The Role of Humanization and Robustness of Large Language Models in Conversational Artificial Intelligence for Individuals With Depression: A Critical Analysis.

Unlabelled: Large language model (LLM)-powered services are gaining popularity in various applications due to their exceptional performance in many tasks, such as sentiment analysis and answering questions. Recently, research has been exploring their potential use in digital health contexts, particularly in the mental health domain. However, implementing LLM-enhanced conversational artificial intelligence (CAI) presents significant ethical, technical, and clinical challenges. In this viewpoint paper, we discuss 2 challenges that affect the use of LLM-enhanced CAI for individuals with mental health issues, focusing on the use case of patients with depression: the tendency to humanize LLM-enhanced CAI and their lack of contextualized robustness. Our approach is interdisciplinary, relying on considerations from philosophy, psychology, and computer science. We argue that the humanization of LLM-enhanced CAI hinges on the reflection of what it means to simulate "human-like" features with LLMs and what role these systems should play in interactions with humans. Further, ensuring the contextualization of the robustness of LLMs requires considering the specificities of language production in individuals with depression, as well as its evolution over time. Finally, we provide a series of recommendations to foster the responsible design and deployment of LLM-enhanced CAI for the therapeutic support of individuals with depression.

[The Basel model of principle-oriented clinical ethics consultation 2.0 : An introduction for psychiatry]. / Das Basler Modell der prinzipienorientierten klinischen Ethikkonsultation 2.0 : Eine Einführung für die Psychiatrie.

BACKGROUND: The effectiveness of clinical ethics counselling in medicine in terms of satisfaction of the participants, support of ethical decision-making, perceived impact on clinical care, moral competence and quality of communication is now empirically well validated. In psychiatry, more and more institutions have structures for clinical ethics consultation as well; however, there is still a lack of evaluative accompanying research on the benefits and differential indications of the various forms and models of clinical ethics counselling in psychiatry. AIM OF THE WORK: The article presents the principles and the step by step application of the principle-oriented clinical ethics consultation according to the Basel model 2.0. MATERIAL AND METHODS: The article is based on material and procedures developed at the Clinical Ethics Unit of the University Hospital Basel and the University Psychiatric Clinics Basel for the principle-oriented clinical ethics consultation according to the Basel model 2.0. RESULTS AND DISCUSSION: Principle-oriented clinical ethics consultation according to the Basel model 2.0 is suitable for dealing with moral issues and conflicts in psychiatric practice and for their ethical reflection; however, further evaluative accompanying research is needed on the benefits and differential indications of this and other models of clinical ethics support in psychiatry.

Scoping review of end-of-life care for persons with anorexia nervosa.

BACKGROUND: End-of-life (EOL) care is the part of palliative care intended for persons nearing death. In anorexia nervosa (AN), providing EOL care instead of coercing life-sustaining measures is controversial. The existing literature has not been synthesized yet. To clearly delineate differing views and identify open questions as well as areas of possible consensus, we conducted the first-ever synthesis of the existing literature. METHODS: We searched EMBASE, PubMed, PsycInfo, and Web of Science for scientific publications on forgoing coerced life-sustaining measures and/or providing EOL care for persons with AN who refuse life-sustaining measures, typically artificial nutrition. Palliative care outside of the EOL context and medical assistance in dying were not reviewed. As very little quantitative studies were identified, we qualitatively analyzed conceptual questions, ethical reasoning, legal aspects, stakeholder attitudes, practical aspects, stakeholder needs, and outcome. RESULTS: We identified 117 eligible publications from 1984 to 2023, mainly case reports (n=26 different cases) and ethical analyses. Conceptualizations of key terms such as terminality, futility, and decision-making capacity (DMC) in AN varied widely and were often value-laden and circular. Ethical reasoning centered on weighing the preservation of life versus quality of life in the context of uncertainty about DMC and likelihood of clinical remission. Studies on stakeholder attitudes reflected this challenge. In some cases, courts ruled against coerced life-sustaining measures and/or in favor of EOL care for persons with AN. While eligibility criteria were contested, recommendations for deliberating about and providing EOL care were consistent. We identified only one study on stakeholder needs and none on outcome. Case reports described quality of life under EOL care as good and death as the most frequent outcome but engagement in voluntary treatment and (partial) clinical remission in some. CONCLUSIONS: The debate around EOL care in AN needs consented, coherent terminology whose value base is reduced to a minimum and made transparent. While more empirical research into decision-making in AN and predictors of outcome might help reduce uncertainty, fundamental normative questions need to be addressed, for example regarding the ethico-legal significance of treatment refusals, the weighing of quantity versus quality of life and the appropriateness of diagnosis-based ethico-legal exceptionalism such as hard paternalism. More research is needed on outcome of and stakeholder needs in EOL care for persons with AN.

