METHODS. Keywords: family needs; satisfaction; comprehension; leaflet; information; intensive care

Transcription

1 Impact of a Family Information on Effectiveness of Information Provided to Family Members of Intensive Care Unit Patients A Multicenter, Prospective, Randomized, Controlled Trial ELIE AZOULAY, FRÉDÉRIC POCHARD, SYLVIE CHEVRET, MERCÉ JOURDAIN, CAROLINE BORNSTAIN, ANNE WERNET, ISABELLE CATTANEO, DJILALI ANNANE, FRÉDÉRIC BRUN, PIERRE-EDOUARD BOLLAERT, JEAN-RALPH ZAHAR, DANY GOLDGRAN-TOLEDANO, CHRISTOPHE ADRIE, LUC-MARIE JOLY, JEROME TAYORO, THIBAUT DESMETTRE, ETIENNE PIGNE, ANTOINE PARROT, OLIVIER SANCHEZ, CATHERINE POISSON, JEAN-ROGER LE GALL, BENOÎT SCHLEMMER, and FRANÇOIS LEMAIRE for the French FAMIREA Group Intensive Care and Biostatistics Departments of the Saint-Louis Teaching Hospital and Paris 7 University; Intensive Care Unit of the Cochin Teaching Hospital, Paris; Intensive Care Unit of the Henri Mondor Teaching Hospital, Créteil, France; and the Ethics Committee of the French Society for Critical Care Comprehension and satisfaction are relevant criteria for evaluating the effectiveness of information provided to family members of intensive care unit (ICU) patients. We performed a prospective randomized trial in 34 French ICUs to compare comprehension of diagnosis, prognosis, treatment, and satisfaction with information provided by ICU caregivers, in ICU patient family representatives who did (n 87) or did not receive a family information leaflet (FIL) in addition to standard information. An FIL designed specifically for this study was delivered at the first visit of the family representative: it provided general information on the ICU and hospital, the name of the ICU physician caring for the patient, a diagram of a typical ICU room with the names of all the devices, and a glossary of 12 terms commonly used in ICUs. Characteristics of the ICUs, patients, and family representatives were similar in the two groups. The FIL reduced the proportion of family members with poor comprehension from 40.9% to 11.5% (p ). In the representatives with good comprehension, the FIL was associated with significantly better satisfaction (21 [18 to 24, quartiles] versus 27 [24 to 29, quartiles], p 0.01). These results indicate that ICU caregivers should consider using an FIL to improve the effectiveness of the information they impart to families. Keywords: family needs; satisfaction; comprehension; leaflet; information; intensive care An important task for intensive care unit (ICU) physicians is to provide family members with the appropriate, clear, and compassionate information they need to cope with their distress and to participate in making decisions about patients who are unable to speak for themselves. Previous studies, from both the United States and Europe, have shown that the quality of information provided to family members of ICU patients can be assessed based on two criteria, namely, comprehension (1) and satisfaction with the information provided by ICU caregivers to family members (2 5). Family members must understand the diagnosis, prognosis, and treatment in (Received in original form August 29, 2001; accepted in final form October 12, 2001) Supported by a grant from the French Society for Critical Care. Correspondence and requests for reprints should be addressed to Élie Azoulay, M.D., Service de réanimation médicale, Hôpital Saint-Louis, 1 avenue Claude Vellefaux Paris, France. elie.azoulay@sls.ap-hop-paris.fr This article has an online data supplement, which is accessible from this issue s table of contents online at Am J Respir Crit Care Med Vol 165. pp , 2002 DOI: /rccm OC Internet address: the patient if they are to participate in management decisions (6 8) and to speak for the patient. Moreover, good comprehension helps the family cope with the psychological stress associated with the ICU admission (1). Satisfaction of family members promotes favorable interactions with caregivers seeking to meet the needs of the family (2 5, 9). Previous studies of the quality of information provided to family members of ICU patients were mainly descriptive, although they also identified the determinants of comprehension and satisfaction in the family. Comprehension was associated with patient s age, delivery of a family information leaflet (FIL), and caregivers perception of poor comprehension (1). Satisfaction was greater when the patient/nurse ratio was less than 3; when the family felt they received no contradictory information, knew the specific role of each caregiver, or were helped by their usual physician; and when the interview time allowed to the family increased. There are no published studies of the effects on comprehension and satisfaction of interventions designed specifically to improve the effectiveness of information provided to families of ICU patients. In previous studies, which were not randomized, handing an information leaflet to the family members during the