Democratizing Conscientious Refusal in Healthcare

Abstract

Settling the debate over conscientious refusal (CR) in liberal democracies requires us to develop a conception of the healthcare provider’s moral role. Because CR claims and resulting policy changes take place in specific sociopolitical contexts with unique histories and diverse polities, the method we use for deriving the healthcare norms should itself be a democratic, context-dependent inquiry. To this end, I begin by describing some prerequisites—which I call publicity conditions—for any democratic account of healthcare norms that conflict or jibe with CR. Next, drawing on Ronald Dworkin’s jurisprudence and Tom Beauchamp & James Childress’s approach to bioethical reasoning, I briefly introduce one method for generating healthcare norms that is faithful to the publicity conditions and has potential to constructively, and democratically, derive important boundaries for CR. Finally, I argue that many critics of CR fail to similarly ground their accounts of healthcare norms in healthcare professionals’ sociopolitical contexts, often relying instead on their own interpretation of a generally stateable healthcare norm. This leads to their misconstruing both the value judgments on which their own approaches rest and the public, political values that are often invoked in favor of CR.

As a foreseeable result of the rapidly changing technological possibilities, social conditions, and legal landscapes surrounding healthcare systems, conscientious objection or conscientious refusal (“CR” hereafter) has generated considerable interest over the last two decades. CR occurs when a healthcare provider refuses, for moral or religious reasons, to provide a healthcare service that is (a) within the scope of his or her competence, (b) legally permitted, (c) desired by the patient, and (d) accepted as permissible by a substantial portion of the relevant medical community. Physician-assisted dying (PAD), for instance, has been legalized in some U.S. jurisdictions and recognized as a constitutional right in Canada. And more recently, the United States Supreme Court’s decision in Dobbs v. Jackson Women’s Health Organization, 597 U.S. (2022) leaves American citizens in a state of uncertainty about the legality (and availability) of abortion-related services throughout the nation. For each of these services, physicians disagree over whether taking part in them is obligatory, permissible, or immoral, given their various understandings of their professional, moral roles. Some providers understand the preservation of life to be their non-abrogable duty and decline to participate in PAD or abortion, while others take the intentional alleviation of suffering or the safeguarding of bodily autonomy to be central (or, at the very least, not antithetical) to their role. To provide a more prospective example, gene editing through technologies like CRISPR/Cas9 blurs the line between treatment and enhancement, inviting debates about whether the latter is an appropriate (or obligatory) medical practice. The question that CR presents in each instance is what we should do about these disagreements: allow providers to decline to participate in some services or else require them, upon a patient’s request, to offer the full array of services that have been legally and professionally recognized. Answering this question, in turn, depends on our own conclusions about the moral aims of medicine and the consequent moral duties of individual providers to the public. Thus, the first task in constructing an argument for or against CR is developing a conception of the healthcare provider’s moral role.

In much of the existing literature on CR, an argument proceeds in a way that seems appropriate to this task by: (a) Identifying one or more healthcare norms that are central to the provider’s moral role, and (b) Demonstrating why CR is either prohibited or permissible in light of those norms. Mark Wicclair (2011) has offered the following labels for the three types of answers that have taken shape: (1) Conscience Absolutism (“Ab”), proponents of which hold that providers ought to have virtually unfettered liberty to sincere CR invocations of CR, (2) The Incompatibility Thesis (“IT”), so-called because they find CR always or almost always incompatible with the healthcare provider’s role, and (3) The Compromise Approach, which includes various positions that endorse some restrictions on CR.¹ Conscience absolutism does not have many proponents in the philosophical and medical literature, but it has been represented in various legislative and regulatory measures in the United States.² Its proponents place a high premium on the value of moral and religious liberty among health care professionals and cite the long-existing liberties that professionals have in choosing who to accept as patients and which services to offer. While the precise rationale for IT varies, the following norms are often cited in support of influential IT arguments: duties inherent in medical practice to put the patient’s interests first (i.e., above one’s private moral interests) or foster the patient’s autonomy (by maximizing her healthcare choices),³ fidelity to a public oath to provide the full array of services within one’s competence,⁴ or the paramount social value of a predictable and stable healthcare system (which is disrupted by widespread CR and the vagaries of individual professionals’ consciences).⁵ In short, there is some determinate healthcare norm that the professional violates by imposing, through his or her CR, a tyranny of his or her subjective moral beliefs on the public or the individual patient.

A shortcoming of both IT and Ab accounts is that they omit or inadequately develop their method for deriving healthcare norms that conflict or jibe with CR. Such a method is necessary because debates and resulting policy changes about CR occur in a particular democratic context with its own history, healthcare system, and diverse polity. Any easily stateable healthcare principle (such as “respecting the patient’s autonomy” or “protecting professional integrity”) is subject to different interpretations and in tension with other widely recognized values in healthcare practice. Moreover, a healthcare principle’s meaning and weight might look different across areas of practice, types of medical interventions, and other circumstances. Thus, an account of the healthcare professional’s role that purports to be democratic—rather than merely an extension of the author’s worldview—must show its work; that is, such an account should show how its conclusions about healthcare obligations are rooted in its democratic milieu and how it resolves tensions between widely recognized values. Since IT and Ab accounts often trivialize the challenge of showing their work in this way, they end up resembling the sort of ahistorical (and undemocratic) approach that they ascribe to their opponents, insofar as they simply impose their conception of “good” healthcare on a morally diverse polity. One significant consequence of this is that these accounts often misunderstand (and misrepresent) the political value of concepts often cited in favor of (or against) CR.

My aim in this paper is to cast doubt on the possibility of any democratic account of healthcare norms (i.e., one that “shows it works” in the way I have suggested) yielding a context-independent and generalized answer regarding CR.⁶ This is, incidentally, a defense of the compromise position, which holds that some allowance of CR is the best means of balancing the competing interests implicated by CR.⁷ My critique of opposing positions will primarily focus on a family of IT accounts that fail to democratize healthcare norms, but only because this position is much better articulated and represented in the CR literature than Ab is. Nonetheless, the kind of methodological shortcoming this paper aims to correct is neither universal among IT approaches nor unique to IT: until recently, relatively little work on either side of the CR conversation has thoroughly engaged resources from democratic theory and jurisprudence in developing an account of healthcare’s normative structure.⁸ Thus, in addition to providing an account of what it takes to democratize healthcare norms, and showing why accounts from all three types of answers to the CR question fail to do so, I demonstrate the necessity of such resources for having fruitful conversations about the moral aims of medicine and CR. Section “The Publicity Conditions” offers a set of conditions that any (meta)normative account of CR must satisfy, while  section “Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms” will offer one such account, which draws heavily from Ronald Dworkin’s theory of jurisprudence. The institutional mechanism for employing such a framework is beyond the focus of the paper, though I suspect it is consistent with such proposals as the licensing board solution from Lynch (2008) or the Uber Conscientious Objection in Medicine Committee proposed by Ben-Moshe (2021). Section “Why Influential IT Accounts Fail to Democratize Healthcare Norms” will then explain why some influential IT accounts fail to democratize healthcare norms pursuant to the constraints outlined in section“The Publicity Conditions”.⁹

The Publicity Conditions

One point of ostensible, widespread agreement in the CR debate is that healthcare’s normative content should be constituted by and reflective of the democratic system to which it belongs.¹⁰ Both sides of the debate, however, begin with different conceptions of what it takes to reflect (or flout) the background democratic culture. IT accounts often assume that reflecting the public will is primarily about maximizing access to legalized medical services (often on the basis of the physician’s fiduciary duty or fidelity to their public oath); the public will is flouted where such options are narrowed by the idiosyncratic beliefs of an essentially oligarchic professional class.¹¹ Some compromise or Ab accounts, by contrast, presume (to differing degrees) that one necessary condition of reflecting the public culture is some degree of equal representation and freedom of conscience among the ranks of the professional class.¹² While both sides argue that the other fails to accord sufficient weight to a healthcare value (e.g., patient autonomy or the professional’s freedom of conscience), few accounts thoroughly ground their preferred value’s priority in a healthcare system’s democratic milieu. In other words, despite wide recognition that healthcare values are derived from democratic considerations outside of healthcare, most commentators spend inadequate time engaging one another at this level.¹³ As such, each commentator presumes their own conception of democratic faithfulness and talks past their interlocuters. Fortunately, there is a wealth of mostly untapped resources from democratic theory and jurisprudence, which recent CR work has begun to engage, that might help remedy this conversational impasse.

