Abstract
Background:
Many patients who receive chronic hemodialysis have a limited life expectancy comparable to that of patients with metastatic cancer. However, patterns of home palliative care use among patients receiving hemodialysis are unknown.
Objectives:
We aimed to undertake a current-state analysis to inform measurement and quality improvement in palliative service use in Ontario.
Methods:
We conducted a descriptive study of outcomes and home palliative care use by Ontario residents maintained on chronic dialysis using multiple provincial healthcare datasets. The period of study was the final year of life, for those died between January 2010 and December 2014.
Results:
We identified 9611 patients meeting inclusion criteria. At death, patients were (median [Q1, Q3] or %): 75 (66, 82) years old, on dialysis for 3.0 (1.0-6.0) years, 41% were women, 65% had diabetes, 29.6% had dementia, and 13.9% had high-impact neoplasms, and 19.9% had discontinued dialysis within 30 days of death. During the last year of life, 13.1% received ⩾1 home palliative services. Compared with patients who had no palliative services, those who received home palliative care visits had fewer emergency department and intensive care unit visits in the last 30 days of life, more deaths at home (17.1 vs 1.4%), and a lower frequency of deaths with an associated intensive care unit stay (8.1 vs 37.8%).
Conclusions:
Only a small proportion of patients receiving dialysis in Ontario received support through the home palliative care system. There appears to be an opportunity to improve palliative care support in parallel with dialysis care, which may improve patient, family, and health-system outcomes.
What was known before
Patterns of palliative service use and associated outcomes among Canadians on chronic dialysis were previously unknown.
What this adds
In this study, we found very low overall rates of home palliative service use among Ontarians on chronic dialysis. Patients who died while on chronic dialysis used acute inpatient, emergency, and intensive care services. Measures reported in this study can be used as a baseline against which to track future improvements in palliative and end-of-life care provided to chronic dialysis patients.
Introduction
Accessible and high-quality palliative care remains a Canadian national priority. Since the recognition of suboptimal palliative care capacity 2 decades ago, 1 policy reforms and, ultimately, the Canadian Federal Health Accord of 2003 led to major federal and provincial investments. These were aimed at shifting end-of-life care from acute care hospitals to preferred settings, by enhancing system capacity and improving the coordination of care.2,3 In Ontario, efforts to bolster the palliative care system have largely focused on cancer patients, who have historically constituted 85% of palliative care referrals. 4 The establishment of Cancer Care Ontario in 1997 led to a formal End-of-Life Care Strategy and the creation of various structures and collaborations that have translated into better outcomes and less resource use. 4 In 2011, the Declaration of Partnership and Commitment to Action was developed and articulated a shared vision of palliative care in Ontario that was disease agnostic. 5
End-stage kidney disease (ESKD) is a debilitating condition with limited life expectancy and symptom burden that rivals that of many cancers. 6 Recognizing the progressive nature of kidney disease, and the high burden of concomitant frailty and comorbidity in the ESKD population, all aspects of palliative care planning and delivery have become a focus of attention in the kidney community.
The Ontario Renal Network (ORN) oversees the funding and quality of kidney care services for patients in Ontario with chronic kidney disease (CKD). Over 11 000 Ontario residents receive some form of dialysis care. Through engagement of patients and other stakeholders, the ORN has developed a 5-year strategic plan, published in the 2015-2019 Ontario Renal Plan (ORP), to improve access to palliative care services; enhance kidney health-system integration with primary care, home, and community services; and improve provider competency in facilitating patient education and shared decision-making. 7 To guide quality improvement initiatives and obtain baseline measures against which to track progress, we undertook a current-state analysis of home palliative care (hPC) service use and outcomes in Ontario’s dialysis population.
Methods
Study Design and Setting
We conducted a retrospective descriptive cohort study using linked healthcare databases held at the Institute for Clinical Evaluative Sciences (ICES) in Ontario. Ontario residents receive universal healthcare with most encounters captured in administrative data. Funded palliative services range from home care (nurse, physician, personal support worker, or multidisciplinary teams), palliative care clinics, hospice care facilities, and inpatient palliative consultative services. Given the central role of home services in Ontario’s palliative care strategies, we restricted our analyses to administrative codes representing hPC services. This study was approved by the research ethics board at Sunnybrook Health Sciences Centre (Toronto, Canada). Manuscript preparation adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (Appendix A). 8
We used an integrated knowledge translation approach to develop our research protocol and interpret our findings with the Ontario Renal Network Palliative Care Priority Panel—the advisory group that oversees the provincial renal palliative care strategy and includes palliative care physicians, the full spectrum of healthcare provider stakeholders, administrators, patients, and families.
