Patient Care Planning Discussions for Patients at the End of Life: An Evidence-Based Analysis

S Baidoobonso

. 2014 Dec 1;14(19):1–72.

Patient Care Planning Discussions for Patients at the End of Life: An Evidence-Based Analysis

S Baidoobonso

PMCID: PMC4561366 PMID: 26357528

Abstract

Background

Ontario spends about 9% of its health budget on care for people at the end of life (EoL), most of whom die from chronic, prolonged conditions. For many people, patient care planning discussions (PCPDs) can improve the quality and reduce the cost of care.

Objectives

This evidence-based analysis aimed to examine the effectiveness of PCPDs in achieving better patient-centred outcomes for people at the EoL.

Data Sources

A systematic literature search was conducted in MEDLINE, Embase, CINAHL, and EBM Reviews to identify relevant literature published between January 1, 2004, and October 9, 2013.

Review Methods

Peer-reviewed reports from randomized controlled trials (RCTs) and observational studies were examined. Outcomes included quality of life (QoL), satisfaction, concordance, advance care planning (ACP), and health care use. Quality of evidence was assessed using GRADE.

Results

While the effects of PCPDs on QoL are unclear, single-provider PCPDs were associated with family members being very satisfied with EoL care (odds ratio [OR]: 5.17 [95% CI: 1.52, 17.58]), improved concordance between patients’ and families’ wishes (OR: 4.32, P < 0.001), fewer episodes of hospital care (mean difference [MD]: −0.21, P = 0.04), spending fewer days in hospital (MD: −1.8, P = 0.03), and receiving hospice care (OR: 5.17 [95% CI: 2.03, 13.17]). Team-based PCPDs were associated with greater patient satisfaction (standardized mean difference [SMD]: 0.39 [95% CI: 0.17, 0.60]) and fewer outpatient visits (MD: −5.20 [95% CI: −9.70, −0.70]). Overall, PCPDs were associated with more ACP and more optimal health care use.

Limitations

Most of the RCTs were unblinded, intervention was measured or described inadequately in some studies, and the term “usual care” was often undefined.

Conclusions

Patients at the EoL and their families benefited from PCPDs. Furthermore, PCPDs occurring earlier in the course of illness were associated with better outcomes than those occurring later.

Plain Language Summary

In 2009–2010, about 88,000 people in Ontario were near the end of their lives; nearly all of these people were adults who died from chronic illnesses such as cancer, heart disease, stroke, diabetes, and Alzheimer's disease. Providing better-quality health care for people at the end of life has become a priority and research suggests that the quality of care can be improved by patient care planning discussions among health care providers, patients, and families. These discussions focus on designing care for a particular patient on the basis of disease progress, treatment options, preferences, goals, values, and other related considerations. This review was conducted to examine the effects of patient care planning discussions on the quality of care provided and the use of health care resources at the end of life.

This review found that patient care planning discussions with a team of providers from multiple professions were beneficial for patients and their families. High-quality evidence indicates that discussions with a single provider can improve families’ satisfaction with care at the end of a loved one's life and increase agreement between the wishes of the patient and his or her family. These discussions can also reduce the likelihood that patients will need care in hospital and reduce the number of days a patient spends in hospital. Finally, discussions with a single provider increased care planning and the use of hospice services. Moderate- to high-quality evidence shows that patient care planning discussions with a team of providers from multiple professions led to increases in care planning, fewer days in intensive care, and fewer visits for outpatient services.

Background

In July 2013, the Evidence Development and Standards (EDS) branch of Health Quality Ontario (HQO) began work on developing an evidentiary framework for end of life care. The focus was on adults with advanced disease who are not expected to recover from their condition. This project emerged from a request by the Ministry of Health and Long-Term Care that HQO provide them with an evidentiary platform on strategies to optimize the care for patients with advanced disease, their caregivers (including family members), and providers.

After an initial review of research on end-of-life care, consultation with experts, and presentation to the Ontario Health Technology Advisory Committee (OHTAC), the evidentiary framework was produced to focus on quality of care in both the inpatient and the outpatient (community) settings to reflect the reality that the best end-of-life care setting will differ with the circumstances and preferences of each client. HQO identified the following topics for analysis: determinants of place of death, patient care planning discussions, cardiopulmonary resuscitation, patient, informal caregiver and healthcare provider education, and team-based models of care. Evidence-based analyses were prepared for each of these topics.

HQO partnered with the Toronto Health Economics and Technology Assessment (THETA) Collaborative to evaluate the cost-effectiveness of the selected interventions in Ontario populations. The economic models used administrative data to identify an end-of-life population and estimate costs and savings for interventions with significant estimates of effect. For more information on the economic analysis, please contact Murray Krahn at murray.krahn@theta.utoronto.ca.

The End-of-Life mega-analysis series is made up of the following reports, which can be publicly accessed at http://www.hqontario.ca/evidence/publications-and-ohtac-recommendations/ohtas-reports-and-ohtac-recommendations.

▸
End-of-Life Health Care in Ontario: OHTAC Recommendation
▸
Health Care for People Approaching the End of Life: An Evidentiary Framework
▸
Effect of Supportive Interventions on Informal Caregivers of People at the End of Life: A Rapid Review
▸
Cardiopulmonary Resuscitation in Patients with Terminal Illness: An Evidence-Based Analysis
▸
The Determinants of Place of Death: An Evidence-Based Analysis
▸
Educational Intervention in End-of-Life Care: An Evidence-Based Analysis
▸
End-of-Life Care Interventions: An Economic Analysis
▸
Patient Care Planning Discussions for Patients at the End of Life: An Evidence-Based Analysis
▸
Team-Based Models for End-of-Life Care: An Evidence-Based Analysis

Objective of Analysis

This evidence-based analysis (EBA) aimed to examine the effectiveness of patient care planning discussions (PCPDs) in achieving better patient-centred outcomes for people at the end of life (EoL). This EBA is particularly focused on the effectiveness of having a discussion that includes a single health care provider, having a discussion that includes an interprofessional team of health care providers, and the timing of discussions about EoL care. This EBA is part of a mega-analysis focused on EoL care.

Clinical Need and Target Population

End-of-Life Population

It is difficult to know in advance when patients are definitely at the EoL; hence, EoL is defined on the basis of assumptions about a person's risk of dying. These assumptions tend to be related: the period in which the person is expected to die; eligibility for certain health services (e.g., hospice care, palliative care); or the part of life in which patients, family members, informal caregivers, and health care providers struggle with the implications of a chronic illness that has progressed to advanced stages. (1)

Because each of the commonly used definitions of EoL is quite narrow, this EBA and the mega-analysis aimed to be inclusive by defining EoL broadly. Thus, the EoL population was defined as adult patients with advanced disease not expected to recover from their condition or stabilize. This population included people who were seriously or terminally ill, those whom clinicians believed were at the EoL, and those receiving palliative care. This definition aims to incorporate the 3 approaches to identifying patients at the EoL.

From 2007 to 2009, the 3 most recent years for which complete data were available, 264,503 persons died in Ontario. Of these, 261,135 persons (98.7%) were adults aged 18 and older, and chronic, prolonged conditions accounted for most deaths in this group. In Ontario, the top 10 causes of death in 2009 were cancer (29.5%), diseases of the heart (20.9%), cerebrovascular diseases (6.2%), accidents (4.4%), chronic lower respiratory illnesses (4.2%), diabetes mellitus (3.2%), Alzheimer's disease (2.5%), influenza and pneumonia (2.4%), kidney-related diseases (1.4%), and suicide (1.4%). (2) Most of these are advanced chronic conditions, so the above definition is appropriate for identifying the EoL population in Ontario.

Quality of End-of-Life Care and Communication

Quality of EoL care is a domain of the multidimensional quality of dying and death construct (QODD). The QODD consists of 7 broad and overlapping domains: physical, psychological, social, spiritual or existential, nature of health care, life course and death preparation, and circumstances of death. The quality of EoL care depends on the extent to which the health care patients receive affects outcomes in those domains. (3) Additionally, good-quality EoL care must be informed by the best available evidence when appropriate. (4) The health care received should thus be related to a plan of care; the patient's and family's goals, values, needs, and so forth; receiving services that are consistent with the care plan; and applying the best available evidence, among other things. (4)

Open and honest communication between patients, families, and health care providers is essential for achieving good-quality EoL care. Communication exchanges information between patients and providers, (5) and it can help to inform patients about their condition, prognosis, and treatment options as well as elicit patients’ goals, values, and preferences. Given that PCPDs include advance care planning (ACP) and goals of care discussions, this type of intervention can help to improve communication about EoL care. They are important for obtaining informed consent and for sharing decision-making.

