The intersectional implications of a quantitative epistemology in pain care and research.

In his 2019 ethnography, 32 Lars Johannessen provided insight into the ways in which pain practitioners—in this case, triage nurses in an Oslo hospital—prioritize their own observations (what the nurses deemed “objective”) over the assertions of their patients (“subjective”). Johannessen noted only five instances (out of 342 assessments) wherein the nurse evaluating the patient asked for a self-report (a score of 0 to 10, with the former being no pain and the latter being the worst): “Nurses commonly claimed that they had stopped using the ‘subjective’ [numerical rating scale value] unless they considered the ‘objective’ score ambiguous or inconclusive.” 32 Johannessen went on to argue that the nurses’ reliance on the objective approach could be attributed to their belief that it is “the most valid, reliable and efficient proxy for scoring patients’ pain.” 32 In this section, we intend to demonstrate how validity, reliability, and efficiency—all effectively statistically substantiated qualities—have come to embody the ideal pain measurement instrument, ultimately implicating both clinician/researcher and patient/participant in a system that privileges quantitative data over qualitative—the aforementioned epistemological entrenchment. In the late 19th century, physiologists were trying to determine whether pain was a sensory modality, served by its own neural apparatus, or an attribute of other sensations, like the stimulation applied to pressure preceptors. 33 This latter theory led physiologists in Sweden, Germany, and the United States to map out sensory qualities elicited by stimulation on the surface of the body. Sense-meters through algometric measurement happened by using a device (the von Frey hair, or the dolorimeter) to apply measurable stimuli to trained participants . In the case of the dolorimeter, designed in 1940, a 1000-W bulb projected a strong beam of light onto a small area of a person’s forehead. Researchers found that it worked best when a small number of reliable participants who had been trained were used to quantify the intensity of felt pain. The researcher of pain was placed squarely in control of the evaluation. Henry Beecher complicated the power of the researcher, however, when he argued that a patient’s experience of pain was highly subjective—to the point that the small- n dolorimeter study was wholly inadequate. In his landmark 1959 work, Measurement of Subjective Responses: Quantitative Effects of Drugs , 34 the anesthesiologist and army veteran changed the landscape of pain research by pointing out how a patient’s experience of pain is predicated on a wide swath of variables. In his outline for chapter 8, “Factors Said to Produce Variation in the Pain Threshold Other Than Analgesics,” he lists: Race—Sex—Aging—Autonomic nervous system—Circulatory change—Skin temperature—Sweating—Elevation of carbon dioxide tension—Hyperalgesia—Other forms of trauma—Nausea—Fatigue—Anxiety and fear—Training (man)—Training (animals)—Bias—Adaptation—Distraction, inattention, lethargy—judgment impaired by drugs—Suggestion and emotion—Warmth and cold—Multiple stimuli and extinction phenomena—Placebos—Diurnal variation—The passage of time—Miscellaneous factors—Lowered pain threshold. 34 Race—Sex—Aging—Autonomic nervous system—Circulatory change—Skin temperature—Sweating—Elevation of carbon dioxide tension—Hyperalgesia—Other forms of trauma—Nausea—Fatigue—Anxiety and fear—Training (man)—Training (animals)—Bias—Adaptation—Distraction, inattention, lethargy—judgment impaired by drugs—Suggestion and emotion—Warmth and cold—Multiple stimuli and extinction phenomena—Placebos—Diurnal variation—The passage of time—Miscellaneous factors—Lowered pain threshold. 34 As Tousignant 33 pointed out, Beecher’s method required an overhaul not just in conceptualization of pain research but in measurement logistics. “For all Beecher’s talk of simplicity, the analgesic clinical trial was more expensive, time-consuming, and difficult to coordinate than the dolorimetric method,” 33 she wrote, adding that “statistical experts” 33 were required to be added to budgets. Nine years before Measurement was published, Beecher and his colleagues Arthur Keats and Frederick Mosteller 35 acknowledged that patient self-reporting was not something readily trusted, noting that “subjective data” could be doubted for its reliability, thus turning researchers back “to the