Time-Limited Trials for Patients With Critical Illness

Abstract

Topic Importance

Since the 1990s, time-limited trials have been described as an approach to navigate uncertain benefits and limits of life-sustaining therapies in patients with critical illness. In this review, we aim to synthesize the evidence on time-limited trials in critical care, establish what is known, and highlight important knowledge gaps.

Review Findings

We identified 18 empirical studies and 15 ethical analyses about time-limited trials in patients with critical illness. Observational studies suggest time-limited trials are part of current practice in ICUs in the United States, but their use varies according to unit and physician factors. Some ICU physicians are familiar with, endorse, and have participated in time-limited trials, and some older adults appear to favor time-limited trial strategies over indefinite life-sustaining therapy or care immediately focused on comfort. When time-limited trials are used, they are often implemented incompletely and challenged by systematic barriers (eg, continually rotating ICU staff). Predictive modeling studies support prevailing clinical wisdom that prognostic uncertainty decreases over time in the ICU for some patients. One study prospectively comparing usual ICU care with an intervention designed to support time-limited trials yielded promising preliminary results. Ethical analyses describe time-limited trials as a pragmatic approach within the longstanding discussion about withholding and withdrawing life-sustaining therapies.

Summary

Time-limited trials are endorsed by physicians, align with the priorities of some older adults, and are part of current practice. Substantial efforts are needed to test their impact on patient-centered outcomes, improve their implementation, and maximize their potential benefit.

An enduring challenge in critical care medicine is recognizing and navigating the limits of prolonged life-sustaining therapy for patients who may not recover. The concept of a time-limited trial has been discussed in the health care literature since at least 1998,¹ as an approach to patient care with potential to address this challenge.² A time-limited trial is a collaborative plan among physicians and a patient and/or their surrogate decision-maker(s) to use life-sustaining therapy for a defined duration, after which the patient’s response to therapy informs the decision to continue care focused on recovery, transition to care focused exclusively on comfort, or extend the trial’s duration.³ These trials are described as a middle-ground, positioned between indefinite life support and comfort-focused, end-of-life care. This middle ground reflects the reported goals of many patients, particularly older adults—to extend life if possible and to avoid prolonged, invasive therapies if death is near.^1,4,5 This approach also reflects a routine, instinctive process of critical care—trialing interventions and evaluating the patient’s response.⁶ However, time-limited trials are distinguished from routine critical care by the explicit collaboration with patients and/or their surrogates to create a structured evaluation plan.³

Time-limited trials are also endorsed and discussed beyond critical care, including in the fields of palliative care, nephrology, neurology, and surgery.^{7, 8, 9, 10} In palliative care, a position statement on withholding and withdrawing nonbeneficial medical interventions from the American Academy of Hospice and Palliative Medicine states that, “when the outcome of a proposed intervention is uncertain, clinicians should consider a time-limited trial of the specific intervention.”¹¹ In nephrology, clinical practice guidelines from the Renal Physicians Association and the American Society of Nephrology recommend consideration of a time-limited trial of dialysis for patients, “who have an uncertain prognosis or for whom a consensus cannot be reached about providing dialysis.”⁸ In a 2006 survey, approximately 80% of nephrologists in Canada and the United States reported using time-limited trials of dialysis.¹² The recognition of time-limited trials has even extended to the lay media, with a 2021 article published in The New York Times describing this approach in the ICU context.¹³

Although time-limited trials are now well described in critical care medicine and beyond, the evidence base to support their use remains limited and sporadic. In this narrative review, we aim to synthesize the existing empirical work on time-limited trials for patients with critical illness, establish what is known about this approach to patient care, and highlight important remaining questions.

