Introduction In a landmark 2013 decision, the Supreme Court of Canada (SCC) ruled that the withdrawal of life support in certain circumstances is a treatment requiring patient or substitute decision maker (SDM) consent. How intensive care unit (ICU) physicians perceive this ruling is unknown.
Objectives To determine physician knowledge of and attitudes towards the SCC decision, as well as the self-reported changes in practice attributed to the decision.
Methods We surveyed intensivists at university hospitals across Canada. We used a knowledge test and Likert-scale questions to measure respondent knowledge of and attitudes towards the ruling. We used vignettes to assess decision making in cases of intractable physician-SDM conflict over the management of patients with very poor prognoses. We compared management choices pre-SCC decision versus post-SCC decision versus the subjective, respondent-defined most appropriate choice. Responses were compared across predefined subgroups. We performed qualitative analysis on free-text responses.
Results We received 82 responses (response rate=42%). Respondents reported providing high levels of self-defined inappropriate treatment. Although most respondents reported no change in practice, there was a significant overall shift towards higher intensity and less subjectively appropriate management after the SCC decision. Attitudes to the SCC decision and approaches to disputes over end-of-life (EoL) care in the ICU were highly variable. There were no significant differences among predefined subgroups.
Conclusions Many Canadian ICU physicians report providing a higher intensity of treatment, and less subjectively appropriate treatment, in situations of dispute over EoL care after the Supreme Court of Canada's ruling in Cuthbertson versus Rasouli.
- End of Life Care
- Bills, Laws and Cases
- Clinical Ethics
- Demographic Surveys/Attitudes
Statistics from Altmetric.com
The Supreme Court of Canada decision in Cuthbertson vs Rasouli (hereafter referred to as ‘the SCC decision’ or ‘ the SCC ruling’) is a landmark case in recent Canadian medical jurisprudence.1 Since 2010, the case has drawn attention in medical journals and been the subject of over 150 articles in the Canadian news media.2 ,3 The patient, Mr Rasouli, was 57 years when he developed meningitis after elective neurosurgery, leaving him in a minimally conscious state that required a prolonged stay in the intensive care unit (ICU). His attending ICU physicians felt that there was no medical benefit to ongoing mechanical ventilation given the severity and permanence of the patient's brain injury, and told the substitute decision maker (SDM) that the team would move to withdrawal of life support (WDLS). The patient's SDM applied to the Ontario Superior Court for an injunction to stop WDLS, but the court ruled that the injunction was in fact unnecessary, as the physicians required consent for WDLS. The Ontario Court of Appeal and the SCC upheld the lower court's decision. The SCC ruled that the WDLS constitutes a plan of treatment (as defined by Ontario's Health Care Consent Act), and thus required patient or SDM consent.1 ,4
The SCC stated that their decision should apply narrowly to cases like Mr Rasouli's.1 Since the ruling is a statutory interpretation of an Ontario law, it does not obviously apply in other Canadian provinces. However, there was anecdotal concern that the case would change the practice of physicians across Canada, even in cases that were not analogous to Mr Rasouli's and outside the jurisdiction of Ontario's Health Care Consent Act and Consent and Capacity Board (See Definitions). Physician understanding of medical case law affects practice patterns across all specialties—even if the understanding of the law is incomplete or inaccurate.5–7 How physicians interpret medical jurisprudence in large part determines its impact. The literature on the SCC decision has so far consisted of medicolegal and medical ethics commentary, but the reaction to the decision in the critical care community is unknown.8–13 The SCC decision is most relevant to intensivists, who are the physicians most involved in decisions about the withholding and WDLS for incapable patients. We conducted a survey of Canadian academic intensivists to evaluate their attitudes towards the SCC decision and how they perceived the ruling to influence their practice.
The primary objective was to determine self-reported changes in physician practice following the SCC decision. Secondary objectives included the determination of self-reported attitudes towards and knowledge of the SCC ruling, as well as demographic and other factors associated with personal views of the SCC ruling.
We conducted a self-administered, email survey of attending ICU physicians at seven academic centres across Canada. Our survey was divided into four sections: demographics; attitudes towards the SCC ruling and similar cases; knowledge of the SCC ruling; and self-reported practice change attributed to the SCC decision (see online supplementary appendix 1).
