In their recent article ‘Social consequences of assisted dying: a case study’,¹ Rhona Winnington and Roderick MacLeod raise concerns about the societal implications of legalising assisted dying (AD) in New Zealand. They present the results of an interview they conducted with an individual whose family member had an assisted death in a country where it is legal. Based on this interview and other evidence, they argue that there is the possibility for broad “community and social discord”:

“…we must also consider the potential of fracturing of our New Zealand communities and broader social settings. Even if we can ‘fix’ such fractures after AD is legislated and implemented, we may not be able to return them to their former status, thereby changing the supportive and intimate nature of the New Zealand social landscape.”¹

Though they say little about what practical outcomes should follow from their paper, they say that “evidence from this study must be factored into the New Zealand debate before the referendum on the End of Life Choice Act in 2020”.¹ Since voters can only vote yes or no in the referendum, and since the specific legislation being voted on cannot be changed before then, the authors seem to believe that evidence from the study should serve as at least partial justification for voting against the motion to legalise AD.

The authors are correct that social consequences are important. No medical decision occurs in a vacuum, and just as in other end-of-life situations, AD will affect healthcare providers, families and others. Therefore, it is worth considering what these effects might be in case AD becomes legal in New Zealand. I encounter these effects regularly in my role as a clinical ethicist in Canada, where AD has been legal for five years. Winnington and MacLeod focus on three concepts to make their case: that legalising AD will lead to a “slippery slope” in which, among other effects, essential restrictions on accessing AD will be lifted over time; that there is stigma surrounding AD that will affect patients, families and providers; and that AD has the potential to cause what the authors refer to as a ‘contagion effect’. Despite the strong language the authors employ, they fail to provide compelling evidence for their conclusions. In what follows, I consider each of their points in turn.

Slippery slope

Winnington and MacLeod argue that legalising AD will result in a “slippery slope”, a term typically used to mean that something is acceptable in the beginning but, over time, leads to unacceptable results. The authors have two slopes in mind. The first is that patients will feel pressured to die:

“Despite the prevalence of right-to-die narratives that support those who are concerned about loss of dignity and quality of life, medicine (in collaboration with law) has maintained an unwavering stance that the right-to-die produces a ‘slippery slope’ effect, whereby some patients may be obligated to terminate their lives.”¹

The claim that either medicine or the law maintain an ‘unwavering’ belief in the slippery slope is false. Demonstrating the falsity of this claim requires looking no further than the conclusion of their own paper, where they say “There remains ongoing debate with regard to the slippery-slope effect”.¹ Though the authors give no citations here, a cursory look at the literature shows that there is considerable debate about the existence of any slope and what effect its potential existence should have on policy.^2,3

The claim that the law has maintained a similar unwavering stance is also false. The slippery slope concern is explicitly discussed in Carter v Canada, the unanimous Supreme Court of Canada decision that struck down the prohibition on AD:

“The trial judge, after an exhaustive review of the evidence, rejected the argument that adoption of a regulatory regime would initiate a descent down a slippery slope into homicide. We should not lightly assume that the regulatory regime will function defectively, nor should we assume that other criminal sanctions against the taking of lives will prove impotent against abuse.”⁴

In other words, the trial judge who originally heard the case found no compelling evidence that a well-regulated AD policy would lead to a slippery slope. After surveying the evidence for themselves, all nine Supreme Court justices agreed with her. Therefore, it is far from being the case that medicine and the law are unwavering in their belief that a slippery slope is inevitable.

The evidence that Winnington and MacLeod present to bolster their claim that patients may be obligated to terminate their lives is the following: “[Evidence for the slippery slope] can be seen in the Netherlands, where one in 30 individuals died by euthanasia in 2012 compared with one in 90 in 2002”.¹ The authors say that this increase “could suggest” that death is becoming a duty for some people. This is true. However, there is no reason to favour this explanation over the explanation that more people are receiving assisted deaths because more people are aware of it and believe that it is the right choice for them. It is an unjustified leap to a negative conclusion. As the Supreme Court of Canada justices admonish, “We should not lightly assume that the regulatory regime will function defectively.” In order to conclude that widespread coercion is taking place, we need some real evidence. When we look for it, we find empirical research showing that no slippery slope has occurred in the Netherlands.³

The second type of slippery slope Winnington and MacLeod are concerned with is the more traditional one in AD discussions: that, over time, stringent eligibility criteria will be loosened to the detriment of the vulnerable. They say that “Research suggests that globally, similar bills have initially restricted eligibility to those with terminal illnesses to make AD more palatable, but the eligibility criteria subsequently became more flexible”.¹

As a factual claim, this is correct. While increased flexibility is not inevitable—Oregon’s criteria have not changed since 1997—as Canada shows, some jurisdictions do change their eligibility criteria. Following the Carter ruling that a complete prohibition on AD was unconstitutional, parliament passed a bill permitting AD in certain circumstances. (That bill is similar in many ways to the one being voted on in New Zealand.) For reasons having to do with the Carter decision, instead of following the Oregon model, which requires that patients must be likely to die of their illness within six months, the parliament of Canada required that death must be “reasonably foreseeable”. Last year, the Superior Court of Quebec (a Canadian province) ruled that the reasonable foreseeability criterion is unconstitutional, and the Minister of Justice and Attorney General of Canada introduced a bill to amend the criminal code to get rid of the unconstitutional criterion.