This section will provide a detailed explanation of what it takes for an account of healthcare norms to reflect the liberal democratic milieu (or fail to do so). In other words, I will provide an account of what it means to democratize healthcare norms relevant to CR. First, I identify relatively uncontroversial moral and epistemic features of liberal democracies. Next, drawing on resources from bioethics and democratic theory, I will offer some minimal requirements—which I will call publicity conditions—for any thoroughly democratic account of healthcare norms and CR; those approaches that cannot satisfy one or more of these conditions ought to be ruled out. Finally, drawing on resources from Ronald Dworkin and Tom Beauchamp and James Childress (“B&C” hereafter), I will sketch one approach that passes muster under the publicity conditions and offers some further boundaries on the use or restriction of CR.

An especially important sociological feature of liberal democratic systems for our purposes is what John Rawls called the “fact of reasonable pluralism,” according to which a diverse set of religious, moral, and philosophical doctrines is a “permanent feature of the public culture of democracy” (Rawls, 2005, 36).¹⁴ These doctrines often stand at the core of individual citizens’ identities and life plans. The explanation Rawls gave for this persistent doctrinal diversity is the “burdens of judgment,” which are the “many hazards involved in the correct (and conscientious) exercise of our powers of reason and judgment in the ordinary course of political life” (Rawls, 2005, 56). If indeed religious and moral pluralism is an enduring feature of liberal democracies, then we would do well to understand our most important social practices (such as healthcare) in light of this condition. Put otherwise, assuming Rawls is correct about reasonable pluralism—and I am unaware of any CR accounts that argue to the contrary—part of our task is to resolve the CR debate in a way that is cognizant of it.

Precisely how we should account for our compatriots’ plural worldviews is answered, in part, by a central moral feature of liberal democracies: an understanding of all citizens as free and equal moral agents or, at the very least, a willingness to treat them that way for the purposes of law and public policy. If our compatriot is a moral equal, and her life plan features identity-conferring moral beliefs and projects, then these facts about her should affect how we pass laws and enact public policies regarding our social institutions (or at least those institutions that affect her interests). This does not mean, of course, that our compatriot is always owed the public policy outcome she wants; this would often result in our privileging her priorities over those of other citizens. What it must mean, however, is that public policy outcomes should generally aim to do one or more of the following: (a) Be supportable by reasons that are intelligible to her, (b) Make reasonable accommodations for her view where doing so is feasible, or (c) Include a process that is otherwise fair and permits her voice to be represented. If a public policy process is devoid of such features, then the outcomes it produces would seem to simply reflect our own worldviews entirely at her expense. It would therefore seem to “browbeat” (Gaus, 1996,  123) her or treat her as a mere subject of legislation (see Talisse, 2009, 64) rather than an equal moral agent with a life plan of her own.

Insofar as this last claim invites any controversy, it is worth briefly reciting three reasons for accepting it. First, the assumption of equality is internal to the idea of a liberal democracy; to borrow another Rawlsian term, it is a “considered conviction” to which any tenable public policy proposal in a liberal democracy must conform.¹⁵ Second, a sincere commitment to democracy includes a desire to keep dialogues of public importance open which, as I suggest later, makes reasonable moral and religious pluralism within our social institutions necessary. If we are committed to truth-seeking in public life, and have sufficient epistemic humility to recognize that we are not the last word on controversial public debates, then certain of our most important practices should leave open the possibility that the moral majority is wrong after all. Indeed, in the tradition of J.S. Mill, much recent work in “new diversity theory” demonstrates the myriad ways in which all of society benefits from such diversity in our social institutions.¹⁶ As Roger Trigg (2017) has similarly argued in the CR context, our democratic commitments recommend the nourishment, not the permanent expulsion, of the moral minority on certain controversial medical practices. Third, reasonable citizens are interested in keeping beneficial social practices, like healthcare, stable and thriving. If we again assume that the fact of moral and religious pluralism is a permanent feature of liberal democracies, then our social institutions should be structured (and justifiable) such that they enjoy the support and reflection of a wide swath of the reasonable citizenry. Otherwise, we risk fomenting divisions that can destabilize these mutually and enormously beneficial social practices. In sum, we have moral, epistemic, and prudential (i.e., self-interested) reasons for public policy determinations that accord some weight to our compatriots’ interests, even where we find them wrong or misguided.

These general reflections on liberal democracies are instructive in fleshing out the democratic spirit in which we now undertake the task of deriving healthcare norms that are relevant to CR. Our aim is to construe “good” healthcare in a way that does not simply presume our own ideological orientation, but attempts to both argue from shared values and grant a hearing to others’ perspectives. At the same time, our manner of deriving healthcare values must be sufficiently rich that it is useful to the CR conversation. In other words, it must be able to specify what is obligatory, laudable, permissible, ignoble, or prohibited in professional conduct, especially insofar as these judgments reveal what we should do about various instances of CR.

The reason we first needed to characterize the “spirit” in which we should construe healthcare norms is that such norms within liberal democracies are not just derived from disparate sources, but are also often vague, incomplete, and in tension with one another. While professional codes of ethics and statutes governing healthcare are perhaps the most easily identifiable sources of professional obligations within a given context, such obligations are also derived from, to different degrees, regulations and other administrative acts, judicial precedent, institutional policies and procedures, hospital ethics committees, democratic principles, and even public opinion. Thus, with multiple sources and interpretations of healthcare norms available to us in many cases, we need a heuristic that ensures we are not just selecting the interpretation among these that aligns with our own worldview.

Another way of framing this is that, where a healthcare norm is indeterminate in its meaning or its importance, we need a way of both specifying and balancing the varied and pertinent norms. For example, a norm that is undeniably central in all democratic healthcare systems, like “do no harm,” is often ambiguous in meaning and scope. Thus, when applying it to a particular case, we must first engage in specification, which on B&C’s account means “reducing the indeterminacy of abstract norms and generating rules with action-guiding content” (Beauchamp and Childress 2013, p. 20). As B&C suggest, the generally stateable concept of “doing no harm” is, by itself, useless in determining the permissibility of physician-assisted dying. Even where a norm is sufficiently determinate, there often remains a tension between it and other applicable norms, as relatively few healthcare norms are absolute. This second difficulty requires balancing, pursuant to which we consider the strengths and weaknesses of each pertinent norm in the case before us, resulting in a judgment about which norm should prevail, and to what extent. Cases before hospital ethics committees often weigh the relative importance of respecting patient (or parental) autonomy versus protecting the patient from harm. In order to ensure that balancing involves an unbiased and rigorous reasoning process, as I have suggested it must in liberal democracies, B&C propose a variety of constraints on how balancing can take place in a fair manner, some of which inform the publicity conditions I have set out below.

To recap our task as we have summarized it so far: Given the fact of reasonable pluralism among free and equal citizens—together with the vague, disparately sourced, and often conflicting healthcare norms—we need a fair decision procedure for identifying relevant healthcare norms from our milieu and interpreting them in a way that tells us what to do in specific cases of CR. While this process will draw from principles within healthcare, these principles are themselves determined by public priorities outside of healthcare. Put otherwise, our construction of the “good” of healthcare requires the consideration of sources and principles that are facially non-medical. Pursuant to these observations, I offer four, minimal conditions that any account of healthcare norms must meet in liberal democracies. These conditions follow from features of liberal democracies I have identified so far and from principles of good argument and epistemic humility. In a Rawlsian vein, I will call these publicity conditions, though I intend them to be consistent with more models of justification than the Rawlsian idea of public reason. These conditions are not sufficient to construct a full account of healthcare obligations by themselves, but they are fruitful in identifying, and ruling out, undemocratic accounts.

First, we cannot build an assessment of religious claims or of conscientious objectors into the test we use to evaluate them. This might come about if one were to suggest that “a willingness to put one’s conscience aside” is the non-negotiable healthcare virtue that CR violates. Since the whole point of the CR debate is to demonstrate the propriety or impropriety of conscience-based exemptions, and whether one should put one’s conscience aside, we cannot simply exclude it by definitional fiat. In other words, the claim that “putting one’s conscience aside” is a virtue must be justified by reference to other, shared values (i.e., shared with those who doubt whether it is a virtue). Nor will it do to define medicine’s normative structure in a way that depends on a specific attitude toward religious professionals or religion itself, for that would also be to similarly beg the question about which attitudes belong in healthcare professions. After all, the consequence of debating CR in a liberal democracy—given the fact of reasonable pluralism—is that we are tasked with making our case intelligible to those with radically different religious attitudes than us.¹⁷ What these observations suggest, to put the first publicity condition succinctly, is that we must offer independent reasons for CR’s propriety or impropriety.

