Place of death in patients with dementia and the association with comorbidities: a retrospective population-based observational study in Germany

The proportion of people developing a dementia-related disease increases with increasing age. Older people’s state of health is also usually characterized by comorbidity — i.e., they suffer from several diseases simultaneously.

In Germany, it is estimated that about 1.6 million people are currently diagnosed with dementia [1]. The absolute numbers of affected people have been estimated as 8.7 million in Europe in 2013 [2] and 46.8 million worldwide in 2015 [3]. Due to the age-dependency of the disease process and continually rising life expectancy, particularly in Western industrialized countries, the prevalence of the disease will increase further in the future. On the basis of predicted population trends in Germany, the number of patients with the condition will increase by around 40,000 annually and will rise to about 3 million by 2050 [1].

There is currently no treatment for dementia and the condition usually progresses very slowly. The duration of the disease cannot be reliably predicted in the individual case. Overall, the age-specific mortality rate is at least double that for individuals without dementia [4, 5]. Sampson et al. demonstrated in a prospective cohort of people with advanced dementia in the UK that 37% of these people died during a 9-months observational period [6]. Similar mortality rates have been reported in other countries [7‐9].

During the course of the disease, people with severe dementia lose almost all their learned skills and abilities. They consequently require extensive nursing and medical support in many life situations. This represents a major health-policy and social challenge. It also affects end-of-life care. The disease is increasingly regarded as life-limiting by physicians, and the need for palliative care in patients in the advanced staged of dementia has been noted [10, 11].

Place of death is regarded as a kind of quality indicator for evaluating end-of-life care. Surveys on place of death show that most people clearly prefer to die at home rather than in institutions [12‐14].

The place of death is not listed in official statistics in Germany, since the information given on the death certificate is not further analyzed by the relevant authorities. Studies on place of death for the general population in Germany show that hospitals are by far the most frequent place of death, followed by the home environment, retirement or nursing homes, hospices, and palliative care units [15].

Hardly any data regarding place of death are available for individuals with dementia in Germany. Escobar Pinzon et al. showed that in the federal state of Rhineland–Palatinate in 2008 42.4% of those with dementia died at home, followed by nursing homes (26.9%), hospitals (26.2%) and palliative institutions (hospices and/or palliative units; 3.2%) [16]. International studies show that individuals with dementia mainly die in institutions, with nursing homes and hospitals to some extent, being the most frequent place of death in most countries [16‐22].

Older people with dementia often suffer from multiple additional diseases [23]. On average, two to eight other chronic diseases are present [24, 25]. In 3971 patients with dementia aged over 64 receiving care from family physicians in Spain, at least three other diagnoses were present in 70% with the most frequent being arterial hypertension, osteoarthrosis (in both women and men), as well as anxiety disorder/neurosis in women and benign prostate hypertrophy in men [26]. In the UK, arterial hypertension (53.4%), chronic pain (33.5%), depression (23.5%), presbyacusis (22.3%), coronary heart disease (21.6%), and chronic renal failure (20.8%) were the most frequent comorbidities in 4999 patients with dementia [27]. Overall, patients with dementia have a higher prevalence of complex situations that indicate functional limitations (including immobility, dysphagia, and impaired hearing), depression, and frailty syndrome (reduced physical activity, weakness, fatigue, weight loss) [28]. In addition, these patients have more often emergency hospital admissions compared to patients without dementia, and the number of hospital admissions increases with the severity of the disease [29, 30]. The reasons for hospital admission are often bronchial and urogenital infections, falls, or fractures, as well as delirium [31, 32]. Although it appears obvious from the clinical point of view that comorbidities contribute to the place of death for patients with dementia, hardly any scientific evidence is available on the topic.

The aim of the present study was to describe the place of death of patients with dementia in Germany on the basis of analyzed death certificates and to investigate the extent to which specific comorbidities are associated with the place of death.

A total of 10,364 death certificates of patients who had died at the age of 65 or over were analyzed. Dementia was described in 1646 cases, representing a relative frequency of 15.9%. A larger proportion of women (19.5%) than men (11.4%) were affected. The standardized prevalence of all individuals with a dementia-related disease was 8.0% (women 9.0%, men 6.9%; data not shown). One in ten deceased persons aged 75–79 suffered from dementia. In the 95 or older age group, one in four men and one in three women were affected by the disease (Fig. 1).

