Advertisement

Determinants of the Place of Death in the Brussels Metropolitan Region

Open AccessPublished:April 06, 2009DOI:https://doi.org/10.1016/j.jpainsymman.2008.05.014

      Abstract

      The place where people die is not only considered an indicator of quality of death, but also has implications for health care costs and the organization of end-of-life care. Advancing urbanization, combined with social fragmentation, poor social conditions, and concentration of inpatient care in large cities make it relevant to study the place of death in a metropolitan context. The objective of this article is to examine determinants of place of death (home, care home, hospital) in a Belgian metropolitan region (Brussels) for patients suffering from chronic diseases eligible for palliative care. Using death certificate data, we describe place of death and associated factors for all deaths after chronic diseases in 2003 in Brussels (n=3672). Of all chronically ill patients, 15.1% died at home, 63.0% in hospital, and 21.6% in a care home. Of those residing in care homes, 23.8% died in hospital. Noncancer patients and residents of districts with higher socioeconomic status had a higher chance of dying at home or in a care home if they resided in one. Home death was also more likely for patients not living alone. Care home death was more likely with increasing age. Compared with other parts of Belgium and other big cities worldwide, few patients eligible for palliative care in Brussels died at home. Both the overall low proportion of people dying in familiar surroundings and the inequality between different districts in Brussels imply that a health policy aiming at facilitating dying in the place of choice might also need to develop specific approaches for metropolitan cities.

      Key Words

      Introduction

      Place of death is an important issue in end-of-life care research. The place where someone dies is often considered an indicator of quality of the end of life.
      • Clark D.
      Between hope and acceptance: the medicalisation of dying.
      • Steinhauser K.E.
      • Christakis N.A.
      • Clipp E.C.
      • et al.
      Factors considered important at the end of life by patients, family, physicians, and other care providers.
      • Grande G.E.
      • Farquhar M.C.
      • Barclay S.I.
      • Todd C.J.
      Caregiver bereavement outcome: relationship with hospice at home, satisfaction with care, and home death.
      Hospital death, although preferred by a few patients and sometimes clinically inevitable, is often considered as suboptimal
      • Barbera L.
      • Paszat L.
      • Chartier C.
      Death in hospital for cancer patients: an indicator of quality of end-of-life care.
      • Bowling A.
      The hospitalisation of death: should more people die at home?.
      compared with home and care home death, because the latter are more in accordance with the wishes of patients, their caregivers, and their health professionals.
      • Teno J.M.
      • Clarridge B.R.
      • Casey V.
      • et al.
      Family perspectives on end-of-life care at the last place of care.
      • Gott M.
      • Seymour J.
      • Bellamy G.
      • Clark D.
      • Ahmedzai S.
      Older people's views about home as a place of care at the end of life.
      Although most people in Western societies prefer to die at home,
      • Foreman L.M.
      • Hunt R.W.
      • Luke C.G.
      • Roder D.M.
      Factors predictive of preferred place of death in the general population of South Australia.
      • Beccaro M.
      • Costantini M.
      • Rossi P.G.
      • et al.
      Actual and preferred place of death of cancer patients. Results from the Italian survey of the dying of cancer (ISDOC).
      • Tang S.T.
      When death is imminent—where terminally ill patients with cancer prefer to die and why.
      • Tang S.T.
      • McCorkle R.
      Determinants of congruence between the preferred and actual place of death for terminally ill cancer patients.
      a study in six European countries showed that between 33.9% and 62.8% of all deaths in 2003 occurred in hospitals.

      Cohen, J, Bilsen J, Addington-Hall J, et al. Population-based study of dying in hospital in six European countries. Palliat Med 2008;22(6):702–710.

      Additionally, home death supported by home care services is believed to reduce the overall cost of care by reducing the number of hospital inpatient days.
      • Hearn J.
      • Higginson I.J.
      Do specialist palliative care teams improve outcomes for cancer patients? A systematic literature review.
      The discrepancy between actual and preferred place of death, and the budgetary consequences,
      • Burke K.
      Palliative care at home to get further funds if it saves money.
      • Chochinov H.M.
      • Kristjanson L.
      Dying to pay: the cost of end-of-life care.
      • Miller S.C.
      • Intrator O.
      • Gozalo P.
      • et al.
      Government expenditures at the end of life for short- and long-stay nursing home residents: differences by hospice enrollment status.
      make the place of death and the organization of dying a point of particular medical and societal interest.
      Although studied in many countries,

      Cohen, J, Bilsen J, Addington-Hall J, et al. Population-based study of dying in hospital in six European countries. Palliat Med 2008;22(6):702–710.

      the place of death has not been studied specifically in a metropolitan setting in continental Europe before, even though this could possibly be very relevant for health policy. Nevertheless, there is proof of some variation in place of death between population groups in rural and urban areas, possibly indicating differences in the organization of end-of-life care; there is also a growing urban population worldwide, and urban regions have a typical population mix and a concentration of institutionalized care. The Brussels metropolitan region, for instance, consists of 46.3% people of foreign descent

      Brussels Health and Social Observatory. Atlas of Welfare and Health of Brussels Capital Region 2006 [in Dutch]. Brussels: Common Community Commission, 2006.

      and has, compared with the rest of Belgium, 15% more inhabitants of over 80 years, nearly twice as many single people, and about twice as many poor;

      Brussels Health and Social Observatory. Barometer of Welfare of Brussels Capital Region 2007 [in Dutch]. Brussels: Common Community Commission, 2007.

