Painful Subjects: Treating Chronic Pain among People Living with HIV in the Age of Opioid Risk.

Public narratives often attribute the opioid overdose epidemic in the United States to liberal prescribing practices by health care providers. Consequently, new monitoring guidelines for the management of opioid prescriptions in patients with chronic pain have become recognized as key strategies for slowing this tide of overdose deaths. This article examines the social and ontological terrain of opioid-based pain management in an HIV clinic in the context of today's opioid overdose epidemic. We engage with anthropological analyses of contemporary drug policy and the nonverbal/performative ways patients and clinicians communicate to theorize the social context of the opioid overdose epidemic as a "situation," arguing that the establishment of new monitoring strategies (essentially biomedical audit strategies) trouble patient subjectivity in the HIV clinic-a place where that subjectivity has historically been protected and prioritized in the establishment of care.

Introduction

Dr. Seles, an HIV specialist in a large, urban safety‐net hospital, recently denied a request for opioids. Long‐term opioid therapy (LTOT) is what Dr. Seles, and several physicians before him, had used for years to help this patient function, despite the patient's chronic pain from past injuries. But recently, a routine urine drug test (UDT) indicated that the patient was not taking his medication.

Dr. Seles: I can tell when he [this patient] is uncomfortable. And I think we had him on a fairly good [opioid] regimen that wasn't excessively high dosage. But then his urines weren't showing that he was taking the [prescription opioid] drugs and they were showing up other things. … It turns out he was shooting heroin and got hooked on it and was diverting our drugs to get the heroin.

Jennifer J. Carroll: And how did you end up resolving that issue? Or is it resolved, I guess, is the question.

Dr. S: Well, no, it's still not resolved. I told him that just because he was showing other substances [in his urine] and not the prescribed substances that I couldn't continue to prescribe [to] him. So, I just forced upon him a tapering schedule and said, “The option is to go to Dr. Gianopoulos [an addiction medicine specialist] to get Suboxone [an FDA‐approved medication for treating opioid use disorder].” He did and he was on it for a while. It's still not controlling his pain.

JJC: Do you feel this patient might eventually be able to return to opioid therapy for his pain, or is that off the table?

Dr. S: Oh, that's what he wants, yes.

JJC: Oh, that's certainly what he wants, but in your opinion as his provider?

Dr. S: No, right, I mean, part of me wants to do that right away. I want to make this guy comfortable. But fool me once, shame on you. I just don't feel like I can put him in a situation he can't handle.

An estimated 100 million American adults live with chronic pain (Tompkins et al. 2017), and approximately 10 million (about 3% of American adults) receive LTOT for non‐cancer pain (Boudreau et al. 2009). Today, physicians often find themselves feeling “caught in the middle—acknowledging the national crisis of opioid addiction and wanting to adhere to the new [safe prescribing] guidelines, but also wanting to decrease patients’ pain” (Serafini 2018). Dr. Seles's account highlights the deeply relational nature of these concerns. The reality of his patient's pain is something that Dr. Seles “picks up on” through social interaction as much as it is something he “sees” through the medical gaze (Foucault 1994). He tracks it through his patient's behavioral embodiment and verbal externalization of that pain experience—through what we might call a clinical habitus of bodily pain (Bourdieu 1977). Dr. Seles still sees his patient's pain through the clinical choreography of that habitus, just as he did before. Yet that knowledge was disrupted by a UDT, clearly indicating his patient was not taking the opioids as prescribed. When faced with a perceived choice between effective pain control and patient safety, Dr. Seles chose the latter, but he remains ambivalent about it.

Questions of pain care are especially fraught in the biomedical sphere of HIV care. At one time, before the development of effective HIV treatments, all pain care provided to patients living with HIV was considered palliative (Barton 1991), risks of substance use disorders, drug diversion, and overdose “rendered largely moot by the imminence of death” (Crowley‐Matoka and True 2012). Today, effective antiretroviral therapies have transformed HIV into a chronic manageable condition. As a result, new clinical urgencies have emerged, like helping persons with HIV achieve and remain virally suppressed on antiretroviral therapy, a goal that requires keeping patients in treatment, which many clinicians admit may be supported through liberal opioids prescribing (Tsui et al. 2019).

This article examines today's intertwined moral and ontological conflicts in the realm of chronic pain care. Previous scholarship has examined the moral ambiguities of pain care (Buchbinder 2011, 2015; Stonington 2015) and the pressures that these ambiguities place on the professional and moral identities of health care professionals offering (or denying) requests for pharmacological interventions in an era of enhanced pharmacovigilance (Carroll et al. 2018; Crowley‐Matoka and True 2012; Knight et al. 2017). This article contributes to these conversations by examining institutional responses to the opioid overdose epidemic and changing clinical norms. Though opioid therapy has been historically less regulated as a medical and commercial practice in the United States than elsewhere in the world (Meyer et al. 2020), new opioid prescribing guidelines for U.S. practitioners have urged close patient monitoring through a variety of clinical tools intended to produce concrete, objective indicators of a patient's adherence to doctor's orders (Dowell et al. 2016). Pain, however, remains a fundamentally subjective phenomenon made meaningful not through objective clinical measures but through therapeutic encounters that acknowledge the social reality of a lived pain experience (Buchbinder 2015).

