Assisted Suicide and Slippery Slopes: Reflections on Oregon

Abstract

Slippery slope argumentation features prominently in debates over assisted suicide. The jurisdiction of Oregon features prominently too, especially as regards parliamentary scrutiny of assisted suicide proposals. This paper examines Oregon’s public data (including certain official pronouncements) on assisted suicide in light of the two basic versions of the slippery slope argument, the empirical and moral-logical versions. Oregon’s data evidences some normatively interesting shifts in its assisted suicide practice which in turn prompts consideration of two elements of moral-logical slippage that are not widely discussed. One is slippage from an initial autonomy-based public justification for assisted suicide which does not include burden-based concerns within its operative account of voluntariness to an evolved public justification that does. The other is an expansion of a terminal illness ground to include chronic illnesses effectively rendered terminal via a refusal of treatment.

KEYWORDS:

• Assisted suicide
• slippery slope
• Oregon
• euthanasia
• terminal illness
• chronic illness

Introduction

The ‘slippery slope’ argument features prominently in debates over assisted suicide.¹ It comes in two basic versions: the moral-logical argument (Jones 2011, Keown 2022) and the empirical argument (Keown 1995, High Court of Ireland 2013). According to the moral-logical version the central assisted suicide rationale cannot be rationally (consistently, non-arbitrarily) limited to a narrow ground permitting access to assisted suicide. Therefore, given the claimed absence² of strong countervailing reasons for limiting access to assisted suicide, the central rationale for the practice justifies broad access – access beyond cases of terminal illness and including at least cases of serious chronic illness (and perhaps beyond these, towards cases of psychiatric suffering, non-voluntary euthanasia, etc.). For the purposes of the argument’s basic form it matters little whether the central assisted suicide rationale is a morally thick account of individual autonomy (e.g. Dworkinian authenticity), or a medically-informed account of beneficence, or a combination of both. While criticism of the moral-logical version has been made (Lillehammer 2002) its central claim has been accepted as part of an argument in favour of the morality of assisted suicide (Raz 2013). According to the empirical version of the slippery slope argument, a jurisdiction’s legalization of assisted suicide tends to cause expansion of assisted suicide cases therein, whether in terms of increasing numbers accessing the practice, or increasing grounds of access to the practice, or both. (This summary presentation of the empirical argument is broader than that proposed by Keown [2021] who identifies the central form of the argument as the near inevitability that safeguards will prove to be ineffective.) The empirical version is difficult to deny wholesale, though (as discussed below) Oregon is often invoked as a counter-example against it.

The present paper focuses in the first instance on the empirical version of the slippery slope argument, and in particular on the case of Oregon. Emerging from that analysis will be two considerations proposed as bearing significantly on important yet underappreciated junctures within the moral-logical argument. Though obviously distinct from one another, since logical arguments are not contingent on empirical evidence, the moral-logical and empirical versions of the slippery slope argument are often invoked together in analyses of assisted suicide. This is quite understandable. If the moral-logical implications of the claimed in-principle permissibility of assisted suicide were not serious then there would be little moral case against expansion of assisted suicide access. Thus the empirical version of the slippery slope argument presupposes a wider moral critique of assisted suicide. And if there was no evidence of assisted suicide expansion decades after its initial, limited legalization, then the motive for investigating the moral-logical argument would thereby diminish (which of course says nothing about the intrinsic validity of the moral-logical argument).

The case of Oregon

Data in support of the empirical slippery slope argument is not hard to find. Most countries that have had legalized ‘assisted suicide’ (here understood to include voluntary euthanasia) for five years or more³ and thus have significant data sets available have seen significant rises in annual cases.⁴ And most of those which enumerate specific grounds for access to assisted suicide⁵ have witnessed significant widening of such grounds (whether occurring de facto or de jure or both).⁶

Oregon, however, is often touted as an exception to the expansionist dynamic. Under Oregon’s Death with Dignity Act 1997 assisted suicide (not voluntary euthanasia) is accessible only on grounds of ‘terminal disease’ defined as ‘an incurable and irreversible disease that has been medically confirmed and will, within reasonable medical judgment, produce death within six months’. Oregon’s assisted suicide practice was cited as a rejoinder to claims of an assisted suicide slippery slope at an evidence session of the House of Lords Select Committee on Assisted Dying for the Terminally Ill Bill (2005). There Baroness Helene Hayman and Lord Joel Joffe pointed to Oregon in order to challenge Prof. John Finnis’ contribution to the Committee, one that relied in part on slippery slope argumentation. Prof. Finnis argued that the status quo in Oregon at the time was ‘a very intermediate stage’ and added,

if the Oregon experiment or regime were to be adopted in the wider United States it would soon prove impracticable to hold the line in the way that it has been held in the few years since Oregon’s law has been in force, and we would have a movement towards the Dutch experience, extensively – perhaps not as rapidly as in Holland because in certain respects American culture retains a bulwark against cynicism, a bulwark that may conceivably not exist in Holland.

