Silveira MJ, DiPiero A, Gerrity MS, Feudtner C. Patients' Knowledge of Options at the End of LifeIgnorance in the Face of Death. JAMA. 2000;284(19):2483-2488. doi:10.1001/jama.284.19.2483
Author Affiliations: The Robert Wood Johnson Clinical Scholars Program at the University of Washington, Seattle (Drs Silveira and Feudtner); and the Department of General Internal Medicine at Oregon Health Sciences University, Portland (Drs DiPiero and Gerrity).
Context Effectiveness of legislation promoting advance directives and legalizing
physician-assisted suicide depends on patients' understanding their legal
options about end-of-life care. However, outpatients' understanding of their
legal options at the end of life has not been studied.
Objectives To estimate the percentage of outpatients who are informed about 4 areas
relevant to end-of-life care: refusal and withdrawal of lifesaving treatments,
physician-assisted suicide, active euthanasia, and double effect; and to determine
whether authoring advance directives, experiencing illness, acting as a proxy
for health care decisions, and caring for an ill loved one are associated
with better knowledge in end-of-life care.
Design Cross-sectional survey.
Setting and Participants One thousand consecutive English-speaking, adult patients attending
1 university-based internal medicine clinic and 3 community-based, university-affiliated,
mixed internal medicine and family practice clinics in Oregon during May and
Main Outcome Measures Percentage of correct responses in the 4 topic areas and total knowledge
score, adjusted for demographic (eg, age, race, educational level, income
level, marital status) and experiential (eg, health, proxy decision making,
advance directives, and death of a loved one) factors.
Results Of the 1000 patients invited to participate, 728 (73%) consented and
completed the questionnaire and were included in the analysis. A total of
69% of respondents answered correctly regarding refusal of treatment, 46%
for withdrawal of treatment, 23% for assisted suicide, 32% for active euthanasia,
and 41% for double effect. Sixty-two percent of respondents did not distinguish
between assisted suicide and euthanasia. After adjustment for other covariates,
better knowledge was significantly associated with white race (odds ratio
[OR], 2.3; 95% confidence interval [CI], 1.3-4.2), having at least a college
degree (OR, 3.0; 95% CI, 1.4-6.7), and having been a proxy for health care
decisions (OR, 1.8; 95% CI, 1.2-2.6). Personal experience with illness (OR,
1.0; 95% CI, 0.6-1.5), death or illness of a loved one (OR, 1.6; 95% CI, 1.0-2.7),
and authoring an advance directive (OR,1.3; 95% CI, 0.9-2.0) were not associated
with better knowledge.
Conclusions A significant proportion of outpatients at university-affiliated clinics
in Oregon appear to misunderstand options in end-of-life care. Our results
suggest that greater public knowledge about end-of-life care is needed, and
advance care planning must be preceded by education about options in end-of-life
For the past 25 years, US citizens have sought increased control over
their experience of dying.1,2
In the courts, they have ensured their rights to refuse and withdraw life-sustaining
treatment, while in the legislature, they have had laws passed to promote
the use of advance directives (the 1990 Patient Self-Determination Act)3 and to legalize physician-assisted suicide (the 1994
Oregon Death with Dignity Act).4
A firm belief in informed consent has justified and legitimated these
efforts to enhance patients' control over end-of-life care. For example, the
Patient Self-Determination Act aims to create "clear and convincing evidence"
regarding every admitted patient's end-of-life care choices, assuming that
these anticipatory choices would be fully informed.5- 7
Similarly, the Oregon Death with Dignity Act allows terminally ill patients
to request lethal prescriptions only "after being fully informed by the attending
physician of . . . the feasible alternatives, including, but not limited to,
comfort care, hospice care and pain control."4
Rarely has the US populace assumed so strongly that patients would be
well informed regarding all care options before making a health care decision,
yet this assumption may not be warranted. If patients mirror the general public,
then many are misinformed about terminal palliative care. A national poll
conducted by the American Medical Association8
in 1997 found that 40% of respondents did not know it is legal to give pain
medicine that could have the additional effect of hastening death (double
effect), and 35% were not familiar with the terms hospice or palliative care. Furthermore, studies
suggest that some patients do not distinguish between physician-assisted suicide
Therefore, we sought to assess outpatients' knowledge of their legal
options near the end of life. Our primary goal was to estimate the percentage
of outpatients who are informed about 4 areas relevant to end-of-life care:
refusal and withdrawal of life-saving treatments, physician-assisted suicide,
active euthanasia, and the doctrine of double effect. Our secondary goal was
to test whether authoring advance directives, experiencing personal illness,
acting as a proxy for health care decisions, and caring for an ill loved one
are associated with better knowledge in end-of-life care. We chose to study
outpatients specifically because they represent a convenient population at
risk of hospitalization who, by virtue of their access to a physician, can
be educated about and encouraged to participate in advance care planning.