study of pain experimentally produced and measured” (i.e., the dolorimetric). They suggested, however, that “adequate controls” needed to be put in place in data collection to reduce a patient’s subjectivity’s irrational effect on results and spent the bulk of their paper explaining the statistical mechanisms one might utilize to do so: primarily, adjusting control variables and implementing weighting systems. These controls—and their underlying probabilistic machinations—work together as a stand-in for the dolorimeter’s dials, ultimately reintroducing the researcher’s expertise. Reviewing the introduction and evolution of some of today’s most popular pain measurement instruments, it is clear that using statistics to prove a tool’s validity and reliability—and, to a great extent, efficiency—is a prerequisite to its widespread use. When Bond and Pilowsky 36 introduced the VAS into pain studies in 1966, for instance, they specifically cited Clark and Spear’s 37 1964 assessment of the VAS’s reliability. As a direct descendant of the graphic rating scale, the VAS could also boast over 4 decades’ worth of testing prior to Bond and Pilowsky’s innovation: in his 1922 introduction and analysis, Paterson bragged that the graphic rating scale’s design “eliminates the restrictions on natural judgements which other ratings methods impose.” 38 Further, as Schaffzin 39 previously discussed, the tool itself was developed with rater instinct and overall efficiency in mind. Another exemplary case study is that of the face-based pain scale, a graphic instrument for measurement that came to fruition in the latter part of the 20th century and was primarily developed in the context of the pediatric pain clinic. Arguing that there existed little research into the validity and reliability of children’s pain scales at the time, Wong and Baker, 40 eponymous originators of the widely used and recognized Wong-Baker FACES Scale, tested six pediatric scales in 1988. They cited Judith Beyer (eventual publisher of the seldom used photographic “Oucher Scale” 41 ), who, along with Thomas Knapp, 42 argued that “just because subjects are able to provide scores or make a mark on some pain intensity scale does not necessarily mean that the scale validly measures pain intensity.” 42 As such, Wong and Baker published their own efforts to statistically endorse the tools’ varied effectiveness. Considering the historical context in which the face-based pain scale proliferated, however, one might deem our third metric, efficiency, to be just as important as—or perhaps even more important than—reliability and validity. As part of the Reagan-Thatcher regime of neoliberalism, the last 20 years of the 20th century also featured the rise of the health maintenance organization in the United States, a byproduct of the widespread capitalization of health care. Concurrently, second-wave feminism encouraged women in the workplace to seek out more authority, and nurses—primarily a position held by women—were no exception. 43 So, when the expedient discharge of patients meant higher margins for hospitals, and nurses were put in charge of evaluating the severity of a patient’s condition, those nurses developed evaluative tools that were as fast to use as they were reliable. 44 Thus, efficiency was enacted through a scale that, for Wong and Baker, at least, provided “a quicker way” to determine a patient’s pain, partly thanks to the ability to abbreviate instructions. 45 Reliability, validity, and efficiency, then, have been and continue to constitute the ideal of the tried-and-true measurement instrument—exemplified here in only two case studies from the post-Beecher era. If the researcher working with dolorimetrics held the privilege of controlling the participant’s pain, the Beecherian researcher relies on the authority of the statistical, a phenomenon summed up (though extensively elaborated upon) by Theodore Porter in his 1995 Trust in Numbers : “reliance on numbers and quantitative manipulation minimizes the need for intimate knowledge and personal trust.” 46 So often, pain evaluation techniques and technologies are promoted using the language of purported objectivity, offering that this time , researchers and caretakers can know how much pain a patient is in with little to no worry of the individual embellishing. 47–49 Returning to Johannessen’s nurses, 32 and considering the intense pressures placed on them to evaluate their patients’ pain quickly and accurately, it is no wonder.