By ‘Dutch experience’ Prof. Finnis had particularly in mind the practice of voluntary euthanasia: the principles underpinning the introduction of assisted suicide also justified voluntary euthanasia of the sort that had by then already become routine (on grounds beyond terminal illness) in the Netherlands. No precise timeline of this movement happening was offered by Prof. Finnis. While he further stated that the line between assisted suicide and voluntary euthanasia would be ‘rapidly gone through’ once the former was (widely) legislated for in the United States, he qualified this remark by adding that the movement would be slower than what transpired in the Netherlands (partly on account of the cultural bulwarks extant in US culture and partly too, argued Prof. Finnis, on account of countervailing political pressures in the United States against the legalization of euthanasia).

18 years later at a different parliamentary committee Oregon was still being cited as an exception to the empirical slippery slope trend concerning expansion of assisted suicide. During proceedings of the Houses of the Oireachtas (Irish Parliament) Joint Committee on Assisted Dying (2023a) Prof. Richard Huxtable cited Oregon as a jurisdiction wherein assisted suicide ‘has not resulted in a slippery slope’. Prof. Huxtable’s appeal to Oregon against the empirical slippery slope argument is clearly not explicable as a reflex against the argument itself for in the same contribution he cited the Netherlands and Canada as examples that support the empirical slippery slope argument. There certainly appears to be a significant level of stability to Oregon’s assisted suicide practices since their legalization 25 years ago, a stability which poses a challenge to the empirical argument. Voluntary euthanasia has not been introduced in Oregon nor has its assisted suicide ground been subject to explicit de jure expansion. Those who reference Oregon in challenging the empirical argument make a prima facie plausible case.

Social normalization and public justification

However, when one examines the Official Data Summaries of the Oregon Death with Dignity Act (Oregon Health Authority 2023, Regnard et al. 2023) one finds that an obvious element of assisted suicide expansion has materialized: substantial increase in annual case numbers. In the first year reported (1998) 16 died by assisted suicide and in the second (1999) and third (2000) years 27 died.⁷ Compare with: 259 deaths in 2020, 255 in 2021, and 278 in 2022. (The final figure for 2022 is likely to be revised upwards by the 2023 report, probably in the region of 15–20 going by previous corrections.) This represents over a ten-fold increase in 25 years with no signs of the increase flatlining. Oregon’s population rose from c. 3.28 m to c. 4.24 m between 1998 and 2021, so the population increase of c. 30% accounts for only a small part of the increased cases. In fact the rate of increase is itself accelerating: it took 17 years for the annual figures to pass 100 cases (1998–2014), five years thereafter for the annual figure to pass 200 (2015–2019), and on the current trajectory it will take three years thereafter for the annual figure to pass 300 (2020–2023).⁸ Given that the introduction of Oregon’s law came through a public vote and gained widespread national and international attention the increase in cases is unlikely to be merely a matter of improved advertisement of the facility. Rather, increasing cases and acceleration of the increase is strong evidence for the social normalization of the practice such that a new norm – the moral permissibility of assisted suicide – is being embedded in end-of-life decision making in Oregon.

