Introduction
Psychiatric advance directives (PADs) are documents that enable individuals with mental health conditions to specify their treatment preferences for future mental health crises (Henderson et al. Reference Henderson, Swanson, Szmukler, Thornicroft and Zinkler2008). Introduced to enhance service user autonomy, PADs allow individuals to exert agency over their care even when they are unable to make decisions based on their deeply held values and preferences (Scholten et al. Reference Scholten, Gieselmann, Gather and Vollmann2019). The benefits of PADs are well-documented in the literature. PADs are associated with enhanced service user autonomy, improved therapeutic relationships, and the prevention of psychiatry-related harms (Braun et al. Reference Braun, Gaillard, Vollmann, Gather and Scholten2023). Evidence from randomized controlled trials further indicates that they can significantly reduce involuntary hospital admissions (de Jong et al. Reference de Jong, Kamperman, Oorschot, Priebe, Bramer, van de Sande, van Gool and Mulder2016; Molyneaux et al. Reference Molyneaux, Turner, Candy, Landau, Johnson and Lloyd-Evans2019; Tinland et al. Reference Tinland, Loubière, Mougeot, Jouet, Pontier, Baumstarck, Loundou, Franck, Lançon and Auquier2022). Despite these advantages and high interest among service users, PAD adoption remains low (Braun et al. Reference Braun, Gaillard, Vollmann, Gather and Scholten2023). This is due not only to practical implementation challenges but also to a lack of “buy-in” among professionals (Shields et al. Reference Shields, Pathare, van der Ham and Bunders2014).
Although many professionals endorse PADs in theory (Elbogen et al. Reference Elbogen, Swartz, Van Dorn, Swanson, Kim and Scheyett2006; Morrissey, Reference Morrissey2015; Swanson et al. Reference Swanson, Swartz, Hannon, Elbogen, Wagner, McCauley and Butterfield2003; Swartz et al. Reference Swartz, Swanson, Ferron, Elbogen, Dorn, Kim and Scheyett2005; Thom et al. Reference Thom, O’Brien and Tellez2015; van Dorn et al. Reference Van Dorn, Swartz, Elbogen, Swanson, Kim, Ferron, McDaniel and Scheyett2006; Wilder et al. Reference Wilder, Swanson, Bonnie, Wanchek, McLaughlin and Richardson2013), they seldom provide service users with the necessary support to draft a PAD in practice (Braun et al. Reference Braun, Gaillard, Vollmann, Gather and Scholten2023). Beyond practical challenges such as lack of knowledge and time, professionals also have fundamental concerns about PADs (Shields et al. Reference Shields, Pathare, van der Ham and Bunders2014). In particular, clinicians have raised concerns about service users employing PADs to refuse all treatment or document inappropriate treatment requests, service users’ capacity to draft PADs, poor-quality PAD content, changes of mind after PAD completion, risk of violence due to adhering to treatment refusals documented in PADs, and potential liability issues (Atkinson et al. Reference Atkinson, Garner, Stuart and Patrick2003; Elbogen et al. Reference Elbogen, Swartz, Van Dorn, Swanson, Kim and Scheyett2006; Farrelly et al. Reference Farrelly, Lester, Rose, Birchwood, Marshall, Waheed, Henderson, Szmukler and Thornicroft2016; Morrissey, Reference Morrissey2015; O’Connell & Stein, Reference O’Connell and Stein2005; Srebnik & Brodoff, Reference Srebnik and Brodoff2003; Swanson et al. Reference Swanson, Swartz, Hannon, Elbogen, Wagner, McCauley and Butterfield2003; Swartz et al. Reference Swartz, Swanson, Ferron, Elbogen, Dorn, Kim and Scheyett2005; Thom et al. Reference Thom, O’Brien and Tellez2015; van Dorn et al. Reference Van Dorn, Swartz, Elbogen, Swanson, Kim, Ferron, McDaniel and Scheyett2006; Wilder et al. Reference Wilder, Swanson, Bonnie, Wanchek, McLaughlin and Richardson2013;).
Treatment refusals outlined in PADs can create ethical conflicts between service user autonomy and professionals’ duty to promote the service user’s well-being or the state’s duty to protect the interests of others (Appelbaum Reference Appelbaum2004; Reference Appelbaum2006; Swanson et al. Reference Swanson, McCrary, Swartz, Elbogen and Van Dorn2006b). The legal case of Nancy Hargrave highlights the complexities of such refusals. Hargrave, a woman with schizophrenia and a history of multiple hospital admissions in Vermont, documented in her PAD a refusal of “any and all antipsychotic, neuroleptic, psychotropic, or psychoactive medications” (Appelbaum, Reference Appelbaum2004, 751). In Hargrave v. Vermont, the U.S. Court of Appeals for the Second Circuit ruled that Vermont’s legal provisions for overriding PADs of involuntarily committed individuals constituted unjustified differential treatment, violating the federal Americans with Disabilities Act. Appelbaum feared that if many service users were to use PADs for extensive treatment refusals, “hospitals might well begin to fill with patients whom they could neither treat nor discharge” (2004, 752).