Palliative psychiatry: research, clinical, and educational priorities.

BACKGROUND: Palliative psychiatry has been proposed as a new clinical construct within mental health care and aims to improve quality of life (QoL) for individuals experiencing severe and persistent mental illness (SPMI). To date, explorations of palliative psychiatry have been largely theoretical, and more work is needed to develop its approaches into tangible clinical practice. METHODS: In this paper, we synthesize existing literature with discussions held at a one-day knowledge user meeting titled "A Community of Practice for Palliative Psychiatry" to generate priorities for research, clinical practice, and education that will help advance the development of palliative psychiatry. RESULTS: Palliative psychiatry will benefit from research that is co-produced by people with lived experience (PWLE) of mental illness, that clarifies contested concepts within mental health care and wider medicine, and that adapts existing interventions that have the potential to improve the QoL of individuals experiencing SPMI into the mental health care context. Specific methods and tools might be developed for use in clinical spaces taking a palliative psychiatry approach. More work must be done to understand the populations that might benefit from palliative psychiatry, and to mitigate mental health care providers' (MHCPs') anxieties about using these approaches in their work. As palliative psychiatry is developed, current MHCPs, trainees, individuals experiencing SPMI, and their loved ones will all require education about and orientation to this novel approach within mental health care. CONCLUSIONS: There are several priorities in research, clinical practice, and education that can help advance the development of palliative psychiatry. All future work must be considered through a human rights-based, anti-oppressive lens. Research projects, clinical models, and educational initiatives should all be developed in co-production with PWLE to mitigate the epistemic injustices common in mental health care.

Overestimating prevalence? Rethinking boundaries and confounders of moral distress.

Moral distress denotes a negative reaction to a morally challenging situation. It has been associated with adverse outcomes for healthcare professionals, patients and healthcare institutions. We argue that existing definitions, along with measures of moral distress, compromise the validity of empirical research. First, the definition and measurement of moral distress conflate moral events and psychological distress, even though they are distinct phenomena that should be assessed independently. Second, in many studies, there is a lack of clarity in distinguishing between moral and non-moral events. Finally, prior research on moral distress often overlooks the substantial body of evidence demonstrating the impact of diverse work-related factors, beyond moral events, on both distress and job retention. These challenges might undermine the effectiveness of interventions aimed at alleviating moral distress. We outline a comprehensive research agenda that encompasses conceptual clarifications, the refinement of data collection instruments, the design of studies and the application of appropriate statistical methods.

Comparing attitudes towards compulsory interventions in severe and persistent mental illness among psychiatrists in India and Switzerland.

