The call to shift the site of care for dying patients has occurred largely in the absence of information on patients' preferences. It has been acknowledged that home care is difficult for patients living alone or those with family members working outside the home (6). The few studies of preferences for site of terminal care have involved patients selected according to receipt of specialized services (7-9) or diagnosis of cancer (10).

Preferences for site of terminal care among older adults are particularly important because most deaths occur among adults 65 to 89 years of age (1). Furthermore, the extensive utilization of Medicare home health care means that many older adults have experienced care in their homes. We studied older adults recently hospitalized with congestive heart failure, chronic obstructive pulmonary disease, and pneumonia to describe preferences for site of death among a group of older persons with illnesses potentially amenable to both home and hospital treatment.

Methods

Top Methods Results Discussion Author & Article Info References

The study consisted of two phases: 1) quantitative interviews to establish the frequency of preferences for site of terminal care and 2) qualitative interviews with a separate group of patients to explore reasons for preference. This work was part of a larger study conducted to determine preference for home or hospital among patients with nonterminal illness; the qualitative results of that study have been published elsewhere (11). The Human Investigations Committee of Yale University School of Medicine approved the study.

Phase 1: Quantitative Interviews

Participants were consecutive community-dwelling persons 65 years of age or older who were hospitalized with a primary diagnosis of congestive heart failure, chronic obstructive pulmonary disease, or pneumonia in two urban teaching hospitals. They were interviewed 2 months after hospitalization. Patients were excluded if their physician did not provide consent or if they had difficulty hearing, did not speak English, had moderate to severe dementia, did not have a telephone, lived out of state, were discharged to another facility, were in other longitudinal studies, or were hospitalized at the time of the interview.

Of 962 persons screened for participation, 1438 did not meet eligibility criteria and 109 died before the interview. Of 415 eligible participants, 15 could not be reached and 154 declined to participate. This resulted in 246 participants and a response rate of 60%. Nonparticipants did not differ from participants by sex, ethnicity, or diagnosis of congestive heart failure or pneumonia but were more likely than participants to be older (P < 0.001) and were less likely to have a diagnosis of chronic obstructive pulmonary disease (P = 0.023).

In a telephone interview, participants were asked, using the questions shown in Table 1, to choose home or hospital as their preferred treatment site.

Phase 2: Qualitative Interviews

Participants were persons 65 years of age or older who had been hospitalized in the past 6 months with congestive heart failure, chronic obstructive pulmonary disease, or pneumonia and were receiving home care services through a single home care agency. The sample size was determined according to the concept of theoretical saturation (12). Interviews continued until additional participants did not introduce new concepts. Saturation was reached after 29 interviews. Thirty-four patients were approached for participation, of whom 5 (15%) declined.

Open-ended questions were used to elicit preferences for site of care by asking participants to think about their recent illness and then about an illness from which they were not likely to recover. They were asked to explain the reasons underlying their preferences. All interviews were conducted in participants' homes by a single investigator and lasted between 30 and 60 minutes. Interviews were audiotaped and transcribed. Two investigators analyzed the transcripts by multiple close readings according to the constant comparative method of Glaser and Strauss (12).

Results

Top Methods Results Discussion Author & Article Info References

Phase 1

Characteristics of the 246 participants are shown in Table 2. When recovery from the illness was unlikely, 106 (43%) participants preferred treatment at home, 118 (48%) preferred treatment in the hospital, and 22 (9%) did not know. When recovery from the illness was likely, 113 (46%) participants preferred treatment at home, 132 (54%) preferred treatment in the hospital, and 1 did not know. Preference was not associated with age, sex, ethnicity, education, diagnosis, or functional status. One third of participants changed their preference according to whether recovery from the illness was likely. Of the 113 who preferred home treatment when recovery was likely, 60 (53%) also preferred home when recovery was unlikely and 36 (32%) instead preferred the hospital. Of the 132 who preferred hospital treatment when recovery was likely, 82 (62%) also preferred the hospital when recovery was unlikely and 46 (35%) instead preferred home.

Phase 2

The 29 qualitative interviews were conducted with persons ranging in age from 65 to 89 years (mean ± SD, 77 ± 7 years), of whom 21 (72%) were female, 18 (62%) were white, and 11 (38%) were African-American. Seventeen (59%) lived alone, 6 (20%) lived with children, 4 (14%) lived with a spouse, and 2 (7%) lived with other relatives. In the case of terminal illness, 7 participants preferred home, 7 preferred the hospital, 5 preferred a nursing home, and 1 preferred hospice; 9 would not talk about their preference. Of the 14 who preferred home or hospital in the case of terminal illness, 6 had a different preference in the case of nonterminal illness.

The 3 participants who preferred home in the case of terminal illness but otherwise preferred the hospital spoke of the importance of having their family around them. One woman stated,

I prefer dying at home. I would at least have my children around me, my relatives, you know.

On the other hand, the 3 participants who preferred the hospital in the case of terminal illness but otherwise preferred home had concerns about their families' ability to care for them if they were dying. They believed that they would require the care of physicians and nurses in the hospital. One woman said,

When you're that bad, go [to the hospital]. This way you're right there. If you need them [the nurses], they're there. That's why they're supposed to be called the angels of mercy.

The participants who preferred the nursing home had similar concerns about their families' ability to care for them. However, rather than thinking about an acute episode of care, they were concerned about burdening family members with their long-term care needs. One woman stated,

They usually put you in a convalescent home if you are so sick there is no help for you  ... . You wouldn't want to be a burden on your family. I mean down to the point where there is no help for you, I don't want to be a burden on my family.

Many participants did not want to provide a response. One woman, who had metastatic breast cancer in addition to congestive heart failure, replied to the question of whether she had considered her preferred site of death:

No, because I don't think that it's going to come to that. If it does, I'll face it when it does, you know?

Discussion

Top Methods Results Discussion Author & Article Info References

In contrast to the 17% of adults 65 years of age or older who died in their homes in 1990 (1), 43% of participants in our study expressed a preference for receiving terminal care at home. For one third of participants, preference for treatment site changed according to whether they were likely to recover from their illness. This finding suggests that many older persons have unique concerns about and preferences for terminal care. The reasons for these preferences, although based on the responses of a small number of patients, suggest that the current debate about home and hospital as the ideal site for terminal care may not capture these concerns.

No participants expressed concern about the quality of terminal care in the hospital. Many participants worried about the additional help they would need during a terminal illness (even if they could rely on family members during a nonterminal illness), and they felt that they would receive this help in the hospital.

When the answer to the question of preferred site of terminal care was not constrained to "home" and "hospital," "nursing home" became a frequent response. For participants choosing a nursing home, terminal illness was conceptualized as the exacerbation of a chronic disease, and their preference reflected concerns about how they would receive the care required by chronic illnesses.

The many participants who did not want to think about their preferred site of terminal care contrasts with the infrequent discomfort reported with advance directives (13). Perhaps discussions about site of terminal care feel less hypothetical than discussions about other treatment decisions and are therefore more difficult for many older persons.

Patients were interviewed alone; preferences may change as a result of family discussion. Nonetheless, the variability in preference for the site of end-of-life care supports a system of care equipped to treat older patients across a spectrum of treatment sites.