Ontario law specifies that, even when an advanced care plan or do-not-resuscitate (DNR) order is in place, decisions about treatments cannot be made without informed consent, which requires health care providers to discuss care options with patients or their substitute decision makers. (6) Patient care planning discussions are thus a necessary and important component of decision-making in health care. Advance care plans and DNR orders are outputs from the PCPDs, and they should be updated if a patient's wishes, values, or beliefs related to care change in any way. More recent expressions of care preferences take precedence over older ones, even if the older ones are written and the more recent ones are verbal. Communication about EoL care has thus become an area of focus for the Canadian Researchers at the End of Life Network (CARENET), a group focused on improving palliative and EoL care. Their conceptual model for improving communication about EoL care is presented in Figure 1.

Figure 1: — Abbreviations: ACP, advance care planning; CARENET, Canadian Researchers at the End of Life Network; MOST, medical orders on scope of treatment; POLST, physician orders for life-sustaining treatment.

Results from recent systematic reviews show that these discussions can be beneficial. For instance, family meetings are associated with lower health care use, and palliative care teams are associated with better use of health care. (7) Palliative care consultations and conferences are associated with increased family satisfaction, improved decision making, and reduced health care use. (8) Ethics consultations also show benefits, because they lead to improved health care use. (7;8) Last, ACP interventions are associated with an increase in ACP documentation, a reduction in the use of health care services, and an increase in hospice use. (9) This evidence indicates PCPDs have the potential to improve the quality of EoL care and reduce health care costs for the province.

Ontario Context

No population-based estimates for the prevalence of PCPDs in Ontario were located, but estimates from 2 hospital-based studies conducted in Canada might provide some insight. A study published in 2013 was conducted with a convenience sample of hospital inpatients in Ontario and 3 other provinces. All patients in the sample had chronic diseases or were aged 80 years or older. Within the sample, 20.1% of 278 patients and 33.2% of 224 family members indicated that they had ever discussed the patient's prognosis and remaining time to live with their physicians. Further, 47.9% of patients and 52.2% of family members indicated that they or their loved one had written documents specifying their treatment preferences; and only 11.9% of patients and 14.7% of family members preferred to receive aggressive care at the EoL. (10) In another study published in 2009, 25.5% (24/108 in Kingston and 13/37 in Toronto) of Ontarians in the sample reported having had a discussion about their prognosis with a physician. This was a multicentre study with a convenience sample of 412 Canadians in 5 tertiary care teaching hospitals. Within the sample as a whole, 18.0% (74/412) of participants had reported having these discussions. The study also showed that people who had prognostic discussions with physicians had higher scores for overall satisfaction and satisfaction with communication and decision making. Their families reported higher overall satisfaction and higher satisfaction with communication and decision making and social support. (11) Although these estimates are based on convenience samples, they are consistent in suggesting that fewer than 30% of Canadians, including Ontarians, are having PCPDs.

Patient care planning discussions also have the potential to meaningfully reduce health care costs in Ontario. Care at the EoL is quite costly for Ontarians, and these costs are expected in increase as the population ages. In 2009–2010, Ontario's health budget was approximately $44.8 billion. The province spends an estimated 9% of its health budget on EoL care. (Ba’ Pham, personal communication, March 19, 2014) Additionally, a report from the Canadian Institute for Health Information states that care for seniors older than 65 years accounted for 44.0% of average health care spending at the national level in 2000 and 45.0% in 2011. This was because of the high cost of EoL care and because health care use for chronic conditions increases with age. (12)

Technology/Technique

“Patient care planning discussions” is an umbrella term used to describe discussions that usually lead to a written medical and nursing care program specifically designed for a particular patient. It encompasses advance care planning or goals of care conversations (i.e., discussions with patients and/or their substitute decision makers about the goals and desired direction of their care). (13) There are many interventions for patient care planning, but this EBA focuses on discussion interventions for reasons described above.

Discussions are complex interventions, because their multiple components can affect their efficacy. Although they can take different forms depending on the context, they adhere to an underlying structure. (14) These interventions can vary by setting, health care personnel, frequency, topics discussed, intensity, structure, and so forth, and therefore must be broken into their constituent parts when they are being evaluated, and each part must be assessed separately. This EBA has deconstructed PCPDs to assess the number of providers involved and the timing of discussions.

Evidence-Based Analysis

Research Question

Which approaches to patient care planning discussions (PCPDs) optimize the quality of end-of-life (EoL) care for patients with advanced disease, informal caregivers, and providers?

Research Methods

Literature Search Strategy

A literature search was performed on October 9, 2013, using Ovid MEDLINE, Ovid MEDLINE In-Process and Other Non-Indexed Citations, Ovid Embase, EBSCO Cumulative Index to Nursing & Allied Health Literature (CINAHL), and EBM Reviews, for studies published from January 1, 2004, to October 9, 2013. (Appendix 1 provides details of the search strategies.) Abstracts were reviewed by a single reviewer and, for those studies meeting the eligibility criteria, full-text articles were obtained. Reference lists were also examined for any additional relevant studies not identified through the search.

Inclusion Criteria

English-language full-text publications
published between January 1, 2004, and October 9, 2013
randomized controlled trials (RCTs), systematic reviews (SRs), meta-analyses, and observational studies
adult patients (aged 18 and older) with advanced disease or who are seriously ill and whose health is likely to continue to deteriorate
adult patients at the EoL
communication between providers and patients
communication between providers and families
communication between patients and their families
goals of care or PCPDs, conferences, conversations, or consultations

Exclusion Criteria

related to sudden or violent death
non-SRs, case reports, editorials, letters, comments, conference abstracts
include children only (younger than 18 years)
most participants are children

Outcomes of Interest

primary outcome is quality of life (QoL)
secondary outcomes are the following:
- patient satisfaction
- family satisfaction
- concordance between patient's wishes and care received
- concordance between patient's wishes and family's or substitute decision maker's wishes
- completion of advance care planning processes or documentation
- health care use

Statistical Analysis

Data Extraction

Data were extracted from the studies with a standardized data form. The form collected information about the following:

Source (i.e., citation information, contact details, study type);
Methods (i.e., study design, study duration and years, participant allocation, allocation sequence concealment, blinding, reporting of missing data, reporting of outcomes, and whether or not the study compared 2 or more groups);
Outcomes (i.e., outcomes measured, number of participants for each outcome, number of participants missing for each outcome, outcome definition and source of information, unit of measurement, upper and lower limits [for scales], and time points at which the outcome was assessed);
Participants (i.e., number of participants, population, diseases and conditions represented, setting, country, age, sex, proportion of racial minorities, proportion of immigrants, proportion of Canadian Aboriginals, marital status, education level, and income or poverty status);
Intervention (number of intervention groups, type of intervention, individuals and professional groups present for the discussions, number of participants in each arm of the study, initiator of the discussion, use of a communication tool to facilitate the discussion, structure of the discussion, timing of the discussion, content of the discussion, and frequency of discussions);
Results (i.e., summary data for each intervention group, effect estimates, and confidence intervals or P values for each effect estimate); and
Other information (i.e., funding source, key conclusions, other information to guide the review, and whether the study's authors needed to be contacted).

Authors of the studies were contacted to provide unpublished data when required for comparisons and meta-analysis.

Assessment of Risk of Bias in Included Primary Studies

The risk of bias assessment was guided by a modified version of the tool in the Cochrane Handbook for Systematic Reviews of Interventions. (15) For randomized controlled trials, bias assessment considered selection bias (i.e., allocation concealment), performance bias (i.e., blinding of participants and health care providers), attrition bias (i.e., incomplete outcome data), reporting bias (i.e., selective outcome reporting), and other limitations (e.g., related to study design). For observational studies, the tool considered selection bias (i.e., appropriate eligibility criteria, adequate control for confounding), measurement bias (i.e., appropriate measurement of exposures and outcomes), and attrition bias (i.e., incomplete follow-up). The results of bias assessment are presented in Appendix 2 and were used to assess the overall quality of evidence for each outcome.

Assessment of Publication Bias

Funnel plots were used to assess potential publication biases, and the results of these assessments were used to rate the quality of the evidence for each outcome (Appendix 2).

Data Synthesis

The studies were divided into 2 subgroups on the basis of whether or not they included patient care planning discussions (PCPDs) with one provider (single-provider) or an interprofessional team of providers working in a coordinated way (team-based). Within the subgroups, studies were pooled if they employed the same study design and used (or did not use) a tool to facilitate discussions.