While the idea of social normalization was not directly adverted to in the aforementioned exchanges between Baroness Hayman, Lord Joffe, and Prof. Finnis increased assisted suicide cases driven by social normalization surely qualifies as assisted suicide expansion. Whether it qualifies as a sound premise in a slippery slope argument is a distinct question. Against it so qualifying is the fact that a legal permission necessarily counts as a permitting reason, and the social introduction of a new permitting reason necessarily amounts to a form of social normalization. Hence some level of social normalization is logically entailed by the bare legalization of a practice. In itself this could hardly count as ‘slippage’ from what was initially legislated for to some subsequent, morally distinct and morally more dubious practice or set of attitudes. On the other hand, it is surely possible that social normalization can outstrip the reasons which justified the introduction of the initial legal norm. Take, for example, a case where legalization of assisted suicide was publicly justified on the basis of accommodating a minority of patients who wished to, in effect, act against social norms that valued their continued medical care and the intrinsic, equal worth of their lives. That initial public justification would become diluted in a changed context whereby an increasing percentage of patients choose assisted suicide in order to remove from others the undesirable burden of caring for them (in view here are only cases where the ‘others’ really do perceive the continued care as an undesirable burden on themselves, i.e. a patient’s false beliefs on the matter are excluded). In both contexts – that which was initially envisaged and that which transpired – autonomous choices for assisted suicide are protected. But what transpires in the second, re-normed context are choices for assisted suicide not only not intended in the initial justification for the permission but partially discordant with that initial justification. Discordant, because the initial justification presupposed a voluntariness motivated against the grain of others’ care for and valuing of the patient’s life, whereas what transpires now covers voluntariness motivated by others viewing the patient’s care and life as an undesirable burden. Voluntariness in the former case has the character of a self-assertive independence from prevailing social views, whereas in the latter case it takes on a very different character: self-abnegating accommodation of prevailing social views. Note, too, that the motivating content of external views has also altered: from being affirmative of the patient’s continued medical care as worthwhile to being dismissive of the patient’s continued care as a burden. (The point will not be pressed here, but it is worth reflecting on whether typical public justifications for legalization of assisted suicide contain principles and presuppositions which are logically oriented towards society viewing care of the seriously ill, aged, and dying as an undesirable burden. Obviously, if so, this itself would constitute a slippery slope.)

This changed texture in the patient’s voluntariness vis-à-vis its motivating reasons amounts to a form of normative slippage insofar as it was not captured by the initial autonomy-based justification⁹ for legalizing assisted suicide. In itself the slippage is from a voluntariness contemplated by an initial public justification for assisted suicide to the materialization of a different sort of voluntariness expressed by individuals after assisted suicide is enacted. Assuming the voluntariness contemplated per the initial justification remains prominent in ongoing practice such slippage may be of limited public relevance to the justifiability of the law at issue.¹⁰ But perhaps the slippage gestures towards a change in public justification: either from a sanitized, more carefully framed version of the initial justification to the more complete version of the initial justification (now more clearly visible by virtue of what is tolerated without comment), or from a restrictive initial justification to an evolved, more expansive justification. Without here taking a definite position on the matter there is reason to think that Oregon’s public justification evidences some slippage and that it falls into one of these two categories. The idea that Oregon’s initial public justification for assisted suicide did not appeal to (or deliberately capture) voluntariness motivated by burden-based reasons is supported by its first official report (1999) upon the first year of legalization (1998). Therein the voluntariness involved in assisted suicide is presented as self-assured and self-affirmative, and as serenely distant from wider social views which would influence the patient. The report presents negatively (and downplays the occurrence of) assisted suicide choices where the patient ‘feels pressured, either internally or through external forces’ (this phenomenon is distinguished in the report from that of patients being ‘coerced’ outright). Conversely, when recently giving evidence before the Houses of the Oireachtas Joint Committee on Assisted Dying (2023c) Dr. Tom Jeanne of the Oregon Health Authority stated that the percentage of those availing of assisted suicide with express burden-based concerns is ‘only [emphasis mine] 47% or 48%’.¹¹ Dr. Jeanne went on to comment of the percentage figure, ‘I do not see that as particularly concerning … among a handful of concerns that somebody with a terminal illness might cite at the end of his or her life, being a burden to his or her family would seem a reasonable one’. While these statements offer but a snapshot of Oregonian public justification for assisted suicide they cannot be dismissed lightly. Between these two official sources, at least, a justificatory shift seems to have occurred. The shift is mirrored in an increase in burden-based concerns for those accessing assisted suicide in Oregon. Being a ‘burden on family, friends/caregivers’ was a concern for an annual average of 53% of assisted suicide patients between 2018 and 2022, up from 34% for between 1998 and 2002.

It is of course open to a legal community to assert an evolved justification for a practice already legally established (or to more completely articulate the initial public justification in question). The present point is simply this: an assertion of an evolved or more complete justification can amount to moral-logical slippage in the normative underpinnings of the law at issue, thus entailing moral-logical slippage in the normative reach and normative trajectory of that law. There is reason to think that this may have happened in Oregon.

