“I think what makes me feel the most uncomfortable is, in some ways, it feels like at this point it doesn’t matter what the results will be,” Mike told Liana during a conversation in his office one afternoon.¹ For three years, he had helped to conduct an RCT of a mental health treatment approach called Peer-Supported Open Dialogue (POD) in the UK’s National Health Service (NHS). While the study adhered closely to established “gold standard” methods, Mike stressed that it was unlike any trial he had been involved in before. Where clinical trials call for clearly defined and replicable interventions, POD was often described as an ethos, a way of life, or an ethical stance. While researchers claimed it would be years before they had enough evidence to draw conclusions about POD, many practitioners were already lobbying passionately for its rollout across the country.² “They’ve already made their mind up,” Mike said, “Like, what’s the point of all this work? Because they’ve already decided it’s good.”

Mike’s comments reveal conflicting approaches to the “good” in public mental health care that were brought into intense focus in debates surrounding the POD trial. Where supporters advocated the widespread adoption of POD based on its alignment with shared values, researchers like Mike maintained that only evidence of real-world impacts could tell us if this was the right way forward. This latter stance reflects the dominant public health paradigm of “evidence-based medicine,” in which particular types of quantitative data are considered the only sound justification for adopting a new therapeutic approach. In the absence of such evidence, practitioners’ enthusiasm for POD was sometimes described as “cult-like” or “evangelical”—a matter of blind faith.

This was a debate, then, about the role of evidence in healthcare planning and policymaking. Yet it did not break down along the usual lines. Anthropological critiques of evidence-based medicine over the past two decades have primarily focused on the paradigm’s epistemological blind spots. They have explored how culture and social relationships condition what counts as “evidence” (e.g., Adams et al. 2005; Biruk 2018), highlighted vital aspects of lived experience obscured by quantitative metrics (e.g., Adams 2016; Adams et al. 2016), and illuminated the disconnect between “effectiveness” within controlled clinical trials and the demands of care in complex, real-world settings (e.g., Mol 2006). Across this literature, such epistemological limitations are framed as the crux of tensions between the clinic and the lab.

In this article, we suggest that new frameworks are needed to understand contemporary debates about the role of evidence in mental health policymaking. Clinicians implementing the trial of POD did not object to the epistemological assumptions and knowledge practices of evidence-based medicine nor were trial researchers ignorant of their limitations. Rather, the debate on POD reflected two different tacit moral philosophies at work within public health decision-making today. While researchers like Mike expressed a utilitarian stance in which the right decision is the one that yields the most positive effects for the greatest number of people, practitioners expressed a position more closely aligned with anthropological formulations of virtue ethics (Roberts and Reich 2002). This latter stance emphasizes the embodiment of collectively held notions of the good, or virtues, as an end in itself (Mattingly 2014).

While the utilitarian approach has long predominated in evidence-based medicine, growing concern with “values-based” and “rights-based” psychiatric care is forcing decision-makers to reconsider how they account for the moral dimensions of service delivery. In this article, we use our ethnography of the POD trial to throw new light on this dilemma. While some anthropologists have called for distinguishing the “moral” from the “ethical” (defining these terms in various, sometimes contradictory ways), we follow Mattingly and Throop (2018) in using these terms interchangeably to refer to the domain of human life concerned broadly with the pursuit of the good, right, or just.

We first introduce the POD model and explore the moral stakes of the clinical trial from the perspective of practitioners advocating the approach. We draw on and extend Mattingly’s (2014) concept of “moral laboratories” to consider how the trial became a site of experimentation with practices that held potential to foster morally better futures for individual clinicians and British society more broadly. Next, we turn to the complex moral deliberations of trial researchers in that role to pinpoint differences in how researchers and practitioners reasoned about POD’s future. Our analysis reveals ethical dimensions of mental health caregiving that are rendered invisible within the utilitarian framework of evidence-based psychiatry. These “moral blind spots” include aspects of service delivery that exceed standardized outcome measures—such as affirming the humanity of people in crisis and the moral personhood of care-providers—as well as the potential contributions of therapeutic practice to larger projects of systemic and societal change.