This article also responds to the work of Jarrett Zigon (2015) and Jeanette Pols (2005) by considering therapeutic encounters with pain patients as “situations” in both the scalar sense (wherein large‐scale events, like the opioid overdose epidemic, and small‐scale events, like therapeutic encounters, mutually influence each other) and the performative sense (wherein risk, responsibility, and pain experiences are socially produced). Widely circulating social narratives about prescription drug misuse have promoted reasonable concern for patient safety and interest in adopting safer opioid prescribing practices (Dowell et al. 2016). At the same time, we argue, these narratives promote biological audit practices that “subtract the individual” (Foucault 1994) and can trouble patient subjectivity in the HIV clinic—a place where that subjectivity has historically been protected and prioritized.

To demonstrate this finding, we begin by describing the evolving bioethics of pain care in the United States; we then discuss the impacts of institutional responses to the opioid overdose epidemic on chronic pain care in an urban HIV clinic. Lastly, we discuss how these changes have forestalled strategies of interpellation, or relational practices of subject‐making, that have long been privileged in therapeutic encounters for pain and for HIV.

Methods and Participants

The data presented here were collected in conjunction with the Targeting Effective Analgesia in Clinics for HIV (TEACH) study, a randomized trial designed to test the effect of a collaborative care model on HIV care physicians and their monitoring of patients receiving LTOT for chronic pain (ClinicalTrials.gov Identifier: NCT02564341). Recruitment for the parental trial and for this qualitative study took place in an outpatient HIV clinic located in a large, urban, safety net hospital. The parent trial randomized physicians to either: (1) an intervention group (which received an initial didactic session on safer opioid prescribing practices, academic detailing to support evidence‐based patient care, access to pain specialists, and the support of a nurse care manager [NCM]), or (2) a control group (which received an informational brochure but no additional support). The analysis presented here emerged from the opportunity created by the parent study to engage in meaningful, semi‐ethnographic research in the rapidly evolving social worlds of pain care.

Data collection was completed between spring 2016 and summer 2017 both alongside the TEACH intervention and after the trial concluded. Semi‐structured interviews were conducted with six intervention physicians, four control physicians, three patients of intervention physicians, and three patients of control physicians. Participants were selected for interviews by the NCM and the study coordinator through a range‐maximizing convenience sampling method. All who were invited to an interview consented to participate. A medical anthropologist (J.J.C.) carried out the informed consent process and all qualitative data collection.

Analysis of interview data began after data collection had ceased. All authors reviewed the transcripts and participated in open discussions about emergent themes and noteworthy observations. These discussions facilitated mutual education and coordination of concepts across this interdisciplinary team of professionals in medicine, epidemiology, public health, and cultural anthropology, among other disciplines. Major themes deemed worthy of further scrutiny included perceived changes in clinical practice, patient risk and safety, pain care, and patient–physician relationships—all of which appeared to be connected to discourses circulating about the ongoing opioid crisis in the United States. Subsequent coding exercises were carried out by J.J.C. to further explore perceived changes to therapeutic encounters between patients and physicians as well as the environments in which those encounters took place. The findings presented here emerged in this final stage of analysis.

All names are pseudonyms. This study received ethical approval from the Institutional Review Board at Boston Medical Center.

A Brief History Pain Treatment in the United States

In the 1980s, two emergent trends laid the groundwork for contemporary methods of pain management. First, health care professionals began to campaign for the recognition of pain care as an ethical issue (Donovan et al. 1987; Sriwatanakul et al. 1983). As one publication in the New England Journal of Medicine put it, “Pain is soul destroying … few things a doctor does is more important than relieving pain” (Angell 1982). Simultaneously, the belief that LTOT placed patients at negligible risk of opioid addiction gained traction. An enormously influential letter published in the New England Journal of Medicine reported that, out of 11,882 patients under review who had been hospitalized and given opioids for pain, only four subsequently experienced opioid addiction (Porter and Jick 1980). Though barely 100 words in length, methodologically questionable, and subject to minimal peer review, this letter has been cited more than 600 times since its original publication (Leung et al. 2017).

