Patient preference predictors (PPPs) promise to provide medical professionals with a new solution to the problem of making treatment decisions on behalf of incapacitated patients. I show that the use of PPPs faces a version of a normative problem familiar from legal scholarship: the problem of naked statistical evidence. I sketch two sorts of possible reply, vindicating and debunking, and suggest that our reply to the problem in the one domain ought to mirror our reply in the (...) other. The conclusion is thus conditional: if we think the problem of naked statistical evidence is a serious problem in the legal domain, then we should be concerned about the symmetrical problem for PPPs. (shrink)

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. (shrink)

In this paper, I set out two ethical complications for Rid and Wendler’s proposal that a “Patient Preference Predictor” (PPP) should be used to aid decision making about incapacitated patients’ care. Both of these worries concern how a PPP might categorize patients. In the first section of the paper, I set out some general considerations about the “ethics of apt categorization” within stratified medicine and show how these challenge certain PPPs. In the second section, I argue for a (...) more specific—but more contentious—claim that proper respect for the autonomy of incapacitated patients might require us to act on reasons which they could endorse and show how this claim places important limits on the categories employed by an ethically acceptable PPP. The conclusion shows how these concerns about apt categorization relate to more familiar worries about Rid and Wendler’s proposals. (shrink)

Except in extraordinary circumstances, patients' clinical care should reflect their preferences. Incapacitated patients cannot report their preferences. This is a problem. Extant solutions to the problem are inadequate: surrogates are unreliable, and advance directives are uncommon. In response, some authors have suggested developing algorithmic "patient preference predictors" (PPPs) to inform care for incapacitated patients. In a recent paper, Earp et al. propose a new twist on PPPs. Earp et al. suggest we personalize PPPs using modern machine learning (...) (ML) techniques. In this paper, I argue that, contrary to Earp et al.'s claim, personalized patient preference predictors are neither technically feasible nor ethically desirable. (shrink)

Patient preference predictors take us from known demographic descriptors to unknown facts about patients’ preferences over treatment options. However, the use of PPPs to make treatment decisions on behalf of incapacitated patients faces an apparent normative problem: their use in certain contexts appears to involve treating patients paternalistically. In this paper, I consider whether PPPs can find a home in the context of military medicine. On the assumptions that military organizations sometimes permissibly treat their members paternalistically, I (...) identify the specific circumstances in which PPPs can play a role in making treatment decisions on behalf of patients. I show that these circumstances are not only the natural home of PPPs, PPPs – or something very much like them – are precisely the sort of device required for it to be permissible to paternalistically override a patient’s expressed preferences. (shrink)

The Patient Preference Predictor (PPP) is intended to improve treatment decision making for incapacitated patients. The PPP would collect information about the treatment preferences of people with different demographic and other characteristics. It could be used to indicate which treatment option an individual patient would be most likely to prefer, based on data about the preferences of people who resemble the patient. The PPP could be incorporated into existing US law governing treatment for incapacitated patients, although (...) it is unclear whether it would be classified as evidence of a specific patient’s preferences or those of a reasonable person sharing certain characteristics with the patient. Ethical concerns about the quality and significance of PPP choices could influence legal decision makers’ views of the PPP. (shrink)

The standard approach to treatment decision making for incapacitated patients often fails to provide treatment consistent with the patient’s preferences and values and places significant stress on surrogate decision makers. These shortcomings provide compelling reason to search for methods to improve current practice. Shared decision making between surrogates and clinicians has important advantages, but it does not provide a way to determine patients’ treatment preferences. Hence, shared decision making leaves families with the stressful challenge of identifying the patient’s (...) preferred treatment option. To address this concern, the present paper proposes to incorporate the use of a “Patient Preference Predictor” (PPP) into the shared decision-making process between surrogates and clinicians. A PPP would predict which treatment option a given incapacitated patient would most likely prefer, based on the individual’s characteristics and information on what treatment preferences are correlated with these characteristics. Use of a PPP is likely to increase the chances that incapacitated patients are treated consistent with their preferences and values and might reduce the stress and burden on their surrogates. Including a PPP in the shared decision-making process therefore has the potential to realize important ethical goals for making treatment decisions for incapacitated patients. The present paper justifies this approach on conceptual and normative grounds. (shrink)