      it also has a larger concentration of inpatient care facilities. These aspects, making the implementation of principles of palliative care in an inner city extra difficult,
      • Hughes A.
      Poverty and palliative care in the US: issues facing the urban poor.
      • Murphy P.A.
      • Price D.M.
      Dying and grieving in the inner city.
      may influence the place of death.
      • Zallman L.
      • Sanchez J.
      • Choi H.
      • Selwyn P.
      Unmet need for chronic disease and end-of-life care at urban family health centers in the Bronx, NY.
      • Born W.
      • Greiner K.A.
      • Sylvia E.
      • Butler J.
      • Ahluwalia J.S.
      Knowledge, attitudes, and beliefs about end-of-life care among inner-city African Americans and Latinos.
      • Koffman J.
      • Burke G.
      • Dias A.
      • et al.
      Demographic factors and awareness of palliative care and related services.
      In Belgium and Brussels, palliative care is integrated into the existing health services and is, therefore, mainly provided by family physicians in a home care situation (sometimes supported by a regional multidisciplinary palliative home care team) or by specialized palliative care units in hospitals. Unlike other countries, hospices have not been established in Belgium. Palliative care in long-term care facilities (i.e., care homes) is provided by the patient's family physician, supported by nurses, the facility's palliative support team and the inpatient palliative care unit of an associated hospital. The family physician, thus, plays a central role in the provision of palliative care at home or in a home-substitute institution. Only a family physician can authorize supplementary compensation for the cost of palliative home care and authorize a family member to take compassionate leave for palliative care.
      The objective of this article is to examine determinants of place of death in the Brussels metropolitan region for patients who died after chronic illness, and thus possibly could have benefited from receiving palliative care. The determinants of place of death were analyzed for patients residing at home and patients residing in a care home separately. We tried to answer the following research questions: First, where do people with chronic illness die in the Brussels metropolitan region? Second, are there any important differences in patterns of place of death depending on illness, sociodemographic characteristics, or social support available to the patient? All the latter factors have been identified as determinants of place of death in the literature.
      • Mor V.
      • Hiris J.
      Determinants of site of death among hospice cancer patients.
      • Gomes B.
      • Higginson I.J.
      Factors influencing death at home in terminally ill patients with cancer: systematic review.

      Methods

      This study analyzed all deaths from chronic diseases in the Brussels metropolitan region in 2003 (n=3672). The autonomous Brussels metropolitan region is the capital region of Belgium and consists of 19 municipalities with, in 2003, approximately 1 million inhabitants. The data were extracted from the death certificates. In Belgium, these data are primarily used for public health purposes, such as the production of vital statistics. The first part of the death certificate, including the cause of death, the sex of the deceased, and the time and place of death, is completed by a physician. Information on sociodemographic characteristics of the deceased, including residence, age, education, occupation, nationality, civil status, and living situation, is completed by the municipal administration. The authority responsible for the death certificates (in Brussels: the Brussels Health and Social Observatory) processes all information and codes the causes of death conforming to the International Classification of Diseases (ICD-10).
      To identify deaths from chronic disease eligible for palliative care, the operationalization of the palliative care subset by Rosenwax et al. was used.
      • Rosenwax L.K.
      • McNamara B.
      • Blackmore A.M.
      • Holman C.D.
      Estimating the size of a potential palliative care population.
      This subset consists of people who died from one of 10 chronic conditions: neoplasm, heart failure, renal failure, liver failure, chronic obstructive pulmonary disease, motor neuron disease, Parkinson's disease, Huntington's disease, Alzheimer's disease, and human immunodeficiency virus/acquired immunodeficiency syndrome.
      • Rosenwax L.K.
      • McNamara B.
      • Blackmore A.M.
      • Holman C.D.
      Estimating the size of a potential palliative care population.
      • McNamara B.
      • Rosenwax L.
      Factors affecting place of death in Western Australia.
      The outcome variable in all analyses was the place of death, with four possible outcome categories: home, hospital, care home (nursing home or residential home for older people), and elsewhere (other than previous places, e.g., public road, workplace).
      Seven independent variables available by means of death certificates can be classified into three main categories: variables related to illness (one variable), variables related to personal characteristics (four variables), and variables related to social support (two variables).
      • Gomes B.
      • Higginson I.J.
      Factors influencing death at home in terminally ill patients with cancer: systematic review.
      The variable related to illness in our analysis is the underlying cause of death, as reported on the death certificate. The 10 categories of chronic disease of the palliative care subset were recoded in two categories: died of a cancer condition and died of a noncancer condition. The variables related to personal characteristics were patient's sex, age, nationality (coded in five categories: Belgians, Northern and Western Europeans, Southern Europeans, Northern Africans and Turks, and others) and the socioeconomic status (SES) of the district of residence (coded in four categories: low, middle-low, middle-high, and high). The classification of the 19 Brussels municipalities into four districts with similar SES was performed using a cluster analysis with four variables: average individual income, percentage of the residences without private bathroom or toilet, unemployment rate, and educational attainment. The variables that refer to the social support of the deceased are marital status (unmarried, married, widowed, and divorced/separated) and living arrangement (single, multi-person private household, and institution).
      In bivariate analyses, we analyzed the distribution of the place of death according to the cause of death and age. Because of the confounding effect of age with place of death in many associations, the crude percentages of the distribution of the place of death with respect to sex, nationality, SES of the district of residence, marital status, and living arrangement were standardized for age, using direct standardization. Associations were considered significant when P<0.01 for Pearson's Chi square.
      In the multivariate analysis, we first analyzed the determinants of home death vs. hospital death of patients residing at home (i.e., care home residents and care home deaths were left out of the analysis). Subsequently, we analyzed separately the place of death of the care home residents to determine the factors associated with care home death vs. hospital death. The small number of persons living at home but dying in a care home and those living in a care home but dying at home were not included in the analyses. The independent variables were considered to have a significant association with the outcome variable if P<0.05 for the likelihood ratio Chi square in a forward stepwise logistic regression. SPSS 15.0 (SPSS Inc., Chicago, IL) was used for all statistical computations.

