Clinical Investigations
Barriers and Facilitators to Initiating and
Completing Time-Limited Trials in Critical Care*
Courtenay R. Bruce, JD, MA1,2; Cecilia Liang, BA1; Jennifer S. Blumenthal-Barby, PhD, MA1;
Janice Zimmerman, MD3,4; Andrea Downey, MS, RN5; Linda Pham, MSSW, LCSW6;
Lisette Theriot, LMSW6; Estevan D. Delgado, BA1; Douglas White, MD, MAS7
Objectives: A time-limited trial is an agreement between clinicians
and patients or surrogate decision makers to use medical therapies
over a defined period of time to see if the patient improves or deteri-
orates according to agreed-upon clinical milestones. Although time-
limited trials are broadly advocated, there is little empirical evidence
of the benefits and risks of time-limited trials, when they are initiated,
when and why they succeed or fail, and what facilitates completion
of them. Our study objectives were to 1) identify the purposes for
which clinicians use time-limited trials and 2) identify barriers and
facilitators to initiating and completing time-limited trials.
Design: Semistructured interviews: We analyzed interviews using
qualitative description with constant comparative techniques.
Setting: Nine hundred-bed, academic, tertiary hospital in Hous-
ton, Texas. Interviewees were from open medical, surgical, neuro-
surgical, and cardiovascular ICUs.
Subjects: Thirty healthcare professionals were interviewed (nine
surgeons, 16 intensivists, three nurse practitioners, and two
“other” clinicians).
*See also p. 2676.
1
Department of Medicine, Center for Medical Ethics and Health Policy,
Baylor College of Medicine, Houston, TX.
2
Department of Medicine, Houston Methodist Hospital System, Houston
Methodist Hospital System Biomedical Ethics Program, Houston, TX.
3
Weill Cornell Medical College, New York, NY.
4
Division of Critical Care, Department of Medicine, Houston Methodist
Hospital, Houston, TX.
5
Division of Palliative Care, Department of Nursing, Houston Methodist
Hospital, Houston, TX.
6
Division of Social Work, Department of Social Work and Case Manage-
ment, Houston Methodist Hospital, Houston, TX.
7
Department of Critical Care Medicine, Program on Ethics and Decision Mak-
ing in Critical Illness, University of Pittsburgh Medical Center, Pittsburg, PA.
Supplemental digital content is available for this article. Direct URL citations
appear in the printed text and are provided in the HTML and PDF versions
of this article on the journal’s website (http://journals.lww.com/ccmjournal).
Dr. Zimmerman served as a board member for Decisio Health and is
employed by Houston Methodist. The remaining authors have disclosed
that they do not have any potential conflicts of interest.
For information regarding this article, E-mail: crbruce@bcm.edu
Copyright © 2015 by the Society of Critical Care Medicine and Wolters
Kluwer Health, Inc. All Rights Reserved.
DOI: 10.1097/CCM.0000000000001307
Critical Care Medicine www.ccmjournal.org 2535
Copyright © 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Interventions: None.
Measurements and Main Results: Interviewees reported initiating
time-limited trials for three different purposes: to prepare surrogates
and clinicians for discussion and possible shifts toward comfort-care
only therapies, build consensus, and refine prognostic information.
The main barriers to initiating time-limited trials involve clinicians’ or
surrogate decision makers’ disagreement on setting a time limit.
Barriers to completing time-limited trials include 1) requesting
more time; 2) communication breakdowns because of rotating call
schedules; and 3) changes in clinical course. Finally, facilitators to
completing time-limited trials include 1) having defined goals about
what could be achieved during an ICU stay, either framed in narrow,
numeric terms or broad goals focusing on achievable activities of
daily living; 2) applying time-limited trials in certain types of cases;
and 3) taking ownership to ensure completion of the trial.
Conclusions: An understanding of barriers and facilitators to initi-
ating and completing time-limited trials is an essential first step
toward appropriate utilization of time-limited trials in the ICUs, as
well as developing educational or communication interventions
with clinicians to facilitate time-limited trial use. We provide practi-
cal suggestions on patient populations in whom time-limited trials
may be successful, the setting, and clinicians likely to benefit from
educational interventions, allowing clinicians to have a fuller sense
of when and how to use time-limited trials. (Crit Care Med 2015;
43:2535–2543)
Key Words: communication; decision making; end-of-life; ethics;
family meetings; substituted judgment; surrogates; time-limited
trials
C
linicians, patients, and surrogate decision makers fre-
quently engage in discussions and decision making
about whether to continue invasive life-prolonging
therapies for critically ill patients or, alternatively, to shift to
palliative-only measures. Especially in the critical care envi-
ronment, patients (or more likely surrogate decision mak-
ers on behalf of incapacitated patients) are asked to make
complex treatment and end-of-life decisions (1–3). Where
outcomes are uncertain, patients, surrogates, and clinicians
Bruce et al
have to balance tensions between shifting to full comfort-
care measures at a time they feel is too early in the course of
patients’ illnesses versus continuing burdensome treatments
for a time they perceive as too long (1, 4). To accommodate
this tension, a third strategy has been proposed as an alterna-
tive to all-or-nothing approaches, often referred to as a “time-
limited trial” (TLT).
A TLT is defined as “an agreement between clinicians
and patient/surrogate decision makers to use medical ther-
apies—such as mechanical ventilation, enteral feeding, or
dialysis—over a defined period of time to observe if the
patient improves or deteriorates according to agreed-upon
clinical outcomes” (1). Quill and Holloway (1) proposed
a five-step framework for initiating TLTs, consisting of a
meeting between the care team and patient/surrogate deci-
sion makers to 1) define the patient’s acute care needs and
prognosis; 2) clarify the patient’s goals and preferences; 3)
identify objective markers for improvement or deteriora-
tion; 4) suggest a time frame, ranging from a few days to a
month or more, for reevaluation of the patient’s condition;
and 5) define potential actions to take at the end of the trial
to ensure its completion or, if complications arise during the
trial, actions that should be taken (1), with similar commu-
nication models being proposed by other authors (2). In this
way, TLTs offer opportunities for reevaluation and iterative,
outcome-focused decision making within the broader con-
text of patient’s overall priorities, thereby alleviating some
of the burden of having to make discrete all-or-nothing deci-
sions where outcomes are uncertain (5).
Although TLTs are broadly advocated (1, 2, 4–7), there
is little empirical evidence of the benefits and risks of TLTs
(8, 9), when they are initiated, when and why they succeed or
fail, and what facilitates completion of them. An understanding
of barriers and facilitators to initiating and completing TLTs is
an essential first step toward promoting appropriate utilization
of TLTs, as well as developing targeted educational or commu-
nication interventions with clinicians to encourage TLT use
and ensure effectiveness. Without this foundational under-
standing, TLTs might continue, at best, to be used infrequently
(5) or, at worst, to be used inappropriately or ineffectively.
Our study objectives were to 1) identify the purposes for
which TLTs are used in four open ICUs with mandatory inten-
sivist comanagement within one hospital and 2) identify bar-
riers and facilitators to initiating and completing TLTs. To our
knowledge, this is the first qualitative study on TLTs, one of
only two empirical studies on the subject (5), and the first to
study barriers and facilitators.
MATERIALS AND METHODS
The Baylor College of Medicine and Houston Methodist Hospi-
tal Institutional Review Boards approved this study. We selected
qualitative methodologies because the focus of this study is to
answer the “how” and “why” questions (10). To explore the
barriers and facilitators to initiating and completing TLTs, we
2536 www.ccmjournal.org December 2015 • Volume 43 • Number 12
Copyright © 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
wanted to cover the interpersonal context in which clinicians
operate, making qualitative methodologies ideal.
Study Design and Participants
We used purposive sampling to recruit a diverse sample of clini-
cians, with a particular emphasis on recruiting intensivists in dif-
ferent ICUs (surgical ICU [SICU], cardiovascular ICU [CVICU],
medical ICU [MICU], and neurosurgical ICU [NICU]). Our ratio-
nale for focusing on intensivists was that they would likely have
robust experience initiating and completing TLTs because they
coordinate care and frequently organize and participate in family
conferences. To allow for some preliminary comparisons, we also
purposively sampled surgeons. The other interviewees (one pallia-
tive care specialist and one neurologist) were purposively selected
because of their experience and familiarity with TLTs.
Interview Guide Development and Interviews
Between December 2014 and March 2015, C.R.B., C.L., L.T.,
A.D., and L.P. conducted in-person interviews using a struc-
tured interview guide. Five investigators reviewed the initial
interview guide based on an extensive literature review, expert
opinion, and familiarity with medical decision making. Inter-
view protocols were modified iteratively throughout the pro-
cess of data collection, so that questions with diminishing
informational returns were gradually replaced by questions
that emerged inductively from interviewees’ responses.
The final interview guide (Appendix 1, Supplemental
Digital Content 1, http://links.lww.com/CCM/B444) consisted
of open-ended questions exploring five domains: 1) experi-
ences with TLTs; 2) attitudes about TLTs; 3) content of TLTs;
4) barriers and facilitators to initiating and completing TLTs;
and 5) advice on interventions to promote appropriate use
of TLTs and facilitate completion of them. C.R.B., C.L., L.T.,
A.D., and L.P. conducted all of the interviews after obtaining
oral consent from interviewees. Interviews lasted between 20
minutes and 45 minutes. Interviews were audiorecorded, tran-
scribed verbatim by an independent transcription agency, and
redacted. Audiorecordings were destroyed after transcription.
Analysis
C.R.B, J.B.B., and C.L. analyzed the interviews using qualitative
description with constant comparative techniques. Qualitative
descriptive studies aim to provide a comprehensive summary
of events. Our aim is not to provide thick description (ethnog-
raphy), theory development (grounded theory), nor interpre-
tative meaning of an experience (phenomenology); rather, our
goal is to provide a straight description of clinicians’ experi-
ences in using TLTs in a language similar to their own (11, 12).
To provide this description, we developed a codebook collab-
oratively through discussions between our research team members
to ensure consistent application of codes. To do this, C.R.B., J.B.B.,
and C.L. organized potential codes by the above domains after
reviewing three transcripts. Code assignments to six preliminary
transcripts were made independently by C.R.B., J.B.B., and C.L.,
each generating new emergent codes as needed, and later com-
pared and discussed until consensus in coding styles was reached.
 Clinical Investigations

