[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 35.175.191.168. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Figure 1.
Proportion of Decedents Dying in Intensive Carea
Proportion of Decedents Dying in Intensive Carea

Forest plot depicting the ratio of the proportions of decedents dying in intensive care units comparing recent immigrant with long-standing resident cohorts (relative risk). Relative risks greater than 1 correspond to an increased relative risk of recent immigrant compared with long-standing resident decedents being in the intensive care unit at death. The size of each square is proportional to the precision of the relative risk estimate.

aThe denominator for each row is the total number of decedents in each cohort in each subgroup by row, ie, the denominator for each cell in the corresponding cell in Table 1.

bDenotes the extent to which an area is urbanized, with “strong” being the most urbanized.

Figure 2.
Proportion of Decedents Dying in Intensive Care: Recent Immigrant Characteristicsa
Proportion of Decedents Dying in Intensive Care: Recent Immigrant Characteristicsa

Forest plot analogous to Figure 1 depicting the ratio of the proportions of decedents dying in intensive care units comparing recent immigrant with long-standing resident cohorts (relative risk). In contrast to Figure 1, this figure focuses on subgroups defined only among the recent immigrant cohort including region of origin, language ability on arrival, education level on arrival, immigration class, and time since immigration. The proportion of recent immigrant decedents dying in intensive care within each subgroup is compared with the proportion of long-standing resident decedents dying in intensive care (92 270 of 919 499 [10%]), and so a relative risk greater than 1 corresponds to an increased relative risk of recent immigrant compared with long-standing resident decedents being in the intensive care unit at death. Note that percentages are based on the size of each subgroup by row, not based on the overall analytic sample size for recent immigrant decedents of 47 514. The size of each square is proportional to the precision of the relative risk estimate.

aThe denominator data for the percentages are the total numbers of recent immigrant decedents in each subgroup, ie, the corresponding cells in eTable 3 in the Supplement.