Alongside these evolving conversations about pain, AIDS (the result of advanced HIV disease) emerged as an acute syndrome that spread rapidly. In the United States, the number of new infections annually peaked at about 130,000 in 1984–85 (Moore 2011), abating only with the development of combination antiretroviral therapy in the late 1990s, which extended the life expectancy patients living with HIV several decades past their initial diagnosis (Gueler et al. 2017). The clinical management of HIV was, from its inception, developed on the assumption that AIDS‐related death was imminent and that improving the quality of whatever life patients had left was paramount (Glare 1994). Though combination therapies now allow life expectancies for people living with HIV that rival those of the general population (Broder 2010), HIV continued—and continues—to hasten early death in absence of antiretrovirals, thus keeping the urgency of effective palliative care at the forefront of HIV medicine well into the 2000s (Livingstone 2003).

By the 1990s, effective pain management was considered a fundamental patient right and a primary responsibility of health care providers in the United States, regardless of prognosis. In 1992, the Agency for Health Care Policy and Research published new pain management guidelines that asserted that “the ethical obligation to manage pain and relieve the patient's suffering is at the core of a health care professional's commitment” (U.S. Agency for Health Care Policy and Research 1992). At the turn of the century, the American Pain Society launched a campaign to treat pain as the fifth vital sign, or “P5V” (Tompkins et al. 2017). This meant monitoring pain “as carefully and regularly as we measure heart rate, blood pressure, respiratory rate, and temperature,” and, “view[ing] the pain with the same degree of urgency as we would view a fever or any other vital sign change and immediately take steps to relieve it” (Winslow 1998). The P5V campaign was quickly adopted by the Veterans Health Administration (VA) and was further institutionalized in 2001 when the Joint Commission on Accreditation of Health Care Organizations mandated that accredited health care organizations assess pain in every patient to receive federal funding (Tompkins et al. 2017).

In practical terms, implementing P5V transformed patients’ own assessments of their pain into key clinical indicators. Myriad clinical tools were developed in an attempt to standardize such pain reporting. Medical anthropologist Kelly Ray Knight has noted that one of the most ubiquitous strategies for representing patient pain in the clinic at this time was the FACES pain scale, which used crude drawings of human faces expressing a spectrum of embodied experiences ranging from joy to agony. Knight and colleagues argue that the use of such pain scales not only fostered increased use of opioids to manage pain, but also (ironically) reified the biomedical orientation to pain as fundamentally patient specific, even as attempts were made to assess pain in a more objective fashion (Knight et al. 2017).

In recent decades, the growing profit motive of the U.S. health care system has also selectively pressured clinicians to simplify their approach to pain care. More and more forms of service became subject to narrowly defined billing codes and fee‐for‐service payment schemes (Tompkins et al. 2017), all of which undermined more holistic approaches to pain. The administrative rules for Medicaid reimbursement and the ethical imperative to see everyone who needs to be seen often leaves providers feeling like too much is “squeezed into the 15 or 20 minute appointments allowed” (Satterwhite et al. 2019). In this context, concurrent pressures to treat pain aggressively and to treat as many patients as possible normalized the use of pharmaceutical analgesics (opioids in particular) as front‐line pain management tools. They were an easy and effective solution that any provider of any specialty could use.

Today, enthusiasm for opioids as a front‐line tool for chronic pain management has dampened. Research from the VA indicates that pain outcomes in hospitalized patients did not significantly changed in the first decade after instituting P5V in 1995 (Mularski et al. 2006). The 2016 Opioid Prescribing Guidelines, published by the U.S. Centers for Disease Control and Prevention (CDC), offered a review of current medical literature, finding scant evidence that LTOT is effective for managing chronic pain (Dowell et al. 2016). Additionally, the CDC's 2016 guidelines brought new waves of public support for the legal enforcement of pharmacovigilance. Specific audit practices recommended by the CDC for reducing the risks posed to patients by prescription opioids include: checking a patient's history of prescriptions for controlled substances in their state's prescription drug monitoring program (PDMP) at the initiation of LTOT and regularly thereafter; and UDT at the initiation of LTOT and at least annually thereafter (Dowell et al. 2016). By 2017, 26 states had enacted legislation mandating the strict adherence to some or all of the recommendations made by the CDC in the 2016 guidelines (Davis et al. 2019), despite the CDC's express statements that the recommendations were intended to be “voluntary, rather than prescriptive standards” (Dowell et al. 2016, 2019).