After the codebook was developed and agreed upon, coding
for the remainder of the interviews was done independently
by members of the research team and later merged for analy-
sis. Throughout this process of coding, emergent patterns and
themes were stored electronically (using “Comments” and
“Memos,” in the language of ATLAS.ti [13]) as a way to pro-
vide analytical depth, identify illustrative quotes, and highlight
points for team discussion. Interrater reliability was assessed
on a random sample of 32 quotes taken from six transcripts.
RESULTS
The majority of intensivists that were approached agreed to
participate (16–18). Three nurse practitioners working in
the ICUs were approached, and all agreed to participate. Ten
surgeons were approached, and nine were able to participate
(Table 1). All nine of them actively conduct surgery, four of
whom are transplant or vascular surgeons and five of whom
are acute care/trauma surgeons.
Specialties of our interviewees included internal medi-
cine critical care (11 interviewees), surgery or surgical critical
care (10 interviewees), anesthesia critical care (five inter-
viewees), or other specialties (one interviewee) (Table 1). In
some circumstances, it was difficult to discretely classify a
clinician’s specialty. One full-time intensivist is board certi-
fied and trained in surgery (categorized as an “intensivist” for
the purposes of this project; Table 1), and some surgeons are
trained in several subspecialties, including surgical critical
care. To provide clarity and consistency in reporting, we use
clinicians’ self-identified primary specialty for categorization
purposes. Specifically, four surgeons function as intensivists in
the SICU in addition to performing surgery. They are identi-
fied as surgeons for reporting purposes, as are the remaining
five surgeons who do not function as intensivists. Where there
are discernible differences between surgeons who function in
an intensivist capacity some of the time and those who do not,
we distinguish them as intensivist surgeons and nonintensivist
surgeons. Participant demographics are provided in Table 1.
We reached saturation of findings at 30 participants. This
is a typical number for reaching saturation when variability
across participants is low (14). Below, themes are arranged
according to domains that were the focus of the interview
guide: purposes, barriers, and facilitators to TLTs. Illustrative
quotes from the interviews are provided in Table 2.

