A Framework for Understanding Medical Epistemologies

George Khushf

In decision-making concerning the diagnosis and treatment of patients, doctors have a responsibility to do this to the best of their abilities. Yet, we argue that the current paradigm for best medical practice - Evidence-Based Medicine (EBM) – does not always support this responsibility. EBM was developed to promote a more scientific approach to the practice of medicine. This includes the use of randomized controlled trials (RCTs) in the testing of new treatments and prophylactics and rule-based reasoning in clinical decision-making. But critics of EBM claim that such a scientific approach does not always work in the clinic. In this article we build on this critique and argue that rule-based reasoning and the use of general guidelines as promoted by EBM does not accommodate the complex reasoning of doctors in clinical decision-making. Instead we propose that a new medical epistemology is needed that accounts for complex reasoning styles in medical practice, and at the same time maintains the quality usually associated with “scientific”. The medical epistemology we propose conforms the epistemological responsibility of doctors, which involves a specific professional attitude and epistemological skills. Instead of deferring part of the professional responsibility to strict clinical guidelines, as EBM allows for, our alternative epistemology holds doctors accountable for epistemic considerations in clinical decision-making towards the diagnosis and treatment plan of individual patients. One of the key intellectual challenges of doctors is the ability to bring together heterogeneous pieces of information to construct a coherent “picture” of a specific patient. In the proposed epistemology, we consider this “picture” as an epistemological tool that may then be employed in the diagnosis and treatment of a specific patient.

Philosophy of medicine has traditionally examined two issues: the scientific ontology for medicine and the epistemic significance of the types of evidence used in medical research. In answering each question, philosophers have typically brought to bear tools from traditional analytic philosophy. In contrast, this volume explores medical knowledge from the perspective offered by social epistemology. While many of the same issues are addressed, the approach to these issues generates both fresh questions and new insights into old debates. In addition, the broader purview offered by social epistemology opens up opportunities to address new topics such as the role of consensus conferences, epistemic injustice, the value of medical knowledge, continuing medical education, and industry funding. This article situates and summarizes the contributions to this special issue.

Since the term "evidence-based medicine" (EBM) first appeared in the scientific literature in 1991, the concept has had considerable influence in many parts of the world. Most professional societies, the public,and funding agencies have accepted EBM with remarkable enthusiasm. The concept of evidence-based practice is now applied in management, education, criminology, and social work. Yet, EBM has attracted controversy: its critics allege that EBM uses a narrow concept of evidence and a naive conception of the relationships between evidence, theory, and practice. They also contend that EBM presents itself as a radical restructuring of medical knowledge that discredits more traditional ways of knowing in medicine, largely in the interests of people with a particular investment in the enterprise of large-scale clinical trials. Because EBM proposes aspecific relationship between theory, evidence, and knowledge, its theoretical basis can be understood as an epistemological sys...

Journal of Medicine and Philosophy, 38: 461–486, 2013 doi:10.1093/jmp/jht044 A Framework for Understanding Medical Epistemologies GEORGE KHUSHF* *Address correspondence to: George Khushf, PhD, Department of Philosophy, University of South Carolina, Columbia, SC 29208, USA. E-mail: khushf@sc.edu What clinicians, biomedical scientists, and other health care professionals know as individuals or as groups and how they come to know and use knowledge are central concerns of medical epistemology. Activities associated with knowledge production and use are called epistemic practices. Such practices are considered in biomedical and clinical literatures, social sciences of medicine, philosophy of science and philosophy of medicine, and also in other nonmedical literatures. A host of different kinds of knowledge claims have been identified, each with different uses and logics of justification. A general framework is needed to situate these diverse contributions in medical epistemology, so we can see how they fit together. But developing such a framework turns out to be quite tricky. In this survey, three possible frameworks are considered along with the difficulties associated with each of them. The essay concludes with a fourth framework, which considers any epistemology as part of a practice that is oriented toward overcoming errors that emerge in antecedently given practices where knowledge is developed and used. As medicine indirectly advances health by directly mitigating disease, so epistemology indirectly advances knowledge by directly mitigating error. Keywords: clinical reasoning, epistemology, error, evidencebased medicine I. THE TASK OF DEVELOPING A GENERAL FRAMEWORK The task of developing a general framework for medical epistemology might be initially posed as follows: casting our net broadly, we search diverse databases for articles and books on the topic. As we do this, we also keep in mind © The Author 2013. Published by Oxford University Press, on behalf of the Journal of Medicine and Philosophy Inc. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 University of South Carolina, Columbia, South Carolina, USA 462 George Khushf 1. The patient history and physical exam as strategies for gathering evidence 2. Pattern recognition and expert knowledge 3. Heuristics and biases, and their role in clinical reasoning 4. Bayesian reasoning 5. Differential diagnosis 6. Causal reasoning 7. Mechanistic reasoning 8. How best to use information systems when you need knowledge related to a specific case 9. Justification and use of guidelines and clinical pathways 10. Systems based approaches to eliminating error and advancing quality 11. Theories and practices of measurement and testing; their use and abuse 12. Knowing patients as persons, rather than just regarding them as instances of diseases This is meant to be just the start of a list, given with no prioritization, and involving somewhat heterogeneous categories. It was generated by sifting through some of the materials at hand, with an initial effort to organize these under common headings that expressed something essential about materials grouped under them. The categories used to group the materials partly depend on the initial search terms that were used and assumptions about the kinds of things that are important in a general account of medical epistemology. To this extent, such categories only make explicit what was initially implicit before the search was conducted. But new patterns are also discovered, so there can be a refinement and development of the initial, partial knowledge. For example, after presenting the list, we might explore relations between items on the list. Some items on the list might better be regarded as subcategories of others on the list; for example, mechanistic reasoning as a species of causal reasoning. In that way, step by step it might be possible to work out a rich texture of relations. Anyone familiar with recent developments in medicine will be aware that there is a huge amount of activity related to medical epistemology that comes under headings like Evidence-Based Medicine (EBM) and Medical Decision-Making (MDM), so these naturally offer themselves as general Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 conditions under which knowledge is developed and used in medicine; for example, epistemic practices associated with management of a clinical case or a population of patients. After utilizing multiple search strategies that are informed by those medical epistemic practices we have in mind, we get a large number of books and articles. How can we categorize these diverse contributions, so we can get a relatively simple map of the topic and so we can properly understand the relations between the different contributions? As a first step in organizing the materials, we might group them under some general headings. By such initial groupings, something like the following list might be generated: A Framework for Understanding Medical Epistemologies 463 II. FIRST APPROXIMATION: DESCRIPTIVE, NORMATIVE, AND META-EPISTEMOLOGIES As a first approximation, we might use a three-fold scheme similar to that often used for ethics, where descriptive, normative, and meta-ethics are distinguished. Medical epistemologies might then be subcategorized as follows: 1. Descriptive epistemology involves empirical investigation into how people come to know the things they know, and how they use this knowledge. Here emphasis is on the actual reasoning and decision-making processes used by epistemic agents. These kinds of studies might involve empirical study of the differences between the way novice and expert practitioners reason through a case, use of heuristics and their associated biases, and psychological studies of the way physicians and patients appraise statistical information. Or one might study the way social systems and institutions evoke certain patterns of reasoning. This descriptive work uses diverse methods, ranging from introspective reflection of experts on their own reasoning processes to more traditional experimental methods used by psychologists for understanding cognitive and emotional processes and intelligence (Chapman and Sonnenberg, 2003). 