How Does Medical Education Affect Students’ Views of Euthanasia/assisted Dying? A Mixed- method Study


 Background Previous studies report a majority of the general public support euthanasia/assisted dying (EAD), while a majority of doctors are opposed. In considering policy decisions about EAD, some may discount the views of doctors because they take them to be based on personal values or tradition, rather than reasons that the general public might share. One way to explore this notion is to examine whether medical students’ views change during medical education. The objective of this study was to learn how New Zealand medical students view EAD and whether their views change during medical education. MethodsAn on-line survey of undergraduate medical students was conducted. Quantitative data was analysed using unadjusted and multiple logistic regression. Thematic analysis was conducted with the qualitative data.Results A total of 326 students replied to the survey. The overall response rate was 28%. 65% of 2nd year students were supportive of EAD, compared to 39% in 5th year. The odds of 5th year students supporting a law change compared to 2nd year was 0.30 (95%CI 0.15-0.60). The predominant themes found in the qualitative results indicate that medical students support or oppose EAD for reasons similar to those found in the wider debate, and that their views are influenced by a range of factors. However, several at all year levels cited an aspect of medical school as having influenced their views. This was mentioned by participants who were supportive of, opposed to, or unsure about EAD, but it was the type of influence most often mentioned by those who were opposed. ConclusionsThe quantitative findings suggest that aspects of medical education made students less likely to support EAD. The qualitative findings indicate that this is not related to a single educational experience at Otago Medical School but a range. Together, these findings suggest that the lower support found among doctors is in part related to medical education and medical work rather than age, personality, or social context. They indicate a practical basis to ethical and professional formation that should be considered in curriculum design, policy decisions and when interpreting public opinion surveys on this issue.


