4: Describing a cardiac arrest, CPR, LST, and palliative care

Depending on the goals of care, it may or may not be necessary to describe a cardiac arrest, CPR, and=or LST. In cases where the goals of care
clearly do not include CPR or LST, this explanation is unnecessary. If an explanation is necessary or requested, the clinician should start by
asking what the patient=family knows about each issue, as this will guide the explanation.
A cardiac arrest is a situation in which the heart stops beating and the patient dies. This is often because the underlying disease (e.g., cancer,
chronic obstructive pulmonary disease [COPD]) has become severe or irreversible. CPR is when a team of health professionals attempts to
restore the heartbeat using chest compressions, a tube down the throat to help push oxygen into the lungs, and possibly intravenous medications or an electric shock. Life-sustaining therapies is a broad term referring to therapies that support the human body. This can include
basic therapies such as intravenous fluids or medications. It can also include advanced therapies such as mechanical ventilation for respiratory
failure, medications for blood pressure, or dialysis for kidney failure.
Palliative care should be presented as a valid treatment plan, rather than simply doing nothing or just keeping you comfortable.
Palliative care is an active treatment plan in which the health care team provides symptom control and comfort measures while a natural process
takes place.

Statement 5: What happens after a cardiac arrest?

When explaining CPR, clinicians need to explain what happens if resuscitation is successful and the patient is transferred to the ICU. This is a
key component of the consent process. Once again, clinicians should start by asking what the patient=family knows about this scenario, as this
will help guide the subsequent explanation.
Patients who survive a resuscitation attempt are transferred to the ICU, where they will receive mechanical ventilation via a tube in their
throat, and probably medications to support their blood pressure. They will typically be unconscious or sedated, and usually unable to
communicate with loved ones. They are very sick, and there is a high likelihood that they will die despite LST (usually approximately 90%).
Among those who survive, many suffer from significant neurological damage and functional decline. There is often physical damage from
the resuscitation attempt, such as broken ribs. In addition, the underlying disease (e.g., cancer, COPD) will still be present and possibly worse as
a result of the cardiac arrest.

Statement 6: How should the clinician modify their approach to reflect the patients condition or beliefs?
Although clinicians must be sensitive to the clinical and cultural context, the essence of the discussion should remain the same regardless of
condition or beliefs. Clinicians should adopt a patient=family-centred approach by asking what the patient=family knows (and wants to know)
about the medical situation, as this will help guide the conversation.
The patient may be very ill, or wish not to be involved in the discussion=decision, in which case a substituted decision-making process can be
used. Some people are not comfortable with detailed medical discussions, in which case a simplified conversation may be appropriate. If the
patient is moribund or actively dying, or if the goals of care clearly emphasize comfort measures, it may not be appropriate to discuss CPR or
LST as a treatment option.

(continued)

HOW SHOULD WE ASK ABOUT CODE STATUS?

191

Appendix A. (Continued)
Statement 7: Offering a prognosis
A discussion of CPR=LST should address the likelihood that each will succeed. Again, this should be a patient-centered prognosis, focusing
on the likelihood of CPR=LST achieving the goals of care established earlier in the conversation.
Clinicians should explain the meaning of prognostication in simple terms. For example, A cardiac arrest is usually a sign of severe illness, so
very few people are able survive it long enough to leave hospital. Patients should also understand the factors that affect prognosis: People
who are older and have metastatic cancer tend to do much worse than a younger person who is otherwise healthy.
Prognostication is sometimes difficult, particularly for less invasive forms of LST. If a clinician is uncertain about the prognosis, he=she
should explain this to the patient=family. The clinician should be careful to correct the misconceptions about CPR=LST that the television
programs often create.