In Part I, we gathered demographic data, including location of practice, previous legal case involvement, and base specialty. For Part II we generated 5-point Likert-scale questions to assess attitudes towards the decision. We also asked about the relevance of the decision to respondent practice, and respondent attitudes towards the withholding and WDLS in general. In Part III, using expert input, we created nine true-false questions about the SCC ruling to measure physician knowledge of the facts and legal outcomes of the case.
In Part IV, we used direct questions and realistic vignettes (table 1 and see online supplementary appendices 1 and 2) to assess decision making in cases of physician-SDM conflict. In each scenario we asked what the respondent would have done before and after the SCC decision, and what management the respondent considered most appropriate independent of legal considerations. The definition of ‘appropriate’ was subjective and respondent-defined. The five vignettes presented intractable disagreements about withholding of life support (WHLS) or WDLS for patients with very poor prognoses, and forced the respondent to choose between initiating or continuing life support, WHLS/WDLS or a time-limited trial of life support. While consensus between physicians and SDMs is the most common outcome in clinical practice, these vignettes presented irreconcilable disputes. One vignette was directly analogous to the case in the SCC decision; the others presented different scenarios of physician-SDM dispute in the management of critically ill patients (see online supplementary appendix 2). In each vignette, the SDM advocated for escalation of treatment or against the withdrawal/withholding of treatment. We included a single open-ended question asking respondents how the decision might have changed practice.
A representative group of collaborators revised the questionnaire before administration. We obtained ethics approval through the primary investigator's institutional Research Ethics Board (REB) and, where required, REBs of collaborating institutions. The questionnaire was pilot-tested on a group of 12 intensivists to evaluate its pertinence, usability, redundancy, scope and relevance using a validated tool.14 Ninety-two per cent of respondents scored the questionnaire the equivalent of ‘fair’ or higher in all sensibility questions.
For further validation, each vignette was paired with a Likert-scale question. These questions directly asked respondents to quantify the impact of the SCC decision on their practice in situations of physician-SDM dispute identical to those in the paired vignettes. The Likert-scale questions and vignettes were then compared for intrarater reliability using the γ test-retest statistic. Agreement between vignette responses and Likert-scale responses ranged from 59% to 72%.
We administered the questionnaire to intensivists in the seven academic centres using SurveyMonkey. Recipients received two electronic reminders at 1-month intervals following the initial electronic invitation. Consent to participate was implied by completing the questionnaire.
Seven 5-point Likert scale questions were used to determine favourability to the SCC ruling. The questions, which were weighted equally, addressed topics like the SCC decision's effect on healthcare expenditures, ICU bed availability and rates of futile care (see online supplementary appendix 1). One question asked respondents to what degree they agreed with the decision (reported separately in figure 1). The options for each question were assigned numerical values ranging from −2 (least favourable) to +2 (most favourable). The favourability score was the sum of scores for all seven questions; a score of zero signified neutral views; negative scores, unfavourable views; and positive scores, favourable views. A single 5-point Likert-scale question asked respondents how relevant the SCC decision was to their practice. Responses were grouped into ‘relevant’ and ‘not relevant/neither’ categories.
Nine true/false questions were used to measure knowledge of the ruling. We calculated the mean score and sorted respondents into low score (0–3/9; poor understanding), medium score (4–6/9) and high score (7–9/9; good understanding) groups. The survey text provided no facts about the SCC case.
We used clinical vignettes to measure the self-reported impact of the SCC decision on clinical practice. For the case vignettes, management options were ranked from 0 to 2 in order of degree of treatment intensity (DTI). For example, declining to perform CPR on a patient with metastatic cancer was scored 0/2, while performing CPR at the direction of the SDM was scored 2/2. For each respondent, DTI scores (0–2) were calculated for individual vignettes and for all five vignettes (pooled score range 0–10). Scores were calculated for before and after the SCC decision (table 1).