While professionals generally prefer non-binding PADs and perceive treatment refusals as a major obstacle to their implementation (Elbogen et al. Reference Elbogen, Swartz, Van Dorn, Swanson, Kim and Scheyett2006; Morrissey, Reference Morrissey2015; Swanson et al. Reference Swanson, Van McCrary, Swartz, Van Dorn and Elbogen2007; Thom et al. Reference Thom, O’Brien and Tellez2015; van Dorn et al. Reference Van Dorn, Swartz, Elbogen, Swanson, Kim, Ferron, McDaniel and Scheyett2006), their support increases when they are granted discretion in overriding PADs (Elbogen et al. Reference Elbogen, Swartz, Van Dorn, Swanson, Kim and Scheyett2006). Conversely, empirical evidence suggests that service users favor legally binding PADs and perceive override provisions as a barrier to completing them (Braun et al. Reference Braun, Gaillard, Vollmann, Gather and Scholten2023). Though this situation seems to present an impasse, the prospects for PAD implementation may not be so dim. Two decades after Hargrave, a systematic review found that PAD content generally aligns with professional practice standards and that only a very small minority of service users use PADs to refuse all psychiatric medication in advance (Gaillard et al. Reference Gaillard, Braun, Vollmann, Gather and Scholten2023).
The evidence on professionals’ perspectives on PADs in Germany is limited. Available evidence suggests low uptake and slow implementation (Borbé et al. Reference Borbe, Jaeger, Borbe and Steinert2012; Radenbach et al. Reference Radenbach, Falkai, Weber-Reich and Simon2014; Weide et al. Reference Weide, Vrinssen, Karasch, Blumenröder, Staninska, Engemann, Banger, Grümmer, Marggraf, Muysers, Rinckens, Scherbaum, Supprian, Tönnesen-Schlack, Mennicken, Zielasek and Gouzoulis-Mayfrank2023), despite relatively high professional endorsement (Gieselmann et al. Reference Gieselmann, Simon, Vollmann and Schöne-Seifert2018; Grätz & Brieger, Reference Grätz and Brieger2012; Hotzy et al. Reference Hotzy, Cattapan, Orosz, Dietrich, Steinegger, Jaeger, Theodoridou and Bridler2020; Radenbach et al. Reference Radenbach, Falkai, Weber-Reich and Simon2014). While professionals’ concerns have not been extensively studied, existing research suggests worries about the quality of PAD content (Gieselmann et al. Reference Gieselmann, Simon, Vollmann and Schöne-Seifert2018; Hotzy et al. Reference Hotzy, Cattapan, Orosz, Dietrich, Steinegger, Jaeger, Theodoridou and Bridler2020). One survey indicated that professionals do not perceive PADs as limiting their therapeutic options and generally find preferences regarding ECT useful (Gieselmann et al. Reference Gieselmann, Simon, Vollmann and Schöne-Seifert2018). A published case report described a refusal of ECT outlined in a PAD leading to a situation where the patient could neither be treated appropriately nor discharged (Anonymous, 2014), sparking debate among ethicists and legal scholars (Gather et al. Reference Gather, Henking, Juckel and Vollmann2016; Gather & Vollmann, Reference Gather and Vollmann2014; Müller et al. Reference Müller, Gather, Gouzoulis-Mayfrank, Henking, Koller, Saß, Steinert and Pollmächer2024; Olzen & Lilius-Karakaya, Reference Olzen and Lilius-Karakaya2014).
The aim of the present study was to explore professionals’ perspectives on ethical conflicts arising from treatment refusals in legally binding PADs. To our knowledge, it is the first in-depth study on this topic.
Methods
Design
We employed a qualitative research design using semi-structured interviews and thematic data analysis to gain in-depth insight into professionals’ perspectives on ethical conflicts arising from treatment refusals in PADs. We prompted discussions about these ethical conflicts using a case report from the literature (Anonymous, 2014). An English translation of the original German case report can be found in the Supplementary Materials. The interdisciplinary research team comprised expertise in bioethics, medicine, philosophy, psychiatry, and sociology. The team included members with clinical experience (AG, JG, and JV) and lived experience (AW). The research team endorsed the increase of service user autonomy in mental healthcare and held generally favorable attitudes towards PADs based on their reading of the literature. At the same time, the team was open to hearing about concerns regarding PADs and challenges in their implementation. We used the COREQ checklist to guide reporting (O’Brien et al. Reference O’Brien, Harris, Beckman, Reed and Cook2014).
