Introduction

Given that medical assistance in dying (MAID) is a controversial issue involving conflicting values, public attitude surveys play an important role in policy discussions in modern pluralistic societies. Although such surveys may provide insights about the public’s general attitudes, there is also a need to assess the public’s perspective on aspects of specific policies to best inform and capture the complexity of MAID policy and practice.

In Canada, MAID was legalised in 2016 and underwent an expansion in eligibility requirements in 2021.1 There have recently been a number of reports of Canadians applying for and receiving MAID after refusing recommended treatment or because they lacked access to standard treatment or resources.2–7 For instance, a young man in his 30s with a curable cancer refused standard treatments and instead requested MAID which he later received.2 In other cases, the requestors chose MAID after failing to access needed treatments or resources that they said would make their lives bearable. One early case was Sean Tagert, 41, who struggled to afford an additional 8 hours per day of home care for his motor neuron disease.8 Non-terminally ill persons with disabilities have chosen to end their lives due to lack of resources, as in the case of a woman with chronic environmental allergies who could not afford housing with proper ventilation to mitigate her symptoms.5 Many disability organisations, activists and academics have raised concerns about MAID for those who request it due to lack of access to care.9–12 In contrast, some authors have argued that MAID—even under conditions those authors consider unjust—is permissible,13 and Health Canada documents acknowledge that such types of MAID cases are not prohibited.14 15 Although a recent survey demonstrated that 73% of Canadians support the MAID policy in general,16 these controversial cases continue to draw criticism.17 18 A recent poll showed that most of the public believe MAID eligibility should not be expanded without first improving access to mental healthcare.19 Legalisation of MAID for mental illnesses again has been postponed20 amid concerns about the lack of adequate mental health support for psychiatric patients.11 19 21 22

The purpose of this study was to assess the Canadian public’s attitudes toward MAID when effective remedies are either refused by MAID requestors or are inaccessible to them. Since MAID in such cases is not prohibited by the current law, we also assessed the public’s knowledge of the implications of the MAID law.

Materials and methodsRecruitment

We recruited 2140 unique participants using vendor CloudResearch’s Prime Panels to create a quota sample23–25 reflecting the Canadian adult population in the 2021 census,26 by age, gender, level of education, household income and province. The anonymous cross-sectional survey was hosted on SurveyMonkey between 8 August and 23 August 2023.

Survey

The survey development included feedback from the Empirical Research Laboratory of the Department of Bioethics at the National Institutes of Health (NIH) Clinical Center and pretesting on Amazon MTurk (selecting for Canadians, using open-ended feedback on survey questions). A French-Canadian version (via NIH Library translation services) was available for respondents.

The full text of the survey can be found in online supplemental file 1 and is summarised in tables 1–3. The first part of the survey covered the following areas: participants’ knowledge about the Canadian MAID law (one question for subjective assessment of one’s knowledge level which we refer to as ‘subjective knowledge’ and five true/false questions which we refer to as ‘objective knowledge’), general attitude toward Canadian MAID policy (using a question from a previous Canadian public survey as a validation item16 and as one of the primary outcome questions), opinion on acceptable circumstances for MAID and MAID for mental illnesses.

Table 1

Respondent characteristics and baseline knowledge about Canadian MAID law, n=2140

Table 2

Attitude toward Canadian MAID law in general, acceptable circumstances for MAID and MAID for mental illness

Table 3

Responses to questions regarding four scenarios depicting MAID in context of refusal of treatment or lack of access to resources

Participants were then randomised to two arms (to reduce respondent burden). One arm depicted two scenarios based on a physical condition (a man with cancer who refuses a treatment with a 70% chance of cure; a woman with severe chemical sensitivities who cannot find medically appropriate, affordable housing). The other arm depicted psychiatric conditions (a man with a history of trauma, anxiety and depression without response to multiple treatments who refuses ketamine treatment recommended by his psychiatrist; a woman with well-controlled post-traumatic stress disorder and depression who relapses due to lack of access to continued care); the respondents in this arm were informed of the planned legalisation of MAID for mental illness, and were asked to answer assuming the law was in effect. Each scenario was followed by five case-specific questions that covered: legality of MAID in the scenario, support for MAID in the scenario (a primary outcome question) and opinion regarding three different statements that a provider might make when presented with the scenario.

