Introduction

When patients die, family members (FMs) and close friends experience grief related to their loss. Grief reactions can be relatively brief, but can also be prolonged and associated with impaired social function as well as psychological and medical morbidity.1,2,3 These grief reactions, variously referred to as complicated grief (CG),4 prolonged grief disorder,5 or persistent complex bereavement-related disorder,6 are common among FMs of patients who die in the intensive care unit (ICU)1 with a reported incidence as high as 52% (depending on the criteria applied).7,8,9

Bereavement support has been identified as a research and clinical priority within the ICU setting10,11 and many bereaved FMs of those who die in the ICU express the desire for bereavement support.9,12 Nevertheless, there is little research about structures, processes, and outcomes for routine bereavement support in this context. While effective treatments for severe grief reactions exist,13,14 there are barriers to identifying and supporting bereaved FMs with severe symptoms. For example, FMs may be reluctant to be contacted or to accept bereavement support. In a mixed-methods study of 105 bereaved FMs, we previously found that 19% met criteria for CG, but only half were willing to receive formal support.15 Notably, there was no correlation between the presence of CG and the willingness to receive support. In the quantitative part of this study, we focused on identifying symptoms and predictors of CG in FMs after a loss. This paper describes the bereavement experience of those who developed CG, including their coping strategies and sources of support, and their advice for approaching future FMs at risk of CG.

Methods

The protocol was initially approved (September 4, 2014) by the University Health Network (UHN) Research Ethics Board (REB) for a mixed-method study. Following the UHN approval, REB approval was also obtained from the eight following participating centres: UHN - Toronto Western Site, Toronto; Sunnybrook Health Sciences Centre, Toronto; St. Michael’s Hospital, Toronto; St. Joseph’s Hospital, Hamilton; Foothills Medical Centre, Calgary; Peter Lougheed Centre, Calgary; Rockyview General Hospital, Calgary; and South Health Campus, Calgary.

Study design and sample size

The overall study used a two-phase sequential explanatory design involving FMs and ICU healthcare providers. In the quantitative portion, we recruited ICU staff and bereaved FMs to complete surveys including the inventory of complicated grief (ICG). The FMs ICG scores were measured at three and six months to identify those with symptoms of CG, and to identify those who were willing to participate in the qualitative portion of the study. For the qualitative portion of the study, we used purposive sampling to ensure a good representation of all study sites and demographics from the 105 FMs who enrolled in the quantitative portion of the study. Of the 37 FMs who were interviewed at three months after the death of their loved one, eight FMs were found to have an ICG score above 25 at six months after the loss (Appendix B). For this report, we analyzed the three- and six-month interview data from these eight FMs who met criteria for CG because we are interested in developing a bereavement program for those with CG.

Data collection

Two qualitative researchers (C.K., A.R.) conducted the semi-structured interviews. The interview guide was developed to move from open to increasingly structured questions. We started with an open-ended question designed to elicit participants’ stories that allowed them to control the direction, content, and pace of the conversation. We then used probing questions to explore their narrative in greater depth, asking about topics they had raised and incorporating words they had used. Lastly, we used semi-structured questions developed from the research team’s expertise, a literature review on bereavement, and topics frequently raised by the participants (Appendix A).

Participants were first invited to share their bereavement narratives following the death of their loved one in the ICU. The latter would enable participants to control the direction of their stories and to set the pace of the conversation.16 Afterwards, specific follow-up questions were presented to understand discrete time frames of their ICU experiences: 1) the time leading up to death, 2) the time after death, 3) commemoration, and 4) existing and future need of support (Appendix A). Interviews were audio-recorded and transcribed.

Data analysis

A grounded theory approach was employed. First, four authors (J.D., T.S., C.K., A.R.) used thematic analysis as articulated by Braun and Clarke to analyze the data.17 Analysis consisted of inductive and iterative open coding that identified emergent themes that would be labelled from terms relevant to the literature or used by participants. Second, the emerging codes were discussed and organized into core themes. NVivo software (QSR International, Doncaster, Australia) was used to facilitate qualitative analysis. Finally, the team organized the main themes to articulate a comprehensible understanding of the analyzed data. Methodological rigor was assured using qualitative researchers with expertise in qualitative methods and bereavement, including critical care, palliative care, and anthropology.

