Algorithm-Aided Prediction of Patient Preferences - An Ethics Sneak Peek

Artificial Intelligence to support ethical decision-making for incapacitated patients: a survey among German anesthesiologists and internists

BACKGROUND

Artificial intelligence (AI) has revolutionized various healthcare domains, where AI algorithms sometimes even outperform human specialists. However, the field of clinical ethics has remained largely untouched by AI advances. This study explores the attitudes of anesthesiologists and internists towards the use of AI-driven preference prediction tools to support ethical decision-making for incapacitated patients.

METHODS

A questionnaire was developed and pretested among medical students. The questionnaire was distributed to 200 German anesthesiologists and 200 German internists, thereby focusing on physicians who often encounter patients lacking decision-making capacity. The questionnaire covered attitudes toward AI-driven preference prediction, availability and utilization of Clinical Ethics Support Services (CESS), and experiences with ethically challenging situations. Descriptive statistics and bivariate analysis was performed. Qualitative responses were analyzed using content analysis in a mixed inductive-deductive approach.

RESULTS

Participants were predominantly male (69.3%), with ages ranging from 27 to 77. Most worked in nonacademic hospitals (82%). Physicians generally showed hesitance toward AI-driven preference prediction, citing concerns about the loss of individuality and humanity, lack of explicability in AI results, and doubts about AI's ability to encompass the ethical deliberation process. In contrast, physicians had a more positive opinion of CESS. Availability of CESS varied, with 81.8% of participants reporting access. Among those without access, 91.8% expressed a desire for CESS. Physicians' reluctance toward AI-driven preference prediction aligns with concerns about transparency, individuality, and human-machine interaction. While AI could enhance the accuracy of predictions and reduce surrogate burden, concerns about potential biases, de-humanisation, and lack of explicability persist.

CONCLUSIONS

German physicians frequently encountering incapacitated patients exhibit hesitance toward AI-driven preference prediction but hold a higher esteem for CESS. Addressing concerns about individuality, explicability, and human-machine roles may facilitate the acceptance of AI in clinical ethics. Further research into patient and surrogate perspectives is needed to ensure AI aligns with patient preferences and values in complex medical decisions.

A Personalized Patient Preference Predictor for Substituted Judgments in Healthcare: Technically Feasible and Ethically Desirable

When making substituted judgments for incapacitated patients, surrogates often struggle to guess what the patient would want if they had capacity. Surrogates may also agonize over having the (sole) responsibility of making such a determination. To address such concerns, a Patient Preference Predictor (PPP) has been proposed that would use an algorithm to infer the treatment preferences of individual patients from population-level data about the known preferences of people with similar demographic characteristics. However, critics have suggested that even if such a PPP were more accurate, on average, than human surrogates in identifying patient preferences, the proposed algorithm would nevertheless fail to respect the patient's (former) autonomy since it draws on the 'wrong' kind of data: namely, data that are not specific to the individual patient and which therefore may not reflect their actual values, or their reasons for having the preferences they do. Taking such criticisms on board, we here propose a new approach: the Personalized Patient Preference Predictor (P4). The P4 is based on recent advances in machine learning, which allow technologies including large language models to be more cheaply and efficiently 'fine-tuned' on person-specific data. The P4, unlike the PPP, would be able to infer an individual patient's preferences from material (e.g., prior treatment decisions) that is in fact specific to them. Thus, we argue, in addition to being potentially more accurate at the individual level than the previously proposed PPP, the predictions of a P4 would also more directly reflect each patient's own reasons and values. In this article, we review recent discoveries in artificial intelligence research that suggest a P4 is technically feasible, and argue that, if it is developed and appropriately deployed, it should assuage some of the main autonomy-based concerns of critics of the original PPP. We then consider various objections to our proposal and offer some tentative replies.

