Better quality of end-of-life care for persons with advanced dementia in nursing homes compared to hospitals: a Swedish national register study

The white paper on dementia care from the European Association for Palliative Care concludes that “transfer to the hospital and the associated risks and benefits should be considered prudently in relation to the care goals and taking into account also the stage of the dementia” [1]. Despite these recommendations, hospitalisation of patients with advanced dementia is common [2].

A study by Mitchell et al. of predictors and risk factors associated with end-of-life stage in dementia clearly demonstrates the difficulties in identifying this stage [3]. In a Swedish study, dementia was a predictor for readmission to hospital within 90 days of discharge [4]. In a Belgian study, 20% of patients with dementia were hospitalised during the last month of life, and slightly fewer than 5% were cared for in an intensive care unit during the last month [5]. Pneumonia, febrile episodes, and eating problems are common for advanced dementia and are associated with high six-month mortality rates [6].

It is difficult to know who will benefit from hospitalisation and who will not. There is evidence that people with dementia can experience a range of adverse outcomes in hospitals; they are hospitalised longer, and more often readmitted to hospital within 3 months, compared to patients without dementia [7]. For patients dying from dementia-related causes, death at a hospital has been shown to be more common for men, married compared to non-married people, and patients 80 years or younger, compared to older patients [8]. According to Sampson et al., patients with dementia in hospitals were less often prescribed symptom management medication and less often admitted to palliative care compared to hospitalised patients without dementia [9].

Old age (regardless of diagnosis) seems to be a factor that influences the palliative care content, regarding both place of death and utilisation of care. Gielen et al. showed that older persons were more likely to die in a nursing home, less likely to be transferred between care settings and stayed fewer days in hospital [10]. Old age has been associated with a lower chance of receiving specialised palliative care [11]. In a Danish population study, women compared to men and patients less than 40 years old compared to 80 years and older were more likely to be admitted to specialised palliative care [12].

Some studies have shown that age has an impact on palliative care for patients with cancer: younger patients with cancer were more likely to receive specialised palliative care in Spain and Italy [13], old age was associated with higher probability of death in hospital [14], a study in Canada on women with breast cancer showed that young age was associated with lower probability of receiving a palliative care–oriented model of health care services [15] and older people who died of colorectal cancer had a lower chance of receiving a palliative care programme [16]. A study using data from the Swedish Register of Palliative Care (SRPC) showed that younger patients with cancer were more often informed about their disease; evaluated for pain and other symptoms; and prescribed medications as needed against pain, anxiety and nausea. Younger patients also more often received palliative care consultation, and their next of kin more often received information and were offered support after death of the patient, compared to those of older patients with cancer [17].

Studies comparing the quality of palliative care for persons with advanced dementia in hospitals and nursing homes are scarce.

A total of 18,610 adults who had died from dementia in hospitals or nursing homes were identified in the SRPC database during the study period. Out of these, 1397 persons were reported to have died unexpectedly, and 365 persons had “Don’t know” reported regarding expected or unexpected death, resulting in missing data, and those were excluded. A total of 16,462 patients, 4752 with Alzheimer’s disease and 11,710 with other or unspecified dementia, remained in the study. A smaller proportion, 967 patients (5.9%), had died at a hospital, and 15,495 patients (94.1%) had died in a nursing home. The mean age for the hospital group was 85.6 years (median 86, range 55–101). The mean age for the nursing home group was 87.4 (median 88, range 41–110). Death at a hospital compared to death in a nursing home was more likely for men than women and more likely in patients aged 65 to 84 years compared to patients 85 years and above (Table 1).

In this national register study of 16,462 adults who died an expected death from dementia in hospitals or nursing homes we have shown that death at a hospital was associated with poorer quality of end-of-life care compared to death at a nursing home for all but three of the measured outcomes. Also, place of death was shown to have a larger impact on end-of-life care outcomes for patients dying expectedly from dementia, compared to age and gender. To the best of our knowledge, this is the largest study so far investigating the quality of end-of-life care in advanced dementia in nursing homes and hospitals, and the only one examining a wide range of quality-related care outcomes.