2. Normative epistemology involves accounts of how people ought to arrive at knowledge and use it. Such accounts are often developed as refinements and extensions of ordinary, given reasoning, and with an effort to eliminate errors and unjustified beliefs. Here we might place many of the contributions in EBM and MDM (Kassirer and Kopelman, 1991; Eddy, 1996; Sackett et al., 2000; Sox et al., 2006). These contributions are normative in at least two senses. First, they involve claims about how clinicians ought to reason and make decisions. In doing this, they involve logics of justification considered necessary for distinguishing genuine knowledge from presumed Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 headings. These categories have thus already been used to group a large amount of the material we might identify, and standard textbooks on these topics provide nice reviews of that material (Sackett et al., 2000; Jenicek and Hitchcock, 2005; Sox et al., 2006; Guyatt et al., 2008). But these headings are not sufficiently general for organizing all materials relevant to medical epistemology: they are both too broad, including things we would not want to include, and they are too narrow, leaving out important areas that we would consider important. The question is now whether there is a higher-order general scheme that can be used to sort all the various lower level groupings. The sought-for general framework would provide a kind of synoptic overview that enables us to see all the issues in proper relation to one another. This framework should be of use in organizing the things in the partial list above, and it should also enable proper sorting of any other categories that anyone else might develop. 464 George Khushf Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 knowledge (Jenicek, 2003). Second, EBM guidelines reflect substantive normative judgments about what is good and valuable and how to rank outcomes (Mills and Spencer, 2003). There are ethical obligations to avoid harm and to do what is best for the patient, which also inform EBM. How patients and clinicians understand what is good involves background normative assumptions. Similarly, EBM invokes the goal of avoiding waste and using resources effectively. Efficiency judgments involve normative assumptions about how to prioritize what is of benefit to individuals and populations (Blumenthal-Barby, 2013). There can be tensions between the norms associated with the justification of knowledge and the norms that reflect judgments about what is good, and a normative epistemology might work to clarify and resolve these tensions. 3. Meta-epistemology involves investigation into higher-order concepts that are used in both the descriptive and normative epistemologies; for example, the nature of knowledge and evidence, logics of justification of knowledge, kinds of inferences, and so on. These are the issues normally dealt with in philosophy under the heading of “epistemology.”1 The “meta-” prefix was added to distinguish this more abstract and general reflection on medical epistemology from both the traditional philosophical reflection, on one side, and the descriptive and normative medical epistemologies, on the other side (Ashcroft, 2003; Howick, 2011). Such a three-fold division seems to capture something important. The headings naturally map to the distinctive styles and methods of investigation and argument we find in the heterogeneous medical literatures related to epistemology. In areas like neuroscience, psychology, and sociology, and in hybrid areas like cognitive science, empirical methods are used to make sense of how humans know, how knowledge is used, and how real-world agents function in real-world environments (or at least how they function in the artificial, hypersimplified environments associated with some psychological experiments). There is now an extensive body of such work associated with the study of physicians and patients, and this work naturally fits under the “descriptive epistemology” heading. Alternatively, much of what comes under the headings of EBM and MDM seeks to prescribe how physicians ought to reason rather than describe how they do reason. The prescriptions are often developed as compensations for the ways ordinary medical reasoning goes astray. We then see a gap between what is and what ought to be, and this gap has many of the same features found in the gap between actual behavior and normative accounts of what should be done. An appreciation of this similarity between what is given and ideal in ethics and epistemology has inspired one prominent, so-called deontological account of the nature of justification associated with knowledge (Alston, 1989). Finally, there are higher-order questions usually not asked by physicians and biomedical scientists, but of special interest to philosophers; for example, regarding what counts as evidence, how claims are justified, what distinguishes genuine from presumed knowledge, and the nature of causal and mechanistic reasoning A Framework for Understanding Medical Epistemologies 465 Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 (an outstanding example is given in Howick, 2011).These again have a style that is distinctively different from both the descriptive and the normative work, and they might naturally be associated with the meta-epistemology (what is called epistemology in philosophy). Thus one of the advantages of the three-fold schema relates to the way it helps us descriptively sort and make sense of the diverse contributions associated with clinical reasoning and decision making. Another advantage of the three-fold classification scheme concerns the role it might play in aligning classical philosophical work on epistemology with those reflections that arise in medical and scientific contexts. One of the most exciting developments in recent years concerns the way leading philosophers of science have now moved to a consideration of medicine as a representative area where deep philosophical questions about science can be explored.2 This work takes the actual reasoning and practices of physicians and scientists seriously but at the same time brings to this work a refined set of tools and critical capacities for reflecting on things like evidence, causal analysis, and mechanism. But it has always been tricky to align this kind of work in epistemology related to the philosophy of science and the classical philosophical project of epistemology associated with justified true belief. Efforts to carefully describe, understand, and critically refine the epistemologies of the natural sciences have often gone under the name of “naturalized epistemology,” and since Quine (1969), they have been contrasted with the traditional project in epistemology. But even strong supporters of Quine have felt that there is something about the classical epistemological questions that is legitimate, and efforts at developing an epistemology grounded on science often try to work out those accounts in relation to classical philosophical questions in epistemology (Feldman, 2012). By means of the three-fold schema of descriptive, normative, and metaepistemology, we have a framework that might be used to bring together disparate sets of reflections that often have not been in communication with one another. If we imagine the three subcategories as corners of a triangle and take the sides of the triangle as lines of communication between these subcategories of medical epistemology, then we could envision additional edges that run from each node outward to nonmedical fields. The node associated with meta-epistemology would run over to classical philosophical debates in epistemology. That associated with descriptive epistemology would run over to more general empirical work in the cognitive sciences, psychology, sociology, and anthropology. And the node associated with normative epistemology would run over to areas like logic, statistics, computer sciences, engineering, and business management. By reflecting on the epistemologies integral to a specific set of practices like medicine, we can see how all these general, nonmedical domains need to be aligned. Further, we also can gain a critical perspective on what was incomplete when armchair epistemology sought to make sense of knowledge without careful consideration of how knowledge arises and is used in practice contexts. The 466 George Khushf Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 three-fold schema thus enables us to appreciate the contributions that classical epistemology might make to medical epistemology and, conversely, the contributions that might be made by medicine to classical epistemology. The value of such a dialogue cannot be overstated. In a strange way, the extensive, vibrant empirical and normative work on epistemology in medicine has until recently gone largely unnoticed by philosophers. There are huge fields of medical decision making and whole industries associated with organizing evidence so it is accessible to clinical practitioners. Medicine has been one of the central areas where information systems have been used to support complex tasks of reasoning and decision. And there are now major efforts at organizing systems so they evoke, sustain, and refine the right kinds of reasoning. With the exception, perhaps, of the military, business management, and use of information systems, there is no area where more work has been done on epistemic processes related to a practice domain. But all this valuable work has rarely been informed by, or informed, classical work on epistemology. Whereas the reasons for this are complex and include a historically interesting antipathy of medicine toward philosophy, one of the significant barriers has been the absence of a general framework for bridging the classical philosophical and medical work. By providing such a framework, the three-fold schema can help facilitate a dialogue that is already emerging. Despite these potential benefits of the three-fold rubric, there are some significant problems with it. First, it seems to include too much. There was a time when science was called “natural philosophy,” but no one would today associate philosophy of science with the full range of contents associated with science. In the same way, we cannot associate epistemology with the full range of methods, strategies, and tools for advancing and using knowledge in practice settings or for empirically studying these. Such an overreaching categorization would imply that much of psychology and cognitive science are epistemology, and that physicians are trained in epistemology so they can properly diagnose patients. Whereas the etymology of epistemology might suggest such inclusive approaches to the study of knowledge, we must today take a far narrower reading. As philosophy of science involves a kind of second-order, critical reflection on science, so likewise medical epistemology (at least of the so-called naturalized variety) involves a kind of critical, second-order reflection on knowledge acquisition and use in medicine. Here the “meta” is really already integral to the meaning of epistemology. This narrowing