A case for preference-sensitive decision timelines to aid shared decision-making in intensive care: need and possible application

In the intensive care unit, it can be challenging to determine which interventions align with the patients' preferences since patients are often incapacitated and other sources, such as advance directives and surrogate input, are integral. Managing treatment decisions in this context requires a process of shared decision-making and a keen awareness of the preference-sensitive instances over the course of treatment. The present paper examines the need for the development of preference-sensitive decision timelines, and, taking aneurysmal subarachnoid hemorrhage as a use case, proposes a model of one such timeline to illustrate their potential form and value. First, the paper draws on an overview of relevant literature to demonstrate the need for better guidance to (a) aid clinicians in determining when to elicit patient preference, (b) support the drafting of advance directives, and (c) prepare surrogates for their role representing the will of an incapacitated patient in clinical decision-making. This first section emphasizes that highlighting when patient (or surrogate) input is necessary can contribute valuably to shared decision-making, especially in the context of intensive care, and can support advance care planning. As an illustration, the paper offers a model preference-sensitive decision timeline-whose generation was informed by existing guidelines and a series of interviews with patients, surrogates, and neuro-intensive care clinicians-for a use case of aneurysmal subarachnoid hemorrhage. In the last section, the paper offers reflections on how such timelines could be integrated into digital tools to aid shared decision-making.

Integrating equity, diversity and inclusion throughout the lifecycle of AI within healthcare: a scoping review protocol

INTRODUCTION

Artificial intelligence (AI) has the potential to improve efficiency and quality of care in healthcare settings. The lack of consideration for equity, diversity and inclusion (EDI) in the lifecycle of AI within healthcare settings may intensify social and health inequities, potentially causing harm to under-represented populations. This article describes the protocol for a scoping review of the literature relating to integration of EDI in the AI interventions within healthcare setting. The objective of the review is to evaluate what has been done on integrating EDI concepts, principles and practices in the lifecycles of AI interventions within healthcare settings. It also aims to explore which EDI concepts, principles and practices have been integrated into the design, development and implementation of AI in healthcare settings.

METHOD AND ANALYSIS

The scoping review will be guided by the six-step methodological framework developed by Arksey and O'Malley supplemented by Levac et al, and Joanna Briggs Institute methodological framework for scoping reviews. Relevant literature will be identified by searching seven electronic databases in engineering/computer science and healthcare, and searching the reference lists and citations of studies that meet the inclusion criteria. Studies on AI in any healthcare and geographical settings, that have considered aspects of EDI, published in English and French between 2005 and present will be considered. Two reviewers will independently screen titles, abstracts and full-text articles according to inclusion criteria. We will conduct a thematic analysis and use a narrative description to describe the work. Any disagreements will be resolved through discussion with the third reviewer. Extracted data will be summarised and analysed to address aims of the scoping review. Reporting will follow the Preferred Reporting Items for Systematic Reviews and Meta-analyses Extension for Scoping Reviews. The study began in April 2022 and is expected to end in September 2023. The database initial searches resulted in 5,745 records when piloted in April 2022.

ETHICS AND DISSEMINATION

Ethical approval is not required. The study will map the available literature on EDI concepts, principles and practices in AI interventions within healthcare settings, highlight the significance of this context, and offer insights into the best practices for incorporating EDI into AI-based solutions in healthcare settings. The results will be disseminated through open-access peer-reviewed publications, conference presentations, social media and 2-day workshops with relevant stakeholders.

Should Artificial Intelligence be used to support clinical ethical decision-making? A systematic review of reasons

BACKGROUND

Healthcare providers have to make ethically complex clinical decisions which may be a source of stress. Researchers have recently introduced Artificial Intelligence (AI)-based applications to assist in clinical ethical decision-making. However, the use of such tools is controversial. This review aims to provide a comprehensive overview of the reasons given in the academic literature for and against their use.

METHODS

PubMed, Web of Science, Philpapers.org and Google Scholar were searched for all relevant publications. The resulting set of publications was title and abstract screened according to defined inclusion and exclusion criteria, resulting in 44 papers whose full texts were analysed using the Kuckartz method of qualitative text analysis.