The EoL population was defined broadly, and this systematic review is concerned with the average effectiveness of PCPDs on EoL care, so a great deal of heterogeneity in the results was expected given the diversity in the EoL population. Hence, while the I² statistic was used to assess heterogeneity in the meta-analyses, it was not used to determine whether or not the results should be pooled. According to the Cochrane Handbook for Systematic Reviews of Interventions, heterogeneity of 0–40% may not be important, heterogeneity between 30% and 60% is moderate, heterogeneity from 50% to 90% is substantial, and heterogeneity between 75% and 100% is considerable. (15) Notably, the importance of I² in assessing heterogeneity depends on the magnitude and direction of the effect estimates and the strength of the evidence for heterogeneity. Given the diversity in the population, certain judgments needed to be made about heterogeneity.

Results were pooled only when the confidence intervals for their effect estimates overlapped. If the I² statistic was greater than 50% and the confidence intervals overlapped, then a random effects model was used to pool the estimates. If the I² statistic was greater than 50% and the confidence intervals did not overlap, the estimates were not pooled. Additional, steps were taken to explain potential reasons for heterogeneity in the studies. For instance, the diseases and conditions in the population, the country and setting, the outcomes measured, and the demographic characteristics of the samples were considered.

Quality of Evidence

The quality of the body of evidence for each outcome was examined according to the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) Working Group criteria. (16) The overall quality was determined to be high, moderate, low, or very low using a step-wise, structural methodology.

Study design was the first consideration; the starting assumption was that randomized controlled trials (RCTs) are high quality, whereas observational studies are low quality. Five additional factors—risk of bias, inconsistency, indirectness, imprecision, and publication bias—were then taken into account. Limitations in these areas resulted in downgrading the quality of evidence. Finally, 3 main factors that can raise the quality of evidence from observation studies were considered: large magnitude of effect, dose-response gradient, and accounting for all residual confounding factors. (16) For more detailed information, please refer to the latest series of GRADE articles. (16)

As stated by the GRADE Working Group, the final quality score can be interpreted using the following definitions:

High	High confidence in the effect estimate—the true effect lies close to the estimate of the effect
Moderate	Moderate confidence in the effect estimate—the true effect is likely to be close to the estimate of the effect, but may be substantially different
Low	Low confidence in the effect estimate—the true effect may be substantially different from the estimate of the effect
Very Low	Very low confidence in the effect estimate—the true effect is likely to be substantially different from the estimate of effect

	Study Design	Number of Eligible Studies
RCTs
Systematic review of RCTs		1
Large RCT		10^a
Small RCT		3
Observational Studies
Systematic review of non-RCTs with contemporaneous controls		3
Non-RCT with contemporaneous controls		11
Systematic review of non-RCTs with historical controls		6
Non-RCT with historical controls		3
Database, registry, or cross-sectional study		3
Case series
Retrospective review, modelling		14
Studies presented at international conference
Expert opinion
Total		54

Author, Year	Country, Setting	Study Design	Length of Study	Sample Size, Intervention/Control	Diseases and Conditions	Age, Intervention/Control	Gender, Intervention/Control	Ethnicity, Intervention/Control
RCTs
Kirchhoff et al, 2012; (40) Kirchhoff et al, 2010 (41)	United States, multiple outpatient clinics	Multicentre RCT	3 years, 7 months	Patients 160/153 Surrogates 160/153	CHF, ESRD	Mean (SD): Patients 71.4 (12.2)/70.6 (11.6) Surrogates 59.5 (13.9)/57.4 (13.6)	Patients 96 (60%) male/89 (58%) male Surrogates 43 (27%) male/41 (27%) male	Patients 149 (93%) white/146 (95%) white Surrogates 153 (96%) white/147 (96%) white
Au et al, 2012 (42)	United States, hospital	Clustered RCT	3 years, 11 months	194/182	COPD	Mean (SD): 69.4 (10.0)/69.4 (10.0)	98% male/96% male^a	85% white/87% white^a
Sampson et al, 2011 (51)	United Kingdom, palliative care unit in hospital	RCT	6 months	Patients 22/10 Carers 22/10	Dementia	Mean (SD): Patients 87 (6.1)/85 (6.6) Carers 60 (13.0)/57 (12.0)	Patients 5 (23%) male/1 (10%) male Carers 14 (64%) male/3 (30%) male	Patients 20 (91%) white/7 (78%) white Carers 20 (91%) white/8 (80%) white
Detering et al, 2010 (39)	Australia, multiple medical units in hospital	RCT	8 months	154/155	Mixed	Median (IQR): 85 (82–88)/84 (81–87)	83 (54%) male/64 (41%) male	Includes minorities^a
Gade et al, 2008 (43)	United States, hospitals	Multicentre RCT	1 year, 7 months	275/237	Mixed	Mean (SD): 73.6 (12.6)/73.1 (13.2)	113 (41%) male/116 (49%) male	NA
Clayton et al, 2007 (44)	Australia, multiple out-patient palliative care clinics	Multicentre RCT	1 year, 11 months	92/82	Cancer	Mean (SD): 65.5 (12.6)/64.6 (14.1)	56 (61%) male/49 (60%) male	NA
Lautrette et al, 2007 (45)	France, ICU in hospital	RCT	9 months	Patients 63/63 Surrogates 63/63	Mixed	Median (IQR): Patients 74 (56–80)/68 (56–76) Surrogates 54 (47–58)/54 (46–64)	Patients 33 (52%) male/37 (59%) male Surrogates 17 (30%) male/12 (23%) male	Patients French descent: 58 (92%)/56 (89%) Surrogates French descent: 48 (86%)/46 (88%)
Gilmer et al, 2005 (46)	United States, ICUs in multiple hospitals	Multicentre RCT	2 years, 2 months	252/247	Mixed	Mean (SD): 67.7 (17.5)/68.5 (17.3)	133 (53%) male/135 (55%) male	155 (62%) white/157 (62%) white
Casarett et al, 2005 (48)	United States, nursing home	RCT	1 year, 1 month	Patients 107/98 Surrogates 88/85	Mixed	Mean (range): Patients 84 (66–102)/83 (54–101) Surrogates 59 (29–88)/57 (23–91)	Patients 27 (25%) male/25 (26%) male Surrogates 22 (25%) male/25 (29%) male	Patients 78 (73%) white/74 (76%) white Surrogates 64 (73%) white/59 (69%) white
Song et al, 2005 (49)	United States, cardiac surgery clinic	RCT	1 year	Patients 16/16 Surrogates 16/16	Cardiac disease	Mean (SD): Patients 69.8 (8.6)/68.0 (8.0) Surrogates 64.4 (11.6)^b	Patients 8 (50%) male/9 (56%) male Surrogates 34% male^b	Patients 100% white/100% white Surrogates 100% white/100% white
Nicolasora et al, 2006 (47)	United States, medical unit in hospital	RCT	3 months	136/161	Mixed	Median: 65/69	54% male/46% male^a	77% white/70% white^a
Dyar et al, 2012 (50)	United States, hospital	RCT	9 months	12/14	Cancer	Mean (SD): 66.7 (16.3)/64.9 (7.5)	25% male/36% male	NA
Jones et al, 2011 (52)	United States, multiple outpatient clinics	Multicentre RCT	1 year, 9 months	Preference cohort 21/14 Randomized cohort 22/20	Cancer	Mean (SD): Preference cohort 62.0 (11.0)/67.7 (7.9) Randomized cohort 58.6 (8.1)/60.2 (13.3)	Preference cohort 48% male/50% male^a Randomized cohort 57% male/50% male	Preference cohort 95% white/93% white^a Randomized cohort 86% white/95% white^a
Observational with contemporaneous controls
Evangelista et al, 2012 (54)	United States, hospital	Prospective case-control	5 months	36/36	Mixed	Mean (SD): 53.9 (8.3)/53.3 (8.7)	26 (72%) male/25 (69%) male	22 (61%) white/22 (61%) white
Jacobsen et al, 2011 (55)	United States, hospital	Non-RCT	7 months	517/382	NA	Mean: 62.9/63.5	308 (60%) male/217 (57%) male	NA
Engelhardt et al, 2009 (56)	United States, health insurance network	Multicentre non-RCT	NA	Patients 198/205 Informal Caregivers 100/85	Mixed	Mean (SD): Patients 66.0 (12.1)/68.4 (12.1) Informal Caregivers, 59.9 (14.6)/63.8 (13.1)	Patients 63 (32%) male/108 (53%) male Informal Caregivers 40 (41%) male/9 (12%) male	Patients 167 (85%) white/183 (91%) white Informal Caregivers NA
Rabow et al, 2004 (57)	United States, home and community	Non-RCT	NA	50/40	Cancer, COPD, advanced CHF	Mean (SD): 67.9 (13.9)/69.4 (11.2)	13 (26%) male/19 (47%) male	22 (44%) white/26 (65%) white
Mack et al, 2012 (58)	United States, health insurance networks and Veterans Affairs health network	Multicentre prospective cohort	2 years	1,231^b	Lung or colorectal cancer	Range: No. (%) 21–54: 172 (14%)^b 55–59: 149 (12%)^b 60–64: 157 (13%)^b ≥65: 753 (61%)^b	766 (62%) male^b	935 (76%) white^b
Mack et al, 2010 (59)	United States, multiple hospital sites	Multicentre prospective cohort	5 years	332^b	Cancer	Mean (SD): Stratified by race^b Black: 55.6 (11.1) White: 60.5 (11.9)	185 (56%) male^b	261 (79%) white^b
Wright et al, 2008 (60)	United States, outpatient cancer clinics	Multicentre prospective cohort	5 years, 6 months	123/209	Cancer	Mean (SD): 57.5 (12.0)/58.0 (12.3)	66 (54%) male/117 (56%) male	84 (68%) white/128 (61%) white
Casarett et al, 2008 (61)	United States, Veterans Affairs health network	Multicentre retrospective cohort	10 months	296/228	Mixed	Mean (range): 72 (27–93)/72 (26–100)	99% male/97% male^a	77% white/74% white^a
Morrison et al, 2005 (62)	United States, nursing home	Non-RCT	1 year, 10 months	43/96	Mixed	Mean (range): 87 (75–100)/86 (65–102)	16% male/16% male^a	67% white/77% white^a
Zhang et al, 2009 (53)	United States, multiple hospitals	Multicentre prospective cohort	5 years, 4 months	188/415	Cancer	Mean (SD): 59.8 (12.9)/58.6 (13.2)	51% male/51% male^a	74% white/70% white^a
Briggs et al, 2004 (63)	United States, multiple units in a hospital	Non-RCT	NA	Patients 13/14 Surrogates 13/14	Heart failure, renal failure	Mean (SD): Patients 68.7 (9.2)^b Surrogates 50 (14.8)^b	Patients 39% male/79% male^a Surrogates 23% male/29% male^a	NA
Observational with historical controls
Lamba et al, 2012 (64)	United States, surgical ICU in hospital	Prospective before and after	1 year, 1 month; 1 year, 3 months	31/21	Mixed	Mean: 54/46	19 (61%) male/12 (57%) male	NA
Norton et al, 2007 (65)	United States, medical ICU in hospital	Prospective before and after	1 year, 1 month	126/65	Mixed	Mean (SD): 66.3 (16.3)/68.8 (15.4)	57 (45%) male/33 (51%) male	95 (75%) white/51 (79%) white
Lindner et al, 2007 (66)	United States, nursing home	Prospective before and after	6 months	107/117	Mixed	Mean (SD): 72 (12.2)/71 (12.3)	91% male/89% male^a	70% white/57% white^a
Cross-sectional
Leung et al, 2012 (67)	United States, Veterans Affairs health network	Multicentre cross-sectional	2 years, 2 months	55/321	COPD	Mean: 70.9/69.1	52 (95%) male/313 (98%) male	44 (80%) white/247 (77%) white
Mori et al, 2013 (68)	United States, hospital	Cross-sectional	2 years	20/29	Cancer	Mean (SD): 60.1 (9.8)/67.5 (11.7)	10 (50%) male/14 (48%) male	NA
Heyland et al, 2009 (11)	Canada, multiple hospitals	Multicentre cross-sectional	NA	Patients 74/338 Family members 46/107	Mixed	Mean (SD): Patients 67.6 (7.2)/71.9 (9.4) Family members 56.6 (14.0)/57.4 (13.5)	Patients 51% male/51% male^a Family members 40% male/34% male^a	NA