Each vignette response was scored based on whether the selected management plan (before and after the SCC decision) corresponded to the respondent's subjective definition of appropriate management; a score of 0 was given if the selected plan did not match the ‘appropriate’ plan, while a score of 1 was given if the selected and self-defined appropriate plan were the same. Pooled scores (0–5) for all vignettes were recorded (table 1). In this way, we calculated DTI scores and subjective appropriateness scores for each respondent for before and after the SCC decision. We used modified χ2 testing to compare responses before and after the SCC decision for each vignette. We used the Wilcoxon signed-rank test to compare the difference between the means of the pooled scores before and after the SCC decision. To exclude bias related to the SCC case, we repeated pooled calculations of DTI scores and subjective appropriateness scores for vignettes 2–5 (ie, the vignettes not directly analogous to the case of Mr Rasouli). The section on behavioural change also included a paired 5-point Likert scale question asking respondents to rate the likelihood of bringing a dispute to the Consent and Capacity Board (CCB)/courts before and after the SCC decision. Responses were collapsed into ‘likely’, ‘unlikely’ and ‘neither likely nor unlikely’ categories. ‘Pre-’ and ‘Post-’ responses were compared using modified χ2 testing.
Changes in DTI scores and subjective appropriateness scores were measured in six predefined subgroups: province of practice, previous legal or CCB end-of-life (EoL) care case involvement, knowledge score, self-reported relevance of decision to respondent practice, favourability towards decision, and years of practice. We used the Mann-Whitney U test to measure differences in behavioural change (DTI and subjective appropriateness) in these subgroups. We also analysed regional differences in likelihood of bringing cases before the courts/CCB using modified χ2 testing.
Our survey included a single free-text question asking respondents how the SCC decision had changed their practice. We performed limited analysis by identifying common themes and sentiments.
The survey was sent to 299 intensivists at eight academic centres in six provinces (Alberta, British Columbia, Manitoba, Nova Scotia, Ontario, and Quebec, Canada) (table 2). We excluded responses from one site because the survey was sent to participants that did not meet inclusion criteria. After exclusion, there were 197 recipients at seven sites; we received 82 responses for a response rate of 42%. Data were analysed with and without the excluded responses and there were no significant differences in primary or secondary outcomes between the two groups.
Response rates per question ranged from 85–100% (excluding the single free-text question). Seventy-seven respondents (94%) completed all five case vignettes; we excluded the remaining 6% from the pooled analysis of vignette responses.
Most respondents had unfavourable views of the SCC ruling. Significantly more Ontario respondents (88%) versus non-Ontario respondents (57%) believed the decision was relevant to their practice (p<0.05) (figure 1). Just over half of the respondents believed that, in their opinion, WDLS was a medical decision that should not require consent, and the remainder believed that WDLS should require SDM/patient consent. Respondents with favourable views were more likely to believe that WDLS should require SDM/patient consent than those with unfavourable views (77% vs 23%, p<0.05).
For the knowledge test (n=73 respondents; score range 0–100%), the mean score was 56% (figure 1).
At baseline (ie, pre-SCC decision) we measured DTI scores for vignette 1—the case directly analogous to Mr Rasouli's. Among all respondents, 26% opted for the lowest DTI (WDLS), 27% chose to apply to the CCB or courts to obtain consent for WDLS, and 46% chose to follow the SDM's direction by providing ongoing life support. Post-SCC decision, 9/77 (12%) reported a shift towards higher DTI score, 6/77 (8%) shifted towards lower DTI score and 62/77 (80%) reported no change (McNewmar-Bowker test p=0.03).
We also calculated subjective appropriateness scores (table 1) for vignette 1. Pre-SCC decision, 51/77 (66%) selected a management plan matching their definition of appropriate treatment, while 26/77 (34%) selected a subjectively inappropriate plan. There was no significant difference in the post-SCC decision responses. We likewise calculated change in DTI and change in subjective appropriateness score for vignettes 2–5 separately. There were no significant differences in either measurement after the SCC decision.
We pooled DTI scores and subjective appropriateness scores for vignettes 1–5 and compared pre-SCC and post-SCC decision pooled scores. For DTI scores, there was a significant increase in mean score after the SCC decision. Twenty-seven per cent of respondents reported a shift towards higher DTI, 65% reported no change and 8% reported a shift towards lower DTI (table 3).