      Results

      Place of Death Related to Illness and Age

      Of all 10,108 deaths in Brussels in 2003, 18.4% died at home, 55.6% in a hospital, 24.6% in a care home, and 1.4% elsewhere. In total, 3672 deaths (36.2% of all deaths) were from a chronic disease amenable to palliative care. Of this subset, 15.1% died at home, 63.0% in a hospital, and 21.6% in a care home (Table 1).
      Table 1Place of Death by Cause of Death and Age for Death from Chronic Disease in the Brussels Metropolitan Region, 2003
      Cause of Death and AgeNo. of Deaths (% of Deaths from Chronic Disease)% Occurring at Home% Occurring in Hospital% Occurring in Care Home% Occurring Elsewhere
      “Elsewhere” includes all other places of death (e.g., public road, workplace).
      Deaths after chronic disease367215.163.021.60.2
      Cause of death
      P<0.001 for Chi-square test for association between age and place of death and between cause of death and place of death (among and within cause of death groups).
       Died of cancer condition2431 (66.2%)15.772.411.60.3
       Died of noncancer condition1241 (33.8%)13.944.641.30.2
      Heart failure393 (10.7%)17.835.147.10.0
      Renal failure94 (2.6%)5.363.829.81.1
      Liver failure10 (0.3%)40.060.00.00.0
      COPD484 (13.2%)13.057.928.90.2
      Motor neuron disease20 (0.5%)15.045.040.00.0
      Parkinson's disease66 (1.8%)19.728.851.50.0
      Huntington's disease4 (0.1%)0.050.050.00.0
      Alzheimer's disease155 (4.2%)9.016.874.20.0
      HIV/AIDS15 (0.4%)6.793.30.00.0
      Age
      P<0.001 for Chi-square test for association between age and place of death and between cause of death and place of death (among and within cause of death groups).
      (years)
       <65761 (20.7%)17.978.83.20.1
       65–801511 (41.1%)16.769.613.30.3
       >801400 (38.1%)11.847.440.60.2
      COPD=chronic obstructive pulmonary disease; HIV/AIDS=human immunodeficiency virus/acquired immunodeficiency virus.
      Presented percentages are row percentages except those in brackets (column percentages).
      a P<0.001 for Chi-square test for association between age and place of death and between cause of death and place of death (among and within cause of death groups).
      b “Elsewhere” includes all other places of death (e.g., public road, workplace).
      Two-thirds of patients who died of a chronic disease died of cancer. Chronic obstructive pulmonary disease (13.2%) and heart failure (10.7%) were the most prevalent noncancer causes of death. Of all cancer patients, 72.4% died in hospital, which is more than that in any other disease. Of all noncancer patients, 41.3% died in a care home. The proportion of home deaths was larger in deaths after liver failure, Parkinson's disease, and heart failure than other causes of death. Death from a neurodegenerative disease (Alzheimer's disease, Parkinson's disease, and motor neuron disease) and from heart failure occurred more in a care home.
      Nearly 80% of deaths from chronic diseases were in people 65 years or older, and more than 38% were in those older than 80 years. Younger people died more often at home or in hospital. About 40% of people over 80 years died in a care home (Table 1).

      Place of Death Related to Personal Characteristics

      Standardized for age, there were significant differences in place of death by sex, nationality, and SES of the district of residence (Table 2). Men died more often at home or in hospital, and women died more often in a care home. Of all nationalities, North Africans and Turks more often died at home. Belgians and Northern and Western Europeans died more often in a care home, and Southern Europeans and the people from other nationalities died more often in hospital. People living in communities with a high or (in particular) middle-high SES died less often in hospital. People living in communities with low SES more often died in hospital.
      Table 2Place of Death by Sociodemographic Characteristics for Death from Chronic Disease in the Brussels Metropolitan Region, 2003, Standardized for Age
      Social SupportNo. of Deaths (% of Deaths from Chronic Disease)% Occurring at Home% Occurring in Hospital% Occurring in Care Home% Occurring Elsewhere
      “Elsewhere” includes all other places of death (e.g., public road, workplace).
      Sex
      P<0.001 for Chi-square test for association between sex and place of death and nationality and place of death.
       Male1738 (47.3%)16.666.316.90.1
       Female1934 (52.7%)14.161.224.40.3
      Nationality
      P<0.001 for Chi-square test for association between sex and place of death and nationality and place of death.
       Belgians3308 (90.1%)15.262.821.80.2
       Northern and Western Europeans111 (3.0%)15.758.724.70.9
       Southern Europeans126 (3.4%)14.166.816.92.2
       Northern Africans and Turks72 (2.0%)22.559.113.64.8
       Other nationalities54 (1.5%)17.671.010.21.2
      SES of the district of residence
      P<0.01 for Chi-square test for association between SES of the district of residence and place of death.
       Low SES728 (19.8%)11.466.522.00.1
       Middle-low SES1708 (46.5%)14.564.920.30.3
       Middle-high SES355 (9.7%)20.056.723.00.3
       High SES881 (24.0%)16.860.122.80.2
      Presented percentages are row percentages except those in brackets (column percentages).
      a P<0.001 for Chi-square test for association between sex and place of death and nationality and place of death.
      b P<0.01 for Chi-square test for association between SES of the district of residence and place of death.
      c “Elsewhere” includes all other places of death (e.g., public road, workplace).