This first condition is like the idea of public reason that appears throughout the Rawlsian tradition, pursuant to which coercive state action should generally be supportable by reasons that all reasonable citizens could be expected to accept.¹⁸ Benjamin Zolf (2019) has argued that a policy allowing CR cannot be justified by public reason because CR claims are generally motivated by religious doctrines that make controversial claims. But this rests on a misunderstanding of how Rawlsian public justification works: the appropriate standard is not whether the religious professional’s worldview is accessible to the public, but whether there are accessible reasons supporting a policy that allows CR. Zolf also claims that the coercive state action that triggers public reason concerns is the licensure of individual professionals who have a particular belief on a controversial matter in healthcare. But if there are non-objecting professionals who that patient might seek out instead, it is not clear how they are coerced. Moreover, on this reasoning, would not the alternative—foreclosing the licensure of those with differing beliefs on a controversial matter in healthcare—be significantly more restrictive? While this is a somewhat tangential point for our discussion, I mention it because it is worth emphasizing that the idea of public reason, insofar as it applies to CR, works in both directions: a policy must be justifiable to all parties whose interests it might affect: patients, prospective and current healthcare workers, and the diverse interests of the citizenry.

Second, similar cases (i.e., those involving similar parties or competing interests) should be decided similarly. Many of the arguments against CR depend on the notion that there are potential harms to patients, the public, or the profession that are sufficiently contrary to good medicine that health professionals should avoid conduct that even risks bringing those harms about. Assuming there is such a duty (i.e., to avoid the risk and not just the occurrence of harm), then presumably any act or omission that creates risk, whether or not that act or omission is religiously or morally motivated, should be prohibited or discouraged. Many of the arguments in favor of CR depend on the notion that more religion in the public square is an unequivocally positive development. To borrow some pertinent terms from First Amendment jurisprudence in the United States, democratic norms about the refusal of medical interventions must be viewpoint-neutral and generally applicable, rather than targeting specific worldviews or motivations for refusal. In short, our account must be marked by equal treatment of various citizens’ worldviews.

Third, a normative foundation for healthcare, like any social practice, must be traceable to considered convictions that form part of our shared history. The four principles B&C identify—beneficence, non-maleficence, autonomy, and justice—are central to Western medicine, even if (a) we might sometimes disagree about the application and weight of these principles and (b) their relative ordering in our bodies of positive law change over time. Experiments at Auschwitz and Tuskegee provide paradigmatic examples of what is to be avoided when experimenting on human subjects, leading to the development of contemporary research ethics in opposition to those atrocities. Through EMTALA, the United States, like many other countries, guarantees universal access to emergency care regardless of ability to pay, a judgment that I take Paul Menzel to be quite right in deeming a fixed part of our moral fabric. Each profession has long-recognized specific, patient-facing duties—such as maintaining clinical competence, keeping patient confidence, or explaining procedures in an honest yet sympathetic register—about which it is difficult to imagine disagreement. Each of these examples is easily locatable within professional codes of ethics, enshrined in law, or the product of universal consensus among competent practitioners and public stakeholders. An account of healthcare’s normative structure ought to be supportable by principles that are universally shared (or very nearly so) and consistent with the evolutionary path our healthcare system has followed. In the following section, I will provide a framework that further illustrates how this condition might be satisfied.

Fourth, a democratically situated healthcare system, though it must have substantive moral priorities and aims, should not be blind to reasonable parties’ competing claims and interests. Recall that we are generally loath to coerce our compatriots for reasons that are inaccessible to them, as doing so treats them as mere subjects of legislation rather than autonomous (and equal) moral agents. In specific circumstances where multiple parties’ putative rights are at stake, we ought to avoid all-or-nothing policies where feasible, i.e., those in which one party is an outright or permanent loser. While political or moral principles often supply reasons to accord more weight to one kind of party (e.g., a patient or pro se litigant) or one kind of claim (e.g., violation of a fundamental right), this does not imply that the less-prioritized party is accorded no weight at all. Even where taking up a profession is a voluntary act, such that placing conditions on entering that profession is less problematic than it is for universally obligatory acts (like military conscription), the former can still be quite problematic.

The “reasonableness” qualification in this fourth condition recognizes that we cannot (and should not) include all conceivable perspectives in balancing affected parties’ interests. Our construction of healthcare’s normative structure need not be addressable to unreasonable persons. The unreasonable person, in Rawlsian parlance, is marked by a zeal to embody the whole truth (their version of it) in politics at the expense of others. To borrow an example from Jonathan Quong, we need not make our laws concerning theft intelligible to someone who claims a religious right to steal my computer, as the assertion of such a right contradicts the environment in which equal rights are possible (Quong, 2004, 333). The reasonable person, by contrast, is one who exhibits a willingness to (a) abide by fair terms of cooperation provided that other citizens do so, (b) recognize that citizens of good will might disagree on matters attributable to their worldviews, and (c) make their reasons for their political actions intelligible to others and be willing to listen to others’ reasons and accommodate them where feasible. Not every instance of living in accordance with one’s worldview, however, constitutes this kind of indifference towards others’ moral lives.¹⁹ Speaking to the matter at hand, there is a significant difference between the physician who declines to offer a specific service, yet refers potential patients to a physician who does offer it, and one who actively aims to prevent patients from obtaining that service elsewhere. Moreover, the objecting professional, or the Ab proponent who supports her, is not the only one who risks behaving unreasonably, an assumption which is implicit in many IT accounts. The policymaker who wishes to impose an excessive degree of moral uniformity on the healthcare profession, by foreclosing the possibility of dissent or pluralism, is also an unreasonable actor with an anti-democratic motive.

Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms

The publicity conditions are a preliminary test in determining precisely which of our judgments about good medicine could conceivably translate to universal obligations for professionals of a given type. Thus, I have offered them as an initial check on any method of deriving healthcare norms, even if they are too thin to supply such a body of norms by themselves. For a more substantive evaluation of whether various cases of CR fit the professional’s role, we require additional conceptual equipment, some of which I briefly sketch here and fully develop elsewhere. The Dworkinian, two-stage model I propose aims for an account of medical ethics that is internally coherent, fair, and historically grounded. It also does not purport to be the only appropriate model for deriving professional obligations. Thus, while the publicity conditions are offered as prerequisites to any construction of healthcare obligations in liberal democracies, this two-stage model is perhaps one of several conceptions that might satisfy them.²⁰

One of Dworkin’s most influential metaphors—that of the chain novel—might be helpful in providing an overall characterization of the model this section proposes. Dworkin asks us to imagine that Charles Dickens never wrote A Christmas Carol and that we have been furnished with the first part of the novel, with several authors having written these earlier chapters. It is then our task, as author of one or more later chapters, to make it the best novel that it can be, giving it a structure and purpose that it would have if written by a single author. In doing so, we are somewhat constrained by the data we already have before us—the chapters already written—and are cognizant of the guidance we are providing to authors of the chapters after ours. While it is likely the case that there are several ways of interpreting the themes, purposes, or currents of character development taking place in these earlier chapters, our choices are far from limitless if we want to give the novel a high degree of integrity. In general, if we are being faithful authors and literary critics, we will not leave unexplained some “dominant and repeated metaphor” or “a subplot treated as having great dramatic importance” from the preceding chapters (Dworkin, 1986, 230). At the same time, our chapter(s) need not give full effect to every minor element of the preceding chapters: certain elements can be identified as accidental or as mistakes in light of the novel’s overall purpose. Thus, if we were tasked with picking up the novel after Scrooge has already been visited by the three spirits and repented for his past callousness and cruelty, it would make little sense to begin the next chapter with the revelation that Scrooge is irredeemably wicked after all; the only interpretations available to us would seem to be ones that set Scrooge on a path to redemption of some sort. If “our” authored chapter took place earlier in the novel, however, then this interpretation of Scrooge’s character might still be available to us. In short, our job is neither a wholly mechanical one nor is it one in which we are to begin a new novel of our own: we are authoring a chapter that is the result of our faithful interpretation of the novel as a whole. Insofar as two interpretations are available to us, each of which render the text coherent and purposeful, we might then engage our own, substantive view about what makes for a good novel in choosing between those two.

Just as Dworkin suggests that the chain novel provides a fitting metaphor for the tasks of judges interpreting the law, so too does this section suggest it provides a fruitful way of constructing the normative structure of healthcare professions. When we ask whether a particular medical intervention is obligatory, or when we ask what the obligations of healthcare professionals are (perhaps in order to answer the first question), we are inevitably engaging in an interpretive process. Some rather particular obligations are set in stone or locatable in a singular source (e.g., a federal law), while others are more diffuse or amorphous, as recurring “themes” in our story. Where two different conclusions about a healthcare practice both explain the elements and direction of a healthcare system, we might then consult our own moral and political ideals in choosing between these two conclusions. In short, as this section will explain in greater detail, we want to tell a story about healthcare norms that is both faithful to the direction of the novel as told by its authors so far (healthcare professionals, political actors, citizens, patients, etc.) and that aims to make the healthcare profession(s) the “best” kinds of practices they can be. Following Dworkin, I will call the first of these stages “fitness” and the second “justification” (which Dworkin also refers to as “substance”).

