Journal of Critical Care 29 (2014) 311.e9–311.e16

Contents lists available at ScienceDirect

Journal of Critical Care journal homepage: www.jccjournal.org

The desirability of an Intensive Care Unit (ICU) Clinician-Led Bereavement Screening and Support Program for Family Members of ICU Decedents (ICU Bereave)☆ James Downar, MDCM, FRCPC a,⁎, Reeta Barua b, Tasnim Sinuff, MD, PhD, FRCPC c, 1 a

Divisions of Critical Care and Palliative Care, Department of Medicine, University of Toronto, Toronto, ON, Canada M5G 2C4 Queen's University, Kingston, ON, Canada c Department of Critical Care, Department of Medicine Sunnybrook Health Sciences Center and Interdepartmental Division of Critical Care, University of Toronto, Toronto, ON, Canada M4N 3 M5 b

a r t i c l e Keywords: Intensive care unit Bereavement Grief Social support Health surveys Caregivers Adjustment disorders

i n f o

a b s t r a c t Purpose: Many bereaved family members (FMs) of intensive care unit decedents experience symptoms of complicated grief (CG) or social distress, but there is no standard screening or follow-up for these individuals. We determined the desirability and need for an intensive care unit–based bereavement screening and support program for these FMs. Methods: We surveyed bereaved FMs to measure symptoms of CG, prolonged grief disorder, and social difficulties and the desire for support; and staff physicians and nurses at 2 teaching hospitals in Toronto, Canada, to determine comfort and interest in providing routine bereavement support. Results: We could not contact 69% of FMs largely because of inaccuracies in the patient record. Of the 64 who were contacted, 32 (50%) agreed to be surveyed a mean (SD) of 7.4 (2.2) months after the loss of their relative. Among eligible staff, 57 (61%) of 94 completed the questionnaire. Nine (28%) FMs met subthreshold criteria for CG or prolonged grief disorder, and 7 (22%) met criteria for social distress. Only 10 (31%) had received professional support for emotional symptoms, and 2 (6%) received professional assistance for their social difficulties. Fiftyeight percent supported routine screening, and 68% wanted to receive (or receive more) support. Fifty-five percent of FMs expressed a strong willingness to meet with the medical team to review events surrounding the death of the patient, which was the type of support that the health care staff were most comfortable providing. Most staff (85%) reported providing emotional support at the time of death, but few provided any support afterwards. Fifty-six (98%) of 57 would be willing to support or participate in a formal bereavement screening and support program. Respondents cited the need for training and dedicated time to carry out such a program. Conclusions: An ICU-based bereavement screening and support program for FMs of ICU decedents is both needed and desirable, although there are important needs and barriers. Future studies should evaluate the effectiveness for such a program at improving outcomes among bereaved FMs. © 2014 Elsevier Inc. All rights reserved.

1. Introduction Almost 240,000 Canadians died in 2009 [1], representing 0.7% of the population. When people die, their close family members (FMs) and friends normally experience “grief,” which is an emotional reaction to the loss. In most cases, grief is relatively mild and resolves in less than 6 months without serious social, psychological, or medical consequences. However, severe or prolonged grief may be complicated by a psychiatric or medical disorder that results in declining health, increased use of health care resources, and even death [2,3]. Prolonged grief disorder (PGD)

☆ Conflicts of interest: none to declare. ⁎ Corresponding author. Tel.: +1 416 340 4800x8577; fax: +1 416 340 4823. E-mail addresses: [email protected] (J. Downar), [email protected] (R. Barua), [email protected] (T. Sinuff). 1 Tel.: +1 416 480 6100x2895. 0883-9441/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcrc.2013.11.024

and complicated grief (CG) are terms applied to severe grief-related disorders that persist beyond 6 months. They are distinct from other psychiatric disorders (Appendix A and Shear et al [4]) and respond to specific forms of cognitive behavioral therapy [4,5]. Bereaved FMs are also at risk of social distress, which is marked by difficulties with activities of daily living, financial matters, and interactions with other people [6]. Mortality is common in the intensive care unit (ICU). Intensive care unit mortality is associated with a higher incidence of CG reactions, with symptoms of psychiatric illness in 34% to 67% [7,8] of surviving FMs. However, there is no standard screening or follow-up to identify those at risk. Even when FMs have an identified psychiatric illness or social difficulty, they are often unaware of or unable to access existing services to address these needs. Many bereaved FMs would like a bereavement support service to be available [9]. Moreover, Critical Care Societies have identified bereavement support as a clinical and research priority [10,11].

