Quality of life 9 years after an intensive care unit stay: A long-term outcome study

Journal of Critical Care (2011) 26, 379–387

Quality of life 9 years after an intensive care unit stay: A long-term outcome study☆,☆☆,★ Kay H. Stricker MD a,b,1 , Susanne Sailer MD c,1 , Dominik E. Uehlinger MD d , Hans U. Rothen MD, PhD b , Regula M. Zuercher Zenklusen MD e , Sonia Frick MD f,⁎ a

Department of Anaesthesiology, Bern University Hospital and University of Bern, 3010 Bern, Switzerland Department of Intensive Care Medicine, Bern University Hospital and University of Bern, 3010 Bern, Switzerland c Department of Surgical Intensive Care Medicine, Zürich University Hospital and University of Zürich, 8091 Zurich, Switzerland d Department of Nephrology, Bern University Hospital and University of Bern, 3010 Bern, Switzerland e Medical-Surgical ICU, Hospital Neuchâtel-Pourtalès, 2000 Neuchâtel, Switzerland f Medical-Surgical ICU, Triemli Hospital, 8063 Zürich, Switzerland b

Keywords: Outcome assessment; Quality of life; Intensive care; Mortality; Sickness Impact Profile; Prognosis

Abstract Purpose: The purpose of the study was to assess long-term mortality after an intensive care unit (ICU) stay and to test the hypotheses that (1) quality of life improves over time and (2) predictions of outcome made by caregivers during an ICU stay are reliable. Materials and methods: Data from a 6-bed university medical ICU were reviewed. Telephone assessment of mortality and interviews/questionnaires 9 years after an ICU stay were performed. Comparison of caregivers' predictions of survival/quality of life with reported outcome was done. Results: Of 409 patients surviving 6 months after ICU, 334 were included and 146 of these had died. Age, diagnostic group, and severity of illness were significant factors for mortality (P b .0001 for all 3). Of all survivors, 59% described their overall quality of life as good and 35% as fair. Physical dependency was significantly related to length of hospital stay (P b .01), whereas quality of life was related to admission age (P b .05). Caregivers' predictions concerning both survival and quality of life seemed reliable, with physicians' predictions being more reliable than nurses' (P b .05). Conclusions: Mortality is high 9 years after ICU stay. Quality of life may deteriorate for some individuals; however, overall quality of life for most survivors remains acceptable and may even improve. Long-term outcome predictions made by caregivers during the ICU stay seem accurate. © 2011 Elsevier Inc. All rights reserved.

☆

Name of institution where work was performed: Department of Intensive Care Medicine (formerly Medical ICU), Bern University Hospital, Bern, Switzerland. ☆☆ Conflict of interest: No conflicts of interest have been reported by any of the authors. ★ No financial or other support was received for the design of the study, the evaluation of data, or the writing of the manuscript. ⁎ Corresponding author. Triemlispital Zürich, Birmensdorferstrasse 497, 8063 Zürich, Switzerland. Tel.: +41 44 466 12 92; fax: +41 44 466 21 06. E-mail address: [email protected] (S. Frick). 1 Contributed equally to the manuscript. 0883-9441/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jcrc.2010.11.004

380

1. Introduction The financial burden facing today's health care systems has increasingly focused attention on resource allocation and outcomes after critical care [1-4]. Health status and quality of life (QOL) after an intensive care unit (ICU) stay have joined decreasing mortality as relevant outcome measures for assessing the effectiveness of ICU therapy [3,5-7]. However, predicting outcome during an ICU stay remains difficult because of various confounding pre-, intra- and post-ICU factors [2,5,8-11]. Several studies have been published about short-term outcome—that is, mortality and QOL—after ICU stay [12-14]. However, there are few data repeatedly assessing ICU outcome of the same ICU patients over prolonged periods of time [15-17]; and a recent editorial pointed out the importance of long-term follow-up of patients beyond the doors of our units [6]. In addition, no data exist to assess caregivers' prediction of outcome over very long periods of time. Ten years ago, we investigated mortality and QOL 6 months after ICU admission, as well as accuracy of outcome prediction by caregivers, that is, critical care nurses and physicians [10]. The ICU and in-hospital mortality were 8.6% and 9.4%, respectively; and postdischarge mortality was 3.5%. Overall quality of life (OQOL) and health-related quality of life (HRQOL) at that time were very good. It was not known whether the good short-term outcome results in this group of former ICU patients would persist over a prolonged period. Furthermore, the restricted accuracy of short-term prognostication might improve over prolonged periods. Thus, the primary aims were to assess the survival rate 9 years after ICU admission and to test the hypothesis that QOL improves over a prolonged period following ICU discharge. The secondary aim was to test the hypothesis that caregivers' predictions of patients' long-term outcome during an ICU stay are reliable.