Literature Search

We conducted literature searches in March and July 2023, using the health care databases of Cumulated Index to Nursing and Allied Health Literature Plus, MEDLINE via PubMed, and Scopus. We used the following search terms: time-limited trial and time-limited trials, which were applied to all available search fields. We limited the search to articles available in English. We included all article types (eg, editorial, reviews, original research) in the initial search, except conference abstracts given their lack of detail. We also reviewed the reference lists of included articles and elicited suggested articles from the coauthors of this review, adding those pertaining to review objectives in the final synthesis. We excluded articles not related to patients with critical or serious illness or that only tangentially remarked on time-limited trials. This strategy identified 71 articles (Fig 1). We organized these articles into three categories: expert opinion and reviews (n = 38), empirical studies (n = 18), and ethical analyses, including both normative and empirical ethics articles (n = 15). The opinion-based articles informed our introduction, and the following narrative review focuses on the empirical and ethical articles. Table 1^{1,3,5,6,14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28} provides a summary of the 18 reviewed empirical articles, and e-Appendix 1 contains the complete 71-article reference list.

Figure 1.

Literature search results.

Table 1.

Summary of Empirical Studies on Time-Limited Trials in Patients With Critical Illness

Study Design, Population, Context	Primary Conclusiona	Considerationsa	Study
Studies examining current time-limited trial practices
An observational study of the routine triage process in one French medical ICU, which included a 5-d ICU trial for patients with cancer.	Of those patients with cancer admitted for an ICU trial, 45% died before day 5 despite maximal therapies, 33% died in the hospital on day 5 or later, and 22% survived the hospital stay.	•No comparison of the ICU trial with other triage or care strategies. •This ICU trial differs from what is typically described as a time-limited trial in other articles.	Lecuyer et al14
Ethnography of care delivery in two ICUs in the United States, with known variation in the unit-level intensity of end-of-life care (high and low intensity).	Time-limited trials are used in both ICUs but were more common in the low-intensity unit (27% of patients) compared with the high-intensity unit (8%).	•Lack of detail on how the investigators defined or labeled care as a time-limited trial.	Barnato et al15
Qualitative study of audio-recorded ICU family meetings, during which limitations on life-sustaining therapy were discussed. Conducted in five ICUs within two hospitals in California.	13% of family meetings included a discussion of time-limited trials. These discussions were incomplete in informing the family about the elements of a time-limited trial.	•Highly selected sample of patients and family meetings. •Unable to determine if time-limited trials were followed through after the meeting.	Schenker et al16
Qualitative interview study with ICU physicians and nurse practitioners who practice in an academic medical center in Texas.	ICU physicians all described experience using time-limited trials and report multiple purposes and various barriers and facilitators to success.	•Families requesting more time were seen as a barrier to time-limited trial “success”; however, this is an accepted feature of time-limited trials.	Bruce et al17
Qualitative interview study with ICU physicians and nurses in the United Kingdom.	Physicians described how the passage of time can help reduce their uncertainty about prognosis, build their relationships with patients and families, and prepare patients and families for end of life.	•Generally focused on the role of time in the ICU; did not query participants directly on the structured time-limited trial approach.	Lonergan et al6
Mixed-methods study of ICU physicians from two academic medical centers in the United States.	65% of physicians are familiar with the concept of time-limited trials but report marked variation in how they conduct trials.	•Small sample size and focused on pulmonary/critical care in academic medical centers, therefore may not generalize to other populations.	Viglianti et al18
Process mapping study of ICU care delivery at two academic medical centers in the United States, using focus groups with interprofessional physicians and semi-structured interviews with patients/surrogate decision-makers.	Time-limited trials are part of the current process of care in study ICUs; some trials are not followed through after being initiated.	•Focused specifically on patients with acute respiratory failure. •Study was not primarily focused on time-limited trials.	Kruser et al19
Modified Delphi process, including a two-round survey of 100 stakeholders from 10 countries.	Strong consensus reached about the 16 essential elements for conducting a time-limited trial in patients with critical illness.	•Results were informed by the stakeholder survey and refined/finalized within a smaller committee of 28.	Kruser et al3
