Preferred and Actual Location of Death: What Factors Enable a Preferred Home Death?

Introduction

Many populations are aging and people are living longer with advanced chronic disease. The majority of deaths are not sudden or unexpected. This longer time frame of advanced illness offers opportunity for patients, family, and health care providers to plan how and where to provide care for the dying. ¹

Most people prefer to die at home,^2,3,4–8 but the majority of deaths occur in an institutional setting. ⁹ Determinants of location of death, and home death in particular, have been described^4,10–13 and are suggested as an interplay between individual, illness-related, and environmental risk factors. ¹⁰ Congruence between preferred and actual location of death^3,14–16 is reported to vary widely depending on diagnosis and source of patient identification.^16,17 Most studies are limited to cancer patients^3,5,6,18 and/or patients in special clinical programs or institutions.^6,14 Few include people dying of multiple chronic diseases, in varied care settings or with knowledge of death location preferences.

In this paper we (1) report the congruence between the last preferred and actual location of death among Nova Scotians who died from chronic disease, and (2) identify individual, illness-related, and environmental factors associated with achieving a preferred home death.

Methods

Data are from a population-based, mortality follow-back survey examining the experience of end-of-life care among all adults who died nonsuddenly of advanced chronic disease in the Canadian province of Nova Scotia (population 950,000) between June 2009 and May 2011.^19–21 All physician, hospital, and some home care services are publicly financed. There are a limited number of specialized inpatient palliative care beds but no free-standing hospices. Participants were 1316 death certificate identified informants (next-of-kin) knowledgeable about the care provided to the decedent during their last month of life. Details pertaining to the main study are published elsewhere. ²¹

To assess congruence, data were limited to decedents whose informant had knowledge of their last preferred location of death. To explore factors predictive of achieving a preferred home death, data were further limited to decedents known to have expressed a preference for death at home and who had spent at least one day ‘at home’ during their last month of life. Long-term care (LTC) residents were excluded from the latter analysis for conceptual reasons. Ethical approval was provided by the Capital Health Research Ethics Board, Halifax, Nova Scotia.

Results

Overall, 56.8% of decedents died in hospital, 25.2% in LTC, 17.9% at home, and 0.2% in transit (which was excluded from further analyses). Almost half of the informants (46%) were knowledgeable about the decedents' last voiced preferred location of death.

Of the 606 decedents who voiced a preferred location of death, home was the most preferred (73.9%), followed by the hospital (15.9%) and LTC (10.3%). For these decedents, the actual location of death was 30.4% home, 51.5% hospital, and 17.9% LTC. Congruence between decedents' preferred and actual location of death was 51.9% with an overall kappa statistic of 0.29, indicating slight to fair congruence (see Table 1). Achieving a preferred hospital (90.6%) or LTC (83.9%) death was attainable for many; 39.2% of decedents preferring to die at home did so.

Of the 447 decedents who preferred to die at home, 142 were excluded (LTC residents, n = 62 and no time at home (n = 80). Of the remaining 305 decedents, 56.4% (n = 172) achieved their wish. A significantly greater proportion of decedents who achieved a desired home death were aware of their approaching death, had a signed living will or advance directive in place, died of cancer, and had their emotional support needs met (see Table 2). They were also more likely to have received nursing visits in the home, family physician home visits, and specialized palliative care program involvement during the last month; they were married or not living alone and spent the majority of their last month in the home and had an informant who was aware of their approaching death.

Following adjustments for significant covariates (see Table 3), illness-related and health service factors providing significant independent contributions to achieving a preferred home death included the decedent's emotional needs being met (AOR 2.5; 95% CI 1.2, 5.5); the receipt of nursing visits in the home (AOR 2.3; 95%CI 1.3, 4.5); family physician home visits (AOR 1.8; 95% CI 1.0, 3.1); and specialized palliative care program involvement (AOR 2.1; 95%CI 1.1, 3.9). Decedents who spent the majority of their last month at home were 3.2 times more likely than those in hospital to achieve a preferred home death (AOR 3.2; 95% CI 1.4, 7.1). No individual risk-factors were significant contributors.

Discussion

Just over half of decedents who voiced a preference were able to die in their preferred location. This overall congruence was driven primarily by decedents who achieved a preferred death in hospital or LTC. Almost half of those preferring to die at home experienced a hospital death. This disparity is not unusual, and similar findings among countries with publicly financed health systems have been reported.^3,6 It is acknowledged that some hospitalizations may be necessary and unavoidable. However, poorer congruence with home as the preferred place of death has implications for access to hospital services. Decedents who died in hospital but did not wish or plan to do so, could result in increased wait times and emergency room overcrowding for patient populations.^24–26

Independent contributing factors to achievement of a preferred home death stress the importance of integrated care in the home and include meeting emotional needs, nursing support in the home, family physician home visits, specialized palliative care involvement, and being at home the majority of the last month. Decedents with emotional support needs met were more than twice as likely to achieve a preferred home death. Providing emotional support with acknowledgment of approaching death is reported to better prepare the decedent and family for care at home. ²⁷ How best to offer emotional support requires investigation. ²⁸ Nursing care in the home, independent of specialized palliative care, had a major impact on the achievement of a desired home death, more than doubling the odds of doing so. Often the effect of nursing care is not considered as a separate indictor. As such their contribution in our health care system may have been previously underestimated. Research targeting the contribution of home care nursing and an examination of how many hours each week are needed to help enable a preferred home death is required. Family physician home visits as a predictor of home death is reported.^4,13,29 As we plan for the demographic trend of many people with advanced chronic disease cared for principally at home, policies to support family physicians in their home visiting role will be important. How to achieve this in innovative, interprofessional models of primary health care will need thought and evaluation. ³⁰ Specialized palliative care involvement aiding home death achievement is consistent with the literature.^6,15,31 However, past research generally examined specialized palliative care as a package, with additonal home services. In this study, we show specialized palliative care involvement as an independent predictor for preferred home death.

Although our final results stress the importance of services provided to decedents in their homes, we cannot overlook the major contribution of family members. For most, it is family and other informal caregivers who advocate for and enable the provision of home services as they support the decedent's preferrence to die at home.