As stated by Linda Emanuel,10 "[Advance care]
planning with young and healthy patients may be considered analogous to screening
for illness: the chance of illness is low, but the effort may be worth it."
Moreover, studies have shown that patients prefer to discuss advance directives
early in the patient-physician relationship11- 14
and that advance care planning can be effectively introduced in the outpatient
During May and June of 1999, 1000 consecutive English-speaking outpatients
18 years or older who were checking in for their clinic appointment were invited
independently to participate in the study. All qualified patients were offered
the self-administered survey during check-in at 4 primary care sites: 1 internal
medicine clinic at Oregon Health Sciences University in Portland (400 patients
approached) and 3 mixed, internal medicine and family practice clinics affiliated
with Oregon Health Sciences University in Beaverton, Sellwood, and Tigard
(200 solicitations each). The sample size was sufficiently large to yield
95% confidence intervals (CIs) of ± 3% or less around prevalence estimates.
Throughout the study period, occasional audits of the distribution and
collection process were conducted to ensure proper selection of patients and
full confidentiality. The study was approved by the Oregon Health Sciences
University institutional review board, which considered completion of the
survey as giving informed consent.
The survey was developed by one of us (M.J.S.), written at a sixth-grade
reading level, and then pilot tested for clarity with a group of 20 randomly
selected outpatients from the internal medicine clinic at Oregon Health Sciences
University. The questionnaire presented a sequence of events in an unfolding
clinical vignette that traced the medical saga of "John," a hypothetical patient
(see "Survey Vignettes and
Questions" box). John eventually dies of cancer
but not before his case raises several issues and presents dilemmas regarding
refusal of treatment, withdrawal of treatment, euthanasia, assisted suicide,
and double effect. The vignettes did not use the terms assisted suicide or euthanasia to avoid the
ambiguity and emotional reaction that these terms might provoke. After each
portion of the vignette, subjects were posed questions in the following format:
"In Oregon, does a patient who has less than six months to live have the legal
right to . . . ?" Subjects could answer, "Yes in all cases," "Yes in some
cases," "No," or "Don't know."
Subjects were instructed to read the vignettes that follow and answer
the 9 questions by circling 1 of the following 4 answers: yes in all cases;
yes in some cases; no, no legal right; and don't know.
John has cancer. He has been told that he has less than 6 months to
live and is terminally ill. His cancer doctor offers him a form of medical
treatment, called "Chemotherapy," that may cure him. He decides that he does
not want the treatment.
1. In Oregon, does a patient like John, who has less
than 6 months to live, have the legal right to refuse treatment that might
cure him or save his life?
John does not get any treatment and the cancer worsens. At one point,
John loses the energy to eat or drink on his own. If John continues without
food or water, he will soon die. His physician offers him intravenous fluids
(water through the veins) and a feeding tube (a tube that goes directly into
the stomach and allows the physician to feed him). John is still able to speak
for himself and make his own decisions. He decides that he does not want the
intravenous fluids or the feeding tube.
2. In Oregon, does a patient like John, who has less
than 6 months to live, have the legal right to refuse intravenous fluids even
if it means that his death will occur sooner?
3. In Oregon, does a patient like John, with less
than 6 months to live, have the legal right to refuse a feeding tube even
if it means that his death will occur sooner?
John rethinks his situation and changes his mind. He decides that he
does want the intravenous fluids and the feeding tube. His doctor orders these
and, for a brief time, John grows stronger. One morning, a nurse finds John
unconscious and not breathing. Physicians successfully bring John back to
life but only after putting him on a ventilator (a machine that breathes for
John by pumping air into his lungs). John now needs the ventilator to live.
He stays on the ventilator for a week without much improvement. He is conscious
but very ill. Finally, he writes a note to his doctor saying, "Turn off the
4. In Oregon, is it legal for a physician to turn
off a ventilator when requested by a patient like John, with less than 6 months
to live, even if it means the patient might die?
The ventilator is turned off, but John miraculously lives. He is not
expected to live long, but he cannot bear his illness any longer. He asks
his physician to help him die. He asks, "Doc, give me a shot or put something
through the IV to make me die right away."
5. In Oregon, is it legal for a physician to inject
a medication that would cause a patient's immediate death if a patient like
John, with less than 6 months to live, requests it?
6. In Oregon, is it legal for a physician to refuse
to inject a medication that would cause a patient's immediate death if a patient
like John, with less than 6 months to live, requests it?