Studies examining patient and/or family perspectives on time-limited trials
Interview study with people living in Seattle, Washington, evaluating concordance in their reported preferences for life-sustaining treatment.	Participants who seemed to have “discordant” preferences described preferring a “trial” of a treatment instead of the dichotomous choice between accepting or refusing it.	•Participants spontaneously raised the notion of time-limited trials, without the investigators suggesting it. •Given this was an emergent finding in the data, details are limited.	Patrick et al20
Survey study about future care preferences among older adults living in nursing homes in New York.	Older adults favor a time-limited trial of mechanical ventilation over open-ended ventilation or immediate comfort-focused care.	•Patients’ hypothetical, future care preferences may not always reflect in-the-moment preferences. •Lack of detail in the description of a time-limited trial.	Berger et al1
Survey study about future care preferences among surrogate decision-makers for older adults living with dementia in Massachusetts.	31% of surrogates think the patient would choose a time-limited trial over open-ended life-sustaining therapy or immediate comfort-focused care.	•Specifically focused on older adults with dementia. •Lack of detail in the description of a time-limited trial.	Givens et al5
Qualitative interview study with surrogate decision-makers for patients with severe acute brain injury in the Northwestern United States.	Time-limited trials were valued by surrogates who felt agency in medical decision-making but not by those surrogates who saw their role as passive.	•Focused specifically on acute brain injury population.	Goss et al21
Studies examining outcomes of time-limited trials
Pre-post evaluation of a quality improvement project, designed to teach ICU physicians to communicate about time-limited trials for patients at high risk of death. Conducted in three hospitals in California.	Postintervention, patients had shorter ICU and hospital length of stay. The reduction in length of stay appeared to be greater in the patients who ultimately died in the hospital than patients who survived.	•Unclear to what extent or how time-limited trials were conducted postintervention (nor preintervention). •An accompanying editorial22 highlights that study results may be mostly due to the more frequent and earlier meetings with families, not time-limited trials specifically.	Chang et al23
Studies quantifying change in prognostic certainty over time or describing time-to-death
An evaluation of the predictability of hospital mortality over time in the ICU, according to serial machine learning models. Conducted with a data set from 11 ICUs in the United Kingdom.	The predictability of ICU mortality improved over time, with the best machine learning model performance on ICU d 2 when trend data are included.	•Focused on model predictions during early ICU days. •Includes all ICU patients, which may not represent a typical time-limited trial population.	Meiring et al24
An evaluation of the predictability of hospital mortality over time in the ICU, according to serial deep learning models. Conducted with an ICU data set from Massachusetts.	The performance of a deep learning model to predict hospital mortality improves hourly over the first 48 h of ICU care.	•Focused on model predictions during early ICU days. •Includes all ICU patients, which may not represent a typical time-limited trial population.	Deasy et al25
State-transition microsimulation decision analysis using four data sets (one test, three validation); limited to patients with both cancer and critical illness.	For patients with advanced cancer because of a solid tumor, a time-limited trial duration of up to 4 d may be sufficient.	•The optimal duration of a time-limited trial was defined by the investigators as: “How long does a trial have to be such that any longer does not increase the chance that [the patient] will be alive at 30 days?”	Shrime et al26
State-transition modeling for patients aged ≥ 80 y, using data from an international cohort study.	Uncertainty in the model’s prediction decreases over the first ∼7 d in the ICU and then becomes stable, suggesting time-limited trials should last at least 7 d.	•Focused on patients ≥ 80 y of age. •Small number of clinical variables in data set. •Model findings vary according to patient factors. •Optimal duration of a time-limited trial defined by investigators as the point at which prognostic uncertainty becomes constant.	Beil et al27
Retrospective cohort study, analyzing time to death in a cohort of older adults receiving mechanical ventilation after intubation in the ED.	The median time from intubation to death was 3 d, but this time varied among principal diagnosis subgroups.	•Focused specifically on older adults who are intubated in the ED. •Unable to discern whether underlying clinical condition or patient preferences influenced timing of end-of-life decisions.	George et al28

^aAs related to time-limited trials, given that time-limited trials were a secondary focus of some studies.