Expansion of the terminal illness ground?

A more common issue in slippery slope argumentation is empirical slippage occurring through expansion of qualifying grounds for assisted suicide access. It would seem at first blush that slippage of this sort has not occurred in Oregon since no amendment has been made to the law expanding access beyond terminal illness as defined under the 1997 Act. Yet when one examines Oregon’s data two features emerge as noteworthy in addition to the accelerating increase of annual assisted suicide cases. It is here hypothesized that part-explaining these features is a possible de facto widening of Oregon’s terminal illness ground,¹² an expansion that is moral-logically interesting in its own right even if, arguendo, empirical evidence does not support its actual occurrence in practice. The two features will be taken in turn.

First, Oregon’s official data provides evidence that misprognosis is occurring in relation to the assisted suicide ground and that the margin of error may have increased drastically over the past four years. From 2020 to 2022 there were 35 reported cases of patients outliving their six-month prognosis, which is half the reported total (69) of all such cases from 1998 to 2019 (though the annual rates have remained relatively constant across the two periods). The figure of 35 patients from 2020 to 2022 is out of a total of 792 assisted suicide patients in those years. This does not mean that the misprognosis rate was only 35/792 (4.4%). It means that the misprognosis rate was 4.4% at an absolute minimum.¹³ It is impossible to tell how prevalent misprognoses were for those that died by assisted suicide before the six-month prognosis point (95.6% of patients). What can be ascertained is that marginal misprognoses by no means exhaust the cases of misprognoses we know occurred. Oregon provides (very limited) data on the duration between the first request for assisted suicide and its eventual occurrence, the most relevant of which data is the outer ‘range’ of the duration, i.e. the longest duration between the two time points. The average annual farthest end of the range for the years 1998–2018 was 536 days (c. 18 months). That is to say, on average per each annual report the longest duration between an initial request for a lethal prescription and death by assisted suicide (not by the terminal illness) was around a full year beyond the six-month terminal prognosis. Every year from 1998 to 2018 bar 1998 itself had a longest duration well in excess of six months. The average longest duration for the years 2019–2022 constitutes a very sharp increase: 1384 days (c. 46 months). For these four years the annual longest duration was on average over two years beyond the six-month terminal prognosis.¹⁴ Again, death occurred by assisted-suicide, not by diagnosed terminal illness. Of course, each of the annual ‘longest duration’ figures proves but one case. Yet it is surely very unlikely that only one case of a badly misprognosed illness is occurring every year. The data indicates a proven case of a badly askew misprognosis for every year and indicates multiple certain cases of misprognosis happening per year (especially over the course of the past four years where the average number of proven misdiagnoses is 10 per annum). And so it is very unlikely for the years 2019, 2020, 2021, and 2022, where there were longest duration assisted suicide cases of, respectively, c. 50 months, c. 36 months, c. 37 months, and c. 62 months, that every other of the roughly nine misprognosis cases those years (that we know of) involved a death only a few months beyond the six month mark.

Second, qualifying illnesses recorded in the official reports have expanded in a direction away from normally (or at least predictably) terminal. In the first three official reports (detailing 1998–2000) the illnesses listed were as follows: cancer, AIDS, congestive heart failure, chronic obstructive pulmonary disease, amyotrophic lateral sclerosis, multi-system organ failure, and scleroderma. Cancer accounted for 88%, 63%, and 78% of all qualifying terminal illnesses from 1998 to 2000 respectively. Two¹⁵ of the three most recent official reports (detailing 2020–2021) include the following additional illnesses: non-COPD respiratory disease, endocrine/metabolic disease (e.g. diabetes), gastrointestinal disease (e.g. liver disease), anorexia,¹⁶ arthritis, arteritis, blood disease, complications from a fall, hernia, kidney failure, medical care complications, musculoskeletal system disorders, sclerosis, and stenosis.¹⁷ Cancer accounted for 66%, 62%, and 64% of all qualifying illnesses from 2020 to 2022 respectively. The annual average of cancer-based assisted-suicide deaths from 2018 to 2022 was 66% compared to 81% from 1998 to 2007. Thus suggested is a loosening of what is considered to constitute a terminal illness defined as ‘an incurable and irreversible disease that has been medically confirmed and will, within reasonable medical judgment, produce death within six months’.