      Place of Death Related to Social Support

      Married people more often died at home and less often died in a care home than people who were not or no longer married (Table 3). People living in a multi-person private household more often died at home than those living alone. The crude percentages (i.e., not standardized for age) show that 23.8% of people who resided in a care home died in a hospital (not shown in the table).
      Table 3Place of Death by Social Support for Death from Chronic Disease in the Brussels Metropolitan Region, 2003, Standardized for Age
      Social SupportNo. of Deaths (% of Deaths from Chronic Disease)% Occurring at Home% Occurring in Hospital% Occurring in Care Home% Occurring Elsewhere
      “Elsewhere” includes all other places of death (e.g., public road, workplace).
      Marital status
      P<0.001 for Chi-square test for association between marital status and place of death and living arrangement and place of death.
       Unmarried409 (11.1%)9.763.625.61.1
       Married1499 (40.9%)20.466.512.90.1
       Widowed1341 (36.5%)14.459.226.10.2
       Divorced/separated420 (11.4%)13.063.223.40.4
      Living arrangement
      P<0.001 for Chi-square test for association between marital status and place of death and living arrangement and place of death.
       Single1169 (32.2%)14.873.611.50.1
       Multi-person private household1741 (48.0%)21.069.49.40.2
       Institution715 (19.7%)2.928.068.70.4
      Presented percentages are row percentages except those in brackets (column percentages).
      a P<0.001 for Chi-square test for association between marital status and place of death and living arrangement and place of death.
      b “Elsewhere” includes all other places of death (e.g., public road, workplace).

      Multivariate Analysis

      We analyzed the determinants of the place of death for patients residing at home and those residing in a care home separately. Both subsamples (n=2553 and n=679) differ significantly for age, sex, nationality, SES of the district of residence and marital status (P<0.001) (data not shown in the table).

      Factors Associated with Home Death

      With the care home residents left out of the multivariate binary logistic regression analysis, home death (vs. hospital death) was associated with illness, personal characteristics, and social support (Table 4). People living in a district with middle-high SES had more than twice the chance of dying at home compared with people living in communities with low or middle-low SES (odds ratio [OR]=2.21). Noncancer patients had better chances of dying at home than cancer patients (OR=1.61). People living in a multi-person private household were more likely to die at home than those living alone (OR=1.53).
      Table 4Adjusted ORs for Home vs. Hospital Death and Care Home vs. Hospital Death for Death from Chronic Disease in the Brussels Metropolitan Region, 2003
      The variables, sex and nationality, were entered in the regression but were found not to be significant, and consequently, eliminated by the stepwise procedure.
      ,
      A problem of multicollinearity between living arrangement and marital status made us omit the latter in the home vs. hospital model. Marital status was not significant in the care home vs. hospital model.
      Home ResidentsCare Home Residents
      Home vs. Hospital Death (n=2553)Care Home vs. Hospital Death (n=679)
      AOR95% CI for AORAOR95% CI for AOR
      Cause of death
       Died of cancer condition1.001.00
       Died of noncancer condition1.611.30, 2.001.751.21. 2.53
      Age (continuous)NS1.041.01, 1.06
      SES of the district of residence
       Low SES1.001.00
       Middle-low SES1.300.99, 1.720.990.62, 1.59
       Middle-high SES2.211.51, 3.231.220.65, 2.27
       High SES1.601.18, 2.171.941.11, 3.37
      Living arrangement
       Single1.00
      Was not entered in the regression comparing care home and hospital death, because living environment was constant (i.e., living arrangement of all patients included in the regression analysis was institutional).
      Was not entered in the regression comparing care home and hospital death, because living environment was constant (i.e., living arrangement of all patients included in the regression analysis was institutional).
       Multi-person private household1.531.24, 1.88
      Was not entered in the regression comparing care home and hospital death, because living environment was constant (i.e., living arrangement of all patients included in the regression analysis was institutional).
      Was not entered in the regression comparing care home and hospital death, because living environment was constant (i.e., living arrangement of all patients included in the regression analysis was institutional).
      AOR=adjusted odds ratio; CI=confidence interval; NS=not significant.
      Presented figures are AORs and 95% CIs.
      a The variables, sex and nationality, were entered in the regression but were found not to be significant, and consequently, eliminated by the stepwise procedure.
      b A problem of multicollinearity between living arrangement and marital status made us omit the latter in the home vs. hospital model. Marital status was not significant in the care home vs. hospital model.
      c Was not entered in the regression comparing care home and hospital death, because living environment was constant (i.e., living arrangement of all patients included in the regression analysis was institutional).

      Factors Associated with Care Home Death

      Place of death of care home residents was also associated with illness and personal factors (Table 4). Again, the SES of the community of residence was most predictive for care home death. People residing in a care home in communities with a high SES were almost twice as likely to die in the care home compared with care home residents in communities with lower SES (OR=1.94). Chronically ill noncancer patients were more likely to die in the care home (instead of in a hospital) than cancer patients (OR=1.75). Finally, the likelihood of dying in the care home increased with age.