Introduction
Euthanasia or assisted dying (EAD) [1] is a long-standing and highly contentious issue in medical ethics.
Internationally, numerous studies have consistently reported majority support for EAD among the general public but minority support among doctors (1,2). A recent study synthesizing the evidence on public opinion in New Zealand calculated a weighted average of 68% support for EAD (3). In comparison, a study of New Zealand general practitioners found 41% (3). One study of Australian doctors reported that 75% were opposed to EAD, while 87% would be unwilling to personally participate in the practice if it became legal (4). Other studies show that the doctors most opposed are palliative care specialists (5)(6)(7).
In recent years a number of countries and states have legalized various forms of EAD (1). Currently, EAD is illegal in New Zealand, but in September of 2020 a binding referendum was held on an Act that would make it legal. This Act received majority support and will come into force in November 2021. In a democracy, the views of the general public should always be considered in deciding policy, and are directly determinative in a binding referendum. Furthermore, given that the healthcare system places a high value on patient autonomy, the views of patients, i.e. the general public, should be taken seriously in considering which options are available. At the same time, it is important that policy decisions are informed by people with expert knowledge of the issue -in this case doctors -as they may be aware of implications that a policy may have which non-experts do not see, and it is possible that if the general public knew what doctors know, they may also be less likely to support EAD. Then again, the views of experts may be discounted if they are thought to be based on reasons that are not shared by the general public. It is possible, for example, that doctors are more likely to be opposed to EAD because they are more religiously inclined (8,9), or less ethnically representative of society (9).
One way to examine whether doctors' views of EAD are based on their individual values or reasons that could be shared amongst the general public is to examine the views of medical students over time.
Medical students are in a transitional state between the general public and doctors. Exploring their views, and whether they change as they progress through medical training, may illuminate how doctors' views on EAD are formed. In addition, it may indicate how the future workforce of doctors will be affected by EAD legislation, for example, by indicating whether more or less doctors will consider the practice ethically troubling, and how many may be willing to participate in the process.
Internationally, studies of medical students' views on EAD report a variety of results. These studies are highly variable in design and thus are di cult to compare directly. Several indicate that medical students are more supportive of EAD than doctors (10,11), and multiple studies nd majority support among medical students (South Africa, Munich, Switzerland) (7,12,13). In contrast, two studies from Poland and the UK found a majority were opposed (14,15). One study surveying medical students at an Austrian university at three time points suggests that the nding of increased support over time re ects the increase in support in Austrian society (16). Societal factors may contribute to the variability in the levels of support found across countries. For example, the lower level of support in Poland is possibly a re ection of the importance of Catholicism in Poland, which is traditionally opposed to EAD. Religious belief has been consistently associated with lower support for EAD (1). In comparison, the higher level of support in Switzerland may re ect a greater cultural acceptance, connected to the fact that assisted suicide has been legal and practiced there for many years. A comparative study of Belgian students reported that philosophy students and law students are more likely than medical students to 'unconditionally accept euthanasia' with 56% and 47% respectively, compared to 31% for the medical students (17).
A small number of studies in other countries have investigated how medical education affects student views. A 1995 study at the Puerto Rico School of Medicine found that 83% of 1st year students were supportive of "alleviat[ing] suffering when they cannot treat successfully an illness and death is imminent" compared to 61% of 3rd year students (18). A survey undertaken by Clemens et al, distributed among German medical students, found that as the students advance through their studies support for EAD drops from around 45% in 3rd year to approximately 35% in 6th year (19). A similar change was found in a study of medical students at the Chinese University of Hong Kong, which found that support for euthanasia dropped from 50% in 1st year to 23% in 5th year (20). In contrast, a survey of rst and nal year medical students at an English Medical School, designed to investigate how religion affects students' views of end-of-life decisions, found that being a nal year student was weakly associated with a greater likelihood of agreeing with EAD (15). However, this study found that the majority of students were opposed to EAD at both stages of learning. Overall, these studies appear to show that medical education makes people less likely to support EAD.
[1] Disagreement over which terms should be used to describe particular actions is part of the wider debate of this issue. 'Euthanasia' is the more traditional terms, while proponents often prefer the term 'assisted dying', or variations of this. Throughout this discussion we use the abbreviation EAD to refer to the range of actions that might be considered within the debate, including administering a drug with the intention of ending a person's life, and supplying a person with the means to end their own life. The survey question that we used speci ed a particular form of action (see below).

Methods
The study used an online survey and subsequent follow-up interviews with 30 students to learn what in uenced their views on EAD and whether they have changed during medical education. In this paper we report the students' level of support and opposition as indicated through the online survey, the reasons they gave for their answers, and experiences that they identi ed as in uencing their views. The results of the interviews will be reported in a subsequent paper.

Participants
The Otago Medical School program is six years long and includes a competitive Health Sciences First Year (HSFY), two years of Early Learning in Medicine (ELM), and three clinical years of Advanced Learning in Medicine (ALM). Around 70% of students enter the program from HSFY, usually following secondary education, while the remaining 30% of students enter ELM after completion of an undergraduate degree or through an 'Alternative' category, comprising older applicants from a range of backgrounds, such as other health related professions (21).
All medical students at Otago Medical School during the end of the 2018 academic year were invited to participate in the research. The listed year categories are for those who have completed that speci ed year, due to the timing of the research. Those who graduated at the end of 2018 were excluded because they ceased being students while the study was undertaken and so could not be contacted through the University system. A Qualtrics questionnaire was distributed via student email addresses and o cial class Facebook groups. A prize draw incentive of two $50 supermarket vouchers was offered. Participation was voluntary and anonymous, though to participate in the draw respondents were asked to provide their student ID number. Two subsequent reminders were sent, spaced 2-3 weeks apart.

