Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Crit Care Med. Author manuscript; available in PMC 2014 February 10.
Published in final edited form as:
PMCID: PMC3918894

Neglect of quality-of-life considerations in intensive care unit family meetings for long-stay intensive care unit patients

Sara L. Douglas, PhD, RN, Barbara J. Daly, PhD, FAAN, and Amy R. Lipson, PhD



To examine the frequency with which quality of life and treatment limitation were discussed in formal family meetings for long-stay intensive care unit patients with high risk for mortality and morbidity.


Descriptive observational study.


Five intensive care units.


One hundred sixteen family surrogate decisionmakers of long-stay intensive care unit patients who participated in an intensive communication system that aimed to provide weekly meetings with family decisionmakers. The structure of each meeting was to address medical update, patient preferences, treatment plan, and milestones for evaluating the treatment plan.



Measurements and Main Results

We audiotaped initial family meetings for 116 family decisionmakers for a total of 180 meetings. On average, meetings were 24 mins long with a majority of time being devoted to nonemotional speech and little (12%) spent discussing patient preferences. Quality of life was discussed in 45% and treatment limitation in 23% of all meetings. Quality-of-life discussions were more likely to occur when patients were admitted to a medical intensive care unit (odds ratio [OR], 5.9; p = .005), have a family decisionmaker who is a spouse (OR, 9.4; p = .0001), were older (OR, 1.04; p = 01), have a shorter length of stay (OR, 0.93; p = .001), and have a family decisionmaker who was a spouse (OR, 5.1; p = .002). For those with a treatment limitation discussion, 67% had a do-not-resuscitation order, 40% were admitted to a medical intensive care unit, 56% had a family decisionmaker who had been their caregiver, and 48% of their family decisionmakers were their children.


To guide discussion with families of the subset of intensive care unit patients with high risk of mortality and long-term morbidity, quality of life was not consistently addressed. Continued efforts to assist clinicians in routinely including discussions of quality-of-life outcomes is needed.

Keywords: communication, decisionmaking, long-stay intensive care, quality of life, treatment limitation

The challenges of communication with families of long-stay intensive care unit (ICU) patients and surrogates facing end-of-life decisionmaking have been a focus on numerous investigations and professional guidelines (15). Despite this attention, research continues to reflect the difficulty of implementing consistent approaches that are effective in addressing informational and support needs of surrogate decisionmakers. This is particularly true in situations in which there is considerable prognostic uncertainty and questions about treatment limitations, a common occurrence for long-stay ICU patients.

Regular family meetings have been recommended as an effective approach for assuring that families receive updated information, have an opportunity to get questions answered, and receive support for shared decisionmaking with the clinical team (6). Although reports suggest having a system of regular meetings can facilitate decisionmaking, minimize use of ineffective interventions, and improve family satisfaction (710), there is evidence that style and elements of communication (e.g., topics, empathic statements) are equally important. Curtis (11) reported that in almost one-third of family conferences where bad news was discussed, there were missed opportunities to provide information or support. We also reported on a trial of an “intensive communication system” of regular weekly meetings with families of long-stay ICU patients (12) where meetings were not associated with length of stay, aggressiveness of care, or resource use.

In this report we describe findings from the audio recordings of 116 family meetings held as part of a larger parent study that examined the impact of a structured family meeting intervention on ICU resource use. Description and outcomes from the parent study have been presented elsewhere (12). For the descriptive study presented here, we were interested in examining the frequency with which quality of life (QOL) was discussed because of the uncertain trajectory of recovery and high mortality and morbidity associated with the long-stay ICU patient population. As part of this investigation, we wanted to describe the components of the conversation, including the frequency of emotionally oriented comments and specific meeting topics (e.g., patient preferences, medical update, prognosis) and to examine whether there were patient or family decisionmaker (FDM) characteristics that were associated with discussions of QOL. A secondary purpose was to describe the extent of treatment limitation discussion and key characteristics associated with those discussions.


The parent study from which these data were obtained was conducted in five ICUs in two academic medical centers in Cleveland, OH, from November 2005 through April 2006. Study sites included a surgical intensive care unit, medical intensive care unit, and neuroscience ICU at a university-affiliated, private not-for-profit medical center and a medical intensive care unit and surgical intensive care unit at a university-affiliated public medical center in the same city. Institutional review board approvals from the two hospitals were obtained before study initiation.