RESULTS

Artificial Intelligence might increase patient autonomy by improving the accuracy of predictions and allowing patients to receive their preferred treatment. It is thought to increase beneficence by providing reliable information, thereby, supporting surrogate decision-making. Some authors fear that reducing ethical decision-making to statistical correlations may limit autonomy. Others argue that AI may not be able to replicate the process of ethical deliberation because it lacks human characteristics. Concerns have been raised about issues of justice, as AI may replicate existing biases in the decision-making process.

CONCLUSIONS

The prospective benefits of using AI in clinical ethical decision-making are manifold, but its development and use should be undertaken carefully to avoid ethical pitfalls. Several issues that are central to the discussion of Clinical Decision Support Systems, such as justice, explicability or human-machine interaction, have been neglected in the debate on AI for clinical ethics so far.

TRIAL REGISTRATION

This review is registered at Open Science Framework ( https://osf.io/wvcs9 ).

Ethics of the algorithmic prediction of goal of care preferences: from theory to practice

Artificial intelligence (AI) systems are quickly gaining ground in healthcare and clinical decision-making. However, it is still unclear in what way AI can or should support decision-making that is based on incapacitated patients' values and goals of care, which often requires input from clinicians and loved ones. Although the use of algorithms to predict patients' most likely preferred treatment has been discussed in the medical ethics literature, no example has been realised in clinical practice. This is due, arguably, to the lack of a structured approach to the epistemological, ethical and pragmatic challenges arising from the design and use of such algorithms. The present paper offers a new perspective on the problem by suggesting that preference predicting AIs be viewed as sociotechnical systems with distinctive life-cycles. We explore how both known and novel challenges map onto the different stages of development, highlighting interdisciplinary strategies for their resolution.

Aneurysmal Subarachnoid Hemorrhage and Clinical Decision-Making: A Qualitative Pilot Study Exploring Perspectives of Those Directly Affected, Their Next of Kin, and Treating Clinicians

BACKGROUND

Exploring the experience and impact of aneurysmal subarachnoid hemorrhage (aSAH) from three perspectives, that of those directly affected (AFs), their next of kin (NoK), and treating clinicians, is a way to support and empower others to make informed medical decisions.

METHODS

In a Swiss neurosurgical intensive care unit (ICU), eleven semi-structured interviews were conducted as part of a Database of Individual Patient Experiences (DIPEx) pilot project and thematically analyzed. Interviews were held with two clinicians, five people experiencing aSAH, and four NoK 14-21 months after the bleeding event.

RESULTS

Qualitative analysis revealed five main themes from the perspective of clinicians: emergency care, diagnosis and treatment, outcomes, everyday life in the ICU, and decision-making; seven main themes were identified for AFs and NoK: the experience of the aSAH, diagnosis and treatment, outcomes, impact on loved ones, identity, faith, religion and spirituality, and decision-making. Perspectives on decision-making were compared, and, whereas clinicians tended to focus their attention on determining treatment, AFs and NoK valued participation in shared decision-making processes.

CONCLUSIONS

Overall, aSAH was perceived as a life-threatening event with various challenges depending on severity. The results suggest the need for tools that aid decision-making and better prepare AFs and NoK using accessible means and at an early stage.

From text to interaction: The digital advance directive method for advance directives

Advance directives allow people to specify individual treatment preferences in case of decision-making incapacity involving decisions of utmost importance. There are many tools that provide information on the topic, digital forms for structured data input, or platforms that support data storage and availability. Yet, there is no tool supporting the innermost process of an advance directive: decision making itself. To address this issue, we developed a visual-interactive, semi-quantitative method for generating digital advance directives (DiADs) that harnesses the potential of digitalization in healthcare. In this article, we describe the DiAD method and its app lined with the exemplary narrative of user Mr S. linking the theory to an exemplary use case. The DiAD method is intended to lower barriers and increase comfort in creating an advance directive by shifting the focus from heavily text-based processes to visual representation and interaction, that is, from text to reflection.