311.e10

J. Downar et al. / Journal of Critical Care 29 (2014) 311.e9–311.e16

Published palliative care guidelines suggest that bereaved FMs should be contacted in the weeks to months after a death and that each family should have a customized bereavement care plan [12]. Previous authors have suggested that ICU-based health care providers are well positioned to organize a bereavement care plan [3,13]. An ICU-based bereavement screening and support program would be one potential response to calls for improved bereavement support. This program could screen bereaved FMs for symptoms of severe grief, social distress, or informational needs, and provide support or refer them for appropriate follow-up based on their needs. As preliminary work to develop such a program at our institution, we conducted surveys to determine whether an ICU-based bereavement screening and support program would be necessary and desirable from the perspective of FMs and clinicians, and whether ICU-based clinicians would want to participate in such a program. 2. Methods We wanted to understand the information needs and the burden of CG and social distress among FMs of ICU decedents and the proportion who would accept support for these problems. We also sought to determine whether ICU clinicians are currently involved in bereavement support and if they are interested in participating in a formal follow-up and support program for bereaved FMs. We conducted both a telephone-administered survey of bereaved FMs and a self-administered survey of staff physicians and nurses. Patients were identified using local databases at the University Health Network (UHN), a tertiary university–affiliated hospital in Toronto, Canada; and the next-of-kin or emergency contact was identified using the Electronic Patient Record (EPR). For the purpose of this study, nonrelated emergency contacts were considered FMs when no FM was eligible. Bereaved FMs were sent a brief letter explaining the study and notifying them to expect a telephone call in the next 2 weeks, as well as instructions for opting out of the study. The interviewer contacted the FM to offer participation, and the survey was administered to those who provided informed consent. This approach had previously been used in the bereaved, and 80% of those contacted agreed to participate [8]. All attending physicians in 2 medical-surgical ICUs at 2 hospitals in Toronto, Canada (Toronto General Hospital and Toronto Western Hospital), were contacted by email and given the opportunity to complete the survey online via SurveyMonkey. We sent a reminder email after 2 weeks to those who had not responded. We also approached nurses from 2 consecutive 12-hour nursing shifts (daytime and nighttime) in the same ICUs and offered paper copies of the survey. We collected responses for the nurses 48 hours after distribution. We included any adult English-speaking FM or close friend listed as “emergency contact” of any patient who died in an ICU at the UHN in 2011 to 2012 within 3 and 12 months of screening. If the next of kin was not English speaking, we asked for the closest relative who was English speaking provided that relative was close to the deceased (either lived with the patient or was involved in the patient's care and decisions during the relative's hospital visit). We also included any staff physician or nurse working in the medical-surgical or neurosurgical ICU at UHN. We excluded FMs who were non–English speaking, younger than 18 years, unable to provide informed consent to participate, or not close to the deceased (did not live with the patient and was not involved in the patient's care or decisions during their hospital visit) and those who refused to provide consent to participate. We collected general demographic information about all participants (ICU staff were allowed to withhold their demographic information). We used the following survey instruments for FMs: 1. Inventory of Complicated Grief—Revised (ICG-r) [14]: This tool has been used by different authors to diagnose either PGD [14] or CG [15]. These conditions are similar in that they both require

the loss of a loved one more than 6 months previously and subjective social impairment, but they differ in the criteria used to diagnose separation distress and cognitive, emotional, or behavioral symptoms. In this study, we present the results using both sets of criteria. Because we enrolled FMs as early as 3 months postloss and we are measuring social difficulty separately, we also report the number of FMs who fulfill “subthreshold” criteria [8] (ie, those who meet all criteria for PGD and/or CG except for subjective social difficulty and waiting 6 months from the loss) (see Appendix B for full criteria). 2. Social Difficulties Inventory (SDI) [6,16]: This tool has been used to diagnose “social distress” on the basis of a score of at least 10 on a subset of 16 questions (SD-16) where individual responses are assigned a score of 0 to 3. 3. Questionnaire developed by the study investigators to assess both the use of existing services and the desire for additional supports during bereavement. For ICU clinicians, we developed the following questionnaire: 1. Questionnaire based on items in the ICG-r and SDI to assess clinicians’ perception of the burden of CG and social distress, and their comfort with providing support for these problems. The questionnaire included items focused on current involvement in bereavement support, and interest in further involvement and training to participate in a program of bereavement screening and support. The questionnaire was developed by the authors and underwent sensibility testing and was piloted in a group of ICU physicians and nurses at the study site prior to the survey. 2.1. Statistical analysis Our analysis was primarily descriptive. For the FM survey, we compared responses from bereaved FMs with and without subthreshold CG/PGD. For the clinician survey, we compared nurse and physician responses. Because of the small sample size for each survey, we converted 5-point Likert scales into dichotomous outcomes and constructed contingency tables by combining the “slightly” and “very” willing responses into a single category, and did the same for the “unwilling” responses, and the “agree” and “disagree” responses. To compare social difficulties between FMs with and without subthreshold CG/PGD, we combined “very much” and “quite a bit” into a single category [17] and “a little” and “no difficulty” into a single category for contingency analysis. We used Fisher exact test for all categorical comparisons. We also compared SD-16 scores (derived from the SDI) of FMs with and without subthreshold CG/PGD using the MannWhitney U test and the proportion who had an SD-16 score of at least 10 (the threshold for “social difficulty” [6]) using Fisher exact test. This study was approved by the UHN Research Ethics Board (Protocols 12-5081-AE and 12-5078-AE). 3. Results 3.1. Bereaved FMs We present eligibility and enrolment data in Fig. 1. We were unable to contact 148 (69%) “emergency contacts” because of incomplete or inaccurate contact information, or loss of contact after initial contact. In total, 32 FMs participated (1 terminated the interview early but did not want to withdraw from the study), representing 15% of decedents but 50% of those who were contacted and met inclusion criteria. Table 1 shows demographic information for the FMs and relevant clinical information about the decedents. The average participant was a well-educated female in her late 50s, grieving the loss of her