      Discussion

      In the Brussels metropolitan region, 15% of people with a chronic disease died at home, 63% in a hospital, and 22% in a care home. Of the care home residents, 24% died in a hospital. For patients residing at home, the place of death depended on the SES of the district of residence, illness, and living arrangement. For care home residents, place of death depended on the SES of the district of residence, illness, and age.
      For the first time, this study analyzed the determinants of place of death in a continental metropolitan region. Only the chronically ill population was selected, because it could potentially benefit from receiving palliative care, and hence was more amenable to planning and organization of end-of-life care. Hospital death was analyzed not only in relation to home death for the subsample of people residing at home, but also in relation to care home death for the sociodemographically distinct subsample of people residing in a care home (for which the care home can, hence, be seen as home replacement). Finally, this study described patterns for a whole population (and not just a sample), resulting in more statistical power.
      This study also has some limitations, which are inherent in the use of death certificate data.
      • DesHarnais S.
      • Carter R.E.
      • Hennessy W.
      • Kurent J.E.
      • Carter C.
      Lack of concordance between physician and patient: reports on end-of-life care discussions.
      • Cardenas-Turanzas M.
      • Carrillo M.T.
      • Tovalin-Ahumada H.
      • Elting L.
      Factors associated with place of death of cancer patients in the Mexico City Metropolitan area.
      • Cohen J.
      • Bilsen J.
      • Hooft P.
      • et al.
      Dying at home or in an institution using death certificates to explore the factors associated with place of death.
      • Bruera E.
      • Sweeney C.
      • Russell N.
      • Willey J.S.
      • Palmer J.L.
      Place of death of Houston area residents with cancer over a two-year period.
      • Polissar L.
      • Severson R.K.
      • Brown N.K.
      Factors affecting place of death in Washington State, 1968-1981.
      • Cohen J.
      • Bilsen J.
      • Miccinesi G.
      • et al.
      Using death certificate data to study place of death in 9 European countries: opportunities and weaknesses.
      • Gruneir A.
      • Mor V.
      • Weitzen S.
      • et al.
      Where people die: a multilevel approach to understanding influences on site of death in America.
      Death certificates provide no information about a number of factors that, evidence shows, affect the place of death and that were, therefore, not involved in our analysis, for example, length of disease, functional status, patient and caregiver preferences, ethnic background, availability and quality of home care, and previous admissions to hospital.
      • Gomes B.
      • Higginson I.J.
      Factors influencing death at home in terminally ill patients with cancer: systematic review.
      Another limitation related to the use of death certificate data is the sometimes-unreliable classification of causes of death on the death certificate. A third limitation regards the use of nationality as a proxy variable for ethnic background, knowing that in recent years, many immigrants obtained Belgian nationality and were considered Belgian in our analysis. Finally, partly because of the many missing values for education on an individual level, ecological variables were used for socioeconomic status.
      The proportion of patients eligible for palliative care who died at home in Brussels is very low. Of the cancer patients, only 15.7% died at home (and 72.4% in hospital) in Brussels, whereas a previous study showed that in Flanders, the northern region of Belgium, 29.1% of cancer patients died at home (and 61.9% in a hospital).
      • Cohen J.
      • Bilsen J.
      • Hooft P.
      • et al.
      Dying at home or in an institution using death certificates to explore the factors associated with place of death.
      We assume these findings to be related both to the demographic and social reality of Brussels, which has a poorer, more socially fragmented and older population than Flanders, as well as to the way palliative care in Brussels is organized, with a central role for the family physician. The lower use of primary care in Brussels (compared, for instance, with Flanders) probably helps to explain the lower proportion of home deaths. In Brussels, 83% of families have a family physician, compared with 97% in Flanders, and the average number of contacts with the family physician is 3.7 per year in Brussels as against 4.6 in Flanders.

      Belgian Scientific Institute of Public Health, Department of Epidemiology. Health Survey Belgium 2004 [in Dutch]. Brussels: Belgian Scientific Institute of Public Health, 2006.

      Not surprisingly, in 2003, the proportion of Brussels patients receiving compensation for the cost of palliative home care was four times lower and the proportion applying for compassionate leave to provide care to a relative five times lower than that in Flanders.

      National Institute for Health and Disability Insurance Belgium and National Employment Office. National Insitute for Health and Disability Insurance Belgium (www.riziv.fgov.be) and National Employment Office Belgium (www.rva.be).

      On the other hand, the number of beds in specialized palliative care units attached to hospitals in 2003 was 5.4 per 1000 deaths in Brussels against 3.2 in Flanders.

      National Institute for Health and Disability Insurance Belgium. National Institute for Health and Disability Insurance Belgium (www.riziv.fgov.be).

