Introduction

End-of-life decision-making and subsequent policymaking are subject of debate worldwide.1 2 Several countries have established laws and regulations stating what is and is not allowed regarding end-of-life interventions at the explicit request of an individual.3 Physician-assisted dying (PAD) includes both euthanasia, where a physician administers lethal medication to the patient, and physician-assisted suicide, in which case the patient self-administers lethal medication provided by a physician.4 Some countries have legal regulations on both euthanasia and physician-assisted suicide: the Netherlands, Belgium, Luxembourg, Colombia, Canada, Spain and all states of Australia, but not the territories. In several states of the USA only physician-assisted suicide is regulated by law. In Switzerland and Germany there is no legislation on euthanasia or physician-assisted suicide, but under certain circumstances assisted suicide is decriminalised.3 5

In the Netherlands, assisting in suicide is forbidden and punishable for anyone but doctors, for whom such assistance is regulated by the Termination of Life and Assisted Suicide (review procedures) Act, which came into force in 2002.6 7 However, there are also individuals who wish to intentionally end their life without a physician’s help.8 In 2007, Chabot published a study into what he called ‘auto-euthanasia’: a form of life termination in which an individual independently ends his own life in dialogue with one or more confidants, but without a physician performing a lethal act.9 He described and studied two methods of such self-directed dying, this being voluntary stopping eating and drinking (VSED) (Chabot referred to this method as ‘voluntary refusal of foods and fluids’ (VRFF).) and independently taking lethal medication attended by a confidant (ILMC) (Chabot defined this method as ‘death by a combination of a medication that is lethal in overdose with an overdose of one or more sleeping drugs’ (see also: Auto-euthanasia, p13). In our current study we broaden ‘a medication’ to ‘medication and/or other substances’.). His study was based on the experiences of confidants who were involved in the process. Chabot estimated that in the period 1999–2003, 2.1% of all deaths in the Netherlands were the result of VSED, and 1.1% were the result of ILMC. He later adjusted the estimate for VSED to 1.7%, to take into account that some of these deaths represented dying patients who lost their appetite in the dying process, instead of a deliberate decision to end life by stopping eating and drinking.10 11

There have been other estimates of the frequency of the number of deaths by means of VSED or ILMC, based on the 5 yearly death certificate studies on end-of-life decision-making practices in the Netherlands. These data are based on the perspective of physicians. The most recent estimate of the frequency of self-directed dying in that study suggests that in 2015 0.5% of all deaths were the result of VSED, whereas 0.2% were the result of ILMC.12 In 2021, the frequency of ILMC was again estimated at 0.2%.13 In countries outside the Netherlands, there is little information available about the prevalence of VSED or ILMC. A Swiss study showed that 1.1% of all deaths in Swiss nursing homes and private homes were the result of VSED,14 and a Belgium study showed that in 2013 0.5% of all deaths in Flanders were the result of VSED.15 A study among hospice nurses in Oregon (USA) showed that 33% of them had cared for a patient who deliberately hastened death by VSED in the past 4 years.16

In the past 15 years there have been a number of important developments relevant for self-directed dying in the Netherlands. In 2014, the Royal Dutch Medical Association together with the nurses association, issued a guideline for physicians and nurses on how to care for people who voluntary stop eating and drinking.10 This guideline, which is recently revised, states that caring for someone who voluntarily stops eating or drinking is not considered assisting in their suicide, but is ‘normal’ medical practice that falls outside the scope of the Dutch (PAD-)law.11 Another development is the continuous societal and political debate about the self-chosen death. Studies into the Dutch public opinion show that people consider values of control and preservation of independence to be important for ‘a good death’,17 and that almost two-thirds of Dutch citizens support the statement that everyone has the right to self-determination over their own life and death.13 The Dutch minister of health and a Dutch member of parliament both submitted blueprints for a new law for assisted death for healthy people over 75 years of age who feel their lives are completed.18 19 Research into the public opinion shows that there is some support for this idea: 58% and 36% of Dutch citizens, respectively support the statement that elderly people who are ‘tired of living’ should be eligible for PAD, and that elderly people should obtain prescribed lethal medication from a physician to end their own lives.13 20 Various societal organisations such as the Cooperative Last Will, the Dutch Right to Die Society and The Einder Foundation, lobby for more opportunities for self-directed deaths, independent from medical help. Furthermore, Chabot published a book that is publicly available and provides information on the preparation and execution of a death by VSED or ILMC.21

These developments may have had an influence on the frequency of self-directed dying. Therefore, this study aims to answer the following research questions:

How many people currently die each year in the Netherlands by VSED or ILMC?

Has this number changed since 2007, and if so: what are possible explanations for changes in frequencies?

What characterises the decision-making process and quality of dying of people who choose self-directed dying, according to their proxy’s?

