Mental Illness Is Already Inside MAID

The committee this article addresses was reconvened on March 10, 2026, for a specific purpose: to examine readiness before the automatic expansion already encoded in the Criminal Code. Under Bill C-62, which received Royal Assent on February 29, 2024, the exclusion of mental illness as a sole underlying medical condition expires automatically on March 17, 2027. The exclusion does not require Parliament to act to end it. It ends by operation of law unless Parliament introduces and passes new legislation before that date. This committee's report will either support legislation to prevent that automatic expansion — or it will not. This article argues the constitutional case for the former. The debate before this committee is therefore not hypothetical. It is not about whether to create a new pathway. The pathway is already legislated. It activates automatically in less than a year. The constitutional argument is urgent.

Mental illness has been inside Canada's assisted dying regime since its creation — in both tracks. Under Track 1, where a clinician assesses that natural death has become reasonably foreseeable, people with mental illness qualify where that illness is contributing to a physical decline a clinician can characterize as placing death within a reasonably foreseeable horizon: severe treatment-resistant anorexia producing organ failure; serious mental illness driving a person to refuse life-sustaining treatment. Track 1's "reasonably foreseeable" standard is assessed subjectively, requires no terminal diagnosis, no fixed prognosis period, and no objective medical anchoring to imminent death — its constitutional questions are different from Track 2's and are addressed in the next article in this series. Track 1 has reached mental illness from the beginning. The founding case for Track 2 confirms what the regime has never been required to address. The government's own expert witness in the Lamb litigation, Dr. Madeline Li, identified in her expert submission a hypothetical eligibility pathway for Julia Lamb under the original Track 1 framework: if Lamb were to indicate an intent to stop using her BiPaP ventilator at night and refuse treatment for the chest infection that would inevitably follow, she would likely meet the reasonably foreseeable threshold — and she would not need to actually refuse the ventilator, only express a certain intent to do so. The case was adjourned. The constitutional challenge was never decided on its merits. Two things follow from that record. Track 2 was not constitutionally necessary to reach her — Track 1 already did, through a pathway the government's own expert identified. And the pathway that reached her was built around a hypothetical device refusal she would not need to act on — a standard whose adequacy as a constitutional matter was never tested because the case never proceeded. The constitutional necessity argument for Track 2's existence has never held.

Under Track 2, where natural death is not reasonably foreseeable, people with mental illness have qualified since 2021 wherever a physical co-condition is present alongside it: treatment-resistant depression with chronic pain qualifies; severe anxiety with a degenerative neurological condition qualifies; psychological suffering is cited as a driver across both tracks in the government's own data. Mental illness is not excluded from Track 2. A specific subset is — those whose mental illness stands entirely alone, without a physical co-condition present, and without a foreseeable death. The MI-SUMC exclusion holds back precisely the population for whom the required clinical determinations are most difficult to make — where there is no physical anchor for the irremediability assessment, where the condition generating the request is the same condition generating the subjective experience of intolerability, and where the medical literature confirms the mental/physical divide that defines the category does not reflect biological reality.

Professor Daphne Gilbert's April 21, 2026 testimony before the AMAD Special Joint Committee therefore advances a very specific equality claim: not that disabled people are excluded from MAID — they are not, in either track — but that one subgroup of disabled people is being denied the death eligibility already granted to others. Her argument assumes, without establishing, that qualifying for assisted death is the superior outcome — that the state's lethal pathway is a benefit worth demanding under section 15 of the Charter.

Gilbert opened her testimony by noting that this committee's mandate is limited to MI-SUMC and not to all of Track 2. That framing is worth examining. A committee cannot constitutionally recommend expanding a regime without first establishing that the regime itself is constitutional. Gilbert invoked section 15 of the Charter — Andrews, Fraser, Withler, the full substantive equality framework — to argue for expansion. Once that framework is in the room, its application cannot be limited by the committee's terms of reference. Constitutional tools do not stop at the boundary of the mandate that summoned them. Applied honestly and completely to the structure Gilbert came to defend rather than the exclusion she came to challenge, her own framework generates a different constitutional conclusion. Section 15 asks whether a distinction reinforces the historical disadvantage of the protected class. The historical disadvantage disabled people have suffered is not exclusion from death pathways — it is the encoding in law and medicine of the assumption that disability-linked suffering warrants a lethal response rather than the support, accommodation, and inclusion section 15 obliges the state to provide. Track 2 does not counter that historical pattern. It operationalizes it. The gateway is disability. The subjective intolerability standard — applied nowhere else in health or social policy — is applied here, for this outcome alone. The dominant documented drivers of Track 2 requests are the socially-produced conditions section 15 requires the state to remedy. And the mental/physical boundary that defines who is inside the regime and who is excluded — the comparator Gilbert's argument depends on — does not reflect biological reality. Fabiano et al. (JCPP Advances, 2026) establishes this with researchers from nine international institutions: comorbidity is the rule, the boundary cannot be reliably located, and the same patient qualifies differently depending on assessment method. Without a stable comparator, Gilbert's discrimination argument loses its clinical foundation. Without a reliable eligibility screen, the regime cannot satisfy Bedford arbitrariness. Without a constitutional obligation to create the pathway, Parliament is free to end it. Gilbert's framework, applied without the constraint she placed on it, does not justify expansion. It establishes that Track 2 cannot presently satisfy the constitutional obligations section 15 imposes. That is the prior question the committee must answer before it can responsibly address the question it was given.

This is not a conventional disability rights claim. Follow Gilbert's argument to its logical conclusion and it says something more stark: no disability should be a barrier to accessing the death pathway. Mental illness is a disability. Physical conditions are disabilities. If any distinction between disability types for death eligibility violates section 15, then the section 15 argument does not stop at MI-SUMC. It demands equal access to death across all disability categories — or it demands that Track 2 not exist at all. Those are the only two positions the framework can sustain. Gilbert chose the first. The disability rights movement, the UN CRPD Committee, and the clinical record all point to the second. The limiting constraint Gilbert placed on her own argument — stopping at MI-SUMC — is not supported by the framework she built. Applied fully and honestly, it does not justify expansion. It condemns the structure she is defending.

Carter did not constitutionalize Track 2. The Supreme Court did not rule that the Charter requires MAID where natural death is not reasonably foreseeable. Track 2 was created by Parliament after an unappealed Quebec Superior Court decision. Parliament made a policy choice. Parliament can reverse it.

The criminal law makes the stakes precise. Under section 241.2(2) of the Criminal Code, a person has a grievous and irremediable medical condition only if they have a serious and incurable illness, disease or disability; are in an advanced state of irreversible decline in capability; and that illness, disease or disability causes them enduring physical or psychological suffering that is intolerable to them and that cannot be relieved under conditions they consider acceptable. The strongest case for expansion — a competent adult with a grievous and irremediable mental illness who meets every other Track 2 criterion — collapses at the irremediability gate. That gate is not a procedural checkbox. It is the constitutional linchpin. And the clinical record before this committee establishes that clinicians cannot presently make these determinations reliably — not only where mental illness is the sole condition, but across all of Track 2, including where physical, neurological, and neurodevelopmental conditions are present alongside mental illness. The Fabiano finding confirms why: the mental/physical boundary the gate depends on does not reflect biological reality. Comorbidity is the rule. The same patient may qualify differently depending on assessment method. A competent adult who genuinely meets every criterion is a category the evidence shows cannot currently be reliably identified — separated from the person in a treatable crisis, the person failed by inadequate care, the person whose suffering is produced by conditions the state could remedy. The gate is not working across any track. The constitutional problem is not confined to the MI-SUMC boundary.

Section 15 Requires Substantive Equality — Not Identical Treatment

Gilbert's argument rests on a misstatement of what section 15 actually requires. She argues that any distinction between mental illness and physical conditions for MAID eligibility is inherently stigmatizing and therefore unconstitutional. That is formal equality — identical treatment regardless of context or consequence. It is not the standard section 15 imposes.

Section 15 requires substantive equality — that distinctions not reinforce, perpetuate, or exacerbate the disadvantage of historically marginalized groups. This was confirmed in Fraser v. Canada (SCC, 2020), Withler v. Canada (SCC, 2011), and Andrews v. Law Society of BC (SCC, 1989). The question is never whether two groups are treated differently. The question is whether the distinction compounds historical disadvantage. The protected class exists precisely because disabled persons have historically been devalued, institutionalized, neglected, and pressured toward premature death when society withheld the supports necessary for dignified life. Different treatment is not automatically discrimination. The inquiry is directional — does the distinction harm the group section 15 was built to protect, in the direction that has historically harmed them?

Section 15 does not impose a freestanding duty to eradicate all societal inequalities — Auton v. British Columbia (SCC, 2004) and Sharma v. Canada (SCC, 2022) confirm this limit. But that limit does not help Track 2. Once the state chooses to intervene in the domain of end-of-life decision-making for a disadvantaged group, it must do so in a manner that does not compound their historical disadvantage or encode the stereotype that disabled lives are inherently less worth living. The state has already intervened. It created Track 2. The question is whether it did so in a manner consistent with section 15. The answer requires examining what the regime actually does — not what it intends.

Apply that standard to what Gilbert is actually arguing. She claims the MI-SUMC exclusion reinforces the stigma that mentally ill people's suffering is less real than physical suffering. But that framing assumes the relevant comparison is between disabled people with different diagnoses — some of whom qualify for the death pathway and some of whom do not. Section 15 does not ask whether one disabled subgroup is treated differently from another disabled subgroup. It asks whether the state's treatment of the protected class reinforces the historical disadvantage that class has suffered. The historical disadvantage disabled people have suffered is not exclusion from death pathways. It is the encoding in law and medicine of the assumption that disabled lives are less worth living — that disability-linked suffering is a problem best resolved by ending the life rather than supporting it. Track 2 does not counter that assumption. It operationalizes it. The question substantive equality requires is not whether all disabled subgroups have equal access to the death pathway. It is whether the state can construct a disability-linked death pathway at all without reinforcing the devaluation section 15 was built to prevent.