first interview improved comprehension (1), although the effect on satisfaction was inconsistent (2). These studies were conducted each in a single ICU, in small numbers of families (1). Some used but did not evaluate the benefits of an information leaflet (10, 11). Moreover, delivery of the information leaflet to the family members was not strictly controlled in any of these studies (2, 12). We sought to determine whether a standardized FIL improved satisfaction and comprehension of the information provided to family members of ICU patients. METHODS Study Design ICU selection. From a total of 114 French ICUs, members of the French Society for Critical Care, we randomly selected 51 ICUs, stratifying on university versus community hospital and on the city (Paris area versus others). An invitation to participate in the study was sent to the heads of these 51 ICUs. A physician in each center (the investigator) collected ICU and patient characteristics on standardized forms. Inclusion criterion. The only eligibility criterion for each patient was the expected length of ICU stay, which was required to be at least 48 h. Each participating ICU included six consecutive patients meeting this criterion, starting on July 1, The number of patients per center was limited to six (including three with no FIL) because a

2 Azoulay, Pochard, Chevret, et al.: Family Information 439 previous study showed that receiving an FIL had an independent favorable effect on comprehension by family members of ICU patients (1), and to avoid any center effect on study results. Exclusion criteria. Patients were excluded from the study if (1) they died within 48 h after ICU admission, (2) their family representatives refused to participate in the study, or (3) no family members visited them within 5 d after their admission to the ICU. Randomization procedure. Computer-generated random-number tables were used to assign patients to the FIL or control group in blocks of six stratified on ICU. Each participating ICU received six sealed envelopes, of which three contained an FIL and three a blank sheet. To ensure full concealment of allocation, a biostatistics department independent from the study ICUs carried out the randomization procedure, prepared the envelopes, and sent them to the study ICUs. Consecutive patients whose length of stay in the ICU was expected to be longer than 48 h were randomized at admission. Family representatives of control patients were informed according to routine practice in the ICU before the study ( standard information, including at least one meeting with a physician each day during the first week of ICU admission, with information on the diagnosis, prognosis, and treatment). Family representatives in the FIL group received standard information and an FIL. At the first visit of the family representative to the ICU, an ICU caregiver other than the investigator opened a sealed envelope and, if it contained an FIL, handed this document to the family representative. Then, assessment of comprehension (1) and satisfaction (see Appendix E1 in the online data supplement) (2, 3), and of anxiety and depression, were performed between the third and the fifth day. Further details on data collection, description of the standardized family information leaflet, and instruments used to measure the effectiveness of information provided to family representatives can be found in the online data supplement. Statistical Analysis Results are expressed as quartiles. Comprehension in the two groups was compared using the chi-square test, or the Fisher exact test when appropriate. Comparison of satisfaction in the two randomized groups was performed using the nonparametric Mann-Whitney test. All p values were two-sided, with values of 0.05 or less indicating statistical significance. All statistical tests were done using the SAS 6.12 (SAS Inc., Cary, NC) package. Sample Size Sample size was computed on the basis of previous results, suggesting that the FIL would improve comprehension by approximately 20% but would not influence satisfaction (1). To detect a 20% difference from a 50% noncomprehension rate without the FIL (i.e., 50% versus 30%), with a type I error of 0.05 and a power of 0.80, 204 patients had to be recruited, 102 in each group. RESULTS A total of 34 ICUs (442 beds) participated in the study and included a total of 204 patients (Figure 1). Eight patients died before their family members could be invited to complete the questionnaire, and 15 additional patients received no visits within the first 5 d after ICU admission. This left 181 family representatives, of whom 175 (97%) accepted and 6 (3%) declined to be interviewed about their comprehension and to complete the Critical Care Family Needs Inventory (CCFNI) questionnaire. Thus, the study data were obtained from 175 family representatives. A single ICU had two excluded patients (one in each randomized group); each of the other ICUs had one excluded patient