These two dynamics are consistent with and may be partly-explained by a de facto change in how terminal illness is understood by the relevant Oregon authorities. According to an analysis of the Oregon law undertaken by the Swedish National Council on Medical Ethics (2018)¹⁸ the six month terminal illness ground is interpreted in practice to include ‘cases where no treatment will be given to inhibit progression of the illness’. Subsequently the following questions were put to the Oregon Health Authority and received an affirmative response from it, ‘If the doctor suggests, to an eligible patient, a treatment that possibly could (a) prolong life, or (b) transform a terminal illness to a chronic illness, or (c) even cure the disease – and if the patient doesn’t give his/her consent to the proposed treatment is he/she still eligible to take [sic] use of the Act?’; and, ‘If a patient with a chronic disease (for instance, diabetes) by some reason decides to opt out from the life-sustaining medication/treatment and by doing so is likely to die within 6 months, thereby transforming the chronic disease to a terminal disease – does he/she then become eligible to take [sic] use of the act?’ (Care Not Killing 2018;¹⁹ see also Richardson 2018²⁰).

Chronic illness and refusal of treatment

The evolved interpretation of ‘terminal disease’ under Oregon law – endorsed by the Oregon Health Authority and (to my knowledge) nowhere denied by relevant Oregon officials – is not opposed to the text of the relevant law. Oregon’s Act defines ‘terminal disease’ for the purpose of assisted suicide access as ‘an incurable and irreversible disease that has been medically confirmed and will, within reasonable medical judgment, produce death within six months’. A disease that is genuinely ‘incurable and irreversible’ may be treatable but not so in a way that would render it less ‘incurable and irreversible’. In this sense the treatability of a disease will not bear upon its incurability and irreversibility. Curability and reversibility of disease do not themselves serve to distinguish chronic from terminal illnesses, of course. Rather, what distinguishes the two illness types is that regarding terminal but not chronic illness it is inevitable that the illness will directly cause death within a foreseeable, identifiable timeframe. Oregon’s law does not define ‘terminal disease’ thus. Instead, it has a two-part definition: (a) an incurable and irreversible disease that (b) will produce death within six months. The definition does not exclude (b) being brought about by a merely chronic disease under (a) that goes untreated. No part of the definition requires of a patient applying for assisted suicide either that they have pursued treatment for their illness or that they will pursue treatment for their illness.²¹ And so the evolved interpretation of ‘terminal disease’ is itself permitted by a reasonable reading of the Oregon Act.

Statutory interpretation aside, there is the moral-logical question of whether it is consistent (i.e. non-arbitrary) to affirm a ground of terminal illness and yet exclude from that ground chronic illnesses effectively rendered terminal by a refusal to access treatment.²² A key principle here is a patient’s right to refuse medical treatment – a basic principle of medical law and medical ethics. Could it be that expanding the terminal illness ground to include voluntarily untreated chronic illness is warranted as a function of a patient’s right to refuse medical treatment? It would seem so if the central rationale for assisted suicide is autonomy-based. Under that rationale access to assisted suicide is protected out of respect for personal autonomy as it bears upon the most morally and indeed existentially important matters facing an individual. If autonomy justifies assisted suicide it justifies at least as clearly the refusal of medical treatment – the two practices share a common central rationale on this view.²³ And if the latter is practiced there will be cases where chronic illnesses are effectively transformed into terminal illnesses. To exclude such cases of ‘indirect’ terminal illness from the legitimating ground for assisted suicide access would be to set the value of personal autonomy against itself. The exclusion would rationally require an important, morally relevant difference between (i) illness where only the illness itself will directly cause death, and (ii) illness where the illness plus a voluntary omission to treat it will cause death. What distinguishes (i) from (ii) is how (ii) includes a voluntary omission as a contributing factor to death causation. If the central rationale for assisted suicide is autonomy-based then it is surely irrational to consider a voluntary omission as counting to exclude an illness rendered terminal from qualifying as a terminal illness. The presence of voluntariness could hardly lessen the permissibility of assisted suicide.