These new regulatory technologies constitute what Julie Netherland and Helena Hansen have termed “the war on drugs that wasn't” (Netherland and Hansen 2017), a new assemblage of audit practices developed in response to a substance use “crisis” that has been predominantly coded as a “White” problem, which sits in stark contrast with the punitive drug control policies that have disproportionately pulled Black and Brown individuals into the criminal justice system. Hansen's work, in particular, represents the vanguard of anthropological critique of the racial coding of the “opioid crisis” in the United States (Hansen et al. 2016; Netherland and Hansen 2016). While significant ethnographic attention has been given to the experiences of racial minorities (predominantly Black and Hispanic persons) within and beyond institutions intervening on substance use (Bourgois 2003; Carr 2010; Garcia 2010), Hansen turns her analysis toward the ways in which “white race is encoded into biomedical technologies and practices” (Netherland and Hansen 2017), especially in the context of substance use and addiction, often serving (in public discourse at least) to shift blame away from White individuals for their substance use onto other figures, including physicians, drug companies, and even the drugs themselves (Netherland and Hansen 2017).

In this social climate, defined by competing medical and moral imperatives, many health care providers are now reconsidering the aggressive, pharmaceutical pain care strategies they were taught as medical students and residents (Tsui et al. 2019), asking themselves a new set of set of seemingly unanswerable questions. Will restricting access to opioids harm the patient–provider relationship? Will that stress cause a patient with HIV to lapse in treatment and become infectious? What remedy exists for these competing narratives that require patients’ subjectivities be simultaneously acknowledged and denied? Physicians from the TEACH study struggled with these very questions.

Scalar Situations

Public health responses to opioid misuse have generally sought to produce knowledge about human health through population‐wide surveillance practices such as the collection and curation of data on overdose events, emergency department admissions, prescriptions, fatalities, and many more—a biopolitical variation on what Marilyn Strathern has called “audit culture” (Strathern 2000). The CDC has dedicated millions of dollars toward the improvement of these modes of surveillance for opioid‐related health events in 32 U.S. states (CDC 2018). How, then, can ethnographic methods trace this new enthusiasm for public health responses through the clinical encounters happening in Dr. Seles's clinic?

It is possible to make sense of events at both scales (the national and the clinical) if we conceive of them, collectively, as “a situation.” Here, we predominantly use this term as proposed by Jarrett Zigon (2015), yet we also employ several insights from Jeannette Pols (2005). Zigon developed his particular conception of a situation while studying the global drug war, finding himself in need of a theoretical concept capable of framing the varied yet interconnected effects of the drug war on individual lives. As he observed, the drug war is neither “a singular issue or a totalized strategy,” nor should it be considered as “localized manifestations [as] in parts of Colombia, Mexico, or American inner cities” (2015: 501). As he proposed, the situation:

allows us to consider that which is widely diffused across different global scales as a nontotalizable assemblage, yet in its occasional and temporary local manifestation allows us to understand how persons and objects that are geographically, socioeconomically, and culturally distributed get caught up in the shared conditions emerging from the situation. Becoming caught up in the shared conditions of a situation, in turn, significantly affects the possible ways of being‐in‐the‐world of those persons and objects that get caught up. (Zigon 2015: 502)

Framing the drug war as a situation, then, allows us to view events on either a local scale or a global scale as a specific, if contingent, manifestation that the overarching situation, in a multi‐scalar sense, directly renders possible.

The U.S. opioid overdose epidemic can be conceived as a situation in precisely this sense. In Zigon's parlance: “In the current global configuration [of the drug war], complexity is at least as knotted nonlocally as it is locally, and this increasingly—so I contend—local complexity emerges within the shared conditions set by this diffused complexity” (2015: 502). Likewise, the face of the U.S. opioid overdose epidemic is hardly the same in rural regions as it is in urban centers. Yet these distinct local patterns are given shape by the same, shared macro‐conditions, such as historical shifts in pain treatment, the illicit drug market, and public discourse about opioids. Similarly, the features of the opioid overdose epidemic on a nationwide (or larger) scale and individual therapeutic encounters in the HIV clinic never fully overlap but are always enabled by and folded into each other.

Macro‐level public health surveillance technologies give shape to diverse clinical encounters by assigning clinicians responsibility for monitoring patient safety through the surveillance of individual patients. Indeed, though the primary outcomes of the TEACH study included two or more UDT per patient annually, use of the state PDMP, and the occurrence of early refills for opioid prescriptions, the study intervention explicitly targeted physicians, aiming to encourage them to enhance monitoring of patients’ bodies and behaviors. Further, just as indicators of risk in epidemiological data are meant to prompt population‐level interventions, current regulatory discourse holds that the identification of risk in a patient on LTOT (such as filling multiple prescriptions or producing unexpected UDT results) is supposed to trigger the physician to alter that patient's pain management strategy and, if indicated, withdraw LTOT completely.