BACKGROUND: Psychiatrists face a major ethical challenge when deciding whether to make use of coercive measures in the treatment process of patients suffering from severe and persistent mental illness (SPMI). As India and Switzerland show major cultural, political and financial differences, it is hypothesized that attitudes towards coercive measures among Indian and Swiss psychiatrists will vary too. Exploring differences in attitudes between cultures strengthens the critical reflection on one's own stances and in consequence, on our way of action. Especially when it comes to situations involving power imbalances between patients and health practitioners, self-reflection is essential to prevent ethically inappropriate behavior. METHODS: An online survey on aspects of care for patients with SPMI was sent to 3'056 members of the Indian Psychiatric Society between April and June 2020 and to 1'311 members of the Swiss Society for Psychiatry and Psychotherapy between February and March 2016. The respondents' answers were compared. This article deals with the questionnaire's items on autonomous decision making and the implementation of coercive measures in clinical practice. More precisely, participating psychiatrists were asked to rate the importance of patient's autonomy in general and their willingness to apply coercive measures regarding two specific case vignettes depicting a patient with schizophrenia and one with depression. The statistical analysis, namely descriptive data analysis and calculation of arithmetic means, Shapiro Wilks tests and Mann-Whitney U tests, was carried out using IBM SPSS Statistics version 27. RESULTS: Answers were received from 206 psychiatrists in India and 457 psychiatrists in Switzerland. Indian participants tended to value autonomous decision making as slightly less important than Swiss participants (62.2% vs. 91%, p =.01). Regarding a case of severe and persistent depression, psychiatrists in the Indian group were on average more in favor of acting against the wishes of the patient (55% vs. 34.1%, p <.0001) as well as of accepting a temporary decrease in quality of life due to coercion (40% vs. 23%, p =.008). Answers concerning a case of schizophrenia revealed that Indian participants were more in favor of acting against the patient's wishes than Swiss participants (39% vs. 37%, p =.007), whereas the comparison whether to accept a temporary decrease in quality of life regarding this case showed no significant difference (p =.328). CONCLUSIONS: The significant difference in attitudes towards coercive measures among Indian compared to Swiss psychiatrists found in this study might arise from a predominantly more collectivist society in India compared to Switzerland. Moreover, differences in financial resources, the organization of the health care system, and the historical background might have an influence. Continuous and critical reflection on one's own views and behavior is essential, especially if ethical principles and individual rights could be violated through a power imbalance, as in the case of coercive measures.

Palliative psychiatry for a patient with treatment-refractory schizophrenia and severe chronic malignant catatonia: case report.

BACKGROUND: Palliative psychiatry is an emerging field that suggests a role for palliative interventions in the management of severe and persistent mental illness (SPMI). Current literature describes using a palliative approach for patients with severe anorexia nervosa. To our knowledge, this is the first case report describing end-of-life care in a patient with treatment-refractory catatonic schizophrenia. CASE DESCRIPTION: We describe the case of a 49-year-old man with schizophrenia and severe chronic agitated/malignant catatonia who was hospitalized for ten months. Multiple treatment trials including medication such as neuroleptics and benzodiazepines, electroconvulsive therapy, and empiric interventions such as intravenous immunoglobulins were either not tolerated or did not result in clinically significant improvement. The patient continued to intermittently require intubation and sedation to control intractable behavioral and psychiatric disturbances. Ultimately, with collaboration of psychiatry, neurology, ethics, intensive care, and palliative care teams, the patient's parents decided to forgo further diagnostic testing and life-sustaining treatments. The patient died weeks later of aspiration pneumonia with good symptom control. CONCLUSIONS: This case permits discussion of palliative interventions in patients with SPMI such as treatment-refractory psychotic disorders who likely cannot achieve a quality of life that is acceptable to them. Here, it can be justified to prioritize relief of suffering and prevention of further burdensome interventions over treatment of the SPMI symptoms such as catatonia and even over keeping the patient alive.

Optimizing treatment expectations and decision making through informed consent for psychotherapy: A randomized controlled trial.