Overall, 67.3% of the deceased patients were women. Of deaths in retirement or nursing homes from dementia, two-thirds were women. Of deaths in hospital, just under half were men. The mean age was 86.3 ± 6.9 years. On average, 1.5 ± 1.0 other comorbidities were present. No comorbidity was present in 14.9% of cases, one additional condition was present in 40.6%, and three comorbidities were present in 10.6% of the cases. The most frequent accompanying disease was cardiac failure (43.6%). Pneumonia was documented in one-quarter of the cases, and cancer in just over one in ten. Patients with dementia who died in hospital suffered significantly more often from pneumonia, aspiration, sepsis, and renal failure compared to patients with dementia who died at home. Patients with dementia who died in hospices had malignant tumours more often compared to those who had received terminal care in the home environment (71.4% vs. 12.8%; p < 0.01) (Table 1).

Patients with dementia dying in hospital had a high proportion of infectious diseases (such as pneumonia or sepsis), aspiration, and renal failure in comparison with other places of death that were investigated. In contrast, septic conditions were only rarely noted in death certificates for those who died at home (1.2%), while cardiac failure was the most frequent in that location with 45.9%. In retirement or nursing homes, the frequency of documented pneumonia was similar to that for deaths in the home environment (19.9%) which was much lower in comparison with hospitals (39.1%). In hospices, dementia patients mainly died of tumours (Fig. 2).

Retirement or nursing homes were by far the most frequent place of death. Approximately every second patient died there. Hospitals represented the second most frequent place of death (28.7%). Only one in five deaths occurred in the home environment. Palliative care units and hospices played a subordinate role, with a total of 1.3%. Stratified by sex, women died more often in retirement or nursing homes, while men by contrast died more often in hospital (Fig. 3).

In the multivariate regression analysis, very elderly patients with dementia (86 years, ≥ 0.5 quantile) and women had a higher odds of dying in retirement or nursing homes compared to the home setting (OR 1.59 [95%-CI 1.21–2.08], p = 0.001; OR 1.55 [95%-CI 1.16–2.08], p = 0.003) and a lower odds of dying in hospital compared to retirement or nursing homes (OR 0.50 [95%-CI 0.38–0.64], p = 0.001; OR 0.50 [95%-CI 0.38–0.66], p = 0.001). There was a statistical association between pneumonia, sepsis, and renal failure and hospitals as a place of death. In comparison with deaths at home, the odds of dying in hospital was three times higher when there was a medically documented pneumonia (OR 2.96 [95%-CI 2.01–4.35], p = 0.001), while in the presence of renal failure it was twice as high (OR 1.93 [95%-CI 1.26–2.97], p = 0.003) and with sepsis 14 times higher (OR 13.73 [95%-CI 4.88–38.63], p = 0.001). Similarly, in comparison with deaths in retirement or nursing homes, the odds of dying in hospital was also higher in the presence of pneumonia (OR 2.38 [95%-CI 1.75–3.25], p = 0.001), sepsis (OR 7.34 [95%-CI 4.21–12.78], p = 0.001) or renal failure (OR 1.65 [95%-CI 1.18–2.32], p = 0.003). There was also a correlation between the diagnosis of myocardial infarction and an increased probability of dying in hospital (OR 2.52 [95%-CI 1.45–4.36], p = 0.001) (Table 2).

Patients with dementia most often died in retirement or nursing homes, followed by hospitals and the home environment. Palliative care units and hospices as places of death played only a minor role. An association was seen between selected comorbidities and an increased likelihood of dying in hospital.