Results

We interviewed eight bereaved FMs with CG at three months post-loss, of which four were able to complete a second interview at six months. These eight participants represented five of the study sites, and their demographics appear in Appendix C. Thematic analysis identified five common themes related to the FMs’ bereavement experience and their coping or support.

Ante-mortem experience

Although the interview questions focused on bereavement, FMs often spoke of experiences before death, including staff support. Nevertheless, most expressed concerns about the care received or the structure of the ICU, and some made a direct link between care delivered and the FM’s bereavement experience. One woman who lost her husband in the ICU explained how a doctor told her, “to take [my husband] off life support, literally after having met me 30 sec before, and I ran down the stairs” followed by a “kind” respiratory therapist who apologized for the physician’s directness.

In addition, many expressed disappointment with the communication and support offered in the ICU. One FM stated that his uncle’s illness had not been taken seriously until the situation had become dire and he was subsequently dissatisfied with the information communicated. In general, FMs were particularly frustrated with the inaccessibility of physicians and the lack of explanations they received.

Communication received by FMs in the ICU was not necessarily aimed at supporting grieving FMs. One FM who was also a physician and whose wife died in the ICU explained how, “ICU people were friendly, objective, knowledgeable, but they weren’t supportive in terms of grief”. This would leave dealing with the emotional elements of the dying process entirely up to the FM. Participants also expressed frustration with the lack of continuity between attending ICU staff and the resultant disjointed communication.

Post-mortem experience

Family members reported specific frustrations with the hospital and the care team. These experiences took place following the death of their loved one and often included interactions with the hospital or ICU staff, or their lack thereof. One FM described a “ridiculous” situation of trying to obtain an autopsy report of her deceased uncle that was sent to his family doctor; later being offered the documents at “a dollar [cost] a sheet”. This left the family unsure about the cause of death, which negatively influenced their bereavement experience. Another participant expressed “a total loathing of the foundations of these [hospital] organizations”, who continued to send letters to his deceased wife asking for donations because she had been a patient there. He explained how receiving these persistent letters would re-ignite his anger and impair his recovery.

Participants also described challenges they experienced related to commemoration ceremonies. Some reported feeling irritated by the behaviour of others at the funeral. In addition to the emotional intensity of the ceremony, some found it challenging to create a monument worthy of the deceased. One of the FMs was unable to hold a ceremony for over a year following his wife’s death.

Most participants described specific and unpredictable events that would “trigger” an uncontrollable emotional reaction. These trigger events were so upsetting that FMs would limit their exposure to and interactions with others as a coping strategy. One participant explained the unpredictability of these triggers. She recounted how, months after her husband’s death, she was at a luau party in Hawaii and suddenly felt “crumbling tears” from the strong emotional response. She elaborated that the, “[luau] has nothing to do with my husband except I’d been to one with him before”.

In addition to unpredictable triggers, participants described a volatile bereavement trajectory, fluctuating between extremes of functional capacity. A woman whose husband died explained how, “there’re days when I’m completely normal and then there’re days that you feel kind of like you can’t function”. This contributed to difficulty in coping with the new reality for the bereaved, who were often taking on new responsibilities in response to the death. Some commented that fear of severe emotional responses hindered their ability to move on.

For many, their loved one’s death coupled with limited social interaction led them to feel an intense loneliness. One participant who lost his wife said, “I think the worst thing is the loneliness. It’s so fierce. And that we had plans that we didn’t follow through”. As this interviewee described, the combination of social isolation and regret of unrealized plans further intensified the feelings of loss.

Coping strategies

Most participants reported efforts to distract themselves from their grief. This included returning to work or taking on a greater workload. A physician participant who lost his wife acknowledged that he was “deliberately driving the thought [of his wife] away” through going back to work. As a form of coping, some bereaved FMs would find ways to maintain a connection with the deceased. For many, the deceased remained a constant presence in their lives. Sometimes, this would take the form of an everyday conversation with the deceased or in events that took place after the death.