John's doctor refuses to inject John with a medication that would cause
him to die. John asks if his doctor would instead prescribe a medication that
John could give to himself to end his life. He asks, "Can't you prescribe
something I can take to die? Don't worry, I could take it without anyone's
7. In Oregon, is it legal for a physician to prescribe
a medication and provide advice so that the patient can take medications that
will end his life if a patient like John, with less than 6 months to live,
8. In Oregon, can a physician refuse to prescribe
a medication and provide advice so that a patient like John, with less than
6 months to live, can take medications that will end his life if the patient
John's physician refuses his request, but promises John that she will
do everything in her power to keep him comfortable. She puts John on a painkiller.
John is very sleepy but appears restless and uncomfortable. She increases
the dose of the painkiller to comfort him even though she knows that the painkiller
at high doses can cause patients to stop breathing. Finally John looks comfortable,
but suddenly he stops breathing. He dies peacefully soon thereafter.
9. In Oregon, is it legal for a physician to give
patients like John, with less than 6 months to live, pain medications with
the goal of relieving pain and suffering even if death may occur sooner as
The questionnaire inquired about the participants' age, sex, marital
status, employment status, income level, ethnicity, religion, educational
level, and health status. The questionnaire also assessed experience with
end-of-life dilemmas by inquiring about the following experiential factors:
personal illness, authoring an advance directive, experiencing the death or
grave illness of a loved one, and participating in proxy decision making.
Finally, respondents rated the clarity of the questions as well as their confidence
level regarding their answers on scales of 1 to 5.
After calculating the percentage of correct responses to the 4 topic
areas, we generated a total knowledge score for each respondent according
to a point system whereby fully correct answers were awarded 1 point, partially
correct answers were given a half point, incorrect answers had a point deducted,
and anything else (including "Don't know") had no value. With this formula,
knowledge scores could range from −8 to 8. A respondent was considered
knowledgeable if his/her score was above the median for the survey population
Demographic factors associated with being knowledgeable (ie, scoring
above the median) were initially identified using χ2 testing.
The strength and independence of the associations was tested using a stepwise
logistic regression procedure whereby a parsimonious set of strong independent
predictors of knowledge were identified from among the set of demographic
variables (eg, race, educational level, income level, and marital status).
Using that model, each experiential factor (such as health, proxy decision
making, authoring advance directives, and experiencing the death of a loved
one) was examined separately while adjusting for the relevant sociodemographic
variables using logistic regression. Results were verified using ordered logistic
regression. Statistical analysis was conducted using STATA 6.0 software (Stata
Corp, College Station, Tex).
Of the initial 1000 outpatients approached, 728 (73%) consented, qualified
to participate, and completed the questionnaire. Two hundred sixty-eight refused
to participate, mistakenly were not offered the survey, or failed to return
the survey. No information was collected on nonrespondents. Subsequently,
2 respondents who were younger than 18 years and 2 who had companions complete
the survey on their behalf were excluded as ineligible, leaving a total of
728. These outpatients ranged from ages 18 to 99 years (median age, 46 years),
were predominantly white (88%), and spoke English as a first language (95%)
(Table 1). Sixty-four percent
were Christian; 37% worshipped regularly. Their median income was between
$20,000 and $39,999. Most subjects were high school educated with many reporting
some college education.
Many respondents had prior personal experience with illness; 21% rated
their health as "poor." One third had authored an advance directive. Many
respondents had had vicarious experience with end-of-life issues, with 87%
having experienced the death or grave illness of a loved one, and 26% having
acted as proxy decision makers at some point.
In the 3 separate segments of the vignette depicting withdrawal of care,
approximately two thirds of respondents correctly understood that competent
patients in Oregon can legally refuse life-saving or life-sustaining treatment
in all cases (Table 2). Respondents
had less knowledge about withdrawal of treatment, with only 46% correctly
identifying that competent patients in Oregon can legally withdraw life-sustaining
treatment in all cases. Only 23% of respondents correctly identified assisted
suicide (defined as "prescribing a medication and providing advice so that
a patient can take medications that will end his life") as a legal option
for competent terminally ill adults in Oregon in all cases. Thirty-two percent
of respondents recognized that active euthanasia (portrayed as "injecting
a medication that would cause a patient's immediate death") is illegal. Forty-one
percent of respondents recognized that double effect (described as "giving
pain medications with the goal of relieving pain and suffering even if death
may occur as a result") is legal in Oregon. When asked if physicians could
legally refuse to satisfy a patient's request for assistance in dying, 340
(48%) of 708 correctly answered that physicians could refuse.