Pooled subjective appropriateness scores significantly decreased after the SCC decision. Twenty-three per cent selected ‘less appropriate’ treatment options, 68% reported no change and 9% selected more subjectively appropriate treatment options (table 4).
We repeated these pooled calculations for the vignettes that were not analogous to the case of Mr Rasouli (#2–5). There was a significant shift towards less subjectively appropriate treatment (pre-SCC decision mean score 3.12 vs post-SCC decision score 2.91, p=0.03) and higher DTI (pre-SCC decision mean score 2.42 vs post-SCC decision core 2.75, p=0.01).
We compared pre-SCC versus post-SCC decision DTI and subjective treatment appropriateness scores for individual respondents (see online supplementary appendix 3: tables 2A and 3A). Twenty-nine (37.7%) of respondents reported providing ‘appropriate’ care in all five scenarios, pre-Rasouli decision and post-Rasouli decision, while the remainder of respondents opted for subjectively inappropriate care in at least one scenario.
More respondents reported being likely to bring cases to the Consent and Capacity Board or the courts after the SCC decision (44% vs 25%, McNemar-Bowker test p<0.01). This effect was limited to Ontario respondents, with 13/40 (33%) reporting they were more likely to bring cases before the CCB/courts in the wake of the SCC decision; for non-Ontario respondents, there was no significant change in likelihood of bringing cases before the CCB or the courts.
For change in DTI and change in subjective appropriateness scores, there was no significant difference in any of the major subgroups (province of practice, previous legal or CCB EoL care case involvement, knowledge score, self-reported relevance of decision to respondent practice, favourability towards decision and years of practice).
There were 40 responses to the free-text question, which asked how the SCC decision had changed practice (see online supplementary appendix 2). We identified 14 themes, 6 occurring five or more times. The most frequent (n=12) theme was that the decision has not affected practice. Nine respondents wrote that similar cases should be resolved outside of court. Six respondents stated they were less likely to allow time-limited trials of ICU admission or life support, because the SCC ruling made withdrawal of a therapy more problematic than its initiation: “I am less inclined to admit patients to ICU. I think [it] is easier to not initiate therapy as opposed to withdrawing therapy.” Six respondents argued the decision weakened the position of physicians and privileged SDMs in EoL treatment disputes: “Unfortunately this case has taken any decision making power around withdrawal (and in my opinion, treatment offering) out of the hands of clinicians and put it in the hands of the SDM, regardless of how futile.” Five respondents noted that the decision does not apply outside Ontario. Five respondents stated that the appellants should have gone to the CCB rather than appealing the initial ruling.
Two respondents stated that SDMs rather than physicians should direct WDLS/WHLS decisions. One wrote that “…at the end of the day, sometimes we just need to…provide life support even when it would not be what I would choose as a person, MD or member of society…In Canada in the 21st century, this includes seeking consent to withdraw or not resuscitate 100% of the time.” Other themes (n=2) included: the SCC decision will lead to greater moral distress; the ruling defends patient interests; the ruling will alter the behaviour of non-ICU physicians; the ruling has created uncertainty about EoL care in the ICU; the ruling has created more certainty about EoL care in the ICU; the CCB is cumbersome and impractical.
When presented with realistic clinical vignettes, we found that academic ICU physicians in Canada reported a significant overall shift towards more intensive treatment in the wake of the SCC decision, even though most individual respondents did not indicate a change in their practice. There was also a significant shift away from treatment the respondents defined as appropriate. These trends held even in cases not analogous to Mr Rasouli's. The results show marked variation in approaches to WHLS and WDLS and the management of patients like Mr Rasouli, as well as in views of the SCC ruling. Even though roughly half of respondents believed that WDLS should require patient/family consent, only 18% agreed with the Supreme Court's decision. Fewer than half reported the ruling made them more likely to bring a similar case before the CCB or the courts. We found important gaps in understanding about the legal aspects of the decision; less than a third of respondents had high levels of knowledge about the ruling.