In May 2023, researchers, primarily from the University of California San Francisco, published a paper in Nature Neuroscience wherein they asserted, “The development of personalized pain biomarkers will be central to accurate diagnosis, tracking prognosis and for future therapeutic drug and device development.” 69 Building on decades of neurological technologies, the authors used machine learning models to interpret data from neural implants to “successfully [predict] intraindividual chronic pain severity scores from neural activity with high sensitivity.” 69 The study only involved four participants, but it made quite the splash nonetheless, receiving write-ups in the New York Times , 70 the Wall Street Journal , 71 and MIT Technology Review , 72 among others. In the Times , Priyanka Runwal spoke with a neuroscientist, who explained, “In addition to advancing our understanding of what neural mechanisms underlie the pain … such markers can help validate the pain experienced by some patients that is not fully appreciated—or is even outright ignored—by their doctors.” 70 By turning specifically toward efforts to locate pain in the brain, we can begin to recognize what Anne Beaulieu, 73 citing Dumit (2001), 74 wrote is “a culturally based longing for insight into what subtends our personhood.” 73 Beaulieu’s study used interviews with both clinicians and imagers (i.e., those who operate functional magnetic resonance imaging [fMRI] and positron emission tomography [PET] machines) to understand the relationship between the data-based (quantitative, digital) brain and the map-based (graphic, representational) one. Ultimately, by interrogating the graphic output of the fMRI or PET, Beaulieu highlights the critical nature of the translation that must occur between technological sensing and the resulting output used for sense-making. Building on Beaulieu, Melissa Littlefield delved into the electroencephalogram (EEG) and its “brain wave ideologies,” 75 unpacking the ways in which expertise was inscribed onto the operators and physicians who used these machines to garner “machine-mediated seeing.” 75 What neither Beaulieu nor even Littlefield could anticipate, however, was a new sort of middleman added between sensing and sense-making: the large language model to feed an artificial intelligence system. Machine learning and artificial intelligence systems are the ultimate generalizers. After being fed large—sometimes massive—collections of data, statistical regressions are performed to identify relevant patterns. In the same way that a natural language model like generative pretrained transformer predicts sequences of words based on previously analyzed prose, 76 a neurological pain model tries to predict measured pain scores based on brain wave patterns that had previously been extrapolated from patient data. So, whereas the statistically valid and reliable measurement instruments considered above necessitated some sort of interaction with the patient (even in the case of the Oslovian nurses, who had to observe the patient face-to-face to make an assessment), the dream of this new crop of neurological technology negates even that requirement. Lopez-Martinez et al., 77 for instance, suggested that an EEG reading of a patient under anesthetic might allow a physician to administer more agent. 78 Further, the authors suggeste, the technology may “help advance the development of automatic analgesia administration systems for hospital settings that currently rely on subjective self-reported pain measures.” 77 Efficient, indeed. 2024 marked 100 years since Hans Berger made the first recording of human brain activity with an EEG. Since then, the EEG, PET, fMRI, and, more recently, advances in neuroprostheses (such as those described by Shirvalkar et al. 69 ) have been employed to situate pain in the brain, effectively removing patient subjectivity from the equation in the name of objectively identifying whether, how much, and where a patient might hurt. Positioning this technology as somehow separate from the patient risks further masking a pain treatment regime already fraught with racial and gender-based inequalities, 79–81 exacerbated further by what Crowley-Matoka et al. identified as a “mind/body dualism.” 82 They elucidated a “scientific tendency to categorize and compartmentalize” wherein privilege is given “to direct observation and hence ‘objective’ over that which is unobservable and hence ‘subjective.’” 82 Here, then, we can easily apply Mol’s project 20 : neuroimaging technology enacts pain for the practitioner—makes it real via data, separate from (though, at times, in concert with) the patient’s own claims. Though we acknowledge that pain research and clinical practice are different activities, they influence each other in practice. Critical data scholars have highlighted concerns around the development of algorithmic tools. 83–87 These studies identified disparities in representation. Meredith Broussard wrote on the topic that “it’s not a glitch that Black, Indigenous, and people of colour (BIPOC) voices are marginalized when algorithms mediate online discourse; it’s a feature of real-life power structures replicating online.” 88 One danger of applying data sets and language models to understand and identify pain is that racism and sexism are prominent within data sets. 89 , 90 It is tempting to suggest that the data sets can be adjusted to be more representative of minority populations, 59, 91 but the answer is not necessarily to collect more data. As Williams noted, “There is a long history of measuring Black bodies, turning those measurements into data, and then building systems of values, beliefs, and predictions off of that data.” 92 Such practices are reminiscent of craniometry and phrenology, measurement practices conducted in service to the foregone conclusion that the White male European form was standard and superior to all others. More data are not always the ethical path forward. Further, Kaiser pointed out that a world designed around what he called the computable subject “caters to the neoliberal governmentality and enables granular biopolitical control of individuals.” 