Data Sources
We ascertained baseline characteristics and outcomes using a variety of linked databases (details below and Appendix B). We obtained death date and other vital statistics from the Registered Persons Database. We identified individuals on chronic dialysis using the Canadian Organ Replacement Register (CORR) and the Ontario Renal Reporting System (ORRS). We captured nurse and physician hPC codes in the Ontario Health Insurance Plan (OHIP), Canadian Institute for Health Information Discharge Abstract Database (CIHI-DAD), National Ambulatory Care Reporting System (NACRS), Home Care Database (HCD), Continuing Care Reporting System (CCRS), Resident Assessment Instrument (RAI)–Contact Assessment (CA), and RAI–Home Care (HC). All databases were linked using unique encoded identifiers and analyzed at ICES.
Population
We identified chronic dialysis patients who died between January 1, 2010, and December 31, 2014, in CORR and ORRS, and used their death date as the index date. We looked back 365 days from this index date to determine their trajectory of care in the last year of life. Patients who had received a kidney transplant required a minimum survival on dialysis of at least one year after transplant failure before they could enter the cohort. We reported outcomes separately for patients who received and did not receive hPC services in the last year of life. We also identified a subgroup of patients who electively discontinued dialysis within 30 days of death (Figure 1). Home palliative care service use was defined as receiving at least one administrative code from OHIP, HCD, RAI-CA, or RAI-HC at any time in the last year of life (Appendix C). We counted the number of days that a patient had received at least one palliative care code to define a palliative visit (ie, multiple codes on the same day were counted as a single encounter). We did not count hospital or clinic-based palliative care visits in the main analysis. We ascertained home deaths using OHIP codes.
Cohort flow diagram.
Variables
We ascertained demographic information at the time of death. We looked back to 1981 in CORR (inception) to identify complete dialysis history. We defined dialysis discontinuation using ORRS codes for patients marked as having withdrawn from dialysis, counting only those who died within 30 days of dialysis discontinuation to limit misclassification where dialysis was stopped due to kidney recovery.
We captured comorbidities within 2 years of death using OHIP, CIHI-DAD, and NACRS. We used the Johns Hopkins Adjusted Clinical Group (ACG) system to score comorbidity. 9 The ACG estimates an individual’s expected service use. The International Classification of Diseases, Ninth Revision (ICD-9) and The International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes are categorized into 32 groups, called Ambulatory Diagnostic Groups (ADGs), then further reduced to 12 “Collapsed ADGs” or Collapsed Aggregated Diagnosis Groups (CADGs). 10
A panel of content experts prioritized candidate outcomes based on importance and feasibility, given available data. These outcomes were (1) emergency department (ED) visits, (2) intensive care unit (ICU) visits, (3) time to death (starting at dialysis initiation), and (4) place of death. Place of death was coded as ICU, acute care hospital, or ED (counted together because most patients that died in hospital had a corresponding ED visit), long-term care, complex continuing care, home, or unattended death/unknown. If a patient had an admission to ICU, hospital, and Emergency Room (ER), they were considered to have had a hospital death with an ICU stay. We coded patients as having died at home only if they had an associated OHIP code for pronouncement of death at home.
Statistical Analyses
We reported baseline characteristics as frequencies and proportions, medians and interquartile ranges (IQRs) as appropriate. We compared patient characteristics using standardized differences across categories of hPC service use and considered a difference of ⩾10% significant. We reported outcomes as counts and proportions, and medians with corresponding first/third quartiles for continuous variables. We used SAS software, version 9.4 (SAS Institute Inc, Cary, North Carolina) for statistical analyses.
Results
Patient Characteristics
We identified 9611 chronic dialysis patients (Figure 1) who died between 2010 and 2014. Sociodemographic characteristics, dialysis modalities, and comorbidities did not differ by year of death (data not shown). The characteristics of the entire cohort are presented in Table 1. At the time of death, the median age in the overall cohort was 75.0 years, 40.6% were women, and 11.5% resided in rural locations. Comorbid conditions included high rates of diabetes (65.0%), ischemic heart disease (55.3%), congestive heart failure (55.5%), and dementia (29.6%). Patients had received dialysis for a median of 3.0 years prior to death, and 4.0% had a history of transplant, with a median time from transplant to reinitiation of chronic dialysis being 13.0 years (IQR: 13). A total of 912 (19.9%) discontinued dialysis in the last 30 days of life.