The magnitude of the problem of physician-SDM conflict over EoL care is not clear. In two studies, Huynh et al15 ,16 showed that ‘futile care’ was common and associated with significant financial cost, but there is little evidence that physician-SDM disagreement drives these types of cases. Chidwick et al17 reviewed physician-SDM EoL disputes brought to Ontario's Consent and Capacity Board. Since 1996, only 26 cases out of 1367 reviewed involved physician-SDM dispute about a patient's EoL treatment, although the incidence of these cases is growing. The SCC ruling is one of many recent, well publicised cases of disagreement between physicians and SDMs about patient management. Whereas in the past it was the physician's role to decide what degree of life support was appropriate for a given patient, there is a trend of SDMs demanding a treatment not considered appropriate by the patient's physicians.18 This conflict between a patient/SDM's perceived right to demand a treatment and a physician's right not to offer treatment outside professional standards, has played out recently in cases in Manitoba and California, among others.19 ,20
Most cases of physician-SDM dispute are resolved outside the courts or quasi-judicial bodies like the CCB, either by consensus, or by physician or SDM acquiescence.8 When physicians accede to SDM directions, they risk providing treatment they consider inappropriate. In our survey, only 38% of respondents selected self-defined appropriate management plans in all scenarios presented. Whether or not the SCC decision changed practice for most respondents, there is a gap between the treatment intensivists believe is appropriate and the treatment they provide.
However, our survey suggests that this ‘appropriateness gap’ is more problematic for some physicians than for others. The qualitative responses, in particular, articulated this conflict. Many respondents lamented the cultural trend towards patient/SDM autonomy in EoL decisions, stating that the SCC decision would lead to futile and inappropriate care. But others argued that patient/SDM autonomy is the reality of practicing medicine in Canada in the 21st century—and that the physician's perception of appropriateness is a lesser consideration. This view aligns with the majority opinion in the SCC decision: “[W]hile a physician may feel that the legal obligation not to withdraw life support is in tension with their professional or personal ethics, such tensions are inherent to medical practice.”1
The SCC ruling was intended to be narrow in scope, rather than ‘resolv[ing] the philosophical debate’ over whether an SDM might overrule a ‘physician's interest in not being forced to provide non-beneficial treatment’.1 But physicians have speculated how it might apply to similar but not directly analogous cases. Downar et al9 reviewed the SCC decision's potential application to jurisdictions outside of Ontario, as well as to clinical scenarios like stopping CPR, withholding CPR and life support, the cessation of ongoing therapies (like haemodialysis) and time-limited trials of life support. Our survey has shown that the SCC decision might affect the treatment provided by Canadian intensivists, even in cases unlike Mr Rasouli's and outside Ontario.
There are several limitations to this study. By framing the questions around the time point of the SCC decision, the survey may have biased respondents towards assigning significance to the ruling. The relationship between self-reported practice patterns and actual behaviour is not clear—observational studies are needed to measure the true impact of the SCC decision on clinical practice. Our response rate was moderate, which we attribute to the survey length and the controversial nature of the research question. Our metrics of behavioural change—DTI scores and subjective appropriateness scores—are logical but not validated tools.
This study suggests there is significant variation in how intensivists approach EoL disputes, within and between regions of Canada—more research is needed to explore this important issue. In open-ended responses, several recipients described how the SCC ruling had affected the practice of non-intensivists. This is meaningful, since these physicians are often responsible for writing resuscitation orders for patients on admission, and patients may deteriorate before intensivists can revisit the orders. It is worth examining whether the SCC decision has led surgeons, internists and emergency physicians to show greater deference to SDM/patient wishes in disputes over EoL treatment.
In case vignettes, Canadian ICU physicians reported providing a higher intensity of treatment, and less subjectively appropriate treatment, in the wake of the SCC ruling in Cuthbertson versus Rasouli. Physician views of the ruling are generally negative but highly variable, as are intensivist approaches to cases of physician-SDM dispute.
The authors thank Anna Danielson for her help with statistical methodology.
Contributors All authors contributed to the conception and design, data collection, critical revision, and final approval. DC drafted the initial manuscript. DC and JD share primary responsibility for analysis, interpretation, and statistical analysis of the data. JD obtained funding and assumes overall responsibility for the study.
Funding Toronto General/Toronto Western Hospital Foundation.
Competing interests None declared.
Ethics approval University Health Network Research Ethics Board (UHN REB 14–7552).
Provenance and peer review Not commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.