An AI-based “patient preference predictor” (PPP) is a proposed method for guiding healthcare decisions for patients who lack decision-making capacity. The proposal is to use correlations between sociodemographic data and known healthcare preferences to construct a model that predicts the unknown preferences of a particular patient. In this paper, I highlight a distinction that has been largely overlooked so far in debates about the PPP–that between algorithmic prediction and decision-making–and argue that much of the recent philosophical disagreement (...) stems from this oversight. I show how three prominent objections to the PPP only challenge its use as the sole determinant of a choice, and actually support its use as a source of evidence about patient preferences to inform human decision-making. The upshot is that we should adopt the evidential conception of the PPP and shift our evaluation of this technology towards the ethics of algorithmic prediction, rather than decision-making. (shrink)

Background: 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. -/- Methods: 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. -/- Results: 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. -/- Conclusions: 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. (shrink)

Like many, I find the idea of relying on patient preference predictors in life-or-death cases ethically troubling. As part of his stimulating discussion, Sharadin1 diagnoses such unease as a worry that using PPPs disrespects patients’ autonomy, by treating their most intimate and significant desires as if they were caused by their demographic traits. I agree entirely with Sharadin’s ‘debunking’ response to this concern: we can use statistical correlations to predict others’ preferences without thereby assuming any causal claim. (...) However, I suspect that, for at least some of us, our unease about PPPs stems from a different kind of ‘autonomy’ concern. In this commentary, then, I will explore this concern, and show how it relates to Sharadin’s work. Very many of our preferences are caused, ultimately, by facts which are outside our control, such as our demographic features. However, I suggest that we can still act autonomously on the basis of such preferences, when they are preferences which we endorse. Imagine, for example, that Jane has grown up in a church-growing environment, which has shaped many of her preferences, …. (shrink)

The question of how to treat an incapacitated patient is vexed, both normatively and practically—normatively, because it is not obvious what the relevant objectives are; practically, because even once the relevant objectives are set, it is often difficult to determine which treatment option is best given those objectives. But despite these complications, here is one consideration that is clearly relevant: what a patient prefers. And so any device that could reliably identify a patient’s preferences would be a (...) promising tool for guiding the treatment of incapacitated patients. The patient preference predictor is just such a tool—an algorithm that takes as inputs a patient’s sociodemographic characteristics, and outputs a reliable prediction about that patient’s treatment preferences.1 But some have worried that the use of such a tool would violate or fail to appropriately respect patients’ autonomy. There are, I think, two ways to understand this kind of criticism. First, globally—as a worry that any systematic implementation of the PPP would be problematic on the grounds that it would result in significant or pervasive autonomy violations. Second, locally—as a worry that in some important range of cases, certain uses of the PPP would be problematic on autonomy grounds. Jardas et al, as I read them, address global autonomy-based criticisms, arguing—convincingly, in my view—that there’s no reason to suspect the autonomy concerns raised by the PPP would be so significant and pervasive as to render any implementation of it generally problematic.1 But even with the global criticisms rebuffed, there remains work to be done. Any ethically acceptable implementation of the PPP must be sensitive to the more local autonomy-based criticisms, with restrictions and safeguards in place to ensure respect for autonomy in the kinds of cases in which the use of the PPP might otherwise threaten …. (shrink)

There are substantial data establishing that surrogates are often mistaken in predicting what treatments incompetent patients would have wanted and that supplements such as advance directives have not resulted in significant improvements. Rid and Wendler’s Patient Preference Predictor (PPP) proposal will attempt to gather data about what similar patients would prefer in a variety of treatment choices. It accepts the usual goal of patient autonomy and the Substituted Judgment principle for surrogate decisions. I provide reasons for questioning (...) sole reliance on autonomy when surrogates, not patients, are deciding and for giving some independent weight to patient well-being. Moreover, that family members, typical surrogates, will know best what the patient if incompetent would have wanted is not the only reason why they are chosen. The more pressing problem is that the PPP would fail to remove the more serious mistakes that empirical psychology over the last few decades has shown to infect such decision making. (shrink)

The future of medicine will be predictive, preventive, personalized, and participatory. Recent technological advancements bolster the realization of this vision, particularly through innovations in...