OBJECTIVE: The objective of this research was to determine the efficacy and safety of an optimized informed consent (OIC) consultation for psychotherapy. METHOD: We performed a randomized controlled superiority online trial involving 2 weeks of treatment and 3 months of follow-up. One hundred twenty-two adults with mental disorders confirmed by structured interview currently neither in out- nor inpatient psychotherapy (mean age: 32, gender identity: 51.6% female, 1.6% diverse), were randomized. Participants received an information brochure about psychotherapy for self-study (treatment as usual [TAU]; n = 61) or TAU plus a one-session OIC utilizing expectation management, contextualization, framing, and shared decision making (n = 61). The primary outcome was treatment expectations at 2-week follow-up. RESULTS: At 2-week follow-up, participants receiving OIC showed more positive treatment expectations compared to those receiving TAU only (mean difference: 0.70, 95% CI [0.36, 1.04]) with a medium effect size (d = 0.73). Likewise, OIC positively influenced motivation (d = 0.74) and adherence intention (d = 0.46). OIC entailed large effects on reduction of decisional conflict (d = 0.91) and increase of knowledge (d = 0.93). Participants receiving OIC showed higher capacity to consent to treatment (d = 0.63) and higher satisfaction with received information (d = 1.34) compared to TAU. No statistically significant group differences resulted for expected adverse effects of psychotherapy. Results were maintained at 3-month follow-up. Data sets for n = 10 cases (8.2%) were missing (postassessment n = 4, 2-week n = 6, 3-month follow-up n = 8). CONCLUSIONS: Explaining to patients how psychotherapy works via a short consultation was effective in strengthening treatment expectations and decision making in a nonharmful way. Further trials clarifying whether this effectively translates to better treatment outcomes are required. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Capacity to consent to psychotherapy: Reliability of the newly adapted German version of the MacArthur Competence Assessment Tool for Treatment for psychotherapy.

BACKGROUND: Patients' capacity to consent to treatment (CCT) is a prerequisite for ethically sound informed consent in psychotherapy. The MacArthur Competence Assessment Tool for Treatment (MacCAT-T) is a reliable instrument for assessing CCT. A German version was adapted to the psychotherapeutical context (MacCAT-PT) to investigate its reliability and possible influences of age, education and prior experience with psychotherapy on CCT in a mixed clinical sample. METHODS: N = 108 patients with indication for psychotherapy were recruited. The MacCAT-PT was administered by trained psychologists, took 20 min on average and was rated by the administering psychologist and an independent rater. Reliability statistics were investigated and regression analyses were conducted on MacCAT-PT scores and sociodemographic variables. RESULTS: Sufficient to moderate inter-rater reliability (ICC = 0.80) and internal consistency (α = 0.80) were found for the total sum score of the MacCAT-PT and its scales, Understanding (ICC = 0.79, α = 0.77), Reasoning (ICC = 0.57, α = 0.65) and Making a Choice (ICC = 0.57). Appreciation featured an unacceptable inter-rater reliability (ICC = -0.01). Regression analyses indicated no significant effects. CONCLUSION: These findings suggest that the MacCAT-PT is a reliable tool for assessing patients' overall CCT in psychotherapy. Psychometric properties of three scales were of good quality, while Appreciation needs to be reanalysed in patient samples with lower motivation for psychotherapy or limited CCT. The CCT may be suggested to be independent of age, education and prior experience. Future research should provide analyses focusing on structural and clinical validity in multiple clinical samples.

[Kulturelle Kompetenz in der klinischen Ethik]. / Kulturelle Kompetenz in der klinischen Ethik.

INTRODUCTION: Language, cultural identity, values, or migration experiences of patients, relatives and medical professionals have a significant influence on health care. In individual cases, this can lead to the perceptions of moral problems and conflicts, and to requests for clinical ethics support service. What is the role of clinical ethicists in the face of such perceptions? How does a culturally competent clinical ethics support service look like? In the present contribution, this question will be examined on the basis of reflections on the relationship between culture and ethics as well as the discussion about culturally competent medical ethics. On that basis, recommendations for a culturally competent clinical ethics support are derived.

Moral Intuitions About Futility as Prompts for Evaluating Goals in Mental Health Care.

Mental health professionals' moral intuitions about futility should prompt reevaluation of goals of care and care plans. Mostly, it will suffice to improve the care plan and/or slightly adjust the goal of care (eg, lower expectations), which is standard practice. Sometimes, however, all care plans that seek to reduce core symptoms (ie, that pursue a curative goal) are most likely futile and thus should not be imposed. Here, it may be in the patient's best interest to change the goal of care toward palliation (ie, harm reduction, relief of suffering, and best possible quality of life). Thus, futility can function as a moral counterweight to the duty to treat, helping mental health professionals find the right balance between over- and undertreatment.

Comparison of analog and digital patient decision aids for the treatment of depression: a scoping review.