Evidence Review

Do We Agree About What a Time-Limited Trial Is?

In 2023, the American Thoracic Society³ sponsored a workshop committee to evaluate and build consensus on the definition of a time-limited trial in critical care. The committee conducted a modified Delphi process that included a survey study with 100 stakeholders representing 10 different countries and multiple critical care roles and professions. The committee identified a high level of consensus within the critical care community about how time-limited trials should be conducted. However, the committee found less consensus about why time-limited trials should be used. The Delphi process identified 16 essential elements for conducting a trial in critical care (Fig 2), and the committee established the following operational, consensus definition: “A time-limited trial is a collaborative plan among clinicians and a patient and/or their surrogate decision-maker(s) to use life-sustaining therapy for a defined duration, after which the patient’s response to therapy informs the decision to continue care focused on recovery, transition to care focused exclusively on comfort, or extend the trial’s duration.”³

Figure 2.

The 16 essential elements for conducting a time-limited trial in critical care, according to a modified Delphi process sponsored by the American Thoracic Society. (Reprinted with permission of the American Thoracic Society. Copyright © 2024 American Thoracic Society. All rights reserved. Kruser JM, Ashana DC, Courtright KR, et al. Defining the time-limited trial for patients with critical illness: an official American Thoracic Society Workshop Report, Ann Am Thorac Soc.2024;21(2):187-199.³ Annals of the American Thoracic Society is an official journal of the American Thoracic Society.)

Are Time-Limited Trials Part of Standard ICU Practice?

Observational studies from medical ICUs demonstrate that time-limited trials are part of current practice in the United States, with frequency of use ranging from 8% to 27% of patients in the ICU.^{15, 16, 17, 18, 19} In one study, investigators analyzed 72 audio-recorded family meeting discussions from five different ICUs in San Francisco and found time-limited trials were proposed in 13% of meetings.¹⁶ The optimal frequency of time-limited trials is unknown, but an ethnographic study in two US ICUs suggests their current use likely varies across medical ICUs based on a unit’s culture.¹⁵ In this study, the investigators purposively sampled one ICU that provided higher-intensity life-sustaining care for patients near death and one lower-intensity ICU, based on unit-level measures from Medicare data found in the Dartmouth Atlas. Investigators found both ICUs used time-limited trials, but the approach was more common in the lower-intensity ICU where 27% of sampled patients aged ≥ 65 years received a time-limited trial compared with 8% in the higher-intensity ICU.

These estimates of time-limited trial frequency are likely imprecise because they are based on two qualitative observational studies from 2012 and 2013 with small samples of highly selected patients. In both studies, time-limited trials were identified and classified as such post hoc by the investigators based on their study observations. Given that time-limited trials cannot be easily identified in electronic health records or administrative data sets, the only feasible method for measuring time-limited trials in clinical practice has been these resource-intensive, qualitative observations of care. To our knowledge, there have been no studies to estimate how often time-limited trials are used outside of the United States. In a 2007 observational study from France, an ICU trial strategy was described as part of the routine triage process for patients with cancer, suggesting the time-limited trial concept does extend beyond the United States.¹⁴

Evidence on the proportion of ICU physicians familiar with time-limited trials is limited to a 2022 study in two academic medical centers in the United States,¹⁸ in which 65% of medical ICU physicians reported being familiar with time-limited trials, and of those with familiarity, 77% had experience using time-limited trials in clinical practice. All the survey participants endorsed the time-limited trial approach as appropriate in some patients, after being provided with a definition.

How (and How Well) Are Time-Limited Trials Currently Used?

Although ICU physicians may be familiar with the time-limited trial approach, evidence suggests their use of this approach in clinical practice is highly variable. In the previously described 2022 ICU physician survey,¹⁸ investigators provided participants with hypothetical patient cases and found variation in how participating physicians would implement a time-limited trial for the same hypothetical patient. For example, 74% of the physicians felt a time-limited trial of invasive mechanical ventilation would be appropriate for the same patient with acute respiratory failure, and among those 74%, the proposed durations of how long the trial should last ranged from 3 to 7 days.

