Emanuel EJ, Fairclough DL, Wolfe P, Emanuel LL. Talking With Terminally Ill Patients and Their Caregivers About Death, Dying, and BereavementIs It Stressful? Is It Helpful?. Arch Intern Med. 2004;164(18):1999–2004. doi:10.1001/archinte.164.18.1999
Discussing end-of-life issues with terminally ill patients is often considered distressing and harmful. This study was conducted to assess whether interviewing terminally ill patients and their caregivers about death, dying, and bereavement is stressful and/or helpful.
Patients from 6 sites in the United States who were estimated to have 6 months or less to live were interviewed in person and reinterviewed 2 to 6 months later. Their caregivers were interviewed separately. At the end of the interviews, patients and caregivers were asked how stressful and how helpful the interview had been. Of 1131 eligible patients, 988 (87.4%) were interviewed, and of 915 eligible caregivers, 893 (97.6%) were interviewed.
At the end of the first interview, 1.9% of the patients reported having experienced a great deal of stress, 7.1% some stress, and 88.7% little or no stress from the interview. Among the caregivers, 1.5% reported a great deal of stress, 8.4% some stress, and 89.7% little or no stress. Slightly more stress was reported to have been caused by the reinterview. Overall, 16.9% of the patients reported the initial interview as very helpful, 29.6% as somewhat helpful, and 49.6% as offering little or no help. Among the caregivers, 19.1% reported the initial interview as very helpful, 34.3% as somewhat helpful, and 44.9% as offering little or no help. The reported helpfulness of the second interview was slightly less. Patients experiencing pain (odds ratio [OR], 1.26; 95% confidence interval [CI], 1.02-1.56), more personal meaning in dying (OR, 3.05; 95% CI, 2.02-4.59), and less ease with talking about the end of life (OR, 1.32; 95% CI, 1.09-1.60) were significantly more likely to report stress. Patients who were from an ethnic minority (OR, 1.85; 95% CI, 1.31-2.63), anxious about the end of their life (OR, 1.39; 95% CI 1.16-1.67), more spiritual (OR, 1.30; 95% CI, 1.06-1.61), and serene (OR, 1.25; 95% CI, 1.08-1.45) were significantly more likely to report the interview helpful. There was no relationship between stress and helpfulness.
Terminally ill patients and their caregivers can discuss death, dying, and bereavement in a structured interview with minimal stress and report that the interview was helpful. Institutional review boards should not preemptively restrict surveys with terminally ill patients without reliable evidence that they will be stressful or otherwise harmful.
The need to attend to dying patients and provide them with better quality of care has become a national priority.1 One important aspect of providing better care is better communication,2,3 but this is not easy to achieve. Many people believe that speaking with patients about death, dying, and their terminal condition is likely to cause stress to both patients and their families, and this belief often leaves health care providers, relatives, and friends of dying patients asking, "What should I say?"
Another important aspect of providing better care for dying patients and their caregivers is understanding their experiences, symptoms, and concerns. It requires research into what patients and their caregivers perceive to be the greatest problems with the dying process, the care they receive from the health system, and other issues. Yet many argue that terminally ill patients are "vulnerable" and need protection from enrollment in research studies.1,4,5 Institutional review boards (IRBs) have expressed concern about permitting such studies, especially surveys, with dying patients and their caregivers for fear that it would be intrusive, alarming, and time-consuming, and that it may increase stress and suffering when dying patients have little time left.1 Other IRBs have restricted or prevented surveys of dying patients and bereaved family members because they perceive such research to be of no benefit and potentially burdensome.
However, there are few published data on how dying patients and their caregivers experience participation in research. The Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment noted that, in a pretest of survey instruments, nurses who rated emotional change in 18 patients reported that most experienced no change (1 experienced negative and 1 experienced positive change).6 Because we have frequently been challenged about the potential inconvenience and stress of our research on death and dying, we sought to determine systematically whether speaking with dying patients and their caregivers about issues related to death and dying was stressful and whether it was helpful from the patients' and caregivers' own perspectives.
This prospective cohort study surveyed and followed up patients designated as terminally ill by both their physicians and their caregivers in 5 randomly selected statistical metropolitan areas and 1 rural county: Worcester, Mass; St Louis, Mo; Tucson, Ariz; Birmingham, Ala; Brooklyn, NY; and Mesa County, Colorado. Physicians were asked to identify terminally patients, and the participating patients were then asked to identify their primary caregivers. The overall method of this study has been previously delineated in detail.7,8
In each of the 6 sites physicians were randomly selected from state boards of medical registration, state medical societies, and membership lists from the American Society of Clinical Oncology, American College of Cardiology, American Gastroenterological Association, and the American College of Chest Physicians. Selected physicians were mailed a letter requesting their participation to identify patients who had "a significant illness and a survival time of 6 months or less, in your opinion." A total of 383 physicians referred patients.
