Article body

introduction

A recent New York Times interview with Canadian children’s author Robert Munsch revealed that he struggles with dementia and Parkinson’s disease, and that several years ago he was approved for Medical Assistance in Dying (MAID) (1). The article received much attention in Canadian media and public debate, fueling ongoing discussion about advance requests for MAID. Advance requests for MAID have already been introduced by provincial law in Quebec, even though it remains prohibited under the Criminal Code. Advocacy organization Dying with Dignity jumped on the occasion to renew its call to legalize the practice (2). Yet, advance requests raise profound ethical, legal and policy challenges (3,4), which I will touch upon here mostly indirectly.

For me, three components of the Canadian reporting on Munsch’s choice for MAID stood out. Most seemed to accept MAiD as an obvious, or even the only reasonable choice for people who, like Munsch, are facing physical or cognitive decline and fragility. All except one commentator (5) reported uncritically Munsch’s in-jest, but rather ableist comments about why he preferred death over decline with dementia: to avoid that his wife would be “stuck with me being a lump.” (6) Most did not critically question why he could be approved for MAID several years earlier, even though he and his family declared that even now, he still has a reasonable quality of life.

The media report about Munsch coincided with the publication of a new Ontario Chief Coroner’s MAID Death Review Committee (MDRC) report on dementia and MAID (7). The MDRC, of which I am a member, is an interdisciplinary committee set up by the Chief Coroner’s office. It reviews the files of a selection of MAID cases. A team of trained professionals review all of Ontario’s reported MAID cases. A sub-set of these cases are then shared with the MDRC to explore the common issues they raise. The MDRC’s review and discussion informs Coroner’s recommendations for various stakeholders. The Coroner’s office has so far prepared seven reports that contain narrative summaries of cases, anonymized discussions of these cases, and recommendations for policy and guidance on various topics. My comments here reflect my personal views.

The MDRC dementia report, and earlier reports, document numerous examples of questionable MAID practices that should be taken seriously by policy makers, health professional organizations, and anyone concerned about patient safety. They confirm in my opinion, concerns that have been raised in the past about Canada’s MAID practice, including in media reports and expert commentary and analysis (8): 1) that people are at times approved for MAID based on remarkably flexible interpretations of legal access criteria and following minimalistic capacity and informed consent procedures; 2) that unaddressed mental health or substance use issues and lack of social supports are not seen as a hindrance for providing MAID; and 3) that overall, Canada’s MAID policy and practice focus on facilitating broad access to death as a form of therapy, not on protection against premature death. This latter concern is evidenced in various guidance documents issued by the Canadian Association of MAID Assessors and Providers (CAMAP), a semi-professional organization that has received significant funding from the federal government to develop guidance and education.

In what follows, I illustrate these issues using examples of the various MDRC reports, with a particular focus on the last dementia report. Worth noting is that earlier MDRC reports also contain some narrative cases of persons with dementia opting for MAID, in circumstances seemingly comparable to those of Mr. Munsch.

How persons with dementia are approved for MAID under current MAID practice

The MDRC report confirms that MAID for reasons of dementia occurs frequently in Ontario, even if it has until now, at 1% of cases, not been a dominant ground for MAID. In Ontario, 103 people received MAID for reasons of dementia in the 2023-2024 period, with nearly 2/3rds of the cases in 2024 (7). All but one obtained access under ‘Track 1’ and were thus considered to be approaching their natural death. All likely involved euthanasia, i.e., a lethal injection, since there are nearly no assisted suicide cases (self-administration of prescribed drugs) in Canada.

In the context of a disease that inherently involves a progressive loss of cognitive function, which becomes particularly pronounced closer to the person’s natural death, these MAID approvals evoke several questions: 1) Since about all were assessed as Track 1 cases, with situations of “reasonably foreseeable natural death” (RFND), were there no problems with capacity for decision-making? How thorough were the capacity assessments? How did people provide informed consent? 2) In the cases where capacity for decision-making was abundantly clear, and thus their death not likely to occur in the near future, why were these applications not considered under Track 2? 3) How do persons with an early diagnosis of dementia satisfy the MAID access criteria, which require under both Track 1 and 2 an advanced state of irreversible decline of capability and intolerable suffering?

