The Preferred Place of Last Days: Results of a Representative Population-Based Public Survey

Introduction

The location or place of death is of considerable interest worldwide. ¹ Not only does the place of death or end-of-life (EOL) care impact the quality of death and dying, ² but there are many direct and indirect costs and other considerations associated with the place of death and dying. ³ Although much has been made of the high cost of dying in hospitals and nursing homes, ⁴ home-based dying often involves significant out-of-pocket costs in addition to difficult tasks and responsibilities for family caregivers.^5,6 Most EOL care in the home is provided by family caregivers, often with little assistance from others, including formal home care providers.^5,7,8 Regardless, a shift from hospital to home-based death and dying is occurring in Canada and elsewhere, notably Britain and the United States.^9–11 Our concern is that until the place of EOL care becomes the focus of attention over place of death, major EOL issues such as inadequately supported home-based dying will not be addressed.

This concern is particularly relevant now that studies are finding terminally ill people and their family members often state a preference for home deaths. ² It is also evident that some home deaths are not possible because of factors such as burdens associated with caring for dying loved ones at home. ¹¹ Regardless, having a place preference is often possible now, as most deaths occur in adulthood and following evident declining health. ⁹ With most deaths “expected”^12,16 now, this potential to anticipate death provides more opportunities to plan ahead and thus enhance the possibility of achieving the preferred place of final days. From a public health perspective, it is critically important to take public preferences into account when developing EOL care strategies.

Results

The overall response rate was 21.4% (N=1203). As shown in Tables 1a, 1b, and 2, 70.8% stated a home preference, while 14.7% preferred a hospice/palliative care facility, 7.0% a hospital, 1.7% a nursing home, and 5.7% declined to answer or did not know/were unsure. The home was more often preferred by men, younger people, those with only high school completion, those not widowed, and those not previously involved in a no-treatment or withdrawal of treatment decision. A hospice/palliative care facility was more often preferred by women, older people, those with postsecondary education, widowed persons, and those with previous involvement in a no-treatment or treatment withdrawal decision. Hospitals were preferred by the oldest people, people with education below high school completion, and widowed persons. Ethnic/cultural group, religiosity, household income, number of people in the household, urbanization level, and all other variables were not significantly related to place preference.

Further analysis revealed marital status was the only place predictor (see Table 3). People who were widowed were more likely than married persons to prefer a hospital or hospice/palliative care facility rather than home. After adjusting for marital status, none of the studied life experiences or personal viewpoints were associated with any place. The logistic regression analysis (with model fitting information: Nagelkerke R²=0.086) revealed people who were widowed were more likely than married persons to prefer a hospice/palliative care facility (adjusted OR=3.09, 95% CI=1.44–6.63) and more likely to prefer a hospital (adjusted OR=4.72, 95% CI=1.74–12.80) over home. The odds ratio was adjusted for gender, age, ethnic/cultural background, religiosity, educational level, marital status, number of children in the household, having been involved in the care of a dying person, and involvement in a no-treatment decision.

Discussion

This study revealed that the majority (70.8%) of surveyed Albertans preferred to spend their last days at home, as compared to a hospice/palliative care facility (14.7%), hospital (7%), or nursing home (1.7%). Only 5.7% had no preference or were unwilling or unable to indicate one. These findings show most surveyed residents of Alberta, the province with the youngest median age and smallest proportion of citizens over 65 were not only willing to answer a question about their preferred place of death but also had a preference. ²⁵ These findings suggest death and dying understandings that shape place preferences are gained through a wide range of everyday events or processes. Younger surveyed adults had place preference, which could be based on childhood influences.