of epistemology doesn’t by itself invalidate the three-fold schema. We could refine it by again considering our triangle, with its spokes extended outward from each node. Only now we introduce a third dimension, and have additional spokes descending from the nodes perpendicular to the plane of the triangle. We now shift downward the contents originally associated with each corner of the triangle, and we take descriptive, normative, and meta-epistemology as the second-order, critical reflections on those A Framework for Understanding Medical Epistemologies 467 Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 lower level domains. Each of these relates to their base in roughly the way that a philosophy of a special science relates to the special science it studies. Medical meta-epistemology then relates to epistemology in the way a philosophy of a special science like physics relates to the more general field of philosophy of science. This seems to get us closer to the framework we need, but there is still something that doesn’t quite work. The problem becomes clear if we move to meta-epistemology and, from within that domain try to clarify the relation between the descriptive and the normative. As soon as we do this, we notice that the very distinction between descriptive and normative is implicated. What are we doing when we seek to describe how human agents come to know and use knowledge in practice settings? In doing this, our own apparatus for coming to know something is deployed in a double sense. First, it is deployed in our effort to describe and categorize the diverse contributions made to medical epistemology. Second, it is deployed normatively, as we present evidence and sift claims and organize contents so they cohere in the right ways. In framing our descriptions of epistemic processes, we do this in relation to our own norms regarding how knowledge ought to arise and be used. These norms play a role from the start in how we select what counts as relevant material. Alternatively, if we seek to advance a normative epistemology, we do this by providing a kind of description, albeit often in a rather artificial language and often in the subjunctive. In either case, all accounts of epistemology take place at the intersection of the descriptive and the normative. Even at the most abstract level, where the focus is on justified true belief, we still see a shadow of the descriptive/normative distinction: first, belief is posited, arising from some fuzzy, unspecified process (the descriptive, given level); then there is the question of how we discriminate that special set of justified, true beliefs arising from faculties somehow fit for the circumstances.3 This second stage associated with justification is explicitly normative. At this most abstract level of reflection the intricate complexities of descriptive and normative epistemologies are lost, and these are handled summarily under the headings of belief and the normative (or technically nondeontic) machinery that assures such beliefs are true and justified or at least reliable. But when we consider the rich texture of epistemic processes associated with a field like medicine, we can no longer handle all this with sweeping summary terms. When we descend into the details, it becomes clear that one’s own normative commitments inform any effort to categorize contributions into the three domains. The net effect of all this is that we are not going to fit fully intersubjective categorizations of epistemologies into the three subcategories. What one person takes as normative, another will take as descriptive. The three-fold schema can thus, at best, be used as a rough first approximation. A quick glance at the literature on EBM will immediately confirm this. For many physicians, EBM is a catch phrase for the normative ideals regarding the 468 George Khushf III. SECOND APPROXIMATION: PURE VERSUS APPLIED EPISTEMOLOGIES For our second effort at developing a general framework for organizing contributions to medical epistemology, we begin with a distinction between pure and applied epistemologies and focus on the kinds of knowledge deployed in medical epistemic practices. Classical epistemology has focused on cognitive content and has assumed that knowledge can be propositionally encoded. Beliefs are then associated with the content of propositions, and the evidence that sustains beliefs is likewise such that it can be expressed propositionally. Truth then concerns the relation between the propositionally encoded content and worldly states of affairs, and justification is understood in terms of the inference chains that link up diverse propositions. We can take epistemologies concerned with such inference chains as “theoretical” or “pure” epistemologies. At the outside limit of such pure epistemologies, there is a kind of knowledge that has a qualitatively different character. This is often associated with the primitive, so-called atomic propositions that encode the bits and pieces of evidence. For example, when a physician notices red blotches on the skin while conducting a physical exam, this might be propositionally encoded as “the patient has red blotches indicative of Herpes zoster.” This propositional expression involves a complex combination of what is immediately seen— the red blotches—and the diagnostic category that may be associated with it, Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 ways physicians should ascertain, sift, and utilize knowledge to effectively manage the diseases of patients. General overviews of EBM often begin with descriptive accounts of how physicians have traditionally approached clinical decision making, and why such approaches are deficient. Then, in response to these descriptions of problematic clinical reasoning, EBM is advanced as a corrective. But for many philosophers, EBM constitutes the given, real-world, knowledge-oriented practices they seek to describe. They then critically reflect on the claims integral to EBM, for example, related to notions of evidence or mechanistic reasoning (Howick, 2011). On the basis of these critical reflections, a normative account is provided that is supposed to refine EBM. Philosophical accounts thus relate to EBM in the same way that EBM relates to the antecedently given, typical patterns of clinical reasoning. At each stage, a normative account provides a refinement of an antecedently given set of knowledge-oriented practices. As soon as such normative accounts inform the practices themselves, this reconstructed set of practices becomes the target of new efforts at reconstruction. The process is iterative and ongoing, and thus destabilizes any categorizations according to the three-fold schema. What is normative at one stage becomes part of what is taken as given and described in a subsequent stage. A Framework for Understanding Medical Epistemologies 469 Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 and that is used to determine the perceived red blotches as a sign of a given disease. The sign as categorized percept, implies a complex differential system that sorts by inclusion and exclusion this instance of red blotches in relation to other possible variants of red blotches. Somewhere in the midst of this complex encoding, there is a primitive relation to an external world that is mediated by the visual apparatus of the physician. This involves a kind of knowledge of what is immediately given, and the function of atomic propositions is to encode this propositionally. There will then be propositions that have this kind of immediate, qualitative, external support. These will be associated with “evidence.” And there will be other propositions that arise from complex inference chains associated with such evidence together with various higher-order categories, empirical generalizations, and so on. Much of the classical work in epistemology concerns the complex processes that lead to the encoding of such primitive empirical relations and the inference chains that link more complex propositions back to their evidential base. When we consider the everyday contexts where medical knowledge is deployed, it becomes immediately apparent that there are many kinds of knowledge that do not fit within this theoretical account, where the only nonpropositional knowledge concerns the primitive perceptions associated with encoded evidence. A physician might know her way around a hospital. She might know a patient personally, rather than just know of her case. She might know how to perform a given procedure. Or she might not know how to effectively use information systems as part of her regular clinical management of patients. She might be poor at interfacing with other health care professionals, and such failure might be partly due to an absence of knowledge about what other professionals can do and partly due to some moral trait such as pride. All of these kinds of knowledge cannot be captured by the standard epistemologies that focus on justified, true, propositionally encoded beliefs (the classical criticism is given in Ryle, 1949). But all these kinds of knowledge are essential for the practice of medicine, and many of the important developments in medical epistemology concern these nonpropositional (or transpropositional) kinds of knowledge. Additionally, it is important to remember: “Diagnosis is not knowledge for knowledge’s sake. It is knowledge for the sake of action. Medicine exists in order to cure, to care, to intervene, or in limiting cases, to know when not to intervene. Medicine is not a contemplative science” (Mainetti, 1992, 79). We might group epistemologies that are concerned with know-how and these existential, situated, and instrumental kinds of knowledge under the heading of “applied epistemology.” This kind of knowledge was traditionally associated with what has variously been called phronesis, practical wisdom, or the art of medicine. When we try to clarify how such situated knowledge and know-how count as knowledge, then things get tricky. What exactly is known? In answering this, we might again start with the classical mantra of justified true belief 470 George Khushf Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 and ask how this might be generalized to nonpropositional cases. “Belief” involves a mental state that somehow relates to the way an agent takes a stance in relation to the world. It thus has an “about relation” to some state of affairs independent of the agent. “Truth” then concerns a match between the belief and the state of affairs it is about. “Justified” then