An anthropological study of POD (APOD)

First developed in Finland in the 1980s, Open Dialogue is today internationally recognized as a rights-based model of mental health service delivery (World Health Organization 2021). In simplest terms, it is a method of responding to psychiatric crisis that involves the social networks of the “person of concern” (Olson et al. 2014).³ Therapeutic encounters called “network meetings” take the form of a dialogue facilitated by two or more practitioners trained in the approach. The practitioners’ role is not to diagnose or solve the problem but to facilitate a dialogue such that every voice in the room is heard and responded to; radically, dialogue, rather than the practitioner, is construed as the agent of therapeutic change. Meetings continue as and when requested by the network until collective decisions are reached about how to move forward.

Open Dialogue is explicitly framed as a “dialogical” counterpoint to Foucault’s famed characterization of psychiatry as a “monologue of reason about madness” (Olson et al. 2014). Practitioners are encouraged to abdicate their usual position of authority and to be present in their “full humanity,” including sharing their own experiences of mental distress where relevant and appropriate (Hopfenbeck 2015; Olson et al. 2014). They strive for transparency by avoiding all discussion of clients in their absence—a marked contrast with standard practice in the NHS, where decisions are often made at meetings attended by professionals alone. POD intends to support client autonomy and reduce involuntary admissions, for example, by responding to crisis within 24 hours, offering home visits, and encouraging professionals to “tolerate uncertainty.” The “peer-supported” variant of the approach embeds practitioners with lived experience of services use within clinical teams, while aspiring to a flattened hierarchy (Hopfenbeck 2015). Various reflective practices built into the model facilitate this way of working, including a form of intensified group mutual supervision called “intervision.”

Since its introduction in the UK in 2014, POD has found robust support among psychiatric service user/survivor and carer groups and amassed a dedicated following of clinicians. It has also piqued the interest of policymakers concerned with cutting costs, with research in Finland demonstrating significant reductions in expensive hospital admissions (e.g., Seikkula et al. 2011).⁴ In 2018, this interest culminated in the launch of the world’s first large-scale RCT of POD implemented in five NHS Mental Health Trusts across the UK.

The trial, entitled “Open Dialogue: Development and Evaluation of a Social Network Intervention for Severe Mental Illness” (ODDESSI), compared the effectiveness of POD with that of treatment as usual (Pilling et al. 2022). People experiencing mental health crises who were referred to teams delivering the control (treatment as usual) and experimental (POD) arms of the trial were invited to take part. Those who consented were followed over a period of two years with medical notes and staged questionnaires used to determine “time to relapse” (primary outcome), the “process of recovery,” client satisfaction and other outcome measures. Practitioners from experimental teams underwent one year of part-time training to deliver POD to trial participants (mostly alongside their ongoing usual practice with non-trial clients). Those opting into the training spanned disciplines (psychiatrists, peer support workers, social workers, nurses, etc.) and practice sites (inpatient facilities, crisis/home treatment, and community mental health teams), allowing continuity of POD through different stages of recovery (see Pilling et al. 2022 for the full trial protocol).

Our anthropological study, developed independently from the clinical trial, aimed to generate complementary insights: to shed light on social and institutional factors shaping trial results through immersive ethnography within POD teams. Our team of seven ethnographers—both anthropologists and clinicians (including lived experience as service users or carers)—completed POD training, becoming full members of trial delivery teams.⁵ NHS employed staff who joined our research team were offered ethnographic skills training. Both authors were anthropologists inducted into different POD teams on honorary contracts. We participated in all aspects of POD delivery, including facilitating client network meetings, writing clinical notes, and attending team meetings. Our dual role as practitioner-ethnographers afforded rich insights into factors affecting the trial while demanding intensive ethical and methodological reflection. As a research team, we reflected critically on how our different social and professional positionalities, our personal backgrounds and investments in strengthening NHS mental health services influenced both the knowledge we produced and the clinical care we provided.