Context
A recent international legal comparison revealed high variability in PAD policies across countries and highlighted Germany’s unique position in the international legal landscape ( Gloeckler et al. Reference Gloeckler, Scholten, Weller, Keene, Pathare, Pillutla, Andorno and Biller-Andorno2025). In Germany, advance directives are governed by the federal guardianship law and apply equally to physical and mental health treatment (Henking & Scholten, Reference Henking, Scholten, Kong, Coggon, Cooper, Dunn and Keene2023). The law permits individuals to document treatment requests and refusals for situations in which they lack decision-making capacity (henceforth: capacity) (Sec 1827 no. 1 BGB). A PAD is valid if the person was at least 18 years old and had capacity at the time of drafting, and the document is available in written form and signed by the person and specifies clear treatment preferences and applicable circumstances (Henking & Scholten, Reference Henking, Scholten, Kong, Coggon, Cooper, Dunn and Keene2023). While prior information disclosure by a physician, certification of capacity, regular updating, and notary confirmation were debated in the legislative process, they were ultimately not included as validity requirements (Vollmann, Reference Vollmann2012). Valid and applicable PADs are legally binding in Germany (Gather et al. Reference Gather, Henking, Juckel and Vollmann2016; Grözinger et al. Reference Grözinger, Olzen, Metzmacher, Podoll and Schneider2011; Henking & Scholten, Reference Henking, Scholten, Kong, Coggon, Cooper, Dunn and Keene2023; Müller et al. Reference Müller, Gather, Gouzoulis-Mayfrank, Henking, Koller, Saß, Steinert and Pollmächer2024). The guardianship law does not allow for overriding treatment refusals in PADs, because it stipulates that involuntary treatment must align with a person’s documented treatment preferences (Sec 1832 para. 2 no. 3 BGB).
Sampling method and participants
The inclusion criteria required that participants were mental health professionals with professional experience with PADs or joint crisis plans. We recruited participants via email by distributing an information flyer to mental health hospitals with established expertise in the use of PADs or joint crisis plans. Additionally, we distributed the flyer via the mailing list of an interdisciplinary working group on ethics in psychiatry in Germany. We included a total of 14 participants to ensure a diverse range of perspectives, including four psychiatrists, four mental health nurses, three psychologists, two social workers, and one peer support worker. Nine participants were male and five were female, and they were aged between 38 and 63 years old. Participants were employed in various mental health hospitals across the German states of Baden-Württemberg, Hamburg, Hesse, Lower Saxony, and North Rhine-Westphalia. While two participants were known to some members of the research team via professional networks, none of the interviewers had any personal or professional relationship with the individuals they interviewed.
Data collection
AG, AW, and JG carried out semi-structured interviews between May 2019 and February 2020 in office spaces within the mental health hospitals in which participants were employed. At the time of data collection, AG was a licensed physician and researcher in a medical ethics department, AW was trained as a peer support worker and employed as a researcher with lived experience in a psychiatry department, and JG worked as a psychiatrist at a mental health hospital and as a researcher in a psychiatry department. AG, AW, and JG developed the interview guide based on an informal review of the literature on professionals’ perspectives on PADs and joint crisis plans. It comprised three sections: (1) prior experiences with and perspectives on PADs and joint crisis plans, (2) ethical challenges in their clinical application, and (3) recommendations for their clinical implementation. This article focuses on findings from the second section, and we present the results from the first and third sections in a separate publication.
The duration of the interviews ranged from 22 to 77 minutes, with a mean length of 50 minutes. We continued data collection until we reached data saturation. We prompted discussions on ethical challenges in the clinical implementation of PADs by presenting participants with the case report and using probing questions to explore professionals’ perspectives on the ethical challenges in clinical PAD implementation. The Supplementary Materials contain an English translation of the full case report. We summarize the case report below. Table 1 contains sample questions from the interview guide.
We audio-recorded the interviews, transcribed the audio recordings verbatim, de-identified transcripts by replacing identifiable information with placeholders, and assigned a randomly generated pseudonymization code to the transcripts and the basic sociodemographic data. We did not return transcripts to participants for comments or corrections. MS translated the interview excerpts presented in this article into English after we completed the data analysis. We do not report the professional roles of participants for individual interview excerpts to prevent inferences about individuals and maintain the anonymity of the research participants.