The details of Canada’s legislation are less important here than the justification for the changes. Winnington and MacLeod take removal of the terminal illness criterion to be evidence of a slippery slope (ie, an unwelcome result). However, a superior court judge and the government of Canada take a different view. The case against a terminal illness condition is that an arbitrary time frame means that people who are experiencing the same level of unbearable suffering due to a medical condition will not qualify for relief from that suffering simply because their death is too far away. Far from this being a justified standard, it means that those people will suffer more. Since the point of AD is to ameliorate unbearable suffering, a terminal illness condition is at odds with a just system, not in line with it. Of course, one person’s good outcome can be another’s slippery slope, but we need to see the argument that removing the terminal illness condition constitutes slippage. The ethical argument points the other way.

Stigma

The second key concept Winnington and MacLeod discuss is stigma. One of their concerns is that patients seeking AD will either disclose to their families their intention to have an assisted death and experience a negative response (pressure to stop, guilt, judgement, abandonment). Another is that physicians and other healthcare workers will be stigmatised for participating in AD.

The authors are correct about the existence of AD stigma. As they note, stigma has been documented in research, and, anecdotally, I have seen instances of AD stigma directed at both patients (by families and healthcare staff) and providers. However, the authors once again paint a dire picture that is not warranted by the evidence they provide. Most importantly, they give no evidence that “communities and broader social settings” will be threatened by AD. Even if we grant that some families will be significantly strained by a family member getting an assisted death, the authors give no clear causal path from a fractured family to communities and the fabric of society being irreparably torn. Oregon is not falling apart. Canada is not falling apart. Switzerland is doing fine. Of course, the legalisation of AD in each of these places has required them to undergo significant change, but the burden of proof is on the authors either to produce evidence of social upheaval in these places or to explain why New Zealand is different enough to warrant a different outcome.

The are many prescriptions for stigma. Even if we grant the assertion that stigma will cause fractured families and communities, it does not mean that AD should be prohibited. Consider the stigma around being gay. Unfortunately, coming out as gay can come with many negative consequences, so, as a result, some gay people keep it a secret from their family. Others come out to their family and community only to experience ostracism, judgement, abandonment and, in the worst cases, threats to their safety. No doubt, the result is fractured families and maybe even communities. Nevertheless, the best prescription is not to try to stop the thing that is causing the stigma, but instead to address the source of the stigma and attempt to reduce it.

The same is true of AD. While many patients I have met are open with their families about their plans for AD, some have kept it a secret. (Of course, this is not unique to AD.) Unfortunately, the health system where I work has had to enact policies to protect patients and staff from the stigma around AD. I have seen no evidence that this stigma is threatening communities (and the authors offer none), but even if such evidence exists, it still would not be grounds for prohibiting AD. It might be grounds, but making that case would require an argument that the correct course of action is to restrict the act instead of addressing the stigma. In either case, since patients and doctors are aware of the stigma, there is no reason why people should not be allowed to decide what is best for themselves.

Winnington and MacLeod describe New Zealand as being at a tipping point from which there might be no going back, and they use stigma as potential grounds for continuing the prohibition of AD. They are wrong about both the facts and the values.

Contagion effect

The final concept Winnington and MacLeod discuss is what they refer to as a ‘contagion effect’. What they seem to have in mind is that increased exposure or awareness of AD will lead to more people using AD. For example, they say that “the provision of positive role modelling of AD practices may normalise or even promote this means of death unless assisted deaths are protected by rigorous legislation that supports those involved” before adding that AD is “now becoming the most prevalent mode of death for patients with cancer” in the Netherlands.

Suppose that by ‘contagion effect’ the authors simply mean a rise in the number of assisted deaths. If so, then using the negative term ‘contagion’ is poorly chosen. (Regardless of their aims, to describe AD in such terms during a global pandemic is irresponsible.) Instead, they clearly have something negative in mind. After all, if a restaurant opens and gets more popular over time, it would be strange to claim that the increase in popularity was evidence of ‘contagion’. Instead, the increase in popularity should be described as a value-neutral increase in use.

That more people are getting an assisted death in the Netherlands is only evidence that more people are using it. Similarly, the authors give the assisted deaths of three extended family members of the research participant as evidence of a contagion effect. But, once again, this does not support the ‘contagion’ hypothesis (where that means something negative) over the value-neutral explanation that people are finding out about AD and deciding to use it. Especially in a jurisdiction where AD was recently legalised, a year-over-year increase should be expected. This negative view of AD is also evidenced by the authors’ endorsement of the participant’s framing of AD as infectious. Since the authors give no indication that the participant’s relatives were coerced or manipulated, the contagion claim is unjustified.

They repeat the error: “[The participant’s concern] was consistent with Kheriaty, who observed that exposure to the idea of AD can lead others to seek such assistance”.1 Yet again, instead of evidence of ‘contagion’, this is only evidence of increased use. If the authors have an objection to AD on ethical grounds, they should say so and provide the argument. Otherwise, evidence that people are accessing AD, even after they find out about it from friends and relatives, does not constitute ‘contagion’. This move is problematic because the authors immediately switch from discussing increased AD to describing their belief that society is at risk of being fractured. But, as we have seen, the fracture thesis is not justified by the evidence Winnington and MacLeod provide.

Finally, the authors directly undercut their justified concern about stigma with their claim about contagion. Imagine telling your loved ones that you were seeking an assisted death, only to be told that doing so is the equivalent of spreading a disease. If the authors are genuine in their concern about stigma, they should abandon their stigmatising language.

Conclusion

I have shown that none of the points made by Winnington and MacLeod support their conclusions of social discord and fracturing. This is not to say that the transition to legalised AD is pain free. Stigma exists. Providers have to figure out how to navigate the new landscape. Families are affected. These will be important topics of research for many years, and we will surely continue to figure out how to do it better. Some problems will occur, but dystopic claims require strong evidence. Winnington and MacLeod fail to provide it.