Clinicians participating in this study reported feeling direct pressure from state regulators and hospital administrators to proactively monitor patient adherence—and manage patient risk—by employing these techniques. As Dr. Pole, a medical fellow, observed: “When you prescribe opioids, it requires an extra level that you have to be more—you need to control them: what are they doing; what are they taking; how are they taking.” Dr. Choate, who treats chronic pain in approximately 10% of her patients, described a total institutional shift: “[There have been] changes in the law, the legislature, and in the way the medicine was prescribed, and that it's a lot stricter and a lot more control is expected in patients who are receiving these medications.” Dr. Harris, slightly younger than Dr. Choate with a similar patient load, described feeling “explicit pressure” to discontinue LTOT. This unilateral directive left him with mixed feelings about his responsibilities. He observed, “pressure from like multiple levels of our system, starting with the governor's task force on opiate addiction, all the way down to my nurses, right? Like, who can we get off opiates, who can we get off opiates, which, I think that's appropriate to be constantly asking ourselves, but the right answer is ‘not everyone,’ right?”

Dr. Harris and Dr. Choate differed slightly in their views of how regulatory changes affected the care they could provide. Dr. Choate, ever the optimist, found some relief in mass media narratives:

Remember last year [2015] and two years ago [2014] opiate use was a lot more on the news as well. Lots of newspaper articles and lots of discussions. And so, it was, I think, a comfort. I mean, I was very happy about that because it's something that patients saw from another source. It wasn't coming from me, and so they saw that I wasn't just trying to limit their use or control their use better, control their use more. They could see like, not, this [new set of clinical practices] is, in truth, a new way of prescribing medications.

Dr. Harris, however, worried that each patient's unique needs were at risk of being sacrificed in service of efforts to control the opioid overdose epidemic at a population level. “Sometimes it's moralistic,” he mused.

There is a sense of like addiction is bad, and you should get people off it [opioids], because getting them off it is in and of itself a good thing. Which frankly isn't true for opiates … we'd be hard pressed to explicitly say how [my patient] benefits from stopping opiates right now. I'm not sure he does.

Thus, even as stricter protocols for monitoring opioid prescriptions become more pervasive in the clinic, Dr. Harris openly struggles to make these new monitoring strategies commensurate with other aspects of patient care: specifically, those subjective or relational forms of knowledge—when patient and physician appreciations must “read” and “be read” by each other—that form the bedrock of clinical assessment in chronic pain care.

The internal conflict described by Dr. Harris is also fueled by other, large‐scale situations in which this clinic is situated, like poverty, racism, and HIV. These physicians, like most people, were exposed to public narratives about opioid use and overdose. At the same time, their employment in a safety net hospital that serves a predominantly poor and Black population is likely to have increased their awareness of how race and class can have on quality of care. That social awareness may, ironically, encourage greater use of LTOT out of an interest in reducing racial disparities and/or encouraging economically vulnerable patients to stay in care (Knight et al. 2017). Further, the physicians in this study served as primary care providers for patients living with HIV—a disease with its own robust legacies of racial disparity (Parker et al. 2019), and which presents its own imperatives to keep patients in care (Colasanti et al. 2018). On the one hand, Dr. Harris's perspective on LTOT is shaped by a clear awareness of the role his safety net hospital plays for his patients. Yet, he also acknowledges that what is best for his patient (as determined by his interpersonal “read” of his patient's condition) may not line up with standardized best practices for opioid prescribing precisely because of who his patient is and, as well, how that patient is structurally positioned within the public evoked in public health.

More than this, though, the challenge Dr. Harris describes also indexes conflict between different ways a physician can see and generate knowledge about a patient in the context of the opioid crisis as a multi‐scalar situation. One mode of knowledge production consists of the accumulation of quantifiable, clinical facts: serum levels, range of motion, blood pressure, and so on. A different mode, subjective and relational, has been described by Jeannette Pols as the “enactment of appreciations” (2005). In her research on long‐term care for nonverbal patients, Pols observed that clinical staff perceived and attended to patient “appreciations” of their surroundings (as opposed to their perspective, which would require explication with words) through a variety of symbolic or enacted channels other than straightforward talk. Patients’ desires, feelings, and assessments of their own well‐being could become known through affect, behavior, bodily practice, and linguistic or physical performance.

Pols proposed conceptualizing these relational moments (in which a habitus is developed and performed to generate shared understanding) as situations—moments of interaction in which patients are regarded as independent subjects who possess appreciations of the care they receive that should be taken into account. Pol uses the term situation in much the same way as Zigon. It is necessarily composed of shared cultural, structural, or institutional characteristics that manifest as unique, context‐contingent interactions. For Pols, the interactive nature of the therapeutic encounter is essential, because this is what renders the patient a subject with a unique standpoint. As Pols observes, the patient–subject is “a co‐production, a result of interactions with others and a material world” (2005: 211). In other words, patients are regarded as subjects because they and their doctor collaborate through interaction to make it so.