Table 1. Sample questions from the interview guide
Summary of the case report
A 42-year-old patient with schizophrenia was involuntarily admitted to a psychiatric hospital because of an acute psychotic episode, characterized by persecutory delusions, delusions of poisoning, auditory hallucinations, and severe agitation. Despite high-dose pharmacological treatment, his condition remained largely refractory, leading to frequent aggressive outbursts, physical assaults on staff and fellow patients, and severe self-neglect, including refusal of food and fluid intake, resulting in significant weight loss. The treatment team considered ECT. The patient had previously completed a PAD in which he explicitly refused ECT. The PAD was based on a template from a patient advocacy group that framed ECT as “an especially brutal and dangerous treatment.” The outpatient psychiatrist’s notes indicated that the patient signed the advance directive while having capacity. The patient’s medical history revealed that he had previously benefited from court-mandated ECT and could return to living in the community after the treatment. Attempts to discuss his health status and the treatment options were unsuccessful due to severe formal thought disorder. The patient made notes indicating a strong desire for freedom and independence, but due to his medical situation, he could not return to independent living. The treatment team decided to respect the patient’s refusal of ECT in his PAD. After ten months of involuntary detention, the patient was transferred to a secure long-term psychiatric facility.
Data analysis
We analyzed the data based on the original German transcripts following qualitative content analysis according to Kuckartz (2024). The data analysis combined deductive and inductive coding methods. After familiarizing themselves with all the transcripts, AG and AW developed an initial coding frame based on the structure of the interview guide and the informal review of the relevant literature. They subsequently coded the transcripts independently using the software MAXQDA 18 and adapted the initial coding scheme to accommodate themes emerging from the data. AG and AW compared their coding and discussed coding disagreements first among each other and then with JG, JV, MS, and SP until the team reached a consensus. The research team also discussed the final themes and subthemes in light of the literature.
Results
Overview of the results
Professionals characterized the case as extreme but not uncommon. Although most felt a duty to respect the PAD, they were also inclined to override it to promote the well-being and restore the autonomy of the service user and to protect others. Those leaning toward overriding the PAD tended to question its validity and applicability, raising doubts about the adequacy of information disclosure, the voluntariness of the person’s decision, the capacity of the person at the time of drafting, and the irrevocability of the PAD. We identified no substantial differences in perspectives across professional roles in the qualitative analysis.
First impressions of the case
At the outset of the interviews, we asked participants about their impressions of the case. Most described it as an extreme case, yet many indicated having encountered similar cases in their clinical practice. Only one participant reported never having encountered a comparable case. Participants tended to characterize the case as “tragic” or as an “ethical dilemma,” emphasizing its multifaceted nature and the need to weigh multiple factors to determine the appropriate course of action.
Reasons for PAD compliance: respecting service user autonomy
Respecting autonomy was a central theme in the interviews, though interpretations of the implications of this concept varied among participants. Several participants emphasized the legal force of PADs, arguing that, irrespective of how well-informed the service user was at the time of drafting, the PAD represented an exercise of self-determination that should be upheld. One participant noted, “All in all, I believe that this is also part of free self-determination. […] Everyone is entitled, according to the law, to just refuse certain things, even if they have not been informed about them” (Professional 12). Respecting autonomy, according to these participants, involves being prepared to accept the potential negative consequences of service users’ decisions. As this participant put it, “But that’s the way it is. He obviously made that decision… made that decision for himself” (Professional 12).
Others viewed respect for service user autonomy as requiring PAD adherence in principle unless an immediate risk of severe harm was evident. One participant stated, “Well, I have to adhere to such an advance directive, and I would only not comply with it if I saw an imminent danger” (Professional 10).
Some professionals voiced concerns that the increasing emphasis on autonomy rights in mental healthcare could lead to an increase in ethically fraught situations like the one presented in the case report. This was independent of professionals’ positions on how to resolve the ethical conflict in the case at hand. One participant prepared to adhere to the PAD remarked, “These are the dilemmas that I see coming our way… all of them, all due to the strengthening of the right to self-determination” (Professional 12). The participant thought the case illustrated that recent legal reforms to strengthen service user autonomy achieve precisely the opposite: “I’m convinced that this was not at all the intention behind the introduction of the… um… UN-CRPD [Convention on the Rights of Persons with Disabilities] and… um… the law on advance directives. […] The reality now is that people do not receive treatment that would benefit them… treatment that could put them in a different situation again… and that it ultimately leads to them ending up in closed institutions. I mean, in this case, it’s really a farce, one has to say, that the patient in this case report now ends up in a closed care facility for the mentally ill. That is certainly not what he wanted.” (Professional 12).