One clarification is necessary before proceeding. The argument here is not that mental suffering is less real than physical suffering, or that it deserves less serious treatment. It is not. The question is whether state-facilitated death is the form of seriousness that section 15 compels. Those are different questions. Gilbert conflates them. The disability organizations challenging Track 2 in the Ontario Superior Court do not. Recognizing mental illness as real, serious, and deserving of full medical attention is not the same as concluding that the state is constitutionally required to offer a death pathway as the remedy. The equality obligation runs toward care, housing, and support — the conditions that make death unnecessary — not toward equal access to a lethal pathway built on clinical determinations the evidence shows cannot presently be made reliably.

Where the intolerability of suffering cannot be reliably disaggregated from poverty, inadequate housing, isolation, or insufficient disability supports — conditions that are socially produced and remediable — authorizing death on the basis of disability alone treats the protected class as expendable. It substitutes state-facilitated death for the positive measures the state's own equality obligations require. A regime that cannot make that distinction in practice fails to afford the equal benefit and protection of the law section 15 guarantees. Fabiano et al. — a peer-reviewed open-access editorial perspective published in JCPP Advances by researchers from the University of Ottawa, Harvard Medical School, King's College London, and six other international institutions — documents that comorbidity between mental and other medical conditions is the rule rather than the exception, and that all health conditions share common biopsychosocial risk factors and pathophysiological pathways (Fabiano et al., JCPP Advances, 2026, DOI: 10.1002/jcv2.70123). The paper calls explicitly for medicine to retire the mental/physical distinction entirely: "we must retire the notion of a broad separation between 'physical' and mental health." The eligibility gate the regime depends on — sole mental illness versus mental illness with physical comorbidity — is therefore not a boundary the medical literature supports. The paper provides direct empirical grounding for this conclusion: in a study of 1,146 people with major depression across 15 countries, 45–95% presented with bodily symptoms only, while 11% denied any psychological symptoms until structured diagnostic interviews were administered. The same patient who would qualify as MI-SUMC under one assessment may qualify as dual-diagnosis Track 2 under another, depending entirely on assessment method and the clinical tools used. An eligibility gate the clinician cannot reliably locate cannot ground a constitutionally valid determination with a lethal outcome. The paper also documents that the artificial divide between mental and physical health has produced a 15-year life expectancy gap between people with mental health conditions and the general population — a gap exacerbated among marginalized groups, produced by disproportionate underfunding, longer wait times, and unequal treatment pathways. That is not a background fact. It is the documented social consequence of the same categorical thinking the regime's eligibility structure perpetuates — and it is the context in which the committee is being asked to treat that suffering as irremediable.

Gilbert assumes the constitutional harm is exclusion from a death-eligible category. But section 15 does not ask only whether a distinction exists. It asks whether the distinction reinforces disadvantage in the direction substantive equality was designed to prevent. The disability rights objection — raised consistently by every major disability organization in Canada and echoed by the UN CRPD Committee — is that inclusion within a disability-linked death category is itself the disadvantage. The state is not excluding disabled people from a universal public benefit available equally to everyone else. Non-disabled people are not eligible for Track 2 at all. The regime instead constructs disability, decline, and medically qualified suffering as the gateway conditions for state-facilitated death. Gilbert's argument does not challenge that structure. It accepts the structure and argues that more disabled subgroups must be included within it equally. That is not equality between disabled and non-disabled people. It is equalization inside a disability-linked death framework.

The Autonomy Objection — And Why the Clinical Record Answers It

The autonomy objection to this analysis must be addressed directly. It runs as follows: Carter affirmed that competent adults in irremediable suffering have a constitutionally protected interest in assisted dying. Preventing a competent person from accessing that pathway on the basis of their diagnosis is paternalistic. Clinical assessments are imperfect in every domain of medicine and courts do not use imperfection to deny rights. The clinical screen argument proves too much — it would justify denying MAID to virtually anyone.

That objection is serious. It deserves a direct answer rather than an inference buried in other paragraphs. The answer is not that autonomy is irrelevant. It is that Carter requires courts to tolerate ordinary clinical uncertainty — the same uncertainty present in every prognosis, every capacity assessment, every end-of-life decision. That toleration has a floor. When the clinical bodies that support expansion have themselves placed on the parliamentary record that the foundational determination is accurate at chance level — the best available predictive model correctly identifying chronic treatment-resistant depression after two years in only 47% of cases — the court is no longer being asked to tolerate uncertainty. It is being asked to authorize lethal outcomes under conditions the supporting clinicians themselves describe as worse than random. That is not ordinary clinical uncertainty. It is the kind of admitted systemic failure that crosses the threshold where section 7 — the right to life — requires more than a good-faith procedural screen before the state extinguishes it. The distinction is not between imperfect and perfect clinical judgment. It is between imperfect clinical judgment — which courts tolerate everywhere — and below-chance clinical judgment — which the constitutional framework cannot absorb when the outcome is irreversible and lethal.

Parliament Is Free to Repeal Track 2. Nothing in the Constitution Requires It.

Track 2 rests on a single unappealed Quebec Superior Court ruling from 2019. The Supreme Court of Canada has never ruled that Track 2 is constitutionally required. Carter v. Canada (SCC, 2015) addressed terminal illness. It left non-terminal cases entirely open. Parliament created Track 2 as a policy choice in response to one provincial judgment. It can repeal it the same way.

A government lawyer will argue that Carter's reasoning — the principle that a blanket prohibition on assisted dying for competent adults in irremediable suffering violates section 7 — extends to mentally ill patients regardless of what the Court explicitly held. That argument confuses the holding with the reasoning. Carter's holding was deliberately narrow: the Court struck a blanket prohibition and explicitly invited Parliament to craft a legislative response. Parliament's legislative response is not constitutionally entrenched. It is a policy choice made under the Court's invitation to legislate. Parliament can reverse a policy choice without disturbing the constitutional principle Carter established. The constitutional principle Carter established — that a competent adult in irremediable suffering has a section 7 interest in assisted dying — does not require any particular eligibility structure. It does not require Track 2 to exist. It does not prevent Parliament from concluding that the clinical tools to implement that principle safely for non-terminal conditions do not currently exist and that the eligibility architecture has produced disproportionate lethal outcomes on a protected class. Repeal of Track 2 is constitutionally available and constitutionally defensible.

Parliament can repeal Track 2 tomorrow without disturbing Carter, without creating a constitutional gap, and without violating any Supreme Court ruling. Nothing in the constitutional framework requires its existence. The committee's recommendation to repeal Track 2 is not a constitutional risk. It is a constitutionally available remedy that the evidence, the clinical record, and the rights framework all point toward.

Apply Gilbert's own section 15 two-step to Track 2 itself. Step one: does it create a distinction based on disability? Yes — unambiguously. No person without a qualifying disability can access Track 2 regardless of capacity, suffering, or autonomous choice. The disability is not incidental to eligibility. It is eligibility. Step two: does that distinction reinforce disadvantage? The UN CRPD Committee answered yes. The major Canadian disability rights organizations answered yes. The mechanism is direct: Track 2 takes shared experiences of disability — loss of autonomy, loss of independence, loss of meaningful activity — and treats them not as grounds for remediation and state support but as grounds for state-facilitated death. It rebrands the predictable social consequences of state failure as a compassionate medical standard. It then applies that standard only to disabled people. That is the definition of encoding disadvantage in law. Section 15 was built to prevent exactly that logic.

The Category Is Defined by Diagnosis. The Clinician Cannot Screen What the Law Requires.

Gilbert says capacity — not diagnosis — should determine who accesses MAID. But the MI-SUMC category she is fighting for is defined entirely by diagnosis. Remove the mental illness and the eligibility disappears. Capacity not diagnosis, she says — while arguing for a category that exists only because of the diagnosis.

The competence argument has a deeper problem she never addresses. If competence rather than diagnosis is the governing standard — if any competent adult experiencing intolerable suffering should have access to Track 2 — then the argument does not stop at mental illness. A competent non-disabled person experiencing intolerable suffering from grief, poverty, social exclusion, or trauma cannot access Track 2. Competence does not get them there. Only a qualifying disability does. Gilbert does not hold the position that any competent person in intolerable suffering should have Track 2 access — she accepts medical gating, she accepts that not everyone qualifies. That internal contradiction reveals that competence is not actually the test she is applying. Disability is. She is arguing for competent disabled people to have equal access to death while accepting that equally competent non-disabled people in equivalent suffering do not have that access. That is not a competence argument. It is a disability argument. And when you apply the disability argument correctly — asking whether constructing a lethal pathway specifically for disabled people reinforces the historical assumption that disability-linked suffering warrants a lethal state response unavailable to everyone else — it does not justify expansion. It identifies the discrimination running in the other direction.

Gilbert invokes Bell Let's Talk to argue the law infantilizes people with mental illness. The reference is worth examining directly. Bell Let's Talk exists because mental illness is intersectional and situational — concentrated in populations that already experience structural disadvantage, produced by and exacerbated through stigma, unemployment, housing insecurity, victimization, and social exclusion. The movement was built on the documented reality that mentally ill people are disproportionately victims of crime and power imbalances, not because of any inherent characteristic but because of the structural conditions the world imposes. Many people the mental health awareness movement was built to serve are autistic, neurodivergent, disabled — the same people the disability organizations challenging Track 2 represent. The lesson Bell Let's Talk teaches is investment in care, inclusion, and support — not death access. Gilbert takes the language of that movement and redirects it toward a death pathway. The organized disability community — which includes many people with mental illness — has said consistently that Track 2 is the stigma, not the solution to it. The suffering Bell Let's Talk acknowledges is real. The response that movement demands is care. Gilbert's framework converts that demand for care into a demand for death eligibility and calls it equality. That substitution is precisely what section 15's substantive equality analysis requires the court to examine — not whether Gilbert's argument sounds like a rights claim, but what it does structurally to the people whose rights are at stake.

That contradiction runs deeper than a terminological inconsistency. Gilbert's section 15 argument requires MAID to be rights-based enough that the equality guarantee governs access to it — otherwise her constitutional claim has no jurisdiction. But she also accepts that MAID is medically gated, that safeguards are legitimate, and that not everyone who suffers qualifies. Those two positions are in unresolved tension. A purely rights-based regime would extend to any competent person in intolerable suffering regardless of diagnosis — a position Gilbert does not hold and the law does not support. A purely medical regime makes protected class membership irrelevant to eligibility, and her section 15 argument evaporates. Her position requires the framework to operate in both registers simultaneously without ever resolving how.