In the present study, nearly one in two deaths among patients with dementia occurred in retirement or nursing homes in Westphalia-Lippe (Germany). This finding is not surprising, as it reflects the high level of nursing care required by dementia patients. Relatives who are caring for people with dementia usually have a strong wish to care for and look after the patient in the shared home environment. However, relatives dealing with dementia patients on a daily basis are exposed to a large number of problems and challenges. The time demands involved in caring often conflict with the carer’s own family, and working life. Persons with dementia may also show depressive or even aggressive behavior in the course of their disease, as well as developing restlessness and/or a marked urge for movement. This can lead to a high level of physical and above all emotional burden on caring relatives, which may even cause social isolation [35, 36]. Caregivers usually belong to the patient’s immediate family (first-degree relatives, children, spouses), or more rarely they may be friends or other people linked to the patient [37]. If relatives struggle to cope with the situation they can get potentially support from a home help, a day or temporary nurse or a nursing service. However, these measures are often only effective in the shorter term. Alternative residential forms are available, such as “sheltered housing” or “dementia apartment-sharing,” but a move to a nursing home is often the only practicable way of ensuring care.

This study has shown that (unsurprisingly) it is mainly very elderly people and women who die in retirement and nursing homes. This observation is explained by demographic change and changes in social life. Life expectancy has been increasing for decades in the Western industrialized countries, and this applies to Germany as well. In this country, the mean life expectancy is currently 83.1 years for women and 78.2 years for men [38]. Due to their lower life expectancy, men are more likely to be survived by their partners, and this also increases the probability that they will be cared for by relatives at home at the end of their lives. Also, due to the increasing age, there is a greater likelihood that women will be widowed or living alone when they are elderly and, with increasing physical problems will be dependent on assistance from strangers or institutions. The results of the 2011 population census in Germany also indicate that more and more people are living alone. The proportion of people living alone, for example, increased from 15.6% in 1996 to 19.6% in 2011 [39].

Surveys have shown that most people would prefer to die at home [12‐14] and this also applies to people with dementia-related diseases [16]. The findings of the present study are in contrast to this wish of patients: only one in five persons with dementia died at home. The reasons for this remain speculative, but it may again be linked to excessive stress on relatives caring for the patient and show that there is a need for relevant action to be taken in health care policy. This need has been recognized by the relevant German political decision-making body, and measures have been implemented [40]. The legal meaning of the term “status of requiring care” has been redefined and extended to mental and psychological illnesses. Patients with dementia have consequently had their previous benefit entitlement from the nursing insurance fund upgraded. In addition, relatives who have had to stop working in order to provide care are now receiving improved financial support from the state.

Advanced-stage dementia is increasingly being regarded as a terminal disease leading to death [10, 11, 41, 42]. Palliative care is appropriate in dementia, since it represents a “disease that does not respond to curative treatment” or a “life-threatening disease,” as dementia itself is not curable. The treatment approach aims at achieving improvements in quality of life. The results of the present study show that in-patient palliative and hospice institutions were only playing a minor role in 2011 in comparison with all other places of death. Only 0.4% of all dementia patients who died received end-of-life care in a palliative care unit, and only 0.9% of them received care in a hospice. There might be several reasons for this observation. First, a health economic aspect. There were only 32,1 palliative care unit beds and 40,2 hospice beds in the study region relative to a population of 1 million — corresponding to two-thirds of the maximum number recommended by the European Association for Palliative Care (EAPC) [43]. Thus, there was a need for implementing further inpatient palliative care and hospice services in this region in 2011. Second, the life expectancy of patients with dementia. In many cases, the natural course of the disease often exceeds the official requirements that patients should only be admitted to hospices when the medical estimate of life expectancy is less than 3–6 months.

Several investigations indicate that patients with dementia are at increased risk of hospital admission compared to people without dementia [29, 30, 43‐46]. The reasons for this are complex [6, 31, 32, 47‐49]. The most frequent causes include respiratory and urogenital infections, fall-related injuries, neurological and psychiatric causes (syncope, confusion, delirium), pressure sores, and nutritional disturbances. There is a consensus in the research findings that many of these diseases could have been treatable in home care or in in-patient care institutions, so that hospital admission could have been avoided [50, 51]. Psychosocial factors also affect hospital admissions — for example, when the previous carer suddenly becomes unavailable.

Generally, a hospital stay is a severe burden for many people with dementia and it is also associated with a number of risks. These include prolongation of the hospitalization period, a decline in physical functional abilities, increased frequencies of nosocomial infections, and an increased likelihood of not being able to return to the home environment after the hospital treatment [52]. Sampson et al. concluded that an unplanned hospital stay significantly shortens the median survival time in patients with dementia [53].