Motivations for PAD override
While many participants felt obligated to respect the PAD, they also felt strongly inclined to override it. Their primary motivations for wanting to override the PAD were promoting service user well-being, protecting the legitimate interests of others, and restoring service user autonomy.
Promoting service user well-being
Many participants questioned whether strict adherence to the PAD truly served the service user’s best interests. One participant remarked, “In the case we have here, one really has to question whether um… the advance directive actually um… serves the well-being of the patient” (Professional 1). Participants frequently described a dilemma between providing optimal care and respecting service user autonomy. One participant stated, “Now, it [ECT] is not administered, and he ends up in a closed facility. That’s certainly not the best future outlook one could wish for. And I have experienced various cases where patients were denied treatment due to these emerging ethical issues” (Professional 11).
When discussing the service user’s well-being, participants did not focus solely on medical factors such as symptom remission. They also considered the individual’s preferences, particularly his preference for regaining independence. Several participants pointed out that the documented refusal of treatment had resulted in a situation at odds with the person’s desire to live freely.
While few participants advocated overriding the PAD by administering ECT involuntarily out of concern for the person’s well-being, some stated they would attempt to persuade the service user to reconsider his refusal. One participant explained, “I would leave no stone unturned to provide good treatment. And that, of course, includes… um… considering and accep… um… considering the advance directive that the patient has written. But ultimately, if I still have the feeling […] that a certain treatment helped in the past, then one must certainly make every effort to obtain assent after all” (Professional 10).
Protecting the legitimate interests of others
Most participants argued that overriding the PAD and administering ECT involuntarily would not be justified solely to prevent harm to others. At the same time, they generally agreed that PADs cannot be used to refuse the use of coercive measures like isolation and mechanical restraint in situations of danger to others. One participant said, “I do think there should also be protection of… you know… the… um… general public. An advance directive should not extend so far that it grants someone the right… um… to endanger or harm others” (Professional 1). Participants generally thought that regardless of the person’s PAD, coercive measures should be used as a last resort to prevent harm to others: “The rights of third parties must also be protected […] From my perspective, this means that, if necessary, coercive measures must be applied to prevent harm to others. That… um… that is where the limit of an advance directive is reached… um… when it comes to the endangerment of third parties” (Professional 5).
Views on involuntary commitment were more mixed. Although most participants believed that potential harm to others did not justify overriding the PAD, many saw it as a reason for continued involuntary hospitalization. One participant noted, “If I were afraid that he might attack someone on the street, I wouldn’t discharge him… I would have him committed” (Professional 3). Others questioned the utility of hospitalization if no treatment could be provided, arguing, for example, that “a hospital is not a place to live long-term” (Professional 1). Participants who voiced this opinion were divided on how to proceed from there. One participant entertained the thought that perhaps the person should be discharged, noting that “sometimes we discharge patients where we have concerns [about safety]” (Professional 4). Others saw a transfer to a secure long-term mental health facility as the logical consequence of the PAD and the need to protect citizens against harm.
Restoring service user autonomy
The analysis revealed a paradox regarding service user autonomy: while respect for service user autonomy emerged as the main reason to respect the PAD, restoring service user autonomy emerged as an important reason to override it. Participants agreed that the service user lacked capacity, and some argued that overriding the PAD and administering ECT involuntarily could restore capacity. One participant explained, “I would favor making every effort to bring the patient into a state where they are capable of self-determination again, so that the decision can be reviewed once more. For me, a key goal is always to restore the ability for self-determination” (Professional 1).
Others likewise understood restoring autonomy as a reason for PAD override, yet understood autonomy as freedom of movement rather than choice. They observed that respecting the PAD had led to a situation in which the person’s freedom of movement and independence of living are severely curtailed: “But when I see that, as a consequence […] he… well… ended up in a permanently closed facility […] then, of course, he has completely lost the fundamental value of his individual freedom” (Professional 9). They argued that overriding the PAD by administering ECT could restore autonomy in this sense.
Reasons to doubt PAD validity and applicability
Within the German legislative framework, professionals cannot override valid and applicable PADs. Consequently, those inclined to override the PAD focused on scrutinizing its validity and applicability. Participants raised doubts about the adequacy of the information disclosure and the voluntariness of the decision-making process, the service user’s capacity at the time of drafting, and the irrevocability of the PAD.