Author, Year	Interventions	Comparisons	Outcomes of Interest
Randomized controlled trials (RCTs)
Kirchhoff et al, 2012; (40) Kirchhoff et al, 2010 (41)	Tool: facilitation tool Type: single-provider; nurse, social worker, or chaplain who was trained to deliver the intervention Duration: 1.0–1.5 hours Frequency: NA Topics: patient and surrogate understanding, information about treatment, assistance with documentation of care preferences, surrogate understanding of patient preferences, preparing surrogate to make decisions that honour patient preferences; completion of STP	Usual care: standard AD counseling, assessment of AD on admission, question whether patient wanted more information, referrals according to institutional protocol	Concordance between patient wishes and care received Concordance between patient's and surrogate's wishes
Au et al, 2012 (42)	Tool: automated form Type: single-provider; physician Duration: NA Frequency: NA Topics: barriers and facilitators to EoL communication, patient preferences	Usual care^a	Completion of ACP documents and processes
Sampson et al, 2011 (51)	Tool: framework Type: single-provider; nurse Duration: NA Frequency: up to 4 consultations at least 5 days apart Topics: knowledge; illness severity and prognosis; patient's physical needs; social situation and social support; previous care preferences; cultural, spiritual, health, social, financial needs; role of palliative care; ACP Other: included a needs assessment, discussion with carers and family, summary of discussion in a standard form	Usual care^a	Informal caregiver's QoL Carer and family's satisfaction with care
Detering et al, 2010 (39)	Tool: Respecting Patient Choices model Type: single-provider; nurse who was trained to deliver the intervention; a physician was sometimes included Duration: Median length of discussion was 60 minutes (range 10–200 minutes) Frequency: NA Topics: prognosis, goals, values, beliefs, discuss and document choices for future health care, ACP Other: offered on top of usual care	Usual care: no ACP unless specifically requested	Patient's satisfaction with care Informal caregiver's and family's satisfaction with care Completion of ACP documents and processes
Gade et al, 2008 (43)	Tool: Weismann's palliative care components Type: team-based; physician, nurse, social worker, and chaplain Duration: NA Frequency: NA Topics: symptom management, psychosocial and spiritual support, EoL planning, posthospital care, diagnosis, prognosis, goals of care, ACP	Usual care^a	Patient's QoL Patient's satisfaction with care Completion of ACP documents and processes Hospital LOS
Clayton et al, 2007 (44)	Tool: question prompt list with 112 potential questions that was reviewed 20–30 minutes before the consultation Type: single-provider; physician Duration: NA Frequency: NA Topics: palliative care, physical symptoms, treatment, lifestyle and QoL, prognosis, support, EoL issues	Usual care: routine consultation	Patient's satisfaction with care
Lautrette et al, 2007 (45)	Tool: VALUE mnemonic, a brochure, and detailed procedures Type: team-based; physician and other professionals Duration: Median length of discussion was 30 minutes (range 19–45 minutes) Frequency: NA Topics: NA	Usual care: EoL family conference	ICU LOS
Gilmer et al, 2005 (46)	Tool: process model Type: team-based; ethicist consultant and other professionals Duration: NA Frequency: NA Topics: ethical issues, hospital policy, ethical consensus statements, statutes, laws, measures to improve communication, medical factors, patient preferences and values, QoL, context	Usual care: no ethics consultation was offered	Hospital LOS ICU LOS
Casarett et al, 2005 (48)	Tool: structured interview questionnaire Type: single-provider; physician Duration: NA Frequency: NA Topics: goals of care, life-sustaining treatment, symptoms, palliative care needs (i.e., comfort care, nursing support, practical support, ACP, emotional support, bereavement support, spiritual support, address loneliness)	Usual care: assessment by research assistant	Family's satisfaction with care Hospital care Hospital LOS
Song et al, 2005 (49)	Tool: Representational Approach model Type: single-provider; nurse Duration: NA Frequency: NA Topics: ACP, illness experiences, misconceptions and concerns, informed choices about medical care, treatment, complications, benefits Other: offered in addition to usual care	Usual care: information cards and a booklet if more information was desired	Concordance between patient's and surrogate's wishes
Nicolasora et al, 2006 (47)	Tools: script that provides detailed information about life-sustaining treatments and ADs Type: single-provider; physician Duration: NA Frequency: NA Topics: completion of ACP, CPR status, patient wishes for change, AD creation	Usual care: not approached by physician	Completion of ACP documents and processes
Dyar et al, 2012 (50)	Tool: none Type: single-provider; nurse practitioner Duration: NA Frequency: NA Topics: Five Wishes (a living will), information about hospices, needs assessment	Usual care^a	Patient's QoL
Jones et al, 2011 (52)	Tool: none Type: single-provider; nurse or physician who was trained to deliver the intervention Duration: NA Frequency: NA Topics: patient perceptions, communication with providers and family, hopes and fears for future, health care decision-making Other: family members were not invited to the first consultation, but they were included in subsequent sessions	Usual care^a	Patient's satisfaction with care
Observational with contemporaneous controls
Evangelista et al, 2012 (54)	Tool: none Type: single-provider; nurse or physician Duration: 50–120 minutes Frequency: NA Topics: patient's values, goals, preferences; areas where patient perceived the need for support (e.g., physical, symptom control, emotional support, illness understanding, treatment goals, ACP); establish treatment plan	Usual care: no palliative care consultation	Patient's QoL
Jacobsen et al, 2011 (55)	Tool: cognitive model for ACP Type: single-provider; physician, nurse, or social worker Duration: NA Frequency: NA Topics: information sharing ACP meetings, understanding, values, goals, preferences, selection of surrogate, prognosis, hopes, worries, CPR, place of death	Usual care^a	Completion of ACP documents and processes
Engelhardt et al, 2009 (56)	Tool: checklist Type: single-provider; social worker or health educator Duration: NA Frequency: took place over 6 sessions Topics: understanding, treatment, symptoms, prognosis, communication with providers, ACP, coping, existential issues, caregiving concerns, long-term care planning, home environmental modifications, etc. Other: in-person discussion followed by phone contact	Usual care^a	Patient's QoL Completion of ACP documents and processes Hospital care Emergency department visits Home health visits Outpatient services
Rabow et al, 2004 (57)	Tool: none Type: team-based; social worker, nurse, chaplain, pharmacist, psychologist, art therapist, volunteer coordinator, and 3 physicians Duration: NA Frequency: took place over 3 sessions Topics: NA Other: provided in addition to usual care	Usual care^a	Patient's QoL Patient's satisfaction with care Completion of ACP documents and processes Hospital care Hospital LOS Emergency department visits Urgent care visits Outpatient services
Mack et al, 2012 (58)	Tool: none Type: single-provider; 40% of discussions included a physician Duration: NA Frequency: NA Topics: ACP (i.e., do not resuscitate, hospices, palliative care, or not otherwise specified), venue of dying	No discussion^a	Chemotherapy Hospital care ICU admission Hospice care
Mack et al, 2010 (59)^a	Tool: none (assumed) Type: single-provider (assumed) Duration: NA Frequency: NA Topics: NA	No discussion^a	Concordance between patient's wishes and care received Hospice care
Wright et al, 2008 (60)^a	Tool: none (assumed) Type: single-provider (assumed) Duration: NA Frequency: NA Topics: NA	No discussion^a	Chemotherapy Hospice care
Casarett et al, 2008 (61)	Palliative care consultation Tool: none Type: team-based; physician, nurse practitioner, or both; social worker; nurse; chaplain; volunteers; and other professionals as needed Duration: NA Frequency: NA Topics: NA	Usual care^a	Patient's QoL
Morrison et al, 2005 (62)	Tool: aid to structure discussions Type: single-provider; social worker trained to deliver intervention Duration: NA Frequency: NA Topics: patient's decisional capacity, information on how to appoint proxies, communication with proxies, ACP, goals of care, feedback to physicians on congruence	Usual care^a	Concordance between patient's wishes and care received
Zhang et al, 2009 (53)^a	Tool: none (assumed) Type: single-provider; physician Duration: NA Frequency: NA Topics: preferences^a	No discussion^a	Patient's satisfaction Chemotherapy Resuscitation ICU admission Hospice care
Briggs et al, 2004 (63)	Tool: aid to structure discussions Type: single-provider; nurse, chaplain, or social worker who was trained to deliver the intervention Duration: 1 hour Frequency: NA Topics: patient and surrogate understanding, information about treatment, assistance with ACP documentation, surrogate's understanding of patient's preferences, preparing surrogate to make decisions that honour patient preferences, complete STP	Usual care: ACP literature, referral to trained ACP facilitator	Concordance between patient's and surrogate's wishes
Observational with historical controls
Lamba et al, 2012 (64)	Tool: none Type: team-based; Part I (physician and nurse), Part II (physician, nurse, counsellor, and interfaith pastor) Duration: NA Frequency: NA Topics: Part I (prognosis, ADs, family support, surrogate decision-maker, pain, other symptoms), Part II (likely patient outcomes; treatment options; goals of care; assessment of content on goals of care discussion, family understanding, issues of conflict)	Usual care^a	Completion of ACP documents and processes Hospital LOS ICU LOS
Norton et al, 2007 (65)	Tool: none Type: team-based; nurse practitioner, physician, chaplain, music therapist, massage therapist, and psychologist Duration: NA Frequency: NA Topics: medical history and physical exam, palliative care, involvement with team and family, other providers	Usual care: physician makes a palliative care referral	Hospital LOS ICU LOS
Lindner et al, 2007 (66)	Tool: electronic medical record note Type: single-provider; physician Duration: NA Frequency: NA Topics: information about surrogates, dates and content of prior written ADs, values, preferences, goals of care, wishes regarding resuscitation, artificial nutrition, hydration, hospital transfer, autopsy	Usual care^a	Concordance between patient's wishes and care received Completion of ACP documents and processes
Cross-sectional
Leung et al, 2012 (67)^a	Tool: none (assumed) Type: single-provider; physician Duration: NA Frequency: NA Topics: care planning^a	No discussion^a	Patient's satisfaction with care
Mori et al, 2013 (68)^a	Tool: none (assumed) Type: single-provider; health care provider Duration: NA Frequency: NA Topics: NA	No discussion^a	Family's satisfaction with care
Heyland et al, 2009 (11)^a	Tool: none (assumed) Type: single-provider; health care provider Duration: NA Frequency: NA Topics: NA	No discussion^a	Patient's satisfaction with care Family's satisfaction with care