Patient Characteristics at Time of Death in Overall 2010 to 2014 Cohort, As Well As Stratified by Patients Receiving/Not Receiving at Least One Home Care Palliative Care Visit in the Last Year of Life With Associate Standardized Differences.
Note. EDCs = Johns Hopkins Extended Diagnosis ClustersTM; CIHI-DAD = Canadian Institute for Health Information Discharge Abstract Database; ACG = Adjusted Clinical Group®; ESRD = end-stage renal disease; LTC = long-term care.
Standardized differences (SD) are less sensitive to sample size than traditional hypothesis tests. They provide a measure of the difference between groups divided by the pooled SD; a value greater than 10% is interpreted as a meaningful difference between groups. Some cell values were suppressed for the purposes of privacy and confidentiality.
Rural was defined as population < 10 000.
Income was categorized into fifths of average neighborhood income on death date. Income Quintile category 3 includes the missing values.
Malignant neoplasm categories were defined using Johns Hopkins Expanded Diagnosis Clusters (EDC), which are based on proprietary ICD codes.
Comorbidities were assessed by administrative database codes or by EDC in the previous 2 years from death
Immobility was defined using ICD-10 codes in CIHI-DAD: “R263,” “R2681,” “Z740.”
Number have been altered to suppress small cells.
Standardized difference has been changed to accommodate suppressed cells.
Home Palliative Care Service Use in the Final Year of Life
Among the 9611 chronic dialysis patients who died during the study period, 1258 (13.1%) received at least 1 hPC service in the last year of life. Compared with those who received no hPC, hPC recipients had fewer comorbidities, with the exception of concurrent malignancy (Tables 1 and 2). Among the 7699 patients who did not discontinue dialysis, 742 (9.6%) received hPC services in the last year of life. In contrast, 516 (27.0%) of those who discontinued dialysis received hPC services.
Comparison of Home Palliative Care Service Measures in the Last Year of Life and Outcomes Among Ontario Residents Who Died on or Discontinued Chronic Dialysis.
Note. ED = emergency department; ICU = intensive care unit; LTC = long-term care; CCC = complex continuing care. All values expressed as n (%) unless otherwise specified. Survival time is defined as the number of years between the first-ever ESKD service date and death date; presented values are median (25th, 75th percentile).
Acute care hospital and ED visits were combined to avoid small cells, and for the purposes of ensuring privacy and confidentiality.
Of those who received hPC services, only 382 (30.3%) had one or more visits in the period 31 to 365 days prior to death, while the majority (n = 1036, 82.4%) had visit(s) only in the last 30 days of life. Among those who received an hPC visit, most received a single such visit (n = 685, 54.4% of those receiving hPC) often within the last 30 days of life (n = 876, 69.6%). A total of 564 (44.8%) of these patients received their first visit in the 7 days prior to death.
ED Visits
The proportion of all dialysis patients who died and had ⩾1 ED visit in the last 14 and 30 days of life was 44.1% and 60.0%, respectively (Table 2). Patients had fewer ED visits in the last year of life if they had hPC services.
ICU Admissions
Patients who received hPC services had fewer admissions to ICU in the last days of life (12.2 and 14.9% within 14 and 30 days, respectively), compared with patients who did not receive hPC (40.1 and 41.9% within 14 and 30 days, respectively).
Time to Death
The median (IQR) time on dialysis before death was 2.7 (4.7) years for all patients (n = 9611) (Table 2). Patients who had received hPC services had a marginally shorter time on dialysis (median: 2.0 years; IQR: 3.9) than those who did not have any palliative care visits (median: 2.8 years; IQR: 4.9). Similar results were seen for those who discontinued dialysis. The median time to death from discontinuation of dialysis was 4 days (IQR: 6).
Location of Death
In the overall cohort, 70.2% of patients died in the ED, acute care hospital, or ICU. Patients who received one or more hPC visits (vs none) were more likely to die at home (17.1 vs 1.4%). Patients who received hPC in the last year had a higher rate of deaths in unclassified locations (33.1%) compared with those who received no hPC (9.4%).