Contextual Exceptionalism After Death: An Information Ethics Approach to Post-Mortem Privacy in Health Data Research

In this article, we use the theory of Information Ethics to argue that deceased people have a prima facie moral right to privacy in the context of health data research, and that this should be reflected in regulation and guidelines. After death, people are no longer biological subjects but continue to exist as informational entities which can still be harmed/damaged. We find that while the instrumental value of recognising post-mortem privacy lies in the preservation of the social contract for health research, its intrinsic value is grounded in respect for the dignity of the post-mortem informational entity. However, existing guidance on post-mortem data protection is available only in the context of genetic studies. In comparing the characteristics of genetic data and other health-related data, we identify two features of DNA often given as arguments for this genetic exceptionalism: relationality and embodiment. We use these concepts to show that at the appropriate Level of Abstraction, there is no morally relevant distinction between posthumous genetic and other health data. Thus, genetic data should not automatically receive special moral status after death. Instead we make a plea for 'contextual exceptionalism'. Our analysis concludes by reflecting on a real-world case and providing suggestions for contextual factors that researchers and oversight bodies should take into account when designing and evaluating research projects with health data from deceased subjects.

Risks and Opportunities to Ensure Equity in the Application of Big Data Research in Public Health

The big data revolution presents an exciting frontier to expand public health research, broadening the scope of research and increasing the precision of answers. Despite these advances, scientists must be vigilant against also advancing potential harms toward marginalized communities. In this review, we provide examples in which big data applications have (unintentionally) perpetuated discriminatory practices, while also highlighting opportunities for big data applications to advance equity in public health. Here, big data is framed in the context of the five Vs (volume, velocity, veracity, variety, and value), and we propose a sixth V, virtuosity, which incorporates equity and justice frameworks. Analytic approaches to improving equity are presented using social computational big data, fairness in machine learning algorithms, medical claims data, and data augmentation as illustrations. Throughout, we emphasize the biasing influence of data absenteeism and positionality and conclude with recommendations for incorporating an equity lens into big data research.

Surrogate Perspectives on Patient Preference Predictors: Good Idea, but I Should Decide How They Are Used

Current practice frequently fails to provide care consistent with the preferences of decisionally-incapacitated patients. It also imposes significant emotional burden on their surrogates. Algorithmic-based patient preference predictors (PPPs) have been proposed as a possible way to address these two concerns. While previous research found that patients strongly support the use of PPPs, the views of surrogates are unknown. The present study thus assessed the views of experienced surrogates regarding the possible use of PPPs as a means to help make treatment decisions for decisionally-incapacitated patients.

This qualitative study used semi-structured interviews to determine the views of experienced surrogates [n = 26] who were identified from two academic medical centers and two community hospitals. The primary outcomes were respondents' overall level of support for the idea of using PPPs and the themes related to their views on how a PPP should be used, if at all, in practice.

Overall, 21 participants supported the idea of using PPPs. The remaining five indicated that they would not use a PPP because they made decisions based on the patient's best interests, not based on substituted judgment. Major themes which emerged were that surrogates, not the patient's preferences, should determine how treatment decisions are made, and concern that PPPs might be used to deny expensive care or be biased against minority groups.

Surrogates, like patients, strongly support the idea of using PPPs to help make treatment decisions for decisionally-incapacitated patients. These findings provide support for developing a PPP and assessing it in practice. At the same time, patients and surrogates disagree over whose preferences should determine how treatment decisions are made, including whether to use a PPP. These findings reveal a fundamental disagreement regarding the guiding principles for surrogate decision-making. Future research is needed to assess this disagreement and consider ways to address it.

Supplemental data for this article is available online at.