2. Methods All German-speaking patients 18 years or older admitted for more than 24 hours to a 6-bed university-based medical ICU from December 1997 to November 1998 were included. A first survey assessing opinions about shortterm QOL and survival was published previously [9,10]: patients judged their own OQOL and HRQOL using the Sickness Impact Profile (SIP) 6 months after ICU discharge. In addition, mortality was assessed. Furthermore, nurses' and physicians' judgments concerning outcome and possibly futile care were assessed daily during the ICU stay, and these predictions were compared with the observed outcome. For the present study, the patients were reevaluated 9 years following ICU discharge using the same outcome parameters. Survival was assessed based on

K.H. Stricker et al. information received from the patients' relatives or primary care physicians. The study was approved by the Institutional Review Board and the local ethics committee.

2.1. Quality of life A validated German version of the SIP questionnaire [18] was mailed together with a consent form to all surviving patients. Upon consent, all patients were additionally contacted for the structured telephone interview to assess OQOL and HRQOL. Patients declining to fill out the SIP questionnaire were still invited to take part in the interview. This telephone interview was validated previously [9] and was performed by one of the authors using a procedure identical to the one used 9 years ago. For the assessment of HRQOL, patients were asked about their independence with regard to bodily care, household management, and outdoor mobility (defined as going for walks independently outside the home). Patients who were independent in all 3 activities were considered “completely independent,” those who were limited in 1 or 2 of the 3 activities were considered “partially dependent,” and those who needed help with all 3 activities were considered “fully dependent.” At the end of the interview, patients chose one of the descriptive words good, fair, or poor to judge their OQOL. If a patient was not able to answer the phone because of barriers in language or understanding, the interview was conducted with the help of a proxy living in the same household, who functioned as a translator but not as a substitute. The SIP [19,20] has been validated in various patient populations, including ICU patients [12,21,22]. It is a multidimensional, cumulative health index and consists of 136 questions divided into 12 categories of daily living. Three of these (“ambulation,” “mobility,” and “body care”) form the score “physical dimension.” The categories “social interactions,” “alertness behavior,” “emotional behavior,” and “communication” form the score “psychosocial dimension,” whereas the other 5—“sleep and rest,” “eating,” “work,” “home management,” and “recreation and pastimes”—are taken as “independent categories.” A dysfunction score is attributed to each question. The score for each category is calculated by dividing the sum of the values of that category by the maximum possible dysfunction scores for that category and multiplying by 100. The overall score is calculated by dividing the sum of all values by the maximum dysfunction score for the SIP and multiplying by 100. The higher the score, the more severe the dysfunction: a score of 0 to 5 is found in a healthy population [23], a score of 5 to 15 corresponds to moderate disability, and a score greater than 15 shows significant impairment of QOL.

2.2. ICU caregivers' predictions of patient outcome To identify possibly futile care, ICU nurses and 1 of 10 staff physicians in charge were asked to estimate the

QOL 9 years after ICU stay

381

patient's expected outcome daily during morning rounds throughout the patient's stay in the ICU. For this purpose, the following 2 questions had to be answered: “Do you consider treatment futile with respect to survival?” (ie, survival to hospital discharge) and “Do you consider treatment futile with respect to QOL?” Answer options were “yes,” “no,” or “questionable.” Outcome prediction was assessed in a “pessimistic mode”; that is, if on any single ICU day any one of the nurses or physicians answered either of the 2 questions with “yes” (ie, futile care) or “questionable” (ie, possibly futile care), the expected outcome was assessed as negative for the respective group of caregivers. Of note, no specific time point was designated when asking the questions. Rehabilitation may vary between individuals and depend on admission diagnosis, health care infrastructure, and other parameters. Therefore, a specific chronological time point did not seem appropriate when asking this question.