The deeper structural problem runs beneath this contradiction. The entire Track 2 regime — including the expansion Gilbert is arguing for — is built on one premise: that people with protected disability status should have equal access to a death pathway that exists specifically because of that status. It is not an offer of access to a universal benefit. Non-disabled people cannot access Track 2 regardless of how much they suffer. The protected characteristic is not the exception to eligibility. It is the qualifying condition for it. Gilbert's equality argument does not challenge that structure. It demands that more subgroups of disabled people be included within it equally. That is not equality between disabled and non-disabled people. It is the equalization of death access inside a regime whose existence depends on singling out disability as the gate.

Gilbert invokes the law's subjective suffering standard — that suffering need only be intolerable to the person requesting — as grounds for a consent-based rather than diagnosis-based eligibility gate. But that standard cannot do the work she assigns it when the condition producing the request is the same condition producing the subjective experience of intolerability. The illness shapes what the person experiences as unbearable. The clinician's task is not simply to accept that experience as reported. It is to assess whether it reflects settled autonomous choice or the illness speaking — whether it is irremediable or produced by inadequate care and social abandonment — whether it is free from the structural coercion of a person failed by every system around them. The law's subjective standard does not eliminate those assessments. It makes them harder. And the clinical record establishes they cannot currently be made reliably.

Gilbert's iceberg analogy has two halves. The tip — the small number who would ultimately qualify for MI-SUMC — she invokes to dismiss concerns about volume. But a small eligible population weakens, not strengthens, constitutional necessity. The narrower the excluded class becomes clinically, the harder it is to establish that the exclusion produces a systemic Charter violation rather than an individual hardship requiring an individual remedy. The submerged mass — the thousands allegedly harmed by the language of exclusion — is where her argument does more work. But that harm depends on the premise that mentally ill people are meaningfully excluded from MAID. They are not. Depression alongside chronic pain qualifies. Anxiety alongside degenerative illness qualifies. Psychological suffering is cited across both tracks in the government's own data. If the submerged mass is already substantially reached by the regime, the symbolic harm of the MI-SUMC exclusion cannot be doing what Gilbert claims. The iceberg analogy cuts both ways — and the second cut is deeper. Gilbert's section 1 argument compounds the problem. She argues the exclusion has no clear readiness benchmark and therefore cannot be justified. She is correct that no clear benchmark exists. But the absence of a clear benchmark is not a reason to lift the exclusion — it is a reason to question the entire architecture. If readiness has no defined standard, then the expansion position has no basis for claiming readiness has been achieved either. The absence of a benchmark is the operational impossibility argument in legislative form. Parliament acknowledged it twice in statute. The answer is not to proceed without the benchmark. It is to address why the benchmark cannot be established — which is what the clinical record before this committee explains.

The clinician assessing any Track 2 request faces three determinations the clinical record confirms cannot presently be made reliably — and whose unreliability makes the regime constitutionally indefensible at this gate regardless of what the law requires procedurally. These failures are most acute where mental illness is the sole condition — where there is no physical anchor for the irremediability assessment and where the condition generating the request is the same condition generating the subjective experience of intolerability. But the MDRC's documented failures, the social drivers data, and the Fabiano finding apply across all of Track 2. The constitutional problem is not confined to the MI-SUMC exclusion. It runs through the regime's eligibility architecture. Expansion advocates will point to the Canadian Psychiatric Association's March 2026 guidance as evidence that readiness is approaching. That guidance is examined in detail below. It does not resolve the problem — Crawford assessed it directly before this committee and found it aspirational, with no operational framework, no clear criteria, and no structured tools. The guidance names the task. It does not provide the instrument the Constitution requires before a lethal outcome is authorized.

Before addressing the three determinations, two structural points require explicit statement. First, this is not a question courts must resolve by choosing between competing philosophical frameworks. Courts are not being asked to decide whether MAID is moral, whether autonomy matters, or whether suffering is real. They are being asked to apply what they have already held — that laws whose effects are inconsistent with their stated objectives are arbitrary, and that irreversible state action on a protected class requires the state to demonstrate its safeguards can perform the determinations they claim to perform. That is a question of operational capacity, not moral philosophy. The evidence answers it. Second, the constitutional architecture question is not the same as the individual hard case question. Expansion advocates will invoke individual people whose suffering is genuine, long-standing, and unresponsive to every available intervention. This article does not deny those people exist. It argues the clinical record before this committee establishes that no validated tool currently exists to identify who they are — to separate them, before an irreversible outcome, from the person in a treatable crisis, the person in structural coercion, the person whose suffering is produced by inadequate care the state could remedy. The constitutional problem is not whether genuine cases exist. It is whether the state can presently find them without also approving deaths it cannot constitutionally justify. But there is a second and independent constitutional problem that survives even if the screen improves. A regime that makes disability the qualifying criterion for state-administered death — that offers a lethal pathway exclusively to disabled people, unavailable to the general population regardless of suffering or capacity — encodes in law the assumption that disability-linked suffering warrants a lethal state response that equivalent suffering in non-disabled people does not. That assumption is precisely what section 15 was built to prevent the state from encoding. The operational argument is sufficient for today's recommendation. The structural argument explains why better tools alone will not resolve the constitutional problem.

First: is this request driven by the illness or by settled autonomous choice? Suicidal ideation is a diagnostic criterion of many conditions MI-SUMC would cover. The thing being assessed as an autonomous choice is also a symptom of the condition generating it. The clinician must distinguish a desire for death that is the illness speaking from one that is the person speaking. No validated tool for that determination presently exists. The heads of psychiatry at thirteen Canadian medical schools said so to this committee directly. The Society of Canadian Psychiatry's Brief on MAID and Mental Illness (October 13, 2023, p. 4) relied on Nicolini et al.'s peer-reviewed analysis of precision psychiatry evidence (Psychological Medicine, 2023, DOI: 10.1017/S0033291722002951), which identified Dinga et al.'s machine-learning model of naturalistic depression course as the closest available proxy for real-world long-term prediction. That model correctly predicted chronic treatment-resistant depression after two years in only 47% of cases — at chance level (Dinga et al., Translational Psychiatry, 2018, DOI: 10.1038/s41398-018-0289-1). The SCP — an organization ideologically supportive of expansion — relied on this evidence to acknowledge that meaningful irremediability determinations cannot currently be made. That acknowledgment is not peripheral. It is the clinical foundation of the regime, assessed by its own supporters, and found to be absent.

The assessment problem is deeper than the absence of validated tools. Across multiple psychiatric conditions that MI-SUMC would reach, the desire to die can be produced as a symptom that is experienced by the person as fully rational, internally coherent, and consistent with their settled sense of self. It does not present to the person as an intrusive symptom they want treated. It presents as the most logical conclusion available given their circumstances. Standard capacity assessment asks whether the person understands the relevant information, can appreciate how it applies to their situation, can reason about options, and can communicate a choice. A person experiencing illness-generated suicidal desire in this form can satisfy all four criteria. Their reasoning is internally coherent. Their communication is clear. Their account of their situation is consistent. They pass every formal capacity test the clinical framework knows how to administer. What standard capacity assessment cannot detect is whether the preference being expressed is state-dependent — whether it is a product of a crisis state the illness is generating that would resolve if the crisis resolved, rather than a preference that would persist across clinical states. No validated instrument for that determination presently exists. The CPA guidance Crawford found aspirational before this committee does not provide one. Crawford's PROSPERO-registered systematic review found none in the peer-reviewed literature. The clinical tool that would need to exist — one capable of distinguishing illness-generated state-dependent desire to die from settled preference that survives across clinical states — is precisely the tool the literature confirms is absent.

This matters for Gilbert's argument specifically. Carter v. Canada authorized assisted dying for competent adults making settled autonomous choices about their deaths — choices free from the distorting effects of illness. A desire to die that is generated by the illness as a symptom, experienced as rational within the cognitive and emotional state the illness has created, and that would not persist if that state resolved, is not a Carter-competent autonomous choice. It is the illness expressing itself through the person. Approving death on the basis of that desire is not respecting autonomy. It is acting on a symptom irreversibly — and calling that compassion. Gilbert's section 15 argument requires the request to be a competent autonomous choice in the Carter sense. Standard capacity assessment cannot presently establish that where the desire may be state-dependent. Her equality argument depends on a clinical finding the available tools cannot make. She is arguing for equal access to a pathway whose constitutional validity depends on a determination that cannot presently be verified.

Second: is this suffering irremediable, or is it produced by inadequate care, poverty, isolation, the absence of supports the state has failed to provide? The government's own 2024 data shows 44.7% of Track 2 recipients cite isolation and loneliness as a source of suffering. For mental illness as a sole condition — where suffering is inseparable from the social context in which the person lives — that determination cannot be made reliably. Fabiano et al. (JCPP Advances, 2026, DOI: 10.1002/jcv2.70123) — researchers from the University of Ottawa, Harvard Medical School, King's College London, and six other international institutions — confirms the mental/physical divide is clinically artificial: comorbidity between mental and other medical conditions is the rule rather than the exception, shared biological pathways mean the conditions cannot be cleanly separated for clinical assessment purposes, and the same patient may qualify as MI-SUMC or dual-diagnosis Track 2 depending entirely on assessment method. The MI-SUMC category is therefore built on a clinical distinction the medical literature calls a monumental mistake in medicine that must be retired entirely.