The patient preference predictor is a proposed computer-based algorithm that would predict the treatment preferences of decisionally incapacitated patients. Incorporation of a PPP into the decision-making process has the potential to improve implementation of the substituted judgement standard by providing more accurate predictions of patients’ treatment preferences than reliance on surrogates alone. Yet, critics argue that methods for making treatment decisions for incapacitated patients should be judged on a number of factors beyond simply providing them with the treatments (...) they would have chosen for themselves. These factors include the extent to which the decision-making process recognises patients’ freedom to choose and relies on evidence the patient themselves would take into account when making treatment decisions. These critics conclude that use of a PPP should be rejected on the grounds that it is inconsistent with these factors, especially as they relate to proper respect for patient autonomy. In this paper, we review and evaluate these criticisms. We argue that they do not provide reason to reject use of a PPP, thus supporting efforts to develop a full-scale PPP and to evaluate it in practice. There are no data in this work. (shrink)

A Patient Preference Predictor (PPP)—an algorithm capable of predicting, on the basis of demographic or more personalized data, what an incapacitated patient would prefer were they capacitated—is a...

It has recently been proposed to incorporate the use of a “Patient Preference Predictor” (PPP) into the process of making treatment decisions for incapacitated patients. A PPP would predict which treatment option a given incapacitated patient would most likely prefer, based on the individual’s characteristics and information on what treatment preferences are correlated with these characteristics. Including a PPP in the shared decision-making process between clinicians and surrogates has the potential to better realize important ethical goals for (...) making treatment decisions for incapacitated patients. However, developing and implementing a PPP poses significant practical challenges. The present paper discusses these practical challenges and considers ways to address them. (shrink)

When a patient lacks sufficient capacity to make a certain treatment decision, whether because of deficits in their ability to make a judgement that reflects their values or to make a decision that reflects their judgement or both, the decision must be made by a surrogate. Often the best way to respect the patient’s autonomy, in such cases, is for the surrogate to make a ‘substituted’ judgement on behalf of the patient, which is the decision that best (...) reflects the patient’s values and identity-infused beliefs as they were when the patient last possessed capacity. It is by now well-known that there is troubling evidence that surrogates do only slightly better than chance at making accurate substituted judgements. The use of a patient preference predictor, an algorithm that predicts patient preferences based on characteristics like the patient’s age, gender, socioeconomic status and education level, could help. Rid and Wendler have done important work making a moral case for using the PPP.1 If it were shown to enhance accuracy, it is not at all obvious what objections there could be to its use. Yet several subtle criticisms of the PPP have emerged in the literature. In their feature article, Jardas, Wasserman and Wendler select six of the most prominent objections to the PPP in the literature and defend the PPP against them.2 Along the way they make significant contributions to broader issues concerning the proper uses of statistical evidence and limits on the kinds of preferences there can be reasons of autonomy to respect. Since I found the authors’ answers to the final two objections—that the PPP relies on naked statistical evidence and that it uses non-endorsed reasons—to be especially interesting and provocative, I will examine those answers and raise a few questions about …. (shrink)

In the age where AI is showing increasing potential to solve problems in unprecedented ways, it becomes tempting to see it as the solution for every problem, resulting in a focus on the means (i.e....

The authors respond to four JME commentaries on their Feature Article, ‘Autonomy-based criticisms of the patient preference predictor’.

Annoni’s critique of Personalized Patient Preference Predictors (P4) highlights a fundamental flaw in their current design: they fail to meaningfully respect patient autonomy.1 His argument that P4 risks reducing decision-making to the presumed preferences of incapacitated individuals underscores the need for a better approach. To address this, we introduce the concept of the Personalized Patient Preference Predictor-Assisted Living Will (P4-LW)—a mechanism that allows individuals, while still capacitated, to formally consent to the use of P4 (...) and subsequently validate that their living will aligns with their personal values and evolving preferences. Unlike static living wills or surrogate decision-makers, the P4-LW has the potential to provide clinicians with real-time, patient-specific guidance from individuals who have proactively endorsed its use. In this commentary, we extend the discussion by examining the ethical and practical implications of this model, arguing that it offers a more nuanced and ethically sound alternative to conventional surrogate decision-making, and the proposed P4-aided process. The transfer of autonomy from an incapacitated patient to a P4 raises ethical concerns, but its acceptability depends on how these tools are implemented. A complete and unchecked transfer is likely inappropriate and would constitute what we called in our previous work ‘AI paternalism’.2 Artificial intelligence (AI) paternalism, …. (shrink)