Eligibility criteria for patients were: 1) age ≥18 yrs; 2) 72 hrs of mechanical ventilation and no expectation by the attending physician of extubation or discharge from the ICU within the next 48 hrs; 3) lacking decisional capacity as indicated by a Glasgow Coma Scale <6 and confirmed by the ICU attending physician; 4) not on mechanical ventilation before admission; and 5) having an identified family surrogate decisionmaker present. Family surrogates were eligible if they were 1) identified as the appointed surrogate through a power of attorney or were the next of kin; 2) age ≥18 yrs; 3) available for participation in family meetings; and 4) willing to provide informed consent for participation in the study. In the case of multiple family members, the decisionmaker was defined as the person from whom consent was obtained for medical interventions.

Data Collection

As part of the parent study, the five ICUs implemented a change in standard practice and began weekly structured family meetings for all long-stay ICU patients. This was explained to all potential subjects and the experimental nature of the study was clarified before written consent was obtained. An Advanced Practice Nurse in each hospital was responsible for scheduling and participating in the family meetings to assure that the communication structure was implemented consistently. A formal training session was held for each ICU physician group before instituting the intervention.

The intensive communication system structure, reviewed in all training sessions, included: 1) family meeting within 5 days of ICU admission (or as soon as could be arranged) and at least weekly thereafter as long as the patient remained in the ICU; and 2) each meeting was to address medical update, values and preferences of the patient, goals of care, treatment plan, and milestones for determining if the treatment plan was effective. There was no attempt to alter the communication style of the participants or direct the decisions made.

Advanced Practice Nurses met each patient and family member as part of the family’s routine orientation to the ICU environment. When the patient met eligibility criteria, the Advanced Practice Nurse explained to the family that having routine, once-a-week family meetings, in addition to the informal bedside updates from doctors and nurses, was the standard communication structure of the ICU for long-stay patients but that they were being asked to participate in a study aimed at evaluating its effectiveness. For those who agreed to participate, a research nurse approached the family member, explained the study, and obtained consent. At the end of each meeting, a tentative date that was convenient for the family was set for the next meeting. All families were asked for signed permission to audiotape the meeting. If any family member or clinician refused permission, the meeting was not taped. A research nurse was responsible for taping the meeting.


Demographic and clinical information about each patient was abstracted from their medical records. This included age, gender, race, Acute Physiology and Chronic Health Evaluation (APACHE) III (13), major diagnostic category, and use of interventions that reflected resource use or goals of care such as placement of tracheostomy. Treatment limitation decisions, hospital and ICU lengths of stay, discharge disposition, and ICU and hospital survival were recorded. The APACHE III is an established tool for defining the severity of illness within the first 24 hrs in the ICU and has excellent psychometric properties (13). Data from the APACHE III were used to derive the Acute Physiology Score, which uses basic physiologic variables (APACHE score − [age score + chronic illness score]) and is considered better than the APACHE II at comparing patients with different diseases (14). Location at 2 months after discharge was obtained through follow-up telephone calls to family members. Interrater reliability for medical record abstraction was assessed before data collection began and every 4 months throughout the data collection period. Pearson’s correlations ranged from .6 to 1.0 for continuous variables and κ’s ranged from .6 to 1.0 for categorical variables.

Each audiotape was reviewed by two research staff members and audited for the following: length of meeting, topics discussed and time spent on each topic, percent of speaking time by participants, and presence or absence of discussion of QOL and treatment limitation. Explicit criteria for what constituted a QOL or treatment limitation discussion was used for the coding of each audiotape (Table 1). In classifying QOL limitation topics, we restricted our criteria to those elements of the discussion that were clearly intended to direct the conversation to references to how the patient would experience, respond to, or evaluate potential outcome states. Prognostic statements about specific organ function without description or inquiry about what this level of function would mean to the patient (i.e., would it be acceptable, would it represent an intolerable state, or one not worth pursuing) were not classified as QOL discussion. We chose this approach based on the work of Testa who identified that the individual’s subjective perception of the objective health status was what constituted QOL and satisfaction with life (15). In classifying treatment limitation topics we identified any question, suggestion, or mention of the option to limit life-sustaining interventions or to withdraw current interventions.

Table 1
Criteria for coding quality of life and treatment limitation discussion

Before data analysis, interrater reliability of QOL and treatment limitation coding was conducted between the Principal Investigators (B.J.D., S.L.D.) and the research assistant responsible for listening to and coding all tapes. Interrater reliabilities conducted for randomly selected taped meetings throughout the coding process yielded acceptable κ’s (.6–1.0) and correlations (.8–.9) (16, 17).


