Definitions

A significant problem in the assisted dying debate is terminological ambiguity.12 In this paper, we adopt the definitions recommended by the European Association for Palliative Care Ethics Task Force:

Euthanasia is killing on request and is defined as: A doctor intentionally killing a person by the administration of drugs, at that person’s voluntary and competent request. Physician-assisted suicide is defined as: A doctor intentionally helping a person to commit suicide by providing drugs for self-administration, at that person’s voluntary and competent request.13 (p 98)

We use ‘assisted dying’ as a generic term to encompass euthanasia and physician-assisted suicide, in line with language used by Goldney,14 Hendry et al 15 and in the End of Life Choice Bill 2017.4 It should be noted, however, that the term ‘euthanasia’ is frequently used in a similarly broad sense in media and public discourse.

High-profile media cases

Several assisted dying prosecutions have received significant media attention in New Zealand.16 17 A particularly high-profile case was that of Lecretia Seales in 2015.18 Lecretia Seales was a lawyer diagnosed with an inoperable brain tumour. She sought confirmation of whether her doctor would be acting unlawfully if she administered, or provided a prescription for, a fatal drug to Ms Seales. She also asked the court to determine whether her inability to access assisted dying was a breach of her rights under the New Zealand Bill of Rights Act 1990. In an urgent judgement delivered shortly before Ms Seale’s death in June 2015, Judge Collins acknowledged that Ms Seales had ‘selflessly provided a forum to clarify important aspects of New Zealand law’ but determined that the changes she was seeking could only be made by Parliament.18

Previous research in New Zealand

A systematised review of the past 20 years of New Zealand research found support and opposition for assisted dying among New Zealanders is stable across time; 67.9% and 13.8%, respectively, with 16.1% unsure.11 The opinions of health professionals on assisted dying appear somewhat more divided as represented by professional bodies.19–22 Research on the attitudes of New Zealand doctors and nurses towards assisted dying found that 58% of doctors were opposed to a law change while 67% of nurses were supportive of a law change.23

Previous qualitative research into New Zealand citizens’ views on euthanasia has primarily used written and phone surveys, interviews with individuals and small focus groups. Malpas et al interviewed older New Zealanders opposed to euthanasia to explore their perspectives. Dominant themes included personal experiences with healthcare and death, religious beliefs, concerns about the potential slippery slope consequences of legalisation and worries about abuse.24 Related research on why older New Zealanders support medical practices that hasten death found that fear of being a burden on others is a significant factor in evaluating end of life issues for many healthy older people.25 Māori perspectives towards physician-assisted dying are more nuanced—encompassing different understandings of assisted dying, the power held by kaumātua (elders), whanau (family) and medical professionals in end of life matters, significance of kawa (protocol) around death and dying, the place of whanau in the dying process and the importance of wairua (spirituality).26

An analysis of New Zealand’s public discourse on assisted dying examined nursing, legal and media texts from 2002 to 2004, when assisted dying gained a high profile with the introduction of Peter Brown’s Death with Dignity Bill and the prosecution of Lesley Martin.17 Lewis explored in detail two major competing discourses: sanctity of life discourse and right to die discourse.27 The most recent analysis of personal narratives of seeking a good death and the right to die and shared on social media found patient choice is a constructed illusion within medical discourse that positions patients as deviant.28

Social media

Social media facilitate communication between individuals across societal, cultural, ethnic and national boundaries, representing ‘interplays of worldviews, values and concepts’ and offering new possibilities and challenges to researchers29 (p 150). It has transformed public responsiveness to events, press releases and media announcements. Rather than days or weeks to write letters to newspapers, or queueing on the telephone to talk with talkback radio hosts, the public can respond within minutes and disseminate information through their social media networks on a truly viral global scale.30

The extent to which social media influence public debate is contested.29 30 The unfiltered nature of social media postings means that profundity and nonsense, fabrication and truth sit alongside each other in a bewildering and fragmentary collage.30 Bouvier suggests that online debates of sociopolitical issues tend to frame events and contributions into pre-existing personal perspectives and ontological frameworks.29 This means that the degree to which social media contributors actually demonstrate engagement and openness to perspectives contrary to their own is variable—highly divisive and polarising issues may not result in meaningful exchanges because supporters of one or another position use social media to correspond with each other rather than to conduct meaningful debates with opponents31 (p 44).