J. Downar et al. / Journal of Critical Care 29 (2014) 311.e9–311.e16

Fig. 1. Family member participant flow diagram.

Table 1 Demographics of FMs Total participants

32

Age, mean (SD) Female sex (%) Education level (%) Primary Secondary Postsecondary Undergraduate university Postgraduate university Relationship to deceased (%) Spouse/Ex-spouse Child Parent Sibling Other relative Marital status (%) Widowed Married Divorced/separated Never married Religious affiliation (%) Roman Catholic Protestant Christian Orthodox Christian No religious affiliation Country/region of origin (%) Canada Central/South America Africa Asia Months since death of relative, mean (SD)

59 (12.5) 19 (61%) 1 (3%) 14 (45%) 1 (3%) 11 (35%) 4 (13%) 19 (61%) 6 (19%) 3 (10%) 2 (6%) 1 (3%) 18 (58%) 8 (26%) 3 (10%) 2 (6%) 10 (31%) 14 (44%) 2 (6%) 6 (19%) 25 (81%) 2 (6%) 2 (6%) 1 (3%) 7.4 (2.2)

311.e11

spouse. Twenty-six (81%) described themselves as Christian, and 27 (84%) reported being “somewhat” or “very” religious. Overall, a large majority of FMs were satisfied with the care received by the decedent before he/she died and would recommend the hospital to others (Table 2). There was a low incidence of uncontrolled symptoms from the perspective of the bereaved. Twenty-five (78%) patients had made a living will or power of attorney form, and most FMs recalled making a decision about life support or CPR (18/32, 56%); but very few could recall involvement of the Palliative Care team (5/32, 16%). Notably, only 4 (13%) of 32 recalled receiving any emotional support from hospital staff in the days or weeks after the death, and only 7 (22%) recalled being contacted by a health care team member during that time. The ICG-r revealed that only 1 patient met all criteria for CG (Table 3), but 9 (28%) of 32 met subthreshold criteria for either CG or PGD. None of the latter reported that the quality of care received by their loved one was “poor,” and they were not more likely to state that they would not recommend the hospital to others than FMs without subthreshold CG or PGD (22% vs. 13%, P = .3). Family members who met subthreshold criteria for CG or PGD were more likely to recall having made a decision about life-sustaining treatment for the deceased than those without subthreshold CG or PGD (88% vs. 45%, P = .02). The SDI revealed an overall low burden of social difficulty (median [interquartile range {IQR}] SD-16 score, 3 [6.75]). Seven (23%) FMs met criteria for social distress (SD-16 score ≥ 10). Although FMs with subthreshold PGD or CG had significantly higher SD-16 scores than those without subthreshold PGD or CG (median score, 5 vs. 1; P = .02), they were not more likely to meet the threshold of social distress (33% vs. 19%, P = .32). Ten FMs (32%) reported receiving professional help for mood or anxiety problems; 8 (26%) of 31 reported seeing a psychiatrist/ psychologist, and 9 (29%) of 31 reported seeing a nonpsychiatrist/ nonpsychologist professional. Fifteen (48%) reported only informal support from friends and family, whereas 6 (19%) reported receiving no support. Five of the 9 FMs with either subthreshold CG or PGD had not received any professional support for their symptoms. Only 2 participants reported receiving any formal support with finances, transportation, or self-care. Most were receiving no support on any domain of social function. Of those who received support from a psychiatrist/psychologist for depression or anxiety, 8 of 9 reported that this was helpful, whereas one described this support as “very harmful.” There were no significant differences between FMs with and