2.3. Statistical analysis Statistical analysis was performed with the software package SAS 9.2 for Windows (SAS Institute Inc, Cary, NC). The SAS PHREG procedure (proportional hazard regression) was used to investigate possible predictors of mortality and/or QOL. The influence of the following covariates on the assessment of survival and QOL was tested: age, Simplified Acute Physiology Score (SAPS) II, length of stay (LOS) in ICU, LOS-hospital, and diagnostic group. Only those parameters with a significant influence on survival in the univariate model were used for multivariate analysis with stepwise parameter selection. Results of hazard ratios (HRs) are given with 95% confidence intervals. In the search for covariates with a significant influence on OQOL and HRQOL, death was considered poor QOL for the proportional hazard model. Failure to do so would have resulted in the censoring of all the worst cases and therefore would have introduced a significant bias with respect to the original judgments of future QOL.

Fig. 1 Flow diagram displaying survival and cooperation of patients in assessment of OQOL, HRQOL, and the SIP.

tality was significantly higher for patients with infectious (P b .0001) or gastrointestinal/liver (P b .0001) diseases as compared with patients with a cardiovascular, pulmonary, or neurologic diagnosis. The 75 patients lost to follow-up after 9 years were significantly younger than the 334 follow-up patients (median, 58 vs 61 years; P b .01). However, no differences were observed with respect to admission diagnosis, SAPS II, LOS-ICU, or LOS-hospital.

3. Results

3.2. Quality of life

3.1. Survival

Of the 188 surviving patients, 186 were willing to answer the questions concerning OQOL and HRQOL (Fig. 1, Tables 2 and 3); and 90 participated in the SIP assessment (Table 4, electronic supplemental material). The telephone interview was performed with proxies for 4 patients.

Seventy-five of the 409 6-month survivors from the original study [10] were lost to follow-up. Of the remaining 334 patients, 146 had died, resulting in a 6month to 9-year mortality rate of 44% (Fig. 1). Age and SAPS II were significant predictive factors for mortality after 9 years (both Ps b .0001) (Table 1). In addition, patients were assigned to a predefined diagnostic group upon admission to the ICU, that is, cardiovascular, pulmonary, neurologic, gastrointestinal, infectious, and others (eg, intoxications). Comparing these groups, mor-

3.2.1. OQOL and HRQOL Sixty percent of the surviving patients considered their OQOL as good and 34% as fair (Table 2). When assessing the change in OQOL between 6 months and 9 years, 40 patients experienced a decline, whereas 18 individuals mentioned an improvement. One hundred twenty-eight

382 Table 1

K.H. Stricker et al. Demographic characteristics of study patients

n Age, y (median) Interquartile range Min-max SAPS II (median) Interquartile range Min-max Admission diagnosis Cardiovascular Pulmonary GI/liver Infectious Neurologic Various LOS-ICU (median) Interquartile range Min-max LOS-hospital (median) Interquartile range Min-max

All patients 1997/1998

Survivors after 6 mo

Died between 6 mo and 9 y

Lost to follow-up between 6 mo and 9 y

Survivors after 9y

521 64 51-72 18-95 27 20-38 6-127

409 62 50-71 18-91 24 18-33 6-89

146 68 † 58-78 34-95 29 † 24-38 10-80

75 58 ⁎ 40-67 18-84 22 18-32 6-89

188 57 49-68 18-83 21 15-27 6-79

285 27 30 53 60 66

237 21 19 32 40 60 2 1-3 1-42 15 9-26 1-215

33 7 4 4 7 20 2 1-3 1-15 14 5-29 1-172

119 2 4 16 22 25 2 1-3 1-42 13 7-22 1-215

1-4 1-42 14 7-27 1-215

85 12 11 12 11 15 2 1-4 1-19 17 10-30 3-182

GI indicates gastrointestinal. ⁎ P b .01 between age of the 334 examined patients and the 75 patients lost to follow-up after 9 years. † P b .0001 for the significant predictive factors for mortality after 9 years.