at the most. Randomization of the patients at admission assigned 88 family representatives to the FIL group and 87 to the control group. The ICU physicians and nurses who collected the study data constitute the French FAMIREA group. Characteristics of the 34 ICUs As shown in Table 1, 17 (50%) ICUs were both medical and surgical, 15 (44%) were medical, and two (6%) were surgical. In 16 (47%) ICUs, it was standard practice for the physicians and nurses to meet regularly to discuss family needs. Only in 6 (18%) ICUs was information about a given patient provided by the same physician throughout the patient s ICU stay. None of the collected parameters were significantly different between the two randomized groups. Characteristics of the 175 Patients Median age was 60 yr (range, 48 to 71) (Table 2). Thirty-nine (22%) patients were not of French descent. Median Simplified Acute Physiology Score (SAPS II) at admission was 41 (29 to 55), length of ICU stay was 9 (6 to 17) d, and in-icu mortality was 25% (44 deaths). None of the collected parameters were significantly different in the two randomized groups. Characteristics of the 175 Family Members Who Completed the Questionnaires Although 40 (23%) family members were not of French descent (Table 3), all spoke French. Of the 175 family members, 82 (47%) were spouses. The ratio of the time actually allowed to family members over the time they would have liked to spend receiving information was 0.66 (0.5 to 1), and the difference between the desired and actual times was 5 minutes (0 to 7.75). Twenty-nine (17%) family members reported receiving contradictory information, 89 (51%) did not know the specific role of each caregiver, 90 (51%) were not receiving help from their usual doctor, and 68 (39%) expressed a desire to receive help from a psychologist. One hundred sixty-five family members (94%) were satisfied with the information provided by TABLE 1. CHARACTERISTICS OF THE ICUs (n 34) Parameters Number (%) or Median (quartiles) Figure 1. Study profile. University hospital 15 (44.1) Number of beds per unit 13 ( ) Waiting room available 27 (79.4) Room for family information available 20 (58.8) Number of attending physicians 3 (2 4) Number of junior physicians 2 (1 3.75) Nurse-to-patient ratio 3 (3 3.37) Regular nurse physician meetings about family information 16 (47) Total daily visiting hours, min 120 (97 180) Written protocol for interacting with families 4 (11.8) FIL available (before the study) 26 (76.5) Information to each family always given by same physician 6 (17.6)

3 440 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL TABLE 2. CHARACTERISTICS OF PATIENTS * Variable All Patients No (n 87) p Value Age 60 (48 71) 61 (51 70) 57 (47 73) NS Sex ratio, % women 62 (35.4) 30 (34) 32 (36.8) NS Patients of French descent 136 (77.7) 68 (77.3) 68 (78.1) NS Unmarried 32 (18.3) 12 (13.6) 20 (22.9) NS Unemployed 31 (17.7) 16 (18.2) 15 (17.2) NS Direct admission 43 (24.6) 21 (23.8) 22 (25.3) NS ICU admission for Acute respiratory failure 99 (56.6) 48 (54.5) 51 (58.6) NS Shock 57 (32.6) 29 (32.9) 28 (32.2) NS Acute renal failure 40 (22.8) 20 (22.7) 20 (23) NS Coma 66 (37.7) 31 (35.2) 35 (40.2) NS SAPS II 41 (29 55) 42 (30 55) 40 (27 54) NS Length of ICU stay 9 (6 17) 10 (6 16) 9 (6 18) NS ICU mortality rate 44 (25.1) 21 (23.9) 23 (26.4) NS Definition of abbreviation: NS not significant. junior physicians. According to the Hospital Anxiety and Depression Scale (HADS), 105 (60%) family members suffered anxiety and 68 (39%) depression. None of these parameters was significantly different between the two randomized groups. Comprehension and Satisfaction Failure to comprehend the diagnosis, prognosis, or treatment was noted in 46 (26%) family members; 16 (9%) did not understand the diagnosis, 18 (10%) the prognosis, and 37 (21%) the treatment. The median satisfaction score was 22 (19 to 26.5). Patients who understood the diagnosis, prognosis, and treatment had a significantly lower CCFNI score, indicating better satisfaction (21 [19 to 25] versus 25 [19 to 28], p 0.04). Symptoms of anxiety and depression were associated neither with comprehension nor with satisfaction. Impact of the FIL on Comprehension and Satisfaction Tables 1 3 show that characteristics of the ICUs, the patients, and the family members were not significantly different between the two groups. As shown in Table 4, comprehension was significantly better in the FIL group (11.5 versus 41% of family members with poor satisfaction, p ). Satisfaction did not differ significantly between the FIL and the control groups (median score, 21 [18 to 26] versus 23 [19 to 27], p 0.08). However, among family members with good comprehension, those who received the FIL had a significantly better median satisfaction score than those who did not (21 [18 to 24] versus 27 [24 to 29], p 0.01) (Figure 2). Anxiety and depression were not significantly less prevalent in the FIL group. Moreover, the number of