concerns the grounds of belief, assuring that the truth of the belief is not an accident, but somehow involves legitimate links or inference relations to those contents immediately ascertained and encoded as evidence. Additionally, some kind of contextual aspect is included that is supposed to assure that the perceptive and reasoning apparatus is functioning reliably in the circumstance under consideration. Generalizing from this classical account, we can say that epistemology is concerned with the integrity of knowledge and its use. In any set of practices, a set of capacities is deployed that assures there is a fit between the associated actions and the world within which these actions occur. From the perspective of the agent, knowledge concerns one aspect of this fit between the world, on one side, and the agent’s grasp of the world and actions within it, on the other side. Epistemology is then concerned with the grounds and justification of these given ways of knowing, including the know-how integral to action. At any stage, human agents seek assurance that their assumptions and practices relate in the right ways to the world within which they act, so there are the right kinds of inference chains and practice chains. The inference chains might be modeled in traditional theoretical terms, but the practice chains involve sets of trial and error corrections over previous practice schemas. In each case, there is a complex set of chains to the antecedent grounds on which they are based. This normative question about justification involves a complex alignment between agent goals and all the things the agent takes as given. This second-order, normative component is integral to epistemology. It involves an effort to assure that what is taken as knowledge is aligned with the world in the manner called “true,” and that know-how is aligned with ends in the efficient and effective way called “fit.” We can now ask more specifically about what an agent knows, when an agent knows something. How do we understand the about-relation in the case of nonpropositional knowledge? And how might appreciation of these nonpropositional contents of knowledge lead to refinements of epistemologies that are more narrowly developed in terms of the propositional contents that are normally of interest to science? The first two essays in this journal issue provide partial answers to this question. Ashley Graham Kennedy (2013) considers two cases of differential diagnosis where physicians too quickly move from limited information to either a problematic dismissal of patient complaints or a premature closure with respect to a diagnostic category. Her first case involves a controversial diagnostic category—that of Lyme disease. The symptoms associated with late stage Lyme disease are often nonspecific, the tests for the disease A Framework for Understanding Medical Epistemologies 471 Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 are not specific enough for conclusive diagnosis, and the treatment is relatively expensive. When confronted with the nonspecific symptoms and the absence of conclusive tests, a patient who has the disease can be dismissed, with the inference that it is “all in his head.” Kennedy presents this as an inappropriate overreaching of evidence. In the case she reviews, the patient quickly responded to antibiotic treatment. Kennedy’s second case concerns a patient with Addison’s disease. The primary initial symptom was that of depressed mood in a patient who had recently lost his spouse in a car accident. In this case, the presenting circumstances and symptoms naturally converged upon a diagnosis of depression. But when the symptoms increased and the patient repeatedly sought medical assistance, physicians refused to reconsider the initial diagnostic category. Eventually the case was properly diagnosed and the patient rapidly improved. Kennedy presents this as a problem of premature closure. When initially considering her cases, we might try to understand the epistemic concerns in traditional terms, for example, as involving inferential chains that are insufficiently grounded in the evidence that is available. Even when viewed in such terms, there are some interesting aspects of her two cases. Today, “successful” treatments have a highly ambiguous status in confirming a diagnosis, especially when the symptoms are largely self-reported and nonspecific. But at an earlier time when there were fewer tests available for confirming a diagnosis, effective treatment was explicitly taken as a probabilistic confirmation of a diagnostic hypothesis, and trial and error was seen as an integral part of medical management (Flexner, 1910, 55). Implicitly, Kennedy takes the effective resolution of the presenting symptoms as a confirmation of the diagnoses, especially in the case of Lyme disease. Such pragmatic considerations are usually quite tricky to integrate into the standard accounts of knowledge. But for Kennedy, there are even trickier kinds of knowledge at issue in these cases. Whereas it is hard to name exactly what is known, this knowledge has something to do with the awareness a physician has of uncertainty and his or her own limits in understanding the medical conditions that patients have. The knowledge of one’s own limits then needs to be deployed by the physician in what Kennedy aptly characterizes as a “compassionate suspension of judgment.” She then associates this suspension of judgment with three virtues: a methodological virtue associated with ongoing investigation; an epistemic virtue associated with humility; and an ethical virtue associated with respect for patients. In his essay on emotional intelligence (EI), James Marcum also focuses on practical rationality. His “main thesis is that EI is a practicable or workable form of intelligence, akin to Aristotle’s notion of phronesis, which affords medical practitioners a robust cognitive resource for providing quality clinical care.” (2013, 501) Emotions are organized human responses to events that are judged to have a positive or negative impact on the well-being of an individual (Salovey and Mayer, 1990, 186). Since humans often appraise 472 George Khushf Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 similar events in similar ways, and since they express their positive and negative reactions in similar ways (e.g., by smile or frown), emotions and their expression provide important information about humans and their appraisals of their environment. They also play a role in regulating dispositions and other cognitive processes associated with planning and action. EI is then regarded as a general capacity to accurately monitor, identify, and distinguish these organized human responses in the self and others, and also to utilize this information in a purposeful way to guide thinking and action (Salovey and Mayer, 1990, 189). Viewed in this way, emotions constitute important, cognitively relevant information that needs to be detected and utilized when considering and interacting with people. By following a complex medical case, Marcum considers how EI might be used to discern information about patients and complex nonmedical events that may be at the root of problems that initially present as medical problems. Both Kennedy and Marcum consider cases where there is a complex intersection of medical and nonmedical considerations, and where part of the diagnostic task involves disentangling these strands of information. Both seek to move beyond an epistemology that only registers biomedical facts in ways that cannot be properly related to patients who must be encountered as people and not as instances of disease. In this effort to expand the horizon of the epistemology, their contributions align with other important efforts to incorporate kinds of knowledge that are not sufficiently appreciated in current biomedicine (Engel, 1977a, 1988; Malterud, 1995). But there is some tension between the lessons they draw from their different cases. Kennedy is concerned about the way physicians might dismiss presenting complaints as psychiatric or nonmedical, whereas Marcum highlights how a physician’s EI enabled him to recognize the traumatic event of rape that was behind a presentation initially worked up as irritable bowel syndrome. This difference is sharp enough and their cases similar enough that one might ask how Kennedy’s “compassionate suspension of judgment” and her three virtues might have been (or were) applied to Marcum’s case and, conversely, how Marcum’s EI might have informed management of the cases that Kennedy describes. Beyond any possible tensions, both essays clearly identify a kind of awareness a physician should have about himself or herself as an embodied, limited agent who encounters another embodied, limited agent in need of help. The deployment of cognitive capacities that work with taxonomies of disease to categorize instances associated with manifest signs and symptoms is in each case taken as just one aspect of medical knowledge. Additionally, another kind of practical knowledge is needed to properly situate and appraise the scope and limits of these cognitive capacities, and order them rightly in relation to an effective clinical encounter. The practical knowledge associated with an awareness of the scope and limits of knowledge (Kennedy) and EI (Marcum) thus orients the use of theoretical knowledge in ways that optimize the goal of the practice, A Framework for Understanding Medical Epistemologies 473 which is to provide a fitting response to that person who comes to the physician for help. IV. THIRD APPROXIMATION: INDIVIDUAL VERSUS SOCIAL EPISTEMOLOGY Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 Thus far we have developed epistemologies as if they concern what goes on in the heads and hearts of individuals. Individuals are the ones who reason and use information, and thus the ones who have knowledge of various kinds. But when an individual physician works up a patient’s problem as a biomedical disease, this isn’t an isolated event. That physician deploys an apparatus for making sense of disease in ways that are similar to other physicians. These diverse physicians are conditioned to work up disease in those distinctive ways promoted by educational systems, professional organizations, systems of health care practice, institutional payers and by the technologies available to support the epistemic activities of physicians. One of the more distinctive recent developments in medical epistemology concerns the new (or rediscovered; Fleck, 1979) appreciation for the