Table 1. Participant Characteristics by Unit, Gender, and Experience

Nurse
Variables Total Intensivists Surgeons Practitioners Other

n (%)a 30 (100) 16 (53) 9 (30) 3 (10) 2 (7)

 Medical ICU 8 (27) 5 (31) 0 (0) 2 (67) 1 (50)
 Neurosurgery ICU 6 (20) 3 (19) 2 (22) 0 (0) 1 (50)
 Cardiovascular ICU 11 (37) 8 (50) 2 (22) 1 (33) 0 (0)
 Surgical ICU 5 (17) 0 (0)b 5 (56) 0 (0) 0 (0)
Specialty of physicians
 Surgery 10 (37) 1 (4) 9 (100) 0 (0)
 Internal medicine 11 (41) 10 (37) 0 (0) 1 (50)
 Anesthesia 5 (19) 5 (19) 0 (0) 0 (0)
 Other 1 (4) 0 (0) 0 (0) 1 (50)
Gender
 Female 10 (33) 4 (25) 2 (22) 3 (100) 1 (50)
 Male 20 (67) 12 (75) 7 (78) 0 (0) 1 (50)
Experience (yr) c

 0–5 5 (17) 2 (13) 3 (33) 0 (0) 0 (0)

 5–10 7 (23) 1 (6) 3 (33) 3 (100) 0 (0)
 10–15 9 (30) 7 (44) 1 (11) 0 (0) 1 (50)
 15–25 2 (7) 2 (13) 0 (0) 0 (0) 0 (0)
 25+ 7 (23) 4 (25) 2 (22) 0 (0) 1 (50)
In some instances, percentages may not equal 100 because of rounding.
a

b
Four surgeons in the surgical ICU also work as intensivists on a part-time basis. They are identified as surgeons for reporting purposes. All of them actively
perform surgeries.
Years” is calculated from time of residency completion for physicians and time of graduation for nurses.
c“

Critical Care Medicine www.ccmjournal.org 2537

Copyright © 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Bruce et al

Table 2. Illustrative Quotes of Purposes, Barriers, and Facilitators to Initiating and