93 As others before us have noted, the failure to address cultural variations in pain expression can lead to a misinterpretation of patients’ pain reports. 94 For this reason, those working in the field of pain studies might consider how to challenge the aforementioned tools and create culturally sensitive assessment tools. The proliferation of statistical mechanisms pigeonholes this landscape, which leads to a hermeneutical lacuna, a lack of vocabulary (in the broadest sense) with which to express and understand the multiple pains of this world. Hermeneutical injustice 16 occurs when someone is unable to express, and thereby understand, an important aspect of their experience. To illustrate this concept, Fricker drew on one woman’s account in the 1960s of being given the opportunity to discuss, in a small group of other women, postpartum depression. 16 Not having had the dialogical space to map these feelings and experiences previously, her experience entailed a hermeneutical breakthrough. This leads to a kind of epistemic justice wherein a person steps into a fairer world, one where their experiences have equal opportunity to take form and be communicated. Fricker wrote of how, in hermeneutical injustice, “extant collective hermeneutical resources can have a lacuna where the name of a distinctive social experience should be” and relations of unequal power can “skew shared hermeneutical resources.” 16 In these cases, individuals and groups, including biosocial communities, lack the appropriate means to make sense of their experiences. Such injustice can be avoided when multimodal assessments and dialogic environments are fostered in the pain care environment (see Charette13 for an example). 2 2 Charette, M. 2024. “‘Play!’: Combatting Pathocentric Epistemic Injustice in Chronic Pain Care.” Qualitative Health Research 10497323241300437. doi: 10.1177/10497323241300437. Charette, M. 2024. “‘Play!’: Combatting Pathocentric Epistemic Injustice in Chronic Pain Care.” Qualitative Health Research 10497323241300437. doi: 10.1177/10497323241300437.

There has been a growing body of literature questioning the use of clinical tools for measuring aspects of pain that are increasingly seen as being at odds with what is termed the lived experience of patients. 1 Yet in these debates, the production of pain tools themselves is rarely questioned or explored. This article takes a systemic approach to understanding this development, drawing on science and technology studies (STS), a field predicated on the social scientific and humanist exploration into and around the so-called hard sciences. Increasingly, STS scholars have begun to explore data–human mediations and what it is like to live with quantitative data. 2–6 This work articulates how data require attributing specific weights to parts of human experience. Ruckenstein 7 suggested that charting data and seeing “trends” can lead to important new questions about our selves. Data can be used to form new social identities 8 or cope with evasive and poorly understood phenomena, including conditions like bipolar disorder. 9 On the other hand, data are often incomplete, because the ordinary self-tracker might forget to do so consistently. 10 Drawing on ethnographic fieldwork on neuroprosthetic development, Alexandra Middleton highlighted how the experience of phantom limb pain is hard to describe without metaphors and “sensory articulations.” 11 Writing about women’s pain and endometriosis, Emma Whelan quoted a patient, Andrea, who says that numbers are “empty of any meaning.” 12 Yet Andrea must operationalize them in her own metrics because she is aware of their power “to appear neutral.” Drawing on in-depth interviews with patients who use data to track their pain, Charette 13 highlighted the ambivalence that users feel about the pressure to measure their pain using mobile health applications, table-side notes, and Excel spreadsheets. This ambivalence is experienced by clinicians as well, who are aware of the fact that counting pain counts in the medical encounter, even if they are skeptical or critical of this fact. 14 Some clinicians have suggested that numerical scales only be used conservatively as a labeling mechanism, not as outcome measures or predictors. 15 In short, quantification is seldom a straightforwardly negative or positive transformation. These testimonies should make us question the extent to which such measurement practices constitute a mode of empowerment, because they do not originate from the user. They are accepted, not chosen, as a feature of contemporary pain care. In this article, we take a systemic and unique approach to highlighting the decades-long foundation upon which the tools most in use today have been established. In doing so, we hope to contribute to an already existing discourse around the benefits and detriments of quantitative pain assessment. Specifically, we bring a historical lens to this work that seeks to contextualize the day-to-day goings-on of a pain practitioner or researcher within technological and sociopolitical theories. We begin with a brief historical overview of pain scale development and substantiation. Then we highlight a persistent emphasis on statistically certified tools that may risk marginalizing the clinical patient or study participant. Third, we provide three examples of pain experience that would not be coherently translated by the tools highlighted before concluding with a discussion of how this is being addressed by a mandate to collect more data and why such a solution is unlikely to succeed and is likely to result in further hermeneutical injustice 16 for patients. Ultimately, we believe that bringing this perspective to the readership of this journal will help make the field of pain medicine more equitable, even if the changes we propose may be an extremely small and/or difficult part of that evolution.