Frequencies and measures of central tendency were used to describe the sample. When comparing those with and without QOL discussions and treatment limitation discussions, independent samples t tests for continuous variables and chi square for categorical variables were used. Effect sizes were calculated for all statistically significant differences.

We also conducted logistic regression analysis to examine relationships of FDM and patient characteristics to QOL (0 = not discussed, 1 = discussed). Each dependent outcome variable was defined using the criteria outlined in Table 1. Because there was no prior research to base hypotheses for testing, and because we were not interested in causality but in a more exploratory approach, we used backward stepwise logistic regression (18, 19) for examining predictors of QOL discussions. We included patient and FDM variables that were either significant in our univariate analyses or had been shown in prior work to relate to QOL discussion (2022). There were no multicollinearity or independence of observation concerns and the logistic model assumption of linear relationships between the logit and the continuous covariates was verified for both logistic models. Then, the backward elimination method was used with the cutoff for elimination being a p value of .10. For all analyses (univariate and multivariate), p < .05 was considered to be statistically significant. The statistical program, SPSS Version 19 (SPSS, Chicago, IL), was used for all analyses.


One hundred sixteen subjects’ first taped meetings constitute the sample. A majority (69%) of subjects had only one taped meeting (range, 1–4) throughout their ICU stay and the first meeting took place, on average, 9.75 (7.9) days after the patient entered the ICU. As seen in Table 2, patients were middle-aged with moderate APACHE III scores and long hospital and ICU stays. Less than one-fourth died in the ICU with the cumulative 2-month posthospital mortality rate being 31%. For those who survived their hospital stay (n = 92), 92.4% were discharged to an institutional setting. Two months later, 35 (41.2%) of the 84 patients discharged to an institution were still residing in an extended-care facility and ten (11.8%) had died.

Table 2
Family decisionmaker, patient, and meeting characteristics (n = 116)

FDMs were more likely to be female, a spouse, new to caregiving, and employed. Despite long ICU stays, FDMs had an average of only 2.1 (1.3) formal meetings/patient. Meetings tended to be moderate in length with a majority of meeting time spent in instrumental speech where factual information (rather than relational or emotional topics) was discussed. A majority (71.5%) of all meetings had an attending physician participating (the remaining 28.5% of meetings had a fellow or resident); physicians were predominantly male (88.5%), white (83.5%), and highly experienced (median years as attending physician = 18).

Quality of Life

Less than half (44.8%) of first family meetings had a discussion of QOL. As seen in Table 3, patients whose families had a family meeting where QOL was discussed were more likely to have been admitted to a medical intensive care unit and also to have a treatment limitation discussion and their FDMs were more likely to have been a prior caregiver and to have been a spouse or child of the patient. Advanced Practice Nurses were most likely to be the first to mention the issue of QOL (59.6%) followed by the physician (23.1%) and family (17.3%). We also found that those who died were three times more likely to have had a QOL discussion than those who lived.

Table 3
Characteristics of those with and without quality-of-life discussion at first meeting (n = 116)

Meetings where QOL was discussed tended to be longer and have a greater percentage of time devoted to discussing preferences. In addition, family members spoke a greater percentage of time and there was less time spent discussing medical update in those meetings where QOL was discussed (Table 3). As seen in Figure 1, 18.2% of meeting time was spent, on average, discussing patient preferences for those meetings where QOL was discussed—much less than time spent on medical update, prognosis, or treatment plans. In addition, as seen in Figure 2, although 66.7% of patients who died had a discussion of QOL, only 39.3% of patients who were discharged to an institutional setting had such a discussion at their first meeting (p = .007).

Figure 1
Comparison of percent time spent discussing key topics for first meetings where quality of life was discussed (n = 52) and first meetings where treatment limitation was discussed (n = 27). QOL, Quality of life; Tx Limit, treatment limitation; Tx, treatment. ...
Figure 2
Percentage of subjects who died inhospital (n = 24) or were discharged to an institutional setting (n = 85) who had either a quality of life or treatment limitation at their first family meeting. QOL, Quality of life; Treat Limit, treatment limitation. ...

Finally, we were interested in exploring the relationships between patient and FDM characteristics and QOL discussions. To examine these relationships, we conducted backward logistic regression in which QOL (0 = not discussed, 1 = discussed) was regressed on seven independent variables (patient age, Acute Physiology Score, length of hospital stay, do-not-resuscitate order, admitting unit, FDM relationship, and prior caregiver experience). The Hosmer-Lemeshow test indicated that the calculated model fit the actual data for estimating the outcome variable (chi square [8] = 3.7, p = .88). As seen in Table 4, the most parsimonious model had four variables with all of them making statistically significant contributions (Nagelkerke R2 = .40) with correct classification of 74.3%.