Questionnaire
The questionnaire was constructed to learn whether the students support, oppose or are unsure about EAD, what informed these views, and the reasons behind them (see appendix). Demographic information about participants' ethnicity, religion, age, gender and prior training was collected as part of the survey.
The rst question in the survey (Q1) was as follows: 'Do you think the law in New Zealand should be changed to allow doctors, under certain circumstances, to provide or administer a medicine to a person, at their voluntary and competent request, that will bring about their death?' To minimise the problem of framing bias, this question was taken from a recent 'Citizens' Jury' project concerning New Zealanders' views on EAD (22). As part of the design of this Citizen's Jury, the wording of this question had been discussed and approved by a steering group of four recognised experts with a mix of views on whether EAD should be legal (22).

Analysis
Associations between views on EAD and year of medical education were investigated using both unadjusted and multiple logistic regression. The multiple regression model adjusted for potential confounders including gender, Māori ethnicity, and religiosity. Binary views on EAD were measured using two different outcomes: Yes versus Unsure/No and No versus Unsure/Yes. Variance in ation factors were investigated for the presence of collinearity. Stata software version 16.0 was used for all statistical analyses. The two-sided signi cance level α=0.05 was speci ed for all statistical tests.
Responses to questions about the reasons behind participants' responses to question one, and what particular experiences had in uenced their thinking on this issue (question 2, section 2 in the survey), were analysed using the general inductive approach (23). Independent parallel coding and creation of initial themes was developed by LN, with SW as the second coder. Next, axial coding was used relating concepts to each other and collating the codes into substantial themes by LN and SW. Continuous revision of the themes were re ned by LN, SW and KSH. Counter-arguments about the delegation of themes led to continued discussion with LN, SW, and KSH until concordance was achieved. The themed responses were then grouped into years 2 and 3 (ELM) and years 4 and 5 (ALM), to identify whether students with clinical experience offered different responses.

Results
Quantitative results 326 of the 1168 students contacted responded to the survey, giving a total response rate of 28% (326/1168). The response rates for the earlier years are higher than the latter years (2 nd 35%, 105/x; 3 rd 31%, 91/x; 4 th 22%, 66/x; 5 th 23%, 64/x). Independent demographics (Appendix , Table 4a) are similar across the years with respect to ethnicity and religiosity. Gender distribution is similar across 2nd, 3rd, and 4th years but in the 5th year there is a relative increase in the number of male respondents. Obviously, as there is a progression through the years of medical school education, there are more respondents in the 25+ age group category. When compared to the rest of the cohort (Appendix, have similar proportions. The largest proportional difference for gender is 3 rd year, where there is a 9% difference in proportion of male participants. If there is any other sampling bias, it would be reasonable to expect this to apply across year levels, and so the comparison between those year levels would still hold. Information about the religious a liation of students is not kept by the Medical School, but comparison can be drawn with the general population. 68% of participants reported having 'no religion' compared to 42% of the general population in 2013, while 26% identi ed as Christian compared to 49% (24). This indicates that the participants were less religious as a group than the general population (which is perhaps unsurprising as the general population includes older people who are more likely to be religious), but it is unknown whether they were more or less religious than the rest of the cohort.
The responses to Q1 (section 2) are provided in Table 1 and Figure 1. Table 1 shows an overall 56% (184/326) support for the legalization of EAD, while 22% (71/326) are unsure, and 22% (71/326) oppose. Over the four years of medical education support decreases from 65% to 63% to 52% to 39%, while opposition increases from 14% to 20% to 24% to 34%. Table 1: Participant year levels and answers to question 1 (section 2) Q1: 'Do you think the law in New Zealand should be changed to allow doctors, under certain circumstances, to provide or administer a medicine to a person, at their voluntary and competent request, that will bring about their death?' Year For the support analyses (Table 2a), the univariate and adjusted models produced similar estimates and inferences: both religiosity and year of study were associated with a decreased odds of supporting EAD. Speci cally, after adjustment for the potential confounders, medical students in their 5th year of training had 0.30 (CI: 0.15, 0.60; p=0.001) and 0.33 (CI: 0.16, 0.68; p=0.003) times the odds of supporting compared to students in their 2nd and 3rd years respectively.
The opposition results (Table 2b) were also predominantly comparative between the unadjusted and adjusted logistic regression models with religiosity and year of study again being presented as signi cant variables. After adjustment, medical students in their 5th year of training were associated with a 260% (OR=3.60; CI: 1.53, 8.50; p=0.003) increase in the odds of opposing EAD compared to 2nd year medical students. Although the univariate model suggested that 5th Year medical students were also more likely to oppose than their 3rd year counterparts (OR=2.12; CI: 1.02, 4.41; p=0.043), this signi cant association no longer remained after accounting for all covariates (OR=2.08; CI: 0.90, 4.83; p=0.087).
There were no signi cant associations between gender or ethnicity and support or opposition for EAD.  Eighty ve percent of participants provided answers to the questions about the reasons for their answer to question one, and 69% described experiences that have in uenced their thinking. See Tables 3 and 4 for the themed reasons and experiences respectively, and examples of each divided into year groups. "I do not support a law change due to lack of trust in the system. I see far too many opportunities for health care practitioners (not just doctors) to become 'trigger happy' or forceful about euthanasia. I also envision situations wherein patients feel pressured to accept this as an option. There is also quite a lot of room in the system for error and/or abuse of elders and those with disabilities."