Completing Time-Limited Trials
Purposes Prepare surrogates Subject 027: “Those few days allow things to sink in, rather than just saying. ‘He’s not going
and surgeons to make it; we need to pull the plug today.’”
Subject 026: “Watchful waiting. It can help prepare families…gives them time, something
concrete to hold onto. Otherwise, you can just go along [for days]. This is more proactive.”
Subject 028: “It does give the family more opportunity for acceptance, but us, too. We can
swing toward the pendulum and agree. It allows for closure.”
Build consensus Subject 023: “I think you have to have the courage and the guts to recognize that you, as a
physician, don’t just have to order up people what they want like a short-order cook off the
drive-thru menu. Direct their thought process…set limits.”
Refine prognosis Subject 03: “I’m not 100% sure whether you’re going to get better or not, I do know that in
72 hours we are going to have some pretty good evidence that if they are not better by
that point, than the likelihood drops off significantly.”
Subject 017: “A specific one might be some type of neurologic injury, where a patient had a
cardiac arrest and there was a hypoxic brain injury. With those patients, usually the first 48
hours is the most critical time period in terms of assessing neurologic recovery.”
Barriers to Surrogates do not Subject 07: “More often than not, the family has different thoughts and their own
initiating TLTs agree prejudices… And that becomes very difficult to set timelines because they don’t want to
hear that, especially if they are in the mindset that the care will be abandoned.”
Clinicians do not Subject 026: “There is no data that all of my colleagues can agree upon. We will just re-
agree evaluate. It becomes more of an art than a science at some point. People have different
amounts of risks they are willing to tolerate.”
Barriers to Requesting more Subject 25: “Once it is crunch time, the family wants the estranged son to come in; the
completing time daughter from Colorado….They are looking for a flight next Tuesday. Even it is about time,
TLTs it isn’t time. The family is the biggest barrier.”
Subject 024: “You make this agreement at the zero hour when you start and, by the time
you get to hour 48, now the whole situation is contaminated by all these people who have
been on the Internet. Or they talked to someone. They said he would never speak again
and now he’s in college.”
Subject 007: “We may keep extending the time without really enforcing the decision that
was made 5 days ago. So we keep on changing or moving the goalpost. Is that right or
wrong? I don’t know.”
Communication Subject 010: “It would be helpful if before we even talk to the families…the teams
breakdown sit together and they decide—a team meeting without the families-- to know…our
expectations. Because if the expectations of the surgical team and the ICU team are not
the same, that is what creates problems.”
Clinical course Subject 026: “You have a plan in your head, but things change. Battle plans only last until the
changes first shot is fired. Cases are too individualized.”
Facilitators to Having a defined Broad goal
completing goal  Subject 002: “I go into a meeting and say, well, your mom isn’t waking up, so let’s give her 3
TLTs days and then we will reassess. But, in those 3 days, if she is opening her eyes…the family
will take this as a positive sign…So before you set a TLT, you have set specific goals that are
based on overall quality of life, like at the end of 3 days, is she able to interact?”
Narrow goal
 Subject 011: “If you tell them, okay, let’s give them a week and if by the end of the week,
they’re not off pressors…If they haven’t met these milestones then, it makes it easier to
take the next step...”
Applying TLTs in Subject 003: “Better outside of the ICU where there might be single-system failure. In
certain types of oncology, where you can start a patient on chemotherapy, you just repeat a biopsy and the
cases numbers tell you if the patient is better or worse.”
Taking ownership Subject 022: “The owners of the patients [are responsible]…the physicians of record. The
the people in the ICU are technically consultants. Those [admitting] folks have to buy-in.
Because I’m the physician of record and I’ve known this patient for the last 40 years...That
makes a whole bunch of difference.”
TLT = time-limited trial.