In some cases, participants found themselves providing support to others, such as family and friends affected by loss. This was sometimes therapeutic, as the bereaved felt good about helping those who were struggling with grief. In other cases, participants felt that it was challenging to support others, particularly when being called on to support friends or FMs suffering from the same loss or having to play a social role once belonging to the deceased. Methods of coping with the loss varied across participants, and often incorporated some aspects of the deceased’s life. Most found it helpful to read about or discuss their experiences with grief.

Sources of support

Bereaved FMs identified friends and other FMs as their primary sources of support, yet reactions to support varied. Some bereaved FMs likened the isolation that accompanied the bereavement experience to keeping oneself in a protective “cocoon” to shield themselves from the grief. These FMs often had relatives or friends who would check in on them and ensure they were receiving support. Participants typically appreciated friends and family being able to help. Friends who were not bereaved were considered less helpful; participants felt that such friends did not truly understand their experience. One participant summarized this sentiment as follows:

“when you lose everything you’re back to your childhood state and you’re dependent on other people for your physical needs. And so that’s why those are my friends because they went through the same experience as me. And they would be the most empathetic to my needs, not my work colleagues that absolutely don’t understand.”

At the same time, participants sometimes felt burdened by receiving support from friends and other FMs who were also grieving from the loss, since it had the capacity to re-trigger their grief.

Some participants sought support from a formal bereavement counsellor or support group; these were identified as positive experiences. These sources, unlike the inconsistent support of friends and FMs, were reported as particularly helpful in coping with symptoms of grief and in understanding bereavement.

Perspectives on future ICU bereavement screening and support programs

Participants’ narratives revealed commonalities specific to their observations of a program to screen and support bereaved FMs. Even though only half of participants with CG had received formal support, the majority could see positive aspects of such a program. Several expressed disbelief that support was not already being routinely offered. One participant explained, “I was actually quite surprised that there wasn’t a bit more follow-up other than getting a letter”, stating the “[hospital would] have a memorial service once a month that I could attend at any time. It was very impersonal”. Such examples underscore the lack of recognition of the emotional and social needs of bereaved FMs.

Others stated that routine follow-up would not be necessary and suggested that support was only needed if requested by the bereaved or if their symptoms worsened. Many participants had good experience with support groups, and this was suggested as an important component of any future support system. One participant compared her support group to a “lifeline” whose members were “like-minded and it’s just, good to talk to people that aren’t in my situation”, but who can connect as they are all going through their own challenging experiences.

Other suggestions included a brochure to explain the process of grief, manage expectations, and provide support resources to bereaved individuals. Some requested a meeting with a physician to review events or discuss the results of the autopsy.

There was broad agreement that everyone experiences bereavement differently, and participants thought it would be important to have a personalized approach to bereavement support. Several received standard letters from the hospital after the death, and these were “nice” but not valued. In contrast, one participant greatly appreciated a call she received a week after the death from an ICU nurse. Furthermore, several participants commented that the bereaved should not be approached too soon after the death.

Discussion

We report these qualitative study results by examining the bereavement experiences of FMs who developed CG after a loss in the ICU. We learned about their experiences before and after the loss, as well as their coping strategies, sources of support, and suggestions for a future program of screening and support.

Participants often mentioned frustration with aspects of care or communication prior to their loved one’s death in the ICU. Previous studies have shown that CG is associated with a perception of poor communication.8,15 Curtis et al. highlighted the importance of choosing words carefully when communicating with FMs in the ICU,18 and showed that dedicated communication “facilitators” may be effective for reducing depression among the bereaved.19 Family members frequently report poor end-of-life care in the acute setting.20,21 Nevertheless, our results in both this study and a previous study9 show high rates of satisfaction with care overall, suggesting that a global assessment of quality of care may mask specific events that influence bereavement. Notably, we found that participants who did not want support tended to have had difficult interactions with others (FMs, staff, etc.) during hospital care and bereavement. These difficulties with interactions may be an important predictor of CG meriting further exploration. Nevertheless, we cannot determine if this reaction is an immutable barrier to not accepting support, or something that could be addressed by leveraging existing supportive relationships with other staff members in hospital, other FMs, or friends throughout and beyond the patient’s illness.22