Introduction: Patient decision aids (PDAs) are important tools to empower patients and integrate their preferences and values in the decision-making process. Even though patients with mental health problems have a strong interest in being more involved in decision making about their treatment, research has mainly focused on PDAs for somatic conditions. In this scoping review, we focus on patients suffering from depression and the role of PDAs for this patient group. The review offers an overview of digital and analog PDAs, their advantages and disadvantages as well as recommendations for further research and development. Methods: A systematic search of the existing literature guided by the Cochrane Handbook for Systematic Reviews and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses - extension for scoping reviews (PRISMA-ScR) was conducted. Three electronic literature databases with the appropriate thematic focus were searched (PubMed, PsycInfo, and Web of Science). The search strategy used controlled and natural language to search for the key concepts decision aids and depression. The articles were selected in a two-step process guided by predefined inclusion and exclusion criteria. We narratively synthetized information extracted from 40 research articles. Results: We included 40 articles in our review. Our review revealed that there is more focus on digital PDAs in research than in clinical practice. Digitalization can enhance the benefits of PDAs by developing tools that are more efficient, interactive, and personalized. The main disadvantages of both types of PDAs for the treatment of depression are related to time, dissemination, and capacity building for the health care providers. Digital PDAs need to be regularly updated, effective strategies for their dissemination and acceptance need to be identified, and clinicians need sufficient training on how to use digital PDAs. There is more research needed to study which forms of PDAs are most appropriate for various patient groups (e.g., older adults, or patients with comorbidities), and to identify the most effective ways of PDAs' integration in the clinical workflow. The findings from our review could be well aligned with the International Patient Decision Aids Standards. Discussion: More research is needed regarding effective strategies for the implementation of digital PDAs into the clinical workflow, ethical issues raised by the digital format, and opportunities of tailoring PDAs for diverse patient groups.

With great power comes great vulnerability: an ethical analysis of psychedelics' therapeutic mechanisms proposed by the REBUS hypothesis.

Psychedelics are experiencing a renaissance in mental healthcare. In recent years, more and more early phase trials on psychedelic-assisted therapy have been conducted, with promising results overall. However, ethical analyses of this rediscovered form of treatment remain rare. The present paper contributes to the ethical inquiry of psychedelic-assisted therapy by analysing the ethical implications of its therapeutic mechanisms proposed by the relaxed beliefs under psychedelics (REBUS) hypothesis. In short, the REBUS hypothesis states that psychedelics make rigid beliefs revisable by increasing the influence of bottom-up input. Put differently, patients become highly suggestible and sensitive to context during a psychedelic session, amplifying therapeutic influence and effects. Due to that, patients are more vulnerable in psychedelic-assisted therapy than in other therapeutic interventions; they lose control during a psychedelic session and become dependent on the therapeutic setting (including the therapist). This enhanced vulnerability is ethically relevant and has been exploited by some therapists in the past. Therefore, patients in current research settings and starting mainstream medical settings need to be well informed about psychedelics' mechanisms and their implications to give valid informed consent to treatment. Furthermore, other security measures are warranted to protect patients from the vulnerability coming with psychedelic-assisted therapy.

Conversational Artificial Intelligence in Psychotherapy: A New Therapeutic Tool or Agent?

Conversational artificial intelligence (CAI) presents many opportunities in the psychotherapeutic landscape-such as therapeutic support for people with mental health problems and without access to care. The adoption of CAI poses many risks that need in-depth ethical scrutiny. The objective of this paper is to complement current research on the ethics of AI for mental health by proposing a holistic, ethical, and epistemic analysis of CAI adoption. First, we focus on the question of whether CAI is rather a tool or an agent. This question serves as a framework for the subsequent ethical analysis of CAI focusing on topics of (self-) knowledge, (self-)understanding, and relationships. Second, we propose further conceptual and ethical analysis regarding human-AI interaction and argue that CAI cannot be considered as an equal partner in a conversation as is the case with a human therapist. Instead, CAI's role in a conversation should be restricted to specific functions.

Physician assisted-dying in mentally and somatically ill individuals in Switzerland: Protocol for survey-based study.