Adding Pols's theoretical frame to Zigon's is helpful for conceptualizing these small‐scale manifestations of the opioid crisis in the clinic, especially in the ever‐evolving context of chronic pain care. The dominant tools for assessing the effectiveness of pain care are largely subjective. Clinicians may ask patients to describe their experience through pain scales and other clinical tools, but most tend to search for some sort of observable evidence in their interactions, asking themselves whether the patient seems to be in pain (Bergman et al. 2013). This objective evidence is often limited to the ways in which the patients are able to enact their pain: limping, touching body parts, vocalizing, swearing, demonstrating physical difficulties. The effectiveness of this enactment is determined by the physician's willingness and ability to take the patient's appreciations at face value, to interpret them as genuine and respond accordingly. As Dr. Seles observed of his patient with bilateral hip fractures, “I can tell when he's uncomfortable.” Thus, as Pols theorizes, assessing pain requires attending to the appreciations enacted by the patient through a clinically legible habitus of pain. In other words, it requires the physician and patient work together to establish the patient as a subject capable of experiencing and indicating pain within that therapeutic space. Further, working with economically marginalized patients, mostly patients of color, in a safety net hospital may add further urgency to the physician's ability to correctly read and act on those appreciations.

We argue that it is this very production of patient subjectivity in the clinic that is challenged by new standards of patient monitoring during LTOT. Indeed, some of the physicians in this study seemed haunted by the possibility that reliance on clinical monitoring rather than patient appreciations would de facto label their patient's pain as illegitimate. Did a failure to meet expectations, a troubling UDT result, or multiple requests for early refills necessarily mean that a patient is a “drug‐seeking trickster” (Crowley‐Matoka and True 2012)? Most physicians in this study would say, “No,” insisting that clinical indicators of medication adherence do not map fluidly onto patients’ lived experiences of pain (or anything else). Yet, the top‐down pressure to monitor patient adherence sometimes left physicians feeling obligated to act in ways that they knew might appear punitive or dismissive of patients’ own appreciations of pain.

After the study had concluded, several physicians in both study arms voiced frustration about the way in which opioid prescription monitoring practices could threaten the social and moral careers of their patients—even as they felt conflicted or unjustified in making individual exceptions when implementing those standards. Dr. Harris mused, “Some of my patients really do have pain. So—what, are they supposed to live in pain because they also have a substance use disorder?” Dr. Adams was also concerned by the ability of monitoring indicators to cast their patients as bad people, even if those indicators correctly mapped the problematic circulation of opioid medications into or beyond the patient's physical body:

Even though, like I said, most of my patients use their medications responsibly, anything can happen to anybody that all of a sudden, they're diverting their medication or they're drinking more or whatever. You know, they'll—something will come up that they need the money and so they divert the drug. And not for a foolish thing but because of a housing issue or something like that.

Physician and anthropologist Rudolf Virchow referred to physicians as “the natural attorney of the poor” (Brown and Fee 2006). These physicians adopted a similar stance, appearing unusually inclined to defend their patients against a world that already seemed against them. In the face of concrete evidence that a patient was using an opioid prescription inappropriately, many physicians in this study sought to protect their patient's moral personhood, even as they sought ways to decrease that patient's personal risk—or the risk that patient posed to the provider. In other words, they spoke out in defense of their patients as subjects even as they felt forced to make alternative clinical decisions that could discount the patient's subjectivity in the co‐production of their pain.

Disrupted Subjectivity

Arthur Kleinman once observed that “illness complaints are what patients and their families bring to the practitioner. … Disease, however, is what the practitioner creates in the re‐casting of illness in terms of theories of disorder” (1988: 5). Chronic pain management in this HIV clinic stands out, at least in part, as an exception to this rule in that patients had long been engaged in the co‐production of their pain as clinical fact. In acknowledgment of this long‐standing dynamic, most physicians made efforts to present new monitoring procedures to their patients as tools designed to operate beyond the range of individual appreciations. Dr. Lee, a young attending physician who manages chronic pain in more than half of his patients, recalled:

Well, I mean, there [are] some areas where the change in the state [laws] does hinder you. I mean, if somebody's had [traumatic injury] and is really hurt and needs drugs for a while, it's really hard to get them now. And the patient gets pissed off. … I said look, what can I tell you. Unfortunately, the laws are written for the majority of people. We'll do what we can to get you what you need, but you got to understand, this has caused—people are dying out there.

This move made sense from a perspective that framed the opioid overdose epidemic as a situation in Zigon's sense of the word: large, supra‐institutional forces restricting how physicians could act in local, interpersonal encounters. Yet, to be successful, this messaging required a shared acceptance that the exigencies of that large‐scale crisis could—or should—supersede how clinical habitus allows patient and provider to make meaning, coordinate realities, and choose a mutually agreeable course of action.