Stage 1: Fitness

The first step in this method is to determine, in a way that builds on the work of the third publicity condition (a connection to our “shared history”), whether a proposed rule governing an instance of CR coheres or fits with the physician’s (or other professional’s) role. This determination, as Dworkin said of adjudication, is an interpretive act that aims to resolve a theoretical disagreement about medical practice. We ostensibly agree about the ultimate referent of the CR conversation—medical practice within liberal democracies—but disagree about what that practice requires. As I suggested earlier, this disagreement is not only a consequence of our conflicting worldviews, but the disparate and diverse sources from which we derive professional obligations. A proposed rule demonstrates fitness where it generally reflects, rather than conflicts with, the pertinent parts of these various sources.

The determination of fitness necessarily takes place at multiple levels of specificity, as each jurisdiction includes norms internal to democratic citizenship, healthcare generally, specific healthcare professions, and certain specialties or sub-specialties within those professions. Fitness does not require a proposed rule to explain every one of these sources, some of which might be anachronistic, more aberrant,²¹ less binding, or more controversial than others. But we ought to look with suspicion upon a putative norm that conflicts with or fails to explain several major sources of obligations in a given sociopolitical context, or that makes little attempt to reconcile existing sources of norms that are in a small degree of resolvable tension with one another. A rule that departs too much or too often from these sources is not an attempt to describe medical obligations as they are currently understood, but an attempt to describe an essentially new practice or impose one’s own conception of the good.²² Fitness is not only desirable for obvious democratic reasons, but also because we wish for a principled, purposive basis for our most important social practices, as opposed to what Dworkin refers to as arbitrary “checkerboard” solutions.²³

Suppose we are determining what an OBGYN’s duty is when it comes to either a long-legalized practice, like abortion, or a prospective one like genetic enhancement (in a world where services related to CRISPR are the province of OBGYNs). We might imagine a list of possible norms that describe the OBGYN’s obligation, including norms that take abortion or genetic enhancement to be: exceptionless obligations, generally applicable obligations with limited exceptions, permissible and non-obligatory services, or generally prohibited. When evaluating these candidate norms for fitness, we might begin with and give preference to localized considerations, as these bear directly on the question of what might be expected of individual physicians, as opposed to generalizing about all practitioners in all contexts. If the OBGYN specializes in maternal-fetal medicine (MFM) in the United States, we might look to the position statements and ethical guidance from the Society for Maternal-Fetal Medicine and any licensing or regulatory bodies that oversee MFM. Then, to a progressively lesser (but still significant) extent, we might turn to professional organizations overseeing all OBGYNs, all physicians, and perhaps even other, similar healthcare providers. Parallel guidance from other healthcare systems, especially liberal democratic ones, also deserves consideration as persuasive authority.

This sort of excavation will inevitably produce ambiguities, including the relative weight of certain sources of healthcare norms, that would need to be resolved through whatever administrative mechanism or quasi-judicial setting this procedure runs. For instance, how might we determine the content of certain core duties that apply to all American physicians?²⁴ Counting an American Medical Association’s position statement on an issue as being dispositive would exclude the voices of the overwhelming majority of American physicians, who are non-members and not otherwise bound or represented by the AMA’s views. On the other hand, turning to a simple majority of individual physicians’ practices would be inadequate for several reasons. First, it’s not clear that we can glean providers’ positions on certain types of interventions based on their provision or non-provision of that service: they might not offer a certain intervention out of a lack of interest or they might offer a service without deeming their differently-minded colleagues negligent for not offering it. Second, treating the practices or opinions of a simple majority of practitioners as dispositive would somewhat vitiate one of the oft-cited reasons for CR: ensuring that minority voices are not automatically barred from certain professional opportunities. Moreover, even where there is an obvious consonance between a professional organization’s position statements and a supermajority of practitioners, scenarios are foreseeable in which both sources differ dramatically from the expectations of a supermajority of the citizenry.

While it is outside the scope of this essay to address the various iterations of such scenarios, their inevitably comes with several implications for actually applying this methodology. A provider using this methodology as a guide for their own conduct might simply (albeit in good faith) consult their sense of what their profession’s consensus (or lack thereof) is. Employing this methodology to determine public policy, however, will require an adjudicative or quasi-adjudicative body that is actually in a position to examine the evidence from various sources and assign these sources weight in a non-biased manner. Fortunately, this task is far from foreign to policymaking in healthcare: much of medical malpractice law relies on the testimony of experts about the standard of care within various disciplines. Similarly, it might be that a committee—like Ben-Moshe’s UBCOM Committee or Lynch’s expanded role for licensing boards—setting CR policy ought to be comprised of representatives from the pertinent professional association, non-member professionals, patient advocates, and other stakeholders. In cases of consonance between a professional association (like the AMA) and the bulk of practicing physicians (as determined by an expert witness or representative) about the obligatory status of a particular medical intervention, then an especially strong showing from outside sources (e.g., patient or citizen advocates) would be required to dislodge the presumption in favor of its obligatory nature. Cases of dissonance might require the adjudicative body to look elsewhere (i.e., to more general principles and practices) for guidance in the fitness analysis.

In any event, the aim of “preferring local” sources is to construct an account of what is expected of this sort of practitioner in this sort of context. Only a particularized account like this, as mentioned earlier, can indicate which practices are core, penumbral, central, peripheral, permissible, ignoble, prohibited, or laudable within that area of practice. These vital distinctions are typically absent from most IT or Ab accounts and, insofar as we want to insulate our accounts against arbitrariness and bias, we want to ground them in socially locatable sources.²⁵ Moreover, we generally have good reason to suppose that those on the inside of a profession have knowledge and experience that warrants some degree of deference regarding services that are beyond public understanding or regular public concern.²⁶ As such, insiders’ voices are weighty in the fitness analysis without being dispositive or insurmountable.

Since the presumption for “local” considerations has a specific, democratic purpose, it is conditional and has limits. First, the “local” rule or convention might be so aberrant against the backdrop of less localized sources that the presumption in favor of the local is rebuttable. There may be special interests that afflict a professional organization or government entity, archaic rules that no one has bothered to revisit or correct as public morality has otherwise shifted, or other reasons why less localized sources offer better evidence of how society views that profession’s role. Second, a source can enjoy the same weight as a transparently local consideration where it is clearly intended to apply to all physicians, professionals, or citizens. Third, notwithstanding my use of codes of ethics or regulations as examples, these need not always enjoy preferred status over less explicitly codified norms from the political or medical milieu, especially where the latter are firmly established. The valorization of good bedside manner and admonition against patient abandonment, for instance, would seem to be on equal footing with various, more facially local sources. The specification of patient abandonment (i.e., which actions constitute abandonment), however, might look different in various specialties. Fourth, less localized considerations might be used where the more localized source is silent on the issue at hand, or where a particular understanding of the localized consideration puts it into a high degree of tension with other normative sources. Overall, the idea of local preference should reflect both the morally diverse public’s considered convictions, on the one hand, and the informed application of those convictions by those internal to each specialty, on the other.

The “preferring local” qualification, insofar as it aims to reflect the democratic culture in which medical practice takes place, necessarily includes a temporal component: what I will call the “preferring recent” qualification. Recent normative shifts—or rather, sources of normative obligation that are more recent in time—are generally weightier in the fitness analysis than contrary and preexisting normative priorities, common medical practices, professional rules, etc. As an obvious application of this “preferring recent” qualification, pre-1970s attitudes towards patient autonomy are clearly of much lesser weight, and perhaps no weight at all, given the myriad legislative, judicial, and professional changes of landscape that make patient autonomy central to contemporary healthcare. Sources of this shift in healthcare include not just well-publicized documents or landmark legal developments (like the Belmont Report of 1978, the Patient Self-Determination Act of 1990, HIPAA, Roe v. Wade, etc.), but numerous developments in informed consent law, hospital policies and procedures, medical education, and professional codes of ethics. As public policy and professional ethics became more protective of patient autonomy, this entailed a larger set of enforceable obligations on the part of providers. At the same time, technological developments expanded the possible scope of medical interventions beyond traditional and unanimously endorsed conceptions of healing. Consequently, these were the first decades in which a tension between patient and physician autonomy could have arisen, which led to some of the earliest protections for physician conscience. In short, the circumstances of the last several decades have seen provider integrity and patient autonomy both become entrenched in various legal and institutional sources of western healthcare systems. The same is true regarding an example mentioned in section “The Publicity Conditions”: EMTALA’s passage, the various practices that arose around its implementation, and its bipartisan popularity suggest it is much more an ineradicable part of our moral fabric than the preceding centuries in which healthcare systems included no such guarantee. It would be absurd to view these later measures as aberrations or departures from the healthcare story rather than new chapters that were now the main plot.