Methods and analysisStudy design

We will conduct a replication study of Chabot’s research, using the same panel and sampling. This retrospective cross-sectional population-based study uses a mixed-methods approach to provide insight in the number of people who die each year in the Netherlands by either VSED or ILMC, and the decision-making process that precedes their death. The results of this study will be compared with the study results from 2007.

Study population

The study population consists of Dutch adults over the age of 18 who will be randomly drawn from a large panel of citizens willing to participate in questionnaire research. The panel is maintained by independent research agency Verian, which ensures the quality of the panel through their recruitment methodology (self-registration is not possible) and intensive panel management. Panel members are approached up to three times a month and receive compensation for their cooperation. This results in high response rates. The panel consists of around 90.000 members and is representative of the Dutch adult population in terms of gender, age, education and geographical location.22 23 In order to be eligible to participate in this study, respondents must comply with the following inclusion criteria: be 18 years or older; be able to speak and read Dutch; and have given informed consent to participation.

Sample size calculation

In Chabot’s study, the prevalence of VSED and ILMC was 0.3% and 0.1%, respectively.24 Our sample size calculation is based on these numbers. We aim to estimate the proportion of people who have experience with VSED and ILMC within a relative margin of error of 30% (meaning an absolute margin of error of 0.09 percentage points and 0.04 percentage points for VSED and ILMC, respectively). To obtain this margin of error we would need to include 13 826 persons for estimating the prevalence of VSED and 28 580 persons for estimating the prevalence of ILMC within a relative margin of error of 30%. To this end, the same sample size (n=37 500) and two-step screening procedure will be used as in the original study.

No sample size calculation has been performed for the qualitative interviews since its purpose is to explore and better understand the variety in experiences of individuals, rather than having a representative sample. We aim to interview around 40 respondents, 20 of them on their experiences with VSED, and 20 on their experiences with ILMC. If data saturation occurs with fewer respondents, no more interviews will be conducted.

Questionnaire study

A two-stage screening procedure will be carried out from January to February 2024 to determine whether the experiences of respondents represent a death by VSED or ILMC, and thus fall within the topic of our study. Reminders will be sent by Verian to people who have not yet completed the questionnaire after 1 and 2 weeks after the initial invitation, respectively. After respondents have passed phase 1 of the screening, that is, have answered yes to questions 1 and/or 2, question 3 and question 4 (box 1), they will gain access to an extended second questionnaire. This extended questionnaire starts with five additional screening questions, representing phase 2 of the screening procedure, followed by about 90 prestructured questions and 5 open-ended questions about either VSED or ILMC. These questions cover demographic and clinical characteristics of the deceased person; the planning and course of the dying process; involvement of family members, friends and/or carers; demarcation of VSED from a ‘natural death’ after a few days of not drinking and from a death by continuous sedation without hydration; and demarcation of ILMC from PAD and from solitary or mutilating suicides

Box 1 Screening and non-response questionnaire

Screening questions stage 1:

Over the past 10 years did you ever become closely involved with someone who after discussing it with you ended his/her life by sleeping pills or other medication while no doctor was present? (yes/no)

Over the past 10 years did you ever become closely involved with someone who after discussing it with you hastened death by a deliberate choice to stop eating and/or drinking? (yes/no)

Question 3 is presented to respondents answering ‘yes’ to question 1 and/or 2. Question 4 is presented to respondents answering ‘yes’ to question 3.

Did the deceased discuss the wish to end life or to hasten death and how he/she planned to do that with you? (yes/no)

Are you willing to answer an online questionnaire about that experience? (yes/no)

Non-response question:

What is the main reason you prefer not to complete a questionnaire about your experience? N.B. Please only tick the answer that indicates the main reason.

Remembering would be too painful.

I had promised the deceased absolute secrecy.

I don’t think research into such personal events is appropriate

Other, namely: ………………….

Since non-response may be associated with experiences with VSED or ILMC, it is important to understand the reasons for non-response. Therefore, respondents who indicate they are not willing to complete an online questionnaire about their experiences with VSED or ILMC (‘no’ to screening question 4), will be asked to answer a final question in the response questionnaire (box 1).

Informed consent procedures

Selected members of the Verian-panel will receive an invitation via email from Verian to participate in our questionnaire study. The invitation will be accompanied by a detailed information letter on the study, after which members can decide whether or not to participate. If so, they will receive a link to the corresponding consent form, in which they can provide consent to participation in the study, the use of their research data, and whether they agree to their research data being used for future research. After giving consent, members are given access to the 4-item screening questionnaire.

When the screening data show that participants fall within the target group, they are invited to participate in the extended second questionnaire. Again, their informed consent is sought. If members decide not to participate in the extended questionnaire study, they are asked for their reasons in a non-response questionnaire. Answering this question is not obligatory.