patients reported no change. Full physical dependence was found in 11 and partial physical dependence in 42 patients (Table 3). Compared with patients' assessment after 6 months, both OQOL (P = .07) and HRQOL (P = .36) were not significantly different after 9 years. Poor HRQOL, that is, full dependency after 9 years, was associated with a higher SAPS II score, LOS-ICU, and LOS-hospital (P b .01 for SAPS and P b .001 for

Table 2

the other 2 in the univariate analysis). Combining these covariates into a multivariate model with stepwise selection, only LOS-hospital was independently related to poor HRQOL (P b .001). Using the same proportional hazard model to investigate possible covariates associated with poor OQOL after 9 years, only age at admission was significantly and independently related to poor OQOL (P b .05).

Change in OQOL during the study period (n = 186)

Cells colored green: improvement in OQOL over time. Cells colored orange: worsening in OQOL over time. Overall, OQOL improved in 18 (9.7%) and worsened in 40 (21.5%) patients.

QOL 9 years after ICU stay Table 3

383

Change in HRQOL during the study period (n = 186)

Cells colored green: improvement in HRQOL over time. Cells colored orange: reduction in HRQOL over time. Overall, HRQOL improved in 5 (2.7%) and worsened in 38 (20.4%) patients.

3.2.2. Sickness Impact Profile The overall SIP score of the 90 respondents was 5.0 ± 8.2 (electronic supplemental material). Participating patients had significantly lower SAPS II scores than patients who did not respond (21 ± 10 vs 25 ± 13, P b .05). However, they were comparable with respect to age, LOS-ICU, and LOS-hospital. Only 31 patients completed the SIP twice, in 1998 and in 2007. Based on this sample of individuals, a significant improvement was noted in the category “eating,” whereas significant deterioration was observed for “household work.” No significant change was noted in the other categories of the SIP (electronic supplemental material). Of the 54 patients who filled in a SIP after 6 months but not

Table 4

a

after 9 years, 10 were lost to follow-up and 23 had died. Of the 21 remaining survivors, not a single patient reported poor OQOL.

3.3. Comparison of the caregivers' outcome prediction of survival and QOL during ICU stay and patients' own assessments after 9 years For the large majority of the survivors, physicians and nurses had correctly predicted long-term survival on each day of their ICU stay (182 of 186 patients). For 171 of the 186 patients, none of the caregivers expressed doubts about good future QOL at the time of ICU stay. Of these

Comparison of caregivers' predictions during the ICU stay with regard to QOL, and patients' own assessments 9 years later

Questionable or yes for futility of treatment on at least 1 hospital day.

384 171 patients, 105 (61%) ultimately enjoyed good QOL, whereas 57 (33%) considered their QOL as fair. Comparing the 2 caregiver groups, physicians were more optimistic than nurses with respect to a positive outcome; that is, they expressed fewer doubts than nurses about survival (in 71 vs 96 of all 334 patients on all days; P b .04) or future QOL (88 vs 120 of all 334 patients on all days; P b .02). Table 4 shows the comparison of expected to reported outcome with respect to QOL for all 186 surviving patients. Doubt about long-term survival at the time of ICU stay was significantly associated with death (HR, 7.19 [5.558.79]; P b .001). Physicians' doubts predicted death better than nurses' (HR nurses compared with HR physicians, 0.83 [0.70-0.99]; P b .04). The predictive power of initial doubts by the caregivers remained unchanged when corrected for age, SAPS II, and length of hospital stay (HR, 4.41 [3.45-5.64]). Doubts about future QOL expressed by any one of the caregivers significantly predicted poor QOL or death with an HR of 4.45 (3.75-5.29; P b .001). The predictive power of initial doubts by any of the caregivers remained when corrected for the additional parameters associated with a poor future QOL, that is, age and SAPS II (HR, 2.97 [2.40-3.67]).

