How do trained palliative care providers experience open desire to die-conversations? An explorative thematic analysis

Objectives. Despite the potential benefits of open communication about possible desires to die for patients receiving palliative care, health professionals tend to avoid such conversations and often interpret desires to die as requests for medical aid in dying. After implementing trainings to foster an open, proactive approach toward desire to die, we requested trained health professionals to lead and document desire to die-conversations with their patients. In this article, we explore how trained health professionals experience an open (proactive) approach to desire to die-conversations with their patients. Methods. Between April 2018 and March 2020, health professionals recorded their conversa-tion-experiences on documentation sheets by answering seven open questions. A subsample was invited to offer deeper insights through semi-structured qualitative interviews. Interviews and documentation sheets were transcribed verbatim and analyzed thematically, then findings from both sources were compared and synthesized. Results. Overall, N = 29 trained health professionals documented N = 81 open desire to die-con-versations. A subsample of n = 13 health professionals participated in qualitative interviews. Desire to die-conversations after the training were reported as a complex but overall enriching experience, illustrated in seven themes: (1) beneficial (e.g., establishing good rapport) and (2) hindering aspects (e.g., patients ’ emotional barriers) of desire to die-conversations, (3) follow-up measures, (4) ways of addressing desire to die, as well as (5) patient reactions to it. The interviews offered space for health professionals to talk about (6) content of desire to die-conversation and (7) (self-)reflection (e.g., on patients ’ biographies or own performance). Significance of results. As part of an open (proactive) approach, desire to die-conversations hold potential for health professionals ’ (self-)reflection and a deeper understanding of patient background and needs. They may lead to a strengthened health professional – patient relationship and potentially prevent suicide.


Introduction
Confronted with approaching death due to serious, life-limiting disease or geriatric multimorbidity, patients frequently express a desire to die (Monforte-Royo et al., 2012;Bellido-Pérez et al., 2018;Bornet et al., 2020;Briggs et al., 2021). Recent public discussions tend to see these desires to die only through the narrow lens of (ethically) adequate reactions toward requests for medical aid in dying (MAiD) (Wright et al., 2017). In contrast, we propose a more open approach toward desire to die in palliative care which can take on various forms and is not limited to requesting MAiD (German Guideline Programme in Oncology, 2020). Our broad definition conceptualizes a desire to die as an idiosyncratic and dynamic phenomenon on a continuum of increasing suicidal pressure to act (German Guideline Programme in Oncology, 2020; Kremeike et al., 2021a), including the wish to hasten death (WTHD) and requests for MAiD as only a few of various possible forms of desire to die (Balaguer et al., 2016). While some patients receiving palliative care merely express their acceptance of death or tiredness of life without a WTHD, some harbor latent wishes to die in case of worsening symptoms and only a few may develop acute suicidality (German Guideline Programme in Oncology, 2020; Kremeike et al., 2021a). Our open approach thereby corresponds to other international efforts to re-conceptualize dealing with desire to die or administering MAiD as a "relational care process that occurs over time" (Wright et al., 2017, p. 61). We designed a semi-structured clinical approach for communicatively dealing with desire to die on which we based a training to increase health professionals' self-confidence and capability Kremeike et al., 2020;Voltz et al., 2021).
When it comes to communication about desire to die, a too narrow view among health professionals contributes to widespread insecurity or even reluctance to initiate conversations with patients (Fujioka et al., 2019;Baile et al., 2000). This uncertainty extends to dealing with situations in which patients themselves express a desire to die (Galushko et al., 2016;Kremeike et al., 2021b). There is a long-standing opinion that taboos surrounding the topic of death and dying are still predominant in society, e.g., among health professionals working in curative and palliative care (Wildfeuer et al., 2015;Baile et al., 2000). However, a differing sociological opinion argues that the "death denial" within society is decreasing (Wildfeuer et al., 2015). Potential taboos can come from a widespread misbelief that addressing a desire to die might cause or increase it, which may trigger health professionals to act defensively (Hvidt et al., 2017). A lack of training in communication about desire to die and the fear of being emotionally overwhelmed by such a conversation may lead to an avoidance of the topic altogether (Lenherr et al., 2012;Galushko et al., 2016) even though patients explicitly wish for their health professionals to address these issues proactively Crespo et al., 2021).
Suicidality research shows evidence for the appropriateness of universal screening for suicidality: a recent meta-analysis of 13 prospective studies clearly stated that it does not pose an iatrogenic risk for patients (DeCou and Schumann, 2018). These findings are recently generalized to the proactive assessment of desire to die in patients receiving palliative care. In a pioneering study, 193 oncological patients were asked upon hospital admission about possible desire to die in form of a short, semi-structured interview. The vast majority of participants did not find it upsetting (94.8%) and deemed it important to talk about these issues (79.3%) (Porta-Sales et al., 2019). Moreover, current evidence we previously published elsewhere suggests a decrease in symptom burden such as depressiveness: in 85 patients with various diagnoses requiring palliative care, moderate to severe depressiveness decreased significantly after an open conversation about desire to die with their health professionals (Voltz et al., 2021).
While effects of desire to die-conversations on patients are increasingly researched, we want to focus on the health professionals' experience of these conversations: What are the concrete contents of related conversations and what are beneficial or hindering aspects? What do health professionals think of their own role and performance? How do they address desire to die and how do they experience patients' reaction to it?
In this article, we aim to explore how trained palliative care providers implement and experience conversations about desire to die using an open (proactive) approach (Kremeike et al., 2020).

