A National Survey of Canadian Psychiatrists’ Attitudes toward Medical Assistance in Death

Abstract

Background:

Bill C-14 allows for medical assistance in dying (MAID) for patients who have intolerable physical or psychological suffering that occurs in the context of a reasonably foreseeable death. In Canada, psychiatrist support for MAID on the basis of mental illness and beliefs influencing level of support are unknown. The objectives of this research were to 1) determine if psychiatrists are supportive of MAID under certain conditions and on the basis of mental illness and 2) determine what factors are related to psychiatrist support for MAID on the basis of mental illness.

Methods:

This cross-sectional study was conducted among 528 psychiatrists in Canada using an online survey platform (February 19 to March 11, 2016).

Results:

The response rate was 20.9% (n = 528). Most psychiatrists supported the legalisation of MAID in some circumstances (72%); however, only 29.4% supported MAID on the basis of mental illness. Factors correlating with decreased support for MAID for mental illness were the belief that MAID for mental illness would change the psychiatrists’ commitment to their patients through enduring suffering, having a personal faith, and having had past patients who would have received MAID for mental illness were it legal but instead went on to recover.

Interpretation:

This study found that most psychiatrists do not support the legalisation of MAID for mental illness, despite being quite supportive of MAID in general. Objections seemed to be based upon concern for vulnerable patients, personal moral objections, and concern for the effect it would have on the therapeutic alliance.

With the recent passing of Bill C-14 in response to the Carter decision of the Supreme Court of Canada (SCC), medical assistance in dying (MAID), which was previously referred to as physician-assisted death, has become legal in Canada.¹ Understandably, there has been much public debate about this issue, particularly with regard to what criteria ought to be met in order for a person to be eligible to receive MAID.^2–4 Bill C-14 restricts access to MAID for people who meet the following criteria: a) they have a serious and incurable illness, disease, or disability; b) they are in an advanced state of irreversible decline in capability; c) that illness, disease, or disability or that state of decline causes them enduring physical or psychological suffering that is intolerable to them and cannot be relieved under conditions that they consider acceptable; and d) their natural death has become reasonably foreseeable, taking into account all of their medical circumstances, without a prognosis necessarily having been made as to the specific length of time that they have remaining.¹ Some have argued that these criteria, which essentially exclude patients who seek MAID for reasons of psychiatric illness, are too strict and not in keeping with the SCC decision. Notably, an Alberta court of appeal decision, which predates Bill C-14, ruled that criteria laid out in the Carter decision do not exclude psychiatric illness. In this particular case, the decision granted E.F., a 58-year-old woman with conversion disorder, access to MAID on this basis.⁵ Others have argued that Bill C-14 is nonetheless consistent with the Canadian Charter of Rights and Freedoms and that a more restrictive and cautious approach to MAID is consistent with the Carter decision.⁶ The Minister of Justice and the Minister of Health are committed to reviewing issues relating to requests that currently do not fall within C-14, including where mental illness is the sole underlying medical condition.

Most studies show that physicians tend to be less supportive of MAID (or physician-assisted death as it is often referred to in other countries) than the general population, although support may be increasing in Western countries.^7–9 The few studies considering physician-assisted death on the basis of mental illness have found significantly less support among physicians than in general.^8,10 When psychiatrists’ attitudes towards physician-assisted death for mental illness have been examined, there has been considerable concern about issues of decision-making capacity (often referred to synonymously as “competency”) and the influence of mental illness on a patient’s request.¹⁰ The introduction of Bill C-14 has engendered much critical thought about this challenging issue in Canada. The question of how decision-making capacity is properly assessed in this context is one of active inquiry and debate. A recent article by Charland et al.¹¹ has suggested that the present state of the art relies heavily on the cognitive element of capacity while ignoring the emotional element. They question whether a clinical model of human decision making that does not adequately account for emotions and feelings can “bear the weight” of extending MAID to patients for reasons of psychiatric illness.