physician family meetings per patient was similar in the FIL group and in the control group (3 [2 to 5] versus 3 [1 to 4], respectively). DISCUSSION Effective communication and a collaborative relationship between ICU caregivers and their patients family members are vital components of quality care (1 4, 9, 13, 14) and of compliance with accreditation requirements (4). We report the first multicenter prospective randomized controlled study of an intervention designed to improve the effectiveness of information provided by ICU caregivers to the family members of their patients. A simple, easily produced leaflet significantly improved comprehension by family members. Moreover, in the subset of patients with good comprehension, the leaflet improved satisfaction. This leaflet both provided information and extended an invitation to talk with ICU caregivers. Comprehension of and satisfaction with the information provided by ICU caregivers are essential if family members are to be not simply visitors to the ICU (13), but participants in some of the aspects of the patient s care and a source of information about the patient s wishes (5, 12, 15). Using our criterion of comprehension, about one-quarter of the family members overall did not understand the diagnosis, prognosis, or treatment in the patient. This is only half the proportion in an earlier study (1), suggesting that efforts made in recent years to improve family management have met with some success (10, 15 17). Analysis of comprehension assessed TABLE 3. CHARACTERISTICS OF FAMILY REPRESENTATIVES* Variable All Representatives No (n 87) p Value Age 49 (37 62) 50 (40 61) 46 (35 63) NS Sex ratio, % women 49 (28) 23 (26.1) 26 (29.9) NS Health care professionals 9 (5.1) 4 (4.5) 5 (5.7) NS Family representatives of French descent 135 (77.1) 66 (75) 69 (79.3) NS Relationship with the patient Spouse 82 (46.8) 44 (50) 38 (43.7) Parent 13 (7.4) 6 (6.8) 7 (8) Child 49 (28) 25 (28.4) 24 (27.6) NS Sibling 21 (12) 8 (9) 13 (14.9) Other relatives 5 (2.8) 2 (2.3) 3 (3.5) Not a relative 5 (2.8) 3 (3.4) 2 (2.3) Hospital commuting time, min 30 (18 60) 45 (16 60) 30 (20 45) NS Allowed/desired information time ratio 0.66 (0.5 1) 0.66 (0.5 1) 0.66 (0.5 1) NS Representatives who felt they received contradictory information 29 (16.6) 17 (19.3) 12 (13.8) NS Representatives who would have liked more information about Diagnosis 113 (66.6) 61 (69.3) 52 (59.7) Prognosis 110 (62.8) 56 (63.6) 54 (62) NS Treatment 112 (64) 59 (67) 53 (60.9) Representatives who did not know the specific role of each caregiver 89 (50.8) 51 (57.9) 38 (43.6) NS Representatives not receiving help from their usual doctor 90 (51.4) 50 (56.8) 40 (45.9) NS Representatives who would have liked help from a psychologist 68 (38.8) 36 (40.9) 32 (36.8) NS Number of physician family meetings between ICU admission and comprehension assessment 3 (2 5) 3 (1 4) 3 (2 5) NS

4 Azoulay, Pochard, Chevret, et al.: Family Information 441 TABLE 4. EVALUATION OF THE EFFECTIVENESS OF THE INFORMATION PROVIDED TO REPRESENTATIVES OF ICU PATIENTS* All Representatives No (n 87) p Value Poor comprehension 46 (26.3) 36 (40.9) 10 (11.5) Poor comprehension of the diagnosis 16 (9.1) 13 (14.7) 3 (3.4) 0.02 Poor comprehension of the prognosis 18 (10.3) 11 (12.5) 7 (8) 0.20 Poor comprehension of the treatment 37 (21.1) 31 (35.2) 6 (6.9) Satisfaction score (CCFNI) 22 ( ) 23 (19 27) 21 (18 26) 0.08 Definition of abbreviation: CCFNI Critical Care Family Needs Inventory. This score can range from 14 (extreme satisfaction) to 56 (extreme dissatisfaction) (5). Satisfactory comprehension of the diagnosis was defined as knowledge of which organ was primarily involved in the disease process; satisfactory comprehension of the prognosis as knowledge of whether the patient was expected to survive (not grave) or not (grave); and satisfactory comprehension of the treatment as knowledge of at least one of the major treatments used among the list of 10 given by the physicians (1). by the ICU investigators, who were blinded to group assignment of family representatives, showed that the FIL reduced the proportion of patients with poor comprehension to approximately 11%. However, the leaflet improved the comprehension of diagnosis and treatment but not of prognosis, reflecting the focus of our leaflet on diagnosis and treatment, and confirming previous reports that understanding the prognosis is difficult (18, 19). The prognosis is often more difficult to determine than the correct diagnosis and appropriate treatment, and this uncertainty may complicate the delivery of easily understandable information on this point. The dichotomous classification of the prognosis as grave or not grave used to inform family members in our study can by no means be viewed as a basis for decision-making in the ICU, but may help families to develop reasonable expectations and to set their coping strategies in motion. In our study, the leaflet was used in combination with