AI support for ethical decision-making around resuscitation: proceed with care

Artificial intelligence (AI) systems are increasingly being used in healthcare, thanks to the high level of performance that these systems have proven to deliver. So far, clinical applications have focused on diagnosis and on prediction of outcomes. It is less clear in what way AI can or should support complex clinical decisions that crucially depend on patient preferences. In this paper, we focus on the ethical questions arising from the design, development and deployment of AI systems to support decision-making around cardiopulmonary resuscitation and the determination of a patient's Do Not Attempt to Resuscitate status (also known as code status). The COVID-19 pandemic has made us keenly aware of the difficulties physicians encounter when they have to act quickly in stressful situations without knowing what their patient would have wanted. We discuss the results of an interview study conducted with healthcare professionals in a university hospital aimed at understanding the status quo of resuscitation decision processes while exploring a potential role for AI systems in decision-making around code status. Our data suggest that (1) current practices are fraught with challenges such as insufficient knowledge regarding patient preferences, time pressure and personal bias guiding care considerations and (2) there is considerable openness among clinicians to consider the use of AI-based decision support. We suggest a model for how AI can contribute to improve decision-making around resuscitation and propose a set of ethically relevant preconditions-conceptual, methodological and procedural-that need to be considered in further development and implementation efforts.

Advance care planning evaluation: a scoping review of best research practice

Various indicators have been used to evaluate advance care planning, including completion rates, type of care received, and satisfaction. Recent consensus suggests, though, that receiving care consistent with one's goals is the primary outcome of advance care planning and assessment should capture this metric. Goal concordant care is challenging to measure, and there is little clarity about how best to do so. The aim of this scoping review is to explore what methods have been used to measure goal concordant care in the evaluation of advance care planning. PubMed, Embase, PsycINFO, CINAHL and Cochrane were searched in September 2020 to identify studies that aimed to track whether advance care planning affected the likelihood of patients receiving care that matched their preferred care. 135 original studies were included for review. Studies used retrospective chart review (36%, n=49), questionnaire (36%, n=48) and interview (31%, n=42), focusing on both patients and proxies. Studies considered both actual care received (55%, n=74) and hypothetical scenarios anticipating possible future care (49%, n=66); some studies did both. While the reviewed studies demonstrate the possibility of working towards a solid methodology, there were significant weaknesses. Notably, studies often lacked enough reporting clarity to be reproducible and, relatedly, key concepts, such as end-of-life or preferred care, were left undefined. The recommendations that follow from these findings inform future research approaches, supporting the development of a strong evidence base to guide advance care planning implementation in practice.

Ethics of ICU triage during COVID-19

INTRODUCTION

The coronavirus disease 2019 pandemic has placed intensive care units (ICU) triage at the center of bioethical discussions. National and international triage guidelines emerged from professional and governmental bodies and have led to controversial discussions about which criteria-e.g. medical prognosis, age, life-expectancy or quality of life-are ethically acceptable. The paper presents the main points of agreement and disagreement in triage protocols and reviews the ethical debate surrounding them.

SOURCES OF DATA

Published articles, news articles, book chapters, ICU triage guidelines set out by professional societies and health authorities.

AREAS OF AGREEMENT

Points of agreement in the guidelines that are widely supported by ethical arguments are (i) to avoid using a first come, first served policy or quality-adjusted life-years and (ii) to rely on medical prognosis, maximizing lives saved, justice as fairness and non-discrimination.

AREAS OF CONTROVERSY

Points of disagreement in existing guidelines and the ethics literature more broadly regard the use of exclusion criteria, the role of life expectancy, the prioritization of healthcare workers and the reassessment of triage decisions.

GROWING POINTS

Improve outcome predictions, possibly aided by Artificial intelligence (AI); develop participatory approaches to drafting, assessing and revising triaging protocols; learn from experiences with implementation of guidelines with a view to continuously improve decision-making.

AREAS TIMELY FOR DEVELOPING RESEARCH

Examine the universality vs. context-dependence of triaging principles and criteria; empirically test the appropriateness of triaging guidelines, including impact on vulnerable groups and risk of discrimination; study the potential and challenges of AI for outcome and preference prediction and decision-support.