Access to assisted death for reasons of mental illness is a legal practice in Belgium and the Netherlands. Those who support this practice argue that it allows patients with unbearable suffering a welcome and humane relief and that stricter procedural conditions adequately minimise the risk of its inappropriate application.¹² Although the quite distinct development of this practice in the unique medical cultures of these countries may complicate comparisons, recently several retrospective descriptive reviews have examined outcome measures for patients receiving assisted death for reasons of mental illness.^12,13 The unexpected variety of diagnoses, the procedure going through despite disagreements among assessors that remain unresolved, lack of psychosocial supports among many patients, lack of rigour in capacity assessment, and apparent lack of illness severity requiring hospitalisation in many patients are some of the concerns identified as red flags.^14,15

Previous studies on psychiatrists’ attitudes towards MAID or physician-assisted death for mental illness have been exclusively in the context of concurrent terminal or debilitating physical illness, not as the sole reason for seeking MAID. Both the Canadian Psychiatric Association (CPA) and the American Psychiatric Association have produced public statements regarding MAID for patients with mental illness, the former raising serious concerns and the latter holding that psychiatrists should not take part in the practice.^16,17 However, aside from a few published surveys that do not touch on the specific issue of MAID for patients with mental illness,^7,9 there is little research on what motivations physicians have for their beliefs and virtually no research on Canadian psychiatrists’ attitudes and motivations. Although the literature is sparse, previous studies have identified factors related to support for MAID on the basis of mental illness, including experience working with patients at the end of life¹⁸ and having a personal faith.¹⁹ Further exploration of psychiatrists’ beliefs and attitudes around MAID for mental illness is warranted. Therefore, the objectives of this research were to 1) determine if psychiatrists are supportive of MAID under certain conditions and MAID on the basis of mental illness and 2) determine what factors are related to psychiatrist support for MAID on the basis of mental illness.

Methods

Data and Sample

In collaboration with the CPA, we developed a survey aimed to assess Canadian psychiatrists’ support for and concerns about the potential legalisation of MAID under certain conditions (as had been recommended in the Carter decision) and the potential legalisation of MAID on the basis of mental illness. A link to the survey was sent out via e-mail to all registered members of the CPA (2520), along with 3 reminder e-mails from February 19 to March 11, 2016. Responses were collected anonymously and consent was obtained (n = 528, response rate 20.9%). The survey and research protocol were approved by the University of Manitoba Health Research Ethics Board.

Measures

Support for MAID

Four questions assessed psychiatrists’ support for MAID: 1) Are you supportive of legislation to legalise MAID under certain conditions? 2) Should people with suffering caused by mental illness be offered MAID? 3) Would you be willing to advocate for MAID for a patient with a treatment-resistant Axis 1 disorder? 4) Would you be willing to advocate for MAID for a patient with a treatment-resistant Axis 2 disorder? Responses were collected on an ordinal scale that included yes, probably, probably not, no, and don’t know. To facilitate data analysis and improve the clarity in our data tables, responses were collapsed into the following categories: yes/probably, no/probably not, and don’t know.

Demographic variables

Five demographic variables were included in the analysis. Sex was measured as male/female. Length of time in practice included the following categories: resident, less than 5 years, 5 to 10 years, 10 to 20 years, more than 20 years, and retired. Area of focus was measured as general adult, child, geriatric, forensic, and consult liaison. Respondents were classified into 5 different types of practice: university teaching hospital, community hospital, private office practice, community-based service delivery, and other. Data were collected from the 10 provinces.

Beliefs and experiences of psychiatrists

Based on previous research^8,10,18,19 and clinical experience of the authors, there were several items assessed on the survey that we hypothesised would be associated with support for MAID for mental illness. These include 1) belief that legalisation of MAID on the basis of mental illness would change psychiatrists’ commitment to their patients with enduring mental suffering, 2) belief that legalisation of MAID on the basis of mental illness would change psychiatrist response to suicidal ideation, 3) have had patients in the past who recovered from mental illness but think may have received MAID if it were available at the time, 4) having a personal faith, 5) experience working with people with a life-threatening condition, 6) belief that suicide can be a rational choice, and 7) length of time in practice. Psychiatrist beliefs and experiences for the first 6 items were collected on an ordinal scale that included: yes, probably, probably not, no, and don’t know. Responses were dichotomised into yes/probably and no/probably not to facilitate data analysis and improve clarity in data tables. Those who responded don’t know were excluded from this section of the analysis. Length of time in practice was measured in 5-year increments, with a final category of retired.