Table 1.  
Baseline Characteristics of Deceased Patients (N = 967 013)a
Baseline Characteristics of Deceased Patients (N = 967 013)a
Table 2.  
Location of Death and Care Received in the Final 6 Months of Lifea
Location of Death and Care Received in the Final 6 Months of Lifea
Table 3.  
Modified Poisson Regression for Relative Risk of Each Location of Death (N = 967 013)a
Modified Poisson Regression for Relative Risk of Each Location of Death (N = 967 013)a
1.
World Health Organization. WHO definition of palliative care. http://www.who.int/cancer/palliative/definition/en/. Accessed March 21, 2016.
2.
Heyland  DK, Dodek  P, Rocker  G,  et al; Canadian Researchers End-of-Life Network(CARENET).  What matters most in end-of-life care: perceptions of seriously ill patients and their family members.  CMAJ. 2006;174(5):627-633.PubMedGoogle ScholarCrossref
3.
Heyland  DK, Cook  DJ, Rocker  GM,  et al; Canadian Researchers at the End of Life Network (CARENET).  Defining priorities for improving end-of-life care in Canada.  CMAJ. 2010;182(16):E747-E752.PubMedGoogle ScholarCrossref
4.
Frost  DW, Cook  DJ, Heyland  DK, Fowler  RA.  Patient and healthcare professional factors influencing end-of-life decision-making during critical illness: a systematic review.  Crit Care Med. 2011;39(5):1174-1189.PubMedGoogle ScholarCrossref
5.
Central Intelligence Agency. The World Factbook: country comparison: net migration rate. https://www.cia.gov/library/publications/the-world-factbook/rankorder/2112rank.html. Accessed March 21, 2016.
6.
Statistics Canada. Population and dwelling count highlight tables, 2011 census. http://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=101&S=50&O=A. Published January 7, 2016. Accessed March 21, 2016.
7.
Fuller-Thomson  E, Noack  AM, George  U.  Health decline among recent immigrants to Canada: findings from a nationally-representative longitudinal survey.  Can J Public Health. 2011;102(4):273-280.PubMedGoogle Scholar
8.
Sanmartin  C, Ross  N.  Experiencing difficulties accessing first-contact health services in Canada: Canadians without regular doctors and recent immigrants have difficulties accessing first-contact healthcare services: reports of difficulties in accessing care vary by age, sex and region.  Healthc Policy. 2006;1(2):103-119.PubMedGoogle Scholar
9.
Tu  JV, Chu  A, Rezai  MR,  et al.  The Incidence of Major Cardiovascular Events in Immigrants to Ontario, Canada: the CANHEART Immigrant Study [published online August 31, 2015].  Circulation.PubMedGoogle Scholar
10.
Lackan  NA, Eschbach  K, Stimpson  JP, Freeman  JL, Goodwin  JS.  Ethnic differences in hospital place of death among older adults in California.  Med Care. 2009;47:138-145.PubMedGoogle ScholarCrossref
11.
Bosma  H, Apland  L, Kazanjian  A.  Cultural conceptualizations of hospice palliative care: more similarities than differences.  Palliat Med. 2010;24(5):510-522.PubMedGoogle ScholarCrossref
12.
Unroe  KT, Greiner  MA, Johnson  KS, Curtis  LH, Setoguchi  S.  Racial differences in hospice use and patterns of care after enrollment in hospice among Medicare beneficiaries with heart failure.  Am Heart J. 2012;163(6):987-993.e3.PubMedGoogle ScholarCrossref
13.
Cruz-Oliver  DM, Talamantes  M, Sanchez-Reilly  S.  What evidence is available on end-of-life (EOL) care and Latino elders? a literature review.  Am J Hosp Palliat Care. 2014;31(1):87-97.PubMedGoogle ScholarCrossref
14.
Hanchate  A, Kronman  AC, Young-Xu  Y, Ash  AS, Emanuel  E.  Racial and ethnic differences in end-of-life costs: why do minorities cost more than whites?  Arch Intern Med. 2009;169(5):493-501.PubMedGoogle ScholarCrossref
15.
Smith  AK, Sudore  RL, Pérez-Stable  EJ.  Palliative care for Latino patients and their families: whenever we prayed, she wept.  JAMA. 2009;301(10):1047-1057, E1.PubMedGoogle ScholarCrossref
16.
Chiu  M, Lebenbaum  M, Lam  K,  et al.  Describing the linkages of the immigration, refugees and citizenship Canada permanent resident data and vital statistics death registry to Ontario’s administrative health database.  BMC Med Inform Decis Mak. 2016;16(1):135.PubMedGoogle ScholarCrossref
17.
Juurlink  D, Preyra  C, Croxford  R,  et al.  Canadian Institute for Health Information Discharge Abstract Database: A Validation Study. Toronto, ON, Canada: Institute for Clinical Evaluative Sciences; 2006.
18.
World Health Organization. ICD-10 version: 2016. http://apps.who.int/classifications/icd10/browse/2016/en. Accessed March 21, 2016.
19.
Glazier  RH, Zagorski  BM, Rayner  J.  Comparison of Primary Care Models in Ontario by Demographics, Case Mix and Emergency Department Use, 2008/09 to 2009/10. Toronto, ON, Canada: Institute for Clinical Evaluative Sciences; 2012: 41.
20.
Yelland  LN, Salter  AB, Ryan  P.  Performance of the modified Poisson regression approach for estimating relative risks from clustered prospective data.  Am J Epidemiol. 2011;174(8):984-992.PubMedGoogle ScholarCrossref
21.
Adames  HY, Chavez-Dueñas  NY, Fuentes  MA, Salas  SP, Perez-Chavez  JG.  Integration of Latino/a cultural values into palliative health care: a culture centered model.  Palliat Support Care. 2014;12(2):149-157.PubMedGoogle ScholarCrossref
22.
Ebrahim  S, Bance  S, Bowman  KW.  Sikh perspectives towards death and end-of-life care.  J Palliat Care. 2011;27(2):170-174.PubMedGoogle Scholar
23.
Dobbs  D, Park  NS, Jang  Y, Meng  H.  Awareness and completion of advance directives in older Korean-American adults.  J Am Geriatr Soc. 2015;63(3):565-570.PubMedGoogle ScholarCrossref
24.
Garrido  MM, Harrington  ST, Prigerson  HG.  End-of-life treatment preferences: a key to reducing ethnic/racial disparities in advance care planning?  Cancer. 2014;120(24):3981-3986.PubMedGoogle ScholarCrossref
25.
Higginson  IJ, Gomes  B, Calanzani  N,  et al; Project PRISMA.  Priorities for treatment, care and information if faced with serious illness: a comparative population-based survey in seven European countries.  Palliat Med. 2014;28(2):101-110.PubMedGoogle ScholarCrossref
26.
Sprung  CL, Cohen  SL, Sjokvist  P,  et al; Ethicus Study Group.  End-of-life practices in European intensive care units: the Ethicus Study.  JAMA. 2003;290(6):790-797.PubMedGoogle ScholarCrossref
27.
Phua  J, Joynt  GM, Nishimura  M,  et al; ACME Study Investigators and the Asian Critical Care Clinical Trials Group.  Withholding and withdrawal of life-sustaining treatments in intensive care units in Asia.  JAMA Intern Med. 2015;175(3):363-371.PubMedGoogle ScholarCrossref
28.
Moss  KO, Williams  IC.  End-of-life preferences in Afro-Caribbean older adults: a systematic literature review.  Omega (Westport). 2014;69(3):271-282.PubMedGoogle ScholarCrossref
29.
Sharma  RK, Khosla  N, Tulsky  JA, Carrese  JA.  Traditional expectations versus US realities: first- and second-generation Asian Indian perspectives on end-of-life care.  J Gen Intern Med. 2012;27(3):311-317.PubMedGoogle ScholarCrossref
30.
Wright  AA, Stieglitz  H, Kupersztoch  YM,  et al.  United states acculturation and cancer patients’ end-of-life care.  PLoS One. 2013;8(3):e58663.PubMedGoogle ScholarCrossref
31.
Jonnalagadda  S, Lin  JJ, Nelson  JE,  et al.  Racial and ethnic differences in beliefs about lung cancer care.  Chest. 2012;142(5):1251-1258.PubMedGoogle ScholarCrossref
32.
Kavalieratos  D, Corbelli  J, Zhang  D,  et al.  Association between palliative care and patient and caregiver outcomes: a systematic review and meta-analysis.  JAMA. 2016;316(20):2104-2114.PubMedGoogle ScholarCrossref
33.
Mack  JW, Paulk  ME, Viswanath  K, Prigerson  HG.  Racial disparities in the outcomes of communication on medical care received near death.  Arch Intern Med. 2010;170(17):1533-1540.PubMedGoogle ScholarCrossref
34.
Volandes  AE, Paasche-Orlow  M, Gillick  MR,  et al.  Health literacy not race predicts end-of-life care preferences.  J Palliat Med. 2008;11(5):754-762.PubMedGoogle ScholarCrossref
35.
White  DB, Ernecoff  N, Buddadhumaruk  P,  et al.  Prevalence of and factors related to discordance about prognosis between physicians and surrogate decision makers of critically ill patients.  JAMA. 2016;315(19):2086-2094.PubMedGoogle ScholarCrossref
36.
Baicker  K, Chandra  A, Skinner  JS.  Geographic variation in health care and the problem of measuring racial disparities.  Perspect Biol Med. 2005;48(1)(suppl):S42-S53.PubMedGoogle ScholarCrossref
37.
Bach  PB, Schrag  D, Begg  CB.  Resurrecting treatment histories of dead patients: a study design that should be laid to rest.  JAMA. 2004;292(22):2765-2770.PubMedGoogle ScholarCrossref
Original Investigation
Caring for the Critically Ill Patient
October 17, 2017

Association Between Immigrant Status and End-of-Life Care in Ontario, Canada

Author Affiliations
  • 1University of Toronto Department of Medicine, Toronto, Ontario, Canada
  • 2Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada
  • 3University of Ottawa, Ottawa, Ontario, Canada
  • 4Bruyere Research Institute, Ottawa, Ontario, Canada
  • 5Institute of Health Policy, Management and Evaluation, University of Toronto, Toronto, Ontario, Canada
  • 6Programme in Trauma, Emergency, and Critical Care, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada
  • 7Department of Medicine and Department of Critical Care Medicine, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada
  • 8Li Ka Shing Knowledge institute of St Michael’s Hospital, Toronto, Ontario, Canada
JAMA. 2017;318(15):1479-1488. doi:10.1001/jama.2017.14418
Key Points

Question  Do recent immigrant patients experience different end-of-life care than long-standing resident patients?

Findings  In this cohort study that included 967 013 patients, recent immigrant patients were more likely to be in the intensive care unit when they died and were more likely to receive invasive procedures in the last 6 months of life including hospital admission, intensive care unit admission, mechanical ventilation, tracheostomy, dialysis, or percutaneous feeding tube placement. These outcomes varied most significantly according to region of origin rather than socioeconomic position, language ability on arrival, or education level on arrival.

Meaning  Among decedents in Ontario, Canada, recent immigrants were significantly more likely to receive aggressive care and to die in an intensive care unit compared with other residents. Further research is needed to understand the mechanisms behind this association.

Abstract

Importance  People who immigrate face unique health literacy, communication, and system navigation challenges, and they may have diverse preferences that influence end-of-life care.

Objective  To examine end-of-life care provided to immigrants to Canada in the last 6 months of their life.

Design, Setting, and Participants  This population-based cohort study (April 1, 2004, to March 31, 2015) included 967 013 decedents in Ontario, Canada, using validated linkages between health and immigration databases to identify immigrant (since 1985) and long-standing resident cohorts.

Exposures  All decedents who immigrated to Canada between 1985 and 2015 were classified as recent immigrants, with subgroup analyses assessing the association of time since immigration, and region of birth, with end-of-life care.

Main Outcomes and Measures  Location of death and intensity of care received in the last 6 months of life. Analysis included modified Poisson regression with generalized estimating equations, adjusting for age, sex, socioeconomic position, causes of death, urban and rural residence, and preexisting comorbidities.

Results  Among 967 013 decedents of whom 47 514 (5%) immigrated since 1985, sex, socioeconomic status, urban (vs rural) residence, and causes of death were similar, while long-standing residents were older than immigrant decedents (median [interquartile range] age, 75 [58-84] vs 80 [68-87] years). Recent immigrant decedents were overall more likely to die in intensive care (15.6% vs 10.0%; difference, 5.6%; 95% CI, 5.2%-5.9%) after adjusting for differences in age, sex, income, geography, and cause of death (relative risk, 1.30; 95% CI, 1.27-1.32). In their last 6 months of life, recent immigrant decedents experienced more intensive care admissions (24.9% vs 19.2%; difference, 5.7%; 95% CI, 5.3%-6.1%), hospital admissions (72.1% vs 68.2%; difference, 3.9%; 95% CI, 3.5%-4.3%), mechanical ventilation (21.5% vs 13.6%; difference, 7.9%; 95% CI, 7.5%-8.3%), dialysis (5.5% vs 3.4%; difference, 2.1%; 95% CI, 1.9%-2.3%), percutaneous feeding tube placement (5.5% vs 3.0%; difference, 2.5%; 95% CI, 2.3%-2.8%), and tracheostomy (2.3% vs 1.1%; difference, 1.2%; 95% CI, 1.1%-1.4%). Relative risk of dying in intensive care for recent immigrants compared with long-standing residents varied according to recent immigrant region of birth from 0.84 (95% CI, 0.74-0.95) among those born in Northern and Western Europe to 1.96 (95% CI, 1.89-2.05) among those born in South Asia.

Conclusions and Relevance  Among decedents in Ontario, Canada, recent immigrants were significantly more likely to receive aggressive care and to die in an intensive care unit compared with other residents. Further research is needed to understand the mechanisms behind this association.

Introduction

Optimal end-of-life care combines best medical therapy and symptom control in accordance with patient wishes.1 Many patients receive aggressive in-hospital end-of-life care despite a preference for dying in a familiar setting, free from invasive technology.2 This discrepancy has many contributors, including inadequate advance care planning, suboptimal communication between health care professionals and patients or their families, poor health literacy, uncertainty about imminence of death, and unavailability of nonintensive end-of-life or palliative care resources.3,4

Canada has high rates of immigration relative to many high-income nations, which in turn leads to a diversity of geographic, cultural, and racial/ethnic backgrounds among its residents.5 In 2011 (midway through this study), Ontario, Canada, had a population of 12 851 821, of whom 3 611 365 (29%) were born in other countries and 501 060 (4%) arrived in Canada between 2006 and 2011.6 Immigrants often face challenges in communication, health literacy, and navigation of the health care system.7-9 Although immigrants are on average healthier than age-matched Canadians when they arrive in Canada, they subsequently experience excess morbidity and mortality from chronic medical and psychiatric conditions.7

Preliminary evidence suggests that some immigrants may face cultural and logistical challenges in end-of-life care due to decreased health literacy or language ability, different modes of family-based decision-making and filial responsibility, and decreased access to care due to insufficient financial and social resources.10-15 Some immigrants may have different end-of-life care preferences than many long-standing residents.15 To our knowledge, there are no comprehensive large-scale quantitative studies of end-of-life care in recent immigrant populations. This population-based analysis was conducted to describe end-of-life care delivered to recently immigrated compared with long-standing resident decedents, including magnitude of and factors associated with differences in end-of-life care.

Methods
Study Setting and Oversight

The study was approved by the research ethics board of Sunnybrook Health Sciences Centre including a waiver for individual patient consent because the data sets were linked using unique encoded identifiers and analyzed at the Institute for Clinical Evaluative Sciences.

Identification of Decedents (Study Participants)

All individuals (recent immigrants and long-standing residents) who died in Ontario between April 1, 2004, and March 31, 2015, were identified. Individuals with fewer than 6 months of enrollment in the provincial health care plan were excluded. Data on individuals who received care in Ontario but died in another country or province were not available. Patients with some missing baseline data were included in unadjusted analyses but not in adjusted analyses requiring a missing variable.

Recent immigrants were identified within the data set through previously validated combined probabilistic and deterministic linkage of the list of deceased individuals to the registry of landed immigrants maintained by Immigration, Refugees and Citizenship Canada.16 Deterministic linkage occurs when 1 or more identifiers (eg, health card number and name) are identical, while probabilistic linkage uses probability scores to identify linkages among records where deterministic linkages were not possible. Recent immigrants were defined as those granted permanent residency or citizenship status in Canada between 1985 and 2015 (the years available in the Immigration, Refugees and Citizenship Canada data) and created subgroups according to duration since immigration. All other residents were defined as long-standing residents. Other research has reserved the term recent for immigrants arriving within shorter timeframes, but this broader definition sought to include all available data and acknowledge that some members of the long-standing resident cohort may also be immigrants but have lived in Canada for more than 30 years. Immigration, Refugees and Citizenship Canada data also included information obtained at the time of immigration application on immigration class (economic, family, refugee, and other), education level, language ability, and country of birth. Information on the level of health literacy, religion, and specific cultural practices was not available.

Identification of Health Care Use Prior to Death

A combination of health administrative databases linked at the individual level were used to describe health care service use at the population level in Ontario. These included the Registered Persons Database containing vital statistics on all persons issued a Provincial Health Card, the Ontario Health Insurance Plan containing data on all professional services and procedures resulting in charges to the health care system, the Office of the Registrar General for Deaths, the Discharge Abstract Database containing detailed patient-level information including resources used and procedures performed for all inpatients, and the National Ambulatory Care Reporting System for similar data among ambulatory hospital admissions and emergency department presentations.17,18 Although these databases do not include care provided at community health centers (frequented by some recent immigrants, but overall reaching less than 1% of the population), the databases contain comprehensive coverage of care provided in hospitals.19

Characteristics of Patients

Patient characteristics and demographics are reported including age, sex, socioeconomic position based on postal code census data, and place of residence at time of death. Data are reported on intensive care admissions, chronic conditions including the Charlson Comorbidity Index (Deyo modification), hospitalizations, procedures, and emergency department visits during the final 24 months of life, with emphasis on the final 6 months. The database does not contain specific information on do-not-resuscitate orders, advanced care planning, or overall goals of care but captures the consequences of these decisions with respect to health care delivery.

Outcomes

The primary outcome described end-of-life care according to location of death: intensive care unit, acute care hospital, long-term care facility (or nursing home), and other (including hospice or home). The results are described in terms of relative risk (RR), which in this case refers to the ratio of proportions of recent immigrant compared with long-standing resident decedents that experienced a given outcome. Secondary outcomes assessed whether a patient experienced intensive or invasive interventions in the last 6 months of life including hospital admission, intensive care admission, mechanical ventilation, dialysis, percutaneous gastric/gastrojejunal tube, or tracheostomy. Other secondary outcomes included emergency department, hospital, and intensive care use in the last 6 months of life.

Subgroup Analyses

Prespecified subgroup analyses were performed according to patient demographics (age, sex, urban or rural place of residence, and socioeconomic position), comorbidity (specific diagnostic categories and Charlson Comorbidity Index), and recent immigrant characteristics (immigration class, language ability on arrival, education level on arrival, time since immigration, and region of birth) (eTable 1 and eTable 2 in the Supplement).

Statistical Analysis

Recent immigrant and long-standing resident end-of-life care was compared using χ2 testing for categorical outcomes (location of death and proportion receiving invasive interventions) and rates (emergency department presentation, hospital admission, and intensive care admission), Mann-Whitney tests for comparisons of median duration of stay (hospital and intensive care admissions) and number of episodes (hospital and intensive care admissions and emergency department visits), and t tests for comparisons of mean duration (hospital and intensive care unit admissions).

Separate modified Poisson regression analyses of location of death (intensive care unit, acute care hospital, long-term care facility, or other including home) were conducted among recent immigrants compared with long-standing resident decedents to estimate RRs.20 We also performed separate modified Poisson regression analyses of type of invasive care received in the last 6 months including hospital admission, intensive care unit admission, mechanical ventilation, dialysis, percutaneous feeding tube placement, and tracheostomy. All analyses adjusted for potential confounders of age, sex, income quintile, urban-rural residence, and cause of death. To account for the correlation of outcomes among patients residing within the same geographic area, the analysis implemented generalized estimating equations using an exchangeable correlation structure, clustering by postal code.20 Recent immigrants were separately analyzed according to region of birth, years in Ontario, language ability on arrival, education level on arrival, and immigration class while adjusting for the same covariates as above. Two-sided P values less than .05 were considered significant but were not adjusted for multiple comparisons. Analyses were performed with SAS Enterprise Guide 6.1 (SAS Institute) and R 3.2.2 software (R Foundation). Details of the analysis protocol, regional definitions, causes of death, and further analyses assessing robustness across multiple fixed intervals preceding death can be found in eAppendix 1 in the Supplement.

Results
Demographic and Clinical Characteristics

A total of 967 013 decedents were analyzed, of whom 47 514 (5%) immigrated since 1985. Recent immigrant decedents originated from diverse global regions (eTable 3 in the Supplement). The median age at death was 79 years, with ischemic heart disease, lung cancer, and dementia being the most common causes of death. Compared with long-standing resident decedents, recent immigrant decedents tended to be younger and more likely to live in an urban area and of lower socioeconomic position (Table 1). The median duration in Canada for recent immigrants was 16 years.

End-of-Life Care

Of the 967 013 decedents, 434 783 (45%) died in the hospital including 99 680 (10%) who died in intensive care. Compared with long-standing resident decedents, a higher proportion of recent immigrant decedents died in intensive care (15.6% vs 10.0%; difference, 5.6%; 95% CI, 5.2%-5.9%) (Table 2). This increase persisted after adjusting for differences in age, sex, income, geography, and cause of death (Table 3) (adjusted RR of dying in intensive care comparing recent immigrant with long-standing resident decedents: 1.30; 95% CI, 1.27-1.32; Table 2).

In their last 6 months of life, recent immigrant decedents experienced more intensive care admissions (24.9% vs 19.2%; difference, 5.7%; 95% CI, 5.3%-6.1%), hospital admissions (72.1% vs 68.2%; difference, 3.9%; 95% CI, 3.5%-4.3%), mechanical ventilation (21.5% vs 13.6%; difference, 7.9%; 95% CI, 7.5%-8.3%), dialysis (5.5% vs 3.4%; difference, 2.1%; 95% CI, 1.9%-2.3%), percutaneous feeding tube placement (5.5% vs 3.0%; difference, 2.5%; 95% CI, 2.3%-2.8%), and tracheostomy (2.3% vs 1.1%; difference, 1.2%; 95% CI, 1.1%-1.4%) , even after adjusting for potential confounders (Table 2 and eTable 4 and eTable 5 in the Supplement). These increases persisted across various fixed intervals preceding death (1 month, 6 months, 12 months, and 24 months) (eFigure 1 in the Supplement).

Subgroup Analyses

Our finding that recent immigrant decedents were more likely to die in intensive care was consistent across diverse subgroups including older age at death, sex, socioeconomic status, place of residence, and Charlson Comorbidity Index score (Figure 1). The association persisted across different conditions including colorectal cancer (RR, 1.26; 95% CI, 1.02-1.57), diabetes (RR, 1.99; 95% CI, 1.74-2.28), cerebrovascular disease (RR, 2.12; 95% CI, 1.98-2.28), and dementia (RR, 3.69; 95% CI, 2.66-5.13).

The RR of death in intensive care comparing recent immigrant and long-standing resident decedents was highest among patients older than 80 years, female patients, and patients with a lower comorbidity index (Figure 1). There was substantial variation in end-of-life care according to region of birth and time since immigration (Figure 2). The RR of dying in intensive care (using the overall long-standing resident risk of dying in intensive care as baseline) ranged from 0.84 (95% CI, 0.74-0.95) among decedents born in northern and western Europe to 1.78 (95% CI, 1.66-1.92) among decedents born in western and central Asia, 1.84 (95% CI, 1.70-2.00) among decedents born in Africa, and 1.96 (95% CI, 1.89-2.05) among decedents born in South Asia (eFigure 2 in Supplement). After adjustment for age and other covariates in the recent immigrant population, the increased RR of dying in intensive care persisted among recent immigrant decedents from East Asia, Central America and Mexico, South America, Africa, western and central Asia, Southeast Asia, and South Asia. All other regions, including Northern and Western Europe, did not show statistically significant differences from Northern America (eTable 6 in Supplement). Differences were associated with time in Canada, with an RR of dying in intensive care of 1.42 (95% CI, 1.36-1.48) among those who immigrated 21 to 30 years before death and an RR of 2.03 (95% CI, 1.80-2.29) seen in those who immigrated fewer than 2 years before death. In adjusted analyses, the increased RR of dying in intensive care seen among recent immigrant decedents who immigrated 2 or fewer years before death remained statistically significant relative to recent immigrant decedents who immigrated more than 10 years before death, but the differences between those who immigrated 2 or fewer years before death and those who immigrated between 3 and 10 years before death were no longer statistically significant (eTable 6 in the Supplement). There were no significant differences in the adjusted analysis according to immigration class, language ability on arrival, socioeconomic position, or education level on arrival.

Sensitivity Analyses

The primary analysis does not include recent immigrant decedents who left Ontario before death. However, for the 95% CI of our estimate for proportion of recent immigrant decedents dying in intensive care to overlap with the corresponding quantity among long-standing residents, 26 329 recent immigrants (36%) would have had to leave Ontario and then die outside of an intensive care unit (eAppendix 2 in the Supplement).

A total of 14 758 decedents were excluded owing to fewer than 6 months of health care enrollment, of whom 458 (3%) were recent immigrants. The prevalence of missing data was highest in the Charlson Comorbidity Index score data because of a subset of patients who were never hospitalized, but otherwise the proportion of missing data was small (eTable 7 in Supplement).

Other sensitivity analyses assessed the difference between recent immigrants identified through deterministic as opposed to probabilistic matching. Of the 47 514 recent immigrant decedents identified, 37 046 (78%) were identified with deterministic linking and 10 468 (22%) were identified with probabilistic linking (eTable 8 in the Supplement). The 2 cohorts of recent immigrants were similar with respect to baseline characteristics, unadjusted primary analysis, and adjusted secondary analyses (eTables 3, 9, 10, and 11 in the Supplement).

Discussion

Among decedents in Ontario, recent immigrants were significantly more likely to receive aggressive care and to die in an intensive care unit than long-standing residents. In the last 6 months of life, recent immigrant decedents were more likely to experience intensive care unit admission, hospital admission, mechanical ventilation, dialysis, percutaneous feeding tube placement, and tracheostomy. These increased rates of aggressive care varied substantially according to region of birth, attenuated with time in Canada, and were not explained by differences in age, sex, cause of death, comorbidity, or socioeconomic position.

This study is a large-scale population-level quantitative analysis of end-of-life care provided to patients who have recently immigrated, using comprehensive data on hospital care as a consequence of universal health insurance. The data show extensive global region variation, with recent immigrants from Northern and Western Europe experiencing less-aggressive end-of-life care than long-standing residents, while those from Africa, South Asia, or Southeast Asia experienced the most-aggressive end-of-life care relative to long-standing residents. Qualitative research conducted in multiple cultural settings corroborates the finding that differences in end-of-life care provided to recent immigrants may be associated with region of origin.21-25 Within Europe and Asia, variations in the rate of organ-supporting care at the end of life are significantly associated with both region and the culture or religion of patients and physicians.26,27 The results agree with quantitative research conducted in racial/ethnic minority groups in the United States.21,22,24,25,28,29 The differences in end-of-life care delivery may also attenuate with time since immigration, consistent with other research describing acculturation and end-of-life care.30

The variation in end-of-life care based on region of birth has multiple potential explanations, including patient preferences, cultural differences, clinician behavior, end-of-life care decision processes, or differences in service accessibility.10,31 If recent immigrants live in social and geographic communities relating to their region of birth, this could lead to differential palliative care service availability for certain groups.32 Clinicians may conduct end-of-life care discussions in different ways, or less commonly, based on conscious or subconscious cultural, geographic, or religious perceptions of end-of-life care practices.33 Variation seen across diagnostic categories suggests the possibility of residual confounding due to clustering of disease processes and immigration status, although the associations with more-aggressive care persist after adjustment for category of cause of death. Many recent immigrant patients and families may be more familiar with clinician-directed or family and community–based models of medical decision-making, leading to different outcomes in an environment where patient preferences or shared patient-clinician decision-making guide end-of-life care decisions.15,21,33 The findings in this study might also be explained by differences in health literacy or language ability that could promote more aggressive end-of-life care through delayed clinical presentations, incomplete understanding of medical situations, or even decreased trust of health care professionals.31,34,35 However, the findings did not appear to be explained by English proficiency, level of education, socioeconomic position, nor place of residence at time of death.

Limitations

This study has several limitations. The most important limitation is that the recent immigrant and long-standing resident cohorts differed significantly in terms of age, socioeconomic status, and geography, which leaves the possibility of residual confounding.36 However, comprehensive coverage of all hospital care for every Canadian resident reduces confounding due to economic barriers that may exist to a greater extent in some other jurisdictions and with adjustment of these and other baseline characteristics, the potential for residual confounding should be greatly reduced. Another limitation is that different diseases have different terminal time courses, while the design analyzed fixed intervals preceding death; therefore, some aspects of end-of-life care may have been missed or some care prior to end of life may have been included.37 Health administrative databases are also limited in terms of risk adjustment by disease severity; however, decedent analyses involve inherent severity adjustment through selection of patients who have died. Although data were captured on all decedents in Ontario, no data were available about recent immigrants who returned to their country of origin to die; however, these populations are likely very small (eAppendix 2 in the Supplement). There were no analyses of hospital length of stay prior to intensive care unit admission. There were no data on or analyses of marital status, language ability for long-standing residents (or language ability more recently than arrival for recent immigrants), education level for long-standing residents, or goals of care and preferences for any patients or families.

Conclusions

Among decedents in Ontario, recent immigrants were significantly more likely to receive aggressive care and to die in an intensive care unit compared with other residents. Further research is needed to understand the mechanisms behind this association.

Section Editor: Derek C. Angus, MD, MPH, Associate Editor, JAMA (angusdc@upmc.edu).
Back to top
Article Information

Accepted for Publication: September 1, 2017.

Corresponding Author: Robert A. Fowler, MDCM, MS(Epi), University of Toronto–Sunnybrook Hospital, 2075 Bayview Ave, D478, Toronto, ON M4N 3M5, Canada (rob.fowler@sunnybrook.ca).

Published Online: October 2, 2017. doi:10.1001/jama.2017.14418

Author Contributions: Dr Fowler had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Yarnell, Manuel, Tanuseputro, Stukel, Fowler.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Yarnell, Fowler.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Yarnell, Fu, Stukel, Pinto, Fowler.

Obtained funding: Fowler.

Administrative, technical, or material support: Manuel, Fowler.

Supervision: Tanuseputro, Fowler.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Scales reports grants from the Canadian Institutes of Health Research, the Canadian Frailty Network, and the University of Toronto Integrating Challenge Fund. Dr Fowler reports grants from the Canadian Institutes of Health Research, the Canadian Frailty Network, and the University of Toronto Integrating Challenge Fund. No other disclosures were reported.

Funding/Support: The study was funded by the Canadian Institutes of Health Research, the Heart and Stroke Foundation of Ontario, the Canadian Frailty Network Center of Excellence, and the University of Toronto Integrating Challenge Fund. This study was supported by the Institute for Clinical Evaluative Sciences, which is funded by an annual grant from the Ontario Ministry of Health and Long-term Care.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The opinions, results, and conclusions reported in this study are those of the authors and are independent from the funding sources. No endorsement by the Institute for Clinical Evaluative Sciences; Immigration, Refugees and Citizenship Canada; or the Ontario Ministry of Health and Long-term Care is intended or should be inferred.

Meeting Presentation: This study was presented at the Critical Care Canada Forum; October 2, 2017; Toronto, Ontario, Canada.

Additional Contributions: We thank the following people for helpful comments: Arthur Slutsky, MD, MASc, BASc (vice president, research; St Michael’s Hospital, University of Toronto), Shail Rawal, MD, MPH (Department of Medicine, University of Toronto), Fahad Razak, MD, MSc, BASc (St Michael’s Hospital, University of Toronto), Amol Verma, MD, MPhil (Department of Medicine, University of Toronto), Raisa Mirza, BA (Social Impact Consultant), Kieran L. Quinn, MD, MSc (University of Toronto), and Donald Redelmeier, MD, MSHSR (Department of Medicine, University of Toronto). No compensation was given nor offered for the involvement of the listed individuals.