To overcome certain challenges with surrogate decision-making traditionally understood, technological support tools have been proposed. One such proposal, as presented by Earp et al, is the development of a ‘Personalised Patient Preference Predictor’ (P4).1 This system would leverage patient-specific data to train a personalised large language model which could then—hopefully—accurately predict that patient’s treatment preferences. Using P4 would be compatible with respect for patient autonomy, at least on the substituted judgement standard, in which making the (...) decision the incompetent person would have made if competent is a priority.2 Annoni, however, doubts that P4 can respect the autonomy of incapacitated patients. This is because ‘the satisfaction of individual treatment preferences is questionable whenever such preferences are not actual, explicit and specific.’2 His view is that merely satisfying a patient’s hypothetical preference is not sufficient to respect their autonomy. Instead, fulfilling respect for autonomy requires ‘explicit, first-person, actual and specific consent.’2 This is clearly not possible if the patient in question no longer has capacity. Thus, arguing for the use of P4 in such cases on the basis of respect for patient autonomy is moot; rather, Annoni claims, P4 can be justified on other grounds, such as the ‘honouring of a patient’s unique identity.’2 In my own view, there are two issues with Annoni’s argument. The first is his rationale that respect for autonomy requires the actual consent of the patient, against the possibility of generating hypothetical consent using tools like P4. While he is correct to point out the various …. (shrink)

Patient preference predictors aim to solve the moral problem of making treatment decisions on behalf of incapacitated patients. This commentary on a case of an unrepresented patient at the end of life considers 3 related problems of such predictors: the problem of restricting the scope of inputs to the models (the “scope” problem), the problem of weighing inputs against one another (the “weight” problem), and the problem of multiple reasonable solutions to the scope and weight (...) problems (the “multiple reasonable models” problem). Each of these problems poses challenges to reliably implementing patient preference predictors in important, high-stakes health care decision making. This commentary also suggests a way forward. (shrink)

> Stephan, a middle-aged immigrant, presents with acute sepsis from his diabetic foot ulcer and confusion that significantly diminishes his capacity. His attending physician, Marta, has initiated an aggressive antibiotic regime but is considering amputation. Unsure, whether this aligns with the preferences of Stephan, his family, and his devoted support worker, Emily, she is considering using a P4 AI for substituted judgment using a corpus of Stephan’s emails and online posts. > > When diagnosed with diabetes shortly after immigrating seventeen (...) years ago, Stephan railed against the disease’s ravages online, declaiming that he would rather be dead than blind or immobile. Five years ago, anxiety over disturbing trends in online discourse led to his abruptly ceasing all online communication, turning increasingly to Emily as a confidant and guide to advances in vision treatment and prosthetics. And two years ago, Stephan gained significantly new purpose and joy with the birth of his first grandchild Anya. This case, following Annoni,1 helps illustrate the potential use of P4 artificial intelligence (AI) to advance moral goods in a substituted judgement setting by (1) honouring the patient’s unique identity and past agency and (2) reducing the burdens on human surrogate decision-makers. However, against (2), we argue that decision-making discourse …. (shrink)

Background Although medical ethicists and educators emphasise patient-centred decision-making, previous studies suggest that patients often prefer their doctors to make the clinical decisions. Objective To examine the associations between a preference for physician-directed decision-making and patient health status and sociodemographic characteristics. Methods Sociodemographic and clinical information from all consenting general internal medicine patients at the University of Chicago Medical Center were examined. The primary objectives were to (1) assess the extent to which patients prefer an active role (...) in clinical decision-making, and (2) determine whether religious service attendance, the importance of religion, self-rated spirituality, Charlson Comorbidity Index, self-reported health, Vulnerable Elder Score and several demographic characteristics were associated with these preferences. Results Data were collected from 8308 of 11 620 possible participants. Ninety-seven per cent of respondents wanted doctors to offer them choices and to consider their opinions. However, two out of three (67%) preferred to leave medical decisions to the doctor. In multiple regression analyses, preferring to leave decisions to the doctor was associated with older age (per year, OR=1.019, 95% CI 1.003 to 1.036) and frequently attending religious services (OR=1.5, 95% CI 1.1 to 2.1, compared with never), and it was inversely associated with female sex (OR=0.6, 95% CI 0.5 to 0.8), university education (OR=0.6, 95% CI 0.4 to 0.9, compared with no high school diploma) and poor health (OR=0.6, 95% CI 0.3 to 0.9). Conclusions Almost all patients want doctors to offer them choices and to consider their opinions, but most prefer to leave medical decisions to the doctor. Patients who are male, less educated, more religious and healthier are more likely to want to leave decisions to their doctors, but effects are small. (shrink)

The concept of a Patient Preference Predictor—an algorithm that supplements or replaces the process of surrogate decision-making for incapacitated patients—was first suggested a decade ago (Rid and...