Statistical analysis

Descriptive statistics describing the prevalence of demographic variables in the study population were determined. Chi-squared values were computed to measure the difference between the study sample and all registered members of the CPA for province and sex. Descriptive statistics were computed to determine the prevalence of psychiatrists who support/do not support MAID under certain conditions, MAID for mental illness, MAID for Axis 1 disorders, and MAID for Axis 2 disorders. Logistic regression was used to measure the relationships between psychiatrists’ beliefs and experiences and support for MAID based on mental illness.

Results

Table 1 presents the characteristics of the survey sample. Slightly less than half of the respondents had been practicing for more than 20 years (43.4%) and the majority were general adult psychiatrists (73.2%). The most common type of practice was university teaching hospital (46.3%).

Table 1.

Characteristics of the Survey Sample.

Characteristic	% (n)
Length of time in practice
Resident	11.0 (58)
<5 years	11.0 (58)
5-10 years	8.7 (46)
10-20 years	20.6 (109)
>20 years	43.4 (229)
Retired	5.3 (28)
Area of focus
General adult	73.2 (380)
Child	9.1 (47)
Geriatric	7.7 (40)
Forensic	4.4 (23)
Consult liaison	5.6 (29)
Type of practice
University teaching hospital	46.3 (243)
Community hospital	14.7 (77)
Private office practice	17.9 (94)
Community-based service delivery	12.0 (63)
Other/mixed	9.1 (48)

Table 2 presents the distribution of both survey respondents and all CPA members by province and sex. Almost half of the respondents resided in Ontario (43.6%), and there was a relatively equal proportion of male and female respondents in the sample. Chi-squared statistics for both province and sex indicated that there were no significant sex or provincial differences between the survey sample and all registered members of the CPA. This demonstrates that our sample was representative of all CPA members.

Table 2.

Demographics of Study Sample Compared to All Canadian Psychiatric Association (CPA) Members.

Characteristic	Study Sample (n = 528), % (n)	All CPA Members (N = 2520), % (n)
Province
British Columbia	14.0 (74)	14.3 (361)
Alberta	11.2 (59)	11.0 (278)
Saskatchewan	2.8 (15)	3.0 (76)
Manitoba	5.1 (27)	5.0 (127)
Ontario	43.6 (230)	43.6 (1098)
Quebec	14.4 (76)	14.4 (364)
Nova Scotia	4.4 (23)	4.3 (108)
New Brunswick	1.7 (9)	1.7 (44)
Prince Edward Island	0.4 (2)	0.4 (9)
Newfoundland	2.3 (12)	2.1 (53)
Chi-squared P value	0.98
Sexa
Female	47.8 (257)	44.7 (1107)
Male	51.3 (271)	55.3 (1368)
Chi-squared P value	0.07

^aForty-five individuals had missing data for sex in the CPA data; therefore, the total number for sex does not add to the total population number.

Table 3 shows that most respondents support legalisation of MAID under certain conditions as suggested in the Carter decision (71.8%) but only 29.4% support MAID on the basis of mental illness. Support for MAID was 23.5% for Axis 1 disorders and 9.3% for Axis 2 disorders. In addition, 46% of the sample endorsed the belief that no psychiatric illness is irremediable to the extent that physician-assisted death would be considered, and 57% of the sample reported having patients who recovered from a mental illness who may have received physician-assisted death if it were available at the time (results not shown in tables).

Table 3.

Prevalence of Psychiatrist Support for Medical Assistance in Dying (MAID) under Certain Conditions and on the Basis of a Mental Illness.