information imparted during face-to-face interviews. Consequently, our results should not be construed as meaning that written information used alone can improve comprehension. The criterion for good comprehension used in this study has been validated (1). It involves relatively limited knowledge of the patient s condition. No criterion standard is available for assessing comprehension in family members of ICU patients. We selected a criterion involving limited comprehension because we believe that families cannot be expected to fully grasp all the facets of the patient s condition based only on information given during a brief period after ICU admission. In the ICU, positive interaction between caregivers and family members requires an open exchange of information aimed both at helping family members cope with their distress and at allowing them to speak for the patient if necessary. Our findings indicate that an FIL improves this exchange of information. Earlier data suggest that a leaflet may be perceived by families as a message of welcome extended by ICU caregivers Figure 2. Impact of the FIL on satisfaction in family members with good comprehension. CCFNI score ranged from 14 (entire satisfaction) to 56 (entire dissatisfaction) (5). p 0.01 using the Mann-Whitney test. Bad comprehension, no leaflet; bad comprehension, with leaflet; good comprehension, no leaflet; good comprehension, with leaflet. (1, 11) who place meeting the needs of families among their priorities (6, 15, 16, 20). However, one aspect of the leaflet used in our study was an invitation to talk with ICU caregivers. Talking more with the caregivers may provide opportunities to obtain information in addition to that contained in an FIL, thereby improving comprehension. However, we found no difference between the FIL and control groups regarding the number of physician family meetings. Our finding that the FIL improved comprehension without providing information specific to the patient is in agreement with the hypothesis that efforts should be made upstream to the ICU to improve the general knowledge of family members about ICU operation and vocabulary. Information specific to the patient may be less likely to be grasped by family members. Good comprehension was associated with better satisfaction. Although satisfaction was not significantly associated with delivery of the FIL, among patients with good comprehension those who received the leaflet had significantly better satisfaction scores than those who did not. These data suggest two important interpretations. First, comprehension did not generate dissatisfaction: this runs counter to the hypothesis that poor satisfaction may be related to the seriousness of the patient s condition in family members who would prefer to deny reality, i.e., to an unwillingness of the family to take in the information that is given to them. Second, comprehension is a foundation from which other benefits, such as better satisfaction, can arise: family members with good comprehension are more likely to benefit from the efforts made by ICU caregivers to meet their needs. The prevalence of anxiety and depression in family members may be useful for evaluating the quality of information provided by caregivers to family members of ICU patients. We previously reported that both symptoms were common in family members (21). Nevertheless, the impact of these symptoms on the decision-making capacity of family members has not been evaluated. The present study confirms the extremely high rates of anxiety and depression in family members of ICU patients. Above all, it demonstrates that anxiety and depression measured 3 to 5 d after ICU admission are independent from comprehension and satisfaction. This strongly suggests that family members experience anxiety and depression because the patient is in an ICU, not as a response to information so painful that they cannot cope with it. Meeting the needs of family members of ICU patients and improving their comprehension is of great importance for several reasons. First, more and more families are asking to participate in medical decisions (10, 17, 20). If caregivers and family members are to work together to determine what is best for the patient, then the family members must have a reasonable level of comprehension of the patient s problem and be

