People stay in various different settings at the end of their lives: at home, in a hospital, in residential care for the elderly or at specialised palliative care facilities such as hospices or palliative care units. Dying in the preferred setting is an indicator of good palliative care quality [1]. Most people would prefer to die at home [2]. There is however a discrepancy between the desired and actual places of death. An international comparative study of death certificates in 14 countries has shown that 13% to 53% of people who die from diseases that are known to be associated with palliative care needs relatively often (cancer, organ failure, COPD, neurological diseases, HIV/AIDS) die at home. Between 25% and 85% die in hospital, 1% to 35% in residential care for the elderly and 4% to 13% in a specific palliative care facility [3]. In the Netherlands, patients die at home relatively often, but in this country too there is a discrepancy between the desired and actual places of death: over a third of Dutch people who die from diseases that are known often to be associated with palliative care die at home [3], which is clearly lower than the 68% who indicated in the study by Koekoek that they would like to die at home [4]. Over a third die in residential care for the elderly and a quarter (25%) die in hospital [3].

An interesting issue is whether the perceived quality of care is associated with the place of death. Earlier research has shown that - according to the relatives - people with cancer who remained at home at the end of their lives perceived a lower burden from their symptoms and the end of their life was better than people with cancer who died elswhere [5]. Relatives of people with dementia who stayed at home were more satisfied with the care [6] and relatives of people with chronic conditions who received care at home indicated that they had fewer care needs that were not met [7]. The care for people who die in hospital was rated less highly (in comparison with the care at home or in a nursing home) by the relatives of people with dementia [6]. In addition, it is known that relatives rate the care received by cancer patients who die in a hospice more highly than the care received by people who die in hospital [8].

Existing research therefore shows that patients prefer to die at home and that relatives are more positive about the care at home or in a hospice than in hospitals. Limitations of the existing studies into the quality of care around the time of death are however that only a single specific group of patients was examined (e.g. cancer or dementia [5, 6, 8]), that they were carried more than a decade ago [7], and/or that not all care settings that are relevant for palliative care were included [7, 8]. In this article, we are therefore reporting a study based on questionnaire data from relatives of people with a wide range of diseases (such as cancer, dementia and various cardiovascular problems and lung disease) who died in various care settings.

On the basis of the studies described above, we expect that relatives will perceive the quality of care to be better when the patient dies at home. This is particularly the case for aspects of care in which the bond between the caregiver and the care recipient plays a major role. In the Netherlands, people often have long-term patient/doctor relationships with the same general practitioner. They also sometimes receive care at the end of their lives for weeks or months from nursing staff or carers from the home care services. In addition, we expect that relatives of people who die in specialised palliative care institutions (a hospice or palliative care unit) will have relatively good experiences with the care provided. We expect that this will apply in particular to aspects such as expertise and the provision of information, because of the palliative care skills and specialisation that such settings have.

When various aspects of palliative care are examined, the quality of palliative care for people who die at home is rated more highly by their relatives than is the case for people who die elsewhere. In particular when the patient dies in hospital or in a residential setting for care for the elderly, the quality of care is perceived as less good by the relatives. Care associated with death in a hospice and in other institutions such as those for the intellectually disabled or mental healthcare services is rated almost as highly as the care for people who die at home.

Perceptions of quality of care at the end of life are affected not only by the place of death, but also by background characteristics and disease characteristics. Our research confirms this: the model containing not only characteristics of the backgrounds and the diseases but also the place of death is better able to explain the differences in the quality of care as experienced than the model containing only the background and disease characteristics. In addition, disease characteristics such as the type of disease can affect the place where someone dies. For instance, people with dementia [13–15] and those who have had a stroke [14] mostly die in a nursing home or care home, and people with cancer die most often (in the Netherlands) in their homes [16]. When people are admitted to hospital shortly before death, this is often because of an acute situation with symptoms such as shortness of breath or digestive or cardiovascular problems [17]. The analyses therefore take account of both the burden of symptoms and the patient’s disease; even after these corrections, the differences in the perceived quality of care between the various care settings remains statistically significant. Nevertheless, the ability of the models to explain the differences is no more than 12%, even with the background and disease characteristics added to the place of death. That means that there are other factors that affect the perceived quality that cannot be accounted for by our data.

The finding that the care provided to people who die at home is rated most positively supports the current Dutch policy of letting people live at home as long as possible [18]. In addition, it is in line with the wishes of many people, who would prefer to die at home [2]. Nevertheless, it is not always either possible or desirable for a patient to die at home. Research by Korte-Verhoef et al. [19] has for instance shown that a large proportion of hospital admissions are avoidable, but at the same time that there can still be situations in which the symptoms are so complex that admission is inevitable. It is therefore important that sufficient attention is paid to good palliative care in settings other than home, and in particular those that are not specialised in palliative care, such as hospitals and residential care for the elderly. A consultation team or specialist palliative team can provide support for this [20, 21]. Our results also demonstrate that this specialist palliative care can offer added value. Experiences with the care in a palliative unit in a nursing home or care home are more positive than those in the regular departments of a nursing home or care home. This applies in particular to the aspects involving care for the relatives. Specialist palliative care is also provided at hospices. The care associated with the patient’s death is also rated relatively highly by the relatives of patients who die in this setting.

The principal reasons for people being admitted to a hospice in the Netherlands are when there is a preference for dying in a hospice and when intensive care and support are needed [22]. This specialised set of knowledge and skills results in good care for both the patient and their loved ones. In addition to the institutions already described, there are also institutions where palliative care is given to specific target groups. In our analysis, the ‘other institutions’ group largely comprises institutions for the intellectually disabled and for mental healthcare. The perceptions of relatives of people who die in these settings are no less positively rated (with the exception of a single aspect) than when people die at home. Because people often live in such institutions for a large proportion of their lives, these institutions can also be seen as a home setting to a certain extent. It is however only a small group within our analyses. To obtain a better picture of the experiences with palliative care in these institutions, further research into these specific target groups is needed.

Our research also looked at the differences in specific aspects of palliative care. In all the care institutions included in this study, the aspect of ‘respect/care for autonomy’ was experienced as being less good than the home setting. This aspect looks at the extent to which the loved ones are involved in decisions made by the caregivers about the care, whether they could determine what their own role was within that care, and whether allowances were made for the individual wishes of the relatives with respect to the care for the patient. One possible explanation for this is that the partners of people who die at home often live in the same house and are therefore closely involved in the care and also tend to take on the role of informal caregiver. Previous research has also shown that it is important for the relatives of people who die at home to be included as part of a good professional care team, and that they have the opportunity to tackle the symptoms and administer pain relief [23]. In addition, the fact that the general practitioner is heavily involved may play a role, as general practitioners in the Netherlands will often have known the patient and their family for many years.

The quality of palliative care from the relatives’ perspective is highest when the patient died at home or in a hospice. The experiences are less positive in particular for patients who died in hospital or in residential elderly care. This applies both to the general ratings given to care for the patient and their loved ones and to specific aspects of palliative care such as care for the psychosocial well-being of the patient and their loved ones, the attitudes to them, autonomy, information in the final week before death, and expertise. This can partly be ascribed to differences in the patient population, but it also seems partly to be genuinely related to the care setting itself. This is an argument for letting people die at home, if they so wish and where feasible, as far as possible.