While a thorough fitness inquiry, for either abortion or CRISPR, is beyond the purpose of this paper, I will briefly comment on some likely starting points given the guidelines I have just laid out. A proposed rule for MFM specialists in the United States that would always obligate them to provide either service finds little (if any) any support within the norms and conventions of MFM, obstetrics, medicine, or other, related healthcare professions. Professional organizations and regulatory bodies have either explicitly supported CR (with restrictions) or, in organizations that have not, included general language defending the professional’s moral integrity. In terms of professional practice, an overwhelming majority of OBGYNs do not provide abortions and residency programs commonly accommodate residents’ requests to assist in no or only some abortions. Yet reproductive freedom is also deeply entrenched in the moral and legal framework of Western democracies. Abortion was recognized as a constitutional right in the United States for nearly a half-century. Even though the U.S. Supreme Court’s recent decision in Dobbs v. Jackson seems to have altered the legal landscape, it remains the case that: (1) The long-term legacy of Dobbs is unclear, (2) The Dobbs decision itself does not affect the right of various states to continue providing abortion-related services, and (3) There is opposition to the contrasting frameworks of Dobbs and Roe v. Wade among both the citizenry and medical professionals. This tension seems to come with two implications for a fitness inquiry in a jurisdiction that is relevantly like the United States. First, a general right to an abortion need not imply an absolute obligation for an individual physician; governmental entities or licensing bodies might be the most appropriate bearers of this obligation. Second, this suggests that cases like these simply do not lend themselves to generalized answers like those that IT or Ab proponents offer. In other words, it might be wrongheaded for either side of this debate, at least where a tension between two ostensibly entrenched principles exists, to ask whether there is a right to CR in general or not.

A nuanced fitness analysis can also reveal the folly of arguments that make CR an all-or-nothing affair, i.e., those in which either any imaginable conscience claims are admissible or none are. IT proponents often argue that opening the door to any CR claims at all results in a deluge of CR claims that are manifestly contrary to any tenable conception of medical practice. Alberto Giubilini (2017) has argued, for instance, that there is no principled way of distinguishing between a physician who regards the fetus as a person with inviolable rights from one who thinks the same of bacteria. Thus, if we were to recognize the former physician’s right to not provide abortions, so this argument goes, we could only exclude the latter physician’s CR claims on parochial or arbitrary grounds. But Giublini mistakenly assumes that the reason for accommodating certain worldviews has a principally metaphysical (and prejudicial) basis. On this understanding, allowing CR on grounds of fetal personhood, but not on bacterial personhood, would be either arbitrary or else contingent on the greater social capital of pro-fetus advocates.

If norms about professional obligation are derived from public moral culture, as I have suggested they are, then Giubilini’s claim falls apart. The use of antibiotics, elimination of illness-causing microbes, and curing of human bodies are all firmly entrenched components of medical practice across specialties and jurisdictions. While the concept of human beneficence evolved in twentieth and twenty-first century medical practice—to prioritize, for instance, mental health and palliative considerations—its pursuit has remained the core of all medical specialties. This is not a matter of mere majority belief, but both a quantitative super-super-majority and qualitative judgment about the sine qua non of healing. Whatever our ultimate judgments about CR, there is a clear difference in terms of fitness, and not simply substantive metaphysics or metaphysical neutrality, between treating the fetus as a patient (or at least a being of some medical priority), and behaving similarly towards bacteria. Among other socially locatable indicators of this difference, specialties devoted to fetal health and an infant’s (legal and medical) status immediately after birth attest to this. There is nothing, on the other hand, that is socially locatable regarding bacterial rights nor a political understanding of toleration that could warrant exposing patients to untreated bacterial infection.

By the same token, and even more transparently, a fitness analysis demonstrates the untenability of Ab in the healthcare systems of western democracies, especially the United States. Consider, for instance, a hematologist or critical care professional who declines to participate in blood transfusions. Whatever the value of freedom of conscience or the authority practitioners otherwise have in determining the scope of their practice, there are several reasons that allowing CR in this context would fit very poorly with (for instance) American healthcare’s normative history. First, blood transfusions are often performed in emergency settings where there is insufficient time to find a replacement for the objecting professional. As noted earlier, measures like EMTALA indicate the paramount importance that American healthcare places on guaranteeing emergency care. Second, and relatedly, American law and professional ethics—both within and outside of healthcare—includes myriad admonishments again patient abandonment and negligent rescue. In other words, there generally no duty to perform a rescue (as a citizen) or take on a client-patient (as a professional), but there is a well-recognized duty to see the job through once one has assumed care for that person. In an emergency care setting, there is an ongoing, immediate assumption of care taking place that sets it apart from other healthcare contexts. Third, the notion of wellness at which blood transfusions aim is an uncontroversial and, yet again, super-super-majority one, so much so that it is difficult to imagine clinical hematology or emergency care without it. The failure to expeditiously perform a blood transfusion (and doing so non-negligently), where medically indicated, is commonly deemed an instance of medical malpractice.

There are might well be healthcare contexts and patient circumstances in which abortion, or some other oft-contested service, shares one or more features with cases like this (i.e., those in which an instance of CR recognition has very poor fitness). But there are also cases involving a unique set of circumstances in which a provider’s objection to what are ordinarily “core” or commonplace interventions within their specialty or related specialties does fit. Whatever the outcome of particular cases, this fitness inquiry aims at ensuring that providers and policymakers are writing the next, coherent chapter of our healthcare story, and not embarking on a novel that is wholly their own conception.

Stage Two: Justification

Whatever direction the fitness analysis takes in a particular instance of CR, let us suppose several candidate norms survive a full fitness inquiry. At this point, we exhaust the descriptive account of medical obligation (i.e., the portion of the account that locates moral expectations of physicians and other healthcare professionals in socio-political sources). Thus, we have run out of clearly neutral ground and have little choice but to engage some of our substantive moral convictions in choosing among the remaining candidate norms. The purpose of this second step is to decide which of the remaining candidate norms do the best job of justifying the practice of medicine and making it a practice worth preserving. Judges, Dworkin observes of this second step, are often deciding between different political values—like justice, liberty, and fairness—in deliberating on the appropriate rule to decide the case, though all judges accord some weight to each of these values. One judge will interpret and balance these ideals differently than the next, and most judges will “think that the balance between the opinions of the community and the demands of abstract justice must be struck differently in different kinds of cases” (Dworkin 1986, 250).

For our purposes, in considering an instance of CR, we might take ourselves to be policymakers behaving like judges or healthcare professionals deliberating on their own obligations.²⁷ There are many healthcare desiderata at play in moral conflicts, such as B&C’s four healthcare principles or various principles of political morality (non-coercion, equal opportunity, respecting reasonable pluralism, social stability, etc.). After the fitness inquiry in an instance of CR, remaining reasonable disagreements might include conflicts of interpretation of a healthcare principle (such as non-maleficence), conflicts between a healthcare principle (like autonomy) and a principle of political morality (like preserving moral or religious diversity), and conflicts between principles of political morality (a predictable healthcare system versus one that is responsive to a morally diverse polity). This step leaves us some latitude, or what Daniel Sulmasy calls “discretionary space,” to determine precisely how to interpret and balance them in the wake of our fitness inquiry. I will offer two, brief examples that illustrate some limits and features of our reasoning in this second, justificatory stage, one which bears directly on CR and another that might prove analogically useful.

Though this stage is prescriptive, it remains constrained by the results of the first stage (fitness) in two ways: (1) The eventual rule must come from among the norms that survive the fitness inquiry, and (2) The rule should be justifiable in a way that is intelligible to the affected parties (the current patient, future patients or patients elsewhere, healthcare professionals, the public, etc.). It would therefore remain inappropriate (in nearly all imaginable circumstances) to engage one’s convictions in a way that justifies compelling treatment against a medically competent patient’s will, even if my moral judgments about balancing beneficence and patient autonomy warrant divergence from some of my colleagues on other issues. To borrow an example from Wicclair, suppose a family physician or nurse recently converted to Buddhism and believes that “pain is the working out of life’s karma,” (Wicclair, 2011, 93) and decides that interfering with his patient’s physical suffering is in fact a great spiritual harm. These would be inadequate grounds for conscientious objection to pain medication, Wicclair argues (I think correctly), because controlling pain through medication is not simply permissible and accepted, but a core goal of family medicine or nursing, and therefore obligatory.²⁸ There are circumstances involving related, palliative issues—like preserving quality versus quantity of life, the social impact and justice of certain pain management practices (especially those giving rise to opioid addiction), medical futility, etc.—that present tensions and ambiguities among healthcare goals for which moral discretion is appropriate. More generally, there are ongoing controversies, especially in light of growing technological possibilities in medicine, about certain understandings of “harm.” On my account, these circumstances differ from Wicclair’s Buddhist physician not due to moral or metaphysical beliefs I hold, but from the indeterminate results of their fitness (the first stage of analysis) inquiry.