Table 4
Logistic regression of quality-of-life discussions on patient and caregiver variables (n = 113)a

Treatment Limitation

Less than one-third (23.3%) of all meetings had a discussion of treatment limitation with the odds of dying being 5.1 times more likely (95% confidence interval, 1.9–13.4; p = .001) if there was a treatment limitation discussion at the first family meeting. As seen in Table 5, patients who had discussion of treatment limitation in the first meeting were older, sicker, more likely to have a do-not-resuscitate order written, and more likely to have been admitted to a medical intensive care unit. FDMs of patients who had a treatment limitation discussion were more likely to have been a prior caregiver.

Table 5
Characteristics of those with and without treatment limitation discussion at first meeting (n = 116)

Meetings where treatment limitations were discussed presented more discouraging news as rated by both the physician and FDM, had a greater emphasis on discussions of patient preferences, and spent less time discussing medical update than those meetings without treatment limitation discussions. In over half of the meetings in which treatment limitations were discussed, the physician (70.4%) was the first individual to mention the topic and family were never first to mention treatment limitations. As seen in Figure 1, the second greatest amount of time (25.7%) was spent discussing patient preferences in meetings where treatment limitations were discussed and for patients who died inhospital, 50% of the first family meetings involved a discussion of treatment limitation.


In this study, despite the implementation of a protocol calling for formal meetings every 5 days, we found that, on average, patients had only two meetings during their 17-day stay in the ICU. This reflects the practical challenges of scheduling meetings around both physician and family schedules. For long-stay ICU patients, it was not uncommon for family members to have returned to work or other responsibilities, constraining their ability to attend meetings during the day when the attending physician and Advanced Practice Nurse were available.

More importantly, we found that fewer than half of the first family meetings had discussions of QOL or patient preferences and even in the “best case” scenarios where QOL was discussed, less than one-fourth (16.8%) of those meetings mentioned patient preferences. Our results are in contrast to that of White, who reported that 86% of taped family meetings addressed expected functional status (23). This finding may be the result of differences in criteria used in coding speech in that they classified these comments as an aspect of prognosis, whereas we specifically looked for statements that addressed what the patient would experience rather than objective descriptions of respiratory status or ability to perform specific functions. In addition, we found that the topic of QOL was more likely to be introduced by a healthcare provider and was most often discussed for those patients with the greatest likelihood of inhospital death. Despite guidelines for patient-centered communication (4), we found that our meetings were still heavily focused on the traditional approach in which healthcare providers spoke a majority of the time, family members primarily listened or asked questions, and very little time was devoted to QOL or emotionally supportive comments.

The low incidence of discussion of QOL for long-stay ICU patients highlights an area of concern and identifies an opportunity for improvement. These patients are a subgroup of ICU patients with increased mortality and morbidity and a high likelihood of continued health problems, readmission, and decrement in QOL compared with their preadmission status (2428). Discussion of the impact of treatment, expected care needs after hospitalization, and postdischarge QOL have been found by others to be as important to patients and families as survival estimates (29, 30). It is particularly important for the treatment team to explicitly address prognosis for relevant QOL dimensions and, in many cases, patient preferences for continued use of aggressive life support technologies.

Our findings reflect a clinical norm of linking QOL discussions to situations in which prognosis is poor and risk of death is high. However, regardless of survival prospects, the relevance of discussing patient values and goals early in the course of treatment is crucial for all of these FDMs and suggests little reason to avoid the topic of QOL. For those patients who are clinically stable, it would be appropriate to make a simple statement that explains that all patients who require prolonged critical care are likely to have a prolonged and somewhat uncertain posthospital recovery that may involve continued limitations in functional status. Examining QOL preferences will be helpful as the family start to envision life after the ICU. For those patients who are unstable, it is important to introduce the concepts of patient preferences early so that the family can begin to have a discussion of possible decisions that might need to be made should the patient’s condition deteriorate or an unacceptable outcome become apparent. To be supported in thinking about and discussing these options before a crisis emerges is an opportunity that will benefit patients and their families alike. Without discussions related to preferences, it becomes impossible for families and the healthcare team to share in the decision about life-sustaining treatment—a goal expressed by families as well as healthcare providers (11, 3134).