The irremediability determination faces a further problem that prediction accuracy alone does not capture: treatment options for conditions the regime must assess as irremediable continue to emerge after assessments are made. In January 2025, the FDA approved esketamine nasal spray (Spravato, Johnson & Johnson) as the first approved monotherapy for treatment-resistant depression — defined as inadequate response to at least two oral antidepressants — following clinical trials showing a 54% increased likelihood of remission at eight weeks compared to existing alternatives (FDA supplemental NDA approval, January 21, 2025; J&J press release; AJMC, January 2025). A clinician who declared a patient's depression irremediable before 2019 — when esketamine was first approved as an adjunct therapy — was assessing irremediability without that option available. A clinician who declared it irremediable before January 2025 was doing so before esketamine existed as an approved monotherapy. The regime has no mechanism for reassessing irremediability as new treatments emerge. A death approved on the basis of a pre-2025 irremediability determination was authorized using a standard the treatment landscape has since materially changed. That is not a peripheral concern about clinical uncertainty. It is structural. Irremediability — the statutory predicate the regime depends on — is assessed at a point in time against treatments available at that time. It cannot account for what will exist next year. The expansion advocates' answer is that assessors should incorporate emerging evidence. But the regime provides no mechanism for doing so, no requirement to revisit prior determinations, and no way to correct a death authorized on the basis of a standard that has since been overtaken by new evidence. The outcome is permanent. The knowledge base is not.

Third: is this request free of coercion — not obvious coercion, but the structural coercion of a person failed by every system around them who has run out of other options? A person warehoused in inadequate care, denied supports, told their condition is treatment-resistant — their request may feel entirely voluntary and still be shaped by conditions that, if remedied, would change everything. The clinician cannot reliably assess this either.

The structural coercion argument is not hypothetical. Roger Foley has spinocerebellar ataxia, a severe neurodegenerative disease. He testified to the House of Commons Standing Committee on Justice and Human Rights on November 10, 2020, from his hospital bed in London, Ontario. His testimony: "I have been coerced into assisted death by abuse, neglect, lack of care and threats. At a time when I was advocating for assistance to live and for self-directed home care, the hospital ethicist and nurses were trying to coerce me into an assisted death by threatening to charge me $1,800 per day or force-discharge me without the care I needed to live. I felt pressured by these staff raising assisted dying rather than relieving my suffering with dignified and compassionate care." He recorded those conversations. The recordings were reported by CBC and CTV News. He told MPs directly: "Assisted dying is easier to access than safe and appropriate disability supports to live." He has been offered MAID on four occasions — not because he requested it, but because the system offering death found it easier than providing the care he needed to remain alive. He has filed lawsuits against the hospital, caregiver organizations, and provincial and federal governments alleging illegal denial of care. His case is not an outlier. It is the structural coercion argument made audible — recorded, reported, and entered into the parliamentary record of the same legislative process that created Track 2.

Gilbert argues that questioning these things is itself discriminatory. That requiring additional scrutiny of MAID requests where mental illness is the sole condition treats mentally ill people as less trustworthy. But the scrutiny is not about doubting the person. It is about the clinician's inability to make the determinations the law requires. The person is not the problem. The assessment tool is. A broken screen does not honour autonomy. It exposes people with mental illness to a system that will approve their deaths on the basis of determinations it cannot presently make — and call that their liberation.

Expansion advocates will point to the Canadian Psychiatric Association's March 2026 guidance as evidence that the clinical gap is being addressed and the regime is approaching readiness. Crawford's testimony closes that argument directly on the parliamentary record. She assessed the CPA guidance and found it "largely aspirational": it asks assessors to differentiate acute suicidal crisis from a well-considered wish to die, but offers no operational framework, no clear criteria, and no specification of structured tools. It also fails to incorporate a standardized equity assessment to determine when social determinants and remediable needs are driving suffering. Guidance that names the task without providing the tool does not constitute clinical readiness. It describes the problem in aspirational language and leaves the clinician without the evidence-based procedures the Constitution requires before a lethal outcome is approved. Crawford's systematic review of peer-reviewed literature — registered on PROSPERO (CRD420251121140) — independently confirms what the guidance's absence of tools reflects: no credible evidence that suicidal intent can be reliably distinguished from a MAID request where mental disorder is the sole condition; no validated assessment instruments capable of making that discrimination. The guidance and the systematic review are looking at the same clinical gap from two directions. Both find it open.

The State Cannot Protect and Facilitate Simultaneously

Canada has an entire legal and clinical infrastructure built on one premise: when someone expresses a desire to die, the state intervenes. That is why crisis lines exist. That is why involuntary holds exist. That is why the Mental Health Act in every province authorizes detention to prevent suicide. The desire to die is not unique to mental illness — the CDC's 2018 study across 27 states found 54% of suicide decedents had no known psychiatric condition. The state's crisis infrastructure responds to expressed desire to die across the population, regardless of diagnosis, because the empirical record on survivor outcomes confirms that suicidal crises are frequently temporary and that most people who survive them are glad they did.

MI-SUMC says something different. It says that for some people — specifically those with a qualifying disability — the same expressed desire to die should be facilitated rather than interrupted. The state has not demonstrated a reliable operational mechanism for distinguishing the contexts in which it intervenes from those in which it facilitates. The clinical record before this committee establishes that no validated tool presently exists to make that distinction reliably — and that the two responses are therefore being applied to the same population on the basis of disability status rather than on any validated clinical determination of which desire to die warrants which response.

The evidence base makes this a constitutional argument, not merely a values one. Evidence shows that people with mental illness who express suicidal ideation are statistically more likely to act on it when means are available and legitimized. Offering a legal, state-sanctioned death pathway to people whose illness produces suicidal ideation as a symptom — and simultaneously running suicide prevention infrastructure for those same people — disproportionately exposes that population to a lethal outcome driven by the illness itself. That is not equal treatment. That is disparate impact on a protected class. It discriminates against people with mental illness by offering them the thing their illness makes them most vulnerable to — and calling that respect for their autonomy.

This is not a theoretical concern. The Chief Medical Officer of Canada's 9-8-8 Suicide Crisis Helpline — Dr. Allison Crawford — brought operational data from inside the state's own crisis infrastructure to this committee. Up to 7% of interactions on 9-8-8 mention MAID. Among those callers, 74% endorsed suicidal thoughts in the past two days, compared with 48% among other contacts. The people reaching Canada's suicide prevention service who mention MAID are in acute suicidal distress at nearly double the rate of other callers. Crawford's description: "a real, measurable, elevated risk." This is not modeled. It is not surveyed. It is call log data from inside the system the state runs to prevent suicide — documenting that the people most proximate to the MAID pathway are simultaneously in acute suicidal crisis. The two arms of the state are not operating on separate populations. They are operating on the same people, simultaneously, with contradictory responses to the same expressed desire to die.

Crawford is also the Principal Applicant on a CIHR-funded study — government-funded research — examining MAID and suicide in the crisis line context. The government is funding research to resolve the screen-failure between suicidal ideation and autonomous MAID requests. It is simultaneously being asked to expand the regime that depends on that screen working. It has not waited for the research to conclude. It has not waited for the tools the research is designed to develop. The state has been repeatedly placed on notice — through its own commissioned research, through Crawford's submission to this committee, through the heads of psychiatry letter, through the CIHR funding record itself — that the screen cannot presently perform the determinations it claims to perform. It has funded science to try to build one. And it is proceeding to expand the regime before the science is done. That is not evidence-based governance. It is the use of a research program to create the appearance of rigor while expanding a regime under conditions of admitted non-knowledge.

The research literature on suicidal ideation and recovery compounds the constitutional problem. The vast majority of people who express strong suicidal ideation — including those who attempt suicide — do not subsequently die by suicide and report being grateful to have survived. Seiden's landmark 1978 study of 515 people who survived interrupted Golden Gate Bridge attempts found 90% did not subsequently die by suicide. Kevin Hines, who survived a jump from the same bridge, described the instant he let go of the rail: "I had an instantaneous regret." 19 of 39 documented survivors reported the same instant regret. Mishara, with 50 years of suicide prevention research: "Even in extreme cases where a person is taken to hospital unwillingly, only 10% will attempt again, and only 1–3% will die. The vast majority are happy to have been saved and are usually very thankful to be alive." The belief that a mentally ill person's current desire to die is permanent — that it accurately predicts their future state — is contradicted by the research literature across every study that has examined it. The regime acts on that belief anyway. And the outcome it produces on the basis of that unverified belief cannot be undone.

The deeper constitutional problem is that the regime does not distinguish between different kinds of desire-to-die. It distinguishes between different kinds of people expressing it. Despair, hopelessness, loss of dignity, loss of future orientation, the desire for death — these are the same phenomenology regardless of who experiences them. The regime classifies them as pathology requiring intervention in non-disabled people, and as potentially rational autonomous choice in disabled people. That classification is not grounded in clinical science. It is applied on the basis of protected status. A classification that treats the same human experience as illness in one population and autonomy in another, depending on disability status, is not a clinical distinction. It is a categorical value judgment encoded in law.

Psychiatric literature makes this precise through the concept of egosyntonic suicidality — suicidal desire experienced not as an unwanted alien symptom the person wants treated, but as rational, self-consistent, and an expression of who the person is. Egodystonic suicidality, by contrast, is experienced as unwanted: the person recognizes the desire as illness and wants it gone. Crisis intervention is premised on recognizing egodystonic suicidality and treating it. The MI-SUMC framework must — to justify itself — identify egosyntonic suicidality as the marker of autonomous choice rather than illness. But no validated clinical tool currently exists to make that distinction reliably. The clinical screen that is supposed to separate settled autonomous desire-to-die from illness-driven desire-to-die does not exist. That is not a technical gap the regime can paper over with procedural safeguards. It is the screen the regime's constitutional validity depends on. Without it, the state cannot tell the person whose desire for death reflects genuine settled autonomous choice from the person whose illness is speaking. It will approve both. The irreversible outcome will follow both. And it will call that equality.

Gilbert argues the exclusion sends a message that mentally ill people must be protected from themselves. She is right that paternalism is a constitutional problem. But MI-SUMC expansion produces a different structural effect: it creates a legal death pathway specifically accessible to people whose illness generates the desire to use it. That is not anti-paternalism. That is the most permanent paternalism imaginable — delivered with a clinician's signature and a Charter citation.