In the present study group of deceased individuals in the general population aged 65 or older with dementia, in-patient deaths represented 28.7% of cases. A similar percentage was reported by Houttekier et al. in a European survey in 2003 [18]. The mean percentage of dementia patients aged 65 or over who died in hospital in that study was 27.4%, including all countries investigated (Belgium, Netherlands, England, Wales, and Scotland). The Netherlands showed a very low percentage, with only 2.8% of deaths occurring in in-patients. In this country, some general practitioners (“verpleeghuisarts”) work exclusively in nursing homes enabling them to monitor the state of health of nursing-home patients tightly and offer medical treatment in a timely manner when physical changes occur. In most cases, hospital admissions can be avoided.

The medical information on cause of death that was analyzed in the present study showed that the deceased dementia patients had been suffering from a mean of 1.5 of the selected comorbidities. Cardiac failure was the most frequent comorbidity documented in the death certificates with almost one in two deaths, with the diagnoses of pneumonia, renal failure and malignant tumours following in frequency. Cardiovascular diseases are the most common cause of death in Germany, followed by cancers. The diagnosis of dementia is already in third place [54]. The prevalence of cardiac failure, like that of dementia, increases with increasing age [55, 56], and this may have contributed to the high prevalence of cardiac failure in the present sample. Many patients with advanced dementia also suffer from dysphagia [57], which may make fluid intake much more difficult. This can lead to dehydration and prerenal kidney failure. In addition, there is a risk of aspiration of fluid and food particles potentially resulting in pneumonia leading to sepsis and multiple-organ failure causing death finally [58]. In several autopsy studies, pneumonia was the most frequent cause of death [59, 60]. These findings are supported by clinical data. Mitchell et al. [7], for example, noted in the CASCADE study that 41.1% of the patients developed at least one episode of pulmonary infection during the 18-month follow-up period. The infection was associated with a high mortality rate. In the Netherlands, the three most frequent causes of death in nursing-home residents with dementia were dehydration (38%), cardiovascular diseases (19%), and respiratory infections (18%) [61]. In the present study, the death certificates described pneumonia in 25.2% of cases and aspiration in 10.0%. Pneumonia and/or aspiration were particularly frequent on the death certificates of dementia patients who died in hospital (39.1 and 15.6%, respectively). Compared to home or nursing home deaths, the odds of dying in hospital with documented pneumonia was two to three times higher, and with aspiration by a factor of 1.3 or 1.7 higher. In addition, deceased hospital patients often had sepsis (14.4%), renal failure (19.7%), and myocardial infarction (7.2%). These results suggest that there is a high intensity of treatment in hospital at the end of life in dementia patients. Unfortunately, the study was not able to provide any further information on this.

Care for patients with dementia in the last phase of their lives represents a special challenge, since those affected are often unable to express their treatment preferences directly themselves, while established, evidence-based treatment pathways for this phase of disease are still largely lacking. Research results showing that physical symptoms are widespread in persons with dementia and that they even increase before death [6, 7, 16, 61].

In Germany, a new law was passed in 2015 to improve hospice and palliative care [62]. The law aims to strengthen comprehensive hospice and palliative care in Germany by implementing targeted measures in statutory health insurance and social care insurance. The measures are intended among other things to ensure networking among medical and nursing services, as well as attendant hospice services, and to guarantee cooperation among the health-care providers involved. The aim is to strengthen palliative care and hospice approaches in in-patient care institutions and hospitals and to offer information to health-insurance policy-holders in a targeted way about the hospice and palliative care services available, as well as enabling nursing-home residents to carry out individualized care planning for the last phase of life. The statutory framework conditions have been set out, but they require specific arrangements and a financial basis so that everyone in Germany — and particularly those with dementia — can be offered adequate palliative medical care adapted to their individual needs at the end of their lives.

The most common place of death in people with dementia was the retirement and nursing home, followed by hospital. Only one-fifth died in the home environment.

End-of-life care for people with dementia represents a special challenge and requires a person-centered care approach with staff qualified in palliative care. In this context, existing nursing and medical care services and hospice services need to be further developed and extended to ensure that all individuals with dementia can receive adequate palliative care in accordance with their own individual needs at the end of their lives.