Characteristic	% (n)
Support for MAID under certain conditions
Support	71.8 (379)
Do not support	24.8 (131)
Don’t know	3.4 (18)
Support for MAID on the basis of mental illness
Support	29.4 (155)
Do not support	61.2 (323)
Don’t know	9.5 (50)
Support for MAID for a treatment-resistant Axis 1 disorder
Support	23.5 (124)
Do not support	67.8 (358)
Don’t know	8.7 (46)
Support for MAID for a treatment-resistant Axis 2 disorder
Support	9.3 (49)
Do not support	82.0 (43)
Don’t know	8.7 (46)

Table 4 shows the association between psychiatrist beliefs and experiences and support for MAID for mental illness. Psychiatrists who had the belief that legalisation of MAID on the basis of mental illness would change psychiatrists’ commitment to their patients through their enduring suffering were less likely to support MAID for mental illness, compared to those who did not have this belief (odds ratio [OR], 0.22; 95% confidence interval [CI], 0.1 to 0.3). Psychiatrists who had the belief that legalisation of MAID on the basis of mental illness would change psychiatrists’ response to suicidal ideation were more likely to support MAID for mental illness (OR, 1.5; 95% CI, 1.0 to 2.3). Psychiatrists who have had patients in the past who recovered from mental illness but think may have received MAID if it were available at the time were less likely to support MAID for mental illness, compared to psychiatrists who did not have such patients (OR, 0.23; 95% CI, 0.1 to 0.4). Psychiatrists who report having personal faith were less likely to support MAID for mental illness (OR, 0.49; 95% CI, 0.3 to 0.7). Due to small sample size, the estimates for the relationship between the belief that suicide can be a rational choice and support for MAID for mental illness were unstable and could not be confidently reported. Nearly all psychiatrists who had the belief that suicide could be a rational choice reported supporting for MAID for mental illness. Experience working with people with life-threatening conditions and length of time in practice did not have a significant relationship with support for MAID for mental illness.

Table 4.

Psychiatrists’ Beliefs and Experiences Associated with Support for MAID on the Basis of MI.^a

Characteristic	Support for MAID for MI
Belief that legalisation of MAID on the basis of MI would challenge psychiatrists’ commitment to endure with their patients through their suffering
No, % (n)	53.4 (78)
Yes, % (n)	19.8 (58)
Odds ratio (95% CI)	0.22 (0.1-0.3)**
Belief that legalisation of MAID on the basis of MI would change psychiatrist response to persistent SI
No, % (n)	29.1 (85)
Yes, % (n)	38.3 (59)
Odds ratio (95% CI)	1.5 (1.0-2.3)*
Have had patients in the past who recovered from MI but think may have received MAID if it were available at the time
No, % (n)	54.4 (74)
Yes, % (n)	21.6 (60)
Odds ratio (95% CI)	0.23 (0.1-0.4)**
Personal faith
No, % (n)	40.5 (81)
Yes, % (n)	24.8 (61)
Odds ratio (95% CI)	0.49 (0.3-0.7)**
Belief that suicide can be a rational choice
No, % (n)	—
Yes, % (n)	—
Odds ratio (95% CI)	—
Experience working with people with life-threatening conditions
Little, % (n)	28.6 (50)
Much, % (n)	34.7 (105)
Odds ratio (95% CI)	1.33 (0.9-2.0)
Length of time in practice
Resident, % (n)	32.7 (16)
Odds ratio (95% CI)	1.00
<5 years, % (n)	35.3 (18)
Odds ratio (95% CI)	1.13 (0.5-2.6)
5 to 10 years, % (n)	27.5 (11)
Odds ratio (95% CI)	0.78 (0.3-2.0)
10 to 20 years, % (n)	32.3 (31)
Odds ratio (95% CI)	0.98 (0.5-2.1)
>20 years, % (n)	32.7 (70)
Odds ratio (95% CI)	1.0 (0.5-1.9)
Retired	32.1 (9)
Odds ratio (95% CI)	0.98 (0.4-2.6)

^aCI, confidence interval; MAID, medical assistance in dying; MI, mental illness; SI, suicidal ideation; —, could not be reported due to small sample size resulting in unstable estimates.

*P ≤ 0.05. **P ≤ 0.001.