It may be replied that a chronic illness effectively rendered terminal through refusal of treatment remains, in reality, a chronic illness. It is not itself a terminal illness and therefore does not itself meet the terminal criterion of the qualifying ground. The addition of voluntariness does not serve to exclude an otherwise qualifying illness – the illness in se remains non-qualifying.²⁴ All this is true. Yet the positing of a terminal illness ground is hardly a matter of sheer stipulation. It has a rationale. A central part of that rationale is the near-term inevitability of death given the circumstances facing the patient. Death’s near-term inevitability serves to cast into relief (a) the patient’s tragic predicament and the consequent concern we ought to have for them (inclusive of concern for their stance towards their predicament), (b) their physical suffering and/or psychological angst as a proportionately large part of their remaining life, and (c) their remaining life as a proportionately small part of their existence. It is death’s near-term inevitability that justifies the terminal illness ground, not an arbitrary concern for terminal illness as one category of illness among many others. Now ‘death’s near-term inevitability’ can obtain in a case of principled refusal of treatment for chronic illness and can do so in a manner that equally engages motivating elements (a)–(c). One could argue that ‘inevitability’ here is merely a function of a decision to refuse treatment. Yet this point cannot be pressed by a proponent of an autonomy-based rationale for assisted suicide. Voluntariness is a value within that rationale, not a disvalue. In fact, it is a pre-eminent value therein. Inevitability brought about in part due to voluntary choice ought to be taken no less seriously than inevitability due to exclusively physical causation. And if voluntariness is given the respect due to it within that framework it must then also be admitted that voluntary decisions can be stable and steadfast. Indeed, acceptance of this is presupposed by all legalization of assisted suicide (and no extant assisted suicide law provides for medical dissuasion of a patient’s choice for assisted suicide). One consistent way of challenging the idea of ‘inevitability’ from within the autonomy-based rationale is to point out that the making of any particular choice is not inevitable. This is true on non-deterministic accounts of choice. But the occurrence of a terminal illness is not necessarily ‘inevitable’, either, and no legal provision of a terminal illness ground for assisted suicide requires the screening of terminal illnesses for avoidability. The relevance of near-term death being ‘inevitable’ for the terminal illness ground has to do with the relevant facts now obtaining, not the genesis or pre-history of those facts. A steadfast decision to refuse medical treatment is a relevant fact that can render death ‘inevitable’.

Matters are more contestable if the central rationale for assisted suicide is beneficence-based. (Here, for the sake of conceptual clarity, I presuppose an account of beneficence irreducible to and importantly distinct from the personal good of autonomy.²⁵) On the beneficence-based view assisted suicide can be in a person’s best interests aside from counting as an expression of autonomy. Perhaps the easiest way of conceiving this is the ending of suffering through assisted suicide: cessation of a patient’s suffering is good independent of the value of a patient’s choice to end the suffering. Now a beneficence-based approach to assisted suicide can take available medical treatments into consideration in a way that autonomy-based approaches cannot. Such approaches could thereby assert that a better treatment than assisted suicide is available (i.e. treatment for the chronic illness) and simultaneously deny that autonomy is the central good at stake in medical deliberation (and thus grant no dispositive value to the patient’s voluntary refusal of treatment for their chronic illness or their wish to avail of assisted suicide). So on the beneficence approach to assisted suicide a medic has strong grounds to be able to say that a patient’s refusal of treatment for their chronic illness does not render assisted suicide in their best interests compared to accessing (or continuing with) the treatment for the chronic illness. As such assisted suicide may not be medically (and thus morally) optimal or even permissible.

But it does not seem to follow that the medic could not rationally judge assisted suicide to be in the interests of the patient given a steadfast refusal on the patient’s part to access treatment for their chronic illness. A truly consistent beneficence approach that is carefully distinguished from an autonomy approach cannot reduce ‘best interests’ or beneficence to autonomy, but neither can it ignore facts occasioned by a patient’s legally protected refusal of medical treatment, nor can it suppose that a duty of care to a patient is terminated upon the patient’s choice to refuse one possible course of treatment. Given a scenario wherein a patient is committed to refuse treatment for their chronic illness, and given that in principle (as conceded by the beneficence rationale) assisted suicide can be in a patient’s best interests in alleviating suffering, it seems that a medic has good reason to think of assisted suicide as a medically beneficent option in cases of such ‘indirect’ terminal illness. Perhaps the medic might conceive of assisted suicide as in the patient’s ‘best interests’ given the narrowing of practically accessible options before them on account of the patient’s legally protected treatment refusal. Or perhaps the medic might more rationally judge assisted suicide to be not so much in the ‘best interests’ of the patient as a medically warranted (or at least medically permissible) interest-based compromise in light of the patient’s refusal of treatment. Either way, a patient’s right to refuse medical treatment makes it difficult for even a beneficence-based approach to assisted suicide to non-arbitrarily restrict assisted suicide access to ‘direct’ terminal illness only, and to thereby refuse access to chronic illness rendered terminal on account of a patient exercising a right to refuse medical treatment.