Doubts about information disclosure and voluntariness
Many participants questioned whether the service user had been sufficiently informed when drafting the PAD. Skepticism arose because the service user had received no consultation from a physician and had instead relied solely on information provided by an anti-psychiatric organization. Participants perceived this information as misleading and thought the structure of the template (e.g. the option to tick boxes for blanket treatment refusals) and the explanatory notes on ECT constituted undue influence. One participant remarked, “One can certainly question whether… um… how well-informed the patient actually was, despite everything. Especially considering that ECT is described here as a particularly brutal form of treatment and that it is recommended to reject it in an advance directive. That’s why one might wonder, or even doubt, whether the information available to the patient at the time of drafting was truly neutral” (Professional 2). As a result, some doubted whether the PAD contained a valid refusal of ECT.
Doubts about capacity
Participants identified the retrospective assessment of the service user’s capacity at the time of drafting the PAD as a major difficulty in determining the validity of the PAD. Many participants emphasized that they would trust the physician who had certified the service user’s capacity at the time, refraining from questioning their colleague’s assessment. One participant stated, “Well, I mean, one obviously has to trust the colleague. I would find it quite difficult to ignore that” (Professional 10).
Some participants felt it was necessary to investigate further. One remarked, “Ultimately, it would be necessary to thoroughly examine to what extent this [capacity to consent] was actually present at the time of drafting. That means assessing to what extent there was an understanding of information, insight into the illness and treatment, and ability of reasoning, particularly concerning the possible consequences of not refusing electroconvulsive therapy” (Professional 3). One participant suggested consulting the psychiatrist who had certified the service user’s capacity to gather information on the individual’s mental state at the time of drafting. Others proposed arranging an ethics consultation, including not only the certifying psychiatrist but also relatives of the service user, to achieve a broader understanding of the circumstances.
One participant cautioned against questioning capacity in retrospect because it might be a mere means to invalidate PADs for paternalistic reasons. While accepting that further investigation might be warranted if there were concrete indications for impaired capacity, they warned, “with that logic, one could override almost any advance directive if desired […] And I find that concerning.” (Professional 2).
PAD revocation
Participants also debated whether a service user’s revocation of a PAD should be honored. Most participants agreed that if a service user assented to treatment, they would interpret this assent as a revocation of the PAD, even if the service user lacks capacity. One participant explained, “I don’t want to find myself in a situation… um… where I stand in front of a patient and say, ‘But five years ago, you said you didn’t want this or that,’ while the patient is right there in front of me saying, ‘But I want it now. I need it now.’ In such cases, if the acutely expressed will […] most likely serves the patient’s best interest, I think it makes more sense to rely on what they express in an acute situation (Professional 13). Other participants argued that service users’ contemporaneous preferences always prevail because service users can change their minds. One participant stated, “The current will is always decisive. Because if an advance directive is, let’s say, three or four years old… we all know from our own experience that we sometimes change our minds and think differently about things… Of course, an advance directive is important for us, but obviously any directive […] can be revoked by the patient” (Professional 4).
Other participants argued that there are circumstances in which PAD revocation should not be accepted, particularly when service users lack capacity. One participant stated, “We would discuss the advance directive, and we might say, ‘At the moment, we consider you to be so ill that we will rely on it [the advance directive]’” (Professional 6).
One participant highlighted the ethical tension in selectively accepting or rejecting PAD revocations, depending on whether the revocation is compatible with clinical judgment. He argued, “As professionals, we have to be careful not to expose ourselves to the… um… suspicion of arbitrariness… that we handle things in a way that suits us at the moment. For example, if an advance directive exists, but the person is currently incapable of giving informed consent yet agrees to treatment, and we then accept this agreement as consent despite their incapacity” (Professional 5).
Ways of addressing the ethical conflict
When asked whether or not they would override the PAD and administer ECT, many participants expressed ambivalence. They emphasized that engaging in collegial discussions was crucial for making well-informed decisions. One participant stated, “I wouldn’t know. I think it’s important to hear all perspectives in the ethics consultation and then make a decision together. And… um… I can’t say that definitively right now” (Professional 14). Others argued in favor of PAD override: “I would try to treat him with electroconvulsive therapy and take the required legal steps and… um… only if it is supported by a court decision, of course” (Professional 11). However, even those arguing for a position expressed ambivalence and accepted the possibility of reasonable disagreement. This participant, for example, added, “I can also very well imagine… This is of course a case where different opinions are possible… and that’s why I believe… um… that the alternative approach is just as justifiable” (Professional 11). Conversely, some participants stated that they would not administer ECT over the person’s objection. One participant remarked, “In collaboration with the senior physician I work closely with, we would argue our case. But if the patient refuses, then it doesn’t happen” (Professional 4).