The survey also collected demographic data (age, gender, household income, education, province, race, finances, religion, religious service attendance) as well as their life experiences with disability and mental illness, access to resources and MAID.

Analysis

We conducted descriptive (frequencies (percentage) and means (±SD), as appropriate) analyses of respondent characteristics. We examined the relationship between each of the five variables reflecting attitude toward MAID (ie, support for MAID in general and for MAID in each of the four specific medical and psychiatric scenarios depicting refusal of effective treatments and lack of access to needed resources) and the variables for knowledge, demographics and life experiences—using general linear models for analysis of variance of continuous data and tests for trend for categorical variables (2×n Cochran-Armitage or n×n Jonckheere-Terpstra tests, depending on whether variables were singly or doubly ordered). Data were analysed using SAS V.9.4.

For the open-ended question asking why they chose their answer, WC and IMA independently read through responses from those who supported MAID and those who answered otherwise, for each scenario, to develop codes; it was determined that saturation was reached by about 50 responses per response group so a fresh set of additional 50 participants who supported MAID and 50 participants who answered otherwise, for each of the four scenarios, were selected and these comments (ie, 100 for each of the 4 scenarios) were independently coded, with disagreements resolved by discussion with SYHK to arrive at a consensus.

Patient and public involvement

Patients or the public were not involved in conducting the study.

ResultsRespondent characteristics

A total of 2140 completed the survey, and their personal characteristics are given in table 1. Our respondents were similar to the results of Canadian 2021 census in age (within 8%), gender (within 2%), level of education (within 9%), household income (within 6%) and province (within 3%).22

General knowledge of MAID policy

40.3% of respondents self-reported a low/very low level of knowledge of the Canadian MAID law, and only 12.1% of respondents answered four or five of the five true/false questions correctly. Only 19.2% of participants correctly answered that patients do not need to have a terminal illness to qualify for MAID, 20.7% correctly answered that patients may refuse medically effective treatment and still qualify for MAID, and 42.4% correctly answered that the eligibility criteria for MAID would expand to include mental illnesses in March 2024. The mean Knowledge Index score (# correct/5) was 2.3 (±1.1).

General opinions regarding MAID

73.3% of respondents supported/strongly supported allowing a person to receive MAID in Canada if all the legal conditions were met. This matches the recent national Canadian survey using the identical question in which 73% moderately or strongly supported MAID, thus validating our quota sampling.16 Only 40.4% of respondents supported/strongly supported MAID for patients whose sole underlying condition is a mental illness. Furthermore, almost two-thirds (64.4%) of the respondents responded that MAID should only be offered as a last resort option (34.3%) or if not last resort, then only in exceptional circumstances (30.1%).

Specific opinions about scenarios depicting refusal of or lack of access to effective treatments or remedies

Across all four scenarios, a minority of participants correctly answered that the patients depicted could still qualify for MAID (26.5%, 23.2%, 31.7%, 30.7% for the two physical illness scenarios and two mental illness scenarios, respectively).

Participants demonstrated a low level of support for providing MAID for the requestor (32.0%, 23.2%, 30.9%, 25.7%). We asked several follow-up questions to gain insight into the rationale for their answers.

In the analysis of the open-ended explanations, the two most common reasons for supporting MAID were (1) respect for the patient’s autonomy (expressed by 26–46% of ‘yes’ respondents across the four scenarios) and (2) the belief that MAID can end the patient’s suffering (14–22% of ‘yes’ respondents). The most common reason against MAID (36–68% of ‘other’ respondents) was that ‘things can get better’ for the patient through treatment, government assistance or other resources, and therefore that the scenarios were not ‘last resort’ situations of MAID. Other common justifications varied by scenario. For instance, in both the medical and psychiatric lack of access scenarios, some respondents voiced versions of ‘the government should do more’ (12% of ‘other’ respondents for medical, 20% for psychiatric scenario) to help the patient access resources that would improve their situation, rather than provide MAID. In contrast, 16% of ‘yes’ respondents to the lack of access medical scenario argued that it is ‘not the patient’s fault the system is failing them’, and therefore, MAID ought to be provided to these patients (see online supplemental file 2).

In terms of the three follow-up questions depicting what a clinician might say to justify providing or not providing MAID in the scenarios, the respondents agreed most strongly with the provider’s statement to not provide MAID because MAID is not a last resort option, across all four scenarios (66.5%, 67.4%, 67.2%, 65.1%).