Procedure
As part of a larger sequential mixed methods study, health professionals from all palliative care settings were invited to participate in multi-professional trainings Kremeike et al., 2018Kremeike et al., , 2020. The trained health professionals then selected patients from their palliative care practice to hold a single desire to die-conversation with these patients following our semistructured clinical approach (Kremeike et al., 2020;Voltz et al., 2021). The clinical approach encourages open desire to dieconversations and can be adapted to the health professionals' personal communication style. A corresponding booklet was distributed among all trained health professionals for support in planning the conversation (available for reference as a supplement in Kremeike et al. (2020)). In the course of this study procedure, qualitative data presented inthis article was gathered in two ways: Health professionals documented their experience of these conversations on desire to die on a documentation sheet with seven open questions (A; see Supplementary material 1). The documentation sheets were explained to them during training, then provided digitally. Health professionals sent back their completed documentation sheets to the research team via fax. A subset of these health professionals was invited to participate in individual semi-structured qualitative interviews (B; see Supplementary material 2). Therewith, the experience of talking about desire to die was explored in more depth. Interviews were conducted by all female members of the research team (KK, CR, LG, and KB) who either held a Master's (CR, LG, and KB) or doctoral degree (KK) and all underwent interview training. For an overview of the procedure, see Figure 1.
This study was approved by the Ethics Committee of the University Hospital Cologne (Nr. 17-265) and is registered in the German Clinical Trials Register (DRKS00012988). Research was conducted in accordance with the Declaration of Helsinki.

Recruitment
For initial study participation, we recruited a convenience sample of health professionals from all palliative care settings (for details  Kathleen Boström et al. on recruitment, refer to Voltz et al. (2021)). For participation in a qualitative interview, an interested subsample of these health professionals was invited, again via convenience sampling. Before study participation, health professionals gave written informed consent for the evaluation of the training course and later, if applicable, for participating in qualitative interviews about their experiences of desire to die-conversations. Due to our sampling strategy, available qualitative data was analyzed after study completion, using an explorative approach without aiming at data saturation. This is common when doing research with palliative patients, as the characteristic frailty as well as drop-out by death of this group of patients runs counter to theoretical sampling approaches (Aktas and Walsh, 2011).
Data collection and analysis (A) Collection and analysis of desire to die-conversation documentation sheets Trained health professionals had a desire to die-conversation according to our clinical approach with their selected patients and completed a documentation sheet at their place of work, including the following points of interest: 1. Key data a. Duration of the conversation b. Setting c. Presence or absence of desire to die in the patient d. Proactive (by health professional) or reactive (by patient) addressing desire to die 2. Content a. Type and function of desire to die b. Further (clinical) measures agreed upon 3. Atmosphere of conversation 4. Perception of one's own performance Content and wording of the documentation sheet was set to fit our semi-structured clinical approach (Kremeike et al., 2020). Therefore, we analyzed the documentation sheets following a concept-driven (deductive) approach (Kuckartz, 2019). During this initial summarizing and structuring analysis, the categories of the clinical approach were used for the construction of main themes.