Many participants highlighted the role that ethics consultations could play in resolving cases like this one. One participant noted that ethics consultations provide an opportunity for structured reflection and interdisciplinary discussion: “From my perspective, they offer essentially two main advantages: First, one once more reflects… um… on one’s position. That is … um… why do I actually believe that this would be helpful or not helpful? And they promote discussion among professionals” (Professional 5).
Discussion
Synthesis of findings
This study explored professionals’ perspectives on treatment refusals outlined in PADs. Most professionals described the discussed case as an extreme yet not unfamiliar case and expressed ambivalence about how to address the ethical conflict resulting from it. They interpreted the ethical conflict as one between respecting service user autonomy, on the one hand, and promoting service user well-being and autonomy as well as protecting others against harm, on the other. Many professionals felt inclined to override the PAD despite legal provisions for binding PADs. Aware of these legal provisions, participants justified their inclination to override the PAD by raising doubts about its validity and applicability.
Interpretation of findings
The inclination to PAD override among participants in our interview study aligns with findings in the literature of a preference for non-binding PADs and discretion in PAD override among professionals (Elbogen et al. Reference Elbogen, Swartz, Van Dorn, Swanson, Kim and Scheyett2006; Morrissey, Reference Morrissey2015; Swanson et al. Reference Swanson, Van McCrary, Swartz, Van Dorn and Elbogen2007; Thom et al. Reference Thom, O’Brien and Tellez2015; van Dorn et al. Reference Van Dorn, Swartz, Elbogen, Swanson, Kim, Ferron, McDaniel and Scheyett2006). Our study showed that even if some professionals were prepared to respect the PAD despite potential negative consequences, most seriously considered overriding the PAD out of concerns for the service user’s well-being and the risk of violence – two concerns documented in the literature (Farrelly et al. Reference Farrelly, Lester, Rose, Birchwood, Marshall, Waheed, Henderson, Szmukler and Thornicroft2016; Morrissey, Reference Morrissey2015; van Dorn et al. Reference Van Dorn, Swartz, Elbogen, Swanson, Kim, Ferron, McDaniel and Scheyett2006). Restoring service user autonomy as a reason for PAD override, as found in our study, is an aspect that has not yet been reported in the literature, although Farrelly et al. (Reference Farrelly, Lester, Rose, Birchwood, Marshall, Waheed, Henderson, Szmukler and Thornicroft2016) reported concerns about service users’ future autonomy.
Participants’ perspectives did not notably differ by professional role in our study. While our qualitative research design did not allow for assessing statistical correlations between professionals’ characteristics and their preference for PAD override, an earlier survey suggested that PAD override is more likely among psychiatrists working in emergency departments, and those who were more concerned about violence risk and lack of insight and who were more legally defensive (Swanson et al. Reference Swanson, Van McCrary, Swartz, Van Dorn and Elbogen2007). Our analysis did not reveal legal defensiveness and concerns about liability as factors inclining professionals toward PAD override. An explanation is that participants were aware of the German legal framework that prevents professionals from overriding treatment refusals in valid and applicable PADs. By contrast, earlier studies carried out in other countries generally found a relatively low awareness of PAD regulation among professionals (van Dorn et al. Reference Van Dorn, Swartz, Elbogen, Swanson, Kim, Ferron, McDaniel and Scheyett2006; Wilder et al. Reference Wilder, Swanson, Bonnie, Wanchek, McLaughlin and Richardson2013).
Professionals disputed the validity and applicability of the PAD by raising doubts about information disclosure and voluntariness, the service user’s capacity when drafting, and the PAD’s irrevocability. Earlier studies also found concerns about adequate information provision among professionals (Srebnik & Brodoff, Reference Srebnik and Brodoff2003), particularly regarding service users who had no personal experience with either the mental health condition or the treatment (Atkinson et al. Reference Atkinson, Garner, Stuart and Patrick2003). The service user in the discussed case report had personal experience in both respects. In line with this, professionals’ concerns about inadequate information provision and undue influence were directed at the explanatory notes in the PAD template from the anti-psychiatric organization. Legally speaking, these concerns are unlikely to cast doubt on the PAD’s validity, given that information disclosure by a physician is not a validity requirement for PADs in Germany (Vollmann, Reference Vollmann2012) and the legal threshold for determining undue influence is typically quite high.