Disclosure requirements within clinical trials offer another example in which there are actions that society has emphatically proscribed—such as lying to patients about the purpose of the research²⁹—and others about which there remains disagreement, including details of how to appropriately communicate risks and benefits to clinical trial participants. There remain disputes about the balance between benefits to society (i.e., future patients), on the one hand, and unacceptable risks to current patient-participants (regardless of their consent) or the extent of material disclosures required (to protect patient autonomy), on the other. Wherever this is the case, a fitness inquiry might not always yield a clear answer, such that considerations of political morality, the social purposes of medical research, medicine’s essential virtues, or the value of reasonable pluralism within medicine are appropriate considerations in (a) a physician’s deciding on a rule for her own conduct and (b) an adjudicative body’s deciding between complete uniformity among practitioners versus discretionary choice among clinician-researchers.

In sum, there are many cases for which the fitness inquiry determines the obligatory course of conduct for a physician, or at least narrows his or her permissible options. But in those cases where the fitness inquiry is indeterminate, the justification stage leaves us with numerous desiderata to balance in one way or another. What I will suggest in the following section is that some IT adherents take for granted the non-coercive or maximally democratic footing on which their accounts rest, as though they (unlike the CR proponents they criticize) leave their idiosyncratic, personal convictions aside. To the contrary, even assuming IT passes the fitness inquiry in each case of CR, it constitutes only one way of balancing the various healthcare desiderata, and one they offer few public reasons (pursuant to the first publicity condition) to accept over the alternatives. Put otherwise, IT accounts are neither worldview-neutral nor non-coercive to all stakeholders in CR policy. Their assumptions are often that maximal patient access to maximum services and near-total uniformity within the healthcare system outweigh other desiderata. Thus, not only do they leave other desiderata (or competing interpretations of these desiderata) entirely to the side, but they assume that the same balance of them obtains across jurisdictions, specialties, and individual cases. As we will see, neither of those shortcomings bode well for their accounts, insofar as they aim to be historically grounded and democratically responsive.

Why Influential IT Accounts Fail to Democratize Healthcare Norms

The primary purpose of sections “The Publicity Conditions” and “Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms” was to offer some democratic contours on any methodological approach to deriving healthcare norms (section“The Publicity Conditions”) and show how one such approach might yield fruitful answers to the CR debate (section “Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms”). A secondary purpose of section “Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms” was to cast doubt on the possibility of a democratic approach yielding acontextual or generalized answers to CR. In other words, it is doubtful that such an approach would lead to either the Incompatibility Thesis (IT) or Conscience Absolutism (Ab). This section will more directly demonstrate the ways in which these arguments often run afoul of the publicity conditions outlined in section “The Publicity Conditions”. This section will focus on a certain family of IT accounts not because Ab positions fare any better, but because Ab has virtually no representation in the CR literature. As such, I hope my expeditious dismissal of Ab in section “Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms” will suffice so that section “Why Influential IT Accounts Fail to Democratize Healthcare Norms” can better address the more formidable challenge to compromise accounts.

Though IT accounts often fail to employ a sufficiently democratic method for deriving healthcare norms relevant to CR, this does not mean that that the concerns which motivate these accounts are undemocratic. To the contrary, the foreseeable consequences of unregulated CR—such as patients being harmed by delayed or unavailable care—should be troubling to any conscientious citizen in a liberal democracy. Indeed, the CR literature (including accounts that are critical of IT) includes many unfortunate instances of such consequences, even in jurisdictions with restrictions on CR. Thus, many IT accounts are successful in both raising significant challenges for any CR-accommodating policy proposal and highlighting restrictions on CR claims that any democracy ought to recognize.

Where each of these accounts fall short, however, is in demonstrating that their well-motivated concerns about unrestricted CR require a near-blanket ban on it. There are three broad reasons for this, each of which can be traced to the absence of the sort of inquiry I introduce in sections “The Publicity Conditions”–“Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms”. First, IT accounts provide an inadequate appraisal of the value judgments on which their accounts rest, taking their proposals to be value-neutral (as opposed to the parochial reasons that justify CR accommodation). Second, they provide a superficial or inadequate account of the countervailing democratic values that support CR; they often suppose that CR is intelligible only as a self-serving practice. Third, their arguments typically depend on their interpretation of a generally stateable healthcare norm that they fail to locate in the democratic milieu(s) of the pertinent healthcare systems. In this section, I describe three influential arguments against CR before demonstrating how each of them commits one or more of these errors.

Fidelity to One’s Public Oath

On many IT accounts, CR violates fidelity to the public oath that healthcare professionals take, and on which the public relies, to act as fiduciaries for current and future patients. Upon completing his pre-professional training and entering a healthcare field, the healthcare professional assumes a general “obligation to place the well-being and rights of patients at the center of professional practice” (Stahl and Emanuel, 2017, 1382). Such a commitment includes both duties regarding his interactions with patients (e.g., keeping them informed and engaging them in a compassionate manner) and his clinical behaviors. Regarding the latter, the public expects professionals to maintain a sufficient knowledge base and skillset to discuss and, where they are medically indicated, competently perform the range of interventions that have been legalized and made part of their practice area. In most jurisdictions, there is an oversupply of persons interested in becoming healthcare professionals and willing to assume these duties. Thus, “those who are unwilling to participate in such expected courses of action should not join professions tasked by society with the provision of such services” (Savulescu and Schuklenk, 2017, 165). Those who are aware of such expectations when they join the profession and decline to live up to them when they claim CR, make a “mockery of their graduation promise” (Savulescu and Schuklenk, 2017, 165) and frustrate the public expectations on which their licensure is predicated.

While it is difficult to understand healthcare professions in liberal democracies except as a commitment of this general kind, this argument takes the extent or scope of this commitment for granted. IT accounts often assume that an intervention’s being legalized constitutes a moral judgment from the citizenry that its provision is morally important,³⁰ or that its subsequent acceptance by some professionals in a specialty makes it as much a “part” of that domain as the interventions its professionals already offered. Thus, an intervention’s legalization or medical acceptance makes it part of the package of services included in the professional’s “graduation promise.”

This assumption is unwarranted for two reasons. First, it is difficult to imagine how a sincere, general promise of this sort is possible in the lived experience of most healthcare workers. Professionals progressively discover what an area of medical practice involves (perhaps including its less savory and surreptitious elements), let alone their appraisal of it, as they advance through different stages of education, training, and professional experience. The physician might only gradually understand the nature of certain interventions or other practices that her colleagues perform, practices which are either very infrequent in her employment context or contrary to the conception of healing patients that her specialty otherwise pursues.³¹ She might reasonably take her commitment to apply to a breadth of other practices that her colleagues emphasize in their professional development, but not to this one.

An IT adherent might reasonably reply that such a practitioner, regardless of her subjective expectations, should have known that she was committing to offering this practice, as she is charged with fulfilling society’s expectations, not her own. But this reply leads to the more fundamental problem with this argument: it often cannot locate the content of this putative promise in the relevant jurisdiction’s milieu. The legalization of a practice is often not a moral approval of that practice, but an expression of society’s political values. Some legalized practices, it is true, also include both (a) the right of any citizen to access that resource or practice and (b) the corresponding obligation of some person or entity to provide it; but neither of these components necessarily follow from the bare fact of legalization. All that legalization of a practice—such as physician-assisted dying or genetic enhancement—clearly represents in a liberal democracy is that the practice is now permissible, and not necessarily obligatory or even encouraged. All that the subsequent performance of a legalized practice among some (or even most) practitioners represents is that this medical intervention does not violate good medicine, but this is distinct from a judgment that this practice’s omission violates good medicine. Indeed, it is commonplace in many jurisdictions to sub-specialize or otherwise not offer certain interventions for various, non-moral reasons.³² In short, in many jurisdictions, there simply is no “graduation promise” that includes a commitment to performing every intervention that has been legalized within one’s practice area. While certainly there is a commitment to offer a wide array of interventions and passionately pursue the patient’s wellbeing, each CR case presents the question of whether that commitment includes this particular intervention in this particular context; concluding that a general patient-facing promise, without more, extends to a particular case is question-begging.

The IT adherent remains convincing that we should be careful in carving out widespread medical practices or behaviors from the scope of the healthcare professional’s oath, which we want to be patient- and public-serving above all else. But the precise content of the professional’s commitment—and it is this precision that matters in CR cases—is an unavoidably context-dependent inquiry. Comparisons like Emanuel and Stahl’s, which claims that the obstetrician’s commitment to providing abortions is as self-evident as the hematologist’s commitment to providing blood transfusions, are unsupportable without such an inquiry.³³ To put the difference succinctly, only the non-provision of the latter would constitute sanctionable medical malpractice in most jurisdictions. The fidelity argument, while powerful in justifying restrictions on CR, cannot justify a context-independent ban on it.