Sanctity of life
"…and I think euthanasia could have a strong follow on effect in the community in regards to how illness and life quality is viewed" Slippery slope "Concern over how this could be used and could progress to include mental illness and childhood illness" Not the role of doctor "Not the place of the medical profession to carry out this role; does not align with mission; could jeopardise public trust in the profession, with ow-on effects for public health (e.g. not following health advice, avoiding GP and hospitals)" Themes for being unsure about a law change Example Potential for misuse "It is very complex. On one hand I don't have a problem with easing the end for people with progressive conditions. On the other hand I have concerns about how the system will work and be used. My fear is that those with disabilities might be at risk. Or those with mental illness. Or those that see themselves as a burden, rather than actually wanting to end their own suffering, using it as a tool to "end" the suffering of those around them." Not the role of doctor "Furthermore, this changes the perception of doctors. A profession that willingly participates in euthanasia is not one I signed up for." General "However, I have lived a fortunate life and have not know anyone to experience a uncertainty drawn-out, painful death, so I can't pretend that I know what's best for people in that position -hence on the fence." Relieving suffering "Because it is a complex issue and people's lives and connections to others are also complex as could be the in uences on the person thought to be voluntarily and competently requesting to do this. On one hand this may well be the compassionate and empathetic thing to provide for a person who is in pain, suffering and terminally ill." Palliative care as an alternative to EAD "A comprehensive study regarding the symptoms a person at EOL experience showed that the process of dying itself can be very peaceful with appropriate palliative care input, e.g. pain is less of a concern than the general population would think. We have a robust palliative care system in NZ and I believe we would be demeaning its importance if we legalised euthanasia…" "…Rather than focusing on euthanasia, I think it would be more bene cial in the long-term to focus on a robust palliative care system…" Personal values "I personally believe that life is sacred and that no person should be legally allowed to end another person's life, even at the other person's request" Themes for being unsure about a law change Example Undermines "I would also want to ensure palliative care wouldn't be neglected because euthanasia palliative care was brought in." Not the role of doctor "Doctors could administer the euthanasia, but I don't think they should have a role in assessing or deciding if someone should receive euthanasia." Potential for misuse "i predict that assessing competence and the whole process in general may be fraught with danger and di culties. The person could be persuaded/pressured by family and friends or they may feel pressure to end their life independent of anything the family have actually done." Autonomy "I believe in autonomy -and have met many patients who are terminal and say they want to die. But I think legislating is fraught with di culty." Themes generated answering 'no' to supporting EAD Medical school teaching "Teaching in medical school have also shaped my opinions around palliative care and given me a lot of con dence in it -I believe it could be better, but I think it is the best way forward." Death/suffering of family member or friend "I have a grandmother who was diagnosed with cancer 5+ years ago that at times has looked very unwell including needing a short stay in hospice care 4 years ago. She is the type of person who does not wish to burden others and I am convinced she would elect for physician assisted dying at the point she was admitted to hospice. She is still alive many years later and has managed to continue living at home with my grandfather who otherwise would be in a residential care facility. This has painted a pretty clear picture as to how vulnerable people's lives could be ended early under this law change to the detriment of those around them and society in general." For both year groups (years 2&3 and 4&5), the authors identi ed three main themes for the reasons most often cited by those who supported a law change: autonomy, relieving suffering, and enabling a digni ed death. Several participants also mentioned the relief it would bring family, that it would provide a safe means for people to end their life who would do so anyway, that palliative care could not resolve all suffering, that current palliative care practice already involves EAD, and nancial reasons. The identi ed themed reasons cited by those who were opposed to a law change were more varied. The most common were that there is a potential for the practice to be misused and that EAD is not compatible with the role of doctors. Several students also mentioned sanctity of life, the potential slippery slope, their own personal values, and that palliative care is an adequate alternative to EAD. Many of those who were unsure cited similar reasons to those who were opposed, but cited these as concerns, while many in this group also expressed sympathy for the reasons given by those who support a law change. There was no notable difference between the ELM and ALM groups in terms of the reasons given.