With written consent from colleagues and clients, our ethnographic team kept detailed fieldnotes on our day-to-day experience, including participant observation within network meetings with over 30 clients/families. We also conducted semi-structured interviews with over 100 staff and 22 POD clients. Finally, we studied the conduct of the clinical trial, including recruitment, “fidelity” interviews with POD teams, and the rating of practitioners’ “adherence” to the model.⁶ The resulting body of ethnographic data was analyzed through inductive thematic analysis (see Mosse et al. 2023 for methodological details). Here, we draw on these ethnographic findings to explore how practitioners and researchers factored moral concerns into their deliberations about POD’s future.

A moral intervention in NHS mental health services

“I hate that Open Dialogue does this to me,” psychiatrist Jennifer joked, dabbing at her eyes with a tissue. Her tears, coming at the end of a weekly team intervision, were not of sadness, but arose as the discussion turned to the hope that the POD team inspired in her. Like Ahmed and Marie, Jennifer had become a vocal advocate lobbying to roll out POD beyond the geographic and temporal bounds of the trial. Reflecting on the intense pressures NHS clinicians faced and the “firefighting” that took up most of their time, she thanked her team sincerely for “keeping it [POD] alive.”

As Wright (2022) describes in detail, POD was introduced within and in response to a state of chronic crisis (“firefighting”) within NHS mental health services. Through a decade of austerity measures followed by a global pandemic, mental health professionals in the NHS have contended with worsening shortages of essential resources, from nursing staff to hospital beds. These working conditions have been linked to a surge in “moral injury” among care providers (Dean et al. 2019:400). The staff burnout, departures, and unfilled posts we witnessed at close range, together with findings of internal staff surveys during our fieldwork, all confirmed that clinicians found it increasingly difficult to practice in line with their values.

Yet these struggles had not gone unrecognized. During our fieldwork, several developments gave practitioners hope that meaningful change could be realized at the national level. The NHS began implementing a self-claimed “transformation” framework for Community Mental Health services guided by principles of person-focused and network-based care (NHS England 2019). In a more explicit bid at moral rehabilitation, it also set out new “culture of care standards,” which were intentionally developed in line with some of POD’s key principles, and a national “Culture of Care Programme” to improve the institutional culture of hospital wards for patients and staff (NHS England 2024; Royal College of Psychiatrists 2025). Key decision-makers also signaled their awareness of POD and its alignment with their agenda. NHS England’s National Mental Health Director, for example, gave glowing remarks at a conference on POD in the trial’s final year, despite the fact that results were not yet in. These developments demonstrated not only a growing concern with ethics in the UK’s public mental healthcare landscape but also a growing openness to making policy decisions on the basis of values.

Against this backdrop, the clinical practice of POD teams took on a wider moral significance. Practitioners like Jana, Marie, Ahmed, and Jennifer saw themselves as influential agents of system change. Their day-to-day clinical labors in the POD trial “kept alive” the dream of a future in which people in crisis set the agenda for their own care and treatment decisions were made in a transparent and inclusive fashion. As such, their experiences bring an important new dimension to theorizations of the clinic as a site of everyday moral experimentation.

Moral anthropological work on caregiving has predominantly focused on the fragmented and privatized US healthcare landscape, with interlocutors often far removed from clinical knowledge production and policymaking. Hopper (2006:223), for example, writes about the moral gestures of frontline clinicians in the American welfare system in the idiom of “quiet protest.” Likewise, Brodwin writes about “moments of refusal” through which frontline practitioners manage to carve out “a small zone of freedom” Brodwin (2013:20). Mattingly’s (2014) influential work on moral laboratories focuses on the aspirations of families suffering multiple forms of marginalization within public health and social care systems.