Pain measurement as praxis depends on having an audience that appreciates the coupling between the thing (here, pain) and its referent (the number 10). These couplings are not universal. Race, gender, culture, age, and other social determinants of health impact the practice of expressing pain. 32 , 50–52 In this section, we discuss three contexts in which pain expression does not fit neatly within the boundaries set by the statistical mechanisms outlined above. If such a mode of measurement makes possible understanding the patient or participant in pain, researchers must pay heed to, and find ways to accommodate, pain as ontologically multiple. Although there are countless examples that could be used to demonstrate this point, ours cut across multiple social determinants of health including language, culture, gender, age, and race. One’s cultural upbringing significantly influences how pain is understood and expressed. Expressive differences can refer to facial expressions, body language, and words or utterances. To illustrate this point, we turn to the work of anthropologist Jason Throop in his 2010 ethnography Suffering and Sentiment , 53 which explores “strategies of concealment” among the Yapese who live on the island of Yap in Micronesia. Cultural values have a direct impact on how pain is experienced and expressed in Yap. So too does the landscape and the metaphors drawn from it. For example, the Yapese term galuuf refers to a species of monitor lizard and is also used to denote pain associated with muscle cramps, echoing the phenomena when the lizard’s muscles are paralyzed from the cold. In Yap, Throop writes, knowledge is understood to be largely a private possession. 53 In part due to the historical roots of the islands, the moral framework of Yap is one in which individual expressivity and inner life are disconnected. Individuals are expected to gain control over the disclosure of their emotions. Such a value is manifest in the expression “think before you speak.” However, in Yap, this injunction applies not only to one’s beliefs and opinions but also to their emotions and correlated embodied expressions: “Closely tied to the valuation of deliberate activity is a prevalent emphasis placed upon the significance of opacity and nonexpressivity in relation to an individual’s emotions, feelings, opinions, thought-objects, intentions, and the like.” 53 Throop carefully traced these moral frameworks across Yapese linguistic practices. For example, the concept puuf rogon (“freedom” or “free will”) refers to a way of speaking or acting that accords with one’s personal desires, wants, or inclinations, whereas anthamagil (“endurance” or “striving”) refers to gaining control over ones’ inner states, thoughts, and emotions. 53 Studying pain expression in Yap, Throop noted that patients described their pain as an object that bears little to no relation to the sufferer. That is, they did not necessarily identify their pain as theirs but instead stated “ baaq amiith ” (“there exists pain”) or “ kab ae amiith ngoog ” (“pain came to me”). As Throop pointed out, this objectifies pain without delineating its location or relationship to the sufferer. 53 A similarly distinct case was demonstrated by Lor and colleagues 54 in their work mapping modes of expressing pain in the Hmong population. They focused on the Hmong for three reasons: first, there are many living in the United States, of whom over 85% speak Hmong at home. Further, over 90% of older Hmong individuals have limited English proficiency. 54 Second, the Hmong are a primarily oral society. Because the written language was developed less than one hundred years ago, many Hmong do not know how to read or write in Hmong. Third, there are distinct words in the Hmong language that cannot be easily translated to English, and they also have no discreet word for “pain.” Their qualitative study showed that pain was expressed, and thus enacted, as a narrative. All older Hmong adults responded to being asked to describe their pain by telling a story. Interestingly, many elements of the stories told included information that statistical mechanisms/Western biomedical practitioners collect during pain assessments (location of pain, time of pain, quality and intensity, causal attributions, etc.). However, like the Yapese monitor lizard metaphor, the Hmong have their own culturally salient metaphors, such as using the imagery of a chicken pecking to describe the quality of pain. 54 Lor and colleagues ultimately suggested that using a story approach may be the most effective way of eliciting pain information from Hmong patients. In their institutional ethnography, Rice et al. 55 noted that gender is undertheorized in chronic pain research. For one, pain arises out of different kinds of activities and forms of labor. Men, for instance, are overrepresented in workplace accidents and injuries. 56 Indigenous and multiracial individuals experience the highest pain prevalence. 55 , 58 In their attempts to account for these differences, some pain researchers conduct empirical and psychophysical investigations. 59 , 60 Understanding the communities most affected by chronic pain requires examining the social and structural determinants of health; that is, the nonmedical factors that contribute to health and illness. Social and economic conditions and inequalities in access to resources and services have been shown to have a greater impact on well-being than health behaviors. 