      Probably, the less prominent role of primary care and the more prominent role of inpatient palliative care can partially explain the higher number of hospital deaths and the lower number of home deaths.
      Another striking finding was that, although most of Brussels' chronically ill care home residents die in the care home where they are living, one of four is still transferred to a hospital at the end of life. This seems to be comparable to British Columbia, Canada,
      • McGregor M.J.
      • Tate R.B.
      • Ronald L.A.
      • McGrail K.M.
      Variation in site of death among nursing home residents in British Columbia, Canada.
      where it is slightly more than that in the rest of Belgium,
      • Van den Block L.
      • Deschepper R.
      • Bilsen J.
      • Van Casteren V.
      • Deliens L.
      Transitions between care settings at the end of life in Belgium.
      but considerably more than that in The Netherlands.
      • Cohen J.
      End-of-life decisions and place of death in Belgium and Europe.
      This could point to a lack of palliative orientation in some of Brussels' long-term care facilities or at least a lack of capacity in these facilities to provide adequate end-of-life care, which would prevent transfer to hospital.
      We found that residents of high-SES districts were more likely than people living in low-SES districts to die at home. This also has been found in other research.
      • Gomes B.
      • Higginson I.J.
      Factors influencing death at home in terminally ill patients with cancer: systematic review.
      • Grande G.E.
      • Addington-Hall J.M.
      • Todd C.J.
      Place of death and access to home care services: are certain patient groups at a disadvantage?.
      • Tang S.T.
      • McCorkle R.
      Determinants of place of death for terminal cancer patients.
      This is all the more striking because these differences cannot be ascribed to differences in available end-of-life care services. Brussels can be considered as a health care region (i.e., services throughout Brussels are used by people all across Brussels). Rather than differences in availability of services, other factors, such as lack of money and information or simply distrust, are likely to explain the differences in the place of death for people of lower-SES districts.
      • DesHarnais S.
      • Carter R.E.
      • Hennessy W.
      • Kurent J.E.
      • Carter C.
      Lack of concordance between physician and patient: reports on end-of-life care discussions.
      • Higginson I.J.
      • Jarman B.
      • Astin P.
      • Dolan S.
      Do social factors affect where patients die: an analysis of 10 years of cancer deaths in England.
      • Decker S.L.
      • Higginson I.J.
      A tale of two cities: factors affecting place of cancer death in London and New York.
      Residents of districts with middle-high SES (i.e., the universities' districts), who are more highly educated but have lower incomes compared with the highest-SES district, were most likely to die at home, although the difference between residents of middle-high and high-SES districts was not statistically significant in the multivariate model. Cognitive aspects (e.g., knowledge of services, communication with professional caregivers) may be more important in determining place of death than material aspects (e.g., money), as suggested by previous research.
      • Hanratty B.
      • Burstrom B.
      • Walander A.
      • Whitehead M.
      Hospital deaths in Sweden: are individual socioeconomic factors relevant?.
      Not only the chance to die at home, but also the chance to die in the care home of residence is socially unequally distributed throughout Brussels. We suspect that care homes in the high SES districts may attach more importance to quality in general and quality of end-of-life care in particular, and are possibly better equipped to manage long-term end-of-life care. This could be a matter of resources (i.e., care home residents in more prosperous districts bring in more money) and also of end-of-life care policies. It is possibly also related to the size or management of the care homes by a local authority, by a private not-for-profit organization, or by a private company.
      • McGregor M.J.
      • Tate R.B.
      • Ronald L.A.
      • McGrail K.M.
      Variation in site of death among nursing home residents in British Columbia, Canada.
      Further research is required to determine the nature of the influence of SES on place of death, both for home death and care home death, and to determine whether or not palliative care is implemented, as anticipated in Belgian law, in the same way in all types of care homes across the whole metropolitan area.
      Both home death and care home death were related to illness. In line with Australian research, where death after chronic disease was analyzed separately, death from a noncancer condition was more likely to take place at the place of usual residence, home, or care home, compared with cancer death.
      • McGregor M.J.
      • Tate R.B.
      • Ronald L.A.
      • McGrail K.M.
      Variation in site of death among nursing home residents in British Columbia, Canada.
      As cancer patients are, because of the long and predictable course of the disease, perhaps more eligible for care in the place of usual residence,
      • Cohen J.
      • Bilsen J.
      • Hooft P.
      • et al.
      Dying at home or in an institution using death certificates to explore the factors associated with place of death.
      • Grande G.E.
      • Addington-Hall J.M.
      • Todd C.J.
      Place of death and access to home care services: are certain patient groups at a disadvantage?.
      this is a remarkable finding. It could be an indication of a lack of availability of appropriate specialized palliative care for cancer patients in care homes and at home, or possibly, cancer end-of-life care services have been developed very intramural in Brussels.
      This study reconfirms the importance of social support.
      • Gomes B.
      • Higginson I.J.
      Factors influencing death at home in terminally ill patients with cancer: systematic review.
      • Grande G.E.
      • Addington-Hall J.M.
      • Todd C.J.
      Place of death and access to home care services: are certain patient groups at a disadvantage?.
      • Tang S.T.
      • McCorkle R.
      Determinants of place of death for terminal cancer patients.
      People living in a private household and supported by family members are more likely to die at home than those living alone.
      Finally, we found that with advancing age, the odds of dying in the care home of residence increase. This might suggest that a decision to transfer the patient to hospital (possibly for life-prolonging treatments) is taken less often for very old patients, but it could also suggest that different end-of-life care criteria are applied in the very old.
      The question remains of whether our findings for the Brussels metropolitan region are typical for large metropolitan areas in general. Compared with much larger cities in the world, like Houston, New York, London, and Mexico City, far fewer people with cancer in Brussels died at home and many more died in hospital,
      • Cardenas-Turanzas M.
      • Carrillo M.T.
      • Tovalin-Ahumada H.
      • Elting L.
      Factors associated with place of death of cancer patients in the Mexico City Metropolitan area.
      • Bruera E.
      • Sweeney C.
      • Russell N.
      • Willey J.S.
      • Palmer J.L.
      Place of death of Houston area residents with cancer over a two-year period.
      • Decker S.L.
      • Higginson I.J.
      A tale of two cities: factors affecting place of cancer death in London and New York.
      whereas the proportion of hospital deaths in the rest of Belgium was not higher than that in the United States.
      • Weitzen S.
      • Teno J.M.
      • Fennell M.
      • Mor V.
      Factors associated with site of death: a national study of where people die.
      Not only the metropolitan character, but probably the sociodemographic reality (e.g., people who died of cancer in Brussels were older than those in New York and London, and much less often married than those in Houston
      • Cardenas-Turanzas M.
      • Carrillo M.T.
      • Tovalin-Ahumada H.
      • Elting L.
      Factors associated with place of death of cancer patients in the Mexico City Metropolitan area.
      • Bruera E.
      • Sweeney C.
      • Russell N.
      • Willey J.S.
      • Palmer J.L.
      Place of death of Houston area residents with cancer over a two-year period.
      • Decker S.L.
      • Higginson I.J.
      A tale of two cities: factors affecting place of cancer death in London and New York.
      ) and the organization of palliative and end-of-life care in metropolitan cities may play a large role. An important question that needs to be explored is how the palliative care needs of people, specifically in metropolitan cities like Brussels, can best be met—whether by expanding the possibilities of palliative home care support, by establishing hospices, or by safeguarding good quality of end of life in palliative care units in hospital.
      Only 15.1% of all deaths after chronic disease occurred at home in Brussels in 2003. More than 20% died in a care home, but nearly 24% of care home residents were transferred to an acute hospital bed at the end. These findings contrast sharply with the rest of Belgium and even with much larger metropolitan regions in the world, and raise certain questions with regard to intra- and extramural organization of end-of-life care. Home death was related to illness, social support, and socioeconomic status of the district of residence. Care home death was determined by illness, age, and surprisingly, the SES of the district of residence. At least in some parts of Brussels, the unequal distribution of dying in familiar surroundings is open to discussion.