Artificial intelligence preference predictors, such as the personalised patient preference predictor (P4), are designed to forecast patients’ decision-making preferences and thereby function as a form of substituted judgement. Annoni advocates for both advance directives and substituted judgement, arguing that medical decisions should reflect the values and preferences held by patients when they are capable of deciding for themselves.1 However, he contends that the ethical justification for substituted judgement should not rest on respect for patient autonomy, (...) since basing decisions on a person’s preferences does not necessarily equate to genuinely respecting their autonomy. Annoni illustrates this point with the example of deceiving a patient into using a placebo. Although the placebo produces effects that align with the patient’s desires, this approach fails to respect their autonomy. We further note that Buchanan and Brock distinguish between two dimensions of autonomy. The first is its instrumental value—the idea that individuals are generally best positioned to determine what is in their own interests—while the second is its non-instrumental value, which concerns the intrinsic worth of engaging in the process of decision-making itself.2 For example, if P4 indicates that individual B prefers apples and individual A subsequently provides an apple to B, then P4 has respected B’s autonomy …. (shrink)

In “A Personalized Patient Preference Predictor for Substituted Judgments in Healthcare: Technically Feasible and Ethically Desirable,” Earp et al. (2024) offer what should be considered a potentia...

The use of machine learning, or “artificial intelligence” (AI) in medicine is widespread and growing. In this paper, I focus on a specific proposed clinical application of AI: using models to predict incapacitated patients’ treatment preferences. Drawing on results from machine learning, I argue this proposal faces a special moral problem. Machine learning researchers owe us assurance on this front before experimental research can proceed. In my conclusion I connect this concern to broader issues in AI safety.

The debate about the chances and dangers of a patient preference predictor (PPP) has been lively ever since Annette Rid and David Wendler proposed this fascinating idea ten years ago. Given the tec...

Rid and Wendler propose the development of a Patient Preference Predictor (PPP), an actuarial model for predicting incapacitated patient’s life-sustaining treatment preferences across a wide range of end-of-life scenarios. An actuarial approach to end-of-life decision making has enormous potential, but transferring the logic of actuarial prediction to end-of-life decision making raises several conceptual complexities and logistical problems that need further consideration. Actuarial models have proven effective in targeted prediction tasks, but no evidence supports their effectiveness in the (...) kind of broad spectrum prediction task that is the proposed goal of the PPP. We argue that a more focused approach, targeting specific medical conditions and generating treatment predictions based on the preferences of individuals with actual disease experience, is both more firmly grounded in past research and is a more prudent initial strategy for exploring the efficacy of actuarial prediction in end-of-life decision making. (shrink)

The Patient Preference Predictor (PPP) proposal places a high priority on the accuracy of predicting patients’ preferences and finds the performance of surrogates inadequate. However, the quest to develop a highly accurate, individualized statistical model has significant obstacles. First, it will be impossible to validate the PPP beyond the limit imposed by 60%–80% reliability of people’s preferences for future medical decisions—a figure no better than the known average accuracy of surrogates. Second, evidence supports the view that a sizable (...) minority of persons may not even have preferences to predict. Third, many, perhaps most, people express their autonomy just as much by entrusting their loved ones to exercise their judgment than by desiring to specifically control future decisions. Surrogate decision making faces none of these issues and, in fact, it may be more efficient, accurate, and authoritative than is commonly assumed. (shrink)

The patient preference predictor is a computer-based algorithm devised to predict the medical treatment that decisionally incapacitated patients would have preferred. The target paper argues against various criticisms to the effect that the use of a PPP is inconsistent with proper respect for patient autonomy.1 In this commentary, I aim to add some clarifications to the complex relationship between autonomy and the PPP. First, I highlight one way in which the decision of a surrogate designated by the (...) patient realises respect for patient autonomy in a way that PPP cannot realise by introducing the idea of shared agency. Second, I show how the attempt to predict someone’s hypothetical choice can be disrespectful to her autonomy. Following the treatment decision of a surrogate designated by the patient counts as a way of respecting the patient’s own autonomy. Designating someone as one’s surrogate is a case of entrusting her with the task of making important life decisions on one’s behalf. To entrust someone with a certain matter is to assign her the responsibility for that matter based on the trust that she will take a good care of it. When a patient entrusts a surrogate decision-maker, what is the matter that she is trusted to take a good care of? In many contexts, the surrogate is r egarded as an epistemic proxy, one who is likely to know the patient’s preferences the best. It is often thought that the point of practising surrogate decision making is to respect the patient’s autonomy by following what the patient most likely would have preferred. Using a PPP in the decision-making process has the same goal of making a more accurate prediction of the patient’s treatment preference. However, at least in some cases, designating the patient’s intimates or ones she loves and cares …. (shrink)

In “A Personalized Patient Preference Predictor for Substituted Judgments in Healthcare,” Earp et al. (2024) introduce the Personalized Patient Preference Predictor (P4), an AI model designed to ex...