2538 www.ccmjournal.org December 2015 • Volume 43 • Number 12

Copyright © 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

PURPOSES OF TLTs
Interviewees described using TLTs for three different purposes:
to prepare surrogates and clinicians for discussion and possible
shift toward comfort-care only therapies, build consensus, and
refine prognostic information. Clinicians in every unit and
every specialty described using TLTs in these three ways, both
as a way to address family dynamics and clinician dynamics.
Preparation of Surrogates and Clinicians
Every interviewee (and every specialty) in each ICU men-
tioned the utility of TLTs as a way to prepare surrogates and
families for likely negative outcomes, including discussion
about pursuing comfort-related goals. By conceptualizing and
describing patients’ health incrementally over a period of days
or weeks, interviewees felt they were helping surrogates foster
appropriate expectations in a gradual, anticipatory way. Inter-
viewees described two distinct goals: provide surrogates com-
passion and time to adjust (subject 013: “Approaching families
at the end when the patient is close to death doesn’t give them
any time to adjust.”) while simultaneously aligning expecta-
tions between clinicians and patients/surrogates. Specifically,
interviewees described receiving “unrealistic” responses from
surrogate decision makers whenever the foundation for end-
of-life discussions was not properly introduced by clinicians in
earlier conversations. Clinicians felt that surrogates perceived
end-of-life conversations as jarring and unexpected, resulting
in what clinicians perceived to be a misalignment between cli-
nicians and surrogates regarding goals of care and expectations
of outcomes.
TLTs could also be used to prepare clinicians, especially sur-
geons, for a shift toward comfort-care measures (subject 027:
“And it helps us. We, especially surgeons, have a history [with
the patient]. It changes your dynamic.”). Clinician prepared-
ness was described in all but two surgeon interviews, one of
whom is a nonintensivist, transplant surgeon. It was also men-
tioned in two intensivist interviews.
Build Consensus
Eleven interviewees described using a TLT to build consensus
among family members who disagree with each other by estab-
lishing a timeframe for surrogates/families to make a decision,
subject 007: “The family… they’re divided. They go into a state
of inertia. That is the trial’s biggest value: it forces them to con-
front and make decisions.” Using the TLT to formulate concor-
dance among divided families was mentioned in every unit,
most especially in the MICU and the NICU. This was the most
frequent use of TLTs in the MICU. There were no discernible
differences between specialties.
Conversely, six interviewees, all from the CVICU and SICU,
described using the TLT to formulate agreement among divided
clinicians, as opposed to divided family members. Subject 011:
“Sometimes, there is conflict within the team itself. If that is the
case, you may wait a week to give the team a chance to accept
what is going on with the patient.” Discordance between clini-
cians was generally described as being disagreement between
surgeons and intensivists, raised primarily by the latter group
Critical Care Medicine www.ccmjournal.org 2539
Copyright © 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigations
of interviewees. Only one surgeon (a nonintensivist transplant
surgeon) reported using TLTs to formulate agreement among
surgeons and intensivists about plans of care.
Refine Prognosis
A TLT may be used to gather more data where there is prog-
nostic uncertainty. This was the most frequent use of TLT in
the NICU. Every interviewee in the NICU described using the
TLT after an acute brain injury to help refine prognostication
efforts, with the first 72 hours being the most critical. Within
the day of admission or the day of the brain injury, interview-
ees described the meeting with the family and frequently advo-
cating nonescalation of aggressive measures for 48–72 hours
to assess the patient’s overall health condition and likelihood
of recovery. Clinicians reported meeting again with the fam-
ily after this time period and, based on whether there was any
improvement or deterioration during that 72-hour period,
recommended proceeding with invasive interventions or shift-
ing to full comfort-care measures.
Thirteen interviewees (every unit; each specialty) described
using the TLT as means of prognostic refinement when a family
requested discontinuation or deescalation of aggressive thera-
pies at a time clinicians thought was too premature. This theme
was particularly prominent among nonintensivist surgeons and
intensivist surgeons, surfacing in all surgeon interviews. There
was a near equal distribution among the units. Subject 003:
The patient said he never wanted to be on life-support.
He is going to be able to return to exactly the same level
of functioning as before…With this potential for a very
high level of recovery, I am OK pushing it a little. And I
think a small amount of paternalism is OK.
By requesting more time, this clinician felt he would be bet-
ter positioned to assess the patient’s anticipated outcome. In this
case, the patient’s surrogate decision maker agreed to wait more
time to reevaluate the patient’s condition in the context of his
overall priorities. However, two other clinicians who mentioned
using a TLT as a means of requesting more time reported that
surrogate decision makers were not amenable to creating a TLT.
BARRIERS TO INITIATING TLTs
Surrogates Do not Agree
One reported barrier to initiating a TLT that arose in approxi-
mately five interviews (mix of specialties and units) is that sur-
rogate decision makers may not be receptive to setting a time
period or may interpret a timeline as a means of “pressuring”
them to make a decision. Subject 003: “When the family isn’t
ready, the family isn’t ready. If they aren’t ready for a trial,
pushing it or ‘insisting’ on it more often alienates the family…
and [causes] more harm than good.” A related barrier is that in
the clinical context, it can often be difficult to build a rapport
with the family in the amount of time that would be necessary