It is unsurprising that ICU physicians hold the gamut of views about the appropriateness of time-limited trials in individual patients because these trials are used to accomplish multiple and still-debated objectives and there are no agreed-on criteria for time-limited trial candidacy.³ In one of the only empirical studies examining time-limited trials outside of the United States, ICU physicians and nurses in the United Kingdom described multiple effects of the passage of time during a time-limited trial, including reducing their uncertainty about prognosis, building their relationships with patients and families, and preparing patients and families for end of life.⁶

Observational studies also suggest that time-limited trials, when used in current ICU practice, are often incomplete or not followed through. For example, in the previously described study of audio-recorded family meetings, investigators found that discussions about time-limited trials were often missing key elements.¹⁶ These discussions almost always included a plan for the trial’s duration and a description of the criteria for improvement or deterioration, but none of the meetings included a complete or explicit discussion about the potential decisions to be made at the end of the trial. In two other qualitative studies, interprofessional ICU physicians described participating in time-limited trials that are initiated but not followed through to completion.^17,19 The continuous rotation of ICU staff contributes to this lack of follow through by disrupting established time-limited trial plans.^17,19 However, we note that physicians in one of these studies described families requesting more time as a barrier to time-limited trial follow through,¹⁷ even though the option to extend a trial’s duration is an accepted, essential element of a time-limited trial.³

What Are Patients’ and Families’ Perspectives on Time-Limited Trials?

Few studies directly examine the perspectives of patients, their surrogate decision-makers, or other caregivers and family members on time-limited trials. However, the existing data suggest that some older adults would select a time-limited trial of critical care over the options of indefinite life-sustaining care or care immediately focused on comfort. In at least two studies, potential patients and actual patients’ family members have spontaneously suggested the concept of a time-limited trial, without prompting from investigators or physicians.^16,20 One of the first studies to directly assess patient preferences about time-limited trials (using that terminology) was published in 1998.¹ The investigators surveyed older adults living in nursing homes about their willingness to accept life-sustaining therapies in the event of a future critical illness. Across all surveyed life-sustaining therapies, time-limited trials were rated, on average, as more acceptable than indefinite, open-ended use of the therapy. For mechanical ventilation specifically, the study found most older adults would choose a time-limited trial over care focused exclusively on comfort and over open-ended use of the ventilator. A similar 2018 survey of surrogate decision-makers for people living with dementia in the Northeastern United States found that 31% of surrogates thought the patient would prefer a time-limited trial of life-sustaining therapies, 47% would prefer comfort-focused care, and 11% would prefer maximal life-sustaining therapies.⁵

A 2023 qualitative interview study with surrogates for patients with severe, acute brain injury suggests there are likely a range of patient and family perspectives about the time-limited trial approach. The surrogates who viewed themselves as having decision-making agency found time-limited trials helpful, whereas those who viewed themselves as playing a passive role in the unfolding of medical care did not.²¹

Does Prognostic Uncertainty About Patients With Critical Illness Decrease Over Time?

As previously discussed, one of the most commonly cited objectives for a time-limited trial is to decrease prognostic uncertainty by evaluating the patent’s response to therapy over time.^3,6,17 Two studies have sought to test this fundamental premise using predictive modeling techniques in large US and UK databases of patients admitted to the ICU. Investigators used machine and deep learning models to investigate whether the predictability of hospital mortality increases over time during an ICU admission.^24,25 Both studies found that model accuracy tends to improve over the first 48 h of a patient’s ICU admission, which was simulated in retrospective data sets by repeating the models with the cumulative and trend data that would be available at sequential time points during admission. These studies with a stated goal of evaluating the time-limited trial premise align with other predictive modeling work, demonstrating that trends in a patient’s Sequential Organ Failure Assessment score or Laboratory-Based Acute Physiology Score, version 2, over the course of an ICU stay are an important indicator of hospital outcome.^29,30 Taken together, these studies do suggest that prognostic uncertainty—understood through the lens of quantitative, predictive modeling—does decrease over the initial days of an ICU stay. However, prognostic uncertainty is a more complex notion than a predictive model’s accuracy. Uncertainty is sometimes categorized as either epistemic, which is incomplete knowledge that can be reduced by better models and more data, or aleatoric, which is inherent uncertainty that can never be reduced, even with perfect models and data.^31,32 The uncertainty about what will happen for patients with critical illness is undoubtedly both epistemic and aleatoric, and it is also held and experienced by humans—physicians, patients, and surrogates—who all bring their own lens to prognostic estimates.^{32, 33, 34, 35} The full answer to the question of whether and to what extent uncertainty decreases during a time-limited trial will need to account for these related complexities.