Concerns with Capacity and Informed Consent Assessments

The answer to the first question is straightforward: case narratives and discussions in the MDRC dementia report reveal that there are concerns about capacity for decision-making in at least some of the cases. The MDRC dementia report, but also earlier MDRC reports, document how some MAID assessors followed questionable capacity assessment procedures; and how information about the capacity assessment process is often too limited to determine how it was done.

It is important to emphasize again that the federal criminal law does not yet allow MAID on the basis of advance requests, and that MAID applicants must be capable of consenting at the time of the MAID approval. Moreover, unless they signed a Waiver of Final Consent (WofFC), permitted only under Track 1 (I briefly discuss this further, below), they must confirm consent just prior to receiving a lethal injection.

Obviously, cognitive decline or mental illness do not, in and of themselves, deprive people of capacity (9). Canadian law emphasizes that every person is presumed to have capacity. Yet, severe cognitive decline is among the most common reasons for findings of incapacity. When it comes to informed consent, not only does cognitive decline evoke concerns in the context of dementia, but also the impact of increasing vulnerability, dependence on family members and others, existential fears, isolation, anxiety, and intersecting issues of physical decline and mental health challenges, such as depression and severe anxiety. As Ramona Coelho writes in a commentary on the MDRC dementia report, “perceived loss of dignity, feeling like a burden, emotional distress, and the fear of decline” often characterise the experience of living with dementia (5). This is often particularly pronounced shortly following diagnosis.

It would seem in a dementia context particularly appropriate to take time to evaluate what is going on, whether problems of adjusting to the challenges associated with dementia, family pressure or burn-out, or other factors, are inappropriately influencing MAID requests, and whether people’s suffering could be addressed in other ways. Yet, the reports document cases of people getting MAID while they struggle with adjustment to a long-term care home placement, or to loss of ability to engage in activities; and a case of a woman whose spouse was struggling with caregiver burnout — she had previously explicitly indicated she wanted palliative care rather than MAID, for personal and religious reasons (10).

Some cases combine several troubling components. Take the case of 6F, a woman in her 80s with advanced dementia (7). She was initially hospitalized following multiple falls. During the hospital stay, a family member initiated a first MAID request, allegedly after she expressed a “desire to die”. A MAID ‘navigator’ felt uncomfortable “facilitating a written request,” due to the woman’s severe cognitive limitations. The woman then opted for continued care in a long-term care home. Four months after her admission, a family member initiated a new MAID request. This time, an assessor approved her for MAID during a single interaction. The assessor mentioned “marked existential suffering” and that it was “clear that she did not want to continue to live as she [was]” (7, p.34). They further acknowledged “communication challenges” and that the assessment was “managed in a way that worked for her [i.e. the patient].” The assessment report contained further general statements about capacity, but no details of any in-depth assessment. The second assessment also failed to mention how capacity was determined.

The MDRC report documents additional capacity concerns just prior to the MAID procedure itself, which took place one week after approval. Mrs. 6F appeared “overwhelmed by the presence of additional visitors” (7, p.34), who were therefore asked to leave the room. “Final express consent” was deemed capable and informed based on her “ability to repeat the consent question and via squeezing the provider’s hand.” (7, p.35)

Hand squeezing as expression of informed consent is matched by at least one other case of remarkably minimalistic consent in an earlier MDRC report. A report on “Complex Same Day/Next Day Provisions of MAID” describes the case of Mr. 4C, a man in his seventies diagnosed with metastatic cancer who requested a MAID assessment five days after palliative care admission (10). He was at risk of losing capacity due to brain metastases and opioid analgesics. Before a formal assessment could be conducted, he experienced cognitive decline and could no longer communicate. Even though the palliative care team informed the MAID provider that he no longer had decision-making capacity, the provider “vigorously roused” the patient, who “mouthed yes” when asked if he wanted MAID (10). After withholding pain medication, he appeared more alert, and the MAID provider organized a second assessment, transmitting questions and answers to a second assessor virtually. The patient’s nodding or “mouthing yes” in response to questions was again accepted as expression of capable and informed consent.