4. Discussion This study extends the follow-up period of a previous 6-month outcome analysis of ICU patients [10] to 9 years after ICU discharge. It is one of few investigations that assess outcome at different time points over a very prolonged period. The survival rate was 56% 9 years after ICU stay, and OQOL was good.

4.1. Survival Following ICU mortality of 8.6% and overall hospital mortality of 9.4%, the postdischarge mortality within the first 6 months was low at 3.5% [10]. However, overall mortality increased to 47.5% over the entire 9-year period. The longterm survival rate of our cohort is comparable to that of a study by Flaatten and Kvale [17], who observed a 12-year survival rate of 48.4% in a general ICU population, and to that of Kaarlola et al [24], who found a 6-year survival rate of 49.2%. Contrary to our population of medical ICU patients with low early postdischarge mortality, the studies by Flaatten and Kvale [17] and by Kaarlola et al [24], as well as that by Niskanen et al [25], showed a high mortality rate early after ICU discharge, which approached that of the general population after 2 years [17,25]. An Australian study [15] found a 68.7% survival rate 10 years after and a 54.7% rate 15 years after admission to a mixed ICU. These differences in survival curves may be due to differences in case mix; that is, patients' severity of illness in our cohort was not very high for the majority of patients, and LOS was

K.H. Stricker et al. shorter than in the other groups. Although long-term survival is reasonable in these investigations, mortality rate may exceed 70% after less than 2 years in subgroups of ICU patients [26,27].

4.2. Quality of life To assess QOL, we used 3 different measures, as described previously [9], that is, OQOL, HRQOL, and the SIP. Although 2 of them are highly subjective, they have been validated against the third instrument, the SIP [9]. The surviving patients subjectively assessed their QOL as good, with 59% of individuals reporting good OQOL and 73% being completely independent in daily activities. These ratings are similar to those of the general Swiss population, with 75% of individuals of comparable age rating their OQOL as good or very good [28]. Nine years after ICU discharge, the SIP scores of our patient population are relatively low, indicating good health: in 10 of 15 categories, the scores are comparable to a healthy population, with the total SIP indicating a slightly decreased QOL. The good long-term outcome may be explained by a relatively low severity of illness at ICU admission, with a low median SAPS II score in this cohort of 21 points. In our analysis, age was the only parameter to be independently associated with poorer OQOL. A recent study of Swiss surgical ICU patients used the Short Form–36 [14] to assess HRQOL 12 months after ICU discharge: the mental component summary was comparable to that of an age- and sex-matched population, whereas the physical component summary was significantly lower. Furthermore, older patients and those with a longer LOS-ICU or greater severity of illness reported poorer HRQOL. Thus, the same risk factors were confirmed using 2 different instruments: the SIP and the Short Form–36. Although almost half of all 9-year survivors completed the SIP, 2 consecutive data sets, ie, at 6 months and 9 years, were available for only 31 of them. We suspect that the lack of returned SIP questionnaires was due to the large number of questions in the SIP, as most survivors were willing to participate in the telephone interview. In the case of posttraumatic stress disorder, fewer telephone interviews might be expected because of an avoidance reaction [29]. Statistically, inclusion of age in addition to SAPS in the same analysis could pose a problem of colinearity. However, recent publications reporting outcome for nonsurgical ICU patients do not support such a colinearity: Severity of illness but not age was found to be a major risk factor for poor outcome [30-32]. Furthermore, both age and severity of illness have been found to be independent risk factors for mortality [33]. A recent Finnish study observed good HRQOL in a mixed ICU population: 97% were living at home, and half of them were completely independent. Furthermore, two thirds of respondents reported similar or improved HRQOL over a 12-

QOL 9 years after ICU stay month period; and half of the patients enjoyed equal or better HRQOL as compared with the pre-ICU assessment [34]. Health-related QOL is a very subjective parameter, influenced by other parameters such as age and sex [35], preadmission QOL [36], postdischarge rehabilitation, and others [37]. Elderly people, although fragile, are very adaptive in their way of living and often consider their overall or subjective QOL to be good even though their life is not free from physical handicap. Looking at the HRQOL of the survivors, none of them was totally dependent 6 months after ICU stay, which indicates that the most dependent patients had died. Despite a large body of literature on outcome, the search for effective interventions, including multidisciplinary postdischarge rehabilitation, to improve long-term outcomes in survivors of critical illness is only in its infancy [8,37]. Considering that HRQOL may be lower in ICU patients than in an agematched general population [13,34], the present outcome results are quite pleasing.