By contrast, British POD practitioners saw their individual moral labors as being yoked to wider projects of national change. The UK’s NHS is historically rooted in a social contract between the government and the public that guarantees free health care to residents according to need. In their overt attempts to refashion this core public institution at a moment when it seemed at risk of eroding, the trial of POD became an emotionally and politically charged space for renegotiating societal responses and responsibilities toward some of the country’s most marginalized: people experiencing acute mental health crises. Rather than merely a therapeutic intervention, POD emerged as a moral intervention into a mental health system in crisis. Recognizing these wider moral stakes of the ODDESSI trial is essential to understand the debates that have unfolded in the years since its launch. To stick with Ahmed’s farming metaphor, we might say that practitioners saw themselves as key actors in a movement to improve the treatment of all British farm animals, while the trial was concerned with measuring the quality and quantity of food produced in specific sites. Yet this characterization falls short of capturing the complex ethical commitments and deliberations of trial researchers, to which we now turn.

Rendering the ethical measurable

“Just because there is silence, doesn’t mean that’s a therapeutic dialogical silence,” explained Milly to a group of researchers on the screen in a Microsoft Teams workshop. POD researcher Milly was charged with training an international group of experts (whom David joined online) to rate the adherence of practitioners to Open Dialogue codes of practice. To draw meaningful comparisons and conclusions from trial outcome data, researchers needed to know that practitioners delivering POD were sufficiently adherent to the key elements and principles of the approach (Pilling et al. 2022). To this end, they had developed an Open Dialogue Adherence Rating Protocol/Manual and were trained in the assessment of recordings of network meetings (Lotmore et al. 2023).⁷ Assessment involved classifying practitioner utterances as “dialogical” or “monological” as well as rating the meeting as a whole against each of the 12 key elements of POD on a scale from “excellent” to “not acceptable.”

RCT methodology demanded that POD be defined and distinguished as a technical intervention: an organized set of practices or procedures which adhered to a set of objectively discernable features. But “rendering technical” (Li 2011) a moral intervention is a complex work. As Milly’s comment above suggests, deliberations about fidelity often centered on qualities of the clinical encounter that were difficult to tack down to observable procedures. Indeed, the debate about how to rate silences from which this comment was drawn reveals the uncertain and often explicitly moral nature of the deliberations involved in rating fidelity:

[There is] something which gave me the feeling that they were not really emotionally synchronized, the seconds, or more than seconds [of silence], after [the client] has shared some problematic topic felt like emptiness for me.

Yes, [I] shared doubts about whether they were … more doing a meeting rather than being in the moment. It felt as if they were doing a dialogical meeting, but it wasn’t—it lacked a sense of [practitioners] feeling their own vulnerability or empathy. So that’s why on [my] first superficial listening the scoring was “acceptable.” But now, I’m leaning towards more “unacceptable” because maybe a lack of attunement. I’m still in doubt to be honest…

I resonate with this when I first listened to the tape, but then I felt there was something missing. It’s not only the case of giving space which impressed me in the beginning … I listen to it again and again, and then suddenly I very much [realized] just saying nothing is not necessarily an advantage …

I’m just wondering because for me, it’s a little bit er, I know what you mean, but it’s also complicated, because you said you try to do this rating in an empirical way, but at this point it’s not empirical. It’s—you know—rating if the silence is dialogical or monological is about the impression of the rater, and it relates to what uncertainty can I tolerate, for example, as a rater.

In conversations like these, raters openly acknowledged that some of the defining qualities of POD did not lend themselves readily to empirical observation. These qualities mapped closely onto ethical aspects of the approach most valued by practitioner-advocates: a human and humane relationship with clients, a genuine redistribution of power, a commitment to being alongside someone in distress. Ratings were influenced by the researchers’ sense of practitioners’ “overall connection” or “affective attunement” with clients; whether they were “emotionally synchronized,” and “empathetically engaged.” The counterpart negative overall impressions were of meetings that seemed “empty,” “superficial,” or “emotionally disconnected,” even where initial utterance ratings had high scores. What distinguished genuine POD from other treatment approaches came down to qualities that could sometimes only be detected in the form of “gut feelings,” and yet a commitment to the trial’s positivist knowledge required empirical measurement of the intervention.