Statement 8: Making a recommendation

Clinicians should always offer a recommendation to the patient=family based on the clinical situation as well as the goals of care.
Recommendations are helpful because they offer guidance while helping to shoulder some of the burden of decision-making. Recommendations do not undermine the autonomy of the patient if they are properly disclosed and explained, e.g., Based on your desire to focus
on comfort and independence, as well as the fact that your cancer has spread, I would recommend . . . .
In cases in which the prognosis is uncertain or the clinician is unsure what to recommend, he=she should consult a colleague. In some cases of
uncertainty, a trial of LST can help indicate the likely future clinical course.

Statement 9: Aim for trust and rapport

Good communication may be time-consuming. Avoid medical jargon and check frequently to ensure that the patient=family does not require
any further clarification. The primary goal of any discussion of CPR=LST=goals of care should be to establish trust and dialogue with the
patient=family, rather than an outcome such as a do not resuscitate order. This is important because some patients=families require time or
multiple family conferences to make these difficult decisions.
CPR, cardiopulmonary resuscitation; LST, life-sustaining therapy; ICU, intensive care unit.

Appendix B. Question Guide

In the first Delphi round, physicians were asked 10 open-ended questions about the content of a discussion of resuscitation and life support:
1. When should you discuss code status or life-sustaining treatments with a patient?
2. How should you frame a discussion of life-sustaining interventions and CPR? (e.g., In the context of illness, prognosis,
reasonableness? Values and goals?)
3. Do you feel life-sustaining treatment discussions should be separated from CPR discussions?
4. How should you describe a cardiac arrest and CPR? Please be very specific, and write your description verbatim if possible.
5. How should you describe life-sustaining treatment? Please be specific.
6. Should you describe what happens after a cardiac arrest (assuming resuscitation was successful)? Why or why not?
a. If you do describe what happens after a cardiac arrest, what do you feel you should include in this description? Please be very specific.
7. How should a patients condition, wishes and beliefs affect the content and=or the framing of the discussion regarding
life-sustaining treatments and CPR?
8. Should you attempt to give a prognosis, or tell patients the likelihood of recovery following a code blue or life-sustaining measures?
If so, please indicate how you should explain this to patients.
9. Should you offer a recommendation to your patients about whether or not they should be a full code, or whether or not they
should receive life support?
10. Are there any other considerations that should significantly modify the usual content of your conversation when you are discussing
code status and life support (e.g., culture, language barrier, age, diagnosis, acuity of illness, patient request)?
CPR, cardiopulmonary resuscitation.

(continued)

192

DOWNAR AND HAWRYLUCK

Appendix C. Delphi Round 1 Coding Scheme
Subtheme

Statement #1
When to discuss Code
Status or life-sustaining
treatment with a patient

Sample quotation

Before illness

Process should begin with family physician when death not imminent.

Any admission to hospital

This should be discussed with all acute care admissions especially those
coming to the ICU, and should be recorded formally on the chart.
Whenever it is recognized that a patient has an acute or chronic potentially
life-threatening health condition.
Changes in the patients condition result in significant revisions of the
treatment plan.
When a patient or family has decided to stop further active intervention
and treatment, it is incumbent to discuss code status and clarify which
other treatments would be stopped and which should be continued.
Sometimes the initial palliative consult meeting is not the best time for this
discussion as no relationship of trust between physician and patient has
yet been established and often patients are feeling as if they have been
given up on or dismissed by their oncologist when they are referred to
palliative care.

Diagnosis of serious illness

Clinical deterioration
Change in treatment
strategy
Not always on first
meeting

Statement #2
Theme
How to frame a
Goals and Possibilities
discussion of CPR=LST

Current status and

expectations

Subtheme
Goals of care
Patient values in the
context of medical
possibilities
Prognosis (if poor)

Baseline function

Statement #3
Should LST be discussed Yes
separately from CPR?

No

Maybe
Statement #4
How should you describe Patient-centred
a cardiac arrest?
questions
Physiologic description
How should you describe Attempt at resuscitation
CPR?
Literal description

How should you describe General organization

LST?