Prevention of Harm to Patients and the Public

Even if I am right that health care professionals across jurisdictions cannot be understood to take a sufficiently determinate oath to preclude CR, some IT adherents will reply that this misses the heart of the matter. Since CR is a refusal to offer a service, each instance of it directly causes or risks harms to individual patients and public interests. If we can make any cross-jurisdictional claims about healthcare at all, it is that healthcare practices are unequivocally about the patient’s well-being. Harms to patients include both physical harms, insofar as the unavailability of timely care puts their physical health at risk, and violations of their sense of worth and autonomy, insofar as CR claims can shame them for seeking desired treatment, foreclose treatment options that the law has made available to others, or cause them to incur additional costs in seeking alternate providers. Zolf (2019) cites two particularly troubling examples of this: (1) The case of Cheppudira Gopalkrishna, whose delayed request for physician-assisted dying (PAD) resulted in his continued, bedridden suffering of ALS for months, and (2) The prevalence of physicians who object to abortion in Italy, which results in approximately 20,000 abortions being performed there annually. While Zolf does not mention this sort of case himself, the possibility of CR in response to terminating ectopic pregnancies, which can endanger the mother’s life, is even more troubling.³⁴ As a different kind of example, Ancell and Sinnott-Armstrong (who ultimately defend the compromise position) cite a case in which a same-sex couple is denied access to intrauterine insemination (IUI), causing them not just shame and indignity, but extraordinary expense in seeking alternative sources of treatment.

What compounds the unacceptability of these harms, according to IT, is that they are cognizable in terms of medicine’s aims, whereas the countervailing values supporting CR are not. A patient’s prolonged suffering (in Gopalkrishna’s case), avoidable suffering (in the unsafe, illegal abortion cases), or bearing additional expenses that other patients do not (in the IUI case) are all clearly explicable in terms of B&C’s healthcare principles or bedrock democratic values (like non-discrimination). By contrast, the CR claims that brought those harms about are justifiable only through religious or moral views that are fundamentally distinct from medical judgments; they depend on conceptions of spiritual harm or a particular metaphysical view of fetal personhood, for example. Private, idiosyncratic worldviews cannot be wielded, especially by persons in such an authoritative position, to cause harms that are cognizable from a public, clinical perspective. In fact, allowing such a result violates the very publicity conditions offered in section “The Publicity Conditions”.³⁵

Avery Kolers (2014) offers another compelling version of the private and public harms argument (hereafter, simply the “harm” argument), as he makes a concerted effort to proceed from shared political values. There are features of healthcare systems in most liberal democracies that all reasonable citizens recognize as morally important, including their ability to foster the autonomy of vulnerable persons (i.e., patients) and stably pursue other aims that society has deemed important. These features have not come about, nor will they remain, easily. Instead, they depend on “multiple autonomous agents” (i.e., healthcare professionals) coming together “under a system of rules to achieve morally important results that none of them could achieve by flouting those rules or by acting on their own or in a different context” (Kolers, 2014, 4). Thus, wherever a healthcare professional declines to participate in an intervention ordered by her superior, or where she elects to create her own, limited package of medical interventions that she offers—and does not offer interventions offered by her colleagues—this detracts from the high degree of cooperation a stable healthcare system requires.

If such behavior where universalized, of course, this would result in disorder (i.e., an unpredictable healthcare system offering varied services) and a failure to foster vulnerable patients’ autonomy. To avoid such an outcome, individuals working in healthcare “cannot be unaccountable, and cannot be allowed to subvert the organization’s agenda or throw it into disarray” (Kolers, 2014,  4). Thus, to avoid these risks and serve the public interest, we rightly expect healthcare professionals to not just grudgingly submit to its practices, but to identify with the moral aims of the profession. Those who would seek to invoke CR, by contrast, live in a constant state of tension with the profession and constitute an ever-present risk of instability.

The various iterations of the harm argument make a convincing case for certain restrictions or preconditions for CR claims, many of which have been discussed elsewhere. What is important for our purposes is that such restrictions can be justified democratically, or rather, through the reasoning process I outline in sections “The Publicity Conditions”–“Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms”. First, there should typically be no right to CR when it comes to emergent care and certain other, time-sensitive interventions. To hold otherwise likely violates all four publicity conditions outlined in section “The Publicity Conditions”, but especially the fourth, which requires a balancing of the parties’ interests where feasible (see, supra, p. 7). This would accord no weight to patients’ interests at all, as it is markedly different from circumstances of CR in which patients have the opportunity to seek out an alternate provider at little or no cost (and which therefore allow for meaningful balancing of the affected parties’ interests). This consequence would be difficult, if not impossible, to square with what the IT proponent is correct in describing as a public-serving social institution.

Second, for similar reasons, the publicity conditions would seem to require objecting professionals to provide a degree of advance notice to patients, and perhaps even a duty to refer patients to nearby providers who do provide that intervention. While the literature contains entire debates about these options that I do not have space to address here, the publicity conditions would seem to require that affronts to patient time, energy, and expense are minimized. This might also include a governmental entity or healthcare employer bearing the cost of a patients’ having to seek alternative providers after being referred elsewhere.

Third, any given practice area should be understood to have a substantive view of “good medicine,” complete with distinctions between core interventions under its purview that are obligatory for its physicians to provide versus those that are merely permissible or penumbral to the practice (and about which there is intentional room for a difference of opinion among colleagues). The trouble with IT, as I will further explain momentarily, is not in supposing there is an objective conception of good medicine within practice areas or jurisdictions; as the third publicity condition (comportment with shard history) suggests, resolution of any public issue needs such starting points. The trouble consists in IT accounts’ failures to derive their conception of good medicine from socially locatable sources or take heed of distinctions between permitted or obligatory conduct within practice areas. While these are not the only conceivable limitations that an inquiry along the lines of sections “The Publicity Conditions”–“Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms” might produce, they are three limitations that would seem to cross jurisdictional and other contextual boundaries in all liberal democratic healthcare systems.

Notwithstanding their successful justification of limitations on CR like these, no iteration of the harm argument warrants a blanket (or near-blanket) ban on CR claims. The first reason for this is that, as Maclure and Dumont (2017) have argued, no IT account has ruled out the alternative of managing or limiting CR claims instead. Supposing that there are several ways of managing CR claims without causing patients significant harm, the fourth publicity condition (balancing parties’ interests) recommends the option that accords some weight to CR claimants’ interests. This is especially true if, as I argue below, CR claimants’ interests overlap with public interests. Moreover, while physicians play an important role in ensuring that patients have access to any care to which they are legally entitled, this burden is also shared with the government and other entities. The burden might instead fall on professional associations, state health authorities, or large healthcare employers to ensure that various regions are adequately staffed with professionals who are willing to provide the full range of services. Indeed, this is already the basis on which Certificates of Need are issued in the United States, whereby state or federal agencies ensure that new healthcare facilities meet the healthcare needs of the surrounding community.³⁶ The trick, for democratic purposes (and pursuant to the first and fourth publicity conditions), is picking the burden-sharing route in patient access that is least coercive to all involved, a challenge that IT accounts typically show little interest in meeting.

Proponents of the harm argument might reply that there is simply no good public reason to devote time and resources to managing CR claims, since they are only justifiable by private values that are immaterial to clinical decisions. This brings us to the second reason that the harm argument fails to justify IT: their mistaken assumption that worldview-neutral, clinical judgments are neatly separable from the parochial, moral ones that motivate CR. As both Cowley and Sulmasy have similarly argued, many CR claims are not simply motivated by extra-clinical or self-serving judgments; rather, they are motivated by the objecting professional’s clinical judgments, or more broadly, her understanding of the appropriate ends of medicine given her obligations to patients and the public. Admittedly the range of permissible interpretations of one’s clinical obligations must be constrained within each specialty, but the outer limits of that range—pursuant to the first and second publicity conditions—cannot be determined by the professional’s religious status. Even where such interpretations are religiously motivated, the most common types of CR claims (such as those involving abortion or PAD) are explicable in secular moral terms. The burden is on the IT proponent to explain why a particular CR claimant’s interpretation is outside this permissible range (from the relevant specialty’s perspective), which would require some recourse to the kind of procedure I have proposed in sections “The Publicity Conditions”–“Beyond the Publicity Conditions: A Constructive Approach to Deriving Healthcare Norms”.