(B) Collection and analysis of qualitative interview data
A subset of health professionals participated in individual semi-structured interviews where they were asked about their experience of the desire to die-conversation during study participation and in general. Interviews were recorded and transcribed verbatim. To explore beyond the concept-driven findings from the documentation sheets, interview passages were inductively coded using thematic analysis (Braun and Clarke, 2006).

(A and B) Synthesizing data analysis from both sources
Results from documentation sheets and interviews with health professionals were analyzed separately, then themes were compared and synthesized when there was an overlap. Initial coding was done by KB and discussed between three researchers (KB, KK, and TD), initiating ongoing adjustments and consecutive coding until consensus was reached (Campbell et al., 2013). All qualitative data was analyzed using MAXQDA 20 (VERBI Software, 2019). Coding and explanation of themes was done following quality criteria for thematic analysis (Steinke, 2000).

Sample
Between April 2018 and March 2020, n = 29 from the original 102 trained health professionals (28.4%) completed N = 81 documentation sheets of desire to die-conversations with their patients. The health professionals were M = 46 years of age (SD = 9.7), with 26 women and 3 men. Afterward, n = 13 interviews were conducted for qualitative evaluation. Health professionals participating in the interviews were 10 women and 3 men. For details on the sample, see Table 1.

Overview of all themes identified in documentation sheets (A) and qualitative interviews (B) on desire to die-conversations
Overall, seven themes with 29 subthemes were found within the documentation sheets (A) and the qualitative interviews (B). While three shared themes became present in data from documentations sheets as well as the interviews (A and B: "Beneficial Aspects," "Hindering Aspects," and "Follow-Up Measures," see Table 2), two themes were unique to documentation sheets (A: "Ways of Addressing Desire to Die" and "Patients' Reactions to Addressing Desire to Die," see Table 3) and interview data each (B: "Content of Desire to Die-Conversations" and "(Self-)Reflection," see Table 4). For all seven themes and their occurrence within the two data sources, see Figure 2. Given the explorative nature of our study, themes and categories are not meant to be conclusive, but to provide a descriptive account of the data gathered by our convenience sample of health professionals.

Beneficial and hindering aspects of conversations and follow-up measures (themes from A and B)
Regarding beneficial and hindering aspects of their desire to dieconversations, health professionals reported similar aspects both on the documentation sheets (A) as well as within the interviews (B). Leading desire to die-conversations following an open approach was mainly seen as a positive experience by trained health professionals: They reported beneficial aspects such as feeling rewarded when they felt an improvement for their patients, e.g., by activating their resources or being able to address their previously unmet communication needs. Patients were reported as being open and showing willingness to cooperate which was deemed a prerequisite for a rewarding conversation. Consequently, health professionals valued a strong relationship with their patients which was sometimes strengthened by talking about desire to die. Some health professionals reported their own self-competence increasing through positive conversation experiences (see Table 2, themes 1.1-1.3 and related quotes).