role such social systems and technology play, and how they even condition the way theoretical and practical knowledge are disentangled and related to one another (Goldman, 1999). As soon as one takes this social and material turn, a somewhat unruly literature related to medical epistemology gets downright unmanageable. This turn has an especially destabilizing impact on discourse on truth and justification (Rorty, 1979; for a social epistemology that seeks to preserve the discourse of truth, see Goldenberg, 2006). To ease our way into such a literature, let us start again with the classical epistemology associated with justified, true beliefs and try to extend it with small steps so that we can bring into view the dependence relations between knowledge and social systems. Let’s call “perceptive capacity” the sum of abilities an individual has to take in information from the external world. This includes vision, hearing, touch, and so on. “Perception” then involves a use of perceptive capacity, and we’ll assume that this eventuates in a propositionally encoded piece of evidence that might be considered by an epistemic agent. Let’s now define “reasoning capacity” in a broad way so that it includes the full range of abilities an individual might utilize to filter, organize, and extract some conclusion from multiple pieces of evidence. In traditional theoretical or pure epistemologies, knowledge generation involves a special kind of coupling between the perceptive capacities, for example, evidence provided by vision, and the reasoning capacities. Worries associated with justification involve constraints that assure the bits and pieces all link up in the right way to get genuine knowledge. In the light of this traditional epistemology, we can now identify an important metaconstraint associated with justification. It is important that 474 George Khushf Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 the perceptive capacities associated with evidence and the reasoning capacities that eventuate in some complex judgments remain tightly coupled in a circuit whose transformations can be surveyed by the epistemic agent. Behind the traditional epistemology, we have an intuition something like that expressed by Edmund Husserl in his Logical Investigations (2001): it is assumed that any complex, symbolically encoded proposition could only be accepted as knowledge if it is possible to follow the chains back to primitive acts of perception that are self-evident (Dummett, 1993). We can associate this same view with a negative criterion: if the required chains of justification run outward into some regions of the external world that cannot be surveyed (an aspect of the social that is inaccessible to the epistemic agent) and then, after some time gap, comes back transformed or augmented into the accessible space of the epistemic agent, then this would fundamentally alter the epistemology. Knowledge would then depend on some inaccessible condition that lies outside the agent, and it would require some account of what happens in that social circuit and how those transformations assure that belief dependent upon that circuit remains true and justified. We could summarize this general condition by saying that the circuit between immediate perceptive evidence and knowledge needs to remain accessible to the epistemic agent. By introducing the notion of an accessible epistemic circuit, we provide a tool for disentangling unproblematic from problematic kinds of dependence on social systems. For example, language and concepts are obviously social constructs and they play a crucial role in epistemic processes. But advocates of traditional epistemologies would not find this problematic, because recognition of this kind of social dependence does not imply an inaccessible epistemic circuit. Similarly, if certain kinds of technologies are used to augment perceptive capacities, this may or may not disrupt the accessibility of an epistemic circuit. Whether it disrupts depends on the technology, circumstance, and use. Most physicians, for example, would not view use of a stethoscope to mediate auditory signals as a disruption of the accessibility of an epistemic circuit. There is thus some comfort with minor extrusions of epistemic processes. The key question is then how far such extrusion might run before an epistemic circuit is taken as inaccessible to the epistemic agent.4 We will say that an inaccessible epistemic circuit involves at least one disruptive extrusion. We will speak of three kinds of disruptive extrusions that could take place: extrusions of perception, extrusions of reasoning, and extrusions of value. The first two kinds are associated with the two core elements of traditional epistemology. In a disruptive extrusion of perception, empirical evidence depends on capacities that extend beyond an individual’s perceptive capacities. In a disruptive extrusion of reasoning, the chains that link evidence to conclusions extend beyond an individual’s reasoning capacity. The last kind of disruptive extrusion—that of value—is associated with A Framework for Understanding Medical Epistemologies 475 Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 the practical kinds of knowledge considered in the last section. We say that a social epistemology is required when there is at least one inaccessible extrusion, and thus the justification of the knowledge depends on social conditions that are at least partly inaccessible to the epistemic agent. Although a bit of a caricature, a physician’s clinical reasoning is often presented in a way that roughly tracks traditional epistemology (for representative examples, see the cases discussed in Cutler, 1998). The perceptive component is expanded to include technologically mediated information, for example, blood work or radiographic images. These expansions are often regarded as straightforward augmentations of normal perceptions. Thus laboratory microscopes and x-rays extend human vision, and stethoscopes and ultrasound machines extend hearing. These extensions are not supposed to alter the way evidence arising from these augmented perceptive acts enter into the reasoning process, that is, they are generally presumed to not involve a disruptive extrusion of perception. A medical case is then presented as a time sequence in which bits of evidence are acquired, with interim stages of reasoning that present a physician’s conceptual train of thought, leading to decisions about the next bit of evidence that is required or that spontaneously presents in the ongoing development of the case. Eventually, the case description is appropriately resolved when a given disease category properly makes sense of all the evidence in the case presentation. Our question is now: can we sustain such a traditional account of clinical reasoning or are there disruptive extrusions of perception, reasoning, or value that have been insufficiently appreciated? In this journal issue, Leah Mcclimans, Haley Faust, and Jennifer Blumenthal-Barby each address aspects of this question. Leah Mcclimans’s reflections on Quality of Life measures and Patient Reported Outcome Measures (PROMs) can be contextualized by considering the ambiguous status of patient testimony as evidence in medicine. In the last section, we tried expanding the notion of epistemology so that it incorporated a broader range of contents, and so knowledge was deployed in the right way to accomplish the goals associated with a clinical encounter. When reviewing this expansion, we still placed our emphasis on the physician’s knowledge, for example, a physician’s practical wisdom, which includes awareness of uncertainty and limits, and thus doesn’t overreach evidence or prematurely close off a case, or a physician’s EI. But when we presented the epistemological concerns in that way, we only considered one aspect of case studies presented by Kennedy and Marcum. In their case studies, there was another very important element: patients also had knowledge! Physician failures were not just due to their lack of some kind of knowledge. Beyond that, they failed to listen. They did not regard patients as genuine epistemic agents and thus discounted the epistemic value of their testimony. Alternatively, the more attentive physicians appreciated that patient’s had a kind of privileged, immediate access to their own medical conditions and 476 George Khushf Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 thus could express knowledge that the physician could not directly obtain. But this assumption about the evidentiary status of patient testimony is not trivial. Alvin Goodman rightly notes how assertions about irreducible testimony require a social epistemology: “the problem of testimony is a problem of justification: what makes a hearer justified in accepting a report or other factual statement by a speaker?” (Goldman, 2010, 627). From the perspective of a traditional epistemology, testimony should be reducible. If it is irreducible, then there is an extrusion of the process that fixes the evidence, and knowledge gets distributed among different epistemic agents. Many studies have considered how the economy of medical knowledge orients physicians in ways that discount patient testimony as reliable (Canguilhem, 1991; Foucault, 1975; Engelhardt, 1996). This discounting has usually been associated with a biomedical disease concept (Engel, 1960, 1977b; Cassell, 1985). That disease concept is then associated with a Cartesian mind/body split, where the structures and functions of the parts and processes of the body are legitimate objects of scientific investigation, whereas the patient’s experienced illness is regarded as a subjective matter of mind (Toombs, 1992; Leder, 1992; cf. Simon, 2008). If we now consider this same theme from the perspective of a traditional medical epistemology, we see another driver of this reduction: if patient testimony is regarded as reliable knowledge, a social epistemology is required. To preserve the traditional epistemology, patient testimony must be reducible, and that demotes it to a lower status of evidence. Thus, with a biomedical model of disease and traditional medical epistemology, patient testimony becomes a kind of hearsay. It involves deficient access to the disease that a physician seeks to understand in a scientific way (Reiser, 1978). The problem here is not just whether a physician should take a good patient history. All physicians would appreciate the need to do this, and all would appreciate that the patient’s testimony provides important information