No patient or caregiver was paid for participation. Patients identified by physicians were mailed an explanation of the study with a postage-paid opt-out card. The letter indicated that the purpose of the study was to understand
"the attitudes of patients with a significant illness and their caregivers towards the quality of the patients' health care [and their] perspective on [their] illness experience." If the opt-out card was not returned, the patient was contacted. Patients were eligible to participate if they (1) had any significant illness (excluding human immunodeficiency virus or AIDS), with a survival time of less than 6 months as determined by their physician, (2) spoke English, (3) had no hearing difficulty, and (4) were able to arrange an interview time and place and sign a consent form. Of the 1472 patients referred by the physicians, 194 had died and 116 had become mentally incompetent, and 31 could not speak English or had hearing limitations; and of the 1131 eligible patients, 119 refused to participate and 24 could not be located. A total of 988 patients were interviewed (response rate, 87.4%).
Patients were asked to identify their primary caregiver as the family member, friend, or other person who provided most of their assistance. Caregivers were ineligible if they (1) spoke no English, (2) had hearing limitations, or (3) were not able to schedule an interview and sign a consent form. Of the 988 patients who responded, 70 reported not having a caregiver and 3 reported that their caregiver did not speak English. Of the 915 eligible caregivers, 22 refused to participate. Overall, 893 caregivers were interviewed (response rate, 97.6%).
Two to 6 months after the initial interview (mean, 125 days), patients were reinterviewed, or the caregivers of those who had died were interviewed. Of the 988 patients, 699 were still alive, but 17 were mentally incompetent and 32 could not be located or refused the second interview. A total of 650 patients were reinterviewed (response rate, 95.3%). Of the 289 patients who had died, 3 had not identified a caregiver, 3 caregivers were too ill to be interviewed, and 27 caregivers could not be located or refused the second interview. Overall, 256 caregivers were reinterview (response rate, 90.5%).
Twenty-four interviewers from the National Opinion Research Center conducted all interviews in person between March 1996 and July 1997, at locations determined by the patients.
The patient survey contained 133 questions and the caregiver survey 116 questions examining (1) symptoms; (2) social supports; (3) communication with health care providers; (4) spiritual and personal meaning of dying; (5) care needs; (6) end-of-life plans; (7) economic burdens; (8) sociodemographic characteristics; and (9) preferences regarding end-of-life care, and opinion regarding euthanasia and physician-assisted suicide. The patients' survey included questions such as "Do you talk freely about the end of your life?"
"Do you feel prevented from doing what is necessary because your family or close friends cannot accept that your life may end?" "Do you feel relieved that life, and all its suffering, will end in the foreseeable future?" "When you imagine the end of your life, what things are you afraid of and hope to avoid?" "How much of an economic or financial hardship is the cost of your illness and medical care for you or your family?" Caregivers were also asked questions such as "Does caring for [the patient's name] interfere with your family or personal life?" "Do you feel alone and abandoned?" "When you imagine the end of [the patient's name] life, are there things you are afraid of and hope can be avoided?" "Do you want to be there when [the patient's name] dies?" Bereaved caregivers were asked about the circumstances of the patient's death.
At the end of the survey, respondents were asked the following questions:
"How much stress would you say this interview has caused you?" (answers were on a 4-point scale from a great deal of stress to no stress at all) and "How helpful would you say this interview was for you?" (answers were on a 4-point scale from very helpful to not helpful at all).
The protocol, letters, survey instruments, and consent documents were approved by the Harvard Medical School and Dana-Farber Cancer Institute's IRBs as well as the IRBs of 38 medical institutions in the 6 geographic sites.
In the consent form, participants were informed of the following: "You will receive no direct benefits from participating in this study. However, you may find that by answering these question, you may gain insight into your values, opinions, and relationships." Regarding risks, they were informed that the issues asked about "are personal and may be sensitive. It is possible that answering questions of this type may cause some stress."