Palliative and Supportive Care
Health professionals reported hindering aspects as well which were more present in the documentation sheets than in the interviews: patients who are dismissive or very emotional were reported as demanding. Health professionals themselves sometimes struggled with their own issues regarding the topic, e.g., fearing to be too pushy or insensitive (see Table 2, themes 2.1-2.2 and related quotes).
Within the interviews, health professionals reported in more detail upon what they discussed with their patients regarding follow-up measures for the time after the desire to dieconversation. These mainly included facilitating the desired care for the patient, e.g., in a hospice or through a home service. When desire to die-conversations revealed unmet needs or wishes, health professionals informed their patients about or initiated further support and therapeutic measures. This occasionally included starting a process to fulfill a patient's last wish. Most health professionals used the desire to die-conversation as a starting point for a deepened contact with the patient through further conversations (see Table 2, themes 3.1-3.5 and related quotes).
Ways of addressing desire to die and patients' reactions to it (themes from A: documentation sheets) For analysis of the theme Ways of addressing desire to die from the open-ended questions on the documentation sheet, we deductively applied the categorization for addressing desire to die from our semi-structured clinical approach (Kremeike et al., 2020). These categories present six ways of asking a patient about potential desire to die derived from a Delphi survey with experts conducted previously in our study and published in Kremeike et al. (2020). Health professionals used all six ways of raising the issue of desire to die with their patients. The variety of their phrasings illustrate the  variety of their approaches. Some ask more directly about thoughts of not wanting to live anymore while others ask less direct questions, e.g., about wishes, thoughts, and fears regarding the nearing end of life (see Table 3, themes 4.1-4.6 and related quotes). Correspondingly, patient reactions to raising the issue of desire to die were reported as predominantly open, with some confirmations of possible desire to die. Many patients were described to have an emphatic focus on their will to live, while few showed emotional or negative reactions (see Table 3, themes 5.1-5.6 and related quotes).

Content and reflection of desire to die-conversations and beyond (themes from B: interviews)
Findings from the thematic analysis of the qualitative interviews on desire to die-conversations supported and expanded what health professionals had already reported on the documentation sheets. Themes corresponded largely to survey categories, though go beyond that in significant ways. This becomes evident in the two categories, content of conversation and (self-)reflection. Health professionals used the interviews to talk in depth about how desire to die-conversations could serve to address and explore desire to die. However, these conversations were also experienced and used as opportunities to more thoroughly engage with their patients' situation and suffering by giving them room to express themselves. Sometimes, conversations directly led to the activation of resources (see table 4, themes 6.1-6.4 and related quotes).
Furthermore, setting the interview several weeks after the desire to die-conversation gave health professionals space for reflection beyond mere documentation. Most of them were engaged with their patients' unique personalities and biographies which shaped the desire to die-conversation. They also reflected on themselves, questioning their working style and approaches to desires to die. Some of them were surprised that unconscious expectations regarding their patients (e.g., absence of desire to die) were not met in the conversations. In others, reflection expanded toward conceptual or ethical issues like the term "desire to die" or one's position regarding MAiD (see Table 4, themes 7.1-7.3 and related quotes).

Discussion
Our study presents how trained health professionals conduct and experience open (proactive) desire to die-conversations. The variety in length, setting, content, and tone of these conversations Table 2. Shared themes, their subthemes and exemplary quotes extracted from documentation sheets of desire to die-conversations and qualitative interviews Themes Exemplary Quotes

Patient Attributes and Resources
"Topics could be discussed openly and without 'psychological barriers' in a 'relaxed' atmosphere." (tn87k10, DS*) "The patient is very positive despite her serious illness. She says, she just lets everything happen and that she feels well cared for here." (tn05k1, DS) "The patient has a stable family network that gives him strength and support." (tn45k5, DS)

Perceived Benefit for the Patient
"I felt that the patient experienced the conversation as relieving, since she said to me that she hasn't been able to talk to someone about it before." (tn40k4, DS) "During the conversation, the patient revealed that her husband and her brother had committed suicide. For her, it was clear that suicide was only a last resort because it makes her cry to this day every time she thinks about it." (tn59k6, DS)

Health Professional Attributes
"So, I find that it is a building block of trust, which also affects me, right.