Respondent characteristics and their views

Table 4 is a summary of the direction of the relationship between respondent characteristics (demographics and life experiences) and their general support for MAID and for MAID in the four study scenarios (full results in online supplemental file 3).

Table 4

Relationship between respondent characteristics and attitude towards MAID in general and toward MAID in specific study scenarios

Older age, higher income, more education, lower religious service attendance and being white (vs other) were all associated with higher general support for MAID; however, for the four specific scenarios, none of these characteristics were associated with higher support for MAID and were instead either not associated or were associated with lower support for MAID. For example, older age, while associated with greater support for MAID in general, was associated with lower support for MAID in three of four specific scenarios. In terms of income, 82.9% of those making more than $150 000, compared with 66.4% of those making less than $30 000, supported MAID in general, but for the medical lack of access scenario, support for MAID was only 8.2% in the $150 000+ group vs 25.1% in the <30 000 group (online supplemental file 3). For MAID in general, attending religious service more than once a week was associated with lower support than those who never do (53.9% vs 79.1%), but for the medical lack of access scenario, this was reversed (37.5% vs 18.4% support) (online supplemental file 3).

Personal life experience variables either were not associated with support for MAID (having a disability or not having a needed service unavailable in past 12 months) or were associated with lower support for MAID for some scenarios (having chronic condition, close to someone with disability, having a mental illness or being close to one who has one). One exception was having had a family or friend who received/requested MAID, which tended to be associated with greater support for MAID in general and MAID in both refusal scenarios.

Relationship between knowledge and support for MAID

There was a consistent positive relationship between both subjective (five levels) and objective (six levels) knowledge (average correct score out of five questions) and support for MAID in general and for all four specific scenarios (p<0.001 for all).

Discussion

The purpose of our study was to assess the public’s views regarding cases similar to the recent reports of patients receiving MAID even when refusing effective and recommended medical treatment2 or when their medical conditions were exacerbated by socioeconomic limitations to accessing care.3 5 8 27 28 These cases, while controversial, are legally permissible under the Canadian law: none of the reported cases have been determined unlawful and Health Canada guidance acknowledges their permissibility.14 15

Our findings are consistent with previous surveys16 19 but go beyond them as we used more detailed scenarios. There are several important findings. First, although we confirmed that there is a high level of general support for MAID (73.3%),16 support for MAID was much lower in scenarios resembling the reported controversial cases of MAID, that is, in the context of refusal of treatment or when the requestor lacks access to resources that can reduce suffering. This is supported by the fact that only 23.2–32.0% of respondents supported MAID in such situations. Thus, although Canadians broadly express support for their MAID law, most do not support the practices that are (or scheduled to be) permitted by that law.

The low support for MAID in the four scenarios was consistent with respondents’ answers for other questions that reflect their underlying views about what MAID should be for. The open-ended text explanations show that majority believe MAID is not appropriate when there are identifiable ameliorating interventions. Additionally, nearly two-thirds of respondents state MAID should be offered only in exceptional circumstance, and agree with doctors not providing MAID unless it is a last resort. These findings are consistent with a poll showing that 82% of Canadians agree/strongly agree that ‘MAID eligibility should not be expanded without improving access to mental healthcare first’.19

Second, the level of knowledge (subjective or objective) regarding specifics of the Canadian MAID law and its implications is low. Only 19.2% of participants correctly answered that terminal illness is not required for MAID eligibility, only 20.7% know that refusal of effective treatments is compatible with MAID eligibility and only 23.2–31.7% of participants believed that MAID in the four specific scenarios was not prohibited by law. Only 42.4% correctly answered that MAID for those whose sole condition is a mental illness was scheduled to become legal in March 2024. It is striking that most Canadians who disapprove of the controversial reported cases of MAID are unaware that the law they largely support permits such cases.