Table 2 Experience of care provided to deceased from perspective of FM Care description Overall quality of care Excellent Very good Good Fair Poor Would you recommend this hospital to friends and family members? Yes, definitely Yes, probably No Did the patient experience pain and suffering that was NOT adequately treated? Yes No Unsure Did you have to make a decision about life-sustaining treatment or CPR on behalf of the patient? Yes No

No. of respondents (%) 10 (31%) 14 (44%) 5 (16%) 1 (3%) 2 (6%)

19 (59%) 8 (25%) 5 (16%)

5 (16%) 25 (78%) 2 (6%)

18 (56%) 14 (44%)

311.e12

J. Downar et al. / Journal of Critical Care 29 (2014) 311.e9–311.e16

3.2. ICU nurses and physicians

Table 3 Grief, social distress, and supports received by bereaved FMs ICG-r Met full criteria for PGD Met full criteria for CG Subthreshold criteriaa PGD CG Either PGD or CG SDI Social distress (SD-16 ≥ 10) Social distress among FMs with subthreshold PGD or CG Social distress among FMs without subthreshold PGD or CG SD-16 (median, IQR) SD-16 score among FMs with subthreshold PGD or CG (median, IQR) SD-16 score among FMs without subthreshold PGD or CG (median, IQR) Currently receiving professional help for mood or anxiety Would like (or would like more) formal support for mood or anxiety … among FMs with subthreshold PGD or CG … among FMs without subthreshold PGD or CG

0 1 (3%) 5 (16%) 8 (25%) 9 (28%) 7 (23%) 3 (33%)⁎ 4 (19%)⁎ 3 (6.75) 5 (8.75)⁎⁎ 1 (4)⁎⁎ 10 (32%) 21 (68%) 7 (77%)⁎⁎⁎ 14 (63%)⁎⁎⁎

a Exclude the requirement of waiting 6 months and having subjective social impairment. ⁎ P = .32 for difference, Fisher exact test. ⁎⁎ Z = 2.35, P = .02 for difference, Mann-Whitney U test. ⁎⁎⁎ P = .88 for difference, Fisher exact test.

without subthreshold CG/PGD in the type of support received or the satisfaction with that support. Twenty-one (68%) of 31 were willing to receive support (or more support) for depression/anxiety, and 16 to 18 (52%-58%) would accept support for various aspects of social function (data not shown). Family members with subthreshold CG or PGD were not more willing to receive support than those without subthreshold CG or PGD (data not shown). Twelve (57%) of 21 participants who requested support were not receiving any formal support at all. Participants had a strong willingness to meet with the medical team to review events surrounding the death of their loved one (17/31 [55%] were “very willing” to do so). The majority of participants (18/31, 58%) would have wanted follow-up after the death of a loved one (Fig. 2), whereas only 6 (19%) would not have wanted follow-up. Ten (32%) said that they would find this stressful; but 6 of those 10 would still want to be contacted, whereas only 3 would not want to be contacted.

Ninety-four ICU staff were offered participation in the survey. The overall response rate was 57 (61%) of 94: 10 (50%) of 20 for physicians and 47 (64%) of 74 for nurses. Demographics for the respondents are provided in Table 4. The average respondent had 15 years of clinical experience. The largest proportion (18/52, 35%) reported no religious affiliation, but half of respondents described themselves as “somewhat” or “very” religious or spiritual. When asked about their current involvement in bereavement support (Fig. 3), clinicians reported that they typically provided support at the time of death (“Often” or “Always or almost always,” 47/55 [85%]) and that they provided emotional support to FMs (“Often” or “Always or almost always,” 46/54 [85%]). Otherwise, the majority of respondents reported “Never” or “Rarely” providing support at a later time or assessing whether FMs required assistance with finances or social function. Only 14 (26%) of 54 reported assessing whether bereaved FMs wanted to review events surrounding the death of the patient “Often” or “Always or almost always.” Clinicians believed that bereaved FMs most often required help with emotional disturbances (“Often” or “Always or almost always,” 33/54 [61%]); but fewer than 25% believed that FMs had a similar need for assistance with social function or a need to review events surrounding the death of the patient with the medical team. The majority (31/54 [57%]) reported feeling comfortable providing bereavement support to FMs; only 13 (24%) reported feeling uncomfortable. But when asked about comfort with specific aspects of support (Fig. 4), the majority of clinicians only felt comfortable meeting with the FMs to review the events surrounding the death of the patient (34/49 [69%]). Clinicians felt least comfortable providing support for issues of finance. Nurses felt more comfortable than physicians providing or arranging support to maintain independence at home (P = .04), support for personal care (P = .03), and support for transportation/mobility (P = .04). There was a strong willingness to receive formal training to provide bereavement support (46/54 [85%]), and to participate in a formal support program for FMs (38/52 [73%]). Only 2 respondents (4%) were unwilling to receive support training, and 8 (15%) were unwilling to participate directly in a formal program. If a formal program existed at our facility, 52 (98%) of 53 participants would be willing to participate in the program or cover for a colleague who was providing support for an FM. Nurses were significantly more willing than physicians to receive support training (90% vs. 64%, P = .03).