Discussion

We surveyed all psychiatrists in Canada who are members of the CPA to determine their attitudes towards MAID and, more specifically, MAID for mental illness and also to determine what factors influence psychiatrists’ support or lack of support for this possibility. Seventy-two percent of respondents were supportive of the legalisation of MAID under certain circumstances, but only 29.4% were supportive of MAID on the basis of mental illness. Several factors were identified that correlated with lower support for MAID for mental illness, including belief that this would challenge psychiatrists’ commitment to patients with persistent suicidal ideation, reporting having had a patient in the past who would have received MAID were it available but went on to recover from their mental illness, and personal faith.

While this represents the first study of Canadian psychiatrists’ attitudes regarding MAID, similar studies have taken place in other jurisdictions. A study of 290 forensic psychiatrists reported that while 63% believed that physician-assisted death ought to be legal in some circumstances, 58% believed a diagnosis of major depressive disorder should automatically exclude a patient from physician-assisted death.¹⁰ Although the exact context was different, this finding is similar to our sample, in which a strong majority supported MAID in general to be legal, while a slightly less strong majority opposed MAID for mental illness. In the Netherlands, where physician-assisted death has been legal and commonly practiced for many years, there appears to be more support for MAID for mental illness than support found in this study. A survey of 550 psychiatrists reported that 46% were willing to participate in physician-assisted death for patients’ requests based on reasons of mental illness, and 55% felt it was acceptable at least in some circumstances.²⁰ In our study, only 29.4% were supportive of MAID for mental illness. However, a more recent survey of physicians in the Netherlands, one of the most permissive jurisdictions with respect to eligibility criteria for physician-assisted death, looked at physicians’ attitudes towards physician-assisted death based on cause of suffering and found results not dissimilar from ours. Of the 1456 physicians who responded, the majority of whom had actually performed physician-assisted death, 66% found it inconceivable that they would perform physician-assisted death for psychiatric illness.⁸ For such a MAID-friendly population, this is surprisingly similar to the 61.2% who were not supportive of MAID in our sample (whereas 29.4% were supportive and 9.5% were undecided).

Public opinion in Canada has been quite favourable towards MAID, with 60% to 90% expressing support for the legalisation of MAID, depending on the survey.^21–23 There appears to be much less public support for MAID on the basis of mental illness, with 1 recent survey finding 78% of respondents believing that MAID should not be allowed for patients with severe psychological suffering but no terminal illness.²⁴ It is notable that 90% of these respondents thought that MAID should be allowed in some circumstances, which suggests that this is the most MAID-supportive sample yet polled, and still there was overwhelming opposition to MAID on the basis of mental illness. This finding was also true of our sample, but a less strong majority of 61.2% were opposed to MAID for mental illness.

We expected that psychiatrists’ support for MAID would be similar to that of other Canadian physicians, who in the most recent survey done by the Canadian Medical Association were 44.8% in favour of legalising MAID.⁷ Respondents to our survey showed substantial support (71.8%) for MAID under certain circumstances, suggesting that psychiatrists in general are more supportive of MAID than the general physician population in Canada. We also predicted that support for MAID on the basis of mental illness would be considerably less, which has been the case in previous studies of physician attitudes towards MAID on the basis of mental illness compared to terminal physical illness, and was true of our sample as well.⁸

A number of factors seemed to influence respondents’ lack of support for MAID on the basis of mental illness. Not surprisingly, belief that suicide can be a rational act by a competent person was strongly correlated with support of MAID on the basis of mental illness. As expected and consistently reported in the literature, there was a correlation between personal faith and decreased support for MAID on the basis of mental illness.⁸ We had expected that length of time in practice and experience working with people close to death would both influence psychiatrists’ disposition towards MAID; however, neither was significantly related to level of support. Assumptions that length of time in practice would correlate to particular clinical experiences and the ability to mitigate chronic mental suffering may have played into this expectation.

Respondents who believed that legalising MAID for mental illness would challenge psychiatrists’ commitment to endure through their patients’ suffering were unlikely to show support for MAID for mental illness. The potential effects of MAID on the therapeutic alliance is not a well-studied area, but it is not difficult to conceive that the option would affect the traditional dynamic between psychiatrist and patient. Examining transference and countertransference implications would be an important future direction for research in this field.