Continuity of care is an ongoing challenge in the ICU with frequent handovers between nurses and physicians. This can lead to inconsistency and difficulty creating stable relationships with FMs. As in previous research,9 we found that many participants were eager to meet with the medical team in follow-up to review events around the death. This contrasts with the findings of van der Klink et al.,12 who reported a generally good understanding among FMs of the events leading to death, and that only 35% of participants wanted follow-up with the medical team. One study of physicians and nurses reported their comfort providing medical information after a death,9 which might address FM’s reflection about aspects of care.

Most ICUs do not offer any bereavement follow-up.23 Our participants provided helpful suggestions for designing a screening and support program. Morris and Block reported similar findings when developing a bereavement service at a cancer centre.24 They highlight the favourable impact of the cancer centre acknowledging the death with a condolence letter, 24 although our results suggest that this needs to be personalized to be of benefit.15 A recent mixed-methods study of an ICU-based support program for dying patients and their families found that personalized sympathy cards were valued by FMs.25 Nevertheless, a randomized-controlled trial by Kentish-Barnes et al. showed higher rates of long-term depression among FMs of ICU decedents who received an individualized semi-structured condolence letter.26 The authors of that study, and the participants in our study, have emphasized the importance of timing; a bereavement intervention delivered too early could cause distress. Other studies have shown harm from interventions intended to be supportive that were delivered to FMs of seriously ill patients27,28 and to people exposed to a traumatic event.29 Complicated grief interventions are ineffective for bereaved FMs who do not have CG.14 The integration of electronic medical information systems provides hospitals with the opportunity to improve their outreach efforts and avoid sending inappropriate fundraising letters to deceased patients that can cause harm to FMs. The timing, content, and selection of participants must be carefully considered in designing support programs.

Many participants highlighted the perceived benefits of participating in the interviews. This was previously described in Holocaust survivors,30 and Germain et al. reported similar results in a study of bereaved FMs of patients who died at home,31 identifying key adaptive themes such as the final hours of resilience. Barnato et al.32 piloted a narrative intervention delivered by trained social workers to 32 bereaved FMs approximately one month after their loved one’s death. This intervention was like the narrative interview approach taken in this study and was perceived as beneficial by participants.

Many participants suggested that informational resources (e.g., pamphlet) might help manage expectations, reassure those experiencing severe emotional responses to “triggers”, or suggest coping skills such as keeping busy or talking with other bereaved individuals. Morris and Block described similar components in their bereavement service,24 but a large survey identified that only 39/237 (16%) of ICUs provide such a service.23 We should note that our participants were all experiencing severe grief reactions, so we cannot know whether their coping strategies were effective means of overcoming CG, or maladaptive strategies that exacerbated their symptoms. Nevertheless, ICUs interested in establishing a bereavement support group may benefit from co-designing such an intervention with bereaved FMs.

Strengths of our study were that the study population was drawn from a diverse representation of FMs, as well as the narrative interview approach, which generated rich descriptions. There are also limitations to our study. The study population was selected through purposeful sampling, which is inevitable because of the emotional and logistical challenges associated with recruiting bereaved FMs. The eight participants in this analysis provided varying accounts of their grief, as expected because of the profoundly individual bereavement experience. While this study design limits our ability to draw broader conclusions about how ICU events impact bereavement and the effectiveness of specific coping and support strategies, we plan to develop a bereavement program and then study its feasibility and effectiveness in future research. Future research projects on this topic may benefit from including FMs with bereavement experience as partners to co-design the intervention.

Conclusions

Family members experiencing CG describe many unique experiences and sources of support yet identify several commonalities. The perceptions that FMs offered about supporting FMs at risk of CG provide insight into elements of bereavement support that may be well received following death of a loved one. This research is an important first step in developing initiatives to screen for severe grief reactions and support FMs with CG following a death in the ICU. Future research will need to expand the scope of how bereavement support is provided to FMs and identify which interventions are both feasible in ICUs and effective in reducing CG among bereaved FMs.