Containing Patient Emotions and Concerns regarding Death
"Only a few sentences into the conversation after some questions concerning her general health, the patient started crying, since she is currently under a lot of psychological stress." (tn40k4, DS) "[The] patient always came back to the thought <How can I endure this pain?> and that she did not want to die as miserably as her husband did." (tn45k5, DS)

Institutional Care
"To feel safer, she wishes to advance her care plan: e.g., organize a care home or hospice care for later." (tn38k4, DS)

Care at Home
"I initiated continuing home care through a palliative care team and a team of medical professionals. He decided for home care with a specialized palliative care team." (tn63k6, DS)

Follow-Up Conversations
"Every morning, I offer her a cigarette and a talk. She comes daily and tells me about her day. Her best friend is involved in the process to make it easier for her." (tn49k7, DS) Palliative and Supportive Care 5 mirrors the idiosyncrasy of the existential phenomenon of desire to die (German Guideline Programme in Oncology, 2020; Kremeike et al., 2022). The findings also reveal how desire to die is embedded in other pressing matters at the end of life: questions regarding isolation, the stress of changing care settings, complex interactions with close ones, as well as physical and psychological symptom burden. Asking about desire to die in an open and respectful manner can serve as a door opener to these topics which are highly relevant to the patient, such as unmet needs behind multi-dimensional suffering (Monforte-Royo et al., 2012;Otte et al., 2017). Where appropriate, this might lead to changes in the care plan, such as care in other institutions or employing volunteer spiritual support. In our study, all health professionals were trained in an open approach to desire to die and their documented experiences are likely shaped by this training Kremeike et al., 2020). However, health professionals were free to approach the topic of desire to die in their own style. Contrasting a recent study (Porta-Sales et al., 2019), they did not use a question from a standardized short interview, but delivered insight into their individual phrasings and approaches. Health professionals reported that tailoring the approach to the patients' situation through open questions allows the patients to express even latent desire to die which they have not talked about before. This is in line with delivering person-centered care that has been shown to be beneficial for patients' well-being (Dwamena et al., 2012). By balancing conversations between giving an opportunity to talk without probing too deep for the patient to remain comfortable, health professionals in our study also used techniques described by Strang et al. (2014). This places desire to die-conversations in the broader context of existential conversations which, despite burdensome aspects, are described as a special honor (Strang et al., 2014).

Informing about and Initiating Therapeutic Measures
A high number of health professionals emphasized the positive impact of desire to die-conversations on their relationship with the patient and repeatedly came back for further conversations. Desire to die-conversations here may stand unique in their catalyzing function of addressing existential issues while simultaneously being part of a larger communicative process. Studies on good patient-doctor relationships show that such communication and continuity are essential elements thereof (Ridd et al., 2009). As such, desire to die-conversations may even contribute to suicide prevention by reducing depressiveness (a strong predictor for developing a wish to hasten death) and fostering therapeutic alliance and allowing the patients to report vital information (Rodin et al., 2009;Lewis et al., 2014;Voltz et al., 2021).
However, health professionals also reported on aspects which were hindering to the conversation: both patients' unwillingness to engage on the one side and intense emotions on the other hampered open conversation, a communication challenge known from the literature (Banerjee et al., 2016). Despite some patients displaying a distinctly negative response to the topic of desire to die, most were reported not to take issue with or welcome it. These results strengthen recent research findings (Porta-Sales et al., 2019;Voltz et al., 2021).
Within the qualitative interviews, health professionals used a striking amount of time to reflect on their patients' biography and situation. They displayed a profound knowledge and interest which they used to understand background and function of patients' desire to die. Additionally, they reflected their own role   -Hodges and Swords, 2015). Either effects of the prior training on desire to die or a pre-existing high level of competence in mindful communication in our interviewed health professionals are conceivable explanations. Lastly, our results need to be contextualized within the legalization process of MAiD in our country: the German Federal Constitutional Court revoked the legal restriction on (medical) assistance in suicide in February 2020 (Federal Court of Justice, 2020). Up until then, patients with a wish to hasten death often used services provided in neighboring countries (e.g., Switzerland) to legally receive MAiD (Gauthier et al., 2015). Since our study was conducted mainly before February 2020, it is highly possible that todays' conversations of desire to die have changed in tone and urgency. First reports from health professionals in Canada suggest a tendency of patients narrowing conversations towards options of MAiD with new legal opportunities in this regard (Ho et al., 2021). This leaves health professionals feeling restricted in their exploration of background or underlying needs behind a desire to die. To combat these and other reported challenges in communication about desire to die and foster health professional' self-reflection, communication trainings for an open, respectful and proactive approach toward desire to die becomes all the more important .