that should inform diagnostic reasoning and decision making. As Paul Cutler, the editor of an influential book in EBM, notes: “Like an embryo, the diagnosis is conceived, germinates, develops, and is most often delivered during the dynamics of taking the history, and the entire drama of medical problem solving may unfold without leaving the stage set of two chairs and a desk” (1998, vii). But hearsay is also useful for a detective tracking down evidence, and most of the cases handled by police are handled at the site of complaint and without further incident. At the same time, such hearsay is not admissible in some courts where higher standards of evidence are required. The key question concerns how the patient’s testimony is to be related to other kinds of evidence associated with direct observation, measurement, and tests in cases where what the patient says and what the tests indicate (or fail to indicate) diverge. Here we come to a problem that some physicians associate with “the worried well.” That phrase itself indicates the presumed discrepancy between the patient’s concern about illness (or perhaps A Framework for Understanding Medical Epistemologies 477 Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 suboptimal health) and the physician’s judgment that the patient is, in fact, well. Is the patient’s testimony simply the point of departure for other activities that get at the same evidence in a more direct way or, alternatively, is the testimonial evidence irreducible? There is a kind of mismatch between the testimonial evidence and the other kinds of evidence obtained by the physician. Additional work is needed to bridge them. When physicians use a traditional epistemology, priority is given to nontestimonial evidence and an effort is made to reconstruct testimony so it better fits. When the patient testimony is of value as an indicator of disease, then a physician can confirm that evidence by finding other evidence that is independent of the patient’s testimony. Alternatively, if the patient testimony is somehow irreducible as evidence of an outcome of a medical intervention, then strange “instruments” are needed for disciplining the otherwise unruly evidence: thus the PROMs McClimans studies. McClimans (2013) begins by highlighting how the machinery of medical measurements is usually black boxed. In EBM, a strong emphasis is placed on statistics of mortality, morbidity, and quality of life. To get such statistics, these outcomes need to be measurable. For that, reliable measurement instruments are needed. There are a host of instruments out there, and they are regularly used to report outcomes. These outcomes influence health care decision making in many ways. But few attempt to open the black box and critically reflect on the complex decisions implicated in any use of the instrument (Baird, 2004). She seeks to do this with instruments for measuring quality of life. In doing this, she poses a double challenge: first, generally, she challenges the way evidence associated with measurements and tests enter into clinical reasoning as bits of information. As soon as the black box is opened, it becomes immediately apparent that measurements don’t arise by a straightforward extension of human perception. There are justificatory questions that must be resolved to get a valid measurement instrument, and some measures fall far short of answering these questions. Here there is a disruptive extrusion of perception, and much more work is needed to clarify the social processes and the material work associated with ascertaining those bits and pieces of evidence. Second, she notes that there is something especially troubling about the effort to measure patient reported outcomes related to quality of life. Here the patient’s own testimony enters in a direct, irreducible way. The survey instrument provides a tool to constrain and regularize that testimony so averages over populations might be developed. McClimans highlights how this can do violence to exactly that which the measures are supposed to accomplish, namely, ascertaining what makes a positive difference in the lives of patients. There is additionally an extrusion of values: nontrivial social mechanisms and decisions are implicated in the use of PROMs. By opening the black box, McClimans brings into view these social mechanisms and submits them to critical scrutiny. 478 George Khushf Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 Halley Faust (2013) focuses on problems associated with causal analysis that arise in primary prevention. To appreciate the difficulties addressed by Faust, we might start with a case where the causal logic integral to a preventive intervention is not so distant from that causal logic associated with a traditional therapeutic intervention. A physician wants to make causal inferences about the possible interventions that might make a positive difference in a patient’s life. Let’s assume a physician is considering a patient who has had high blood pressure over several visits to the clinic. The physician might initially recommend changes in diet and exercise and indicate that blood pressure needs to be monitored. If there is no improvement, then some drug may be prescribed to “control” the blood pressure. The pill is regularly taken as prescribed, and on a subsequent visit a new measurement is taken, confirming that blood pressure has been reduced to an acceptable level. Here there is a simple circuit: something is measured, and the value of the measure is considered in relation to some norm. If that measure is outside the norm, an intervention takes place and the measure is again taken. The intervention involves or evokes some mechanism that is supposed to alter the thing that is measured in some predictable way. Subsequent measurement either confirms or disconfirms that this predicted alteration has occurred, and subsequent medical decision making proceeds in a way that depends on this outcome. This simple scenario can only occur because of much antecedent work by biomedical scientists and clinicians who have constructed the conditions under which the associated causal inferences and interventions can take place. As soon as we start to ask about blood pressure, instruments for measuring it, interventions for correcting values outside the normal range, and so on, we find that a host of nontrivial decisions needed to be made before these simple cases of causal reasoning could ever arise (Schwartz, 2008). Is the intervention to control blood pressure a treatment of a kind of inflammatory metabolic disease, hypertension, or is it a prevention of diseases like cardiovascular disease and stroke? In this case, the act of prevention targets some factor that is related by association or cause with diseases that involve partial or catastrophic failure of vital organ systems. The prevention thus mobilizes the same general notions of cause and disease that are mobilized in other cases of medicine. But what happens if we move from secondary and tertiary preventive efforts to cases of primary prevention where there is no elevated risk in the individual who is part of the population that is the target of the primary preventive intervention? If the intervention is successful, there is what Halley Faust has identified as a double metaphysical bind: “first confirming that something didn’t happen, and second proving we caused the absence.” He speaks of these as “the two ‘hard’ questions of prevention: (1) cause without an effect, and (2) absence of action as a cause” (Faust, 2013, 540). For example, if one has a smoking prevention campaign that works, this has the effect that people will be less likely to smoke. A Framework for Understanding Medical Epistemologies 479 Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 The effect is thus an absence of action, that is, not smoking. This, in turn, means there will be people who do not get cancer who otherwise would have. The effect is then no effect in a double way: no smoking and no cancer. When primary prevention is considered in relation to diseases that are prevented, Faust argues that we get a metaphysics of “causing not” that is incompatible with general assumptions about cause in the scientific community. “The standard covariation theories of causation require that there be a positive physical event in relation to the cause and effect. Or the standard causal power theories require that some mechanism be in place to transmit down a causal chain pathway. But how to reconcile that if nothing is done then there is nothing . . .?” (Faust, 2013, 550). To avoid these problems, he proposes changing the way preventive medicine is understood. Instead of focusing on a cause of an absence of a negative event, the disease, he thinks preventive interventions should be seen as an advancement of a positive event, optimal health. When a risk or susceptibility is reduced, then this improves health from some lower state (H1) to some higher state (H2). These health states are ordered in relation to some optimum, and the causal mechanisms need to be understood in terms of the contribution an intervention makes to some incremental increase in health. Faust appreciates that there are a host of values integral to preventive measures, but he seeks to bracket these questions and focus on the epistemic factors related to causal analysis. The huge shift from disease prevention to health promotion is motivated by his concern as a scientist to create the conditions under which causal claims can be properly made. Normally, the logic of cause and its relation to interventions is simply taken for granted as part of an explanatory scheme. In Faust’s deliberations, we see how a master architect of preventive medicine seeks to create those conditions where those explanatory schemes can be taken for granted. Here there is an extrusion of the rationality integral to current preventive medical interventions, together with an effort to construct altered social conditions of practice where the pathways of justification are more transparent. In interesting ways, Faust’s analysis also links back up with that of McClimans and highlights another important aspect of PROMs. Unlike measures of morbidity and mortality, PROMs seem to get at something positive about a patient’s well-being. “Quality of life” would presumably be an aspect of any measure of optimal health. Advancement of Faust’s conditions for causal analysis thus require a simultaneous advancement of measurement instruments like those associated with PROMs. In the final essay of this