Teasing out Artificial Intelligence in Medicine: An Ethical Critique of Artificial Intelligence and Machine Learning in Medicine

The rapid adoption and implementation of artificial intelligence in medicine creates an ontologically distinct situation from prior care models. There are both potential advantages and disadvantages with such technology in advancing the interests of patients, with resultant ontological and epistemic concerns for physicians and patients relating to the instatiation of AI as a dependent, semi- or fully-autonomous agent in the encounter. The concept of libertarian paternalism potentially exercised by AI (and those who control it) has created challenges to conventional assessments of patient and physician autonomy. The unclear legal relationship between AI and its users cannot be settled presently, an progress in AI and its implementation in patient care will necessitate an iterative discourse to preserve humanitarian concerns in future models of care. This paper proposes that physicians should neither uncritically accept nor unreasonably resist developments in AI but must actively engage and contribute to the discourse, since AI will affect their roles and the nature of their work. One's moral imaginative capacity must be engaged in the questions of beneficence, autonomy, and justice of AI and whether its integration in healthcare has the potential to augment or interfere with the ends of medical practice.

Virtual care to increase military medical centre capacity in the primary health care setting: A prospective self-controlled pilot study of symptoms collection and telemedicine

INTRODUCTION

The Singapore Armed Forces (SAF) recognises the potential benefits and looks to harnessing telemedicine for primary health care services. In this prospective self-controlled pilot study, we aimed to evaluate the safety, efficiency and user satisfaction outcomes of virtual care (VC) at a military medical centre.

METHODS

Out of 320 patients seen during the study period, 28 were enrolled in this study and underwent on-premises VC, comprising digital symptoms collection and telemedicine in addition to the usual in-person physician consultation. Safety outcomes were measured based on the diagnostic concordance between physicians. Efficiency was measured based on consultation times, and user satisfaction was evaluated using a standard questionnaire.

RESULTS

There was a higher caseload of both upper respiratory infections and dermatological conditions in our population, in which telemedicine performed well. In terms of safety, telemedicine achieved a mean diagnostic concordance of 92.8% compared to in-person consultations. In terms of efficiency, consultation times were 26.2% - or 2 minutes and 15 seconds - shorter on average with telemedicine (p = 0.0488). User satisfaction was favourable, with 85.5% of patients satisfied with the VC experience.

DISCUSSION

This study has been invaluable in showing that on-premises telemedicine is a safe, efficient and effective means to extend and increase our surge capacity for primary health care. Our results have given us reasonable confidence to explore a larger-scale implementation in our network of military medical centres in the future.

Ethical considerations about artificial intelligence for prognostication in intensive care

Background

Prognosticating the course of diseases to inform decision-making is a key component of intensive care medicine. For several applications in medicine, new methods from the field of artificial intelligence (AI) and machine learning have already outperformed conventional prediction models. Due to their technical characteristics, these methods will present new ethical challenges to the intensivist.

Results

In addition to the standards of data stewardship in medicine, the selection of datasets and algorithms to create AI prognostication models must involve extensive scrutiny to avoid biases and, consequently, injustice against individuals or groups of patients. Assessment of these models for compliance with the ethical principles of beneficence and non-maleficence should also include quantification of predictive uncertainty. Respect for patients’ autonomy during decision-making requires transparency of the data processing by AI models to explain the predictions derived from these models. Moreover, a system of continuous oversight can help to maintain public trust in this technology. Based on these considerations as well as recent guidelines, we propose a pathway to an ethical implementation of AI-based prognostication. It includes a checklist for new AI models that deals with medical and technical topics as well as patient- and system-centered issues.

Conclusion

AI models for prognostication will become valuable tools in intensive care. However, they require technical refinement and a careful implementation according to the standards of medical ethics.

Advance care planning in Parkinson's disease: ethical challenges and future directions

Recent discoveries support the principle that palliative care may improve the quality of life of patients with Parkinson's disease and those who care for them. Advance care planning, a component of palliative care, provides a vehicle through which patients, families, and clinicians can collaborate to identify values, goals, and preferences early, as well as throughout the disease trajectory, to facilitate care concordant with patient wishes. While research on this topic is abundant in other life-limiting disorders, particularly in oncology, there is a paucity of data in Parkinson's disease and related neurological disorders. We review and critically evaluate current practices on advance care planning through the analyses of three bioethical challenges pertinent to Parkinson's disease and propose recommendations for each.