Analysis of variance for age was used to compare the characteristics of the patients who were reinterviewed and those who had died but whose caregivers consented to be reinterviewed for them; the Cochran Mantel-Haenszel χ2 test for ordered categorical variables was used for analysis of education and income; and the χ2 test of independence was used for unordered categorical variables when analyzing sex, race, religious affiliation, marital status, and sort of disease. Potential covariates for the multivariate regression models were selected from bivariate analyses where tests of association were significant at α = .05. Stepwise logistic regression was used to identify the covariates that explained the greatest variation in the following outcomes: the survey was "stressful" and the survey was "helpful." To check specific covariates, such as pain, not identified in the stepwise procedure, a second logistic model was fit to force the covariate of interest into the model.
Table 1 provides the sociodemographic characteristics of the terminally ill patients and their caregivers enrolled in the study. Initially, 50.2% of the patients experienced moderate or severe pain; 17.5% were bedridden more than 12 hours per day; 70.9% experienced shortness of breath while walking 1 block or less; 35.5% had urinary or fecal incontinence; and 16.8% had depressive symptoms. In the 6 months before the interview, 66.5% of the patients had been hospitalized, 36.8% had undergone surgery, and 22.3% had been admitted to the intensive care unit.
Less than 2% of the terminally ill patients reported that completing the initial survey on death and dying caused them a great deal of stress, 7.1% reported some stress, and 88.7% reported little or no stress (Table 2). A slightly larger proportion of patients found the follow-up survey stressful (Table 2). The patients' ratings of stress at the time of the initial interview did not affect their participation in the follow-up interview (P = .2).
Similarly, among the caregivers, few reported a great deal or some stress during the initial interview (9.9%) (Table 2). Among those who participated in the postbereavement interview, substantially more caregivers reported stress. Yet, only 9.4% reported a great deal of stress and 16.4% reported some stress while 73.4% reported little or no stress (Table 2). The response rate in the postbereavement survey was 90.5%.
In multivariate logistic regression, terminally ill patients experiencing pain (odds ratio [OR], 1.26; 95% confidence interval [CI], 1.02-1.56), finding less personal meaning at the end of their life (OR, 3.05; 95% CI, 2.02-4.59), and being less at ease speaking about the end of life (OR, 1.32; 95% CI, 1.09-1.60) were significantly more likely to report stress from the interview (Table 3). Caregivers whose patient was experiencing psychological distress and depression (OR, 1.83; 95% CI, 1.49-2.24), and who perceived greater economic burden from the patient's illness (OR, 1.25; 95% CI, 1.01-1.53), were significantly more likely to report stress (Table 3). Jewish caregivers reported more stress than caregivers from any other religion (OR, 2.75; 95% CI, 1.10-6.85).
Overall, 16.9% of terminally ill patients found talking about death and dying in the interview very helpful, 29.6% found it somewhat helpful, and 49.6% reported little or no help (Table 2). A slightly lower percentage found the follow-up survey helpful. Slightly more caregivers than terminally ill patients found completing the survey helpful. A total of 19.1% of caregivers found talking about death and dying in the survey very helpful, 34.3% somewhat helpful, and 44.9% of little or no help (Table 2). Among bereaved caregivers, 16.8% reported that completing the survey was very helpful, 24.6% somewhat helpful, and 56.3% of little or no help (Table 2).
Multivariate logistic regression showed that patients from ethnic minorities (OR 1.71; 95% CI, 1.22-2.40) were more frightened and anxious about the end of their life (OR, 1.26; 95% CI, 1.06-1.50), and that those who were less religious or spiritual (OR, 1.30; 95% CI, 1.06-1.61) and more serene (OR, 1.25; 95% CI, 1.08-1.45) were significantly more likely to report the interview helpful (Table 4). Multivariate logistic regression also showed that caregivers who spoke less with the patient about spiritual matters and relationships (OR, 0.58; 95% CI, 0.41-0.82), and who found that health care providers did not provide clear information (OR, 0.85; 95% CI, 0.72-1.00), were significantly less likely to report the interview as helpful. However, caregivers who felt a burden because of the patient's fatigue (OR, 1.33; 95% CI, 1.15-1.54) were more likely to report the interview helpful (Table 4). Also, Protestant caregivers were more likely to find the interview helpful than caregivers from any other religion (OR, 1.68; 95% CI, 1.21-2.33).
Finally, there was no relationship between stress and helpfulness. Among patients who did not find the interview stressful, 46.6% found it helpful while 44.9% of those who found the interview stressful also found it helpful (P = .76). Similarly, among caregivers who did not find the interview stressful, 52.9% found it helpful, while among those who did find it stressful 58.0% found it helpful (P = .37).
These data suggest that discussing topics related to death and dying during an interview is not particularly stressful, and may actually be helpful, for many dying patients—including those who found the interview stressful, and even their bereaved caregivers.