Fig. 2. Family members’ feelings about being contacted in the postdeath period. SA = strongly agree, NAND = neither agree nor disagree, D = disagree, SD = strongly disagree.

J. Downar et al. / Journal of Critical Care 29 (2014) 311.e9–311.e16 Table 4 Demographics of nurse and physician participants Participants, N

57a

Age, mean (SD) Female sex (%) Profession (%) Physician Nurse Years since completing medical/nursing degree, mean (SD) Country/region of origin Canada Asia Europe Central America/Caribbean Africa Australia/Oceania South America USA Religious affiliation (%) No religious affiliation Roman Catholic Protestant Christian Other Christian Jewish Buddhist Hindu Other Decline to answer Degree of religiosity/spirituality (%) Very religious/spiritual Somewhat religious/spiritual Minimally religious/spiritual Not religious/spiritual Decline to answer

41.2 (10.4) 36 (82%) 40 (78%) 11 (22%) 15.4 (10.1) 23 (45%) 8 (16%) 6 (12%) 5 (10%) 3 (6%) 3 (6%) 2 (4%) 1 (2%) 18 (35%) 9 (17%) 8 (15%) 7 (14%) 3 (6%) 3 (6%) 2 (4%) 1 (2%) 2 (4%) 10 (19%) 16 (31%) 11 (21%) 12 (23%) 3 (6%)

a Participants were allowed to withhold their demographic information, so not all questions have 57 responses.

311.e13

Participants were asked to indicate what made it difficult for them to support bereaved FMs. The most common barriers mentioned were “not knowing what to say,” dealing with emotion, a lack of knowledge about existing supports in the community, high clinical workload, and a lack of continuity or established relationship with the patient/FM. The most common facilitators mentioned for provision of bereavement support included support training, a list of available support resources, and dedicated time after the death and at a later date to provide support. 3.3. Interpretation In the present study, we used 2 surveys to determine the need and desirability of an ICU-based bereavement screening and support program for recently bereaved FMs of patients who died in the ICU. We found that despite an overall high satisfaction with patient care and a very low prevalence of CG or PGD, an important proportion of bereaved FMs experienced “subthreshold” CG and PGD, as well as social distress. Family members had a strong desire for routine bereavement screening and support regardless of whether or not they were experiencing symptoms of CG and PGD. Although ICU clinicians rarely provided bereavement support beyond the acute postmortem period, we found that they would be willing to participate in further screening and support for bereaved FMs if given appropriate time and training. We also found that the types of support most desired by FMs (emotional support and a meeting to review events prior to death) were those that clinicians felt most comfortable providing. Importantly, there was a large gap between the high prevalence of social difficulties (eg, managing finances, performing activities of daily living) FMs faced and the low comfort and knowledge of ICU clinicians who might be asked to organize social support. We also identified important barriers and

Fig. 3. Current clinician involvement in bereavement support. A = always or almost always, S = sometimes, R = rarely.

311.e14

J. Downar et al. / Journal of Critical Care 29 (2014) 311.e9–311.e16

Fig. 4. Clinician comfort with providing aspects of bereavement support. NCNU = neither comfortable nor uncomfortable, VU = very uncomfortable.

facilitators to a bereavement screening and support program, most notably the high prevalence of inaccurate contact information in the EPR and the need for formal training to address support needs. To our knowledge, no previous study has explored both ICU clinician and FM perspectives regarding the desirability of and need for an ICU-based bereavement screening and support program for bereaved FMs. Siegel et al [8] surveyed 41 bereaved FMs of ICU decedents 3 to 12 months postmortem, finding that 34% fulfilled criteria for at least 1 major psychiatric illness but that only 5% of FMs met full criteria for PGD. As in our study, they enrolled patients before the 6-month mark, which could explain the similarly low prevalence of PGD; 22% met “subthreshold” criteria for PGD. Notably, 14% had undergone psychiatric treatment, compared with 22% in the present study. Our FMs may have had a high prevalence of PGD or CG symptoms because of the large number of spouses in our cohort [8] and a high frequency of involvement in decisions to limit or withdraw life support [18]. Many FMs report concerns about the end-of-life care provided in acute settings [19,20]; but our FMs reported good symptom control and overall satisfaction with care, which may be protective against CG [21]. Our findings underscore the fact that a well-managed death and good support in the predeath period are not sufficient protection against symptoms of CG or PGD. There was a widespread desire among FMs for a bereavement support service that would address a variety of informational, psychological, and social needs. Our findings differed from those of Van der Klink et al [9], who found that only 35% of bereaved Dutch FMs wanted such a service to be available. Of note, participants in the aforementioned study reported a good understanding of the sequence of events that led to death, whereas the majority of our participants wanted to review these events with the medical team. We can only speculate as to why so many of our participants wanted to review the events that occurred prior to death; but we know that FMs of ICU patients often have a poor understanding of the diagnosis and care plan [22], and interventions to improve this understanding have been shown to improve satisfaction levels [23]. Furthermore, FMs who feel that they have received contradicting information are more likely to experience depressive symptoms [24]. These observations suggest