5 442 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL reasonably satisfied with the ICU. Comprehension and satisfaction can be evaluated using previously reported criteria (1 3, 5, 12, 16). Second, few patients in the United States and in Europe have living wills or advance directives (22 24). Consequently, when patients are unable to speak for themselves, their families are asked about their wishes. Third, information helps family members to cope with their distress, to build reasonable expectations about the patient s outcome, and to perceive the ICU as a welcoming place where they can wait with relative peace of mind for events to unfold (1, 10, 18). Moreover, our study indicates that understanding information provided by caregivers does not generate anxiety or depression. We evaluated the impact of our leaflet on only one family member of each patient, selected as the patient s representative by the patient or by all available family members. In everyday practice, we believe that each family member should receive the leaflet and that ICU caregivers should check that the information is passed on to other family members (25, 26). In conclusion, a family information leaflet significantly improved comprehension. ICU caregivers should hand a leaflet to family members at their first visit to the ICU. Further interventional studies are needed to identify and evaluate other means of improving the quality of information provided to family members of ICU patients. Acknowledgment : The authors are indebted to A. Wolfe, M.D., for her help in preparing this manuscript. References 1. Azoulay E, Chevret S, Leleu G, Pochard F, Barboteu M, Adrie C, Canoui P, Le Gall JR, Schlemmer B. Half the families of intensive care unit patients experience inadequate communication with physicians. Crit Care Med 2000;28: Azoulay E, Pochard F, Chevret S, Lemaire F, Mokhtari M, Le Gall JR, Dhainaut JF, Schlemmer B, for the French FAMIREA group. Meeting the needs of ICU-patients families: a multicenter study. Am J Respir Crit Care Med 2001;163: Johnson D, Wilson M, Cavanaugh B, Bryden C, Gudmundson D, Moodley O. Measuring the ability to meet family needs in an intensive care unit. Crit Care Med 1998;26: Harvey MA, Ninos NP, Adler DC, Goodnough-Hanneman SK, Kaye WE, Nikas DL. Results of the consensus conference on fostering more humane critical care: creating a healing environment. Society of Critical Care Medicine. AACN Clin Issues Crit Care Nurs 1993;4: Molter NC. Needs of relatives of critically ill patients: a descriptive study. Heart Lung 1979;8: Lilly CM, De Meo DL, Sonna LA, Haley KJ, Massaro AF, Wallace RF, Cody S. An intensive communication intervention for the critically ill. Am J Med 2000;109: Devlin J. Pain assessment in the seriously ill patient: can family members play a role? Crit Care Med 2000;28: Novaes MA, Knobel E, Bork AM, Pavao OF, Nogueira-Martins LA, Ferraz MB. Stressors in ICU: perception of the patient, relatives and health care team. Intensive Care Med 1999;25: Larson CO, Nelson EC, Gustafson D, Batalden PB. The relationship between meeting patients information needs and their satisfaction with hospital care and general health status outcomes. Int J Qual Health Care 1996;8: Maillet RJ, Pata I, Grossman S. A strategy for decreasing anxiety of ICU transfer patients and their families. Nursingconnections 1993;6: McGaughey J, Harrison S. Developing an information booklet to meet the needs of intensive care patients and relatives. Intensive Crit Care Nurs 1994;10: Henneman EA, McKenzie JB, Dewa CS. An evaluation of interventions for meeting the information needs of families of critically ill patients. Am J Crit Care 1992;1: Molter NC. Families are not visitors in the critical care unit. Dimens Crit Care Nurs 1994;13: Jastremski CA, Harvey M. Making changes to improve the intensive care unit experience for patients and their families. New Horiz 1998; 6: Medland JJ, Ferrans CE. Effectiveness of a structured communication program for family members of patients in an ICU. Am J Crit Care 1998;7: Bernat Adell MD, Tejedor Lopez R, Sanchis Munoz J. [How well do patients relatives evaluate and understand information provided by the intensive care unit?]. Enferm Intensiva 2000;11: Robinson SM, Mackenzie-Ross S, Campbell Hewson GL, Egleston CV, Prevost AT. Psychological effect of witnessed resuscitation on bereaved relatives. Lancet 1998;352: Randolph AG, Guyatt GH, Calvin JE, Doig G, Richardson WS. Understanding articles describing clinical prediction tools. Evidence Based Medicine in Critical Care Group. Crit Care Med 1998;26: Randolph AG, Guyatt GH, Richardson WS. Prognosis in the intensive care unit: finding accurate and useful estimates for counseling patients. Crit Care Med 1998;26: Daly K, Kleinpell RM, Lawinger S, Casey G. The effect of two nursing interventions on families of ICU patients. Clin Nurs Res 1994;3: Pochard F, Azoulay E, Chevret S, Lemaire F, Hubert P, Canoui P, Zittoun R, Le Gall JR, Dhainaut JF, Schlemmer B, for the FAMIREA group. Anxiety and depression in families of ICU patients: ethical considerations for decision making capacities. Crit Care Med 2001;29: Most elderly people inpatients haven t heard of living wills. Br Med J 2000;320:E. 23. Hanson LC, Rodgman E. The use of living wills at the end of life. A national study. Arch Intern Med 1996;156: Small H. Increasing completion of advance directives. JAMA 1994;271: Pochard F, Grassin M, Herve C. Palliative options at the end of life. JAMA 1998;279: Pochard F, Grassin M, Azoulay E. Ethical principles for everyone. Ann Intern Med 2001;134: