1. Introduction
We analyzed the Decedent Questionnaire (DQ) from the Canadian Longitudinal Study on Aging (CLSA) [1, 2] to determine the percent of participants in this cohort study who died in their desired location as well as the factors associated with dying in the desired location, as reported by a proxy.
2. Methods
The CLSA consists of two related cohorts—the “tracking cohort” which is randomly sampled from representative population frames across Canada, and the “comprehensive cohort” which is limited to those living close to one of the study sites, and includes in‐person interviews and physical assessments of health. The combined cohort comprises 51,338 participants aged 45–85 years at enrolment [1]. The original sampling frame was from representative sampling frames with a goal of ensuring that the sample was as representative of community living adults as possible. The sampling frames excluded: adults living with cognitive impairment at recruitment, individuals unable to participate in English or French, individuals living in long‐term care (LTC) facilities at recruitment, active Royal Canadian Mounted Police (RCMP) and military personnel, individuals living on First Nations reserves, the three territories and in some rural and remote regions. The DQ [3] is proxy reported after the participant died. At the time of our analysis, 3672 participants had died, and 1287 (35%) completed DQs. The proxy respondent was a husband/wife for 510 (40%); a father/mother for 97 (8%); and a son/daughter for 434 (34%); with all others making up the remainder. The proxy reported being very close to the decedent in 1178 (92%) of the participants.
For the outcome, we considered the response to the question: “In the last week of [his/her] life, do you feel that [decedent's name] died where [he/she] wanted to?” These were coded as Yes/No, Don't know or refused. We considered age at death and sex. Proxies are also asked about the functional trajectory leading to death [4]—Sudden, Terminal (a rapid decline over 1–2 months), Chronic Illness (illness over a period of years with periods of worsening followed by partial or complete recovery), and Gradual Decline (a slow, gradual decline usually over years with steadily increasing disability before dying). Proxies also report functional status [5], which we categorized as “No Disability; Mild Disability; and Moderate/Severe/Total Disability.” We used the cause of death provided by the proxy. Two investigators (PStJ and RT) reviewed the “Other Causes” of death and recoded these causes of death. We recoded some into an existing category (e.g., Valvular Heart Disease into “Cardiac”), and created two categories—Gastro‐intestinal (which includes liver disease) and Chronic Neurological (which includes Parkinson Disease, Upper Motor Neuron Disease, etc.). We also recoded “Covid‐19” deaths into “Influenza/Pneumonia.” We left some causes of death as “Other”—many of these were deaths from multiple organ failure, or deaths at very old ages. We considered place of death as: Community (their home, another private home, or other non‐institutional places such as workplace, shopping areas, or on roads/trails); Hospital; LTC (Nursing Home, Chronic Care facility or Seniors' Residence); or Palliative Care (Palliative Care Unit or Hospice).
We conducted bivariate analyses using chi square tests for categorical variables and Student's t‐tests or Analysis of Variance (ANOVA) for continuous variables. We then constructed logistic regression models with the participant dying in the location that they wanted as the outcome variable.
3. Results
Of the 1287 DQs, 764 died in the location that they desired, and 425 did not. There were 98 respondents with missing data. The results of the bivariate analyses are shown in Table 1. The results from logistic regression models adjusted for age, sex, and education are shown in Table 2. Compared to those who died in hospital, those who died in the community were markedly more likely to have died where they wanted to die, as were those who died in palliative care settings. Those who died with dementia, with infection, with respiratory tract infection, or due to an accident were all less likely to die where they desired compared to those with cancer. We also found that the trajectory leading to death was associated with the desired location of death: those with a gradual or a chronic disease pattern of decline to death were less likely to die in the desired location compared to those who died suddenly.
TABLE 1.
Baseline characteristics.
| Died where they wanted (N = 764) | Did not die where they wanted (N = 425) | Not included (N = 98) | p value | |
|---|---|---|---|---|
| Age (mean years, SD) | 77.1 (8.8) | 77.6 (9.0) | 78.2 (8.7) | 0.38 |
| Sex (N, % female) | 281 (37%) | 171 (40%) | 37 (38%) | 0.50 |
| Income (N, %) | 0.38 | |||
| < 20,000 | 106 (14%) | 73 (17%) | Suppressed | |
| 20–50,000 | 353 (46%) | 181 (42%) | Suppressed | |
| 50–100,000 | 184 (24%) | 112 (26%) | Suppressed | |
| 100,000 or more | 57 (7%) | 23 (5%) | Suppressed | |
| Missing | 64 (8%) | 36 (8%) | Suppressed | |
| Education | 0.64 | |||
| Less than secondary | 100 (13%) | 51 (12%) | 15 (15%) | |
| Secondary | 90 (12%) | 55 (13%) | 13 (13%) | |
| Some post‐secondary | 75 (10%) | 41 (10%) | Suppressed | |
| Post‐secondary | 497 (65%) | 277 (65%) | 65 (66%) | |
| Missing | Suppressed | Suppressed | Suppressed | |
| Place of death (N, %) | < 0.001 | |||
| Home | 273 (36%) | 32 (8%) | 13 (13%) | |
| Hospital | 301 (39%) | 266 (63%) | 64 (65%) | |
| Palliative care | 136 (18%) | 75 (18%) | 11 (11%) | |
| LTC | 54 (7%) | 52 (12%) | 10 (10%) | |
| Cause of death (N, %) | 0.0019 | |||
| Cancer | 329 (43%) | 160 (38%) | 32 (33%) | |
| Cardiac | 171 (22%) | 72 (17%) | 25 (26%) | |
| Stroke | 48 (6%) | 26 (6%) | 9 (9%) | |
| Respiratory | 37 (5%) | 21 (5%) | Suppressed | |
| Dementia | 12 (2%) | 18 (4%) | Suppressed | |
| Accident | 16 (2%) | 17 (4%) | Suppressed | |
| Kidney disease | 14 (2%) | 13 (3%) | Suppressed | |
| Respiratory tract infection | 20 (3%) | 25 (6%) | Suppressed | |
| Degenerative neurological | 17 (2%) | 9 (2%) | Suppressed | |
| Infection | 15 (2%) | 22 (5%) | Suppressed | |
| Liver/GI disease | 20 (3%) | 10 (2%) | Suppressed | |
| Other | 65 (9%) | 32 (8%) | 9 (9%) | |
| Trajectory to death (N, %) | 0.0067 | |||
| Sudden | 166 (22%) | 74 (17%) | 25 (26%) | |
| Terminal | 364 (48%) | 174 (41%) | 40 (41%) | |
| Chronic illness | 115 (15%) | 97 (23%) | 14 (14%) | |
| Gradual | 96 (13%) | 66 (16%) | 13 (13%) | |
| Other/don't know | 23 (3%) | 14 (3%) | 6 (6%) | |
| Functional impairment (N, %) | 0.12 | |||
| None | 427 (56%) | 225 (53%) | 60 (61%) | |
| Mild | 157 (21%) | 109 (26%) | 21 (21%) | |
| Moderate/severe/total | 65 (9%) | 28 (7%) | 10 (10%) | |
| Missing | 115 (15%) | 63 (15%) | 7 (7%) |
Note: LTC is long‐term care; Suppressed is suppressed due to small cell size.
TABLE 2.
Results of logistic regression models predicting the odds ratio of dying in the place that the proxy reported they desired.
| Model 1 OR (95% CI) | Model 2 OR (95% CI) | Model 3 OR (95% CI) | |
|---|---|---|---|
| Age (per year) | 1.00 (0.99, 1.02) | 1.00 (0.98, 1.01) | 1.00 (0.98, 1.01) |
| Sex (ref = male) | 0.88 (0.68, 1.15) | 0.86 (0.67, 1.10) | 0.86 (0.66, 1.11) |
| Education | |||
| Less than secondary | 1.27 (0.86, 1.88) | 1.11 (0.76, 1.63) | 1.16 (0.80, 1.71) |
| Secondary | 0.94 (0.63, 1.40) | 0.92 (0.63, 1.35) | 0.91 (0.62, 1.34) |
| Some post‐secondary | 0.95 (0.61, 1.48) | 0.94 (0.62, 1.44) | 1.03 (0.67, 1.58) |
| Post‐secondary | Ref | Ref | Ref |
| Place of death | |||
| Hospital | Ref | ||
| Community | 8.13 (5.34, 12.37) | ||
| Palliative care | 1.51 (1.09, 2.12) | ||
| LTC | 0.91 (0.59, 1.40) | ||
| Cause of death | |||
| Cancer | Ref | ||
| Cardiac | 1.18 (0.84, 1.66) | ||
| Stroke | 0.90 (0.53, 1.54) | ||
| Respiratory | 0.93 (0.52, 1.69) | ||
| Dementia | 0.34 (0.16, 0.72) | ||
| Accident | 0.47 (0.22, 0.99) | ||
| Kidney disease | 0.56 (0.25, 1.24) | ||
| Respiratory tract infection | 0.37 (0.20, 0.70) | ||
| Degenerative neurological | 0.93 (0.41, 2.14) | ||
| Infection | 0.33 (0.17, 0.66) | ||
| Liver/GI disease | 0.98 (0.45, 2.15) | ||
| Other | 1.01 (0.63, 1.62) | ||
| Trajectory to death | |||
| Sudden | Ref | ||
| Terminal | 0.91 (0.65, 1.28) | ||
| Chronic illness | 0.50 (0.34, 0.74) | ||
| Gradual | 0.63 (0.41, 0.97) |
Note: Models are adjusted for age and sex. Model 1 considers the place of death; Model 2 considers the cause of death; and Model 3 considers the proxy report of functional trajectory prior to death.
