In this large, Israeli, population-based study, we found an association between place of death and age, male sex, years of education, and time from diagnosis. Rates of death at home were constant among most socioeconomic residential scores (~30% for levels 2-9). Rates of death at home differed only between the highest and the lowest RSS (42.9% deaths at home at level 1 and 15.0% at level 10). Death at home rates were highly variable among the different cancers: the highest rate was seen among patients with brain tumors (37.0%), while the lowest rate was among patients with hematologic malignancies (lymphoma and leukemia, 20.3% and 20.0%, respectively). In addition, a slight increase was seen each year during the last years of the study.
The association found between dying at home and socioeconomic status has been previously reported in other countries . The mechanisms whereby socioeconomic status affects one’s ability to die at home may involve various factors: whether a person lives alone, the ability to provide adequate care at home, adequate access to care services, care-related costs, control over monetary resources, the ability to secure local support, and personal preferences . Socioeconomic status may, however, influence or be influenced by other dominant variables. For example, higher socioeconomic status may contribute to a more stable home environment and increased patient comfort with home death, and more resources may be available to pay for formal caregivers and strong support networks, thus increasing the likelihood of dying at home. Rather than a direct effect on the likelihood of dying at home, higher socioeconomic status may be the result of higher education, which in itself may result in a higher likelihood of dying at home . High socioeconomic status is also characteristic of those seeking physician-assisted suicide as a tool to achieve the sense of control that is of importance to this population [11,12]. Dying at home may present a similar measure of control over the uncontrollable circumstances of approaching end of life. On the other hand, the association between high socioeconomic status and death at home may reflect unequal opportunities for optimal delivery of care according to patients’ wishes at end of life.
The constant rate of dying at home among interim levels of socioeconomic status supports the importance of defining socioeconomic status as an ordinal variable rather than a continuous variable when attempting to understand complex mechanisms whereby socioeconomic status affects one’s ability to die at home. The fact that for most RSS levels in Israel home death rates were constant may have multiple explanations. Israel has National Health Insurance coverage and highly accessible health services , which may decrease the impact of socioeconomic status on accessibility to health care, resulting in differences only between extreme levels of RSS. In addition, the relationship between low socioeconomic status and dying in hospital may be attenuated in Israel by the nature of the population in areas with low RSS. In these areas, there are more ultra-orthodox Jews and Muslim Arabs. These populations are characterized by a preference to die at home, as well as strong social support that may help fulfill the wish to die at home [14-17]. Yet, the current results highlight the need for improvement of end-of-life care at home among patients in the lowest RSSs.
In our study, while death at home occurred at a relatively fixed rate in patients from 31 (20.9%) to 70 (27.9%) years of age, the oldest patients exhibited a significant increase in the percentage of those dying at home: 33.6% among those aged 71-80, 41.1% among those aged 81-90, and reaching 50% in those 91 years old and above. A similar association between age and dying at home has been reported in other countries, primarily in Singapore . However, in Scotland, age has been found to be inversely associated with dying at home [18,19]. This variability likely reflects cultural differences in the care of the elderly as well as availability of personal care workers. The strong association found between age and death at home of patients over 81 years old in our study is remarkable. These results may be related to the fact that the oldest old cancer patients in Israel invest resources in close relations and attribute fundamental importance to family caregivers, turning their thoughts to their surroundings, relatives, and family , and they may therefore prefer death at home in their accustomed environment . Dying at home is fraught with illness and caregiver-related uncertainty . Illness-related uncertainty may be less prevalent for the oldest patients, and indeed many of them may already have formal caregivers, thus reducing caregiver-related uncertainty.
The lowest rates of death at home among patients with hematologic malignancies may reflect a younger sub-population. On the other hand, end-of-life care for patients with hematologic malignancies deficient compared to treatment for other solid tumors . More aggressive treatments at the end of life resulting from the continuing discovery of cures for advanced diseases, as well as the, often rapid, pace of decline near death, have been reported among patients with hematologic malignancies . Another barrier to efficient end-of-life care for this population is unrealistic expectations from both physicians and patients. Also, quality measures developed for solid tumors (such as blood transfusions) may be unacceptable for patients with blood cancers .
In contrast to those suffering from hematologic malignancies, patients with brain tumors have the highest rates of death at home. Swallowing difficulties combined with decreased function and consciousness are frequently seen among these patients during the end-of-life period . These symptoms frequently appear a few weeks before death. During this period, major challenges are to efficiently provide basic adult daily activities and to deal with minor medical problems, which can frequently be done at home.
Rates of death at home gradually increased during the years 2014-2017. We hypothesize that these changes are the result of the National Program for Palliative Patients, launched in 2015.
Strengths and Limitations
The strengths of this study include the use of large national databases which allowed us to adjust the results for potential confounders and to mitigate selection and information biases. However, this study did not include information on patients' or family members' preferences for place of death, both of which have a great impact on the place of death [8, 25].