A related and unexpected finding was that respondents who reported believing that legalisation of MAID for mental illness would change their approach to persistent suicidal ideation were more likely to support it. We had anticipated that respondents concerned that their practice with respect to persistent suicidal ideation would be affected, challenged even, would be less likely to support it, and so this finding was surprising. However, we believe that this can be understood with two considerations. First, those who staunchly oppose MAID for mental illness are unlikely to change their approach to suicidal ideation in a meaningful way, even if such legislation passes, and so they would oppose MAID for mental illness and also report that legislation would not change their approach to persistent suicidal ideation. Second, those who believe that MAID for mental illness should be legalised and are supportive of it as an intervention are almost certain to change their approach to suicidal ideation should such legislation come to pass.

Half of respondents did not think any psychiatric illness could be considered “irremediable” in the sense that MAID would be justifiable, and 57% reported having treated patients who they believe would have received MAID were it legal for mental illness but instead went on to recover. This indicates the substantial uncertainty among highly trained specialists as to how one might balance assessing a patient’s request for MAID against the effect a mental illness has on that request, as well as the prospect for that patient to improve. In one recent review of accepted applications for MAID on the basis of mental illness in Belgium, the author reported that “in all patients, the suffering was chronic, constant and unbearable, without prospect of improvement, due to treatment resistance.”¹² However it is unclear on what basis the determination of “no prospect of improvement” was made and how such an assessment is made. Amongst our respondents, there was substantial variation in what was might be considered “irremediable.” This is understandable, given that when chronically ill patients are followed over longer time periods and given a good standard of care, there is evidence of substantial rates of improvement. This is true of both treatment-resistant depression²⁵ and borderline personality disorder,²⁶ two of the most common disorders in patients requesting and receiving physician-assisted death for mental illness in other countries.^12,13

These issues remain controversial in countries where physician-assisted death for mental illness is legal. Referral to other physicians for second opinions is a relatively common practice in psychiatry and may help practitioners work through what are sure to be very difficult and complex questions. However, a recent examination of a fraction of patients who received physician-assisted death for mental illness in the Netherlands found a considerable number of cases where there was disagreement among assessing physicians, yet physician-assisted death proceeded in most instances with the disagreements unresolved.¹³

Clinical Implications

Our study suggests that psychiatrists have significant reservations about MAID for the mentally ill. Although Bill C-14 has temporarily answered whether this will be a possibility in Canada, debate on this issue is both active and lively. In our opinion, the clinical implications of this debate are quite profound for our patients: some have suffered with chronic mental illness for much of their lives, despite our best efforts and are desperate for relief, and others are in very vulnerable positions and wish to pursue MAID despite legitimate prospects for improvement. However, these issues are complex, and changes to MAID eligibility must be carefully weighed against the imperative of not undermining suicide prevention strategies for the mentally ill. The policy decisions that stem from this debate will have a direct impact on patients who desire to end their lives. Many psychiatrists have indicated it would directly change the way they approach persistent suicidal ideation, and many think that legalisation of MAID for mental illness would pose a fundamental challenge to their work. Ongoing dialogue among psychiatrists will be essential to establish policies and guide our profession, as well as to make recommendations to legislators that balance concern for the suffering of our patients with the need to protect our most vulnerable patients. This survey is part of a broader discourse that informs our profession on this contentious and important issue.

Limitations

Our study has several limitations. The nonresponse rate does raise the possibility of response bias; it is possible that psychiatrists with particular beliefs either for or against MAID for mental illness were disproportionally represented. However, there are very similar demographics between the study sample and the general pool of CPA members, and respondents’ general support for MAID and lack of support for MAID for mental illness are similar to that of the general population.^21,22,24 Another limitation is the failure to capture the many psychiatrists, particularly in Quebec but also in Ontario, who are not members of the CPA.

Conclusion

This study found that most psychiatrists, despite supporting MAID in general, do not support the legalisation of MAID for the mentally ill. Objections appear based on concern for vulnerable patients (particularly those with persistent suicidal ideation), the nature of the psychiatric illness not being easily qualified as “irremediable,” personal moral objections, and concern for the effect it would have on the therapeutic alliance.