Politicians and organisations pay close attention to social media; in our own data set, we found contributions from politicians and public figures sitting within threads alongside the posts of ordinary citizens. Shirky used the example of a social media storm that mobilised citizen protests and caused a consequent change of political action in the Philippines to argue that strong public sentiments can influence the decision-making of politicians.32

Social media cannot be ignored, whether or not one chooses to participate in it. We suggest that social media, like more traditional media, reflect and refract public opinion on topical issues such as assisted dying in complex and highly nuanced ways. Assisted dying is a polarising and emotive issue with strong moral and ethical underpinnings. In this article, we show that the debate in the public social media space includes not only perspectives on what the law should be, and what is the right decision regarding the End of Life Choice Bill, but broader issues about who should make this decision and how it should be made on behalf of citizens, and the ways in which contributors to the debate legitimate their authority to speak to this issue.

Ethical considerations

Ethical approval was not deemed necessary for this project because the data were within the public domain. However, with researchers turning their attention to the internet as a rich source of data, internet ethics has become highly salient. Zimmer claims that, far from being diminished in cyberspace, issues of consent, privacy and autonomy become even more important35 (p 324). Internet research ethics highlight the ‘loaded nature of terms such as ‘public’ and ‘private’ and the difficulty of applying them to the online world’36 (p 323). Conventional binary distinctions between ‘private’ and ‘public’ do not apply in cyberspace, and individuals can engage with social media holding expectations of privacy.37 38 One key ethical issue is that internet users are not expecting that their activities will become part of a research data set; and this unauthorised secondary use of posts could also be considered a violation of privacy.35 38 Another is the issue of potential traceability of social media excerpts published in research papers and reports. How then should researchers approach this field?

Researchers in this space have been described as covert lurkers in that they do not disclose their presence and they do not gain consent from contributors to be present or to use their material.38 One guide to ethical research of social media recommends researchers consider the ethical principles of respect for persons, justice and beneficence on a case by case basis for each project. This includes an obligation to protect the vulnerable and to balance the rights of internet users and those of researchers. It also includes considering the extent to which internet data might be considered to be sensitive, personal and private, or public and freely available for reposting and analysis.37

Moreno et al note that contributors to social media can place limitations on the accessibility of posts using the privacy controls for profile owners—certainly the case for Facebook and LinkedIn.39 von Benzon argues that framing contributors to social media as vulnerable, ignorant and naïve is a highly paternalistic stance that denies contributors’ agency and diminishes their autonomy.38 She and Bruckman suggest that contributors should be recognised as authors in the same way that professional authors of other online and publicly availably works are.38 40 von Benzon supports her argument by citing examples of international prosecutions against individuals who have used social media for hate attacks, thus recognising their agency38 (p 5). Courts have so far upheld the stance that users of social media cannot expect privacy in materials that are posted to social media and which are available to the public.39

Our data comprised selected social media. We excluded specific interest group blogs such as that of End of Life Choice NZ or Care Alliance, and limited our sources of data to sites which were unequivocally situated in the public domain and where postings are clearly for public consumption. Following Pedersen and Lupton,41 the data we collected were not elicited by researchers, nor commented on, were freely available to the public online, identities are not revealed, and we assume that contributors were aware that fora to which they posted were public and content available to non-members. We have not collected personal information about contributors although we are aware that postings may be traceable if searched for.35 39 As von Benzon describes, we view posts as social commentary and therefore a legitimate source of secondary data.38 Digital space is used as a means of accessing information about health, illness and medicine and represents a space in which cultural understandings and practices are contested.

Data collection

We searched our selected sources using the keywords Euthanasia OR ‘assisted suicide’ OR ‘assisted dying’ OR ‘aid in dying’ and date range 8 June 2017 to 2 August 2017. This period of time was chosen because it followed the selection from the Parliamentary Ballot and publicisation of the Seymour Bill, and also provided parameters on the data set. Use of these search terms automatically included texts that mentioned ‘voluntary euthanasia’, ‘medically assisted suicide’ and ‘medical aid in dying’. Where comments were present, these were included for analysis.