What Is the Optimal Duration of a Time-Limited Trial?

Several investigators have attempted to inform decisions about the duration of a time-limited trial through a similar quantitative approach with predictive modeling and descriptive time-to-death analyses.^{26, 27, 28} The underlying premise of these studies is that a trial should last until prognostic uncertainty is minimized. For example, one study used a state-transition model in a database of critically ill patients with cancer. They defined the optimal duration of a time-limited trial using the following question: “How long does a trial have to be such that any longer does not increase the chance that [the patient] will be alive at 30 days?” Using this definition and modeling strategy, they found the optimal duration of a time-limited trial for patients with advanced cancer due to a solid tumor is 4 days. This study and a similar approach in older ICU patients²⁷ help characterize the time course of prognostic estimates for patients with critical illness, but they cannot fully answer the question of how long a time-limited trial should last. In the American Thoracic Society workshop report,³ Delphi survey participants agreed that patients and surrogates also have a stake in determining a trial’s duration. For example, a patient may have a maximum acceptable duration of life-sustaining therapy, after which the benefit of additional statistical certainty is not worth the burdens of treatment.

We can be fairly confident that models predicting whether a patient presenting to the ICU will survive critical illness become more accurate after accumulating several days of clinical data. However, these quantitative findings have methodologic limitations and do not yet address the heterogeneous clinical trajectories of patients with critical illness. Furthermore, even the best possible model to minimize statistical uncertainty can only inform the process of setting the duration of a time-limited trial, which is a collaborative process between physicians, patients, and surrogates and relies on individual patients’ priorities and value judgments.

Do Time-Limited Trials Improve Outcomes?

We have limited evidence about whether and to what extent time-limited trials impact care processes and outcomes for patients with critical illness and their families. To our knowledge, only one prospective study has sought to compare usual care with an intervention to support time-limited trials.²³ This study was a pre-post evaluation of a quality improvement program in three California hospitals, which included training ICU physicians to communicate about time-limited trials with surrogates. After implementing the training program, family meetings were more common, occurred earlier in an ICU stay, and were more likely to include discussion elements indicative of high-quality communication. Hospital mortality was unchanged, and although the primary clinical end point of ICU length of stay was reduced, this end point was analyzed using linear regression which is known to be biased because of its failure to account for informative censoring due to death.³⁶ The intervention was not associated with changes in surrogates’ satisfaction with ICU care nor their decision-making conflict. Although these results are largely supportive of time-limited trials, this single before-after quality improvement study cannot determine whether time-limited trials caused the observed changes or whether the results were primarily driven by earlier, more common, or higher quality family meetings alone.²² Additionally, the investigators did not report whether time-limited trials were actually offered during the family meetings, whether surrogates agreed to participate in offered time-limited trials, the details of the time-limited trials that were planned, or the extent to which time-limited trials plans were carried out during the hospital stay.

What Are the Ethical Considerations About Time-Limited Trials?