Conclusion

Oregon is often cited at parliamentary committees as a challenge to the cogency of the slippery slope argument against legalization of assisted suicide. While acknowledging the main reasons for the citation’s plausibility, the present paper argues that there are features of Oregon’s evolving practice that evidences an empirical slippery slope – albeit one with a hitherto relatively low gradient. Empirical evidence of a morally interesting trajectory is ascertainable from increases in each of the annual case numbers, the percentage of assisted suicide patients citing burden-based concerns, the margin of error in misprognoses of terminal illnesses, and the reporting of qualifying illnesses that are not normally (or at least predictably) terminal in character. Part-explaining these increases may be two aspects of normative slippage worthy of increased attention. One is slippage from a sort of voluntariness contemplated by the initial public justification for assisted suicide, one which intentionally did not include burden-based concerns in its account, to a now widespread assisted suicide voluntariness that is motivated by burden-based concerns. Potentially, the slippage may even be to a newly evolved public justification: to one that affirms burden-based voluntariness in assisted suicide provision. The other aspect of moral-logical slippage raised by Oregon’s practice is expansion of the terminal illness ground to accommodate chronic illnesses effectively rendered terminal due to a legally protected right to refuse treatment. There are reasons to think that this de facto expansion has occurred in Oregon. Even if evidence to the contrary is uncovered, this particular issue is an important and under-appreciated juncture in the moral-logical slippery slope argument. As such it warrants scrutiny. Is there a non-arbitrary rationale for restricting assisted suicide access to ‘direct’ terminal illness cases and excluding cases where a refusal of treatment effectively transforms a chronic illness into an illness where death is a near-term inevitability? I argue that such exclusion is not rationally justified under an autonomy-based account of assisted suicide’s permissibility and proves difficult to consistently (non-arbitrarily) maintain even under a beneficence-based account of same.

Acknowledgements

This paper has been improved by comments from Prof. John Keown and the journal's two anonymous reviewers.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.

Notes on contributors

Thomas Finegan

Thomas Finegan is the Chair of the Governing Board at the Anscombe Bioethics Centre. His previous publications have appeared in Bioethics, Journal of Medical Ethics, Medical Law International, among other journals.

Notes

1 Though assisted suicide differs materially from voluntary euthanasia, for argument’s sake I include the latter under the former (unless otherwise stated). This paper understands both to centrally involve deliberate medical co-operation in a competent person’s autonomous choice to end their own life.

2 An absence largely due to the self-regarding nature of voluntary decisions to proceed with assisted suicide.

3 USA (Oregon will be used as the most important state example in this paper), Netherlands, Belgium, Switzerland, Luxembourg, Colombia, and Canada.

4 Oregon: see below; Netherlands: Houses of the Oireachtas Joint Committee on Assisted Dying (2023b); Belgium: Michas (2023); Switzerland: Montanga et al. (2023); Luxembourg: RTL (2019a, 2019b); Colombia: Cook (2023); Canada: Government of Canada (2023) and BMJ News (2022b).

5 Switzerland has not legislated for assisted suicide and therefore has no grounds to expand or contravene.

6 Luxembourg has witnessed neither a de jure nor a de facto widening of grounds, though the legislative grounds – enacted in 2009 – were wide and subjective to begin with and include provision for advance euthanasia directives; Oregon: see below for reasons to think that a de facto expansion of grounds is there occurring; Netherlands: BBC News (2019) and The Guardian (2023); Belgium: Raus et al. (2021); Colombia: World Federation of Right to Die Societies (2022) and BMJ News (2022a); Canada: Coelho et al. (2023).

7 Providing figures for the first three years corrects against backlog bias.

8 New Zealand’s End of Life Choice Act 2019, based on Oregon’s Act, offers an interesting comparison. The first full-year official report into New Zealand’s assisted suicide figures indicates 328 deaths between April 2022 and March 2023 (Government of New Zealand 2023). In the first year of the operation of Oregon’s Act there were 16 deaths (Oregon’s population in 1998 was c. 3.28m compared with New Zealand’s current population of c. 5.2m).