to initiate dialogue about a TLT. Without this foundation of
rapport, discussion about a timeline would only be met with
resistance.
Bruce et al
Clinicians Do not Agree
Clinicians may not uniformly agree about whether to institute
a TLT trial or, alternatively, the amount of time that should be
set as a variable for the trial. Eleven clinicians (in the CVICU
and SICU; every specialty) discussed public reporting of surgi-
cal outcomes based on 30-day mortality and how this could
impact surgeons’ receptivity to initiating TLTs. Subject 017
(nurse practitioner): “If you have competing interests, it won’t
work. Surgeons…want their patients to survive because they
need their numbers to look good. Intensivists are looking at
the bigger picture.” Surgeons (whether intensivist surgeons or
nonintensivist surgeons) did not report a reluctance to initi-
ate TLTs, even when prompted during the interview, although
two of them (both nonintensivist transplant surgeons) said
outcomes matter (subject 27: “It will count as a death. Like it
or not, we have to pay attention to that stuff. But that doesn’t
mean I will allow my patient to suffer or prolong it.”).
Eighteen clinicians reported that, even where there was
agreement with initiating a TLT in theory, it was difficult to
formulate agreement on the amount of time that should be set
as a variable, particularly when the patient had multilpe organ
failure or where there was little evidence to drive the analysis.
This was most salient in the MICU, a theme that arose in every
interview, and least prominent in the NICU (whose clinicians
uniformly reported 48- to 72-hr trials, with no disagreement
among them). Given that all of the MICU clinicians have back-
grounds in internal medicine, this theme was most prominent
among this specialty, although it also arose in interviews with
surgical specialties.
BARRIERS TO COMPLETING A TLT
Requesting More Time
During a trial, a surrogate may request more time when a
patient deteriorates rapidly, pointing to the original timeframe
as the point at which decision making should occur (subject
006, “Let’s say you gave them a week. In two or three days, the
patient may be deteriorating drastically, and you are coding
3–4 times a day and the family says: ‘But you said 7 days.’”). It is
unclear whether this clinician discussed with surrogates at the
initiation of the trial what would happen if the patient rapidly
deteriorated during the interim of the trial.
Three interviewees (all in the CVICU) described TLTs as
lacking efficacy when clinicians extend timeframes without
any substantive change in clinical status, which they attributed
to a diffusion of responsibility among the team members or a
general lack of desire to “be the one to call it quits” or “it can
just be passed to the oncoming team,” as reflected in this com-
ment, subject 011:
A TLT has backfired when we can’t stick to a timeline.
Let’s wait a week. Oh, there’s no improvement. Let’s wait
another week. And no one is willing to step up and say,
we’ve waited long enough. We kick the can down the
road when there is high turnover among the team.
2540 www.ccmjournal.org December 2015 • Volume 43 • Number 12
Copyright © 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Communication Breakdown
Rotating call schedules and frequent handoff among the team
was cited in all but five interviews as a principal reason for TLT
disintegration, with no notable differences between the units
or specialties. Subject 003, “There are too many teams involved
with a patient in the ICU. Even among ourselves as the intensiv-
ist group, it’s rare that the same doc sees the same patient more
than a few days in a row.” Twenty-three interviewees believed
that rotating call schedules and block scheduling not only
affected their ability to assess whether a TLT is appropriate and
feasible (e.g., building enough rapport to establish a TLT) but
also impacted their ability to know whether a trial had been set,
if the goals or variables of the trial had been determined, and
whether there is a willingness to complete them.
Clinical Course Changes
Approximately half of clinicians (every unit; every specialty)
described patients’ conditions changing so much that complet-
ing TLTs was subject to disintegration, no matter how well for-
mulated the trial was at the outset. This was most prominent in
the MICU, surfacing in all but one interview. Subject 003: “You
tell a family that at the end of 7 days, if the ventilator doesn’t
work, then we should transition to comfort care. But at day 7,
the patient is 60% better and trending up. But now they have an
infection... How do you account for other problems that weren’t
present at the initiation of the trial but come on after it started?”
FACILITATORS TO TLTs
Having a Defined Goal
Clinicians agreed that TLTs can only be successful insofar as
the goals were clearly defined and measurable, a theme aris-
ing in all but two interviews. However, there was considerable
variation on what constitutes an appropriate goal, with some
clinicians focusing quite narrowly on clinical, objective vari-
ables mapped to positive or negative trends in laboratory val-
ues or weaning efforts, whereas others focused more broadly
on outcomes in the context of a patient’s overall priorities.
“Broad” framing centered on whether a patient would be able
to achieve certain activities by a specific time frame, such as
“wakefulness” or “mobility.” “Narrow” framing focused on
numerical values: trending laboratory values or the amount of
ventilatory support to indicate increased or decreased depen-
dence on specific therapies or improvement or deterioration of
a particular organ.
There were discernable differences between units on
whether they defined goals in terms of narrow goals or broad
goals. The MICU and NICU intensivists focused on broad
goals. All CVICU intensivists described mapping goals using
clinical variables, typically liver function tests and drips, and
only one intensivist from this unit described mapping goals
using broad goals. Likewise, this finding could also be mapped
to specialty. Internal medicine critical care specialties (MICU
and NICU) focused on broad goals, whereas anesthesia and
surgery critical care specialties (CVICU and SICU) focused on
narrow ones.