There is a further equality argument that the article's preceding analysis generates and that has not yet been named in the constitutional literature on this regime. Section 15 jurisprudence typically concerns access — the removal of barriers, the provision of accommodation, inclusion in benefits available to others. What this regime does is structurally different. It converts unresolved positive obligations into eligibility criteria for an irreversible state response. The unmet housing, the inadequate care, the isolation, the poverty, the institutionalization, the social abandonment — these are conditions section 15 obligates the state to address. They are also, by the government's own 2024 data, the dominant drivers of MAID requests. The state has not addressed those conditions. It has made them the pathway to an eligibility determination. Track 2 does not remedy the positive obligations section 15 requires. It converts the failure to meet them into the conditions of death eligibility. That is not equality. It is the appropriation of the state's own unmet obligations as the mechanism for their permanent resolution.

The asymmetry at the heart of this regime is most visible when you ask a simple question: in what other domain does the state accept a disabled person's subjective experience of intolerability as the governing standard for state response? The answer is none. A disabled person who finds their housing intolerable waits on a list and meets objective criteria. A disabled person who finds their level of support intolerable navigates assessment frameworks and resource constraints. A disabled person who finds their mental health treatment intolerable encounters clinical gatekeeping, waitlists, and institutional processes. In every domain where the state's response would cost it something — money, resources, political will — the state applies objective criteria, clinical assessment, and institutional gatekeeping. The subjective experience of the person is relevant but not determinative. The state does not simply accept "I find this intolerable" as sufficient grounds for any particular response. Except one. For the death decision alone — the single irreversible outcome, the one that cannot be reviewed, appealed, or corrected — the law says suffering need only be intolerable to the person. Not objectively verified. Not assessed against a population baseline. Not subject to the institutional gatekeeping the state applies to every other resource it allocates to disabled people. Subjective intolerability, reported by the person, is sufficient. The eligibility architecture is structured so that the state's most immediate and unmediated response to a disabled person's subjective experience of intolerability is available for one outcome only: death. For everything that would require the state to provide housing, care, and supports — the conditions that would make death unnecessary — the structure interposes objective criteria, assessment processes, resource constraints, and waitlists. That structural asymmetry is not incidental to the regime. It is encoded in it. And under substantive equality analysis, what matters is not what Parliament intended but what the structure produces: a regime whose eligibility architecture is maximally responsive to disabled people's expressed suffering at precisely the point where that responsiveness is lethal — and minimally responsive at every point where it would require the state to help them live.

The Clinical Record Closes the Constitutional Door

Under Bedford v. Canada (SCC, 2013), a law is arbitrary — and therefore violates section 7 — when its effects are inconsistent with its stated objective. Track 2's objective is the relief of irremediable suffering. If irremediability cannot be reliably assessed, the law's effect — an irreversible death — is not consistent with that objective. The safeguards do not fix this. They describe the procedure around a determination that does not work.

The clinical record before this committee converges across lines of ideology. The organizations that support MI-SUMC expansion in principle and those that oppose it share a single clinical finding: reliable irremediability assessment for mental illness as a sole condition does not currently exist.

The Ontario MAID Death Review Committee — the state's own oversight body — has documented what the regime produces when operating as designed. From 2018 to 2023, the Ontario coroner tracked 428 cases of possible criminal violations in MAID: four were reported to the regulatory college, none was referred to law enforcement. In 2023 alone, 178 noncompliance cases were identified. A quarter of all Ontario MAID providers were flagged by the coroner at least once. In at least one reviewed case, the MDRC concluded the person appeared to have sole mental illness and should not have been eligible — but had been approved anyway. MDRC reports confirmed that some Track 2 deaths were driven by unmet social needs including fear of homelessness. Dr. Trudo Lemmens — a member of the MDRC — documented in peer-reviewed analysis that capacity evaluations were minimalistic and that current practices may circumvent criminal law safeguards. The committee estimated approximately 2% of reviewed cases may not have followed all legally required safeguards. No prosecutions have followed any of these findings. This is not regulatory imperfection awaiting correction. It is the accountability structure operating as designed — identifying concerns, documenting them, and treating them as administratively manageable rather than legally determinative. The Dillon case is one documented individual instance within this systemic pattern. The CPSO found conduct that risked the appearance of coercion, named it, verbally cautioned the physician, and permitted him to continue practicing. The patient whose death occurred under coercion-adjacent conditions is dead. The accountability architecture performed as designed. This committee cannot recommend extending that architecture to a population — sole mental illness — for whom the clinical record confirms the relevant determinations are even harder to make reliably than they were for Dillon.

The government's own Sixth Annual Report on MAID, reporting 2024 data, confirms the deeper problem. The dominant drivers of MAID requests are loss of ability to engage in meaningful activities — cited by 95.1% of Track 1 and 97.5% of Track 2 recipients — followed by loss of independence, loss of dignity, and increasing dependence on others. For Track 2 specifically, 44.7% of recipients cite isolation and loneliness. These are not biological facts fixed by medical processes. They are functional and relational conditions shaped by what supports exist or do not exist. The Canadian Institute for Health Information data presented to this committee confirms that 41% of adults diagnosed with mental health disorders say their needs were only partially met or were completely unmet. The Mental Health Commission has costed the upstream alternative — services that are known to work, that prevent the suffering driving these requests, and that are cost-effective. The committee has been told this directly, in the same hearings, by the same witnesses. This is not a case where the state does not know what to do. It knows. The solution is in the record. The choice to consider expanding the death pathway instead is a policy decision made with full knowledge of the alternative. When the dominant drivers of MAID requests are conditions the state could address through adequate care, housing, community supports, and mental health services — and the state has been told this — the irremediability determination is not a clinical finding. It is a social fact rendered in clinical language, produced by a policy choice, and resolved by an irreversible outcome.

The disproportion is not only by disability status. Elizabeth Sheehy testified at this committee that Track 2 already disproportionately ends the lives of disabled women, and that extension to sole mental illness would disproportionately end the lives of mentally ill women. The government's own data confirms Track 2 recipients are disproportionately female. A regime that applies a lethal outcome disproportionately to women with disabilities — on the basis of their disability — engages the intersection of sex and disability discrimination that section 15 specifically addresses under Fraser v. Canada (SCC, 2020). This committee has that testimony. It is in the record of these hearings.

The disproportion extends further than sex and disability. Statistics Canada's Aboriginal Peoples Survey confirms that 32% of First Nations people living off reserve and 30% of Métis have one or more disabilities — rates higher than non-Indigenous Canadians even after accounting for age, geography, and population centre size (Statistics Canada, Aboriginal Peoples Survey, 2019). Indigenous women are more likely than non-Indigenous women to have disabilities, and more likely to experience severe or very severe disabilities. The state that produced those elevated rates through documented colonial harm — residential schools, forced child removal, the destruction of community supports — is now operating a death pathway that disproportionately captures disabled people, framed as compassion and mercy. In 2021, Indigenous children accounted for 7.7% of children under 15 but 53.8% of all children in foster care — 14 times more likely than non-Indigenous children to be in state care (Statistics Canada, Census 2021). Those children will age into disability status at the rates the data confirms. The regime they age into will be operating. Simultaneously, the federal government cut Jordan's Principle educational funding for disabled First Nations children in Ontario from $122.1 million to $1.2 million in a single year — a 99% reduction — while continuing to operate a death pathway for the adults those children become (Indigenous Services Canada, Jordan's Principle operational data, Ontario, 2024–2025, reported CBC News, January 9, 2026). The structural effect of these simultaneous conditions — elevated disability rates produced by colonial harm, withheld supports, and a death pathway that reaches disabled people at rates disproportionate to the general population — is constitutionally significant regardless of the intent behind any individual policy decision. Under Fraser v. Canada (SCC, 2020), what matters is disproportionate adverse impact on intersecting protected classes. The data establishes that impact. The section 15 analysis cannot be conducted in isolation from the conditions the state has produced and continues to produce for the people the regime disproportionately reaches.

Even if a court found Track 2 violated section 7, the government would argue section 1 saves it. The full Oakes test requires the government to establish four things. First, a pressing and substantial objective — the relief of irremediable suffering in competent adults is accepted. Second, rational connection — a regime that cannot reliably determine irremediability, cannot screen suicidality from autonomous choice, and cannot detect structural coercion is not rationally connected to that objective; it applies death to suffering it cannot assess. Third, minimal impairment — applying an irreversible lethal outcome to a determination that fails more than half the time, on a population for whom the clinical tools do not yet exist, is not the least restrictive means available; investing in those tools, improving care, and providing adequate supports would achieve the objective without the irreversible outcome. Fourth, proportionality of effects — the government's own 2024 data shows the dominant drivers of MAID requests are loss of independence, isolation, and inability to engage in meaningful activities; these are conditions the state produces through inadequate supports and could remedy; the effects of applying death to socially produced suffering that could be addressed by other means are not proportionate to the objective. Section 1 does not save what all four stages of Oakes condemn.

The constitutional analysis does not begin with the Charter. It begins with the Criminal Code. Under section 241.2(2), a person has a grievous and irremediable medical condition only if their illness, disease or disability is serious and incurable; they are in an advanced state of irreversible decline in capability; and their suffering cannot be relieved under conditions they consider acceptable. These are not aspirational standards. They are statutory predicates that must be operationally established before a death can be authorized. The clinical record before this committee establishes that across all of Track 2 — and most acutely where mental illness is the sole condition — none of the foundational predicates can currently be reliably verified. Irremediability cannot be reliably assessed — at chance level — 47% accuracy for predicting chronic treatment-resistant depression, by the SCP's own evidence review. Advanced irreversible decline cannot be reliably established where the dominant drivers of suffering are social and relational conditions the state could remedy — as the government's own 2024 data confirms across Track 2. Voluntariness free from external pressure cannot be reliably verified where structural coercion is documented, undetectable by available tools, and acknowledged by the regime's own regulatory decisions. A regime cannot constitutionally rely on statutory safeguards that are operationally incapable of determining whether the statutory criteria are met. That is not a Charter argument. It is a more fundamental problem: the regime is authorizing deaths it cannot establish meet its own legal threshold. The constitutional argument is the second problem. The first is that the machinery is not presently capable of performing the determinations the statute requires — across the entire track, not only at its most contested boundary.