61 Loss of traditional livelihoods, “whole foods” nutrition—in short, resource alienation—are all contributors to chronic illness rates in Indigenous communities. 57 , 62 , 63 Qualitative studies conducted over the past 30 years have shown that women and men tend to express pain in different ways. 56 , 64 , 65 These linguistic and expressive tendencies can be a product of cultural norms. In their overview of research on gender norms in pain treatment, Samulowitz and colleagues 66 showed how women are often regarded as more sensitive than men, whereas men are viewed as brave and strong. Women often must manage their pain and the demands of their environment simultaneously. They often continue to perform duties around the house while experiencing pain, whereas men more often “handed these duties over to their spouses” while in pain. 66 These ideas about expression can be problematic and can even lead to iatrogenic pain; that is, pain caused by an error made by medical professionals. Consider, for example, the story of Malika Bilal’s pregnant sister-in-law, who, despite telling nurses that her blood pressure was skyrocketing and that she had a pounding headache, was ignored and left partially paralyzed by a stroke. 67 Anushay Hossain’s 2020 The Pain Gap offers chilling, and unfortunately copious, real-world examples that illustrate that “Black, Native American, and Alaska Native women are two to three times more likely to die from pregnancy-related causes than white women.” 67 There are ways to communicate pain that, though perhaps not as statistically legible or efficient as more commonplace scales (at least in the Western context), are still meaningful and informative. Continuing to dismiss these approaches based on their misfitting 68 (a la Rosemary Garland-Thomson 1 1 Here, we acknowledge that Garland-Thomson’s original conceptualization of the term “misfitting” in the context of health and disability refers to the lived experiences of disabled individuals rather than in reference to an inanimate object or tool such as a pain scale. However, given that the patient who is accustomed to using a “non-traditional” scale is often forced to adapt to more normative tools in the name of fitting the rater’s needs and expertise, the term feels apt. ) risks the further proliferation of pain measurement technologies that conceptualize the patient or participant as purely computable, a mass of data points that can inform treatment without the vagaries of discourse. And so, with a reminder of Porter’s assertion that a reliance on the statistical abates the necessity for intimate knowledge, we now highlight a number of pain evaluation technologies that have been recently proposed by researchers in the United States. Here, we acknowledge that Garland-Thomson’s original conceptualization of the term “misfitting” in the context of health and disability refers to the lived experiences of disabled individuals rather than in reference to an inanimate object or tool such as a pain scale. However, given that the patient who is accustomed to using a “non-traditional” scale is often forced to adapt to more normative tools in the name of fitting the rater’s needs and expertise, the term feels apt.

STS scholars examine the production and legitimization of scientific knowledge with a focus on how social, economic, and political elements structure those processes. Many have made significant contributions to the social studies of health sciences by examining how developments marked as “progress” are not linear, nor are they neutral. 17–23 STS research has shone a light on how scientific professionals navigate personal and institutional biases and how these biases are shaped by the political economy that underpins innovative practice. In short, STS scholars set out to highlight subtle and taken-for-granted aspects of scientific medicine. Within STS, there is a subtype of studies that foreground the materiality of scientific work. These studies explore how scientific ideas (diseases, diagnostic categories, ethical frameworks) are enacted and legitimized via the tools that make them possible. This way of thinking is inspired by STS scholar Annemarie Mol’s seminal work in The Body Multiple: Ontology in Medical Practice , 20 where Mol offers a way of studying scientific medicine that pays heightened attention to how materials factor in translational practice. Praxiographic studies, arising from the Latin praxis , or “practice,” involve close-up, detail-oriented examinations of medical practice. Praxiographic accounts highlight how materials highlight certain features of the world at the expense of others. Crucially, these material practices give rise to distinct forms of knowledge. We might say, then, that praxiographic accounts examine the interplay between scientific action and epistemological communities. Just as a pathologist discerns disease through a microscope, with a pointer, two glass sheets that make the slide, the decalcification that allows the technician to cut thin cross sections of vessels, tweezers, knives, dyes, etc., a pain specialist or nurse practitioner discerns pain with the assistance of the visual analogue scale (VAS), numeric rating scale, and/or other quantitative tools. Of course, this comparison is not perfectly symmetrical. For example, depending on the subtype, pathologists typically rely on a single material process as a privileged means to accessing a biological specimen. Unlike pain specialists, pathologists seldom interact with patients. Unsurprisingly, however, data figure no less prominently on the horizon of pathology, because data-fed tools are being used in pathology from training, diagnostics (using digital image analysis, for example), and cross-institutional consultations. 