There are recommendations to, if possible, avoid hospitalisation for patients with advanced dementia [1]. In addition acute settings are not suitable for persons with dementia [23]. Despite this, little is known about where patients with dementia wish to be cared for at the end of life. Also, their families’ opinions and views of the healthcare staff are not fully known [24]. However, carers may be more satisfied with the care given if the patient had died at home compared to having died in a hospital [25].

A marked difference between deaths in nursing homes and hospitals was that 30% died without someone present in hospitals, while the corresponding figure for nursing homes was 10%, although all deaths were reported as having been expected. It is unclear whether this difference is due to lack of resources or if it could be addressed with better routines and awareness of end-of-life care.

Also, there was a pronounced difference between hospitals and nursing homes concerning administration of fluids via enteral tube or intravenously during the last 24 h of life. Over one third of the patients who died an expected death in hospitals received fluids, compared with only 1.4% in nursing homes. According to the white paper on dementia care from the European Association for Palliative Care, permanent enteral tube nutrition may not be beneficial and should as a rule be avoided in dementia, and hydration is inappropriate in the dying phase [1]. The current Swedish national guidelines for palliative end-of-life care advise against routine administration of fluids intravenously or via enteral tubes for most patients who are expected to die within a short timeframe [26].

This study found strong associations between younger age, male gender and hospitalisation for patients dying an expected death from dementia. These findings are in line with the findings from Reyniers et al. [8] and also in line with Houttekier et al., who found that women and patients with higher age were more likely to die in nursing homes than in hospitals [27]. A possible explanation for male gender being a risk factor for hospitalisation could be that men are more often cared for by their spouse at home compared to women, resulting in hospitalisation in case of acute deterioration. Marital status has also been shown as a predictor for hospitalisation [8] but could not be examined in this study because the SRPC database does not contain this data.

Identifying the patient as being in the dying phase can be a key factor in avoiding hospitalisations that do not benefit the patient. Several prognostic tools for survival estimates have been published, but it still remains a challenge to identify the late palliative phase of dementia [3, 28, 29].

Advance care planning has been shown to improve consistency of care with the patients’ goals [30] and may be a way to reduce unwanted hospitalisations and increase the use of palliative care [30]. A do-not-hospitalise order has been shown to reduce the probability of acute hospitalisation [31], and advance directives have been shown to predict less time in hospital for patients with dementia [26]. In the SRPC database, there is no information about these types of orders or why the patient was sent to hospital. Also, advance directives are not legally binding in Sweden. Referral to hospital of a patient may be based on the conclusion that it will be better for the patient or that it is not possible to properly care for the patient in the home care setting or at the nursing home. If the aim is high quality end-of-life care, it may be better for the patient to be taken care of at the nursing home, as the staff know how to communicate and meet the needs of persons with severe dementia. This study demonstrates that a majority of patients with dementia were given end-of-life care in the nursing home, indicating that a majority of patients suffering from dementia can be managed in nursing homes during end of life.

A strength of our study is the large sample size and the population-based data. Despite a large sample size, the group of patients below 65 years of age who had died in a hospital was small, making it hard to draw further conclusions for this subgroup. In the future, it would be interesting to further examine which patients with a dementia diagnosis are hospitalised in the end of life and why. This could possibly be achieved by collecting data, including socioeconomic factors, from medical records and/or the patient registry at the National Board of Health and Welfare.

Death in hospitals was associated with poorer quality of end-of-life care compared to death in nursing homes. Our data support the importance of advance care planning and individual assessments in nursing homes to avoid referral to hospitals during end of life. Despite established recommendations to avoid hospitalisation if possible, there were strong associations between younger age, male gender and hospitalisation in the end of life. Further studies are needed to investigate the role of socioeconomic factors in end-of-life care for this patient group.