      Acknowledgment

      The data for this study were provided by the Brussels Health and Social Observatory.

      References

        • Clark D.
        Between hope and acceptance: the medicalisation of dying.
        BMJ. 2002; 324: 905-907
        • Steinhauser K.E.
        • Christakis N.A.
        • Clipp E.C.
        • et al.
        Factors considered important at the end of life by patients, family, physicians, and other care providers.
        JAMA. 2000; 284: 2476-2482
        • Grande G.E.
        • Farquhar M.C.
        • Barclay S.I.
        • Todd C.J.
        Caregiver bereavement outcome: relationship with hospice at home, satisfaction with care, and home death.
        J Palliat Care. 2004; 20: 69-77
        • Barbera L.
        • Paszat L.
        • Chartier C.
        Death in hospital for cancer patients: an indicator of quality of end-of-life care.
        Palliat Med. 2005; 19: 435-436
        • Bowling A.
        The hospitalisation of death: should more people die at home?.
        J Med Ethics. 1983; 9: 158-161
        • Teno J.M.
        • Clarridge B.R.
        • Casey V.
        • et al.
        Family perspectives on end-of-life care at the last place of care.
        JAMA. 2004; 291: 88-93
        • Gott M.
        • Seymour J.
        • Bellamy G.
        • Clark D.
        • Ahmedzai S.
        Older people's views about home as a place of care at the end of life.
        Palliat Med. 2004; 18: 460-467
        • Foreman L.M.
        • Hunt R.W.
        • Luke C.G.
        • Roder D.M.
        Factors predictive of preferred place of death in the general population of South Australia.
        Palliat Med. 2006; 20: 447-453
        • Beccaro M.
        • Costantini M.
        • Rossi P.G.
        • et al.
        Actual and preferred place of death of cancer patients. Results from the Italian survey of the dying of cancer (ISDOC).
        J Epidemiol Community Health. 2006; 60: 412-416
        • Tang S.T.
        When death is imminent—where terminally ill patients with cancer prefer to die and why.
        Cancer Nurs. 2003; 26: 245-251
        • Tang S.T.
        • McCorkle R.
        Determinants of congruence between the preferred and actual place of death for terminally ill cancer patients.
        J Palliat Care. 2003; 19: 230-237
      1. Cohen, J, Bilsen J, Addington-Hall J, et al. Population-based study of dying in hospital in six European countries. Palliat Med 2008;22(6):702–710.

        • Hearn J.
        • Higginson I.J.
        Do specialist palliative care teams improve outcomes for cancer patients? A systematic literature review.
        Palliat Med. 1998; 12: 317-332
        • Burke K.
        Palliative care at home to get further funds if it saves money.
        BMJ. 2004; 328: 544
        • Chochinov H.M.
        • Kristjanson L.
        Dying to pay: the cost of end-of-life care.
        J Palliat Care. 1998; 14: 5-15
        • Miller S.C.
        • Intrator O.
        • Gozalo P.
        • et al.
        Government expenditures at the end of life for short- and long-stay nursing home residents: differences by hospice enrollment status.
        J Am Geriatr Soc. 2004; 52: 1284-1292
      2. Brussels Health and Social Observatory. Atlas of Welfare and Health of Brussels Capital Region 2006 [in Dutch]. Brussels: Common Community Commission, 2006.

      3. Brussels Health and Social Observatory. Barometer of Welfare of Brussels Capital Region 2007 [in Dutch]. Brussels: Common Community Commission, 2007.