24 Pain medicine as a subspecialty was only recognized in Canada in 2010. During training and education, pain specialists are taught to become proficient in communication and pain assessment tools. 25 In meeting with a patient, they rely on testimony, dialogue, physical examinations, and numerical measures to partially approximate the pain experience. But not all pain specialists take a dialogue or patient-centered approach—many prioritize intervention in the form of nerve blocks, for which there is a growing private industry in Ontario. Multimodal pain treatment is offered in multidisciplinary pain treatment facilities across Canada, but these are concentrated in urban cities and involve lengthy wait times. 26 Studies also show that multidisciplinary pain treatment facilities exclude patients experiencing migraines, fibromyalgia, as well as those who have mental health or substance use disorders. 27 As a result, not all people experiencing pain receive pain care from pain specialists. Within and beyond pain medicine as a field, more work is needed to address the pain epidemic. 28 , 29 Despite slowly increasing numbers of pain specialists, there were only 24 pain specialists in all of Ontario in 2020. 30 So, though we acknowledge that there are well-trained pain specialists who use more than one tool to understand pain, here we focus on settings (triage in emergency care, family medicine, interventionist pain care) wherein the numerical measure remains a prioritized tool due to resource (fiscal and temporal) constraints. Numerical measurement tools help build epistemic communities whereby pain is enacted as a number or intensity. As data-oriented tools become increasingly popular, our point here is that we must preserve an awareness of the need for greater access to multimodal approaches. Numerical measurement in pain bears a resemblance to the stethoscope, which allowed doctors to “use their ears to ‘see’ inside the body, especially the heart and lungs,” at the cost of patients’ stories. 31 Moreover, these enactments are themselves justified via the same tools used to bring pain to foreground. There is a circularity here, a kind of material and epistemological entrenchment. This is not to suggest that new habits and methods cannot be formed to intervene on the well-worn grooves of praxis. But it is only by virtue of tending and understanding the material practices of pain medicine that such can interventions be made and epistemic communities opened up.

To theorize his case study, Johannessen 32 employed Espeland and Stevens’s 95 , 96 commensuration , wherein a common quantitative metric is utilized to compare differences that may otherwise be qualitative in nature. In summarizing part of their argument, he wrote, “We should question the work and conventions that underpin the production of seemingly neutral numbers.” 32 Pain rating—turning a wholly qualitative phenomenon into the quantitative—is a practice consistent with the biomedical field’s emphasis on validity, reliability, and efficiency, especially in the name of higher patient throughput and, consequently, profit. There are alternatives, however, that recenter the patient, even if they eschew the ontological boundaries of statistically sound practices. A narrow focus on data-based pain measurement technologies, however, risks further entrenching both patient/participant and practitioner/researcher in systems of computation and control. We are not arguing here that pain should not be quantified—this would entail a complete overhaul of the entire biomedical complex. Rather, we seek to influence three constituencies with our research project. Firstly, we wish for the patient in pain to understand the historical and social contexts in which the tools that govern their health care were developed and disseminated, perhaps empowering them to advocate for alternative means of communicating their experiences. These could include narrative, 97 arts-based, 98 or multimodal models. 99 Secondly, we encourage social scientists who are concerned with the care offered to patients experiencing pain to study the “regimes of practice” 23 of pain science. And finally, we appeal to practitioners and researchers for a willingness to seek out opportunities within their daily interactions with patients and study participants for pain evaluation techniques that go beyond the quantitative. The habits of measuring pain bear implications for what (whose) pain is observed and understood and what (whose) pain is not. The statistical data set sidesteps, or renders obsolete, the narrative enactment of pain, which may be the only appropriate enactment in certain contexts.