A search on news database Factiva and further Google site searches were performed for documents from NZ Herald, Dominion Post, Otago Daily Times, Waikato Times and The Press. These publications were selected based on readership and geographic representation across New Zealand. Google site searches were also performed for written texts from One News, Newshub, Māori Television, Radio New Zealand and Newstalk ZB websites. Google site searches were performed for documents from new media sites thespinoff.co.nz, thewireless.co.nz, noted.co.nz and newsroom.co.nz. Political blogs were selected for inclusion in consultation with an expert contact in the New Zealand blogosphere. Google site searches were performed for Whale Oil, Kiwiblog, The Standard, The Daily Blog, Public Address, Pundit and E-Tangata.

A Google search was performed in site: www.reddit.com/r/newzealand/. Due to the nature of social media search functions, we employed alternative strategies to collect data on Facebook and Twitter. For Facebook, Google site searching was performed on the Facebook pages for NZ Herald, Stuff, Māori Television, One News, Newshub, The Spinoff and The Wireless using the keywords described above, and eight relevant posts were identified within our date range. Comments were included for analysis. For Twitter, it was not possible to accurately collect only New Zealand tweets for the entire date range, as most Twitter users disable location services. The following searches were used in order to return only New Zealand content:

• Tweets with ‘euthanasia bill’ OR ‘end of life choice’ from 8 June 2017 to 2 August 2017.

• Tweets on the popular New Zealand political hashtags #nzpol and #nzqt, or tweets mentioning the Bill’s sponsor @dbseymour, that included our keywords within our date range.

• Tweets that included our keywords and were sent on the date of a focusing event that generated local discussion about assisted dying, which were as follows:

  • Selection of the End of Life Choice Bill from the ballot on 8 June 2017.

  • Publication of a Colmar Brunton poll on 14 July 2017.

  • Release of the Health Committee report on assisted dying on 2 August.

All Twitter search results were screened by Isabelle Lomax-Sawyers to identify tweets by New Zealanders for inclusion. This resulted in 1256 pages of text, more than 463 000 words of data.

Summary of data sources

Analysis

Coding of data was completed using ATLAS.ti by Isabelle Lomax-Sawyers. A small sample of the data was coded independently by all researchers to establish analytic concordance, and subsequent thematic development and interpretation of findings were discussed over several team meetings. We used a pragmatic inductive method of analysis,42 categorising and grouping initial codes into overarching themes in two consecutive processes. Our emergent themes aligned with Lupton’s textual and contextual thematics of discourse analysis which are described throughout.33

Theme 1: how should assisted dying be debated?

The first theme, pertaining to what Lupton referred to as the textual components,33 on the ways in which assisted dying is debated is presented in table 1. We identified two subthemes. We labelled the first subtheme ‘meta-discussion’ because it encompassed several threads by social media contributors that appeared to discuss the rules for discussing the topic itself. These included the validity of this topic in the contemporary New Zealand sociopolitical context relative to other issues such as child poverty and homelessness, and competing perspectives of progressiveness—that like other progressive societal reforms, the legalisation of assisted dying is inevitable in a just, humane and compassionate society that upholds principles of autonomy. Another category of the meta-discussion was the actual content of the debate. This referred to what the terms used (voluntary euthanasia, assisted dying, physician-assisted suicide) actually mean. It also referred to the supporting moral and ethical frameworks used to justify a contributor’s position, such as Christian, humanitarian and philosophical arguments. This was linked to the next category in this subtheme which concerned who had authority to speak on this issue (and more importantly perhaps, who should carry greater persuasive weight). Should the opinions of the terminally ill and their closest kin, the medical profession, the legal profession, the pastor or ethicist carry more moral authority? The final major category concerns the factors that influence the debate on assisted dying. This related to the framing of assisted dying by interest groups, politicians and media reports, as well as to the influence of socioeconomic and other demographic factors on shaping the public’s views.

The second subtheme concerns the persuasive rhetorical devices that were being employed by contributors. These included reference to comparators such as the historic Nazi Germany’s euthanasia and eugenics programme, and social attitudes and legislation on animal euthanasia, abortion and suicide. There was a notable use of humour in many contributions, as well as use of commonly used metaphors such as ‘the slippery slope’, ‘thin end of the wedge’ and the use of the term ‘murder’ as a synonym for assisted dying.

Theme 2: what should the law be, and how should we decide?