While these cases are clearly at the extreme end of remarkable capacity assessment and informed consent procedures, several other cases in the MDRC reports, some of which I discussed elsewhere (11), raise related capacity concerns. Some also raise concerns about potential pressure or influence by family members, at times with remarkably fast approval and provision of MAID, with one report discussing various cases of same-day or next-day provision of MAID (10). Take the earlier mentioned case of a woman who preferred palliative care, whose husband struggled with caregiver burn-out. In that case, it was the husband who contacted the MAID coordinator service, after a request for placement in hospice care was rejected. She received MAID the same day as the rejection of the request.

Dementia and Flexible Interpretations of a Reasonably Foreseeable Natural Death

Even though some other cases of the MDRC dementia report involve people with severe cognitive decline and questionable capacity for decision-making, with only limited documentation of how capacity was assessed, other cases involve people in earlier stages of dementia with less cognitive decline. But those cases raise the question how they were deemed to fulfill the legal access criteria related to intolerable suffering and an advanced state of irreversible decline of capability, and how nearly all were approved under Track 1.

It is important to emphasize here that to be approved for MAID, a person must have a grievous and irremediable medical condition, which is defined as: a serious illness, disease or disability, with an advanced state of irreversible decline of capability, either of which cause intolerable physical or psychological suffering that cannot be relieved in circumstances that they find acceptable. In addition, to qualify under Track 1, a person’s death must be reasonably foreseeable (12).

When it comes to the qualification under Track 1, some of the dementia cases as well as cases in earlier MDRC reports confirm how a remarkably broad interpretation of the concept of a “reasonably foreseeable natural death” (RFND) has been embraced by some of Canada’s MAID providers. These interpretations are likely influenced by CAMAP guidance documents, which in turn were inspired by some academic scholars who have been putting forward interpretations that facilitate a very broad application of MAID.

According to the CAMAP guidance on RFND, for example, a person whose condition would make them normally only qualify under Track 2, can still fit the RFND criterion to qualify under Track 1,

…if they have demonstrated a clear and serious intent to take steps to make their natural death happen soon or to cause their death to be predictable. Examples might include stated declarations to refuse antibiotic treatment of current or future serious infection, to stop use of oxygen therapy, to refuse turning if they have quadriplegia, or to voluntarily cease eating and drinking

13, p.1

So far, no federal or provincial health authority and none of the health professions’ regulatory colleges have pushed back against this remarkable interpretation, which facilitates a circumvention of the longer and more ‘complex’ (albeit in my view insufficient) review procedures the legislature deemed necessary for Track 2.

MDRC reports reveal that CAMAP’s interpretation, allowing a transfer of Track 2 into Track 1 cases, has found its way into MAID practice. Take the case of Mr. 5B, a man in his 60s with cerebral palsy (14). He was noted to struggle with “profound psychosocial suffering and loneliness due to limited social relationships and isolation from community,” and “anticipatory fear of further dependence” (14, p.22). The MAID assessor emphasized that “they did not counsel Mr. B on Track 1 vs Track 2 determinations to avoid influencing or coaching his personal care decisions,” with respect to “voluntary stop eating and drinking.” (14, p.22) The report does not reveal if anyone else counselled the patient. What we know is that Mr. 5B stopped eating and drinking and was then deemed to qualify under Track 1 because of a RFND.