that the ICU team might be able to meet a large proportion of FMs’ support needs with a short meeting and simple explanations about the medical care received prior to death [13]. Coincidentally, this was the type of support that ICU clinicians felt most comfortable providing. Such meetings would require little training and relatively little time compared with other forms of support, making them a key element of any future bereavement program. At the same time, an important proportion of bereaved FMs experience meaningful psychological symptoms and are at risk of suicidality and poor health-related quality of life. Previous data from Lichtenthal et al [25] suggest that only a minority of these FMs seek mental health services and that the deceased patient’s health care providers may play a vital role in helping FMs connect with mental health services. Screening is important because universal “preventative” grief therapy does not appear to be beneficial for all bereaved individuals [5] and may squander limited supportive resources. However, cognitive behavioral therapy is effective for treating those who have developed CG; and the treatment effects increase in the follow-up period [4,26]. Antidepressant medication may be a useful adjuvant, helping distressed FMs to complete a course of behavioral therapy [27]. More than two thirds of FMs reported that they would want (or would want more) formal support for their mood and anxiety symptoms, and half of those with subthreshold CG or PGD were not receiving any formal psychiatric or psychological support. Interestingly, the presence of subthreshold CG or PGD was not associated with the desire for support, which may partially explain why symptomatic individuals often fail to connect with mental health services. Given that we approached many FMs in the early bereavement period (b6 months), we do not know whether therapy would have improved their symptoms, or whether many would have improved or remained symptomatic regardless of treatment received. The effectiveness of therapy for improving outcomes of those in the early bereavement period has not been demonstrated, but we cannot ignore the fact that so many participants expressed a desire to receive support less than 6 months after their loss. Our surveys did not specifically identify what types of emotional support the FMs would want to receive. Cognitive behavioral therapy

J. Downar et al. / Journal of Critical Care 29 (2014) 311.e9–311.e16

is known to be effective for CG, but other therapies such as psychotherapy are also available. The surveys also did not identify which health care professional(s) would provide the support requested. An ICU-based screening and support program would likely comprise many members within an interdisciplinary team: for example, ICU clinicians could review events that occurred during admission; psychiatrists, psychologists, or spiritual care practitioners could address severe psychological symptoms; and social workers could focus on social difficulties. Our findings suggest that any program for early screening of bereaved FMs may need to consider a screening tool other than the ICG-r to guide referral to support services. Other more sensitive tools could be used to determine the need for psychological or psychiatric support, and the prevalence of CG and PGD may be better suited as outcome measures for such a program. Additional brief assessment tools are available [28,29], although they also have yet to be validated in the early bereaved. We need to acknowledge that bereavement screening and support have the potential to exacerbate symptoms of CG for a proportion of bereaved FMs. In our study, one third of FMs reported that it would be “stressful” to be contacted by hospital staff; and one of the FMs described his/her professional psychiatric care as “very harmful.” However, the risk of harm needs to be viewed in the context of potential benefit: CG can be treated [26], and more than half of FMs who would find it stressful said that they would still want to be contacted in spite of the stress. Nevertheless, any screening and support program should always approach bereaved FMs in a gentle, sensitive manner to avoid exacerbating the trauma of their loss. Selfreferral (“inreach”) is one potential approach to minimize this risk, and it may also be a more effective means of selecting FMs than traditional “outreach” programs (where bereaved FMs are approached by health care workers) [30]. Our study had limitations. We were not able to measure the predeath prevalence of mood disorders, anxiety disorders, and other psychiatric illnesses among FMs. We also did not specifically assess for major depression, posttraumatic stress disorder, or other psychiatric illness because of concerns about the total interview length and the difficulties of making hierarchical diagnoses [15]. We chose to include “subthreshold” FMs who had been bereaved for less than 6 months, so our results may not be generalizable to true CG or PGD (N6 months). Complicated grief is an evolving field of study; and the controversy surrounding the inclusion of “Persistent Complex Bereavement