Table 2 presents both textual and contextual elements specific to the Seymour Bill and the decision-making processes through which it will pass. This theme examines the motives assigned by contributors to the Bill, and consideration on what a law decriminalising assisted dying should address, with consideration given to eligibility criteria, and safeguards. The first subtheme address what the law is for. This covers two categories. The first is that the Bill would merely formalise a covert clinical practice of easing a dying person’s passage through morphine overdose. The rule of double effect has been used to justify this in medical ethics: in brief, where the intention is to relieve pain and not to cause death, then no criminal act has occurred, despite the patient’s death from overdose, because the death is unintended. The second category is that the Bill is driven by hidden motives of which the public are largely unaware, such as the need to reduce the costs of ongoing palliative care of patients with terminal illness by offering euthanasia. This category also links to theme 1 in many contributors’ suspicions that a covert motive for introducing the Bill is population management through eugenics and a means of removing dissidents and problematic citizens. The second subtheme revolves around contributors’ concerns for rigorous eligibility criteria for assisted dying (primarily cognitive capacity for decision-making), the mechanism by which dying is assisted (doctor administered, or patient administered) and appropriate setting. The third subtheme was that of the process by which any decision to change the current (il)legal status of assisted dying should occur. Should the principle of democratic majoritarianism be upheld, and if so, should this be via opinion polls, petitions or referenda? Included in this category were references to the then upcoming government election and the degree to which the assisted dying debate might be manipulated by constituents seeking election.

Theme 3: the consequences of decriminalising assisted dying

Citizens’ concerns about the potential consequences of legalising assisted dying are presented in table 3. This theme reflects the contextual components of our discourse analysis. There were two primary subthemes: individual and broader social consequences. Individual consequences included categories of involuntary death, premature death and the quality of death. Contributors discussed the likelihood that legalising assisted dying could lead to individuals being coerced into requesting assisted dying, or being euthanised against their will. Another potential consequence concerned mistaken terminal diagnosis which could lead to an individual choosing assisted dying unnecessarily. The third category in this subtheme was that potential for improved dying experiences for a dying individual and their family because assisted dying would (1) allow the terminally ill person to exercise control over the manner of their exit from life, and (2) reduce the occurrence of ‘bad deaths’ where a terminally ill individual dies in severe pain. The second subtheme, broader social consequences, concerned the ways in which assisted dying might threaten the relationships between individuals and their healthcare professionals, and between individuals and their society by weakening social bonds and legitimating suicide for seriously depressed individuals. A second category was practical implications with contributors musing over the health resource implications for assisted dying, whether assisted dying might become the solution for the escalating costs of aged care and overpopulation and whether New Zealand might become an international destination for euthanasia tourism. Public health considerations comprised the third category in this subheading; including how harm prevention and suicide contagion would be managed, as well as how inequity of access to healthcare and aged care might impact on who would have access to assisted dying and what the potential harm/benefit analysis might be for patients.

What does our study add?

Formal submissions to parliamentary committees frequently represent dominant societal and perhaps conservative perspectives on assisted dying because they represent those who have the necessary social and cultural capital to compose and submit a formal document to the relevant Select Committee. This might explain the overwhelming opposition to assisted dying reported by the Parliamentary Health Committee.1 At the other end of the spectrum, opinion polls offer a snapshot of public opinion and these indicate strong public support for the legalisation of assisted dying.8 43 Opinion polls have been considered problematic as reliable sources of data because they are limited to those who have landlines which represents a decreasing proportion of the population in New Zealand as many people rely on mobile cellphones.44 45 Opinion polls conducted by telephone also exclude people who cannot afford a landline thus excluding the perspectives of the most socioeconomically deprived citizens. In opinion polls on assisted dying there are also issues of framing effects in the questions asked, and a lack of clarity over the various terms, euthanasia, assisted dying, hastened death and medical aid in dying.46 47