A CAMAP Guidance on Dementia further explicitly claims that a mere diagnosis of Alzheimer’s satisfies the RFND criterion, even though it acknowledges that such persons have a median survival of 4 to 7 years (15). It embraces an interpretation put forward by Jocelyn Downie and Jennifer Chandler, who suggest that a RFND is fulfilled either if the person’s death is “not too remote” or “the cause of the person’s death is predictable” (16). But it is questionable whether this is really what the legislature had in mind. A Canadian government website on MAID states, for example, that RFND requires “the person to be approaching the end of their life in the near term.” (17)

Those who interpret the RFND very broadly tend to invoke an interpretation of the access criteria and RFND in a 2017 Ontario Superior Court decision by Justice Perell (18). However, a careful reading of that judgment reveals that it is doubtful that an early onset of dementia would align with this interpretation. Justice Perell recognized that there may be doubt in some cases whether a person has a RFND. But the case “of an almost 80-year-old woman in an advanced state of incurable, irreversible, worsening illness with excruciating pain and no qualify of life” (18, para 87) clearly satisfied for him the criteria. He suggested that the “proximity or remoteness of death and the duration of suffering must be considered.” (18, para 85) An early dementia diagnosis with years of life left is clearly fundamentally different from the facts in this case.

Dementia and Advanced State of Irreversible Decline of Capability

CAMAP further puts forward that the future loss of capacity associated with dementia results prima facie in the satisfaction of the criterion of an advanced state of irreversible decline (19). In a recent article, Christopher Lyon, Scott Kim, and I suggest that this shows that for CAMAP “the potential future impossibility to satisfy a legal requirement for access to MAiD (resulting from possible loss of capacity …) turns a person’s present health situation into an advanced state of irreversible decline” and “[a] potential ineligibility condition [into] positive evidence of eligibility” (20, p.15). It is puzzling how such intellectual gymnastics have not yet been challenged by our medical professional organizations, which appear to grant CAMAP remarkable freedom in constructing peculiar interpretations of the law.

The MRDC dementia report reveals that committee members had starkly differing opinions on this interpretation. Some opined that to respect the legislative intent, the term should be reserved for situations where a “loss of function is profound and enduring, and significantly compromises [people’s] quality of life.” (7, p.38) They suggested that a subjective interpretation would permit ableist biases to influence assessors’ judgment about the value of life for people who struggle to perform complex tasks. Others, seemingly following CAMAP guidance, argued that the term should be understood within a “person-centred and psychosocial framework”, which recognizes the “subjective meaning persons assign to their cognitive and functional decline.” (7, p.38) In other words, the fear of losing higher-order cognitive functions would be an “advanced state of irreversible decline” for some people with strong cognitive capacities in the first place, but not necessarily for others (7) A retired mathematics professor’s fear of losing the ability to play chess might, according to this approach, thus be evidence of an “advanced state of irreversible decline”, whereas for another person it would be the ability to play bingo. In some of the reported cases, assessors explicitly mention people’s difficulty in practicing their usual hobbies.

In my opinion, interpreting the concept of “advanced state of irreversible decline of capability” as based on a largely subjective and individual experience further erodes an important safeguard. It also opens the door to the influence of assessors’ own subjective interpretations and biases of what a life with cognitive decline would mean to them. The emphasis on such subjective experiences makes it difficult to create the time and space for adjustment and coping strategies in the face of slowly declining cognitive capacities. Further, in the context of dementia, it seems odd to suggest that well-recognized clinical stages of dementia, which range from very mild cognitive decline, to moderate, to severe and very severe cognitive decline,[1] would have no bearing on the determination whether a person is already in an advanced state of decline or not.

Dementia and Waivers of Final Consent: How Flexible Interpretation Appears to be Used to Enable Still Prohibited MAID on the Basis of Advance Requests

CAMAP’s interpretations of RFND and advanced state of decline of capability are particularly relevant in the context of dementia because this has resulted in the construction of a form of advance requests for MAID. As mentioned before, WoFCs are only permitted in the context of a Track 1 RFND case. They were introduced in 2021 for situations where a person who is approved for MAID is at risk of losing capacity and would therefore not be able to confirm consent prior to the procedure. With a WofFC, a specific date in the near future must be identified for when the MAID procedure will take place. An advance request for MAID, which, as stated above, is not legal under the Criminal Code, involves a stipulation by a patient of the future conditions under which a third party can decide that the MAID procedure of the then incapable person can proceed (21).