Disagreement persisted over whether and how gut feeling could be incorporated within the rating process. Raters’ judgments hovered ambiguously over the distinction between the procedural and the ethical. For some researchers, like Milly, rating introduced a degree of systematicity and objectivity to what were ultimately subjective judgments: “a lot of this is gut feeling, but by using these ‘hashtags’ [scoring criteria] does make it a little more empirical.” For other raters, however, it was, “dangerous to go for a gut feeling,” and important to “stick to the manual, the behavioral components.”⁸ Through the back-and-forth of such deliberations, trial researchers were able to find agreement broadly on what counts as acceptable POD practice. The inter-rater reliability required was easier to achieve in some recordings, some criteria and some kinds of utterances than others.⁹ Regardless, for the purposes of a trial there was a necessary commitment to defining POD as an objectively measurable procedure (not just an ethical stance).

Paradoxically, then, trial researchers assessed the technical measures of the “dialogical” in terms of an ethic of emotional connection, often making this assessment by “gut feeling.” In doing so, they strove, as one of the trial researchers put it, “to capture the spirit and the ethos of Open Dialogue” through technical measurements. While POD is distinctive in the degree to which it fuses ethical principles and specific therapeutic practices, this attempt to operationalize and evaluate the core ethos of the approach bears resemblance to other efforts to measure an ethical stance within mental health caregiving—for example, scales for procedural justice and coercion (e.g., Watson et al. 2010). The assumption cutting across these initiatives is that measurement offers an essential window onto how ethical, just, or right a particular treatment approach really is. It is in this core assumption that researchers in the POD trial diverged from practitioners advocating for the approach’s immediate rollout. While ethical judgment was important for both POD practitioners and trial researchers, there was a key difference in how these considerations figured into their reasoning about the future of the approach.

Meta-ethical tensions

“We have pulled our hair at times,” Mike explained to Liana,

“with the clinicians who have really advocated for Open Dialogue to extend its remit from beyond just the [trial team] and for it to be offered on a wider basis. That could be incredibly unethical because if they go ahead and do that and then [the trial] comes back in 2024, 2025 and says, it’s no better … I don’t think there is the evidence at all.”

Mike’s mobilization of the “ethical” here reflects what Gupta (2003) calls the implicit moral mandate of evidence-based medicine: that we ought to pursue health through the most effective means possible, with evidence-based medicine offering the best available methods of identifying those means. By these standards, advocating the national rollout of POD during an ongoing trial was a morally questionable prospect, at best. A 2019 review had determined that existing evidence on the effectiveness of POD was low quality, concluding that further clinical trials were needed to make decisions about the model’s application in clinical services—hence the ODDESSI trial (Freeman et al. 2019). In the absence of trial results, some described practitioners’ advocacy of POD as “fervent,” “evangelical,” “religious,” and “cult-like,” suggesting it was rooted in a blind faith rather than empiricism.

At first glance, these tensions might be put down to conflicting views on the value of “objective” trial evidence compared with the more “subjective” expertise of clinicians. Yet clinicians and researchers we spoke with were both heavily invested in producing evidence through the clinical trial while expressing similarly nuanced perspectives on its limitations—indeed, both welcomed the presence of practitioner-ethnographers within trial teams to capture more subjective dimensions of the approach. Rather than an epistemological misalignment, we suggest the conflicting visions of trial researchers and practitioner-advocates reflect two distinct moral philosophies guiding public mental health decision-making today.

On one hand, Mike’s critique embodies an approach to ethical analysis philosophers call “utilitarianism” (Roberts and Reich 2002). In this view, no action is inherently right or moral. Instead, we must judge each action based on its consequences in the world. The most ethical policy decision is one in which negative outcomes are minimized and positive outcomes are maximized (with both equitably distributed). In the context of mental health care these outcomes are most often objectively assessed indices of clinical recovery, although they may also include subjective measures such as “client satisfaction.” Irrespective of how a positive outcome is defined, utilitarianism generally demands that we measure these effects in a representative sample to ascertain whether adopting a particular treatment is the right thing to do.