The anticipated response to the classification instability argument is that law routinely operates on imperfect medical categories — tax thresholds, disability ratings, competency assessments, psychiatric diagnoses. Borderline cases do not eliminate discrimination analysis. That response is correct as far as it goes. It does not go far enough. The constitutional problem here is not that the mental/physical boundary has fuzzy edges. Many legal classifications do. The problem is what the boundary determines. In ordinary administrative contexts, classification errors produce correctable consequences — a denied benefit can be restored, a licensing decision reversed, a bracket recalculated. The tolerance courts extend to imperfect classification reflects the reversibility of what that classification produces. This regime applies its classification to authorize death. Death cannot be corrected. Death cannot be appealed. Death cannot be reversed. The instability of the mental/physical boundary is not a peripheral imperfection in a regime producing reversible outcomes. It is a foundational unreliability in a regime producing the one outcome the state cannot undo. Under Bedford v. Canada (SCC, 2013), arbitrariness is measured by the relationship between a law's effects and its objective. A regime whose stated objective is the relief of irremediable suffering — and whose foundational clinical classification cannot reliably determine who meets that criterion — is not merely imperfect. It is applying an irreversible lethal outcome to a determination the clinical record confirms it cannot currently make. That is not the kind of imperfection the law tolerates. It is the kind Bedford was designed to catch. And unlike every other context in which courts extend tolerance to imperfect classification, there is no appeal mechanism, no correction process, and no remedy available once the classification error is applied.

Courts have long accepted that constitutional rights may require individualized clinical judgment operating under uncertainty. That acceptance reflects the reality that clinical judgment, while imperfect, is the best available tool for determinations the law requires. It does not extend to a regime whose foundational determinations are acknowledged as presently unresolvable by the clinical organizations that ideologically support it. The Society of Canadian Psychiatry — which supports MI-SUMC expansion in principle — relied on Nicolini et al.'s review identifying Dinga et al.'s model as the best available predictor — which correctly identified chronic treatment-resistant depression after two years in only 47% of cases, at chance level. The heads of psychiatry at thirteen Canadian medical schools told this committee directly that no validated tool exists to distinguish suicidal ideation from an autonomous MAID request. These are not opponents of the regime speaking. They are the regime's own clinical infrastructure admitting, on the parliamentary record, that the required judgment cannot presently be made. Courts extend tolerance to imperfect clinical judgment. They do not extend it to a regime whose own advocates have told Parliament the required judgment cannot presently be made reliably — for a protected class — producing an irreversible outcome.

Parliament itself placed the clinical unreadiness on the legislative record — twice. Bill C-39 extended the MI-SUMC sunset clause by one year. Bill C-62 extended it by three more. The stated rationale in both cases was that provinces and territories raised concerns about the readiness of their health care systems, and that more time was needed to develop policies, standards, guidance, and resources for assessors and providers. That is Parliament acknowledging — in legislation, on the public record — that the clinical infrastructure required to safely administer MI-SUMC MAID did not exist at the time of each extension. The committee this article addresses was reconvened specifically to examine readiness before the automatic expansion. The evidence before this committee does not show a system that has achieved the readiness Parliament acknowledged was absent when it enacted the pause. Crawford's systematic review found no validated tool. The SCP precision modeling placed irremediability prediction at chance level. The MDRC documented 178 noncompliance cases in 2023 alone. The pause was Parliament's own admission. The evidence is the answer to it. And the answer is that the pause did not solve a training problem. The tools the system needs do not exist because the foundational clinical distinctions they would need to make cannot presently be reliably drawn — and in the case of the mental/physical boundary, the medical literature establishes that the distinction does not reflect biological reality at all. Parliament acknowledged the infrastructure did not exist. It still does not. A regime that cannot be safely administered when Parliament paused it, and cannot be safely administered now that the pause is expiring, is not a regime awaiting readiness. It is a regime awaiting repeal.

The question of whether the framework is workable is not hypothetical. The regime has been operating since 2016. Track 2 has been operating since 2021. If the clinical screen were successfully identifying people in genuinely irremediable suffering — separating them from people in socially produced, potentially remediable distress — the outcome data would show it. It does not. The government's own 2024 data shows isolation and loneliness cited by 44.7% of Track 2 recipients. Loss of ability to engage in meaningful activities cited by 97.5%. Loss of independence, loss of dignity, increasing dependence on others — these are the dominant documented drivers. These are not the signatures of irremediable biological suffering. They are the signatures of inadequate care, withheld supports, and social abandonment — conditions the state produces and could remedy. If the screen were working, people whose dominant drivers are remediable social conditions would not be getting through. They are getting through. The screen is not working. This is not a theoretical concern about future risk. It is a description of what the regime is producing right now, documented in the state's own annual reports. The Dillon case makes it concrete and individual: the accountability architecture identified conduct that looked like coercion, named it, and permitted the death anyway. A framework whose outcomes contradict its stated criteria is not a framework awaiting better implementation. It is a framework that is not workable. The expansion advocates' answer — wait for better tools, develop better guidance, trust the process — is answered by five years of outcome data. The process has been running. These are the results.

The Autonomy Claim Cannot Be Verified — And the Default Is Death

The regime's response to every structural objection is autonomy. The state is not deciding for disabled people that their lives are less valuable. It is respecting the competent autonomous choice of individuals who have made that determination for themselves. The clinical panel is not judging the worth of disabled lives. It is a gateway function — verifying that the person meets eligibility criteria before their own decision is honoured.

That framing is the regime's strongest ground. It is also where the clinical record destroys it most completely.

The gateway cannot verify what it claims to verify. The clinical record before this committee — from organizations that support expansion as much as from those that oppose it — establishes that the assessor cannot presently determine reliably: whether the request is driven by autonomous choice or by the illness generating it; whether the suffering is irremediable or produced by inadequate care and social abandonment the state could remedy; whether the request is free of structural coercion or shaped by a system that has failed the person until death appeared to be the only remaining option. These are not peripheral uncertainties at the edges of a functioning system. They are the foundational determinations the regime requires — acknowledged as presently unresolvable by the clinical bodies that built the assessment framework.

When a gateway cannot verify the autonomy it claims to respect, the autonomy framing does not insulate the regime from constitutional scrutiny. It relocates the constitutional problem. The question is no longer whether the state is overriding individual choice. It is whether the state has constructed a mechanism that defaults to death when it cannot verify that autonomous choice is present — and whether it may do so for a Charter-protected class, on the basis of characteristics of that status, producing an irreversible outcome, without an evidence base establishing death as the superior outcome to adequate care.

The answer that flows from the Charter, from the UN CRPD, and from the historical record is no. The structural default encoded in the regime — when verification fails, approve — is not a neutral administrative gap. It is a value judgment about protected class members' lives embedded in the regime's design. It says: when we cannot tell whether a disabled person's desire to die is autonomous or illness-driven, coercion-free or structurally coerced, irremediable or socially produced — we approve the death and call it their choice. The same unverifiable desire to die in a non-disabled person triggers crisis intervention. The same state that cannot verify the autonomy responds differently depending on disability status. That differential response is not justified by the clinical record. It is not supported by an evidence base establishing death as the superior outcome. It is a structural assumption — that unverifiable desire-to-die in a disabled person should default to approval — that has no clinical foundation and that the regime has never been required to justify.

The Cases That Created Track 2 Do Not Support It

Jean Truchon had cerebral palsy with triparesis. His suffering was real and his autonomy was genuine. But the record of his circumstances makes clear that a central driver of his MAID request was the state's failure to provide the community supports that would have allowed him to live independently. The state had institutionalized him. Confining disabled people to institutions while denying them supports for independent living is itself a form of state-imposed disadvantage on the basis of disability — the precise harm section 15 prohibits. The constitutional remedy for a section 15 violation is restoration of the right violated. Not death. The question of whether the state may offer death as the remedy for suffering it produced through its own rights violations has never been put to the Supreme Court of Canada. It must be.

Nicole Gladu had post-polio syndrome — a materially different condition from Truchon's cerebral palsy. She held the identical legal right established by the same ruling. She had independent living and, by her own account, the economic means to access alternatives unavailable to Truchon. She never invoked her right to MAID. She died of natural causes in Montreal on March 27, 2022, at the age of 76. Her close friend confirmed she chose to let nature take its course. The difference between Truchon and Gladu was not medical. It was the material conditions of their lives — conditions the state controls and that section 15 obligates the state to provide on an equal basis. Truchon's irremediability was not a clinical fact. It was a policy choice the state made and then offered to permanently resolve.

Julia Lamb has spinal muscular atrophy. Her constitutional challenge was adjourned in September 2019 after the government's own medical expert, Dr. Madeline Li, submitted a written opinion that Lamb already qualified under the existing Track 1 framework — not under Track 2, which did not yet exist, but under the original "reasonably foreseeable" criterion. Dr. Li's submission identified a specific hypothetical pathway: if Lamb were to indicate an intent to stop using her BiPaP ventilator at night and refuse treatment for the chest infection that would inevitably follow, she would likely meet the reasonably foreseeable threshold. Crucially, she would not need to actually refuse the ventilator — only express a certain intent to do so. The case was adjourned without the constitutional challenge being decided on its merits. Lamb did not use assisted dying. She married. In her own words: "I am comfortable and happy with my days. But it is a huge relief and gives me so much peace to know that when I am ready, I do have a choice." That is not the statement of someone who needed a new death eligibility pathway. It is the statement of someone reassured that the existing one already covered her — through a hypothetical device refusal, constructed by the government's own expert, requiring only expressed intent rather than clinical event. Two things the Lamb case establishes that are constitutionally unanswerable regardless of her diagnosis. First: Track 2 was never constitutionally necessary — the original framework already reached her situation, through the law whose restrictiveness had been the basis of the constitutional challenge. Second: the adequacy of that pathway — built on expressed intent rather than actual refusal, never tested in a merits ruling — is a question the adjournment left permanently unanswered. The constitutional necessity argument for Track 2 collapses entirely when its own founding case was already covered by the law it claimed to be challenging.

What Equality Actually Requires

Gilbert argues that equality means treating mental suffering no differently from physical suffering. At that level of generality she is right. The question is what equal treatment requires in practice.