4.3. Change in QOL A recent study found that recovery of HRQOL starts early after ICU discharge and may reach preadmission levels by 6 months [13], whereas a review of 26 publications found that post-ICU QOL may remain lower than preadmission levels [12]. However, median follow-up time in the latter study was only 7 months. With repeated assessments over a very prolonged period, we noted a tendency toward lower OQOL and higher physical dependence over a period of 8.5 years. However, no significant decline was observed in 11 of 12 SIP categories and in all 3 summary dimensions. Furthermore, although HRQOL may decrease with greater age [35], OQOL even improved in 10% of our patients. Thus, our data indicate that recovery may continue after 6 months and that longerterm outcome assessment is mandatory.

4.4. Caregivers' outcome prognoses Mortality can be predicted using a number of scoring systems [38]. However, risk scoring systems are of limited value for an individual patient [39]. In fact, caregivers' subjective predictions of survival were shown to be superior to and should supplement scoring systems [11,40]. Because a large number of parameters may influence outcome [8,37] and individuals' experience of QOL is very broad, estimating survival or future QOL must be done with the utmost caution. Almost half of ICU caregivers were reported to have felt uncomfortable with life support plans, mostly because these were considered technologically too intense [41]. It may therefore be reassuring that caregivers in our study correctly predicted long-term survival in over 90% of patients and good QOL in two thirds of patients. Our study adds additional

385 information about caregivers' ability to anticipate QOL and confirms prior observations that caregivers' repeated assessments of prognosis may be superior to a single prognosis on ICU admission [40]. Our caregivers' prognosis was superior over prolonged periods of time, that is, 9 years, as compared with a 6-month observational period. Thus, a short follow-up period may be inadequate for outcome assessment. It goes without saying that professionals' apprehension about futility of care should not influence their daily work, and more objective predictors are needed to help us with difficult decisions in the ICU. Finally, although some professional critical care societies recommend that physicians provide guidance with regard to limitation of life support, many surrogates do not wish to receive such information [42]; and outcome prediction should not be forced on them.

4.5. Limitations of the study Seventy-five of the initial 521 patients—that is, 14%— were lost to follow-up between 6 months and 9 years and were therefore not included in our analysis. Some of the study patients were tourists or foreign workers who returned to their countries of origin. Others, such as drug addicts, did not have a permanent address. These patients were younger at ICU admission. Mortality was significantly associated with age and severity of illness, so the overall survival of the entire group may even be underestimated. The same is true for QOL, which was also associated with age. Mortality and QOL were assessed at given time points, that is, 6 months and 9 years. Therefore, a Kaplan-Meier survival curve cannot be calculated. In addition, unlike Norway and Finland, Switzerland does not maintain a central death registry. Therefore, comparison with survival in the general population was not possible. Moreover, we do not know when exactly the changes in QOL occurred. In fact, additional illnesses, independent of the admission diagnosis to the ICU, may have influenced survival and QOL and therefore long-term outcome. However, access to complete medical charts of all patients for the 9 previous years was not possible for logistic reasons. One of the strengths of the study is the use of a validated questionnaire, that is, the SIP. Although half of all 9-year survivors were willing to fill out the SIP, only one third of them had completed the assessment at the first time point. Therefore, the data must be interpreted in this context. In addition, QOL may be influenced by sex [35]. Unfortunately, sex was not registered at the beginning of the study and therefore could not be included in the analysis. Finally, as this is a single-unit study from a small ICU, the findings should be extrapolated only cautiously to other patient populations. The data should be interpreted in the context of other studies cited in the “Discussion.”