Discussion
The themed experiences that participants described as in uencing their thinking were largely similar across the 'yes', 'no' and 'unsure' groups, and also year levels. Several in each group mentioned medical school teaching as in uential, along with discussions with family or friends, discussions with doctors, public lectures and personal study, among others. The death or suffering of a family member or friend was the most common type of experience mentioned by those who were supportive of a law change. Moreover, the death or suffering of a family member was a common theme that in uenced the students' views for all year groups and all responses except 'unsure'.
A high number of those in ELM (years 2 &3) who supported a law change also mentioned their rest home placement, where second year students work as assistant caregivers in residential care facilities. Those who were opposed to a law change most often identi ed medical school teaching as having in uenced their thinking, and a small number directly stated that their views had changed because of what they have learnt in medical school. Several in all groups demonstrated considerable prior and independent thinking on the issues.

Discussion
Quantitative ndings A survey of University of Otago medical students was undertaken to learn their views of EAD, in order to learn how medical education affects medical students' views. The results of the survey show that the odds of medical students supporting EAD at 5 th year compared to 2 nd year is 0.30 (95%CI 0.15-0.60). The level of support found among 2 nd and 3 rd year (65% and 63% respectively) students is close to that found in the general public (68%) (3). However, the level of support among 5 th year students (39%) is closer to that of practicing doctors, for example, general practitioners in New Zealand (41%) (3). Identifying with a religion was also found to have a statistically signi cant correlation with higher levels of opposition and lower levels of support for EAD, which is consistent with other studies (1). To our knowledge, this is the rst study to investigate how New Zealand Medical Students view EAD, and whether their views are affected by medical training.
The level of support among 2 nd and 3 rd year students is generally higher than that reported in several overseas studies of medical students (10,(12)(13)(14)(15), though caution should be taken in drawing precise and direct comparisons as these studies use a variety of questions that are worded differently, and involve students at a range of stages of learning, along with a variety of medical degree structures (e.g. length of degree, undergraduate vs. postgraduate entry).
The similarity between the levels of support for EAD among participants in years 2 and 3 and the general public supports the link that Stronegger et al 2011 identi ed between medical students' views and levels of support in society. However, the decrease in support across year levels suggests that medical education may be another signi cant factor. This suggests that the opposition among practicing doctors is not simply due to social factors at the time they entered medical school, the personality of those who choose to be doctors, or because doctors are naturally more 'conservative' on moral issues. If this is true, then we should not expect doctors to become more supportive of EAD in the future, simply because societal attitudes have changed.
The nding that medical students' support for EAD decreases as they advance through their education aligns with ndings in three other studies (18)(19)(20). It does not align with the results of one study of students at an English Medical School, which reported that being a nal year student was weakly associated with a greater likelihood of agreeing with EAD (15). However, unlike in this study, a clear majority were opposed to EAD at both the rst and nal year of training. It seems likely that medical students support and oppose EAD for a range of reasons, and it is possible that medical education only in uence some of these reasons. An explanation for why this English study found a different effect could be that the students' views were generally less changeable.