The MDRC report on the topic of WofFC illustrates this with the case of a man diagnosed with Alzheimer’s disease and Parkinsonian symptoms, who was distressed by short-term memory loss (21). The MAID assessor approved him under Track 1. The assessor documented that Mr. A experienced “suffering due to current and future anticipatory suffering of dependency,” including “anticipatory fear of admission to long-term care”, but that he continued to enjoy his quality of life (21, p.13-14). With the assessor’s help, Mr. A signed a WofFC with a date for the MAID provision 3.5 year later. The WofFC further identified various factors that should trigger a MAID procedure in the future (e.g., difficulty eating or swallowing, bodily tremors, inability to verbally communicate), as if one would do for an advance request.

One month later, his family physician approved him under Track 2. Two months later, a third assessor, who became the MAID provider, agreed with the first assessor and approved him under Track 1. Less than a year later, Mr. A was hospitalized after a fall. He was delirious and agitated, hallucinated, and was rapidly declining. During “a period of cognitive improvement”, he was deemed capable of re-consenting and, according to the MAID provider “[t]he WofFC was not invoked” before the lethal injection (20, p.15).

In addition to capacity concerns, the case reveals how an inventive use of different CAMAP guidance documents (on dementia, RFND, and waiver of final consent) opens the door to practices that appear to run counter to or at least stretch the contours of the law, in this case by allowing a WofFC that has all the characteristics of an advance request for MAID. That the MAID report states that WofFC was “not invoked” may reflect that the provider realized its validity could have been contested.

Dementia and Track 2 MAID

Mr. A’s case is also relevant for a discussion of why approving persons with dementia for MAID under Track 2 may be in tension with the current law. It is not obvious that Mr. A satisfied the criterion of “an advanced state of irreversible decline of capability” and “intolerable suffering” that cannot be relieved at the time of his approval, since he opted for MAID several years in advance, when he appeared to have a good quality of life. These criteria need to be satisfied for a Track 2 approval. Mr. A’s story reveals that MAID practitioners appear to follow the CAMAP guidance, interpreting a future loss of function as sufficient to conclude a current advanced state of irreversible decline of capability, and the fear of future cognitive decline as creating intolerable suffering already now, with future provision of MAID as the reasonable solution.

The coroner reports do not contain sufficient detail to explore in full what went on during some of these assessments shortly after a dementia diagnosis. But it is fair to also ask questions about the apparent ease by which some people obtain approval for MAID following a first diagnosis of Alzheimer’s or dementia under Track 2.

As some members of the MDRC point out, it is important that people with a diagnosis of dementia receive sufficient “psychological support, existential therapies, and a focus on optimizing quality of life and function.” (7, p.40) Members also emphasized “the critical role of specialized geriatric services in facilitating early and ongoing conversations about diagnosis, prognosis, and care planning.” (7, p.31) In a commentary on the report, one of the MDRC members, Ramona Coelho, emphasizes that only 13.6% of dementia patients who died by MAID received palliative care (5). According to her, this raises concerns about adequate support for people when they are diagnosed with dementia; support that could assist them and family members in adjusting to the challenges of (future) cognitive decline. As some MDRC members frame it, “palliative care should be a foundational component of end-of-life planning for persons with dementia, particularly when MAiD is being considered” and “supportive care options” should be available to mitigate the concerns resulting from a dementia diagnosis (7, p.42). The report refers also to the importance of dignity therapy (7), a therapeutic approach developed by Canadian psychiatrist Harvey Chochinov (23,24) aimed at alleviating the anxiety and distress of people facing terminal illness and at enhancing their psychological, existential, and spiritual well-being.