311.e15

Disorder” in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, has illustrated the lack of consensus on how to define this entity [31]. The low overall FM response rate raises concerns about sampling bias, although it is difficult to estimate how this bias might have affected the measured prevalence of CG and PGD, or the desire for support. The high prevalence of inaccurate contact information in the EPR also limits our interpretation of the data, but it is an important finding. This represents the current limitations of our system, and demonstrates the need to collect this information more accurately and independently as part of any future screening and support program. Our surveys were also subject to social desirability bias; FMs may have overstated their desire to receive bereavement support, and staff may have overstated their willingness to provide it. Our study also has strengths. We were able to conduct telephone interviews of FMs of a very sensitive subject during a very difficult time in their lives. We used a pragmatic approach, focusing on FMs in the early bereavement period. Although it is considered “normal” to grieve during this period, many FMs clearly experience important symptoms and expressed a clear desire for support. Finally, we included ICU clinicians’ perspectives, which showed important complimentary data that will inform future studies of the feasibility of this type of program. In conclusion, we found that an ICU-based bereavement screening and support program for FMs of ICU decedents was needed and desirable to both FMs and clinicians, although we identified important educational and supportive needs and barriers. Because this type of program is a potential response to broad societal calls for improved bereavement support, future studies should evaluate the feasibility and effectiveness of such a program for improving outcomes among bereaved FMs. 4. Conflicts of interest On behalf of all authors, the corresponding author states that there is no conflict of interest. Acknowledgments We are grateful to the generosity of the FMs who participated in this study at a very difficult time and the ICU clinicians who responded to the questionnaires. This study was funded by the Toronto General Hospital Foundation.

Appendix A Complicated grief and prolonged grief disorder PGD

CG

A. Loss of a significant person A. Loss of a loved one N6 mo ago B. Separation distress: feelings of yearning that occur daily or cause B. At least one symptom of acute grief present for longer than expected in the person’s culture: disability • Persistent intense yearning • Frequent intense loneliness/emptiness • Recurrent feelings of unfairness or meaninglessness of life, or a desire to die in order to rejoin the deceased • Frequent intrusive thoughts about the deceased C. At least 5 specific cognitive, emotional, or behavioral symptoms: C. At least 2 of the following symptoms present for N1 mo: • Diminished sense of self; • Rumination about the death and its consequences; • Difficulty accepting loss; • Disbelief, inability to accept death; • Avoidance of reminders of loss; • Feeling of shock, numbness; • Inability to trust others; • Bitterness or anger related to loss; • Bitterness or anger related to loss; • Inability to trust others; • Difficulty moving on with life; • Experiencing pain/symptoms that deceased person experienced; • Emotional numbness; • Intense reaction to memories/reminders of loss • Feeling that life is meaningless; • Disturbing reaction to memories/reminders of loss • Feeling stunned/dazed/shocked by the loss • Excessive avoidance/proximity seeking relevant to deceased D. Timing: ≥6 mo since loss D. Duration: symptoms and impairment lasting ≥1 mo E. Impairment: significant social, occupational, or functional E. Impairment: significant social, occupational, or functional impairment caused by symptoms and not impairment better explained as a culturally appropriate response F. Not better accounted for by other psychiatric disorder. CG and PGD. Adapted from Shear et al [15].