The authority of surrogates—often close family members—to make treatment decisions for previously capacitated patients is said to come from their knowledge of the patient, which they are to draw on as they exercise substituted judgment on the patient’s behalf. However, proxy accuracy studies call this authority into question, hence the Patient Preference Predictor (PPP). We identify two problems with contemporary understandings of the surrogate’s role. The first is with the assumption that knowledge of the patient (...) entails knowledge of what the patient’s choice of treatment would be. The second is with the assumption that a good decision reproduces the content of that choice. If we are right, then the PPP, helpful though it might be in guiding surrogates’ decisions, nevertheless would hold them to the wrong standards and in that way could add to, rather than relieve, the stress they experience as they try to do their job. (shrink)

In their target article, “A Personalized Patient Preference Predictor for Substituted Judgments in Healthcare: Technically Feasible and Ethically Desirable,” Earp et al. (2024) discuss ways in whic...

In their recent paper, Earp and coauthors (2024) argue for the ethical desirability of personalized patient preference predictors (P4s): large-language models (LLMs) finetuned on a patient’s “own p...

Surrogate decision-making is fraught with speculation. Amid fogs of uncertainty, surrogates must ascertain an incapacitated patient’s wishes. From this, they are entrusted to make life-altering or life-ending decisions based on limited information. This process of guesswork, however, is inevitably shaped by subjective interpretation and personal biases. In response, artificial intelligence (AI) tools like personalised patient preference predictors (P4) have been proposed as a means to safeguard the accuracy and reliability of surrogate decision-making.1 Yet, whether AI-driven (...) class='Hi'>preference predictors can uphold autonomy is a point of moral contention. Proponents like Earp et al. argue that AI tools offer a more systematic, data-driven approach to inferring patient wishes, thereby upholding autonomy by mitigating the human errors of traditional surrogates. Others challenge this assumption. Annoni, for instance, argues that AI-driven preference predictors categorically fail to respect patient autonomy, asserting that autonomy cannot be merely reduced to the satisfaction of patient preferences.2 What follows is an argument for reconsideration. This paper does not contend that AI tools definitively respect autonomy, but instead draws forth novel considerations that challenge the notion of their fundamental incompatibility. To this end, I first examine the case study of AI-powered psychotherapy, in which patients demonstrate a greater willingness to disclose personal information to AI systems. Then, I explore how these insights underscore the promise of …. (shrink)

End-of-life decision making frequently occurs in the intensive care unit (ICU). There is a lack of information on how a do-not-resuscitate (DNR) order affects treatments received by critically ill patients in ICUs. The objectives of this study were: (1) to compare the use of life support therapies between patients with a DNR order and those without; (2) to examine life support therapies prior to and after the issuance of a DNR order; and (3) to determine the clinical factors that influence (...) the initiation of a DNR order in ICUs in Taiwan. A prospective, descriptive, and correlational study was conducted. A total of 202 patients comprising 133 (65.8%) who had a DNR order, and 69 (34.1%) who did not, participated in this study. In the last 48 hours of their lives, patients who had a DNR order were less likely to receive life support therapies than those who did not have a DNR order. Older age, being unmarried, the presence of an adult child as a surrogate decision maker, a perceived inability to survive ultimate discharge from the ICU, and longer hospitalization in the ICU were significant predictors of issuing a DNR order for critically ill patients. This study will draw attention to how, when, and by whom, critically ill patients’ preferences about DNR are elicited and honored. (shrink)