Author, Year Instrument Range	Study Design	Tool Used?	Results Intervention	Results Control	Difference in Means (95% CI)
Single-Provider Discussion vs. Usual Care or No Discussion GRADE for highest-quality evidence: Very low
Highest-quality evidence
Dyar et al, 2012 (50) FACT-G (0–108)	Small RCT	No	Mean change in FACT-G score from baseline (SD): 1.2 (12.5)	Mean change in FACT-G score from baseline (SD): −3.9 (5.0)	5.10 (−3.98, 14.18) SMD: 0.47 (−0.58, 1.52)
Dyar et al, 2012 (50) LASA (12–120)	Small RCT	No	Mean change in LASA score from baseline (SD): 2.0 (25.7)	Mean change in LASA score from baseline (SD): −8.8 (21.7)	10.80 (−11.10, 32.70) SMD: 0.44 (−0.52, 1.37)
Evangelista et al, 2012 (54) ^a MLHFQ (0–105)	Obs-cont	No	Mean change in MLHFQ score from baseline: 9.9 (17.3)^b	Mean change in MLHFQ score from baseline: 4.3 (15.6)^b	5.60 (−2.01, 13.21)^b,c,d SMD: 0.34 (−0.13, 0.80)^b
Engelhardt et al, 2009 (56) McGill Quality of Life Questionnaire (0–10)	Obs-cont	Yes	Mean change in score from baseline: 0.1 (1.2)^b	Mean change in score from baseline: 0.1 (1.5)^b	0.01 (−0.25, 0.27)^b SMD: 0.01 (−0.19, 0.20)^b
Team-Based Discussion vs. Usual Care or No Discussion GRADE for highest-quality evidence: Moderate GRADE for all evidence: Very low
Highest-quality evidence
Gade et al, 2008 (43) 11-point Likert scale (0–10)	Large RCT	Yes	Mean change in score from baseline: 2.3 (3.7)^b	Mean change in score from baseline: 2.2 (3.5)^b	0.10 (−0.61, 0.81) SMD: 0.03 (−0.17, 0.23)
Lower-quality evidence
Rabow et al, 2004 (57) Multidimensional Quality of Life Scale—Cancer Version (0–100)	Obs-cont	No	Mean change in score from baseline: −0.4	Mean change in score from baseline: 2.3	−2.7, P = 0.43^e
Casarett et al, 2008 (61) FATE (0–100)	Obs-cont	No	Mean FATE score: 64	Mean FATE score: 54	10, P < 0.001^d,e