Sensitivity Analyses
We repeated the main analysis after restricting the observation window for hPC services to the final 30 days of life. All results were qualitatively similar in direction and magnitude as in the main analysis in which we captured hPC service use in the last 365 days of life. In addition, we analyzed the data using inpatient palliative care service visits (DAD and OHIP). Results were similar for most comparisons. The details of these analyses are available upon request.
Subgroup Analysis
Among the 1912 patients who discontinued dialysis, 516 (27.0%) received hPC in the final 365 days of life. Compared with those who had no hPC, those who received hPC had similar rates of ED visits, but lower rates of ICU admission (30.4 vs 12.4%) and fewer deaths in ICU (23.5 vs 4.7%) or acute care hospital/ED (42.3 vs 22.9).
Discussion
We report a current-state analysis of hPC use among Ontario residents maintained on chronic dialysis. We found that hPC services are infrequently used in this population and that, in most cases, used only within the last days of life.
Our findings suggest that at present, there is infrequent and late collaboration with home-based palliative care teams. We found, much like other studies,11-14 that palliative services were mostly used in end-of-life situations. Even when death is anticipated, patients most often do not appear to receive palliative care. In our cohort, only 27% of patients who discontinued dialysis received home palliative services in the last year of life. In their series, Couchoud et al found that among patients who discontinued dialysis while in hospital, 90% were not admitted to designated palliative care beds. 15 While various patient, hospital, and system-level barriers may limit patients’ access to palliative services, provider attitudes and practices are likely major determinants of palliative care use. In a questionnaire-based study of European nephrologists, only 10% reported involving palliative care services when they started discussions around dialysis discontinuation. 14 In a recent survey of Ontario renal providers (nurses, physicians and social workers), the provincial mean score on a subscale measuring providers’ propensity to engage community palliative services was only 2.5 out of 5.0, suggesting room for improvement (unpublished data, Ontario Renal Network).
A large proportion of patients in our cohort had ED and ICU visits during their final days of life. This is consistent with the high rate of cardiovascular and infection-related death in the dialysis-treated population. 16 Patients who received hPC had fewer ED and ICU visits compared with those who did not. However, in this descriptive study, we did not adjust for confounders that might have accounted for these differences. Interestingly, patients who received hPC had a higher rate of high-impact neoplasia compared with those who did not (39.8 vs 10.0%), suggesting that they may have accessed hPC through the cancer care system.
Developing disease-agnostic palliative care services is a priority in Ontario. These services are intended to extend beyond end-of-life care and provide multidisciplinary symptom and existential support to individuals with life-limiting, progressive, and burdensome disease. When optimally used, palliative care is introduced early in the disease trajectory and is provided in conjunction with disease modifying treatments.17-19 As the disease progresses along the trajectory, the focus of care shifts, gradually, from disease-modification to symptom management and focuses largely on the individual patient’s experience. ESKD is well recognized as a serious and life-limiting illness with a high symptom burden. Treatments often centre around dialysis care, with until recently, little focus on symptom management or the patient experience. We observed that, in Ontario, palliative care was not being used in this way, with most hPC recipients having only 1 hPC visit. This finding suggests it is unlikely patients have sufficient time to establish a trusting relationship with the palliative care team with, likely, a higher reliance on care provided in the acute care setting. We also found a high rate of deaths occurring in hospital, often in association with an ICU stay. Although it is not possible from these data to tease out those deaths that were unexpected from those potentially anticipated, it is likely that a substantial proportion of these patients had predicable chronic disease progression and may have benefitted from hPC services at an earlier stage of their disease trajectory.
Our study has recognized limitations. Our findings may not be generalizable outside of Ontario. We used administrative data, which did not allow for analysis of patient and family experience, quality and appropriateness of care, and the specific types of services provided. We did not analyze referral sources (eg, primary care versus nephrology). Death location was unconfirmed in up to 38% of patients, as it could only be ascertained through an OHIP claim for pronouncement of death at home, or if the death occurred in an institution. However, it is very likely that most undocumented deaths occurred in the home setting; hence, home deaths likely occurred in 16% for the overall cohort and 50% for patients who had one or more hPC encounter, which is in keeping with patients’ preferences as captured in surveys. 20
These limitations notwithstanding, our findings identify opportunities for improvement and provide baseline measures for gauging the success of future efforts. In a comprehensive provincial strategy, we hope to improve patients’ access to and experience of care by introducing interventions including formalized goals of care assessments, patient and provider education, standardized care pathways, and team-based approaches and by improving collaboration with primary care. 21 Developing these interventions in partnership with patients and families, partner agencies, 5 and the primary care community will be critical to the success of the provincial renal palliative care strategy.