Many respondents did not distinguish between assisted suicide and active
euthanasia. For example, of those who correctly believed that assisted suicide
is a legal option for terminally ill competent adults in Oregon, 262 (64%)
of 411 respondents believed that active euthanasia is legal as well. A comparison
of responses for the assisted suicide and euthanasia questions (Table 3) demonstrated that many respondents gave similar answers
for both questions (eg, most respondents believed that both assisted suicide
and active euthanasia are legal when, in fact, only assisted suicide is).
Only 127 (18%) of 705 respondents answered both questions correctly. Finally,
38 (5%) of 705 answered that euthanasia was legal and assisted suicide was
Despite the frequency of "Don't know" and incorrect responses, 626 (89%)
of 703 respondents indicated that the survey was clearly worded and 440 (63%)
of 700 were confident in their responses. Having confidence was associated
with respondents having experienced the illness or death of a loved one (P<.01) or having acted as a proxy for health care decisions
(P<.01). Authoring a living will (P = .51), experiencing personal illness (P
= .14), or being well educated (P = .80) was not
associated with confidence.
Total knowledge scores were significantly higher for respondents who
were white, married, aged at least 30 years, English-speaking, college-educated,
and Jewish, but after multivariable adjustment, knowledge was independently
associated only with white race and college education (Table 4).
We then examined, after adjusting for demographic variables, 4 experiential
factors: personal illness, death or illness of a loved one, proxy decision
making, and authoring an advance directive. Among these markers of personal
life history, only the experience of proxy decision making was independently
and significantly associated with accurate knowledge about options in end-of-life
care (odds ratio [OR], 1.8; 95% CI, 1.2-2.6). Experience with death or illness
of a loved one showed a trend toward significance (OR, 1.6; 95% CI, 1.0-2.7).
Interestingly, personal experience with illness (OR, 1.0; 95% CI, 0.6-1.5)
and authoring an advance directive (OR, 1.3; 95% CI, 0.9-2.0) were not significantly
associated with better knowledge about options in end-of-life care. A secondary
analysis of the data, categorizing knowledge score into quartiles and then
using ordered logistic regression, revealed death or illness of a loved one
to be a statistically significant association (P
= .008); the remaining results did not change.
We found that a significant proportion of outpatients surveyed in Oregon
appears to misunderstand patients' options in end-of-life care. While most
outpatients know that they can refuse life-sustaining medical care, many do
not understand that life support can be withdrawn once started or that pain
can be managed aggressively by increasing medication doses into the realm
of a double effect. In addition, many outpatients appear not to distinguish
between assisted suicide and active euthanasia.
These findings should be interpreted with several caveats. First, although
our survey's instructions specified that all consecutive patients were to
receive a questionnaire, a relatively small proportion was not offered the
questionnaire, possibly introducing some selection bias. Second, we do not
have information on nonresponders. Third, we may have assessed respondents'
opinions rather than their knowledge even though the survey was intentionally
worded to avoid this problem. Finally, we surveyed only outpatients in Oregon.
While this feature potentially limits generalizibility, targeting Oregonians
with the state's unique legislation and surrounding publicity puts these findings
into sharp relief. The ignorance regarding end-of-life care options found
in this survey—especially following so closely on the heels of the 1997
referendum and surrounding media campaign that reaffirmed the Oregon Death
with Dignity Act—suggests that even extraordinary public discussion
regarding end-of-life issues fails to educate typical outpatients about their
basic health care rights. If borne out in other settings, the high prevalence
of misunderstanding found among these patients underscores the need for more
effective physician-directed discussion about end-of-life care.18- 21
How might more effective discussion about end-of-life care occur? Our
findings raise several points that should be considered in formulating a response.
First, no group of patients knows enough about end-of-life care (if our goal
is to inform all patients fully), not even those with better socioeconomic
situations or higher educational levels. Second, having patients go through
the exercise of composing or signing an advance directive appears an insufficient
means of educating patients about options in end-of-life care. Third, authoring
advance directives does not improve patients' comfort with end-of-life deliberations.
Fourth, patients may be more likely to learn about options in end-of-life
care when a loved one is in peril than when they themselves are threatened.
In light of these findings, physicians might consider discussing end-of-life
issues with all patients, using a questionnaire or structured interview to
identify knowledge deficits, then educating accordingly. Furthermore, besides
using hypothetical situations involving the patients themselves, physicians
might examine end-of-life care options with patients by exploring their prior
experiences with a loved one's illness or death (87% of our study population
had such experiences).
In conclusion, advance directives generally and legalized physician-assisted
suicide specifically make it imperative that people properly understand options
in end-of-life care. Our study suggests that the majority of patients do not
understand them. Not only should physicians consider this issue as they discuss
end-of-life issues with patients, but legislators and policymakers must also
consider how well US residents comprehend options in end-of-life care because
this level of comprehension may affect how patients interact with the health
care system and how citizens vote on measures such as the Oregon Death with