Qualitative ndings
The reasons participants provided to explain their answers to question one broadly correspond with what is known about the views of the general public in New Zealand (25), and with the reasons typically put forward in the international debate (26). This suggests that the participants as a group are not substantially different in how they think about EAD from the rest of the population. However, the reasons themselves do not provide a clear indication of why medical students are less likely to support EAD later in their medical education. It could be that during medical education a student becomes more aware of reasons to be opposed, or that such reasons become more compelling because of certain clinical experiences they have.
Overall, the experiences that participants identi ed as in uencing their views also do not indicate a clear answer to that question. The nding that a higher number of those who were opposed identi ed medical school teaching as in uential indicates this is an important factor, and yet this was the case for both ELM and ALM students, and some of those who were supportive also identi ed medical school teaching as having in uenced their views. Taken together, these ndings suggest that it is not any particular part of or event during medical education that changes students' views, but rather that a variety and perhaps accumulation of experiences can change students' views in both directions, and that there is a greater movement toward opposition.

Overall discussion
There are several possible factors that might make medical students less likely to support EAD as they progress through medical education. Exposure to end-of-life situations could be a factor. At Otago Medical School, students only begin learning in clinical environments with direct patient contact from fourth year. One study attributed decreased support to increased personal experience in end of life care (27). However, another found that while personal experience in caring for a dying person correlates with less indecision in terms of views, there was no signi cant correlation between such experience and support or opposition for EAD (28). Hence, it is perhaps not personal experience with care of the dying that reduces support (e.g. caring for a dying relative), but being professionally educated in, and having a sense of the obligations involved in, providing such care. Speci cally, it may be that education in healthcare brings a greater awareness of the treatments and approaches that are available (e.g. how pain can be managed, holistic approaches to suffering), which reduces the sense of a need for EAD (29). It could also be that medical education increases awareness of the practical challenges or risks involved in applying an EAD policy, either for patients or the practitioners (30). Studies have reported that practitioners nd the processes around EAD stressful and di cult (1).
It is well established that medical students learn about the medical profession from the formal, informal and hidden curriculum (31,32). Theoretically, each of these dimensions of the curriculum could in uence how participants view EAD. However, it is not clear what part of the formal curriculum would, or could, have this effect. In a Dutch study, student attitudes remained relatively constant before and after a 'euthanasia and doctor-assisted suicide' conference, though knowledge improved signi cantly by the end (33). When EAD is addressed in ethics classes at Otago Medical School, students are presented with contrasting arguments and encouraged to think independently, and -as noted -medical school teaching was identi ed as an in uential factor by participants representing all three responses ('yes', 'no' and 'unsure').
The concept of the hidden curriculum is that medical education is not just learning about knowledge and skills, but is also a process of socialisation into the medical profession (31,32). It is possible that the students are in uenced by the doctors they learn from in clinical rotations, and that in their change of views they are conforming to a perceived professional norm. If this is the case, then it might be thought to align with the claim that doctors as a group are more socially or ethically conservative. However, it is also possible that this apparent 'conservatism' is grounded in ethical concerns that members of the general public would share if they had a clearer understanding of how doctors perceive this issue. In studies where doctors have been asked why they are opposed to EAD, they often articulate these kinds of reasons, rather than personal values (34,35). Furthermore, it is possible that responsiveness to these reasons depends on certain practical knowledge. Drawing on the work of the philosopher Ludwig Wittgenstein, Amesbury has argued that understanding the normative force of reasons requires some familiarity with the practices in which the reasons are used (36). Perhaps medical students become less likely to support EAD because they gain this familiarity, and so understand the ethical weight of the sorts of reasons put forward by doctors who are opposed. Such practical understanding is often not adequately represented in whatever reasoning a person puts forward to explain their position, and this is perhaps re ected in the absence of any clear difference between the reasons put forward by ELM and ALM participants. If correct, this suggests there is a practical basis to ethical and professional formation, which underpins ethical reasoning.
Further qualitative research may provide deeper insight into which aspects of medical education in uence medical students' views, and why some students change their views during medical education while others do not. The results of the interviews that were undertaken as part of this study will be reported in a subsequent paper. Similar research at other medical schools might help to show whether the difference between year groups found in this study is due to local factors or a common feature of medical education.