Applying TLTs in Certain Types of Cases
Seventeen interviewees reported that it is easier to initiate and
complete a TLT when there are few confounding variables,
most often attributed to cases where a patient has only one
organ-failing or one life-limiting condition rather than sys-
temic failure. In these cases, there may be more evidence to
guide the setting of variables. Subject 028: “Too many factors
and variability…Any obstacles beyond isolated organ failure
are just so hard to overcome and hard to calculate.”
Clinicians also reported that it is easier to complete a TLT
when there is one life-limiting condition because they would
likely defer to the consultant charged with addressing a partic-
ular health problem on whether and how to initiate a TLT. This
was a particularly prominent theme in the SICU and MICU
and most especially among internal medicine specialties.
Taking Ownership
A majority of clinicians claimed that someone taking “owner-
ship” of the trial and ensuring its completion was vital to its
success. However, there was significant disagreement about
who should take ownership. Specifically, clinicians disagreed
about who should be responsible for initiating a trial or com-
pleting a trial: ranging from the primary service, to the sur-
geon, to the person with the strongest relationship with the
family. On some occasions, clinicians thought that the person
initiating the TLT should not necessarily be responsible for
completing it. For instance, at least two interviewees thought
that the primary service or surgeon should be responsible for
initiating it, but they also thought that support services (e.g.,
social worker or palliative care specialists) should be charged
with ensuring trial follow-through.
Despite this discordance, nearly all clinicians agreed that,
in instances where there is isolated organ failure, the person
charged with addressing that underlying issue (e.g., neurology
for an acute brain injury) should drive the timing and content
of TLTs.
DISCUSSION
Our work extends and deepens our understanding of TLTs by
identifying additional purposes, core challenges to TLT imple-
mentation (initiation and completion), and potential strategies
to overcome these challenges (Table 3). Below, we elaborate on
a few of our main findings concerning TLT implementation
challenges, provide clinical implications for the use of TLTs,
and suggest recommendations on how clinicians should oper-
ationalize TLTs.
The most prominent TLT implementation challenge iden-
tified in our study involves fragmented healthcare environ-
ments resulting from shift-based ICU or block-scheduling
staffing patterns (15–17). Although Quill and Holloway (1)
raised this as a theoretical concern, this is the first empiri-
cal evidence highlighting the degree to which rotating teams
and shifting responsibilities affect TLT. Although this is an
exploratory study, there are preliminary data suggesting that
staffing patterns have the potential to undermine the utility
Critical Care Medicine www.ccmjournal.org 2541
Copyright © 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigations
of TLTs: 1) shift-based or block-scheduling staffing patterns
impact clinicians’ assessments of whether a TLT would be
feasible, likely creating a reluctance to initiate a trial where it
might be appropriate; 2) block scheduling impacts success-
ful communication of TLT variables between team members;
and 3) block scheduling impacts follow-through of the trial
because of a diffusion of TLT-completion responsibilities
among off-going and oncoming service members.
A second implementation challenge we identified involves
intrateam discordance. This finding is related to observa-
tions previously described by others that surgeons may be less
inclined to initiate postoperative goals-of-care discussions
than other clinicians (18–20), and, in the context of our study,
these discussions would include initiating and completing
TLTs. However, we found differing perspectives between sur-
geons and intensivists about surgeons’ willingness to engage in
goals-of-care conversations and TLTs. Specifically, our surgeon
interviewees (intensivists and nonintensivists) did not express
a reluctance to initiate TLTs, and they frequently touted the
benefits of a TLT while simultaneously acknowledging a need
to monitor surgical outcomes, a finding that deviates from
previous conceptual work on surgeon reporting and its impact
on TLTs (6). Surgeons did not perceive TLTs and surgical out-
comes as necessarily competing in the way that intensivists did.
Another major barrier to successful TLTs reported by
interviewees is the perception that changes in the patient’s
clinical course sometimes made completing a TLT difficult, a
finding that, to our knowledge, has not been previously iden-
tified. We consider this an important finding because it may
signal an area in which more clinician education is needed
about how to operationalize TLTs, implications of which we
describe further below.
Our work has potential implications regarding how clini-
cians should operationalize TLTs. First, rather than perceiving
a change in clinical course to be an implementation barrier, it
should be viewed as integral to the clinical complexities and
uncertainties that make TLTs an appropriate alternative to
all-or-nothing treatment approaches. One way to appropri-
ately align expectations between clinicians and surrogates is to
describe the possible outcomes before the trial is implemented.
Potential outcomes would include the following: 1) the patient
gets worse and is unable to successfully meet recovery-oriented
milestones, and therefore, the comfort pathway is likely appro-
priate; 2) the patient improves and meets recovery-oriented
milestones, and therefore, life-prolonging measures are appro-
priate; 3) the patient’s likely outcome is still uncertain, and
therefore, another conversation is needed, which could include
establishing new milestones for the TLT.
As part of managing expectations, it is important for clini-
cians to recognize the limits of TLTs, employing them in cases
where they are most likely to be successful and not viewing
them as a panacea for all complex cases. The “ideal” TLT case,
it seems, is one involving a medically straightforward situation,
such as single-organ failure, or involving a single life-limiting
illness, such as cancer (Table 3). The challenge in ICUs is that
patients often have systemic failure. Perhaps, TLTs might be
Bruce et al