The topic of treatment limitation is more sensitive than QOL and was likely not seen as relevant in situations in which the prognosis for survival seemed good. Nevertheless, given the relatively high incidence of postdischarge death and hospital readmission in this population, it is likely that the FDMs will continue to be faced with decisions about aggressiveness of care sometime after discharge. A study by Nelson (29) confirmed that long-stay ICU patients and their surrogates identify discussion of alternatives to treatment as an important domain in communication. Raising the issue of treatment limitation in the ICU as an option that might be appropriate at a later time, depending on the illness trajectory, could be a helpful and supportive educational intervention.

There are several limitations to the present study. First, we were not able to audiotape meetings for all subjects. Although we found no differences on the variables that we assessed (age, gender, ethnicity), there was little variability for these variables and there are many other relevant characteristics that we were not able to assess (e.g., comfort level with communicating with family or healthcare providers, religious beliefs, etc.). Second, we were limited to assessing only the communication that occurred during taped family meetings. Communication that occurred outside of this area (at the bedside, over the telephone, etc.) was not assessed and likely had an impact on the quantity, quality, and type of communication that occurred during the family meetings. In addition, we report only on first family meeting and are unable to extrapolate findings to subsequent meetings or to those who had more than one formal family meeting. A fourth limitation was that the criteria used for identifying QOL and treatment limitation discussions, although reliable over time and among coders, were developed by the study investigators. As noted in reference to the study by White (23, 31), other clinicians and researchers may conceptualize these variables differently, thus obtaining different results. Finally, our sample of FDMs primarily was comprised of middle-aged, white individuals and our sample of physicians was primarily male whites with extensive ICU experience and results cannot be generalized to all ICU FDMs or physicians.


Despite the use of a structured communication protocol that included discussion of values and preferences at each meeting, less than half of all first meetings had any discussion of QOL issues. There is an expectation that hospital personnel inquire of all patients or their surrogates whether they have an Advance Directive; the discussion of QOL issues for long-stay ICU patients should be as commonplace and should be considered part of the standard approach to communicating with critical care unit families. Including information about the likely post-ICU outcomes for long-stay ICU patients in the education and training of ICU physicians and nurses could serve to expand their understanding of the long-term mortality and morbidity associated with such patients and the importance of anticipatory guidance. Continued education and training for ICU healthcare providers is needed to assist them in developing more comfort with having meaningful conversations with families about patient preferences and realistic treatment goals.


The authors thank Mary Leuchtag for her role in the audiotape coding process.

This study was funded by a grant from the National Institute of Nursing Research, Grant Number RO1-NR8941.


The study was performed at Case Western Reserve University, University Hospitals Case Medical Center, and MetroHealth Medical Center, Cleveland, OH.

The authors have not disclosed any potential conflicts of interest.