This utilitarian approach to ethical analysis fits hand in glove with the paradigm of evidence-based medicine, which outlines a reliable method of obtaining and evaluating such measurements (Roberts and Reich 2002). For researchers in the POD trial invested in this wider paradigm, it was not merely the ethical stance of POD practitioners that mattered but that stance’s measurable effects through specified interventions.¹⁰ This position was summed up by Stephen Pilling, Principle Investigator for the ODDESSI trial, at our first APOD conference:

… we are in the business of improving people’s lives and we ought to find ways in which we can objectively measure that, and that is what the ODDESSI program seeks to do … If you do that [measure whether or not people get better and stay better] and build that routinely into practices, then I think you have the basis on which to start making a real argument for change. (APOD 2023)

On the other hand, many POD practitioners we spoke with were more closely aligned with neo-Aristotelian formulations of “virtue ethics” that have recently captivated anthropologists. In this approach, the goodness of actions is judged by the degree to which they embody virtues, or collectively held ideas about what makes a good person and a good society. Mattingly (2014) has elaborated “first-person virtue ethics” as a moral philosophy among caregivers of people with health problems and disability, drawing attention to how gestures of caregiving are imbued with hopes and aspirations for a morally better future. Her account emphasizes moral striving within the messiness of everyday life, where multiple conflicting value systems come to bear on the concrete situations we are called to respond to.

Crucially, first-person virtue ethics sees people as engaged in a continuous process of moral becoming, in which virtuous action is already an end in itself. In the chaos of real life, where causes and effects, beginnings and endings, are rarely clearly defined, the moral value of action cannot be straightforwardly reduced to the question of that actions’ consequences. Indeed, Mattingly (2014) suggests that the most valiant efforts to create morally better lives can be risky wagers, attempts to begin something new “against the odds.” Casting the relationship between moral aspirations and real-world effects as inherently unpredictable and uncontrollable, Mattingly’s work reveals how the attempt to realize a better future can be considered worthwhile even when it is unlikely to bear fruit in the short-term, or at all.

In the context of the clinical trial of POD, this approach suggests that positive results as quantified in the trial were not prerequisite to arguing that POD should be widely adopted in mental health services; rolling out POD could be the right thing to do because it embodied the ideas of a good clinician and a good society that practitioners hoped to realize. As one participant summarized this stance,

Morally and ethically it’s quite hard to argue against the principles of it. So, what exactly are we worried about? Why is there even the ODDESSI trial? Because should we not just be doing this stuff?

While practitioner-advocates were deeply concerned with clinical outcomes and curious about the results of the trial, these did not capture the full range of what was at stake, ethically, in their practice. The framework of first-person virtue ethics makes space for recognizing these moral dimensions of mental health caregiving that are rendered invisible within dominant forms of clinical evidence-making. We turn now to a discussion of some of these “moral blind spots.”

The moral blind spots of evidence-based psychiatry

During an interview over Zoom one afternoon, a nurse named Angela told Liana about a client who felt unable to keep himself safe for a period of time. In line with the principles of the approach, Angela and her co-practitioners sought to “tolerate uncertainty,” holding a series of network meetings in lieu of taking more intrusive steps, such as involuntary admission or referral to the home treatment team. “It’s certainly tough,” she said, “especially when you know someone quite well,”

But, through conversations with him and his network, we were able to just leave it as it is and trusted him and his network to—to navigate that … difficult time without feeling the need to drag someone to hospital against their will or sending a whole other team of people that don’t know him that just turn up for 15 minutes a day … You know, I’m sure the outcome would have been the same. But it felt on reflection like a much better way of helping someone through.

Angela’s words provide a useful entry point for exploring how approaches to ethical analysis concerned with “the outcome” might fail to capture what is “much better” about the POD approach from the standpoint of practitioners.