Distinct issues

Continuum

Quotation
This should always be framed as part of a discussion
about patient values and goals.
Clarify goals of carethis may require a negotiation of
what is possible
There are several circumstances where the patients and
familys values, although respected, will not influence
the decision on life sustaining therapy and CPR. These
circumstances include those where death is predicted
with very near 100% certainty.
The initial discussion should revolve around the
functional status of the patient prior to the current
admission.
Life support is not a single entity, but comprises multiple
acts and actions . . . all of these may be appropriate and
reasonable. But CPR may not be appropriate and
reasonable and the decision process is different.
Even if they represent a different issue, from my personal
point of view, they are different alternatives on a
continuum.

Depends on goals of care

Start by asking what the
patient knows about
cardiac arrest
Explain that the heart or
breathing stops

I would begin by finding out what they already know

and if they want me to explain.

I use the term ATTEMPT AT RESUSCITATION.

Chest compressions
Tube down throat
Medications for heart
Defibrillation
Patient-centred questions To explain life-sustaining treatment, I would again start
by asking what they know
Basic vs. Advanced
Basic life supports would include fluids, nutrition,
standard medications and treatments for
medical=surgical conditions. Advanced life supports
would include resuscitation in the event of cardiac
arrest, ICU admission, aggressive life supports
including ventilation, dialysis, hemodynamic support

(continued)

HOW SHOULD WE ASK ABOUT CODE STATUS?

193

Appendix C. (Continued)
Theme
System-based
How should you
describe palliative
care?
Statement #5
Should you describe
what happens after a
cardiac arrest?

Yes

Subtheme
Listing organ systems
and supports
Active care while dying
process occurs

No

Not necessary after

values discussion

Limitations of LST

Neurological injury

Functional impairment

Does not affect

underlying illness
Statement #6
How does the patients
condition, wishes,
and beliefs affect the
content=framing of
the discussion?

Statement #7
Should you give a
prognosis? How?

General approach

Yes, in general as a principle of consentdisclosure is

always necessary.
Not usually necessary to discuss. If I have done my
job correctly and taken a careful values history from
the patient or SDM, we will have come to an
understanding of the appropriate use of life
sustaining technologies. Going into more detail would
sound like I was trying to talk the patient out of
something . . .
If your heart stop, we might be able to start back,
but you might be left with severe neurological
damage (vegetable state or with important functional
deficit).
If it is successful, you is likely that you will be weaker
than you were before and you may never leave the
ICU.
Because all of my patients are end stage cancer patients, I
also discuss that no matter what the outcome, that the
cancer will still be there and is not reversible

Patient-centred questions

I would begin by finding out what they already know

and if they want me to explain

Similar approach to all

Cultural differences and language can certainly make the

discussion more challenging but I do not think they
should alter the content
In situations where in the context of the illness, CPR
would not likely achieve any benefit or be successful,
the discussion would be more directive at not pursuing
CPR than in situations where CPR is an appropriate
decision.

Limited by medical
reality

Dont always offer

CPR=ICU

Sensitive to context

Less detailed if patients

condition or
language barrier
mandates this
Cultural sensitivity

Guided by patient
concerns

I tell them that it is part of my routine practice to be

certain you know that as we provide comfort while
your loved one is dying, as the goal of care, and are
allowing a natural death to occur

Values

Ethical obligation

How?

Sample quotation

Patient-centred
questioning

Focus on goals of care

Simple estimates

Culture affects language, degree of explicitness and

meaning of illness especially
Yes, it is important that within the Goals of Care
discussion, reversibility of disease and resuscitative
efforts are framed in terms of your best information and
using the patients own language regarding return to an
acceptable, functional lifestyle.
Of course, patients need to know the chances of being
able to achieve the goals of care. Most of the time, this
leads to a conversation about time limits on treatments
or other health conditions which would alter the
patients choice about continuing with life sustaining
treatments.
I would inform them that the likelihood of success is low
or very low (depending on the clinical situation.