The research data will be pseudonymised by the data provider (Verian) before sharing them with the researchers. The data provided to the researchers do not contain any directly identifying data. This way, the identity of the respondents is only to be re-identified by using the pseudonymisation key. This key is stored separately from the data by Verian. The research data will be stored for 10 years in the secured environment of Erasmus Medical Center Rotterdam.

Interview study

The interviews (n=40) will be held from May to September 2024 with a selected sample of questionnaire respondents indicating their willingness to participate in an interview. In that case, the relevant participants are asked for consent by Verian to share contact details (name and telephone number or email address) with the researchers.

The interviews will be based on a topic list and conducted by an experienced interviewer. The interviews will be audio recorded and transcribed verbatim.

Informed consent

At the end of the extended questionnaire on VSED or ILMC, respondents can indicate whether they are willing to participate in a subsequent interview about their experiences with the research topic. If they indicate their willingness to do so, and provide informed consent to share their contact details with the researchers, the researchers will receive the contact details of selected participants to schedule an interview. In the process, a new information letter will be provided. If it comes to an interview, participants can sign a consent form together with the researcher-interviewer.

According to the GDPR/AVG regulations, participants are informed about the processing of their data and their rights through the subject information sheet in the questionnaire, and through the information letter they receive prior to the interview. When a participant refuses to participate, their data will not be collected. When a participant withdraws consent, the data collected up to the moment of withdrawal will be used (according to GDPR/AVG Article 891), but no further data will be collected.

Data analysis planPrimary outcomes

The primary outcome of this study is an estimate of the number of people who die each year in the Netherlands by VSED or ILMC. Frequency estimates will be based on the number of participants that pass all screening stages. The frequency of both types of self-directed deaths in the population can be estimated with the standard weighting method that is used by Statistics Netherlands (CBS) for estimating the annual frequency of relatively rare events such as traffic incidents. We will use an adapted standard weighting. Each informant is weighted by the inverse of 1 (the informant) plus the number of other individuals who had been confided by the deceased person about their intention to hasten death by VSED or ILMC. The estimated number of cases will be computed for the responding part of the sample. For the frequency estimation, we will use of the number of informants on a death by VSED or ILMC in the 5 years from 1 January 2018 to 31 December 2022, because we consider reports about cases before 2018 less reliable. (See also: Auto-euthanasia, p95.) The annual frequency of VSED and ILMC can then be computed as the estimated proportion of reported cases in the sample (ie, weighted sum of informants on VSED deaths and ILMC deaths, respectively divided by sample size corrected for non-response), multiplied by the total number of Dutch adults minus non-western adults (under-represented in the sample) and divided by the 5 years covered by the informants. The Poisson distribution will be used in calculating 95% CIs.

After computing the estimated numbers of deaths by VSED and ILMC, a comparison will be made with the 2007 estimates. The numbers of deaths by VSED and ILMC will be described in frequency tables, using standard statistical analysis. In addition, possible explanations for changes in frequencies will be explored and described.

Secondary outcomes

The secondary outcomes of our study concern frequency estimates of the characteristics of deaths by VSED and ILMC (ie, clinical and demographic characteristics of the deceased person) and the decision-making process (ie, the planning and course of the dying process and involvement of confidants), as well as the perceived quality of dying of people who choose self-directed dying as described by their confidants. In addition, a subgroup analysis will be conducted to check for correlations between deaths by VSED and ILMC, and various variables, such as age, education level and health status. The interviews will be transcribed verbatim and thematically analysed using Nvivo to provide more in-depth insight into the quality of dying of people who choose self-directed dying.

Data management statement

Dataset can become available at reasonable request on completion of the study.

Patient and public involvement

This study concerns a replication study. Patients and the general public were not involved in the design of the study, choice of outcome measures or recruitment procedures. We will set up a focus group to advise on the topic list for the interviews and on the interpretation of the results of the study. The focus group will consist of experts in the field of self-directed dying in the Netherlands, such as legal experts, healthcare providers, ethicists and epidemiologists, and representatives of the general public. The first focus group will take place in May 2024 and the second in September 2024.

Ethics and dissemination

The study will be conducted according to the principles of the Declaration of Helsinki25 and in accordance with relevant laws and regulations. The study has been approved by the Medical Ethical Review Committee of the Erasmus Medical Center (MEC-2023-0689).

We acknowledge the potential burden and risks for respondents to participate in this study. Burden for participants is remembering witnessing the intentional death of a close person (friend or relative). Participation involves no risks. The benefit of participation is limited: some participants may feel relieved by sharing their experiences.

Within 1 year after the end of the study a final report will be written and published in Dutch. The results will be presented to the organisations funding the study, and disseminated through at least one scientific publication and presentations at conferences.