At the same time, many physicians felt a visceral tension between the demands of the opioid overdose epidemic as a large‐scale crisis and the demands of a different, overlapping one: the global HIV pandemic. Dr. Adams expressed explicit fears that disrupting LTOT could, in turn, disrupt HIV treatment, bringing this conflict to a head:

[HIV care] was already complicated before now, because HIV care is not just about the patient but also about everyone else the patient might be in position to transmit to. It's a public health concern as much as an individual health concern, and these two treatments in orbit with each other—the [HIV medication] that they “have” to take and the opioids they “want” to take can create some ethically fuzzy push and pull in clinical interactions.

Evidence from cross‐sectional surveys indicate that as many as half of all HIV physicians believe that LTOT facilitates patient engagement in HIV care (Tsui et al. 2019). Evidence also indicates that only a minority of HIV physicians were regularly using the monitoring tools made available through the TEACH study prior to the intervention. Therefore, regularly implementing these monitoring practices would constitute a marked change in clinical practice. As a result, physicians’ competing obligations to enact incompatible forms of doctoring (more opioids for HIV, fewer opioids for the overdose epidemic) forced the therapeutic interaction deep into the problem space of patient autonomy.

Physicians’ close attention to the effects of their doctoring on larger‐scale public health risks (overdose and HIV) was not mirrored by patients in their interviews for this study. Certainly, patients were capable of understanding the larger context in which changes in LTOT management were taking place. Yet, their priority was their own care. They were especially attuned to how the large‐scale situation (Zigon 2015) of the opioid overdose epidemic disrupted subjectivity and the enactment of appreciations in the socio–relational situation (Pols 2005) of their therapeutic encounters. For example, some patients perceived changes in care as a judgment against their value as partner–agents in their medical care. Mr. Jenkins, a middle‐aged man whose physician was in the control arm of the study, had been on LTOT longer than almost any other patient. He was astutely aware of how the subjective nature of pain could incentivize misdirection on the part of those seeking care. He felt that his own forthrightness about his pain had earned him—and should continue to earn him—his physician's trust. “I'm not trying to be overly—you know—I need more I need more,” he said. “What's your pain, it's a 12, the gauge stops at 10—it's a 13! None of that.” It stung, then, when that trust appeared to be withdrawn, replaced with UDTs, treatment contracts, and what he perceived to be an assumption of guilt by association with LTOT. “It [the monitoring] really starts when we put profit in front and looking at because somebody's an addict that they're a worthless piece of shit … [treated like] being an addict, yea. That's exactly how I feel.”

Mr. Jenkins's feelings were widely shared. Patients consistently expressed that the perceived loss of opportunity to communicate their pain experiences through appreciations, and, in turn, to be interpellated as self‐aware, autonomous individuals, felt like a violation of the social contract they held with their physicians. Mr. Thomas, a man in his early 50s who has been on LTOT for nearly a decade, was taken aback when his physician, Dr. Seles, suggested ceasing his opioid prescriptions because a recent UDT was positive for cocaine:

At first [my doctor's] coming like, “You're getting off the pain medicine.” Aren't you a pain specialist? You're supposed to prescribe pain medicine, but he's … “you know you got to be ‘clean’ [not using illicit drugs] to get, you know, Percocet.” And I got upset. “Why?” You know. “Because I get high? That got nothing to do with my pain.”

Mr. Thomas perceived that the quality of his pain care, which his doctor had apparently deemed satisfactory for years, was forfeited in favor of standardized prescribing policies. It felt as though he, an individual present in the social situation (Pols 2005) of their therapeutic encounter, had been forced to make sacrifices on behalf of an abstract public conjured up by the framing of the opioid crisis as a large‐scale public health situation manifesting in the clinic (Zigon 2015). Mr. Thomas recalled, “And I told the doctor, ‘If you take me off the pain medicine, what's going to happen? I'm just going to get more higher.’” Later, he mused, “[The NCM] was just here to get me off the Percocets, and then she left. I was the first to get shut off.” Similar to Dr. Harris, he heard the message loud and clear: “Get folks off opioids wherever you can.”

Dr. Seles, well aware that patients were taking these changes personally, took steps to minimize the damage. His solution was to allow the demands of the opioid overdose epidemic as a public health issue to universally dominate his clinical decisions:

[The NCM] puts [patients] into high and low risk people and thought [some] didn't need [UDTs] every time. I thought, you know what? Let's just make this a standard so I don't have to make a judgment which category you're in. It's your expectations, you're not being singled out, because many of the patients were very resentful of having to do that. Felt like they were—had been good citizens and were now being treated with suspicion like they were addicts.

In other words, Dr. Seles sought to reduce how severely his patients felt their appreciations of pain were being dismissed by eliminating the role of those appreciations altogether.