Strengths and Limitations
The interpretation of our results faces several limitations: all information on content of desire to die-conversations is mediated by (a) selected health professionals' perception and memory of the conversations and (b) our questions in the documentation sheets as well as within the interviews. A significant number of trained health professionals did not recruit patients for desire to dieconversations despite high recruitment efforts (e.g., monetary incentives, bi-monthly reminders per phone, and information of health professionals' superiors). Stated reasons were lack of time, no support from superiors and a critical institution-wide attitude concerning conversations about desire to die; problems already reported before (Kremeike et al., 2021b). Consequently, our results do not present a close description of desire to die- 7.2.2. Patient personality and biography "( … ) I know that she thinks about things in a differentiated manner and also pays very close attention to every small detail in her own body, to her own feelings, to her dialogue partner ( … ). I knew that she would reevaluate a lot of things and that she still needs to clarify a lot, such as family history, unresolved conflicts with the son and ( … ) also a last wish ( … )." (tn38k4, QI) conversations, but an explorative insight into the experience of a convenience sample of health professionals. It is noteworthy to consider potential effects of the ways health professionals selected patients, lead their conversation and document it. Seeing it as a training exercise in learning to communicate about taboo topics, health professionals might have chosen patients that they considered easy to talk to (Aktas and Walsh, 2011). In our study, no patient reported a serious request for MAiD which can be considered in the context of a potential selection bias through gate-keeping. Additionally, only a highly motivated subsample of all trained health professionals participated in interviews. This might skew results to paint a more positive picture of how desire to die-conversations evolve in general.
Furthermore, health professionals repeatedly remarked on the artificial context of leading a desire to die-conversation within a research project. This unique setting likely had an impact on how these conversations were approached: health professionals could take deliberate time and focus on the topic of desire to die with their patients, yet they had to do so differently than they might otherwise have done (e.g., by planning and documenting their conversation in a fixed time frame). They also reported feeling a role conflict between being a carer and recruiting patients for study participation. Nevertheless, health professionals were free to lead their desire to die-conversations in their own style which is important, given differing requirements, habits and time constraints of different palliative care settings. Thereby, our health professional sample meets the criteria of heterogeneity, making our results generalizable for health professionals trained in dealing with desire to die. Further research might systematically analyze differences between care settings or professions in a larger sample.

Conclusion
When discussing adequate ways of dealing with desire to die, it is important to utilize an open approach that prioritizes the exploration of patients' background and needs and instead of narrowing its focus on requests for MAiD. Health professionals trained in this approach experience conversations about desire to die as mainly positive and use such conversations to negotiate a variety of practical and existential topics. Even in the given context of a study, health professionals experienced desire to die-conversations as a door opener and appreciated their value for (self-)reflection. This insight into health professionals' experience of desire to dieconversations emphasizes the utmost importance of addressing this topic openly, respectfully and with curiosity for the patients' individuality and thereby potentially preventing suicide.
Supplementary material. The supplementary material for this article can be found at https://doi.org/10.1017/S1478951522001006.