(continued)

194

DOWNAR AND HAWRYLUCK

Appendix C. (Continued)
Theme

Subtheme
Precise numbers
Explain what determines
prognosis

Manage expectationsmay not reflect reality

Statement #8
Should you offer a
recommendation?

Support=guidance

Guidance

Assume burden of
decision

Physician
interpretation

Respect for autonomy

Statement #9
Other
considerations=
comments

Communication
techniques

Sample quotation
In your case, the chance that this would be successful is
%
Age, diagnosis, acuity of illness may all impact the
success and subsequent recovery from a
resuscitation and this should be communicated to
the patient, otherwise how can they make an
informed decision.
I might indicate that CPR outcomes are quite different
when presented on TV with a high success rate
compared to those who are sick in hospital.

Patients and families are NOT health care professionals

for the most part. They need guidance with respect to
the medical aspects . . . . Given what you have told me
about your prior quality of life, values, etc., and given
xyz about your condition and prognosis, my
recommendation is . . . .
This is particularly important in dealing with families
since in many cases it is an added burden to leave the
decision up to the family without our
recommendation.

When requested
Interpret patient
Yes. This is based on an understanding of the patients
values to form
wishes and the physicians nuanced judgment of the
a plan
likely outcome of the treatment plan.
Interpret clinical situation It depends on the situation e.g. for patients with
to form a plan
terminal=end-stage illness: In your present state
of health, from my personal experience and from
the evidence of many studies, the risks to your
quality of life are actually greater than
the benefits.
Resolve inconsistencies
I disclose what my understanding is of their disease, and
between goals and
what is reversible and if they choose a full code after
plan of care
telling me they know they can not recover, I simple,
restate both points and ask them to help me understand
their decisions
Recommendations do not It is a mistake to think that respect for patient autonomy
undermine autonomy
means not giving recommendations to patients . . . . If
done skillfully and respectfully, I think one can fully
respect patients autonomy while at the same time
offering a recommendation
Checking in

Communication skill

Avoid agenda

Difference between
hoping and planning
Communicate concern
about pt dying

Check repeatedly for need for clarification and

understanding, and that we give as much time as they
need.
So much depends on the communication skills of the
person speaking with patients and families and not only
on the content.
Very importantly, when having these discussions I do not
start with a pre-determined decision of what I want the
out come of the discussion to be. Rather, I want to
engage in a meaningful dialogue that builds a trusting
relationship.
Hope for the best, but prepare for the worst
Also, I forewarn patients=families=POA=SDM, that I am
afraid we may need to discuss these issues, and ask
them to start thinking about the goals they would want
met should the clinical condition deteriorate.

(continued)

HOW SHOULD WE ASK ABOUT CODE STATUS?

195

Appendix C. (Continued)
Theme

Subtheme
Get whole picture

Explore concerns

ICU, intensive care unit; LST, life-sustaining therapy.

Sample quotation
Understanding the patient, their familys (regardless of
the patients competence) understanding of the
situation. As well, I discuss the situation with other
health care providers to understand treatment options
and expected reversibility of disease
If they are completely unrealistic, but otherwise thinking
clearly, I gently point out the discrepancy and ask them
to explain their logic.
Very interesting answers come back especially after
they gain trust in you. For example they may say Well,
really I would never want to live on a ventilator, but if I
tell you that now, then you will stop caring about me.

This article has been cited by:

1. James Downar, Tracy Luk, Robert W. Sibbald, Tatiana Santini, Joseph Mikhael, Hershl Berman, Laura Hawryluck. 2011. Why
Do Patients Agree to a Do Not Resuscitate or Full Code Order? Perspectives of Medical Inpatients. Journal of General Internal
Medicine . [CrossRef]
2. 2010. Bibliography. Progress in Palliative Care 18:2, 99-128. [CrossRef]