Limitations And Strengths
The response rate of this study is low, and nothing is known about the views of those who did not respond. Most other, comparable types of studies have received a similar response rate (19). The sample is broadly similar to the overall cohort, though females were slightly overrepresented in the third year and those identifying as 'other' ethnicity are underrepresented.
Strengths of the study are that it examines the views of medical students across four year levels and integrates quantitative and qualitative ndings. While the quantitative ndings show a difference in views between year groups, the qualitative ndings suggest that this difference is related to the different experiences that participants have had, particularly in medical education, and the different ways they have responded to these. A follow up study of the 2 nd year cohort once they have completed 5 th year may strengthen the ndings.
The difference in levels of support between students at years two and ve could be due to a range of factors relating to medical education, including exposure to end of life situations in years 4 and 5, increased knowledge of the options that are available through palliative care, awareness of the practical challenges created by EAD policy, and socialisation. It is not clear whether the effect is due to local factors or medical education in general. While there is some evidence from overseas studies to support the claim that it is due to medical education generally, this is di cult to con rm due to the variables involved.

Conclusions
Many studies internationally have found that doctors are less likely to support EAD than the general public. It may be thought that this is because of certain personal attributes assumed to be more common among doctors (e.g. being morally conservative), but it could also be because doctors perceive reasons to not support EAD that others would share if they had the same practical knowledge. This survey of Otago Medical School students' views of EAD found that students at the end of year 5 were less likely to support EAD than those at the end of year 2, and that this difference is partly due to experiences in medical education. These ndings suggest that the difference in support between doctors and the general public is not simply attributable to personal factors, and that the views of the general public may be changeable. These conclusions should be considered when interpreting public opinion surveys on this issue. The ndings also suggest a practical basis to ethical and professional formation, which -if correct -should inform teaching about professionalism and medical ethics. Further research on medical students' views of EAD may reveal more about how those views are formed during medical education, and which aspects of their learning make them less likely to be supportive.

Declarations
Ethics approval and consent to participate Category A ethics approval was granted for this project by the University of Otago Human Ethics Committee (reference code: 18/209). Potential participants were emailed an invitation to participate with an information sheet attached. They were informed in the email that completing the survey would be taken as informed consent to participate. The survey, data management, and interviews (not reported in this paper) were all undertaken in accordance with relevant guidelines and regulations.

Consent for publication
Not applicable.

Availability of data and materials
The qualitative dataset generated and analysed during this current study is not publicly available for privacy reasons. The quantitative dataset, separated from the demographic information, is available from the corresponding author on reasonable request.

Competing interests
No member of the research team has a con ict of interest in regard to this research.

Funding
This project was supported by funding from the Health Research Council of New Zealand.
Authors' contributions LN and SW generated the idea for the study and worked with KSH on the survey design and data gathering. LN undertook an initial literature review, managed the administration of the survey, and wrote a rst draft of the paper. EI provided the statistical analysis and wrote the relevant sections of the paper. LN and SW undertook the thematic analysis of the qualitative data with support from KSH. SW led the writing of subsequent drafts with all authors contributing at various steps. All authors have read and contributed to the nal version.