The barriers to participating in public debate on social media are fewer than for formal consultation processes. Social media can be contributed to by anyone with a cellphone although the same caveats apply to the participation of individuals who do not have a cellphone or the social and cultural capital to participate in social media. While social media offers a forum for many individuals’ unfiltered voices to be heard that otherwise would not be, it is difficult to ascertain the social demographics of contributors from their posts to social media. While any individual with internet connectivity can contribute to social media, we postulate that contributors are likely to represent New Zealanders who are social media literate—and may therefore reflect perspectives of younger, educated, middle classes. This may or may not be the case in other countries. While ordinary people might be more visible in social media, it has been noted that social media debates often follow frameworks laid out by traditional sources of information and dominant discourses because trending topics are often based on mainstream media’s breaking news.48 Our own methodology took advantage of this pattern. Motivations for using social media also vary—while some use its connectivity to contribute to public debates on social and political issues, others use it for entertainment, or social relationships.29 Zimmer notes that social media participants cannot be representative of a larger population and may not reflect the complexity of information contributors post on social networking sites.35 Another consideration is that those who are motivated to contribute to social media debates on high-profile topics may do so because they have a strong investment in the subject under discussion, and conversely that those who are neutral on the subject are not motivated to contribute—although this is likely to also be the case for research involving voluntary participation. Those motivated to contribute to social media need not post a comment, but can comment in what Bouvier describes as a passive way by ‘liking’ or using an emoji option.29

Given that the legality of assisted dying is dependent on a vote in Parliament, it is perhaps little wonder that media representation of this polarising issue has focused on the binary of ‘for’ and ‘against’ positions. Our findings reveal the frequently complex textual and contextual elements underpinning the positions of public contributors to selected social media. This adds another layer of understanding to what it is that ordinary New Zealand citizens see at stake in legalising assisted dying. While individuals who oppose the proposed law change can legitimate their position using well-established moral discourses from the dominant social institutions of Christianity, law and medicine, as well as catastrophic tropes appealing to history (such as Nazism and eugenics), and dystopian consequences (such as elimination of citizens who are a drain on health resources), those who support the proposed law change frequently legitimate their position against their own observations of the ‘bad’ deaths of loved ones and their own personal experiences. They appeal to humanitarian principles of compassion, alleviating or not prolonging suffering, or extending life in futile cases, using comparative tropes such as the compassionate euthanasia of animals.

Many contributors expressed frustration with the politics of representation—revealing viewpoints on the fiduciary relationship between citizens and the State. For those opposed, appeals were made to trust the traditional moral authority of the State and its governance apparatus as wards of its citizenry to act in its citizen’s best interests. For those in support, appeals were made to democratic and utilitarian principles underpinning the State’s responsibilities and responsiveness to its citizens, and that of politicians to their constituents. At issue here were what and whose values should be taken into account and/or privileged among multiple and contested traditional (Christianity, Hippocratic), modernist (progressive, cynical, antiauthoritarian) and neoliberal (individualist, responsibilist) discourses on authority, power and autonomy. This has relevance to all western and neoliberal jurisdictions. The former perspective carried the underlying assumption that citizens are not best positioned to make life and death decisions concerning their own lives, while the latter perspective suggests that it is appropriate for citizens to make life and death decisions concerning their own lives. Both positions ask, to whom does a citizen’s life belong? The State, God, the medical fraternity, or the citizen?

Bouvier suggests that the internet and social media illustrate a diminishing of formal authoritative information and knowledge from powerful disciplinary blocks, and increasing scepticism in the form of challenges to dominant discourses and hegemonies.29 While social media illustrates the interaction of a wide range of citizens over trending topics, it is unclear to what degree of dialogic engagement is really occurring in these threads. Are contributors truly open to others’ perspectives? Exercising agency through clicktivism? Or are they primarily engaging in the production of self-identity through their posts as Murthy suggests?48 Similarly, to what degree can social media influence political and parliamentary decision-making? Do contributors have any expectation of wider influence?

Limitations and strengths

Due to limitations of the search functions on Facebook and Twitter, it is likely that we missed some data within the date range. Also our conservative inclusion criteria regarding location likely resulted in an underaccounting of social media posts. Our analysis was limited by the researchers’ language skills to English language texts, and cannot be taken to reflect the debate on assisted dying within Māori, Pacific and migrant communities. This research was an opportunistic study of public opinion on Seymour’s End of Life Choice Bill following its introduction to the New Zealand House of Representatives in June 2017. The presentation to the House of submissions to the Parliamentary Select Committee has been rescheduled from later in 2018 to early in 2019. It is possible that during this extension, public opinion as evident in social media might shift.

Although we cannot make claims as to the numbers of contributors (few or many), a key strength of this study is the volume and range of the corpus we collected, representative of discussions occurring in the public domain independent of the formal parliamentary submission process for the Seymour End of Life Choice Bill.