In fact, it is commonplace for clinical judgments to include both technical and moral components in other contexts. In some circumstances, there is intentionally no “official” or “preferred” interpretation of what one’s clinical duties require: there is intentional room for discretionary judgment. When it comes to judgments of medical futility, for example, critical care physicians’ judgments vary, even where they agree on the purely technical facts of the matter. Judgments about whether a particular procedure is worth performing, or who should receive priority of care, also appear in organ transplant committees’ decisions globally and in the United Kingdom’s judgments about which care ought to be publicly funded. What is most telling of all, however, is that IT proponents themselves defend the propriety of moral judgments in ways that closely resemble CR. Savulescu and Schuklenk (2017), for example, argue that it would be acceptable for a physician to attempt to talk her patient out of IVF—and instead opt for adoption—on the grounds that the latter better comports with distributive justice. Since their objection to CR is, at least in part, grounded in their claim that professionals should not interfere with or coerce patients’ informed healthcare decisions, this is a surprising recommendation. It is unclear why, given their arguments in favor of IT, it is the province of healthcare professionals to make judgments about which interventions comport with their private conception of distributive justice. My suspicion is that this constitutes their implicit recognition of the place of moral perspectives in clinical judgment: they simply object to those with a religious source.

Portions of Kolers’s argument arguably survive this initial reply, as his account does not as assume the healthcare profession has a morally neutral view of harm, but instead grounds his account in the shared, political value of maintaining a stable, predictable, and publicly beneficial healthcare system. But there are value-judgments that even his iteration of the harm argument takes for granted. First, he seems to assume that a professional’s “identifying” with her profession’s “moral aims” necessarily means that she offers the maximal range of legalized interventions. As I have suggested, however, the moral aims of the profession often intentionally include the opportunity for continued disagreement over some interventions, particularly those that are controversial or novel. Moreover, it is not clear that we want a class of professionals who identify with all the practices common to their profession at a given time, for certainly we want motivated, conscientious healthcare professionals (rather than indifferent, conformist, or enterprising ones). It would be a surprising coincidence if a class of conscientious professionals did not produce vociferous disagreement on certain important matters at any given time. Second, while a stable and predictable healthcare system undoubtedly qualifies as one publicly intelligible value, it is far from the only one. It might be the case that some degree or risk of instability or non-uniformity in professional conduct is permissible on the grounds that it promotes other values, or even that it promotes a more enduring stability than the sort that aims at professional uniformity. Third and finally, it seems that stability, which on Kolers’s account is attained through uniformity in practices, is instrumentally, not intrinsically, valuable. In other words, he takes integrity (and stability) to be valuable because of the goods that such a healthcare system affords us. If there were an effective system for managing CR claims, however, then the instrumental value of stability would no longer constitute an objection to CR.³⁷

Moral Consensus Within Healthcare Professions

A third argument worth examining is better understood as a reply to a value that is invoked in defense of CR: religious and moral diversity among healthcare professionals. The claim that diversity in our social institutions is valuable has been mostly implicit in much of my discussion, though section “The Publicity Conditions” offered some general reasons supporting it (see Kolers, 2014, 4, 19–20). Against this claim, IT proponents have argued that there are affirmative reasons for circumscribing the degree of moral diversity within healthcare professions. We have already encountered an instance of this in one strand of Kolers’s argument. On his account, it is desirable to have sufficient moral uniformity among healthcare professionals for each of them to “identify” with the profession’s moral aims. This is a powerful argument with a Rawlsian legacy. Stephen Macedo (1998) has similarly argued, regarding religiously devout judicial appointees, that their public-serving vocation is a special one that requires some sacrifice of their personal aims in pursuit of distinctly civil interests. Thus, it is inevitable (and even desirable) that we will progressively find a particular range of worldviews within the judiciary and healthcare, especially one that might not reflect that of the citizenry.

Other IT accounts are even less sanguine about the value of worldview diversity among healthcare professionals. In response to the claim that moral and religious diversity is in the public interest, Savulescu and Schuklenk (2017, 164) have argued that diversity should be reflected in the public sphere outside of healthcare, but not in the conduct of healthcare professionals. While they acknowledge that matters of controversy should remain open to public disagreement from those of diverse perspectives, these disagreements should be aired and resolved legislatively. All that should happen within healthcare, they claim, is the provision of all the services to which patients are “legally entitled” (Savulescu and Shuklenk, 2017, 164).

Each of these arguments undervalue professional diversity because they undervalue the influence that professionals have on issues of public concern and the ways that professional diversity serves the public interest. Patients of various worldviews may benefit immensely from being able to seek out providers whose values resemble theirs, whether that is because such a provider is better equipped to foster their autonomy by helping them make important medical decisions (given their moral or spiritual priorities) or because the patient otherwise benefits from certain types of communication that only likeminded professionals might offer. Healthcare professionals are highly influential when it comes to initiating and advocating for legislative change, determining the priorities and direction of healthcare funding, leading and developing new areas for medical research, blowing the whistle on unethical practices within healthcare, and providing insight on the weight and urgency of health needs faced by the populations they serve. They also form part of institutional ethics committees, regulatory bodies, and professional associations. More broadly, the culture of healthcare itself is, and will always be, more than the regulatory and legislative measures that bind it: the moral and epistemic milieu of the field is largely shaped by those within it. For similar reasons, Rawls’s later work recognizes the vital role that the morally and religiously diverse “background” institutions of society play in forming the overlapping public consensus of diverse views that legitimize our system of government. Thus, not only do individuals of various background have an interest joining the ranks of the healthcare profession (an interest which is protected by the political value of equal opportunity), but various population groups’ interests—including the vulnerable patients themselves—are affected by the composition of these professions.

The morally and religiously diverse public is also interested in healthcare having certain priorities and avoiding harmful practices, interests which are best served through a diverse professional class. The many limits of the legislative process mean that these matters are, in large part, entrusted to those who join the ranks of these professions. To suggest the relegation of moral and religious diversity to legislative conversations outside healthcare is to propose an impotent, even illusory, sort of diversity. It is also to render the profession dangerously uniform, as those who dissent from now-permissible practices are permanently foreclosed from entering it and serving as, among other things, a humbling check on the sometimes-unwarranted confidence of professional majority opinion. Both Sulmasy and Weinstock (2014) have, for this reason, argued that even those citizens with no qualms about current medical practices benefit from a diverse perspectival presence. What ultimately causes IT’s misappraisal of diversity’s value, as I suggested earlier, is their failure to take adequate stock of the plural values that are pertinent to deriving healthcare norms. When the professional objects to a practice, it is often far from a selfish, myopic concern for one’s own salvation, but is instead a judgment about precisely what is good for the patient, the public, or the profession’s future.

Conclusion

I have argued that many IT accounts, though they are well-motivated and persuasive in setting important limits on CR, fail to adequately ground their conclusions in their democratic milieus. The contrary position, Conscience Absolutism, even more transparently ignores various norms at the heart of healthcare professionalism. The CR debate should instead seek a framework for deriving healthcare obligations that is fair and sensitive to sociopolitical context. I have given reasons to believe that such analyses will tend to favor some version of the compromise approach, since its many iterations eschew generalized and context-independent answers to CR. This is not, as some IT accounts suggest, tantamount to moral relativism; it is grounded in liberal democratic values that recognize both the importance and limits of enshrining our own judgments about good medicine in law and public policy. In response to the CR question, we ought to be able to ground our position in one or several healthcare norms, and in response to the question of where those norms come from and how much they weigh, we ought to be able to give yet another answer, one that is grounded in something other than a bare expression of our own moral or spiritual worldview. Our answer, to invoke Dworkin’s chain novel analogy once more, should be in the spirit of joint authorship.

Such an account of medical obligations is just as pressing in conversations tangential to CR, especially in an era that increasingly emphasizes social consciousness and public responsibility of healthcare professionals, especially in accounting for social determinants of health in their professional lives. Recently, the COVID-19 pandemic has highlighted the sacrifices and risks we ask of health professionals, as well as the extent to which their expertise might inform limits on citizens’ behaviors well outside the traditional domain of healthcare. If our moral expectations of healthcare professionals continually, even if gradually, shift under our feet, then we must agree on a means of determining what they are at any given moment, especially if such expectations are to be enshrined in law and public policy.

I have also emphasized, in considering the myriad technological possibilities on the horizon, that healthcare professionals play an important role in determining the priorities and aims of our health institutions. Our history is already rife with examples of medical services whose widespread use or funding prioritization led to various forms of public harm, such as the opioid crisis in the United States. Now more than ever, medical research and practice might frequently find itself at numerous crossroads, in which various prospective medical practices or research directions might irreversibly shape our public health institutions for better or worse, and even alter what “better” or “worse” might mean in each case. While important decisions along these lines take place among the voting public, IT and Ab arguments often miss sight of two important considerations: (1) The public interest in intentionally leaving room for moral diversity and discretionary judgment within healthcare, and (2) The extent to which these decisions are unavoidably influenced by the healthcare community itself. Thus, for reasons of caution and democracy, we would do well to both construct our understanding of medical obligations carefully and ensure that we are on broadly justifiable ground wherever we restrict the prospect of dissent (via conscientious refusal) within our healthcare communities.