Though adopted in good faith, Dr. Seles was unable to produce his desired result. Instead, he brought into sharper relief how inextricable the large‐scale and small‐scale situations—the local and the global, the macro and the personal—really are. Mr. Thomas was also aware that these changes were being standardized, noted with contempt: “Everyone on pain medication's going through it.” Rather than feeling singled out, he felt the whole situation was wrong. Pain care strategies that had been standard and (in his view) practiced without incident for so long were being rolled back as a result of the opioid overdose epidemic, causing new and unnecessary burden to patients for the sake of a larger public health concern in which they were not necessarily involved. “Everybody got to [lower their dose],” he said, “and you got to get on every week now. You don't get a month supply [of opioids], you get it weekly, so you had to come [to the clinic] every week.” Notwithstanding his best intentions, Dr. Seles observed after the study had concluded, “I think it harmed the relationship, yeah.”

Despite all these challenges, and despite the absence of any clear resolution for most of them, most physicians still expressed gratitude for the opportunity to be involved in the study and to learn more about pain management and opioid prescribing strategies. Dr. Lee observed: “Let's face it, you've got this humungous problem out there. And a lot of your patients have no idea what they're taking, they don't know. So, the more you learn and the more signs you could pick up and then discuss with them, I mean, it's got to be better.” Though this study offers reason to suspect that the multi‐scalar situation (Zigon 2015) of the opioid overdose epidemic and the social situations (Pols 2005) of clinical pain care are intractably conflicted, the physicians remained grounded in the practical reality that some sort of doctoring would always be required. Good or bad, they had to make some kind of choice—and continually be choosing—in their roles as health care providers. Dr. Lee concluded, “The whole idea is to take as good care of these people as you can and try and keep them as safe as you can, so you're an idiot not to be open to [learning] anything.”

Conclusion

Recent changes in professional best practices for opioid prescribing have shifted the ontological terrain of chronic pain care. Authoritative guidelines recommend that patients be assessed through an assemblage of concrete indicators, designed in response to the opioid overdose epidemic as a global situation (Zigon 2015), mitigating the value of their assessment as subjects with appreciations (Pols 2005) that should be taken into account. A primary concern for HIV physicians is not simply the depersonalized nature of the new standards for patient monitoring in the context of LTOT, but the seemingly significant opportunity costs of implementing those standards in therapeutic relationships that have historically been governed by socially contingent knowledge (produced through enactment and interpellation) about patient pain and the appropriateness of analgesic medications.

Previous ethnographic scholarship on the licit and illicit circulation of opioids in the United States has suggested that “policy debates become part of the problem by shifting political issues into a technocratic register” (Bourgois and Schonberg 2009: 297). In response, we have posed an alternative question: What kind of therapeutic praxis is emerging from subtle, diffuse forms of resistance to or acceptance of those technocratic forms? The data presented here reveal chronic pain care to be a therapeutic problem space in which patient subjectivity is not only protected from the objectifying powers of the medical gaze but remains foundational to the co‐production of clinically legible pain in the first place. The disruption of that subjectivity through the implementation of clinical audit technologies thus proves distressing for both the patient and the physician. It also serves as a generative medium for new therapeutic and interpersonal strategies—for new kinds of doctoring in these scalar situations. Emergent strategies observed in this study included: homogenizing monitoring practices across patients; stalwartly defending the value of LTOT in relation to the risks of ceasing that prescription; and ceasing care of pain patients altogether, transferring them to addiction medicine specialists instead. As public health responses to the opioid overdose epidemic continue to scale up, and as new discourses of opioid risk emerge from this cascade of regulatory change, patients and providers are likely to generate even more novel strategies for navigating that biopolitical field. This analysis suggests that a clearer understanding of this new praxis requires close attention to the symbolic work already at the center of therapeutic encounters.

In light of these findings, we wonder if stricter implementation and enforcement of monitoring practices (whether designed to monitor patients or the clinicians serving those patients) will produce the desired reduction in prescription drug misuse and diversion if they cannot also account for the inviolable meaning of patient appreciations (Pols 2005) in the clinical encounter. Perhaps new strategies to assist clinicians in navigating the morally ambiguous terrain of pain management in the HIV clinic—where patient subjectivity, appreciations of care, and clinical indicators of adherence to prescribed treatment regimens may all contradict one another—are what's urgently needed. We find Dr. Adams's observation that LTOT can create an “ethically fuzzy push and pull” between patient and physician in the HIV clinic is particularly prescient here. It may be impossible to reduce the risks of opioid medications without first resolving the ethically fuzzy push and pull that the particular histories of pain care and HIV care have helped create. We, as a research team, must admit that we have no immediate remedies for this tension at hand; however, simply recognizing this tension—and naming it for what it is—may be an important first step in navigating this tricky terrain of HIV, LTOT, and opioid risk in a way that is fairer and offers more beneficence to all those involved.

All authors declare that they have no conflicts of interest.

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in this study.