Footnotes
Appendix
Palliative Care Codes by Administrative Database.
| Type | Database | Code |
|---|---|---|
| Acute care services | CIHI-DAD | PATSERV = 58, Main patient service, palliative care |
| OHIP | C945, Special palliative care consult hospital inpatient | |
| C882, Terminal care in-hospital general practitioner/family practitioner | ||
| C982, Palliative care | ||
| K023, Palliative care support individual care 1/2 h or major part | ||
| Home care services | RAI-CA | B2c = 1, Referral to initiate or continue palliative services = Yes |
| B4 = 12, Expected residential/living status during service provision = Hospice facility/palliative care unit | ||
| RAI-HC | P2S = 1 or 2, Special Treatments, Therapies, Programs—Hospice Care= Scheduled, full adherence as prescribed OR Scheduled, partial adherence | |
| CC3f = 1, Understanding of Goals of Care—Palliative Care = Yes | ||
| OHIP | B966, Travel Premium—Palliative Care Home Visit | |
| B998, Special visit palliative care home, days or evenings | ||
| C997, Special visit palliative care home, days or evenings (starting Oct 2009) | ||
| G511, Telephone management of palliative care at home | ||
| HCD: Clients | SRC_admission = 95, Service Care goals at time of submission for open admission= End of Life (In-Home) | |
| Service_RPC = 95, Service Recipient Code associated with the care delivery event = End of Life (In-Home) | ||
| SRC_discharge = 95, Service Care goals (service receipt code) at time of discharge = End of Life (In-Home) |
Note. CIHI-DAD = Canadian Institute for Health Information Discharge Abstract Database; OHIP = Ontario Health Insurance Plan; RAI-CA = Resident Assessment Instrument–Contact Assessment; RAI-HC = Resident Assessment Instrument–Home Care; HCD = Home Care Database.
Acknowledgements
This study was supported by the Institute for Clinical Evaluative Sciences (ICES) Western site. ICES is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). Core funding for ICES Western is provided by the Academic Medical Organization of Southwestern Ontario (AMOSO), the Schulich School of Medicine & Dentistry (SSMD), Western University, and the Lawson Health Research Institute (LHRI). This project was conducted with members of the provincial ICES Kidney, Dialysis and Transplantation Research Program (
), at the ICES Western facility, which receives programmatic grant funding from the Canadian Institutes of Health Research (CIHR) in collaboration with the ORN. Parts of this material are based on data and information compiled and provided by the Canadian Institute for Health Information. The opinions, results and conclusions are those of the authors and are independent from the funding sources. No endorsement by ICES, AMOSO, SSMD, LHRI, CIHR, or the MOHLTC is intended or should be inferred. Dr. Amit Garg is supported by the Dr. Adam Linton Chair in Kidney Health Analytics. Parts of this material are based on data and information compiled and provided by CIHI. However, the analyses, conclusions, opinions and statements expressed herein are those of the author, and not necessarily those of CIHI. Parts of this material are based on data and information provided by Cancer Care Ontario (CCO). The opinions, results, view, and conclusions reported in this paper are those of the authors and do not necessarily reflect those of CCO. No endorsement by CCO is intended or should be inferred.
Ethics Approval and Consent to Participate
This project has been approved by the Research Ethics Board at Sunnybrook Health Sciences Centre, Toronto, Canada.
Consent for Publication
As this work included secondary use of administrative health care data we do not require consent here.
Availability of Data and Materials
The data set from this study is held securely in coded form at the Institute for Clinical Evaluative Sciences (ICES). While data sharing agreements prohibit ICES from making the data set publicly available, access can be granted to those who meet pre-specified criteria for confidential access, available at . The full data set creation plan and underlying analytic code are available from the authors upon request, understanding that the programs may rely upon coding templates or macros that are unique to ICES.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The Ontario Renal Network and the Ontario Ministry of Health and Long-Term Care funded this work.