References
1.
World Health Organization. WHO definition of palliative care. http://www.who.int/cancer/palliative/definition/en/. Accessed March 21, 2016.
2.
Heyland  DK, Dodek  P, Rocker  G,  et al; Canadian Researchers End-of-Life Network(CARENET).  What matters most in end-of-life care: perceptions of seriously ill patients and their family members.  CMAJ. 2006;174(5):627-633.PubMedGoogle ScholarCrossref
3.
Heyland  DK, Cook  DJ, Rocker  GM,  et al; Canadian Researchers at the End of Life Network (CARENET).  Defining priorities for improving end-of-life care in Canada.  CMAJ. 2010;182(16):E747-E752.PubMedGoogle ScholarCrossref
4.
Frost  DW, Cook  DJ, Heyland  DK, Fowler  RA.  Patient and healthcare professional factors influencing end-of-life decision-making during critical illness: a systematic review.  Crit Care Med. 2011;39(5):1174-1189.PubMedGoogle ScholarCrossref
5.
Central Intelligence Agency. The World Factbook: country comparison: net migration rate. https://www.cia.gov/library/publications/the-world-factbook/rankorder/2112rank.html. Accessed March 21, 2016.
6.
Statistics Canada. Population and dwelling count highlight tables, 2011 census. http://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=101&S=50&O=A. Published January 7, 2016. Accessed March 21, 2016.
7.
Fuller-Thomson  E, Noack  AM, George  U.  Health decline among recent immigrants to Canada: findings from a nationally-representative longitudinal survey.  Can J Public Health. 2011;102(4):273-280.PubMedGoogle Scholar
8.
Sanmartin  C, Ross  N.  Experiencing difficulties accessing first-contact health services in Canada: Canadians without regular doctors and recent immigrants have difficulties accessing first-contact healthcare services: reports of difficulties in accessing care vary by age, sex and region.  Healthc Policy. 2006;1(2):103-119.PubMedGoogle Scholar
9.
Tu  JV, Chu  A, Rezai  MR,  et al.  The Incidence of Major Cardiovascular Events in Immigrants to Ontario, Canada: the CANHEART Immigrant Study [published online August 31, 2015].  Circulation.PubMedGoogle Scholar
10.
Lackan  NA, Eschbach  K, Stimpson  JP, Freeman  JL, Goodwin  JS.  Ethnic differences in hospital place of death among older adults in California.  Med Care. 2009;47:138-145.PubMedGoogle ScholarCrossref
11.
Bosma  H, Apland  L, Kazanjian  A.  Cultural conceptualizations of hospice palliative care: more similarities than differences.  Palliat Med. 2010;24(5):510-522.PubMedGoogle ScholarCrossref
12.
Unroe  KT, Greiner  MA, Johnson  KS, Curtis  LH, Setoguchi  S.  Racial differences in hospice use and patterns of care after enrollment in hospice among Medicare beneficiaries with heart failure.  Am Heart J. 2012;163(6):987-993.e3.PubMedGoogle ScholarCrossref
13.
Cruz-Oliver  DM, Talamantes  M, Sanchez-Reilly  S.  What evidence is available on end-of-life (EOL) care and Latino elders? a literature review.  Am J Hosp Palliat Care. 2014;31(1):87-97.PubMedGoogle ScholarCrossref
14.
Hanchate  A, Kronman  AC, Young-Xu  Y, Ash  AS, Emanuel  E.  Racial and ethnic differences in end-of-life costs: why do minorities cost more than whites?  Arch Intern Med. 2009;169(5):493-501.PubMedGoogle ScholarCrossref
15.
Smith  AK, Sudore  RL, Pérez-Stable  EJ.  Palliative care for Latino patients and their families: whenever we prayed, she wept.  JAMA. 2009;301(10):1047-1057, E1.PubMedGoogle ScholarCrossref
16.
Chiu  M, Lebenbaum  M, Lam  K,  et al.  Describing the linkages of the immigration, refugees and citizenship Canada permanent resident data and vital statistics death registry to Ontario’s administrative health database.  BMC Med Inform Decis Mak. 2016;16(1):135.PubMedGoogle ScholarCrossref
17.
Juurlink  D, Preyra  C, Croxford  R,  et al.  Canadian Institute for Health Information Discharge Abstract Database: A Validation Study. Toronto, ON, Canada: Institute for Clinical Evaluative Sciences; 2006.
18.
World Health Organization. ICD-10 version: 2016. http://apps.who.int/classifications/icd10/browse/2016/en. Accessed March 21, 2016.
19.
Glazier  RH, Zagorski  BM, Rayner  J.  Comparison of Primary Care Models in Ontario by Demographics, Case Mix and Emergency Department Use, 2008/09 to 2009/10. Toronto, ON, Canada: Institute for Clinical Evaluative Sciences; 2012: 41.
20.
Yelland  LN, Salter  AB, Ryan  P.  Performance of the modified Poisson regression approach for estimating relative risks from clustered prospective data.  Am J Epidemiol. 2011;174(8):984-992.PubMedGoogle ScholarCrossref
21.
Adames  HY, Chavez-Dueñas  NY, Fuentes  MA, Salas  SP, Perez-Chavez  JG.  Integration of Latino/a cultural values into palliative health care: a culture centered model.  Palliat Support Care. 2014;12(2):149-157.PubMedGoogle ScholarCrossref
22.
Ebrahim  S, Bance  S, Bowman  KW.  Sikh perspectives towards death and end-of-life care.  J Palliat Care. 2011;27(2):170-174.PubMedGoogle Scholar
23.
Dobbs  D, Park  NS, Jang  Y, Meng  H.  Awareness and completion of advance directives in older Korean-American adults.  J Am Geriatr Soc. 2015;63(3):565-570.PubMedGoogle ScholarCrossref
24.
Garrido  MM, Harrington  ST, Prigerson  HG.  End-of-life treatment preferences: a key to reducing ethnic/racial disparities in advance care planning?  Cancer. 2014;120(24):3981-3986.PubMedGoogle ScholarCrossref
25.
Higginson  IJ, Gomes  B, Calanzani  N,  et al; Project PRISMA.  Priorities for treatment, care and information if faced with serious illness: a comparative population-based survey in seven European countries.  Palliat Med. 2014;28(2):101-110.PubMedGoogle ScholarCrossref
26.
Sprung  CL, Cohen  SL, Sjokvist  P,  et al; Ethicus Study Group.  End-of-life practices in European intensive care units: the Ethicus Study.  JAMA. 2003;290(6):790-797.PubMedGoogle ScholarCrossref
27.
Phua  J, Joynt  GM, Nishimura  M,  et al; ACME Study Investigators and the Asian Critical Care Clinical Trials Group.  Withholding and withdrawal of life-sustaining treatments in intensive care units in Asia.  JAMA Intern Med. 2015;175(3):363-371.PubMedGoogle ScholarCrossref
28.
Moss  KO, Williams  IC.  End-of-life preferences in Afro-Caribbean older adults: a systematic literature review.  Omega (Westport). 2014;69(3):271-282.PubMedGoogle ScholarCrossref
29.
Sharma  RK, Khosla  N, Tulsky  JA, Carrese  JA.  Traditional expectations versus US realities: first- and second-generation Asian Indian perspectives on end-of-life care.  J Gen Intern Med. 2012;27(3):311-317.PubMedGoogle ScholarCrossref
30.
Wright  AA, Stieglitz  H, Kupersztoch  YM,  et al.  United states acculturation and cancer patients’ end-of-life care.  PLoS One. 2013;8(3):e58663.PubMedGoogle ScholarCrossref
31.
Jonnalagadda  S, Lin  JJ, Nelson  JE,  et al.  Racial and ethnic differences in beliefs about lung cancer care.  Chest. 2012;142(5):1251-1258.PubMedGoogle ScholarCrossref
32.
Kavalieratos  D, Corbelli  J, Zhang  D,  et al.  Association between palliative care and patient and caregiver outcomes: a systematic review and meta-analysis.  JAMA. 2016;316(20):2104-2114.PubMedGoogle ScholarCrossref
33.
Mack  JW, Paulk  ME, Viswanath  K, Prigerson  HG.  Racial disparities in the outcomes of communication on medical care received near death.  Arch Intern Med. 2010;170(17):1533-1540.PubMedGoogle ScholarCrossref
34.
Volandes  AE, Paasche-Orlow  M, Gillick  MR,  et al.  Health literacy not race predicts end-of-life care preferences.  J Palliat Med. 2008;11(5):754-762.PubMedGoogle ScholarCrossref
35.
White  DB, Ernecoff  N, Buddadhumaruk  P,  et al.  Prevalence of and factors related to discordance about prognosis between physicians and surrogate decision makers of critically ill patients.  JAMA. 2016;315(19):2086-2094.PubMedGoogle ScholarCrossref
36.
Baicker  K, Chandra  A, Skinner  JS.  Geographic variation in health care and the problem of measuring racial disparities.  Perspect Biol Med. 2005;48(1)(suppl):S42-S53.PubMedGoogle ScholarCrossref
37.
Bach  PB, Schrag  D, Begg  CB.  Resurrecting treatment histories of dead patients: a study design that should be laid to rest.  JAMA. 2004;292(22):2765-2770.PubMedGoogle ScholarCrossref
×