Author, Year Instrument Range	Study Design	Tool Used?	Results Intervention	Results Control	Difference in Means (95% CI)
Single-Provider Discussion vs. Usual Care or No Discussion GRADE for all evidence: Very low
Sampson et al, 2011 (51)^a EuroQoL EQ-5D (0–100)	Small RCT	Yes	Score at baseline (SD): 69.8 (23.6) Score at 6 wk (SD): 73.6 (11.1) Score at 6 wk (SD): 73.6 (11.1) Score at 6 mo (SD): 80.0 (6.1) Score post-death (SD): 69.3 (24.7)	Score at baseline (SD): 62.7 (37.5) Score at 6 wk (SD): 79.8 (12.2) Score at 6 wk (SD): 79.8 (12.2) Score at 6 mo (SD): 80.8 (13.2) Score post-death: 92.0 (1 person)	NA

Author, Year Instrument Range	Study Design	Tool Used?	Results Intervention	Results Control	Effect Estimate (95% CI)
Single-Provider Discussion vs. Usual Care or No Discussion GRADE for highest-quality evidence: Moderate GRADE for all evidence: Very low
Highest-quality evidence
Detering et al, 2010 (39) 3-point Likert	Large RCT	Yes	Very satisfied: 125/133 (94.0%) Satisfied or very satisfied: 131/133 (98.5%)	Very satisfied: 91/139 (65.5%) Satisfied or very satisfied: 131/139 (94.2%)	OR: 8.24 (3.72, 18.26)^a OR: 4.00 (0.83, 19.19)
Clayton et al, 2007 (44)^b,c 25-item scale (25–125)	Large RCT	Yes	Mean score: 110.1	Mean score: 110.3	MD: −0.2 (−3.4, 2.9)
Jones et al, 2011 (52) 5-item scale (range not reported)	Small RCT	No	Mean change in score from baseline (SD): 0.6 (1.5)	Mean change in score from baseline (SD): 1.9 (1.1)	MD: −1.3 (−2.09, −0.51)^a SMD: −0.96 (−1.61, −0.32)
Lower-quality evidence
Zhang et al, 2009 (53)^d 11-point Likert (0–10)	Obs-cont	No	Mean score (SD): 6.3 (2.7)	Mean score (SD): 5.7 (3.3)	MD: 0.60 (−0.39, 1.59) SMD: 0.20 (−0.13, 0.53)
Leung et al, 2012 (67) 5-point Likert dichotomized as very satisfied versus less satisfied	Cross-sectional	No	NA	NA	OR: 2.02 (1.16, 3.50)^a
Heyland et al, 2009 (11) CARENET's Family Satisfaction Survey (1–100)	Cross-sectional	No	Mean score (SD): 76.1 (9.7)	Mean score (SD): 73.1 (10.6)	MD: 3.00 (0.53, 5.47)^a SMD: 0.29 (0.03, 0.54)
Team-Based Discussion vs. Usual Care or No Discussion GRADE for highest-quality evidence: High GRADE for all evidence: Low
Highest-quality evidence
Gade et al, 2008 (43)^b 11-point Likert (0–10)	Large RCT	Yes	Mean score (SD): 8.0 (1.4)	Mean score (SD): 7.4 (1.7)	MD: 0.60 (0.27, 0.93)^a SMD: 0.39 (0.17, 0.60)
Lower-quality evidence
Rabow et al, 2004 (57) Group Health Association of American Consumer Satisfaction Survey (20–100)	Obs-cont	No	Mean change in score between 6- and 12-month evaluations: 0.5^e	Mean change in score between 6- and 12-month evaluations: −2.1^e	MD: 2.6, P = 0.26^f

Author, Year Instrument Range	Study Design	Tool Used?	Results Intervention	Results Control	Effect Estimate (95% CI)
Single-Provider Discussion vs. Usual Care or No Discussion GRADE for highest-quality evidence: High GRADE for all evidence: Moderate
Highest-quality evidence
Detering et al, 2010 (39)^a 3-point Likert	Large RCT	Yes	Very satisfied: 24/29 (82.8%) Satisfied or very satisfied: 26/29 (89.7%)	Very satisfied: 13/27 (48.1%) Satisfied or very satisfied: 21/27 (77.8%)	OR: 5.17 (1.52, 17.58)^b OR: 2.48 (0.55, 11.10)
Casarett et al, 2005 (48)^c Toolkit Afterdeath Survey (1–5)	Large RCT	Yes	Mean score (SD): 4.3 (1.0)	Mean score (SD): 2.2 (1.5)	MD: 2.10 (1.75, 2.45)^b SMD: 1.67 (1.35, 1.99)
Sampson et al, 2011 (51) Satisfaction with End-of-Life Care in Advanced Dementia Scale (10–40)	Small RCT	Yes	Mean score (SD): 27.6 (8.5)	Mean score (SD): 23.0 (1 person)	MD: 4.6^c
Lower-quality evidence
Mori et al, 2013 (68)^d Toolkit of Instruments to Measure End-of-Life Care (0–10)	Cross-sectional	No	Mean score (SD): 9.7 (0.6)	Mean score (SD): 8.7 (1.4)	MD: 1.00 (0.42, 1.58)^b SMD: 0.84 (0.24, 1.43)
Heyland et al, 2009 (11) CARENET's Family Satisfaction Survey (1–100)	Cross-sectional	No	Mean score (SD): 75.2 (13.3)	Mean score (SD): 70.4 (12.2)	MD: 4.89 (0.41, 9.37)^b SMD: 0.39 (0.04, 0.74)
Pooled estimate (2 cross-sectional studies, FE), I² = 39%					SMD: 0.50 (0.20, 0.80)^b

Author, Year	Study Design	Tool Used?	Results Intervention	Results Control	Odds Ratio (95% CI)
Single-Provider Discussion vs. Usual Care or No Discussion GRADE for highest-quality evidence: High GRADE for all evidence: Moderate
Highest-quality evidence
Kirchhoff et al, 2012; (40) Kirchhoff et al, 2010 (41)	Large RCT	Yes	Concordant: 46/62 (74.2%)	Concordant: 30/48 (62.5%)	1.73 (0.76, 3.90)
Lower-quality evidence
Mack et al, 2010 (59)^a	Obs-cont	No	Concordant: 87/113 (77.0%)	Concordant: 137/219 (62.6%)	2.00 (1.19, 3.36)^b
Morrison et al, 2005 (62)	Obs-cont	No	Concordant: 47/49 (95.9%)	Concordant: 79/96 (82.3%)	5.06 (1.12, 22.87)^b
Pooled estimate (2 obs-cont studies, FE), I² = 24%					2.28 (1.41, 3.70)^b
Lindner et al, 2007 (66)	Obs-hist	Yes	Concordant: 39/40 (97.5%)	Concordant: 38/44 (86.4%)	6.16 (0.71, 53.59)

Author, Year	Study Design	Tool Used?	Results
Single-Provider Discussion vs. Usual Care or No Discussion GRADE for all evidence: Low
Mack et al, 2012	Obs-cont	No	Proportion receiving chemotherapy in the last 14 days of life
			Days between first discussion and death: ≤ 30: 65/311 (20.9%) 31–60: 19/186 (10.2%) 61–90: 9/108 (8.3%) > 90: 23/189 (12.2%) χ²: 17.057, df=3, P < 0.001^a

Author, Year	Study Design	Tool Used?	Results
Single-Provider Discussion vs. Usual Care or No Discussion GRADE for all evidence: Moderate
Mack et al, 2012 (58)	Obs-cont	No	Receiving any hospice care
			Days between first discussion and death: ≤ 30: 152/311 (48.9%) 31–60: 126/186 (67.7%) 61–90: 80/108 (74.1%) > 90: 146/189 (77.2%) χ²: 50.756, df = 3, P < 0.001^a