issue, Blumenthal-Barby (2013) considers scientific and professional judgments about what tests or interventions have value. The American Board of Internal Medicine (ABIM) Foundation has advanced a “Choosing Wisely” initiative to identify and eliminate low-value care. Several medical professional organizations have identified five tests, 480 George Khushf V. A FOURTH FRAMEWORK? EPISTEMOLOGIES AS COMPENSATIONS FOR HEIGHTENED RISKS OF ERROR The question is now how we can link these social epistemologies back to the knowledge that resides in the heads and hearts of individuals, and that is Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 treatments, or services that are commonly used, expensive, and that do not have a solid evidence base to support those uses. Blumenthal-Barby notes how multiple criteria for what counts as low-value care might be used. “Low value” might mean the test, treatment, or service has a small benefit, unlikely benefit, or inefficient benefit. These meanings have different implications for the way patients and physicians might be convinced to not utilize these things that have low value. But in the “Choosing Wisely” initiative, these value judgments are hidden from the public. Instead, we are just given lists, and these are presented as evidence-based judgments of the professional organizations. Blumenthal-Barby seeks to make more transparent the grounds for judgments of low value, so others might see the justificatory links between the claims and the evidence and value judgments that justify these claims. Her study thus makes transparent the complex social processes associated with any claims about what evidence-based practices should involve. In net effect, she seeks to take more seriously the language of “wise choice” in the ABIM initiative. With the extrusion of values associated with the ABIM initiative, it is hard to recognize the nuanced balancing of factors integral to practical wisdom. By attempting to make the social processes of justification explicit, and by providing a framework for the normative judgments about values, Blumenthal-Barby helps bring knowledge about the low-value care options back within the awareness of those epistemic agents who must make decisions about their use in the context of a clinical interaction. Jointly, the contributions by McClimans, Faust, and Blumenthal-Barby make clear how evidence and measurement, clinical causal reasoning, and prioritization decisions regarding tests, treatments, and services all depend on social mechanisms that are often incoherent and opaque. Each highlights a different kind of disruptive extrusion of an epistemic circuit: McClimans emphasizes an extrusion of perception, Faust an extrusion of reasoning, and Blumenthal-Barby an extrusion of values. But each also advances criticism directed toward creating conditions that are more favorable to an implicit, partially expressed ideal regarding how clinical reasoning and decision making ought to occur. They open the black boxes associated with measurement, causal reasoning, and prioritization decisions and work to make transparent and properly justify social practices that currently are at least partly incoherent and insufficiently justified. In doing this, they each express confidence that such critical scrutiny can lead to more reliable knowledge in the domains they consider. A Framework for Understanding Medical Epistemologies 481 Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 concerned with truth and justification. As the perceptive act moves outside individual perceptive awareness and into instruments like PROMs, Magnetic Resonance Imaging procedures (MRIs), and pathology laboratories; and as reasoning capacity moves outside of heart and head and into the information systems and professional organizations that organize the bits of available knowledge and develop guidelines and clinical pathways that inform clinical practices, questions concerning knowledge and its justification get more complex. When the diverse ends of medicine and the practical reasoning that advances them are considered, yet another level of complexity is added. With all this complexity, hopes for a synoptic framework for medical epistemology seem to get more distant. Instead, we just seem to have many provisional accounts of some part of an epistemic process, and any explicit epistemology is largely oriented toward providing a corrective to errors and problems that arise in the existing practices. With the extrusion of perception, rationality, and values into the social sphere, the plurality seems to become irreducible—we no longer have an epistemology but epistemologies. There are many and different social projects of construction that condition the way specific epistemologies settle out. With this fragmentation, we also seem to lose track of a single discourse on truth and justification. Our very effort to find an overarching framework for an integrated epistemological discourse seems to be disrupted by those same extrusions that disrupt what we have called the traditional, pure epistemology (Rorty, 1979, 8). This seems to suggest that the traditional epistemology and the effort to obtain a synoptic viewpoint on epistemologies are co-implicated or perhaps jointly dependent on some other condition. This condition seems to be undermined by social epistemologies which then draw us into the intricate details of specific epistemic processes. With McClimans, Faust, and Blumenthal-Barby, epistemological concerns seem to be disruptively extruded from the heads and hearts of individuals and inextricably intertwined with the specifics of measurement, causal analysis, and prioritization, as each of these are worked out in relation to their specific application domains and by their associated communities. When overwhelmed with the obvious relevance of these details for the questions being addressed can we say anything useful about epistemologies generally? Or do we need to just abandon our initial efforts at a synoptic view, concede irreducible plurality, and then situate any discourses of justification and truth in the appropriate social contexts where they are advanced? Does the social turn wipe out any prospect for the classical project when we are concerned with real-world practices like those associated with medicine? Before revisiting this question, we can make an observation about all of the contributions to this journal issue. None provides a synoptic vision of the epistemological practices they consider. Instead, each advances a corrective of more general epistemic practices that are largely taken as given and that remain implicit in their discourse. Even in this introduction, when the task 482 George Khushf Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 of providing a synoptic vision was presented, there was a gesture toward an empirical base: there were epistemic practices out there, taken as given, and addressed in some way by that literature. This, it turns out, is a pervasive feature of all contributions to medical epistemology. Even the most comprehensive efforts like those associated with EBM always start by taking some antecedent set of epistemic practices as given. They then involve an attempt to systematize and correct certain problematic aspects of the given practices, so errors might be avoided. This, in turn, implies that epistemologies are better understood as strategies for overcoming error, rather than as comprehensive accounts of how we attain justified, true beliefs. When we view epistemologies as general strategies for error correction, there is an interesting parallel between medicine and epistemology, one which George Canguilhem (1991) explicitly considered in one of the most influential contributions to medical epistemology in the past century. In the case of both medicine and epistemology, there is an ontic priority to the positive goods: health and knowledge. But in ordinary practices we are oriented toward these in unproblematic, taken-for-granted ways. When healthy, we don’t worry about health. We simply “live life in the silence of the organs” (this phrase is from Leriche, reviewed by Canguilhem, 1991, ch. IV). Similarly, when our epistemic faculties and practices exhibit to us the world in the natural way we normally encounter it, we don’t think about knowledge and truth. Instead, we simply know things as they are given to us in those ordinary interactions of life. In both cases, health and knowledge only become objectives of critical reflection when they are lost or otherwise at risk. Thus, although there is an ontic priority to health and truth, there is an epistemic priority to illness and error. Medicine and medical epistemologies are the result of our efforts to overcome these failures. Medicine targets illness in a form we can manage scientifically, that is, as disease. Medical epistemologies then target errors, and they do this in relation to the given practices already under development to target disease. Viewed in this way, recent systems oriented efforts by the US Institute of Medicine to eliminate error and advance quality might be regarded as social epistemologies (Corrigan, Kohn, and Donaldson, 2000; Kohn, Corrigan, and Donaldson, 2001). Medical epistemologies are then oriented to strategic error management, just as use of medical disease concepts are oriented to strategic illness management. Although the positive goods of health and knowledge are ontically first things, epistemologically they are the last, and perhaps we never will have anything more than a partial, explicit grasp of them. We can only come to an approximation of the positive goods after extensive efforts to map organismic and epistemic norms from the side of failures. But that shift to the positive end is extremely difficult to make, and it involves challenges that are only partially appreciated and addressed. What is optimal health and how might that inform preventive efforts to advance it? That question is strangely A Framework for Understanding Medical Epistemologies 483 NOTES 1. Steup (2012) provides a nice overview of philosophical epistemology; Korcz (2013) and DeRose (2013) provide annotated bibliographies. 2. Representative examples include Thagard, 1999; Solomon, 2008; Worrall, 2008; and Cartwright, 2011. 3. There are four conditions usually associated with knowledge: (1) justified, (2) true, (3) belief. The last condition concerns (4) fitness of a faculty of knowledge to an environment or context. This condition is needed to address what have been called Gettier problems (Gettier, 1963). Review of proposed solutions to the problem is given in Ichikawa and Steup (2013). 