The home preference finding was striking, as hospitals are readily available as a consequence of Canada's universal, publicly funded health care system. Hospitals, as well as nursing homes and hospices/palliative care facilities, are sites where health care professionals work and technologies are readily available to prevent or address unpleasant symptoms. Furthermore, some studies have indicated death anxiety is common, ²⁶ and that death anxiety can reduce the willingness of people to think about and plan for their own death and dying. ²⁷ It is possible that the public as a whole, or at least the persons surveyed in Alberta, are more aware of and comfortable with the concepts of death and dying now, and are more open to talking about and planning for this eventuality. An earlier study also revealed dying people have clear and articulate views about dying. ²⁸

Although Fukui and colleagues' study showing influences on place preference ¹⁸ suggests that experiences related to death and dying, EOL preparations, and personal viewpoints on death hastening would impact place preferences, widowhood emerged in the current study as the only predictor. Widows were more likely to state a preference for a hospital or hospice/palliative care facility. Widowhood could simply limit place options, as many studies have established that the spouse is the primary caregiver for ill and dying persons.^5,11 People who have experienced the death of their spouse have gained insights into the practical, emotional, and other difficulties with death and dying. Although the care needs of dying people vary and they may need help for only a short time,^29–31 it is likely that widowed people have learned that dying persons need assistance with many matters, including first instrumental and then basic activities of daily living.^30,32 Given limitations in publicly funded home care,^5,7,8 home-based dying may be an impossible choice for widowed persons. Clearly, more research on EOL care needs and knowledge dissemination are needed to help the public plan ahead for the end of life, and help more people achieve their preferred EOL place.

Another finding of significance is that widowed persons rarely chose nursing homes, but instead identified hospitals and hospice/palliative care facilities as preferred places. The hospice/palliative facility choice demonstrates public awareness of these specialized centers. This suggests a major social learning achievement, as free-standing hospices are not yet common across Canada and palliative care units are typically only found in larger hospitals.

This disregard of nursing homes by widowed persons helps explain why an “aging in place” strategy for population aging is common across Canada, ³³ and why nursing home bed numbers have been stagnant despite population aging. ³⁴ Although it is understandable that no one would wish to be debilitated to the point of needing nursing home care, research should determine how nursing homes could become a preferred EOL place. Nursing homes have already emerged in some countries as places where high-quality EOL care is an aim. ¹ Achieving this aim is important, as nursing homes in many countries are already a common place of death.^1,34 Nursing home deaths are likely to increase with population aging and with more people reaching very old ages. ³⁴

Despite the need for qualitative and other research to establish why widowhood is a major place predictor and why most people chose the home over other places for their final days, the widespread home preference indicates public health interventions are needed to promote good home deaths. As palliative care is not yet fully integrated into the public health agenda, it is an area with great potential.

Research is also needed to address the additional challenges posed by advanced aging, as more people are living to be very old. ⁹ Many older persons outlive spouses, and as home-based dying typically requires one or more family member caregiver to be present and capable of providing care, issues associated with widowhood and living alone are barriers to home deaths. In addition, some older couples have no children, while others have children who are not physically or emotionally able to provide EOL care at home. Geographic proximity to informal caregivers is also essential, a factor explaining why some surveyed Albertans chose another home instead of their own.

Research is needed to better understand which if any personal experiences, viewpoints, or EOL preparations, and which (predominant or minority) social values such as autonomy impact not only place preferences but EOL preparations and location of death. An increasing death rate based on the aging population suggests that care provided at home to dying friends and family members needs to be better documented. This information will inform how the options for informal care and care outcomes influence place planning and future actions or outcomes. This research is overdue, as half of those surveyed consider they have already provided EOL care in one form or another. Many have been involved in decisions to stop or not start life-supporting treatment. Many also have developed a living will/personal directive in preparation for the end of life, and developed a viewpoint on legalized assisted suicide.

Conclusion

This population-based survey provides evidence for consideration across and outside of Canada, notably that most adults prefer to be at home in their final days of life. Although past death and dying experiences and death-hastening viewpoints were expected to influence place preferences, widowhood emerged as the only place predictor. Marital status should not be considered as a simple decision factor, however. Home-based EOL care may not be possible for them, as they will have learned that assistance from others is usually needed when dying. For this reason and others such as severe or prolonged illnesses, ³⁵ it is important to realize that preferences are heavily impacted by personal and environmental circumstances. The challenge now will be to determine ways and means to ensure more people can stay at home and are not driven by a lack of EOL care options into another place.