First, Angela suggests that the “outcome” as defined by objective measures of recovery may fail to capture ethical aspects of care delivery (the journey to outcome) that matter greatly to both practitioners and people using services. Kleinman’s (2012:1551) descriptions of care as an existential practice that “changes the subjectivity of both the caregiver and the person receiving care” help to elucidate this. He suggests caregiving embodies individual and collective ideas about both the recipient’s deservingness and the moral personhood of the provider (Kleinman 2012). Good care can involve not only practical assistance toward the goal of ameliorating suffering but also the labor of acknowledging and affirming the humanity of the one suffering. This may be particularly important in the context of psychiatry, in which, as Luhrmann (2000:284–85) describes, “The popularized … medical model invites us to see the mentally ill as not quite human, particularly if their problem is chronic and unremitting.” This dehumanization is not just a matter of individual professionals’ attitudes, but as Armstrong (2024) argues, and POD advocates’ accounts show, something embedded within and sustained through the organization of care within status quo NHS treatment.

The affirmation of clients’ humanity and agency are immediate “goods” within mental health caregiving that can influence clients’ lives in ways that exceed “recovery.” POD practitioners also expounded the benefits for clinicians of practicing in a way they regarded as more “human and humane.” Their accounts clearly embody an orientation toward ethics as a process of moral becoming, in which the way they interacted with clients both reflected and shaped who they were as a person in the world. These ethical dimensions of caregiving are significantly overlooked in the clinical trials of effectiveness that form the basis of most public health decision-making.

While there are some nascent attempts to quantify ethical aspects of service delivery (e.g., coercion) in ways that could conceivably be incorporated in trial methodologies, the POD approach troubles the idea that this myopia can be addressed simply through new and better modes of measurement. As Cubellis’s (2022) rich ethnography of German Open Dialogue practice shows, POD involves a fundamental rethinking of outcome both in affirming the agency of the person of concern to define treatment goals and through its attempt to respond to goals and desires distributed across that person’s network. A core tenet of POD is allowing clients and their networks to set the agenda for care, suggesting that any attachment to pre-defined metrics of success may hinder the therapeutic process. While for some networks avoiding involuntary hospitalization might be the hallmark of ethical care, there are other networks where members consider this the most appropriate and compassionate response to crisis. POD’s relational approach thus raises questions about not only how but also the standpoint from which we might objectively measure how good, just, or right a particular response to a mental health crisis is (Cubellis 2022).

A second, and more profound, challenge to utilitarian ethics apparent in Angela’s account has to do with the uncertain relationship between actions and effects. Angela’s description of the “toughness” of adhering to the approach in this case attests that a commitment to the principles of POD is not at all the same as certainty about the effects clinical practices will have. The key POD principle of “tolerating uncertainty” reflects an acknowledgment that we cannot know in advance what actions will lead to the outcomes desired (whether by a person of concern, their network, clinicians, or the mental health system). Intervisions were peppered with questions reflecting this: “Where is this going?” “Is this the right thing to do?” Practitioners regularly voiced awareness that in specific circumstances our practice was imperfect or flawed, a sense that “network meetings are not enough,” or on the contrary, surprise at positive changes in clients’ lives. If there was “faith” in POD, then, it was often more like “trust.” Amidst tangled and emotionally demanding uncertainty, it was often said, “trust the model,” “trust in the process.” This might mean trust your colleagues or your clients in crisis, when outcomes were ultimately uncertain. Rather than the religious certainty that some observers attributed to them, POD practitioners seemed to hold with conviction the freedom to speak honestly and process the uncertainty they felt in relation to clients.

First-person virtue ethics gives us a framework for understanding how adopting POD can be construed as the right thing to do even as practitioners readily acknowledged that its effects were by no means assured. This framework emphasizes the embodiment of shared values as a valid rationale for action in its own right. It calls for being the change one wishes to see in the world even when there is a real possibility of failing to realize this change in the unpredictable messiness of life-as-lived—or in the case of clinical work, in the face of complex suffering. This philosophy was evident in interactions with individual clients/networks, but it is perhaps brought into sharpest relief when we consider POD practitioners’ speculation about the results of the clinical trial.