Objective: To investigate the correlation between preoperative response to the L-dopa challenge test and efficacy of deep brain stimulation (DBS) on motor function in Parkinson’s disease (PD).Methods: We retrospectively reviewed the data of 38 patients with idiopathic PD who underwent DBS surgery with a median follow-up duration of 7 months. Twenty underwent bilateral globus pallidus interna (GPi) DBS, and 18 underwent bilateral subthalamic nucleus (STN) DBS. The Movement Disorder Society Unified Parkinson Disease Rating Scale-Motor Part (MDS UPDRS-III) was assessed before (...) surgery and at the last follow-up in different medication and stimulation conditions, respectively.Results: Pearson’s correlation analysis revealed a positive correlation between preoperative L-dopa challenge responsiveness and GPi-DBS responsiveness on the total score (R2 = 0.283, p = 0.016) but not on the non-tremor total score (R2 = 0.158, p = 0.083) of MDS UPDRS-III. Such correlation remained significant (R2′ = 0.332, p = 0.010) after controlling for age at the time of surgery as confounding factor by partial correlation analysis. The preoperative L-dopa challenge responsiveness was significantly correlated with the tremor-controlling outcome of GPi-DBS (R2 = 0.390, p = 0.003). In contrast, we found a positive correlation between preoperative L-dopa challenge responsiveness and STN-DBS responsiveness on the non-tremor total score (R2 = 0.290, p = 0.021), but not on the total score (R2 = 0.130, p = 0.141) of MDS UPDRS-III. The partial correlation analysis further demonstrated that the predictive value of preoperative L-dopa challenge responsiveness on the non-tremor motor outcome of STN-DBS was eliminated (R2′ = 0.120, p = 0.174) after controlling for age at the time of surgery as confounding factor.Interpretation: The short-term predictive value of preoperative response to the L-dopa challenge test for the motor outcome of GPi-DBS in PD was systematically described. Our findings suggest: (1) a solid therapeutic effect of GPi-DBS in treating L-dopa-responsive tremors; (2) a negative effect of age at the time of surgery on motor outcomes of STN-DBS, (3) a possible preference of STN- to GPi-DBS in L-dopa-resistant tremor control, and (4) a possible preference of GPi- to STN-DBS in elderly PD patients who have a satisfactory dopamine response. (shrink)

Tensions between respect for autonomy and paternalism loom large in Ferrario et al ’s discussion of artificial intelligence (AI)-based preference predictors.1 To be sure, their analysis (rightfully) brings out the moral matter of respecting patient preferences. My point here, however, is that their consideration of AI-based preference predictors in treatment of incapacitated patients opens more fundamental moral questions about the desirability of over-ruling considered patient preferences, not only if these are disclosed by surrogates, but (...) possibly also in treating competent patients. I do not advocate such an evolution—the moral desirability of that calls for a much broader debate, one in which the meaning of ‘doing good’ in medicine, and how this intersects with normative views on ‘the goal(s) of medicine’ would be central elements. While my aim in this piece is more modest, I nonetheless hope to approach it sideways, by indicating how the contribution by Ferrario et al reopens discussion about paternalism and the normativity of preferences in medicine. This follows from what these authors give as reason for endorsing this technology in care for incapacitated patients. Their argument for employing such tools in case of incapacitation is based on the premise that advice of surrogate decision-makers about care preferences is suboptimal because of their biased perception …. (shrink)

Introduction: Congruence, understood as the agreement between the patient's preferred place of death and their actual place of death, is emerging as one of the main variables indicating the quality of end-of-life care. The aim of this research was to conduct a systematic literature review on levels and determinants of congruence in palliative patients over the period 2010–2021.Method: A systematic review of the literature in the databases of PubMed, Scopus, Web of Science, PsycINFO, CINAHL, Cuiden, the Cochrane Library, CSIC (...) Indexes, and IBECS. Information was extracted on research characteristics, congruence, and associated factors.Results: A total of 30 studies were identified, mainly of retrospective observational design. The congruence values varied substantially between the various studies, ranging from 21 to 100%. The main predictors of congruence include illness-related factors, individual factors, and environmental factors.Conclusion: This review, in comparison with previous studies, shows that treatment-related factors such as physical pain control, marital status, having a non-working relative, age, discussing preferred place of death with a healthcare professional, and caregiver's preference have been associated with higher levels of congruence. Depending on the study, other factors have been associated with either higher or lower congruence, such as the patient's diagnosis, gender, or place of residence. This information is useful for designing interventions aimed towards greater congruence at the end of life. (shrink)