The concerns professionals in our study raised about the service user’s capacity when drafting the PAD also confirm findings from the literature (Srebnik & Brodoff, Reference Srebnik and Brodoff2003; Swanson et al. Reference Swanson, Van McCrary, Swartz, Van Dorn and Elbogen2007; Wilder et al. Reference Wilder, Elbogen, Swartz, Swanson and Van Dorn2007, Reference Wilder, Swanson, Bonnie, Wanchek, McLaughlin and Richardson2013). However, in Germany, as elsewhere, adults are presumed to have capacity regardless of whether they have a mental disorder (Grisso & Appelbaum, Reference Grisso and Appelbaum1998; Srebnik & Kim, Reference Srebnik and Kim2006; Scholten & Haberstroh, Reference Scholten and Haberstroh2024), and it is difficult for evidential reasons to rebut this presumption retrospectively. Empirical evidence suggests, moreover, that people with mental health conditions living in the community generally have the capacity to complete a PAD despite major functional impairments (Srebnik et al. Reference Srebnik, Appelbaum and Russo2004).
Participants in our study understood that if the individual were to revoke the PAD, its validity would consequently be undermined. While professionals generally agreed that the person in the case report lacks capacity, some would try to persuade the person into assenting to ECT. The German law states that PADs “can be revoked informally at any time” (Sec 1827 no 1 BGB). The meaning of this statement has not yet been settled in court. While there is a relatively wide agreement among legal scholars that capacity is required for PAD revocation (Henking & Bruns, Reference Henking and Bruns2014), professionals might have taken the formulation in the law as providing an opportunity to override the PAD based on the person’s assent to treatment.
Policymakers might wonder where to go from here. Appelbaum (Reference Appelbaum2006) argues that the scope of PADs should be limited to prevent ethical conflicts arising from treatment refusals in PADs, stating that “patients who qualify for involuntary commitment should not be able to trump that process by means of an advance directive” (2006, 396). It should be noted, however, that ethical conflicts arising from treatment refusals in PADs are rare. Empirical evidence indicates that the great majority of PADs have content that conforms to professional practice standards (Gaillard et al. Reference Gaillard, Braun, Vollmann, Gather and Scholten2023). Moreover, service users see PAD override as a major obstacle to PAD completion (Braun et al. Reference Braun, Gaillard, Vollmann, Gather and Scholten2023). The legal dictum “hard cases make for bad law” is thus a case in point. Although granting professionals leeway in PAD override may indeed prevent a small number of hard cases, this will at the same time raise barriers preventing many service users from enjoying their autonomy rights. This dilemma must be addressed in policy analysis.
Regardless of these policy questions, professionals seem well-advised to provide service users with the support they need to draft a PAD. Evidence suggests that providing support dramatically increases PAD uptake (Swanson et al. Reference Swanson, McCrary, Swartz, Elbogen and Van Dorn2006a) and uniquely contributes to the realization of some of the key benefits of PADs, notably the improvement of the therapeutic relationship (Braun et al. Reference Braun, Gaillard, Vollmann, Gather and Scholten2023). Providing professional support for PAD completion is also likely to reduce the likelihood that service users use these documents to articulate far-reaching treatment refusals.
Strengths and limitations
To the best of our knowledge, this is the first study to systematically explore professionals’ perspectives on ethical conflicts arising from treatment refusals outlined in PADs. The inclusion of participants from different geographical regions in Germany ensured that findings were not limited to local contexts. While the use of a case report enabled us to prompt detailed and in-depth discussions about the study subject, it may also have framed the interview in a particular way. The study included only professionals with experience working with PADs, who may hold more favorable views of the instrument than those without such experience. We did not report the professional roles of participants for individual interview excerpts to maintain anonymity. The ethical conflict we discussed will have different contours in jurisdictions where PADs are unregulated or where professionals have discretion in PAD override.
Conclusions
Legally binding PADs may lead to ethical conflicts arising from treatment refusals outlined in PADs. While it remains unclear which policy response is appropriate, professionals can prevent these conflicts by providing service users with support in drafting PADs.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ipm.2025.10086.
Acknowledgements
We thank all study participants for their contribution to this work.
Author contributions
All authors contributed to the design of this study. AG, AW, and JG developed the interview guide and carried out the interviews. AG and AW analyzed the data and discussed disagreements and initial findings with all other authors. AG wrote the first draft of the results section. MS wrote the background, methods, and discussion sections and the second draft of the results section. All other authors revised the drafts critically for important intellectual content. All authors agreed to the article’s arguments and conclusions and gave their approval for the final version to be published.
Financial support
This work was supported by the German Federal Ministry of Education and Research (SALUS, grant number 01GP1792).
Competing interests
The authors declare none.
Ethical standard
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2024. The study protocol was approved by the research ethics committee of the medical faculty of the Ruhr University Bochum (Reg. No. 18-6583-BR).
Open access