There was in general a close relationship between greater knowledge of conditions under which MAID is legal and stronger support for MAID (both general attitude toward MAID and for specific scenarios). In a cross-sectional survey, it is not possible to determine the causal relationship for associations between variables, and so one can only speculate regarding the potential direction of influence. On one hand, it may be that as one gains more knowledge of the MAID law or has positive personal experiences with legalised MAID, one develops a more positive view of MAID, or it may be that people are reluctant to support something they believe is currently illegal. Such explanations posit a direction of influence from knowledge of legality, or experiences with legal MAID cases, to opinions about MAID. On the other hand, some may argue that even if such explanations were valid for some people, they underplay the fact that MAID engages people’s deeply held values, which are unlikely to be easily influenced. Perhaps it is more likely that those who already have a positive attitude about MAID tend to follow the issue more closely; further, it may be difficult for people to believe something to be wrong (especially regarding life and death decisions) and also believe that it is something their government would permit. Which of these two explanations has had a greater role will require further study.

Third, the pattern of association between attitudes toward MAID and demographic variables is notable. It is not surprising that secular, older, educated, wealthier white respondents are the most supportive of MAID. But it is striking that for the four specific scenarios, none of these characteristics were associated with higher support for MAID and were instead often associated with lower support for MAID, especially for the lack of access scenarios. Clearly, our respondents make a distinction between the different situations in which MAID might be requested. Even if someone is very supportive of end-of-life MAID, they may not support MAID in situations of treatment refusal or health inequality.

Furthermore, respondents with chronic conditions and those close to people with disabilities expressed lower support for MAID in some scenarios, and those with mental illnesses or close to someone with one were less likely to support MAID in the psychiatric scenarios. However, those with personal experience of MAID in someone close to them were more supportive—but only in the refusal scenarios, not in the lack of resource scenarios. The positive association is not surprising since the respondents may share the same values as the loved one, or because they saw it as a good experience, or because it ‘normalises’ the practice for them.

From a policymaking perspective, there are two issues that may warrant future research and discussion. First, why is there such a large gap between Canadians’ support for the MAID law in general and their support for specific implications and provisions of that same law? Second, how significant is it that the public do not seem to see this gap because they do not believe that the law is as permissive as it really is? Future research might focus on how an apparently popular idea—legalised MAID—came to be implemented in a way not supported by most Canadians. Perhaps the low level of knowledge reflects the rapid changes in the law, going from an ‘end of life’ (reasonably foreseeable death) practice to one that is not—all the while retaining the language of ‘irremediability’ that, unlike in the Netherlands or Belgium, is ultimately defined by the MAID requestor.29–31 Another contributing cause of the policy apparently running ahead of public opinion may be, as reported in a recent investigative report,32 the policy influence exerted by a single, well-funded lobbying group.

Limitations

One limitation is that our respondents were not a probabilistic sample and may not be generalisable to the target population. However, our targeted quota sampling of Canadian adult population in terms of age, gender, household income, education and province resulted in a virtually identical response distribution for the general MAID attitude question used in a previous nationally representative survey.15 Second, although we provided more context than is typical, this was still a cross-sectional survey with its usual limitations on internal validity. The context of the MAID debate is a complex one such that even a vignette-based survey that attempts to go beyond the usual modes of opinion surveys will not be able to capture all the relevant factors that may go into people’s views on MAID. Third, because Quebec has its own MAID laws that do not perfectly coincide with the federal law, the survey may have been more challenging for Quebecois respondents. Fourth, the survey was fielded in August 2023 during intense public debate leading up to scheduled implementation date for MAID for mental illness in March of 2024. Although this means there may have been greater than usual coverage of the topic of MAID in the media, it also means there could have been greater exposure to misleading information from both sides of the debate. Finally, the tests of association between variables were numerous and not adjusted, and the associations found should be seen as suggestive and exploratory.

Conclusions

Although most Canadians support their MAID law—especially among those who are secular, older, educated, wealthier or white—that support may be tied to an incorrect assumption that the law is restricted to persons at the end of life who have access to timely and effective treatments and other resources. Only a minority are supportive of MAID in situations where patients refuse effective treatments or when they lack access to standard resources needed to make conditions manageable. Interestingly, the attitude toward MAID in these scenarios is not correlated in the same manner with demographic factors as it is for general support for MAID, suggesting that other factors—such as shared moral views across demographics—play a greater role. For a topic such as MAID which engages deeply held values on all sides of the debate, it seems important to understand the reasons behind the gap between public policy and public opinion, especially as expansions of the MAID law continue to be debated. Finally, although the survey focused on the Canadian situation, there may be lessons for other jurisdictions debating MAID legislation. In particular, even when polls show that there is general support for MAID, it may be important to understand the public’s understanding of and views regarding the downstream effects of various MAID policy options.