For information regarding this article, SLD4/at/


1. Lautrette A, Ciroldi M, Ksibi H, et al. End-of-life family conferences: Rooted in the evidence. Crit Care Med. 2006;34(Suppl):S364–S372. [PubMed]
2. Curtis JR, Patrick DL, Shannon SE, et al. The family conference as a focus to improve communication about end-of-life in the intensive care unit: Opportunities for improvement. Crit Care Med. 2001;29(Suppl):N26–N33. [PubMed]
3. Nelson JE, Walker AS, Luhrs CA, et al. Family meetings made simpler: A toolkit for the intensive care unit. J Crit Care. 2009;24:626.e7–e14. [PubMed]
4. Davidson JE, Powers K, Hedayat KM, et al. Clinical practice guidelines for support of the family in the patient-centered intensive care unit: American College of Critical Care Medicine Task Force 2004–2005. Crit Care Med. 2007;35:605–622. [PubMed]
5. Gay EB, Pronovost PJ, Bassett RD, et al. The intensive care unit family meeting: Making it happen. J Crit Care. 2009;24:629.e1–e12. [PMC free article] [PubMed]
6. Curtis RJ, Vincent JL. Ethics and end-of-life care for adults in the intensive care unit. Lancet. 2010;375:1347–1353. [PubMed]
7. Lilly CM, DeMeo DL, Sonna LA, et al. An intensive communication intervention for the critically ill. Am J Med. 2000;109:469–475. [PubMed]
8. Schneiderman LJ, Gilmer T, Teetzel HD, et al. Effect of ethics consultations on nonbeneficial life-sustaining treatments in the intensive care setting. JAMA. 2003;290:1166–1172. [PubMed]
9. Campbell ML, Guzman JA. Impact of a proactive approach to improve end-of-life care in a medical ICU. Chest. 2003;123:266–271. [PubMed]
10. Lautrette A, Darman M, Megarbane B, et al. A communication strategy and brochure for relatives of patients dying in the ICU. N Engl J Med. 2007;356:469–478. [PubMed]
11. Curtis JR, Engelberg RA, Wenrich MD, et al. Missed opportunities during family conferences about end-of-life care in the intensive care unit. Am J Respir Crit Care Med. 2005;171:844–849. [PubMed]
12. Daly BJ, Douglas SL, O’Toole E, et al. Effectiveness trial of an intensive communication structure for families of long-stay ICU patients. Chest. 2010;138:1340–1348. [PubMed]
13. Knaus W, Wagner D, Draper EA, et al. The APACHE III prognostic system: Risk prediction of hospital mortality for critically ill hospitalized adults. Chest. 1991;100:1619–1636. [PubMed]
14. Matis G, Chrysou O, Birbilis T, et al. Acute Physiology Score prediction in patients with traumatic brain injury. Pneumon. 2008;21:374–379.
15. Testa MA, Simonson DC. Assessment of quality-of-life outcomes. N Engl J Med. 1996;334:835–840. [PubMed]
16. Bland J, Altman D. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1988;8:307–310. [PubMed]
17. Landis J, Koch L. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159–174. [PubMed]
18. Menard S. Applied Logistic Regression Analysis. Sage; Thousand Oaks, CA: 1995.
19. Field A. Discovering Statistics Using SPSS. Third Edition Sage; Washington, DC: 2009.
20. Meissner A, Genga KR, Studart FS, et al. Epidemiology of and factors associated with EOL decisions in a surgical intensive care unit. Crit Care Med. 2010;38:1060–1068. [PubMed]
21. Holley A, Kravet SJ, Cordts G. Documentation of code status and discussion of goals of care in gravely ill hospitalized patients. J Crit Care. 2009;24:288–292. [PubMed]
22. Cook D, Rocker G, Marshall J, Griffith L, et al. Levels of care in the intensive care unit: A research program. Am J Crit Care. 2006;15:269–279. [PubMed]
23. White DB, Engelberg RA, Wenrich MD, et al. Prognostication during physician–family discussions about limiting life support in intensive care units. Crit Care Med. 2007;35:442–448. [PubMed]
24. Douglas SL, Daly BJ, Gordon N, et al. Survival and quality of life: Short-term versus long-term ventilator patients. Crit Care Med. 2002;30:2655–2662. [PubMed]
25. Chatila W, Kreimer D, Criner G. Quality of life in survivors of prolonged mechanical ventilatory support. Crit Care Med. 2001;29:737–742. [PubMed]
26. Friedrich JO, Wilson G, Chant C. Long-term outcomes and clinical predictors of hospital mortality in very long stay intensive care unit patients: A cohort study. Crit Care. 2006;10:1–9. [PMC free article] [PubMed]
27. Nelson JE, Cox CE, Hope AA, et al. Chronic critical illness. Am J Respir Crit Care Med. 2010;182:446–454. [PMC free article] [PubMed]
28. Daly BJ, Douglas SL, Gordon NH, et al. Composite outcomes in the chronically critically ill: Survival, location, and cognitive status at four months. Am J Crit Care. 2009;18:456–464. [PMC free article] [PubMed]
29. Nelson JE, Kinjo K, Meier DE, et al. When critical illness becomes chronic: Informational needs of patients and families. J Crit Care. 2005;20:79–89. [PubMed]
30. Lloyd CB, Nietert PJ, Silvestri GA. Intensive care decision making in the seriously ill and elderly. Crit Care Med. 2004;32:649–654. [PubMed]
31. Carlet J, Thijs LG, Antonelli M, et al. Challenges in end-or-life care in the ICU. Statement of the 5th international consensus conference in critical care: Brussels, Belgium, April 2003. Intensive Care Med. 2004;30:770–84. [PubMed]
32. Heyland DK, Rocker GM, O’Callaghan CJ, et al. Dying in the ICU: perspectives of family members. Chest. 2003;124:392–397. [PubMed]
33. Evans LR, Boyd EA, Malvar G, et al. Surrogate decision-makers’ perspectives on discussing prognosis in the face of uncertainty. Am J Respir Crit Care Med. 2009;179:48–53. [PMC free article] [PubMed]
34. Nelson JE, Angus DC, Weissfeld LA, et al. End-of-life care for the critically ill: A national intensive care unit survey. Crit Care Med. 2006;34:2547–2553. [PubMed]