386

5. Conclusion In conclusion, approximately 50% of the patients admitted to a university-based medical ICU survived more than 9 years after discharge; and survivors' QOL was generally good. Although outcome prediction remains difficult, professional estimation of patients' future QOL should not be ignored. Finally, rehabilitation after ICU care does not stop after a few months; and changes may occur after prolonged periods of time. Supplementary materials related to this article can be foundonline at doi:10.1016/j.jcrc.2010.11.004.

Acknowledgment The work was supported by the institutions of the respective authors.

References [1] Wunsch H, Angus D, Harrison D, et al. Variation in critical care services across North America and Western Europe. Crit Care Med 2008;36:2787-93. [2] Rothen HU, Stricker KH, Einfalt J, et al. Variability in outcome and resource use in intensive care units. Intensive Care Med 2007;33: 1329-36. [3] Karlsson S, Ruokonen E, Varpula T, et al. Long-term outcome and quality adjusted life years after severe sepsis. Crit Care Med 2009;37: 1268-74. [4] Graf J, Mühlhoff C, Doig GS, et al. Health care costs, long-term survival, and quality of life following intensive care unit admission after cardiac arrest. Crit Care 2008;12:R92. [5] Angus DC, Carlet J. Surviving intensive care: a report from the 2002 Brussels roundtable. Intensive Care Med 2003;29:368-77. [6] Azoulay E, Kenstich-Barnes N, Pochard F. Health related quality of life. Chest 2008;133:339-41. [7] Hofhuis JG, van Stel HF, Schrijvers AJ, et al. Health-related quality of life in critically ill patients: how to score and what is the clinical impact? Curr Opin Crit Care 2009;15:425-30. [8] Herridge MS. Long-term outcomes after critical illness: past, present, future. Curr Opin Crit Care 2007;13:473-5. [9] Frick S, Uehlinger D, Zürcher Zenklusen R. Assessment of former ICU patients' quality of life: comparison of different quality-of-life measures. Intensive Care Med 2002;28:1405-10. [10] Frick S, Uehlinger D, Zuercher Zenklusen R. Medical futility: predicting outcome of intensive care unit patients by nurses and doctors—a prospective comparative study. Crit Care Med 2003;31:456-61. [11] Conti M, Eckert P. Patient âgé aus soins intensifs. Rev Med Suisse 2009;5:2494-8. [12] Dowdy D, Eid M, Sedrakyan A, et al. Quality of life in adult survivors of critical illness: a systematic review of the literature. Intensive Care Med 2005;31:611-20. [13] Hofhuis J, Spronk P, van Stel H, et al. The impact of critical illness on perceived health-related quality of life during ICU treatment, hospital stay, and after hospital discharge: a long-term follow-up study. Chest 2008;133:377-85. [14] Stricker K, Cavegn R, Takala J, et al. Does ICU length of stay influence quality of life? Acta Anaesthesiol Scand 2005;49:975-83. [15] Williams T, Dobb G, Finn J, et al. Determinants of long-term survival after intensive care. Crit Care Med 2008;36:1523-30.