Conclusion

We cannot know if in individual cases — such as that of Mr. Munsch and some of the MDRC reported cases — better supports, early access to palliative and geriatric services, community-based care, and alternative interventions would make a difference in the face of what often appears to be a growing pressure to make MAID as broadly available as possible. But it is concerning, in my view, that such services are clearly not always readily available; that requesting MAID is increasingly presented, including by MAID assessors and in public discourse, as the appropriate response, as the obvious solution to the anxiety and distress of a devastating diagnosis associated with (future) cognitive decline; and that insufficient efforts are made in the context of some MAID assessments to explore different approaches. It is clear from the reports that some people die by MAID following minimalistic capacity and consent procedures, and in situations where concerns are raised about factors that may affect the voluntariness of those requesting MAID, including care-giver burn-out.

One can respond in two contrasting ways to what has been discussed here: 1) Those who support the broadest possible application for MAID and push for its normalization as a medical practice, may diminish the concerns and may suggest that the cases and evidence analyzed in these reports show that MAID access should be made easier, and MAID options expanded. 2) In contrast, those who are concerned about MAID, or take the position that MAID should be treated as a last resort intervention aimed at facilitating the dying process in situations where there are no other reasonable forms of care, will see in these cases a confirmation of troubling aspects of our current MAID practice that require adjustment and regulatory intervention.

Some of what has been discussed here in this paper in relation to the current Canadian practice of MAID for dementia is also directly relevant for the debate about MAID on the basis of advance requests: how can we confidently conclude that people are suffering intolerably when they are no longer capable of communicating their current perception of their lived experiences? Who will decide when the time has come? On what basis will these judgments be made? How can we expect physicians or nurse practitioners to end the life of a person who has no clue as to why they are being sedated or getting a needle inserted in their arm?

Those who are concerned, as I am, about the significant expansion of the MAID practice, and the growing evidence of the at times clearly problematic practices, will find in these MDRC further confirmation that our MAID regime has serious problems. The reports provide further evidence of the use of low standards of capacity and informed consent; of MAID driven at times by structural and contextual pressures (25); of the at times remarkable drive to offer MAID to people who are suicidal and struggle with mental health issues; and of CAMAP guidance that straddles the contours of the law.

The response to these reports should not be to further expand MAID. After all, if we already have such questionable practices in the context of our current legislative regime, one can only imagine what that will mean if MAID is further expanded, in particular with MAID for sole reasons of mental illness and on the basis of advance requests. In both these contexts, the discretionary power of practitioners will be so much more substantial than what the MDRC reports reveal it already is: in the mental health context because of the vagueness of psychiatric diagnoses and the impossibility to reliably predict the irremediability of mental illness (26); in the context of advance requests because of the need to interpret patients’ prior instructions, without option to communicate and obtain clarifications from them, and without being able to determine reliably, at the time of the MAID provision, whether people are suffering intolerably. The fact that much of the practice is dominated by a small number of frequent providers adds to these concerns (20). Indeed, the greater the discretionary power of MAID assessors and providers, the more likely it is that idiosyncratic practices of some will result in the broad application of MAID in contexts where many others would take a much more cautious approach. The report on same day or next day provision of MAID indicates, for example, how the practice is disproportionally present in some geographical areas of Ontario (11). Whether people will be steered towards further care and support options or rather towards MAID, would in a further expanded MAID regime even more depend on the clinician or MAID provider they encounter.

In my view, the most appropriate answer to the problems raised here is to rethink our MAID regime. There is a need to tighten the legislative access criteria, and to impose a stronger independent review of MAID requests. At a minimum, more rigorous screening and capacity procedures should be introduced to remove the overly broad discretionary power of MAID assessors and providers. In the interim, more detailed professional guidance must be developed. Stronger monitoring and scrutiny of the practice is also needed. It is troubling that documented problematic applications of MAID have not yet resulted in either criminal or professional regulatory intervention. In matters of life and death, failure to exercise utmost caution and to provide the most rigorous oversight and enforcement of criminal law-based standards are unacceptable.