All of the practitioners we interviewed readily acknowledged the factors potentially militating against trial success. These included an institutional context that made it almost impossible to uphold key principles of the approach and a wider culture of services in which many clients had come to expect directive, clinician-led treatment. Most openly described some cases where individual clients failed to benefit from, or even explicitly rejected POD. As Ahmed’s eloquent metaphor for the transformations involved in POD’s adoption captures, the model faced initial resistance on all sides. Many clients were not “used to” or comfortable with an approach that handed them the reins, and there were numerous instances when practitioners’ efforts to be dialogical were construed as awkward or unprofessional.

From the standpoint of first-person virtue ethics, these actual and anticipated failures did not undermine the moral worthiness of practitioners’ attempt to bring about systemic and societal changes. Indeed, they were often interpreted as a reflection of the very entrenched problems practitioners were working to address—the interpersonal and institutional resistance that could be expected in the face of a disruption to the status quo. Recognizing these longer-term, collective stakes of therapeutic practice poses a challenge to the idea that the measurement of effects is the only sound basis for moral action. Indeed, evidence-based medicine may be particularly blind to the promise of interventions that meaningfully shift institutional cultures and ways of working—processes that are inevitably gradual, risky, and fraught with resistance.

Conclusion

In the mid-2020s, British psychiatry is in a challenging predicament. On the one hand, it is more committed than ever to ethical practice. From the halls of government officials to the staff rooms of Mental Health Trusts, there is active discussion of the need for moral transformation. This is attested by the growing representation of people with lived experience and carers in clinical guidelines and policy-making processes, in the widespread adoption of “values” by medical institutions and clinical bodies and in initiatives to transform the dominant “culture of care” (Rethink Mental Illness 2017; Rose 2019; Royal College of Psychiatrists 2025). Practices that once occupied the fringes of mental health systems, such as peer-to-peer support, are today being integrated into mainstream services (Rose 2019).

On the other hand, public mental health remains wedded to evidence-based medicine as a means of guiding decision-making. And within the utilitarian approach tacitly embraced by this paradigm, the moral dimensions of clinical interventions become visible only to the extent that they lead to measurable outcomes or effects. In this article, we have drawn on some of the rich work coming out of contemporary moral anthropology to throw light on the shortcomings of this approach. We have considered how caring ethically may call for ceding an attachment to clinical effects and control of the therapeutic process, posing challenges to conventional understandings of intervention and outcomes in mental health services. We have further argued that the moral stakes of mental health caregiving extend beyond individual clients’ treatment experiences to encompass the moral personhood of providers and their attempts to shape wider systems and society.

The POD trial makes a valiant attempt to square calls for values-based care with the demands of evidence-based medicine. It operationalizes and encodes an ethical stance in a set of discrete criteria, and then measures the outcomes of clinical care applied in this way. In order to find validity in the eyes of evidence-based medicine, POD has adopted the language and form of a technical, rather than moral, intervention. And if the results of the trial are as impressive as supporters expect, the divergent moral philosophies explored in this article may not need to be reconciled to decide POD’s future.

Yet the tensions here described have implications for health planning and policymaking beyond the future of POD. As the NHS increasingly subscribes to values- and rights-based approaches to mental health care, decision-makers must reach agreements about how to weigh claims based on observable outcomes against claims based on values. Without such agreement, it is often left to individual clinicians to reconcile these tensions—to carve out space for more ethical ways of working within systems that prioritize the most efficient route to “good outcomes” that can be measured. The high rates of burnout in the NHS reveal the significant emotional and health systems toll of failures to effectively account for the moral dimensions of mental health caregiving in policy decisions. Anthropological theory and research can support better decision-making by shining a light on the moral blind spots of other forms of evidence-making. This may prove particularly important within efforts to remake the “culture of care,” which can provoke initial resistance in the course of more subtle and gradual processes of transformation.

Beyond their implications for public mental health, the experiences of POD advocates in the UK also hold lessons for anthropology. They call for a more robust theorization of first-person virtue ethics that accounts for the collective moral stakes of caregiving within nationalized health care systems. They reveal how mental health trials may become highly influential moral laboratories. Finally, they show how attention to projects of hope and moral aspiration can enrich the anthropology of psychiatry, taking us beyond well-trodden insights into the violence of services to imagine ways of caring otherwise.