The section 15 cases she cites — Eldridge v. British Columbia (SCC, 1997), Fraser v. Canada (SCC, 2020), Attorney General v. G (SCC, 2020) — establish positive obligations on the state to remedy the disadvantaged position of disabled people. Those obligations run toward life, toward support, toward the conditions that make independence possible. They do not run toward a regime that makes disability the qualifying criterion for death.

The Council of Canadians with Disabilities, ARCH Disability Law Centre, Inclusion Canada, and Not Dead Yet Canada have opposed Track 2 consistently and unanimously. These organizations represent the people section 15 was written to protect. They live under this regime. They are served by the social workers, care workers, and community support professionals who see what the regime produces on the ground. They have applied the same equality principles Gilbert invokes to their actual experience and reached the opposite conclusion. Gilbert came before this committee affiliated with a pro-expansion lobby organization, to argue for a regime the major disability rights organizations in Canada have consistently rejected. The rights language she used belongs to that movement. The conclusion she drew from it does not.

What International Law Confirms

Canada ratified the UN CRPD on March 11, 2010. Under the principle Dickson CJC established in Slaight Communications Inc. v. Davidson (SCC, 1989), international human rights instruments Canada has ratified must be relevant and persuasive sources for Charter interpretation.

On January 25, 2021 — before Bill C-7 received Royal Assent — three independent UN mechanisms issued a joint statement titled "Disability is not a reason to sanction medically assisted dying." The statement was issued by the Special Rapporteur on the rights of persons with disabilities, the Special Rapporteur on extreme poverty and human rights, and the Independent Expert on the enjoyment of all human rights by older persons. It stated directly: "Disability should never be a ground or justification to end someone's life directly or indirectly" and "Under no circumstance should the law provide that it could be a well-reasoned decision for a person with a disabling condition who is not dying to terminate their life with the support of the State." Parliament received that warning. It enacted Track 2 anyway.

In March 2025, the UN CRPD Committee issued its Concluding Observations on Canada — document CRPD/C/CAN/CO/2-3 — and called for the repeal of Track 2 in full, including Part 2 of Bill C-7, the planned 2027 MI-SUMC expansion, and any further expansions to mature minors or advance requests. The Committee described Track 2 as based on "negative, ableist perceptions of the quality and value of life for persons with disabilities" and identified it as a violation of Article 5 (non-discrimination) and Article 10 (right to life) of the CRPD. It called on Canada to address the root causes driving persons with disabilities to request MAID — poverty, lack of housing, inadequate services, and systemic discrimination — describing these as the actual source of the suffering the regime treats as irremediable.

These bodies applied Gilbert's own analytical framework without her limiting constraint and reached repeal, not expansion. A recommendation to expand Track 2 would deepen Canada's violation of ratified treaty obligations its courts are required to consider.

The constitutional analysis is not only academic. In September 2024, a coalition of disability organizations represented by ARCH Disability Law Centre launched a Charter challenge in the Ontario Superior Court against Track 2 MAID, arguing it discriminates against persons with disabilities under sections 7 and 15 of the Charter and results in premature deaths driven by unmet needs rather than autonomous choice. The domestic constitutional challenge and Canada's international obligations now converge on the same conclusion. Expansion deepens both simultaneously.

What Parliament Knows About Consent It Is Choosing to Ignore

In 1972, Alberta MLA David King rose in the Legislature to support the repeal of the Sexual Sterilization Act. His argument was not that individual sterilizations were wrong case by case. It was that no panel has the authority to determine who can and cannot reproduce on the basis of disability status. The mechanism itself was the problem — regardless of its framing, regardless of its claimed compassion. That was political speech, not judicial precedent. But it named a constitutional principle that the Charter's section 15 equality guarantee has since formalized: the state cannot construct a clinical panel to make categorical determinations about the worth of disabled lives and apply irreversible outcomes on the basis of those determinations. What King recognized through political accountability in 1972, the Charter now requires as law.

The Sexual Sterilization Act was disproportionately applied to females, to Indigenous and Métis persons, to the institutionalized, to the unemployed, and to the rural poor. These are the same populations Track 2 disproportionately reaches. That is not rhetorical. It is the structural parallel made evidentiary — the same people, the same state, a different mechanism, the same constitutional problem.

Gilbert argues that any distinction between disability types within the MAID regime discriminates against the excluded class. Applied consistently, her logic requires that all disabilities be equally open to the death pathway. But this confirms the King principle rather than defeating it. If no panel can constitutionally determine which disability type is excluded — because any exclusion discriminates — then no panel can constitutionally determine which disability type is included either. The mechanism that requires a clinical panel to assess whose disability-linked suffering is fatal enough to warrant death is constitutionally impermissible for the same reason King named in 1972. Gilbert's argument does not save the mechanism. It indicts it. A mechanism that cannot constitutionally exclude any disability type cannot constitutionally include any either.

The Screen the Regime Claims Is Running Is Not Running

Every assessment under every current track — Track 1 and Track 2 — requires a clinician to confirm that the person's suffering is not predominantly driven by mental illness, that their request is not suicidality produced by a treatable condition, and that their desire to die is a settled autonomous choice rather than a response to remediable circumstances. That screen is the regime's foundational safeguard. It is the clinical work that is supposed to separate a legitimate MAID request from a person in crisis who needs intervention instead.

The Fabiano paper establishes that the mental/physical distinction does not reflect biological reality and that comorbidity is the rule rather than the exception. The psychiatric evidence before this committee establishes that suicidality cannot be reliably screened from an autonomous MAID request. And the CDC data — 54% of suicide decedents had no known psychiatric condition — establishes that the desire to die is not diagnostically bounded at all. It occurs across the entire population, in grief, in existential crisis, in response to loss and circumstances that may resolve with time and adequate support.

These three findings together mean the screen the regime depends on is not functioning in any track. Assessors approving Track 1 and Track 2 requests right now are recording a clinical conclusion — that suffering is not predominantly mental, that the request is not suicidal ideation, that the desire to die is settled and autonomous — that the clinical record confirms cannot currently be made reliably. The safeguard is not a safeguard. It is a documentation practice applied to a determination that does not work. Which means the MI-SUMC exclusion is not keeping mental illness out of the current tracks. It is misclassifying it as physical and approving it anyway — undetected, unscreened, and uncounted in the data the regime uses to justify its own safety. The claim is not that no one currently accessing MAID has made a genuine autonomous choice. The claim is that the regime cannot presently determine which of the approvals it has issued reflect genuine autonomous choice and which reflect the egosyntonic suicidal state, structural coercion, or socially produced suffering the screen was supposed to identify. It cannot know. And it cannot correct the outcome.

The Suffering-Relief Claim, the Egosyntonic Problem, and the Evidence That Doesn't Exist

The regime's central promise is that it compassionately relieves irremediable suffering. That claim requires establishing two things the evidence does not support: that the suffering being addressed is genuinely irremediable rather than produced by remediable social conditions; and that death produces a better outcome than adequate care would for this population. Neither has been established.

The government's own 2024 MAID data establishes what is driving requests. Loss of ability to engage in meaningful activities: cited by 95.1% of Track 1 and 97.5% of Track 2 recipients. Loss of independence: 75–79%. Loss of dignity: 63–74%. Isolation and loneliness: 44.7% of Track 2 specifically. These are functional, relational, and social conditions — not fixed biological endpoints. They are conditions the state produces through inadequate care and could address through adequate support. The regime treats them as irremediable. They are not. They are the predictable consequences of policy choices the state has made and could reverse.

There are no randomised controlled trials comparing MAID outcomes to adequate care and supports for this population. No comparative outcome data establishing that death produces better results than remediated social circumstances. No peer-reviewed literature establishing death is more effective than housing, community support, or mental health services at relieving the suffering the regime treats as irremediable. The positive outcome claim — that death constitutes relief of suffering superior to the alternative — is a philosophical position applied as a clinical finding. It has no evidence base. Under Oakes rational connection, the government must demonstrate the means actually achieves the objective. It cannot make that demonstration because the evidence base does not exist. It has never been required to produce it.

Egosyntonic Suicidality — The Assessment Impossibility the Regime Cannot Name

There is a clinical phenomenon that makes the foundational assessment the regime requires not merely difficult but structurally impossible in a subset of cases that cannot currently be identified. It is called egosyntonic suicidality.

In psychiatric practice, a thought or desire is egosyntonic when it feels consistent with the person's sense of self, values, and rational worldview. It does not present as intrusive or unwanted. It presents as the most logical, considered conclusion available given the person's circumstances. Suicidal ideation in severe mental illness is frequently egosyntonic. The person experiencing it does not perceive the desire to die as a symptom. They perceive it as reason. They will describe it to an assessor as settled, considered, and autonomous — because within the cognitive and emotional framework the illness has created, it genuinely is.

This is the assessment impossibility at the heart of the regime's foundational gate. The assessor must distinguish two things: suicidal desire that is a symptom of illness — which the state's entire suicide prevention infrastructure is built to interrupt — and autonomous settled choice to die — which the regime facilitates. The egosyntonic quality of illness-driven suicidal desire makes those two things phenomenologically identical from the inside. The person cannot tell the difference between them from their own experience, because the illness has made the desire feel rational. The assessor therefore cannot use the person's subjective report to make the distinction. And the external validated tools that could make it from the outside — the instruments thirteen heads of Canadian psychiatry told this committee do not exist — are precisely what would be required to overcome the impossibility the egosyntonic quality creates.

This is not a theoretical objection. The documented testimony of suicide attempt survivors establishes what egosyntonic suicidal desire looks like at the moment of action — and what happens when the egosyntonic state breaks. Kevin Hines survived a jump from the Golden Gate Bridge in September 2000 while struggling with bipolar disorder — a mental illness. His documented testimony: "My hands left the rail and I had an instantaneous regret for my actions... all I wanted to do was live." Of the 39 known survivors of Golden Gate Bridge jumps, 19 have come forward and all reported the same instant regret. Richard Seiden's 1978 study tracked 515 people stopped from jumping at the Golden Gate Bridge: 90% did not go on to die by suicide. The suicidal crisis — even at the point of near-lethal action — is frequently temporary. The egosyntonic quality makes it feel permanent. Survival reveals it was not.