Background: Article 115 of the Swiss Penal Code (StGB) permits physician-assisted dying (PAD), provided it is not performed for "selfish reasons," and thus, occupies a special role in international comparison. However, the Swiss federal law does not regulate who exactly is entitled to access PAD, and there is no universal agreement in the concerned professional societies. Additional uncertainty arises when assessing the wish for PAD of a mentally ill person compared to a somatically ill person. Objectives: This study aims to contribute to the discussion of PAD among the mentally ill and to provide insight into the current situation in Switzerland. Methods: This is a monocentric prospective observational survey-based study. We will conduct an exploratory online/telephone survey about PAD in somatic vs. mental illness in Switzerland. The survey sample will comprise 10,000 Swiss residents of the general population from all three language regions (German, Italian, and French) as well as 10,000 medical professionals working in the seven states ("cantons") of Basel-Stadt, Basel-Landschaft, Aargau, Lucerne, Graubünden, Ticino, and Vaud. Opinions on PAD in mentally and somatically ill patients will be assessed using 48 different case vignettes. Each participant will be randomly assigned a somatic terminal, a somatic non-terminal, and a mental non-terminal case vignette. Furthermore, the attitude toward the ethical guidelines of the Swiss Medical Association of 2004, 2018, and 2022, as well as the stigmatization of mentally ill people will be assessed. Discussion: Physician-assisted dying in mentally ill persons is a highly relevant yet controversial topic. On the one hand, mentally ill persons must not be discriminated against in their desire for PAD compared to somatically ill persons while at the same time, their vulnerability must be considered. On the other hand, treating physicians must be protected in their ethical integrity and need security when judging PAD requests. Despite its relevance, data on PAD in the mentally ill is sparse. To regulate PAD for the mentally ill, it is therefore important for Switzerland-but also internationally-to gain more insight into the ongoing debate. Clinical trial registration: ClinicalTrials.gov, identifier: NCT05492461.

Optimized Informed Consent for Psychotherapy: Protocol for a Randomized Controlled Trial.

BACKGROUND: Informed consent is a legal and ethical prerequisite for psychotherapy. However, in clinical practice, consistent strategies to obtain informed consent are scarce. Inconsistencies exist regarding the overall validity of informed consent for psychotherapy as well as the disclosure of potential mechanisms and negative effects, the latter posing a moral dilemma between patient autonomy and nonmaleficence. OBJECTIVE: This protocol describes a randomized controlled web-based trial aiming to investigate the efficacy of a one-session optimized informed consent consultation. METHODS: The optimized informed consent consultation was developed to provide information on the setting, efficacy, mechanisms, and negative effects via expectation management and shared decision-making techniques. A total of 122 participants with an indication for psychotherapy will be recruited. Participants will take part in a baseline assessment, including a structured clinical interview for Diagnostic and Statistical Manual of Mental Disorders-fifth edition (DSM-5) disorders. Eligible participants will be randomly assigned either to a control group receiving an information brochure about psychotherapy as treatment as usual (n=61) or to an intervention group receiving treatment as usual and the optimized informed consent consultation (n=61). Potential treatment effects will be measured after the treatment via interview and patient self-report and at 2 weeks and 3 months follow-up via web-based questionnaires. Treatment expectation is the primary outcome. Secondary outcomes include the capacity to consent, decisional conflict, autonomous treatment motivation, adherence intention, and side-effect expectations. RESULTS: This trial received a positive ethics vote by the local ethics committee of the Center for Psychosocial Medicine, University-Medical Center Hamburg-Eppendorf, Hamburg, Germany on April 1, 2021, and was prospectively registered on June 17, 2021. The first participant was enrolled in the study on August 5, 2021. We expect to complete data collection in December 2022. After data analysis within the first quarter of 2023, the results will be submitted for publication in peer-reviewed journals in summer 2023. CONCLUSIONS: If effective, the optimized informed consent consultation might not only constitute an innovative clinical tool to meet the ethical and legal obligations of informed consent but also strengthen the contributing factors of psychotherapy outcome, while minimizing nocebo effects and fostering shared decision-making. TRIAL REGISTRATION: PsychArchives; http://dx.doi.org/10.23668/psycharchives.4929. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/39843.