Table 3. Summary of Practical, Clinically Based Recommendations for When

and How to Use Time-Limited Trials
Description Likely Facilitators to Effective TLT Use Likely Barriers to Effective TLT Use

Patient Single-organ failure Conditions without a strong evidence base

for formulating a recovery/deterioration
Single life-limiting illness timeline
Conditions with a strong, consensus-
driven evidence base for formulating a
recovery/deterioration timeframe
Setting Small or closed ICUs Large or open ICUs
ICU staffing models not based on shifts or Shift-based or block-scheduling staffing
blocks of time models

Post–acute care settings ICU team members who rotate or shift

Families Preparation of surrogates or families for
comfort-care discussions
Mitigation of intrafamily discordance about
prognostic information or plans of care
Clinicians Initiation by a clinician charged with
treating a primary life-limiting condition
TLT = time-limited trial.
responsibilities frequently
Surrogates or families who prefer taking
an active/lead role in decision making
Surrogates or families who report feeling
pressured to make decisions
Initiation by a clinician who is unable or
unwilling to close the trial

most appropriate outside the ICU setting, an area ripe for
future study.
Another clinical implication of our work is that we, like
others, believe that TLTs will only be successful insofar as they
have discrete, achievable goals (1, 5). We build upon previous
work by suggesting that narrow clinical variables will be used
in ICUs where clinicians tend to be more anesthesia and surgi-
cally oriented (like the CVICU). These specialties are perhaps
more conducive to numerical values than other specialties,
which supports the use of narrow clinical variables. Broadly
focused goals, on the other hand, are used in ICUs where clini-
cians are often trained in internal medicine. Whether and how
clinicians’ backgrounds influence the content of TLTs, includ-
ing its goals, is worthy of further exploration to determine
whether narrow or broad framing facilitates completing TLTs.
Our final clinically based, practical strategies for overcom-
ing TLT implementation barriers focus on mitigating chal-
lenges associated with shift-based ICU staffing models. To this
end, it would be helpful to ensure continuity of communica-
tion among rotating teams, perhaps using a visible tool in the
patient’s chart to call attention to the creation of a trial or an
added feature on handoff reports that accounts for whether a
TLT has been initiated and the variables of a trial. If communi-
cation interventions and educational sessions are provided to
foster clinicians’ skills in talking about TLTs, perhaps these pro-
grams should develop a phased-in approach for training inter-
ventions, prioritizing training for clinicians most often given
deference in creating trials (e.g., neurologists, oncologists, and
intensivists).
This study has a number of limitations that are inherent to
qualitative methodologies and unit characteristics. For instance,
we recognize that there are limitations in reporting word
frequency and category frequency. Frequencies do not allow for
definitive causal assertions and empirical comparisons because
qualitative interviewing only allows for rich exploration of
themes and topics (8, 12). Despite this, we believe that provid-
ing numbers in areas that can confidently be reported gives the
reader a sense of the prevalence and prominence of themes.
Finally, because this study was done at one tertiary referral insti-
tution, results may not be generalizable. Preliminary findings,
although modest, are important for directing further research.
Future work should also assess surrogates’ or patients’ perspec-
tives on TLTs and their efficacy, as we only provide clinicians’
perspectives, a limitation we acknowledge.
CONCLUSIONS
Our study documents that clinicians perceive TLTs to be ver-
satile. They are used to build consensus among clinicians,
among clinicians and family members, and among family
members. They also have several implementation challenges.
By exploring barriers and facilitators to initiating and com-
pleting TLTs in several different ICUs with clinicians of varied
backgrounds, we are able to provide practical, clinically based
suggestions, allowing clinicians to have a fuller sense of when
and how to use TLTs, as well as when there might be barriers
to implementation.
REFERENCES
1. Quill TE, Holloway R: Time-limited trials near the end of life. JAMA
2011; 306:1483–1484
2. Workman S: A communication model for encouraging optimal care at
the end of life for hospitalized patients. QJM 2007; 100:791–797
3. Berger JT, DeRenzo EG, Schwartz J: Surrogate decision making:
Reconciling ethical theory and clinical practice. Ann Intern Med
2008; 149:48–53

2542 www.ccmjournal.org December 2015 • Volume 43 • Number 12

Copyright © 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

4. Schenker Y, Crowley-Matoka M, Dohan D, et al: I don’t want to be the
one saying ‘we should just let him die’: Intrapersonal tensions expe-
rienced by surrogate decision makers in the ICU. J Gen Intern Med
2012; 27:1657–1665
5. Schenker Y, Tiver GA, Hong SY, et al: Discussion of treatment trials in
intensive care. J Crit Care 2013; 28:862–869
6. Neuman MD, Allen S, Schwarze ML, et al: Using time-limited trials to
improve surgical care for frail older adults. Ann Surg 2015; 261:639–641
7. Bruce CR, Allen NG, Fahy BN, et al: Challenges in deactivating a
total artificial heart for a patient with capacity. Chest 2014; 145:
625–631
8. Rinehart A: Beyond the futility argument: The fair process approach
and time-limited trials for managing dialysis conflict. Clin J Am Soc
Nephrol 2013; 8:2000–2006
9. Creutzfeldt CJ, Holloway RG: Treatment decisions after severe stroke:
Uncertainty and biases. Stroke 2012; 43:3405–3408
10. Yin RK: Case Study Research: Design and Methods. Third Edition.
Thousand Oaks: SAGE Publications, Inc, 2002
11. Neergaard MA, Olesen F, Andersen RS, et al: Qualitative description -
The poor cousin of health research? BMC Med Res Methodol 2009;
9:52
12. Sandelowski M: Whatever happened to qualitative description? Res
Nurs Health 2000; 23:334–340
Critical Care Medicine www.ccmjournal.org 2543
Copyright © 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigations
13. Friese S. ATLAS. ti 7 User Guide and Reference. Berlin: ATLAS. ti
Scientific Software Development GmbH, 2013
14. Strauss A, Corbin JM: Basics of Qualitative Research: Grounded
Theory Procedures and Techniques. Second Edition. Newbury Park:
SAGE Publications, Inc, 1990
15. Hysong SJ, Knox MK, Haidet P: Examining clinical performance feed-
back in patient-aligned care teams. J Gen Intern Med 2014; 29:
667–674
16. Crown DF, Rosse JG: Yours, mine, and ours: Facilitating group pro-
ductivity through the integration of individual and group goals. Organ
Behav Hum Decis Process 1995; 64:138–150
17. Locke E: The nature and causes of job satisfaction. In: Handbook
of Industrial and Organizational Psychology. Chicago: Rand McNally
Publishing Company, 1976, pp 1297–1349
18. Paul Olson TJ, Brasel KJ, Redmann AJ, et al: Surgeon-reported con-
flict with intensivists about postoperative goals of care. JAMA Surg
2013; 148:29–35
19. Penkoske PA, Buchman TG: The relationship between the surgeon
and the intensivist in the surgical intensive care unit. Surg Clin North
Am 2006; 86:1351–1357
20. Bruce CR, Miller SM, Zimmerman JL: A qualitative study exploring
moral distress in the ICU team: The importance of unit functionality
and intrateam dynamics. Crit Care Med 2015; 43:823–831