The irremediability assessment is supposed to screen for precisely this — to identify the cases where the desire to die, however settled it feels, is not permanent. It must distinguish the person whose desire is egosyntonic and potentially transient from the person whose desire reflects a settled and irremediable choice. The clinical record establishes it cannot currently do so reliably. The survivor data establishes what the consequence of that failure is: some proportion of people who present to assessors with egosyntonic suicidal desire that feels settled and autonomous would not have maintained it given time and adequate support. The intervention the regime applies is permanent. The state cannot know, after approving the death, whether the desire would have persisted. It cannot correct the outcome if it would not have.

The clinical detail of egosyntonic suicidality is not peripheral to the constitutional argument. It is the mechanism that explains why the three determinations the regime requires cannot presently be reliably made — and why that operational failure has a constitutional consequence. Standard capacity assessment, the clinical tool the regime deploys at this gate, cannot detect state-dependency. It cannot distinguish a desire to die that would survive resolution of the crisis from one that would not. The regime applies an irreversible outcome to a state it cannot presently characterize. The constitutional problem is not that the tools are imperfect. It is that the tool being used cannot reach the determination the Constitution requires before death is authorized.

The Disproportion Is Structural, Not Clinical — And the State Has Made It Invisible

The regime's defenders argue that most MAID recipients are educated and not living in poverty — that the demographic profile of recipients does not support the disproportion concern. That argument cannot be assessed on the data available, because the data was never designed to assess it. The government's annual MAID reports track age, sex assigned at birth, diagnosis category, and province. They do not track race, ethnicity, Indigenous status, sexual orientation, gender identity, housing status, or income level at the intersectional granularity that a section 15 disparate impact analysis under Fraser would require. The aggregate profile the regime cites to defend itself is drawn from data that was structurally designed to make its worst effects invisible. You cannot use the absence of data you chose not to collect to prove the absence of harm to communities that data was never designed to see.

The Track 2 data specifically tells a different story from the aggregate. The dominant drivers for Track 2 recipients — isolation at 44.7%, loss of meaningful activities at 97.5%, loss of independence at 79% — are social conditions concentrated in communities that face documented structural disadvantage: disabled women, Indigenous people, racialized people, LGBTQ+ people, people in poverty. The "educated and not poor" finding is overwhelmingly a Track 1 profile — terminal cancer patients who are older, have more social resources, and whose clinicians have assessed natural death as reasonably foreseeable. That profile does not describe the Track 2 population the expansion debate is about. Using the aggregate to defend Track 2 is a category error that the data architecture makes possible precisely because it was never designed to separate them.

Even if the state removed all medical gatekeeping entirely — pure autonomous choice, no diagnosis required, no clinical assessment — the pathway would still disproportionately reach women, Indigenous people, disabled people, LGBTQ+ people, and racialized people. Because the conditions producing the desire to die are not randomly distributed across the population. They are structurally concentrated in those communities by documented state-produced harms — colonial dispossession, poverty, inadequate housing, violence, discrimination, withdrawal of supports. The disproportion is not a product of the medical gateway. It is a product of the social conditions the state produces and has failed to remedy. Removing the gateway does not change who dies. It removes the last clinical friction between structurally produced suffering and state-facilitated death. Any future version of this pathway — medicalized or not — must meet the same constitutional test. The structural disproportion does not disappear because the framing changes.

A preference — even strongly held, deeply felt, and autonomously expressed — is not an absolute constitutional entitlement. Preferences yield when acting on them imposes irreversible harm on others in the same category whose screen fails. They yield when formed under conditions of state-produced deprivation the state is obliged to remedy. And they yield when the philosophical framework generating the preference claim is contested — not established as the constitutional standard — and its application in this structural context reinforces the exact devaluation section 15 was built to prevent. Gilbert presents autonomy-maximization as though it were a constitutional holding that forecloses all other analysis. It is not. The Charter requires the state to balance life, liberty, and security. Autonomy is not the master value that ends all other inquiry. The autonomy argument also has a specific history in the disability context: it has been used against disabled people, not for them. Involuntary sterilization was justified on autonomy grounds. Institutionalization was justified on autonomy grounds. Parliament is currently criminalizing coerced sterilization of disabled people — acknowledging that the autonomy framework, applied without structural analysis, has been weaponized against this same population. When the autonomy argument is deployed to expand a regime that produces disproportionate lethal outcomes on a protected class, section 15's substantive equality framework requires the court to ask what the autonomy claim is doing structurally — not just whether it sounds like a rights claim in the abstract. Gilbert never does that assessment. She assumes the answer and calls it constitutional law. Her position is internally inconsistent: she accepts medical gating, she accepts that not everyone qualifies, but she claims any gate applied to mental illness is discriminatory without structural analysis. A framework that accepts gates in general but treats any specific gate as discriminatory without assessing structural effect is not applying section 15. It is asserting a conclusion and working backward.

Gilbert described this committee's work correctly. The exclusion of MI-SUMC creates a distinction. It reinforces disadvantage. It should not survive constitutional scrutiny. She had the diagnosis exactly right. But the patient is not the exclusion. The patient is Track 2.

Track 2 has never been assessed by the Supreme Court. Its clinical screen fails more than half the time for the very determination the regime requires — by the admission of the organizations that built it. Its clinical foundation rests on a mental/physical divide that Fabiano et al. — researchers from the University of Ottawa, Harvard Medical School, King's College London, and six other international institutions — call a monumental mistake in medicine that must be retired entirely. The government's own data shows the dominant drivers of MAID requests are social conditions the state produces and could remedy. The state runs suicide prevention infrastructure and a death eligibility pathway for the same population simultaneously — and the Chief Medical Officer of that crisis infrastructure has told this committee the two cannot coexist without the screen between them, and that the screen does not exist. Parliament is simultaneously advancing legislation to criminalize coerced sterilization of the same population. The institutional knowledge embedded in legislation is not legally inert. Courts have treated Parliament's acknowledgment of a systemic failure — embedded in the statute creating the remedy — as evidence bearing on whether the state can simultaneously claim that failure does not exist in adjacent domains. Parliament cannot in one bill acknowledge it cannot detect coerced consent in medical procedures performed on disabled people, and in another maintain that its clinical infrastructure reliably detects coerced consent in lethal procedures for the same population. That is not a rhetorical coincidence. It is a record of Parliament's own institutional self-knowledge, created contemporaneously with the decision to expand the regime it acknowledges it cannot adequately screen.

The major disability rights organizations in Canada, the UN CRPD Committee, the UN Special Rapporteur, leading psychiatric and clinical witnesses before this committee and its predecessor, the Alberta Legislature — which passed Bill 18 on April 22, 2026, becoming the first province to pass legislation prohibiting Track 2 MAID in Alberta — and the Chief Medical Officer of Canada's own suicide prevention infrastructure — all on the parliamentary record — have reached the same conclusion. Inclusion Alberta, welcoming Bill 18's passage: "Track 2 is discriminatory in that it offers assisted suicide to those not dying and it applies only to persons with disabilities. True and lasting protection requires Parliament to amend the Criminal Code to permanently end Track 2." Gilbert's own analytical framework, applied honestly and completely, reaches the same conclusion.

The recommendation this committee can defensibly make is to repeal Track 2. Parliament created it. Parliament can end it. The Charter does not require it to exist. The evidence does not support its expansion. Gilbert's equality framework — applied correctly, completely, and without the limiting constraint she placed on her own argument — makes the case for ending it. The specific legislative mechanism is available and precise: this committee can recommend that Parliament repeal the provisions of section 241.2 of the Criminal Code that extend MAID eligibility to persons whose natural death is not reasonably foreseeable. That recommendation requires no finding of unconstitutionality. It requires no Supreme Court ruling. It requires only Parliament to exercise the same legislative judgment that created Track 2 — informed now by five years of evidence, two parliamentary committee records, an ongoing Charter challenge, the position of every major disability organization in Canada, the clinical consensus of the country's psychiatric institutions, and the direct testimony of the Chief Medical Officer of the state's own crisis prevention infrastructure. The evidence was not available to the 2021 Parliament that created Track 2. It is available to this committee. The constitutional path is clear.

The regime claims its screen identifies people whose suffering cannot be remedied any other way. The government's own data shows what the screen is identifying. Those two things are not the same. That gap — between what the regime claims to be doing and what the outcome data shows it is doing — is not a philosophical disagreement about the value of MAID. It is the constitutional problem. It is what Bedford arbitrariness looks like in practice. It is what section 15 disadvantage looks like when it is documented in annual government reports rather than argued in the abstract. And it is what this committee has the evidence and the authority to address.

The constitutional questions this analysis raises about Track 2 are specific to that track: non-terminal conditions, no foreseeable death even in the subjective sense the statute requires, the complete absence of a physical anchor for the irremediability assessment, and a clinical screen the evidence confirms cannot presently perform the determinations it claims to perform. Track 1 involves a different statutory structure — natural death must be assessed as reasonably foreseeable — but that assessment is itself subjective, requires no terminal diagnosis, no fixed prognosis period, and no objective medical anchoring to imminent death, and is not subject to validated screening tools or adequate oversight architecture. The physical dimension of Track 1 conditions provides a different evidentiary context that may matter constitutionally. Whether the structural and operational failures documented here apply to Track 1 in the same way and with the same constitutional force is a question the next article in this series addresses directly. The answer to that question is not predetermined by this analysis — but nor is it foreclosed by it. The recommendation here is specific: repeal Track 2. It does not extend further than the evidence before this committee compels. The constitutional problem running through every argument this article advances is not moral disagreement about the value of assisted dying. It is that the state is exercising lethal authority over a specific population — people whose natural death is not foreseeable even subjectively — under conditions of admitted non-knowledge, with clinical tools its own supporting bodies have acknowledged cannot presently do what the statute requires. That is the problem this committee has the power to address. The recommendation available now, on this record, is that Track 2 cannot constitutionally proceed. Not because the suffering it claims to address is not real. But because the state cannot presently verify that its gateway is reaching the people it claims to be reaching, cannot confirm the autonomy it claims to be respecting, and cannot correct the outcome when it is wrong.