311.e16

J. Downar et al. / Journal of Critical Care 29 (2014) 311.e9–311.e16

Appendix B. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jcrc.2013.11.024. References [1] Mortality statistics. http://www.statcan.gc.ca/pub/84f0209x/2009000/t001-eng. htm; 2012 . [Accessed October 12, 2012]. [2] Prigerson HG, Jacobs SC. Perspectives on care at the close of life. Caring for bereaved patients: "all the doctors just suddenly go". JAMA 2001;286(11):1369–76. [3] Cuthbertson SJ, Margetts MA, Streat SJ. Bereavement follow-up after critical illness. Crit Care Med Apr 2000;28(4):1196–201. [4] Shear K, Frank E, Houck PR, Reynolds III CF. Treatment of complicated grief: a randomized controlled trial. JAMA 2005;293(21):2601–8. [5] Mancini AD, Griffin P, Bonanno GA. Recent trends in the treatment of prolonged grief. Curr Opin Psychiatry Jan 2012;25(1):46–51. [6] Wright P, Smith AB, Keding A, Velikova G. The Social Difficulties Inventory (SDI): development of subscales and scoring guidance for staff. Psychooncology Jan 2011;20(1):36–43. [7] Lautrette A, Darmon M, Megarbane B, et al. A communication strategy and brochure for relatives of patients dying in the ICU. N Engl J Med 2007;356(5):469–78. [8] Siegel MD, Hayes E, Vanderwerker LC, Loseth DB, Prigerson HG. Psychiatric illness in the next of kin of patients who die in the intensive care unit. Crit Care Med Jun 2008;36(6):1722–8. [9] van der Klink MA, Heijboer L, Hofhuis JG, et al. Survey into bereavement of family members of patients who died in the intensive care unit. Intensive Crit Care Nurs Aug 2010;26(4):215–25. [10] Truog RD, Campbell ML, Curtis JR, et al. Recommendations for end-of-life care in the intensive care unit: a consensus statement by the American Academy of Critical Care Medicine. Crit Care Med Mar 2008;36(3):953–63. [11] 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 Feb 2007;35(2):605–22. [12] Hudson P, Remedios C, Zordan R, et al. Guidelines for the psychosocial and bereavement support of family caregivers of palliative care patients. J Palliat Med Jun 2012;15(6):696–702. [13] Fauri DP, Ettner B, Kovacs PJ. Bereavement services in acute care settings. Death Stud Jan-Feb 2000;24(1):51–64. [14] Prigerson HG, Horowitz MJ, Jacobs SC, et al. Prolonged grief disorder: psychometric validation of criteria proposed for DSM-V and ICD-11. PLoS Med Aug 2009;6(8):e1000121. [15] Shear MK, Simon N, Wall M, et al. Complicated grief and related bereavement issues for DSM-5. Depress Anxiety Feb 2011;28(2):103–17.

[16] Wright EP, Kiely M, Johnston C, Smith AB, Cull A, Selby PJ. Development and evaluation of an instrument to assess social difficulties in routine oncology practice. Qual Life Res 2005;14(2):373–86. [17] Smith AB, Wright P, Selby P, Velikova G. Measuring social difficulties in routine patient-centred assessment: a Rasch analysis of the social difficulties inventory. Qual Life Res Jun 2007;16(5):823–31. [18] Azoulay E, Pochard F, Kentish-Barnes N, et al. Risk of post-traumatic stress symptoms in family members of intensive care unit patients. Am J Respir Crit Care Med 2005;171(9):987–94. [19] Stajduhar k, Nicol D, Heyland D. Differences in quality of end-of-life care provided in inpatient health care settings. J Palliat Care 2010;26(3):1. [20] Rodriguez Villar S, Sanchez Casado M, Prigerson HG, et al. Prolonged grief disorder in the next of kin of adult patients who die during or after admission to intensive care. Chest Jun 2012;141(6):1635–6. [21] Mularski RA, Heine CE, Osborne ML, Ganzini L, Curtis JR. Quality of dying in the ICU: ratings by family members. Chest Jul 2005;128(1):280–7. [22] Azoulay E, Chevret S, Leleu G, et al. Half the families of intensive care unit patients experience inadequate communication with physicians. Crit Care Med Aug 2000;28(8):3044–9. [23] Azoulay E, Pochard F, Chevret S, et al. Impact of a family information leaflet on effectiveness of information provided to family members of intensive care unit patients: a multicenter, prospective, randomized, controlled trial. Am J Respir Crit Care Med 2002;165(4):438–42. [24] Pochard F, Azoulay E, Chevret S, et al. Symptoms of anxiety and depression in family members of intensive care unit patients: ethical hypothesis regarding decision-making capacity. Crit Care Med Oct 2001;29(10):1893–7. [25] Lichtenthal WG, Nilsson M, Kissane DW, et al. Underutilization of mental health services among bereaved caregivers with prolonged grief disorder. Psychiatr Serv Oct 2011;62(10):1225–9. [26] Wittouck C, Van Autreve S, De Jaegere E, Portzky G, van Heeringen K. The prevention and treatment of complicated grief: a meta-analysis. Clin Psychol Rev Feb 2011;31(1):69–78. [27] Simon NM, Shear MK, Fagiolini A, et al. Impact of concurrent naturalistic pharmacotherapy on psychotherapy of complicated grief. Psychiatry Res 2008;159(1–2):31–6. [28] Ito M, Nakajima S, Fujisawa D, et al. Brief measure for screening complicated grief: reliability and discriminant validity. PLoS One 2012;7(2):e31209. [29] Holland J, Currier JM, Coleman RA, Neimeyer RA. The Integration of Stressful Life Experiences Scale (ISLES): development and initial validation of a new measure. Int J Stress Manag 2010;17(4):28. [30] Schut H, Stroebe MS. Interventions to enhance adaptation to bereavement. J Palliat Med 2005;8(Suppl 1):S140–7. [31] Boelen PA, Prigerson H. Commentary to the inclusion of persistent complex bereavement-related disorder in DSM-5. Death Stud (in press).