K.H. Stricker et al. [16] Cuthbertson BH. Life after intensive care—it's life … but not as we know it! Crit Care Med 2008;36:1668-9. [17] Flaatten H, Kvale R. Survival and quality of life 12 years after ICU. A comparison with the general Norwegian population. Intensive Care Med 2001;27:1005-11. [18] Slama H, Fisch H, Frey F. Quantitative assessment of the quality of life of patients with a kidney transplant. Schweiz Rundsch Med Prax 1993;82:1253-62. [19] Bergner B, Bobbit R, Pollard W, et al. The sickness impact profile: validation of a health status measure. Med Care 1976;14: 57-67. [20] Bergner M, Bobbitt RA, Carter WB, et al. The sickness impact profile: development and final revision of a health status measure. Med Care 1981;19:787-805. [21] Tian Z, Miranda D. Quality of life after intensive care with the sickness impact profile. Intensive Care Med 1995;21:422-8. [22] Short TG, Buckley TA, Rowbottom MY, et al. Long-term outcome and functional health status following intensive care in Hong Kong. Crit Care Med 1999;27:51-7. [23] Lipsett P, Swoboda S, Campbell K, et al. Sickness Impact Profile Score versus a Modified Short-Form survey for functional outcome assessment: acceptability, reliability, and validity in critically ill patients with prolonged intensive care unit stays. J Trauma 2000;49: 737-43. [24] Kaarlola A, Pettilä V, Kekki P. Quality of life six years after intensive care. Intensive Care Med 2003;29:1294-9. [25] Niskanen M, Kari A, Halonen P. Five-year survival after intensive care–comparison of 12,180 patients with the general population. Finnish ICU Study Group. Crit Care Med 1996;24:1962-7. [26] Soares M, Salluh JI, Torres VB, et al. Critically ill patients with cancer and prolonged ICU length of stay. Chest 2008;134:520-6. [27] Gaudino M, Girola F, Piscitelli M, et al. Long-term survival and quality of life of patients with prolonged postoperative intensive care unit stay: unmasking an apparent success. J Thorac Cardiovasc Surg 2007;134:465-9. [28] Bundesamt für Statistik. Schweizerische Gesundheitsbefragung 2007. http://www.bfs.admin.ch/bfs/portal/de/index/news/publikationen. Document.112182.pdf [last accessed September 14, 2010]. [29] Capuzzo M, Bertacchini S, Jones C, et al. Patients with PTSD after intensive care avoid hospital contact at 2-year follow-up. Acta Anaesthesiol Scand 2008;52:113-4. [30] Kaufmann P, Smolle K, Krejs G. Short- and long-term survival of nonsurgical intensive care patients and its relation to diagnosis, severity of disease, age and comorbidities. Curr Aging Sci 2009;2: 240-8. [31] Lesage A, Ramakers M, Daubin C, et al. Complicated acute myocardial infarction requiring mechanical ventilation in the intensive care unit: prognostic factors of clinical outcome in a series of 157 patients. Crit Care Med 2004;32:100-5. [32] Karvellas C, Pink F, McPhail M, et al. Bacteremia, acute physiology and chronic health evaluation II and modified end stage liver disease are independent predictors of mortality in critically ill nontransplanted patients with acute on chronic liver failure. Crit Care Med 2010;38: 121-6. [33] Sakr Y, Vincent J, Ruokonen E, et al. Sepsis and organ system failure are major determinants of post-intensive care unit mortality. J Crit Care 2008;23:475-83. [34] Kaarlola A, Tallgren M, Pettilä V. Long-term survival, quality of life, and quality-adjusted life-years among critically ill elderly patients. Crit Care Med 2006;34:2120-6. [35] Ellert U, Bellach BM. Der SF-36 im Bundes-GesundheitssurveyBeschreibung einer aktuellen Normstichprobe. Das Gesundheitswesen 1999;61:S184-90. [36] Rivera-Fernández R, Navarrete-Navarro P, Fernández-Mondejar E, et al. Six-year mortality and quality of life in critically ill patients with chronic obstructive pulmonary disease. Crit Care Med 2007;34: 2317-24.

QOL 9 years after ICU stay [37] Rubenfeld GD. Improving clinical trials of long-term outcomes. Crit Care Med 2009;37:S112-6. [38] Pettilä V, Pettilä M, Sarna S, et al. Comparison of multiple organ dysfunction scores in the prediction of hospital mortality in the critically ill. Crit Care Med 2002;30:1705-11. [39] Gersbach P, Tevaearai H, Revelly JP, et al. Are there accurate predictors of long-term vital and functional outcomes in cardiac surgical patients requiring prolonged intensive care? Eur J Cardiothorac Surg 2006;29: 466-72.

387 [40] Sinuff T, Adhikari NK, Cook DJ, et al. Mortality predictions in the intensive care unit: comparing physicians with scoring systems. Crit Care Med 2006;34:878-85. [41] Griffith L, Cook D, Hanna S, et al. Clinician discomfort with life support plans for mechanically ventilated patients. Intensive Care Med 2004:30. [42] White DB, Evans LR, Bautista CA, et al. Are physicians' recommendations to limit life support beneficial or burdensome? Am J Respir Crit Care Med 2009;180:320-5.