4. The language of “extrusions of thoughts from the mind” comes from Dummett, 1993, ch. 4. He is concerned with extrusions into language that do not deeply challenge traditional analytic epistemologies. The more disruptive kinds of extrusions are a central concern in cognitive science, where they are associated with embodied cognition. Clark (1998) provides a nice review of these more radical extrusions. ACKNOWLEDGMENTS This essay has greatly benefited from critical comments and suggestions made by Ana Iltis. Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 similar to the epistemic variant: what is knowledge and truth, and how are these advanced in the arena of medicine? We try to get at these questions by aggregating the hard won bits and pieces, but this effort at aggregation gets overwhelmed when we reopen the black boxes associated with earlier codifications and find the issues of justification were never quite cleaned up in the manner we supposed. In the end, it is not clear whether the diverse social discourses of justification might converge or not. Surely the discourse is complex, and there are multiple strands that need to be accounted for if any adequate account of optimal health or knowledge and truth is to be obtained. But can we infer from this complexity that the plurality is irreducible and the pursuit of knowledge and truth subverted? Such inferences too quickly move from our uncertainty to positive claims about what is possible, and this clearly overreaches the evidence and involves premature closure of the discourse about epistemologies. Marcum (2013) already observed that EI might involve one strategy humans have for managing complexity. There are of course other epistemic management strategies that also help us make sense of complex, convergent processes (Clark, 1998; Wimsatt, 2007). Here, Ashley Graham Kennedy’s recommendations to clinicians might be generalized: we need to suspend judgment and keep open the pathways of investigation that are nicely exhibited to us in the diverse contributions to this issue. Whether this discourse converges and, if so, how and to what—these remain open questions in philosophy and in medicine. With our effort to develop a framework for seeing the diverse contributions as part of a common project, we indicate our belief that the general project of an epistemology is worth pursuit, as are the associated goods of knowledge, truth, and practical wisdom. 484 George Khushf REFERENCES Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 Alston, W. 1989. Epistemic Justification. Essays in the Theory of Knowledge. Ithaca, NY: Cornell University Press. Ashcroft, R. E. 2003. Current epistemological problems in evidence based medicine. Journal of Medical Ethics 30:131–35. Baird, D. 2004. Thing Knowledge: A Philosophy of Scientific Instruments. Berkeley: University of California Press. Blumenthal-Barby, J. 2013. “Choosing Wisely” to reduce low value care: A conceptual and ethical analysis. Journal of Medicine and Philosophy 38:559–80. Canguilhem, G. 1991. The Normal and the Pathological. New York: Zone Books. Cartwright, N. 2011. The art of medicine: A philosopher’s view of the long road from RCTs to effectiveness. Lancet 377:1400–1. Cassell, E. 1985. The Healer’s Art. Cambridge, MA: MIT Press. Chapman, G. and F. A. Sonnenberg. 2003. Decision Making in Health Care: Theory, Psychology, and Applications. New York: Cambridge University press. Clark, A. 1998. Being There: Putting Brain, Body, and World Together Again. Cambridge, MA: MIT Press. Corrigan, J., L. T. Kohn, and M. S. Donaldson, eds. 2000. To Err is Human: Building a Safer Health System. Washington, DC: National Academy Press. Cutler, P., ed. 1998. Problem Solving in Clinical Medicine: From Data to Diagnosis. Third Edition. Baltimore: Lippincott Williams and Wilkins. DeRose, K. 2013. The Epistemology Page [On-line]. Available: http://pantheon.yale.edu/~kd47/ e-page.htm (accessed July 23, 2013). Dummett, M. 1993. Origins of Analytical Philosophy. Cambridge, MA: Harvard University Press. Eddy, D. M. 1996. Clinical Decision Making: From Theory to Practice. Sudbury, MA: Jones and Bartlett Publishers. Engel, G. 1960. A unified concept of health and disease. Perspectives in Biology and Medicine 3:459–85. ———. 1977a. The care of the patient: Art or Science? John Hopkins Medical Journal 140:222–32. ———. 1977b. The need for a new medical model: A challenge for biomedicine. Science 196:129–36. ———. 1988. How much longer must medicine’s science be bound by a seventeenth century world view? In The Task of Medicine, ed. K. White, 113–36. Menlo Park, CA: Henry J. Kaiser Family Foundation. Engelhardt, H. T., Jr. 1996. The Foundations of Bioethics. Second Edition. New York: Oxford University Press. Faust, H. S. 2013. A cause without an effect? Primary prevention and causation. Journal of Medicine and Philosophy 38:539–58. Feldman, R. 2012. Naturalized epistemology. The Stanford Encyclopedia of Philosophy [On-line]. Available: http://plato.stanford.edu/archives/sum2012/entries/epistemologynaturalized/ (accessed July 24, 2013). Fleck, L. 1979. Genesis and Development of a Scientific Fact. Edited by T. J. Trenn and R. Merton. Translated by F. Bradley and T. J. Trenn. Chicago: University of Chicago Press. A Framework for Understanding Medical Epistemologies 485 Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 Flexner, A. 1910. Medical Education in the United States and Canada. New York: Carnegie Foundation for the Advancement of Teaching. Foucault, M. 1975. The Birth of the Clinic: An Archeology of Medical Perception. Translated by A. M. Sheridan-Smith. New York: Vintage Books. Gettier, E. 1963. Is justified true belief knowledge? Analysis 23:121–23. Goldenberg, M. J. 2006. On evidence and evidence-based medicine: Lessons from the philosophy of science. Social Science & Medicine 62:2621–32. Goldman, A. 1999. Knowledge in a Social World. New York: Oxford University Press. ———. 2010. Social epistemology. The Stanford Encyclopedia of Philosophy [On-line]. Available: http://plato.stanford.edu/archives/sum2010/entries/epistemology-social/ (accessed July 24, 2013). Guyatt, G., D. Rennie, M. O. Meade, and D. J. Cook. 2008. Users’ Guides to the Medical Literature: A Manual for Evidence-Based Clinical Practice. Second Edition. JAMA Evidence. New York: McGraw Hill Medical and JAMA & Archives Journals. Howick, J. 2011. The Philosophy of Evidence-Based Medicine. Oxford: BMJ Publishing Group. Husserl, E. 2001. Logical Investigations. New York: Routledge. Ichikawa, J. J. and M. Steup. 2013. The analysis of knowledge. The Stanford Encyclopedia of Philosophy [On-line]. Available: http://plato.stanford.edu/archives/fall2013/entries/ knowledge-analysis/ (accessed July 24, 2013). Jenicek, M. 2003. Foundations of Evidence-Based Medicine. New York: The Parthenon Publishing Group. Jenicek, M. and D. L. Hitchcock. 2005. Evidence-Based Practice: Logic and Critical Thinking in Medicine. Chicago: AMA Press. Kassirer, J. P. and R. I. Kopelman. 1991. Learning Clinical Reasoning. Baltimore, MD: Williams and Wilkins. Kennedy, A. G. 2013. Differential diagnosis and the suspension of judgment. Journal of Medicine and Philosophy 38:487–500. Kohn, L. T., J. M. Corrigan, and M. S. Donaldson, eds. 2001. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academy Press. Korcz, K. 2013. The ERG Bibliography [On-line]. Available at: http://www.ucs.louisiana. edu/~kak7409/ERG_Bibliography.html (accessed July 23, 2013) Leder, D. 1992. A tale of two bodies: The Cartesian corpse and the lived body. In Body in Medical Thought and Practice, ed. D. Leder, 17–35. Dordrecht, the Netherlands: Kluwer Academic Publishers. Mainetti, J. A. 1992. Embodiment, pathology, and diagnosis. In The Ethics of Diagnosis, eds. J. L. Peset and D. Gracia, 79–93. Dordrecht, the Netherlands: Kluwer Academic Publishers. Malterud, K. 1995. The legitimacy of clinical knowledge: Towards a medical epistemology embracing the art of medicine. Theoretical Medicine 16:183–98. Marcum, J. A. 2013. The role of emotions in clinical reasoning and decision making. Journal of Medicine and Philosophy 38:501–19. McClimans, L. 2013. The role of measurement in establishing evidence. Journal of Medicine and Philosophy 38:520–38. Mills, A. E. and E. M. Spencer. 2003. Evidence-based medicine: Why clinical ethicists should be concerned. HEC Forum 15:231–44. Quine, W. V. O. 1969. Ontological Relativity and Other Essays. New York: Columbia University Press. Reiser, S. J. 1978. Medicine and the Reign of Technology. New York: Cambridge University Press. 486 George Khushf Downloaded from https://academic.oup.com/jmp/article/38/5/461/899502 by guest on 11 June 2022 Rorty, R. 1979. Philosophy and the Mirror of Nature. Princeton, NJ: Princeton University Press. Ryle, G. 1949. The Concept of Mind. London, UK: Hutchinson. Sackett, D. L., S. E. Strauss, W. S. Richardson, W. Rosenberg, and R. B. Haynes. 2000. EvidenceBased Medicine: How to Practice and Teach EBM. Edinburgh, UK: Churchill Livingstone. Salovey, P. and J. D. Mayer. 1990. Emotional intelligence. Imagination, Cognition and Personality 9:185–211. Schwartz, P. H. 2008. Risk and disease. Perspectives in Biology and Medicine 51:320–34. Simon, J. R. 2008. Constructive realism and medicine: An approach to medical ontology. Perspectives in Biology and Medicine 51:353–66. Solomon, M. 2008. Epistemological reflections on the art of medicine and narrative medicine. Perspectives in Biology and Medicine 51:406–17. Sox, H., M. A. Blatt, M. C. Higgins, and K. I. Marton. 2006. Medical Decision Making. Philadelphia: American College of Physicians. Steup, M. 2012. Epistemology. The Stanford Encyclopedia of Philosophy [On-line]. Available: http://plato.stanford.edu/archives/win2012/entries/epistemology/ (accessed July 24, 2013). Thagard, P. 1999. How Scientists Explain Disease. Princeton, NJ: Princeton University Press. Toombs, S. K. 1992. The Meaning of Illness: A Phenomenological Account of the Different Perspectives of Physician and Patient. Dordrecht, the Netherlands: Kluwer Academic Publishers. Wimsatt, W. C. 2007. Re-engineering Philosophy for Limited Beings: Piecewise Approximations to Reality. Cambridge, MA: Harvard University Press. Worrall, J. 2008. Evidence and ethics in medicine. Perspectives in Biology and Medicine 51:418–31.

RELATED PAPERS

RELATED TOPICS

Log In

A Framework for Understanding Medical Epistemologies

A Framework for Understanding Medical Epistemologies

Related Papers

RELATED PAPERS

RELATED TOPICS