        • Hughes A.
        Poverty and palliative care in the US: issues facing the urban poor.
        Int J Palliat Nurs. 2005; 11: 6-13
        • Murphy P.A.
        • Price D.M.
        Dying and grieving in the inner city.
        in: Doka K.J. Davidson J.D. Living with grief. Hospice Foundation of America, Washington1998: 113-120
        • Zallman L.
        • Sanchez J.
        • Choi H.
        • Selwyn P.
        Unmet need for chronic disease and end-of-life care at urban family health centers in the Bronx, NY.
        Fam Med. 2003; 35: 637-642
        • Born W.
        • Greiner K.A.
        • Sylvia E.
        • Butler J.
        • Ahluwalia J.S.
        Knowledge, attitudes, and beliefs about end-of-life care among inner-city African Americans and Latinos.
        J Palliat Med. 2004; 7: 247-256
        • Koffman J.
        • Burke G.
        • Dias A.
        • et al.
        Demographic factors and awareness of palliative care and related services.
        Palliat Med. 2007; 21: 145-153
        • Mor V.
        • Hiris J.
        Determinants of site of death among hospice cancer patients.
        J Health Soc Behav. 1983; 24: 375-385
        • Gomes B.
        • Higginson I.J.
        Factors influencing death at home in terminally ill patients with cancer: systematic review.
        BMJ. 2006; 332: 515-518
        • Rosenwax L.K.
        • McNamara B.
        • Blackmore A.M.
        • Holman C.D.
        Estimating the size of a potential palliative care population.
        Palliat Med. 2005; 19: 556-562
        • McNamara B.
        • Rosenwax L.
        Factors affecting place of death in Western Australia.
        Health Place. 2007; 13: 356-367
        • DesHarnais S.
        • Carter R.E.
        • Hennessy W.
        • Kurent J.E.
        • Carter C.
        Lack of concordance between physician and patient: reports on end-of-life care discussions.
        J Palliat Med. 2007; 10: 728-740
        • Cardenas-Turanzas M.
        • Carrillo M.T.
        • Tovalin-Ahumada H.
        • Elting L.
        Factors associated with place of death of cancer patients in the Mexico City Metropolitan area.
        Support Care Cancer. 2007; 15: 243-249
        • Cohen J.
        • Bilsen J.
        • Hooft P.
        • et al.
        Dying at home or in an institution using death certificates to explore the factors associated with place of death.
        Health Policy. 2006; 78: 319-329
        • Bruera E.
        • Sweeney C.
        • Russell N.
        • Willey J.S.
        • Palmer J.L.
        Place of death of Houston area residents with cancer over a two-year period.
        J Pain Symptom Manage. 2003; 26: 637-643
        • Polissar L.
        • Severson R.K.
        • Brown N.K.
        Factors affecting place of death in Washington State, 1968-1981.
        J Community Health. 1987; 12: 40-55
        • Cohen J.
        • Bilsen J.
        • Miccinesi G.
        • et al.
        Using death certificate data to study place of death in 9 European countries: opportunities and weaknesses.
        BMC Public Health. 2007; 7: 283
        • Gruneir A.
        • Mor V.
        • Weitzen S.
        • et al.
        Where people die: a multilevel approach to understanding influences on site of death in America.
        Med Care Res Rev. 2007; 64: 351-378
      4. Belgian Scientific Institute of Public Health, Department of Epidemiology. Health Survey Belgium 2004 [in Dutch]. Brussels: Belgian Scientific Institute of Public Health, 2006.

      5. National Institute for Health and Disability Insurance Belgium and National Employment Office. National Insitute for Health and Disability Insurance Belgium (www.riziv.fgov.be) and National Employment Office Belgium (www.rva.be).

      6. National Institute for Health and Disability Insurance Belgium. National Institute for Health and Disability Insurance Belgium (www.riziv.fgov.be).

        • McGregor M.J.
        • Tate R.B.
        • Ronald L.A.
        • McGrail K.M.
        Variation in site of death among nursing home residents in British Columbia, Canada.
        J Palliat Med. 2007; 10: 1128-1136
        • Van den Block L.
        • Deschepper R.
        • Bilsen J.
        • Van Casteren V.
        • Deliens L.
        Transitions between care settings at the end of life in Belgium.
        JAMA. 2007; 298: 1638-1639
        • Cohen J.
        End-of-life decisions and place of death in Belgium and Europe.
        VUBPRESS, Brussels2007
        • Grande G.E.
        • Addington-Hall J.M.
        • Todd C.J.
        Place of death and access to home care services: are certain patient groups at a disadvantage?.
        Soc Sci Med. 1998; 47: 565-579
        • Tang S.T.
        • McCorkle R.
        Determinants of place of death for terminal cancer patients.
        Cancer Invest. 2001; 19: 165-180
        • Higginson I.J.
        • Jarman B.
        • Astin P.
        • Dolan S.
        Do social factors affect where patients die: an analysis of 10 years of cancer deaths in England.
        J Public Health Med. 1999; 21: 22-28
        • Decker S.L.
        • Higginson I.J.
        A tale of two cities: factors affecting place of cancer death in London and New York.
        Eur J Public Health. 2007; 17: 285-290
        • Hanratty B.
        • Burstrom B.
        • Walander A.
        • Whitehead M.
        Hospital deaths in Sweden: are individual socioeconomic factors relevant?.
        J Pain Symptom Manage. 2007; 33: 317-323
        • Weitzen S.
        • Teno J.M.
        • Fennell M.
        • Mor V.
        Factors associated with site of death: a national study of where people die.
        Med Care. 2003; 41: 323-335