Panel Member	Affiliation(s)	Appointment(s)
Panel Co-Chairs
Dr Robert Fowler	Sunnybrook Research Institute University of Toronto	Senior Scientist Associate Professor
Shirlee Sharkey	St. Elizabeth Health Care Centre	President and CEO
Professional Organizations Representation
Dr Scott Wooder	Ontario Medical Association	President
Health Care System Representation
Dr Douglas Manuel	Ottawa Hospital Research Institute University of Ottawa	Senior Scientist Associate Professor
Primary/Palliative Care
Dr Russell Goldman	Mount Sinai Hospital, Tammy Latner Centre for Palliative Care	Director
Dr Sandy Buchman	Mount Sinai Hospital, Tammy Latner Centre for Palliative Care Cancer Care Ontario University of Toronto	Educational Lead Clinical Lead QI Assistant Professor
Dr Mary Anne Huggins	Mississauga Halton Palliative Care Network; Dorothy Ley Hospice	Medical Director
Dr Cathy Faulds	London Family Health Team	Lead Physician
Dr José Pereira	The Ottawa Hospital University of Ottawa	Professor, and Chief of the Palliative Care program at The Ottawa Hospital
Dean Walters	Central East Community Care Access Centre	Nurse Practitioner
Critical Care
Dr Daren Heyland	Clinical Evaluation Research Unit Kingston General Hospital	Scientific Director
Oncology
Dr Craig Earle	Ontario Institute for Cancer Research Cancer Care Ontario	Director of Health Services Research Program
Internal Medicine
Dr John You	McMaster University	Associate Professor
Geriatrics
Dr Daphna Grossman	Baycrest Health Sciences	Deputy Head Palliative Care
Social Work
Mary-Lou Kelley	School of Social Work and Northern Ontario School of Medicine Lakehead University	Professor
Emergency Medicine
Dr Barry McLellan	Sunnybrook Health Sciences Centre	President and Chief Executive Officer
Bioethics
Robert Sibbald	London Health Sciences Centre University of Western Ontario	Professor
Nursing
Vicki Lejambe	Saint Elizabeth Health Care	Advanced Practice Consultant
Tracey DasGupta	Sunnybrook Health Sciences Centre	Director, Interprofessional Practice
Mary Jane Esplen	De Souza Institute University of Toronto	Director Clinician Scientist

#	Searches	Results
1	exp Terminal Care/	86347
2	exp Palliative Care/ use mesz,acp,cctr,coch,clcmr,dare,clhta,cleed	41169
3	exp palliative therapy/ use emez	61505
4	exp Terminally Ill/ use mesz,acp,cctr,coch,clcmr,dare,clhta,cleed	5628
5	exp terminally ill patient/ use emez	5936
6	exp terminal disease/ use emez	4501
7	exp dying/ use emez	5665
8	((End adj2 life adj2 care) or EOL care or (terminal^∗ adj2 (care or caring or ill^∗ or disease^∗)) or palliat^∗ or dying or (Advanced adj3 (disease^∗ or illness^∗)) or end stage^∗).ti,ab.	336689
9	or/1–8	430311
10	exp Patient Care Planning/	79710
11	((goal^∗ adj2 care) or patient care plan^∗ or advance care plan^∗).ti,ab.	5370
12	((communicat^∗ or conversation^∗ or conference^∗ or discuss^∗ or consult^∗) adj2 (strateg^∗ or plan^∗ or model^∗ or intervention^∗ or framework^∗ or care or program^∗ or process^∗)).ti,ab.	83458
13	or/10–12	165666
14	9 and 13	10891
15	limit 14 to english language [Limit not valid in CDSR,ACP Journal Club,DARE,CCTR,CLCMR; records were retained]	10171
16	limit 15 to yr=“2004 -Current” [Limit not valid in DARE; records were retained]	6293
17	limit 16 to yr=“2009 -Current” [Limit not valid in DARE; records were retained]	4023
18	remove duplicates from 17	2853
19	limit 16 to yr=“2004 – 2008” [Limit not valid in DARE; records were retained]	2270
20	remove duplicates from 19	1473
21	18 or 20	4326

#	Query	Results
S1	(MH “Terminal Care+”)	38,947
S2	(MH “Palliative Care”)	19,702
S3	(MH “Terminally Ill Patients+”)	7,636
S4	((End N2 life N2 care) or EOL care or (terminal^∗ N2 (care or caring or ill^∗ or disease^∗)) or palliat^∗ or dying or (advanced N3 (disease^∗ or illness^∗)) or end stage^∗)	52,178
S5	S1 OR S2 OR S3 OR S4	60,161
S6	(MH “Patient Care Plans+”)	6,983
S7	((goal^∗ N2 care) or patient care plan^∗ or advance care plan^∗)	6,438
S8	((communicat^∗ or conversation^∗ or conference^∗ or discuss^∗ or consult^∗) N2 (strateg^∗ or plan^∗ or model^∗ or intervention^∗ or framework^∗ or care or program^∗ or process^∗))	19,731
S9	S6 OR S7 OR S8	29,515
S10	S5 AND S9	2,848
S11	S5 AND S9 Limiters – Published Date: 20040101-20131231; English Language	2,237

Number of Studies (Design)	Risk of Bias	Inconsistency	Indirectness	Imprecision	Publication Bias	Upgrade Considerations	Quality
Patient's Quality of Life
1 (RCT)	Serious limitations (−1)^a	No serious limitations	No serious limitations	Serious limitations (−1)^b	Undetected	Large magnitude of effect (+1)	⊕ Very low
2 (observational)
Informal Caregiver's Quality of Life
1 (RCT)	Very serious limitations (−2)^c	No serious limitations	No serious limitations	Serious limitations (−1)^b	Undetected	None	⊕ Very low
Patient's Satisfaction with End-of-Life Care
3 (RCTs)	No serious limitations	Serious limitations (−1)^d	No serious limitations	No serious limitations	Undetected	None	⊕⊕⊕ Moderate
Family's Satisfaction with End-of-Life Care
3 (RCTs)	No serious limitations	No serious limitations	No serious limitations	No serious limitations	Undetected	None	⊕⊕⊕⊕ High
Concordance Between Patient's Wishes and Care Received
1 (RCT)	No serious limitations	No serious limitations	No serious limitations	No serious limitations	Undetected	None	⊕⊕⊕⊕ High
Concordance Between Patient's and Family's Wishes
2 (RCTs)	No serious limitations	No serious limitations	No serious limitations	No serious limitations	Undetected	None	⊕⊕⊕⊕ High
Completion of Advance Care Planning Documents and Processes
3 (RCTs)	No serious limitations	No serious limitations	No serious limitations	No serious limitations	Undetected	None	⊕⊕⊕⊕ High
Receiving Chemotherapy at End of Life
3 (observational)	Serious limitations (−1)^e,f	No serious limitations	No serious limitations	No serious limitations	Undetected	Large magnitude of effect (+1)	⊕⊕ Low
Resuscitation
2 (observational)	Serious limitations (−1)^e,f	No serious limitations	No serious limitations	Serious limitations (−1)^b	Undetected	Large magnitude of effect (+1)	⊕ Very low
Receiving Hospital Care
1 (RCT)	No serious limitations	No serious limitations	No serious limitations	No serious limitations	Undetected	None	⊕⊕⊕⊕ High
Hospital Length of Stay
1 (RCT)	No serious limitations	No serious limitations	No serious limitations	No serious limitations	Undetected	None	⊕⊕⊕⊕ High
Emergency Department Visits
1 (observational)	No serious limitations	No serious limitations	No serious limitations	No serious limitations	Undetected	None	⊕⊕ Low
Intensive Care Unit Care
3 (observational)	Serious limitations (−1)^e,f	Serious limitations (−1)^d	No serious limitations	No serious limitations	Undetected	Large magnitude of effect (+1)	⊕ Very low
Home Health Visits
1 (observational)	No serious limitations	No serious limitations	No serious limitations	Serious limitations (−1)^b	Undetected	None	⊕ Very low
Other Outpatient Visits
1 (observational)	No serious limitations	No serious limitations	No serious limitations	Serious limitations (−1)^b	Undetected	Large magnitude of effect (+1)	⊕⊕ Low
Receiving Hospice Care
1 (RCT)	No serious limitations	No serious limitations	No serious limitations	Serious limitations (−1)^b	Undetected	None	⊕⊕⊕⊕ High