Abbreviations: CI, confidence interval; LTC, long‐term care; OR, odds ratio; Ref, reference group.
4. Discussion
We have found that dying in the place that people want to die, as reported by a proxy, is dependent upon the primary cause of death and the functional trajectory to death. These factors are difficult to modify with health service delivery changes—such as increased access to palliative care, home care, or geriatric specialty services. Thus, it may prove difficult to ensure that most people die where they would like to die.
Author Contributions
All authors contributed significantly to this work. All authors were involved in the conceptualization. Mr. Nowicki and Dr. St John were responsible for data curation for the Canadian Longitudinal Study on Aging (CLSA) data set received for these analyses. Mr. Nowicki conducted the data analyses and Dr. St John made the graphs. All authors acquired funding and all contributed to the data interpretation. Dr. St John wrote the original manuscript, and all authors provided comments and revisions. All authors take responsibility for the final manuscript.
Funding
This work was supported by Government of Canada through the Canadian Institutes of Health Research (CIHR), the Canada Foundation for Innovation, and the Centre on Aging, University of Manitoba.
Conflicts of Interest
Dr. St John declares no conflicts of interest. His spouse (Dr. Christie Macdonald) provided in kind support by proof reading this. Dr. St John was an unremunerated board member of A and O (Formerly Age and Opportunity Manitoba) and has received speaking honorarium from the Manitoba Gerontological Nursing Association and management consulting fees from the University Health Network (University of Toronto). He has received research grants from the Canadian Institutes for Health Research, from the Manitoba Medical Services Foundation, and from Charitable donations from members of the Manitoba Follow‐Up Study. Dt Tate has received research grants from the Canadian Institutes for Health Research, and from Charitable donations from members of the Manitoba Follow‐Up Study.
Acknowledgments
This research was made possible using the data/biospecimens collected by the CLSA. Funding for the CLSA is provided by the Government of Canada through the Canadian Institutes of Health Research (CIHR) under grant reference: LSA 94473 and the Canada Foundation for Innovation, as well as the following provinces: Newfoundland, Nova Scotia, Quebec, Ontario, Manitoba, Alberta, and British Columbia. This research has been conducted using the CLSA dataset Decedent Questionnaire—Version 1, Baseline Tracking Dataset—Version 4.0, Baseline Comprehensive Dataset—Version 7.0, Follow‐Up 1 Tracking Dataset—Version 3.1, Follow‐Up 1 Comprehensive Dataset—Version 5.0, Follow‐Up 2 Tracking Dataset—Version 1.1 and Follow‐Up 2 Comprehensive Dataset—Version 2.0—Participant Status—Version 2.0 under Application ID 2310003. The CLSA is led by Drs. Parminder Raina, Christina Wolfson, and Susan Kirkland. The authors gratefully acknowledge the time and commitment of the CLSA participants, without whom this research would not be possible. The opinions expressed in this manuscript are the authors' own and do not necessarily reflect the views of the Canadian Longitudinal Study on Aging. AI was not used in any capacity for this manuscript.
St John P. D., Tate R. B., Nowicki S., et al., “Dying in the Desired Location—Findings From the Canadian Longitudinal Study on Aging,” Journal of the American Geriatrics Society 74, no. 10 (2026): 3111–3114, 10.1111/jgs.70567.
This research was presented at the 44th Annual Scientific Meeting (ASM) of the CGS in Toronto, May 29–31, 2025.
Data Availability Statement
Data are available from the Canadian Longitudinal Study on Aging (www.clsa‐elcv.ca) for researchers who meet the criteria for access to de‐identified CLSA data.
References
- 1. Raina P., Wolfson C., Kirkland S., et al., “Cohort Profile: The Canadian Longitudinal Study on Aging (CLSA),” International Journal of Epidemiology 48 (2019): 1752–1753j. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Raina P. S., Wolfson C., Kirkland S. A., et al., “The Canadian Longitudinal Study on Aging (CLSA),” Canadian Journal on Aging 28 (2009): 221–229. [DOI] [PubMed] [Google Scholar]
- 3. CLSA , Decedent Questionnaire v1.2 (2018). [Google Scholar]
- 4. Lunney J. R., Lynn J., Foley D. J., Lipson S., and Guralnik J. M., “Patterns of Functional Decline at the End of Life,” JAMA 289 (2003): 2387–2392. [DOI] [PubMed] [Google Scholar]
- 5. Fillenbaum G. G. and Smyer M. A., “The Development, Validity, and Reliability of the OARS Multidimensional Functional Assessment Questionnaire,” Journal of Gerontology 36 (1981): 428–434. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data are available from the Canadian Longitudinal Study on Aging (www.clsa‐elcv.ca) for researchers who meet the criteria for access to de‐identified CLSA data.