Cases & Commentaries

To Resuscitate or Not?

Commentary By Albert W. Wu, MD, MPH; Peter J. Pronovost, MD, PhD

The Case

A critically ill end-stage AIDS patient was
hospitalized for end-of-life care. Given the state of his disease,
his code status was Do Not Resuscitate/Do Not Intubate (DNR/DNI),
though he was still receiving active care. The patient was
suffering from multiple concurrent infections requiring intravenous

The patient in the room next to him was also
receiving antibiotics. Inadvertently, the antibiotics were switched
at the nursing station, resulting in the administration of
Nafcillin (a penicillin-like antibiotic) to the AIDS patient, who
in the past had a known anaphylactic-shock reaction to penicillins.
The difficulty was, if the AIDS patient developed anaphylaxis to
the Nafcillin, there was a good chance he would suffer a
cardiopulmonary arrest, given the advanced state of his illness. If
this were to occur, he would likely die if heroic measures were not

The nurse taking care of both patients realized
the error after she went to give the other patient his antibiotics
and saw that the printed label did not match the patient′s
name. Unfortunately, the wrong medication had already been given to
the AIDS patient. After noticing the error, the nurse called the
covering physician. The physician spoke to the patient, advised him
that he might possibly suffer anaphylaxis to the errant medication,
and asked him if he wanted to maintain his code status as

The patient was treated prophylactically for the
possibility of allergic/anaphylactic reaction. For the next several
hours, his nursing care became more intensive with more frequent
vital sign checks. Luckily, he did not suffer any adverse outcome
from the medication. By the next morning, it was clear that there
were no sequelae.

The Commentary

Making an error that injures a patient causes
many clinicians the greatest distress of their professional
careers. Clinicians may feel that they have failed and may
experience shock, grief, shame, and guilt. They may also fear the
consequences that the error may have for them, including loss of
status, disciplinary action, or malpractice litigation.(1)
Typical coping mechanisms include hypervigilance, characterized by
excessive attention in what may be a vain attempt to undo, atone
for, or make up for the harm done. In one case, a provider who gave
his patient an overdose of a medication described his response as
′36 hours of rapt attention.′(2)

In some cases, our reactions to our errors may
alter usual clinical decision-making. The cardiac transplant
patient at Duke who received a heart-lung transplant with an
incompatible blood type is illustrative. The patient, clearly dying
as her initial mismatched transplant failed, ultimately received a
second transplant even though the prospects of success were remote.
This action could be interpreted as the clinicians′ attempt
to right their wrong, or to give a second chance to a patient who
had been profoundly unlucky. However, this decision may have
deprived another patient with a greater chance of benefiting from
the organ. One could argue that the clinical decision was flawed
because the clinicians were trying so hard to ′make up′
for their error.

We believe the second transplant at Duke may
represent a more widespread phenomenon than has been previously
appreciated, in which patients receive more aggressive care after
an iatrogenic injury than they would have otherwise received. There
are no data on the incidence of this phenomenon. However, research
suggests that physicians may feel a greater obligation to treat
when an illness results from the physician′s error.(3)
They may also fear that they will be more likely to be sued if they
do not take all possible measures. Unfortunately, more is not
always better, and not always consistent with patients′

If the patient in this case had arrested, the
decision to attempt resuscitation should still have been based on
the patient′s wishes, rather than on the clinician′s
remorse. An ethical analysis suggests that the admonition to
′do no harm′ does not override an autonomous
patient′s refusal of treatment.(4) Nor should the patient be subjected to an undesired
resuscitation as part of a defensive strategy to thwart litigation.
Physicians may believe that DNR orders do not apply in the case of
iatrogenic cardiac arrests. In one survey, physicians responded
that they would be more likely to override a DNR order when an
arrest was caused by a complication of treatment.(3)

When a patient′s detailed wishes are known,
the decision is clear. A thornier issue arises when it is not known
what a patient would want in a specific situation, including
iatrogenesis. For example, what if this patient had suffered an
immediate anaphylactic reaction, and his wishes regarding an
iatrogenic arrest were not known (as they usually will not be)? In
the absence of such knowledge, the clinician may feel obligated to
treat immediate life threats and then obtain additional information
from the patient or a surrogate.