In this study, interviews of terminally ill patients and their caregivers were not particularly stressful despite questions about symptoms, care needs and economic burdens, the personal meaning of and adjustment to impending death, plans for terminal care, social supports, communication with health care providers, and related topics. Less than 5% of patients and caregivers reported a great deal of stress from the initial interviews; and among bereaved caregivers interviewed shortly after the death of a loved one, less than 10% reported a great deal of stress. Indeed, more than 70% of respondents, including bereaved caregivers, reported little or no stress from the interview. This low level of stress does not appear to result from the number of terminally patients or their caregivers who refused to participate in the interview because they might find it stressful; the response rate for patients was greater than 87%, ie, greater than response rates in many public opinion polls and other surveys of healthy individuals on uncontroversial topics. These findings suggest that the common fears of researchers, IRB members, and the public that talking about death and dying with terminally ill patients will be stressful are exaggerated.
While few terminally ill patients and their caregivers reported stress from the survey, more than 40% of them found answering questions about their symptoms, care needs, and relationship with health care providers somewhat or very helpful. This relatively high percentage of positive answers occurred despite the warning that the information would not be shared with their health care providers and that they would "receive no direct benefit" although they
"may gain insight into [their] values, opinions, and relationships." Merely having a chance to answer an organized set of questions about their experience to an interviewer was judged helpful to patients and their caregivers. We do not know in what way patients and caregivers found it helpful, but it may be that answering structured questions helped them better understand their experiences, or that, for many people, having someone, even unknown, be interested in hearing about their personal experiences at this sensitive moment is helpful. Determining the precise nature of the benefit to patients and caregivers of speaking about death and dying experiences requires further research.
We had hoped that the structured interview might prove to be particularly helpful to patients and caregivers who found communication about death and dying difficult. While this appeared true for some patients, it was not true for caregivers as a group. More than 50% of terminally ill patients who reported fear and anxiety, and/or a feeling of being prevented from coming to closure by family and friends who could not accept their death found the interview helpful. These interviews also seemed particularly helpful to patients from ethnic minorities and those who were anxious. Our findings suggest that for some patients who otherwise might be at higher risk of having problematic dying experiences, a structured interview might be particularly helpful. Conversely, the interview did not seem more helpful to caregivers who reported problems in communicating with the patients or health care providers.
This study suggests 2 important implications regarding policy. First, IRBs should not restrict interview research of terminally ill patients and their caregivers unless there are reliable data to indicate that the survey will be stressful. Our findings suggest that survey research with terminally ill patients and their caregivers is rarely stressful and is frequently beneficial. About 40% of patients perceived some helpfulness from the survey compared with less than 10% who experienced stress; even among bereaved caregivers the stress-helpfulness ratio indicates that interviews are helpful. Indeed, our findings suggest that it may be appropriate to include in informed consent documents a statement that approximately 40% of terminally ill respondents find participating in similar surveys helpful and less than 5% find it stressful.
Second, our findings on helpfulness suggest that implementing structured interviews as part of the routine care of terminally ill patients and their caregivers may be important. Not only might such standard interviews obtain critical information about problems such as uncontrolled symptoms or care needs, but merely having the discussion and expressing interest may be therapeutic. And while we do not have detailed cost-effectiveness data, such structured interviews can be comparatively inexpensive, especially when the additional costs associated with formal research data collection are excluded.
This study has important limitations. First, patients fearing stress may have refused to participate; and while identifying patients as terminally ill, physicians may have excluded those for whom such a survey would be particularly stressful. While this is possible, most data suggest that physicians are inaccurate at predicting the psychological state or experiences of their terminally ill patients9- 12; and there is no reason to think that they are good at predicting which patients will have stress in response to a survey. Second, these data are from in-person interviews and may not be generalizable to other interview formats such a telephone or e-mail surveys.
Intuitively, many people think that speaking with terminally ill patients and recently bereaved caregivers about death, dying, and bereavement is stressful. The findings from this study, however, suggest that structured interviews on these topics are rarely stressful and frequently are helpful.
Correspondence: Ezekiel J. Emanuel, MD, PhD, Department of Clinical Bioethics, National Institutes of Health, Warren G. Magnuson Clinical Center, Bldg 10, Room 1C118, Bethesda, MD 20892-1156 (email@example.com).
Accepted for publication December 16, 2003.
This study was supported by grants from the Commonwealth Fund and the Nathan Cummings Foundation.
The views expressed in this study are those of the authors and do not necessarily reflect the opinions or policies of the National Institutes of Health or the Department of Health and Human Services.