Outcome	Single-Provider Discussion vs. Usual Care or No Discussion		Team-Based Discussion vs. Usual Care or No Discussion
	Effect Estimate (95% CI)	GRADE	Effect Estimate (95% CI)	GRADE
Patient's QoL	*SMD* 0.47 (−0.58, 1.52) 0.44 (−0.52, 1.37) 0.34 (−0.13, 0.80)^a,b 0.01 (−0.19, 0.20)	Very low	*SMD* 0.03 (−0.17, 0.23)	Moderate
Informal caregiver's QoL	Estimates NA	Very low	NA
Patient's satisfaction with EoL care	OR Very satisfied: 8.24 (3.72, 18.26)^a Satisfied or very satisfied: 4.00 (0.83, 19.19) MD −0.2 (−3.4, 2.9) *SMD* −0.96 (−1.61, −0.32)^a	Moderate	*SMD* 0.39 (0.17, 0.60)^a	High
Family's satisfaction with EoL care	OR Very satisfied: 5.17 (1.52, 17.58)^a Satisfied or very satisfied: 2.48 (0.55, 11.10) MD 2.10 (1.75, 2.45)^a 4.6	High	NA
Concordance between patient's wishes and care received	OR 1.73 (0.76, 3.90)	High	NA
Concordance between patient's and family's wishes	OR 4.32, P < 0.001^a MD 1.4, P = 0.002^a	High	NA
Completion of ACP documents and processes	*Differences in Proportions* 25% (17, 33)^a 77% (70, 83)^a 13% (5, 19)^a	High	*Differences in Proportions* 22% (15, 30)^a	High
Receiving chemotherapy at EoL	OR 0.50 (0.35, 0.72)^a	Low	NA
Resuscitation	OR 0.13 (0.03, 0.55)^a	Very low	NA
Receiving hospital care	MD −0.21, P = 0.04^a	High	MD 0.40 (−0.24, 1.04)^b	Low
Hospital LOS	MD −1.8, P = 0.03^a	High	MD 0 −3.00 (−6.07, 0.07)^a,b	Low
ED visits	MD −1.27 (−3.34, 0.80)	Low	MD −0.10 (−1.16, 0.96)	Moderate
ICU care	OR 0.44 (0.13, 1.53)	Very low	NA
ICU LOS	NA		MD −2 −1.50 (−3.48, 0.48)	High
Home health visits	MD 0.49 (−2.57, 3.55)	Very low	NA
Urgent care visits	NA		MD −0.30 (−0.61, 0.01)^a,b	Moderate
Other outpatient visits	MD 2.39 (−4.18, 8.96)	Low	MD −5.20 (−9.70, −0.70)^a	Moderate
Receiving hospice care	OR 5.17 (2.03, 13.17)^a	High	NA

Outcome	Results	GRADE
Patient's QoL	Earlier discussions were associated with greater QoL adjusted β = 0.003, P = 0.006^a	Low
Receiving chemotherapy at EoL	Earlier discussions were associated with lower receipt of chemotherapy in the last 14 days of life χ²: 17.057, P < 0.001^a	Low
Receiving hospital care	Earlier discussions were associated with less hospital care in the last 30 days of life χ²: 55.905, P < 0.001^a	Moderate
ICU care	Earlier discussions were associated with less ICU care in the last 30 days of life χ²: 16.606, P < 0.001^a	Low
Receiving hospice care	Earlier discussions were associated with receiving hospice care χ²: 50.756, P < 0.001^a	Moderate

Author, Year	Allocation Concealment	Blinding	Complete Accounting of Patients and Outcome Events	Selective Reporting Bias	Other Limitations
Kirchhoff et al, 2012 (40); Kirchhoff et al 2010 (41)	No limitations	Limitations^a	Limitations^b	No limitations	No limitations
Au et al, 2012 (42)	No limitations	Limitations^c	No limitations	No limitations	Limitations^d
Sampson et al, 2011 (51)	No limitations	Unclear	Limitations^b	No limitations	Limitations^e
Detering et al, 2010 (39)	No limitations	Limitations^f	Limitations^b	No limitations	No limitations
Clayton et al, 2007 (44)	No limitations	No limitations	Limitations^b	No limitations	No limitations
Casarett et al, 2005 (48)	No limitations	Limitations^g	Limitations^b	No limitations	No limitations
Song et al, 2005 (49)	No limitations	Limitations^h	Limitations^b	No limitations	No limitations
Nicolasora et al, 2006 (47)	No limitations	Unclear	Limitations^b	No limitations	No limitations
Dyar et al, 2012 (50)	Unclear	Unclear	Limitations^b	No limitations	Limitationsⁱ
Jones et al, 2011 (52)	No limitations	No limitations	Limitations^b	No limitations	No limitations

Author, Year	Allocation Concealment	Blinding	Complete Accounting of Patients and Outcome Events	Selective Reporting Bias	Other Limitations
Gade et al, 2008 (43)	No limitations	Unclear	Limitations^a	No limitations	No limitations
Lautrette et al, 2007 (45)	No limitations	No limitations	Limitations^a	No limitations	No limitations
Gilmer et al, 2005 (46)	Unclear	Unclear	Limitations^a	No limitations	No limitations

Author, Year	Appropriate Eligibility Criteria	Appropriate Measurement of Exposure	Appropriate Measurement of Outcome	Adequate Control for Confounding	Complete Follow-Up
Observational with Contemporaneous Controls
Evangelista et al, 2012 (54)	No limitations	No limitations	No limitations	No limitations	No limitations
Jacobsen et al, 2011 (55)	No limitations	No limitations	No limitations	Limitations^a	No limitations
Engelhardt et al, 2009 (56)	No limitations	No limitations	No limitations	Limitations^b	No limitations
Mack et al, 2012 (58)	No limitations	Limitations^c	No limitations	Unclear	No limitations
Mack et al, 2010 (59)	No limitations	Limitations^c	No limitations	No limitations	No limitations
Wright et al, 2008 (60)	No limitations	Limitations^c	No limitations	No limitations	No limitations
Morrison et al, 2005 (62)	No limitations	No limitations	No limitations	Limitations^d	No limitations
Zhang et al, 2009 (53)	No limitations	Limitations^e	No limitations	Limitations^d	No limitations
Briggs et al, 2004 (63)	No limitations	No limitations	No limitations	Limitations^f	No limitations
Observational with Historical Controls
Lindner et al, 2007 (66)	No limitations	No limitations	No limitations	No limitations	No limitations
Cross-Sectional
Leung et al, 2012 (67)	No limitations	Limitations^e	Limitations^g	No limitations	No limitations
Mori et al, 2013 (68)	No limitations	Limitations^e	No limitations	No limitations	No limitations
Heyland et al, 2009 (11)	No limitations	Limitations^e	Limitations^g	Limitations^f	No limitations

Author, Year	Appropriate Eligibility Criteria	Appropriate Measurement of Exposure	Appropriate Measurement of Outcome	Adequate Control for Confounding	Complete Follow-Up
Observational with Contemporaneous Controls
Rabow et al, 2004 (57)	No limitations	No limitations	No limitations	No limitations	No limitations
Casarett et al, 2008 (61)	No limitations	No limitations	No limitations	Limitations^a	No limitations
Observational with Historical Controls
Lamba et al, 2012 (64)	No limitations	No limitations	No limitations	Limitations^b	No limitations
Norton et al, 2007 (65)	No limitations	No limitations	No limitations	No limitations	No limitations

PERMALINK

Patient Care Planning Discussions for Patients at the End of Life: An Evidence-Based Analysis

S Baidoobonso

Abstract

Background

Objectives

Data Sources

Review Methods

Results

Limitations

Conclusions

Plain Language Summary

Background

Objective of Analysis

Clinical Need and Target Population

End-of-Life Population

Quality of End-of-Life Care and Communication

Figure 1: CARENET's Model for Improving Communication about End-of-Life Care.

Ontario Context

Technology/Technique

Evidence-Based Analysis

Research Question

Research Methods

Literature Search Strategy

Inclusion Criteria

Exclusion Criteria

Outcomes of Interest

Statistical Analysis

Data Extraction

Assessment of Risk of Bias in Included Primary Studies

Assessment of Publication Bias

Data Synthesis

Quality of Evidence

Results of Evidence-Based Analysis

Figure 2: Citation Flow Chart.

Table 1:

Description of Included Studies

Table 2:

Description of Interventions

Table 3:

Quality of Life

Table 4:

Table 5:

Satisfaction with End-of-Life Care

Table 6:

Table 7:

Concordance

Table 8:

Table 9:

Completion of Advance Care Planning Documents and Processes

Table 10:

Health Care Use

Chemotherapy

Table 11:

Table 12:

Resuscitation

Table 13:

Hospital Care

Table 14:

Table 15:

Table 16:

Emergency Department Visits

Table 17:

Care in Intensive Care Unit

Table 18:

Table 19:

Table 20:

Table 21:

Home Health Visits

Urgent Care Visits

Table 22:

Other Outpatient Visits

Table 23:

Hospice Care

Table 24:

Table 25:

Summary of Findings

Table 26:

Table 27:

Conclusions