Physicians may believe that patients are not
thinking about the possibility of an iatrogenic cardiac arrest when
they consider resuscitation preference. This problem could be
alleviated by eliciting more detailed preferences at the time of
DNR discussions. Patients should understand that DNR orders
generally apply to arrests in which the probability of successful
resuscitation may be either high or low, depending on the clinical
circumstances. They should also understand that arrests can be
expected or unexpected, and may even be caused by medical care.
Incidents that put patients at risk are frighteningly common,
particularly for severely ill patients. Research suggests that
virtually all patients in intensive care units experience
potentially life-threatening errors in their care.(5,6) Patients can be asked directly how they would like
their care handled in each of these situations, some of which could
be acute and reversible. For seriously ill patients, the
circumstances of ′double effect′
could also be mentioned: ′What if I gave you a usual dose of pain
medication, and it made you stop breathing—would you still
not want us to try to resuscitate you?′ The distinction should be
drawn that a DNR order is not tantamount to
abandonment—something that many patients fear. It should be
emphasized that the team will continue to provide care and caring
under all circumstances.

One helpful way to structure these discussions
involves four steps: discuss prognosis with the patient or family,
evaluate the extent to which patients or families want to be
involved in decision making, establish goals based on the
prognosis, and implement therapies that help achieve the goal and
eliminate those that do not. Immediate life threats, if treated,
may not affect prognosis. This process is iterative and decisions
need to be revised frequently. For example, if the patient desires,
physicians may acutely treat an anaphylactic reaction from a
medication error. However, if the event altered the patient′s
prognosis (eg, if after several days the patient was still on
mechanical ventilation and receiving vasopressor therapy), the
goals should be reevaluated.

We are incorporating these four steps (establish
prognosis, establish degree of involvement in decision-making, set
goals, revise therapies to achieve the goals) in a checklist that
is part of our ′ Daily Goals Form. ′ (7) We are pilot testing it on patients in the ICU longer
than 5 days, since the majority of these patients will not be
discharged home. Checklists such as these can help ensure that key
steps in the process are reliably accomplished. Nevertheless, these
discussions should ideally occur with the patient ′ s primary
care physician—who has a better understanding of the
patient′s values and wishes—prior to an acute

Take-Home Points

  • Prospective discussions about what
    may occur as a patient's illness progresses reduce the likelihood
    of surprises. When patients understand the potential course of
    events and prognosis, it is easier to establish the goals of
  • Hope
    for the best, but plan for the worst. Discuss with patients their
    preferences about resuscitation attempts, including in
    circumstances that might be reversible and iatrogenic. Otherwise,
    if the patient's wishes are unclear and an immediate life threat
    occurs, it may be prudent to treat the patient and continue
    discussions when the crisis is resolved.
  • A daily checklist may
    help facilitate these discussions: establish prognosis, establish
    degree of involvement in decision-making, set goals, and revise
    therapies to achieve the goals.
  • Carry out an ongoing discussion about what is
    happening with the patient's health and
  • In end-of-life care,
    the patient's wishes remain

Albert W. Wu, MD, MPH
Associate Professor, Health Policy and Management
Johns Hopkins Bloomberg School of Public

Peter J. Pronovost, MD,
Associate Professor, Anesthesiology and Critical Care Medicine,
Surgery and Public Health
The Johns Hopkins University School of Medicine


1. Wu AW. Medical
error: the second victim. The doctor who makes the mistake needs
help too. BMJ. 2000;320:726-7.[ go to PubMed

2. Wu AW, Folkman S,
McPhee SJ, Lo B. Do house officers learn from their mistakes? JAMA.
1991;265:2089-94.[ go to PubMed

3. Casarett DJ,
Stocking CB, Siegler M. Would physicians override a
do-not-resuscitate order when a cardiac arrest is iatrogenic? J Gen
Intern Med. 1999;14:35-8.[ go to PubMed

4. Casarett D, Ross
LF. Overriding a patient′s refusal of treatment after an
iatrogenic complication. N Engl J Med. 1997;336:1908-10.[ go to PubMed

5. Donchin Y, Gopher
D, Olin M, et al. A look into the nature and causes of human errors
in the intensive care unit. Crit Care Med. 1995;23:294-300.[ go to PubMed

6. Andrews LB,
Stocking C, Krizek T, et al. An alternative strategy for studying
adverse events in medical care. Lancet. 1997;349:309-13.[ go to PubMed

7. Pronovost P,
Berenholtz S, Dorman T, Lipsett PA, Simmonds T, Haraden C.
Improving communication in the ICU using daily goals. J Crit Care.
2003;18:71-5.[ go to PubMed