This study investigated patient preferences for various types of physician persuasion strategies. Four types of persuasion strategies were utilized which involved combination of high and low levels of affectivity and information. In addition, patient variables, receiver apprehension and health beliefs were introduced to predict preference choices by patients. Results indicated that patients are influenced in their decision-making (preferences) by the type of persuasive strategy employed. Further, patients with different characteristics and predispositions prefer different persuasive strategies. The results (...) of this study suggest that the success of physician persuasiveness is dependent upon the type of strategy used and the type of patient being persuaded. (shrink)

Introduction: This study, the first of its kind in Iran, was to assess Iranian patients’ preferences for receiving information and participating in decision-making and to evaluate their satisfaction with how medical information is given to them and with their participation in decision-making at present. Method and materials: 299 of 312 eligible patients admitted to general internal medicine or surgery wards from May to December 2006 were interviewed according to a structured questionnaire. The questionnaire contained questions about patients’ preferences regarding four (...) domains of information and their participation in decision-making. Patients’ responses were measured on a visual analogue scale graded from 1 to 10. Results: The mean (SD) score for desire to receive information was 8.88 out of 10 (1.5) and for participation in medical decision-making was 7.75 out of 10 (3). The desire to receive information was greater in women than men (9.0 (1.5) vs 7.8 (1.4), p = 0.025). It was also correlated with their education (r = 0.2, p = 0.001) and their estimation of the severity of their own disease (r = 0.13, p = 0.027). The score for preference to participate in decision-making was higher in women than in men (7.95 (2.8) and 7.0 (3.2), respectively; β = 0.8, p = 0.022) and was negatively correlated with education (r = −0.14, p = 0.015). Discussion: This study shows that Iranian patients are highly interested in receiving information about their condition and participating in clinical decision-making. No predictive variable for such attitudes was found; therefore, the only way for the physician to recognise patients’ desire is to ask them explicitly. (shrink)

Bioethics, Volume 36, Issue 4, Page 388-402, May 2022.

In this paper, I argue that patient preferences concerning human physical attributes associated with race, culture, and gender should be excluded from public healthcare robot design. On one hand, healthcare should be (objective, universal) needs oriented. On the other hand, patient well-being (the aim of healthcare) is, in concrete ways, tied to preferences, as is patient satisfaction (a core WHO value). The shift toward patient-centered healthcare places patient preferences into the spotlight. Accordingly, the design of (...) healthcare technology cannot simply disregard patient preferences, even those which are potentially morally problematic. A method for handling these at the design level is thus imperative. By way of uncontroversial starting points, I argue that the priority of the public healthcare system is the fulfillment of patients’ therapeutic needs, among which certain potentially morally problematic preferences may be counted. There are further ethical considerations, however, which, taken together, suggest that the potential benefits of upholding these preferences are outweighed by the potential harms. (shrink)

Despite the positive effects of including patients’ preferences into therapy on psychotherapy outcomes, there are still few thoroughly validated assessment tools at hand. We translated the 18-item Cooper-Norcross Inventory of Preferences into German and aimed at replicating its factor structure. Further, we investigated the reliability of the questionnaire and its convergence with trait measures. A heterogeneous sample of N = 969 participants took part in our online survey. Performing ESEM models, we found acceptable model fit for a four-factor structure similar (...) to the original factor structure. Furthermore, we propose an alternative model following the adjustment of single items. The German C-NIP showed acceptable to good reliability, as well as small correlations with Big-Five personality traits, trait and attachment anxiety, locus of control, and temporal focus. However, we recommend further replication of the factor structure and further validation of the C-NIP. (shrink)

Background: In the USA, the Food and Drug Administration waiver of informed consent permits certain emergency research only if community consultation occurs. However, uncertainty exists regarding how to define the community or their representatives.Objective: To collect data on the actual preferences and values of a group—those at risk for stroke—most directly affected by the waiver of informed consent for emergency research.Design: Face to face focused interviews were conducted with 12 patients who were hospitalised with a stroke diagnosis in the previous (...) year. The interviews were audiotaped and a transcript based method was used for their analysis.Results: All 12 participants felt “that it was important that new treatments for stroke be developed”, but they were initially confused about the distinction between “research for stroke” and “emergency research for stroke”. However, after explanation, most expressed willingness to participate in the latter. In the absence of a surrogate to give informed consent in a stroke emergency situation, the majority said they would want the physician to “go ahead and enrol them in the trial”.Conclusions: This study is the first to identify the values and concerns of individuals most directly affected by stroke emergency research. Further interviews and focus groups are needed to develop and test a validated questionnaire on the preferences and values surrounding emergency research for stroke. (shrink)