9 An autonomy-based justification for assisted suicide is distinct from the legal permission established by the justification – a legal permission established by a particular justification may have a side-effect of giving immunity to practices or reasons not intended by, or even not consonant with, its establishing justification.

10 Which is not to suggest that it is morally unimportant. Unless one assumes a simple, unvariegated, absolutist account of autonomy’s significance, such that all expressions of self-regarding autonomy are morally equivalent regardless of variations in underpinning reasons and wider motivations, the normative underpinnings of voluntary choices ought to matter for moral analyses of a dedicated legal facilitation of those voluntary choices.

11 As Dr. Jeanne also pointed out, in Oregon patients can cite numerous concerns motivating a choice for assisted suicide. The top two concerns are loss of autonomy and loss of ability to enjoy life.

12 De jure expansion of the Oregon law has also occurred but not as regards expansion of the terminal illness ground. These expansions, well summarized by Jones (2023), are left aside for present purposes since they are not as morally important as the expansion of the ground(s) for access to assisted suicide.

13 Part of the issue is the difficulty in making accurate prognoses as much as six months out from the limit timepoint. See White et al. (2017). Yet a safeguarding issue also arises because there is an inherent difficulty in safeguarding against potential abuses of a terminal prognosis ground when the ground gives access to an intervention that works to shield the prognosis from scrutiny.

14 One might be tempted to think that Covid explains part of the increase from a longest duration average of 536 days for 1998–2018 to an average of 1384 days for 2019–2022. But the two longest durations were for the years 2019 and 2022, i.e. the two least affected by Covid.

15 The official report for 2022 ceased outlining illnesses that fall under the heading ‘other illnesses’.

16 From this condition to the last condition mentioned in this sentence the combined average numbers of assisted suicide deaths per annum across 2020–2021 was 8.

17 Prior reports included other conditions which qualified for assisted suicide and which are difficult to reconcile as predictably terminal, e.g. the reports for 2014–2016 included under ‘other illnesses’ Parkinson’s and Huntington’s diseases.

18 Citing a statement of Prof. Linda Ganzini, School of Medicine, Oregon Health and Science University made on March 16, 2017. Prof. Ganzini has published widely on Oregon’s assisted suicide practices.

19 Citing the responses of Mr. Craig New, a Research Analyst at the Oregon Health Authority with responsibility for, inter alia, ‘Death with Dignity statistics’.

20 Citing a different official source from within the Oregon Health Authority.

21 In practice there appears to be sufficient medical privacy to allow for treatment refusal of this sort to happen in a way that would not preclude access to assisted suicide. According to Dr. Jeanne (Houses of the Oireachtas Joint Committee on Assisted Dying 2023b), ‘Determinations related to disease treatment and, if appropriate, end-of-life care options are made between the patient and the physician. The law does not include any medical oversight or regulation distinct from what is done for other medical care … . The OHA [Oregon Health Authority] does not investigate whether patients met the Act’s criteria, nor how their diagnoses, prognoses, and treatment options were determined’. This (startling) admission suggests at least the potential for slippage concerning what is permitted by the overall regulatory regime in Oregon.

22 Further illustration of the question’s real-world relevance is provided by the recent development of the concept ‘terminal anorexia’. See Riddle et al. (2022).

23 I do not think that the right to refuse medical treatment is exclusively a function of patient autonomy, especially autonomy understood merely as doing-what-one-wills. Other considerations also seem to inform the right, e.g. the intrinsic personhood of the human body, the more qualified duty to positively help compared to the absolute duty to avoid intentional infliction of harm, practical feasibility and limited medical resources, etc.

24 Perhaps only up to a point, for an untreated chronic illness may produce a terminal condition. This scenario would be impossible to exclude from a terminal illness ground and illustrates a practical difficulty in excluding chronic illnesses from considerations of terminality.

25 Beneficence-based approaches to assisted suicide would logically seem to require moving towards non-voluntary euthanasia. Hence some in favor of medicalized killing proceed via a joint view: autonomy plus beneficence grounds assisted suicide. The argument here leaves all this aside. To the extent that it can be shown that the beneficence-based approach to assisted suicide has reason to think it dubious to restrict the qualifying grounds to ‘direct’ terminal illness only, then the reason(s) applies a fortiori to any joint view.