Some of the earliest mentions of time-limited trials or similar concepts like the ICU trial³⁷ situated this care approach within the longstanding bioethical discussion about withholding and withdrawing life-sustaining care. Ethicists generally agree there is no ethical distinction between withdrawing and withholding this care, but they also acknowledge important pragmatic and psychological challenges that can create a distinction in practice.^38,39 Withholding life-sustaining care based on inadequate prognostic information can lead to preventable deaths, whereas physicians’ and families’ feelings of moral asymmetry and cognitive biases can make it difficult to withdraw care once it is already started (even if ethically justified and aligned with the patient’s priorities).^{40, 41, 42} Some ethicists have promoted time-limited trials as an ethical approach to overcome both of these challenges,^38,39 and many hospital ethics committees recommend and support time-limited trials, particularly in the context of decisions about extracorporeal membrane oxygenation.^43,44 Ethical support for time-limited trials is not uniform, with one scholar arguing that using time-limited trials for patients with low chance of survival exposes patients to the harms of life-sustaining therapies without sufficient chance of benefit.^{45, 46, 47, 48} However, without widely accepted thresholds defining what qualifies as a low chance of benefit, and in the context of caring for diverse populations of patients and families with heterogeneous beliefs about the value of critical care,^49,50 it is hard to envision applying this view in contemporary clinical practice.

It is also noteworthy that at the height of the COVID-19 pandemic, some ethicists argued for time-limited trials as a pragmatic approach to allocate scarce health care resources.^39,51,52 Truog,⁵¹ for example, proposed time-limited trials as an approach to overcome the arbitrariness of the first-come, first-served strategy of scarce resource allocation. Using time-limited trials as an allocation strategy, mechanical ventilation would be offered to those presenting with respiratory failure, but if certain criteria of improvement are not met within a defined period, the ventilator would be withdrawn to allocate it to another patient. Importantly, Truog⁵¹ draws a clear distinction between this allocation strategy and the clinical use of time-limited trials, which—unlike resource allocation protocols—represent collaborative processes that requires agreement by patients and families.

This narrative review has limitations. We focused our search strategy on the term time-limited trial to compare articles referring to a shared construct. However, we acknowledge that other terms are used for the same or similar construct, and some articles using alternate terms (eg, ICU trial) were added to this review based on our review of reference lists. Furthermore, we acknowledge the importance of adjacent evidence focused on prognostication, patient/family/physician communication, and outcomes research for time-limited trials; however, its inclusion in this review was beyond our scope.

Summary

We have a relatively sparse evidence base for time-limited trials, with only 18 empirical studies identified over 25 years that have explored, often indirectly, time-limited trials in patients with critical illness. The existing evidence, which primarily reflects academic medical ICUs within the United States, suggests that time-limited trials are part of current practice, but their use varies based on unit and physician factors. When time-limited trials are used, they are implemented incompletely and challenged by system-level barriers (eg, continuous rotation of ICU staff). Some older adults appear to prefer time-limited trial strategies over unrestricted ICU care or comfort-focused care when presented with hypothetical scenarios, but the experiences and perspectives of actual patients and surrogates considering or using time-limited trials are poorly understood. Prevailing wisdom that prognostic uncertainty tends to decrease over time in the ICU is supported by predictive modeling studies in large data sets of ICU patients, but the questions of decrease for whom, how much, over what period, and how to use this knowledge in support of time-limited trials remain unsettled. Only one study has prospectively compared usual ICU care with an intervention designed to promote and support time-limited trials, with promising but not definitive results. Overall, this synthesis of the evidence demonstrates that time-limited trials are part of current practice in critical care, but substantial efforts will be needed to improve their use, test their impact, and maximize their beneficial impact.

Funding/Support

J. M. K. was supported in part by the National Institutes of Health (NIH)/National Heart, Lung, and Blood Institute (NHLBI) [Grants K23HL146890, R01 HL168474]. E. M. V. was supported in part by the NIH/NHLBI [Grant K23HL157364].

Financial/Nonfinancial Disclosures

The authors have reported to CHEST the following: J. M. K.’s spouse receives honoraria for lectures and speakers bureaus from Astra Zeneca and for consulting from Radialogica. None declared (N. R. N., E. M. V., J. T. C., K. E. S., S. D. H.).