Introduction

Discussions of the proper role of conscience and practitioner judgement within medicine have increased of late, and with good reason. The cost of allowing practitioners the space to exercise their best judgement and act according to their conscience is significant. Misuse of such protections carve out societal space in which abuse, discrimination, abandonment of patients, and simple malpractice might occur. These concerns are offered amid a backdrop of increased societal polarization and are about a profession (or set of professions) which has historically fought for such privileged space (Starr 1982) and possesses a special status (which might allow for an extensive degree of autonomy in the exercise of judgement) that is in need of justification (Buchanan 2009).

There is a great deal that has been and might yet be said here, but in this paper I aim to address one recent thread of this discussion: justification of conscience protection rooted in autonomy. In particular, I respond to an argument from Greenblum and Kasperbaur (2018) and clarify a critique I offered (2016) of an autonomy-based conscience protection argument which Greenblum and Kasperbaur seek to improve and defend. To this end, I briefly recap the central contention of that argument (from Aulisio and Arora (2014)), briefly describe Greenblum and Kasperbaur’s analysis of autonomy and of my critique, and correct what appears to be a mistake in interpretation of both my work and of autonomy-based defenses of conscience protection in general.

Traditional and Contemporary Accounts of Conscience

Aulisio and Arora thoughtfully engage the tension between protecting practitioners’ consciences and respecting the rights of patients to be well informed and to receive care according to the best standards of practice. In “Speak No Evil? Conscience and the Duty to Inform, Refer or Transfer Care,” published in HEC Forum in September 2014, they argue that there is a difference between being complicit in evil and performing an evil action and this distinction is reflected in the differences between traditional and contemporary accounts of conscience protection. Traditional accounts of conscience protection fit well with cases where an agent seeks to avoid the performance of an evil action, whereas the contemporary accounts of conscience protection involve concerns that a practitioner might be complicit in an evil act. They claim that the contemporary account depends upon an account of autonomy, which they understand as acting according to one’s own values without imposing one’s values on others. Additionally, they argue that autonomy is not a sufficient foundation for the contemporary account of conscience protection because it undermines the very claims that appeals to contemporary conscience protection are meant to protect. Protection of practitioner consciences is rooted in autonomy, and consequently practitioners must respect a patient’s autonomy, which on their view only supports the traditional account, or the practitioner loses her justification for what protected her conscience in the first place. The theoretical conclusion is that only traditional accounts of conscience protection are justifiable; the practical conclusion is that practitioners must refer patients directly for particular services that they seek if the practitioner does not wish to satisfy such requests.

In 2016, I argued that Aulisio and Arora’s argument in favor of traditional conscience protection fails because their account requires the adoption of unrealistic views of a person’s life and self-conception.Footnote 1 It does so in at least two ways. First, it reduces interactions between persons to a list of values that might clash with each other; persons are more complicated than this analysis admits. Second, it rules out without argument accounts of moral responsibility which might reasonably be held by persons; how persons might reasonably view themselves to be interacting in the world is more complicated than that argument allows. Though Aulisio and Arora’s argument fails, their inclusive approach is virtuous in at least two respects. First, their argument took seriously both patients and practitioners in attempting to address a tension between them. Second, in paying careful attention to liberal democratic social norms in their account of autonomy, they took seriously what might be described as the importance of reciprocity in relationships, in particular the relationship between patients and practitioners. Though the first point requires little elaboration, the second is in need of elucidation.

Aulisio and Arora describe the concept of liberty, at least in part, as allowing for “maximal play to live in accord with values while not avoiding values” (2014, 260). Influencing this discussion is surely the work of John Rawls, though it is not appealed to explicitly in their argument. Consider, for example, the first principle of justice famously articulated by Rawls, “[E]ach person is to have an equal right to the most extensive scheme of equal basic liberties compatible with a similar scheme of liberties for others” (Rawls 1999, 53). One way to frame Aulisio and Arora’s project is as an attempt to highlight the importance of reciprocityFootnote 2 in the practitioner–patient relationship. Their claim is, essentially, if physicians wish to have their autonomy respected, then they ought to do the same for patients. It would be unfair or a violation of the relationship to expect more from one party than another. Given that the foundation for protection of practitioner conscience is respect for autonomy, that is, the allowance of a maximal play of values, presumably for all parties to the relationship, that foundation could not support respect for the values of one party and not the other.

In spite of these virtues, the argument fails to ameliorate the tension between the expression of practitioner consciences and respect for patient values because their accounts reduce conceptions of the moral life of persons to a mere list of values and moral conflict to adjudication by the prioritization of liberty, neither of which has sufficient argumentative support.

Autonomy Revised

In a paper published in this journal, Greenblum and Kasperbaur (2018) revisit Aulisio and Arora’s attempt, strengthening the account of autonomy and drawing new implications. Greenblum and Kasperbaur take Aulisio and Arora’s liberal political approach and revise the account of autonomy such that it fits more squarely within the standard conversations in that field, explicitly citing Rawls, among others. They write:

Although the right to live in accordance with one’s values is unquestionably morally important, autonomy consists of much more than this. In particular, autonomy requires independence or the capacity to independently set goals for oneself, which practically requires a sufficient understanding of one’s situation and a sufficient level of freedom from controlling influences. (2018, 315)

This account of autonomy, with a focus on independence, allows in better resources to defend the importance of, for example, fully informed consent, which Aulisio and Arora are in favour of, but which—according to Greenblum and Kasperbaur—they cannot defend:

This broader notion of autonomy can help explain clear autonomy violations in ways that Aulisio and Arora’s account cannot. Consider a physician refraining from giving a patient medically relevant information. Suppose the physician is pro-life and knows a pro-life patient will choose to carry her fetus to term and so refrains from informing this patient of her right to abort. According to Aulisio and Arora’s value-based understanding of autonomy, there is no violation of autonomy since the patient chooses in accordance with her values. However, it would seem that the patient’s autonomy has indeed been violated because she was not made aware of the full range of legitimate medical options available to her. The notion of autonomy as independence can explain why this is so: the patient lacks key information necessary to understand her situation and failure to​ ​inform is a controlling influence on her decision (315).

Greenblum and Kasperbaur’s broader account of autonomy follows Savulescu (2006) and others in maintaining that practitioners’ medical advice must adhere to publicly defensible norms and that patients are entitled to predictability, both of which would not be satisfied if practitioners were allowed to conscientiously object.

Greenblum and Kasperbaur’s account, like Aulisio and Arora’s, is virtuous in at least two important, albeit distinct, ways. First, they offer a much more sophisticated account of autonomy in light of their reliance on liberal political philosophy. In this way their position might avoid my original objections to the account offered by Aulisio and Arora. Second, Greenblum and Kasperbaur are open about the fact that physicians will bear a cost, and a significant one, if their account is supported. To the second point, they write, “​Patients, by contrast, enter their physician’s office with an understanding that the physician’s professional judgement, not their privately held values, will determine the advice they receive. Thus, the value imposition on physicians is not only permissible but in some cases necessary for adequate care” (315). Thus, they have alleviated the aforementioned tension between patient and practitioner: values imposition on the practitioner is not an issue of concern. It is worth highlighting here the difference in how Aulisio and Arora, on the one hand, and Greenblum and Kasperbaur, on the other, seek to avoid this tension. The former apply considerations of autonomy to both parties, relying on an implicit appeal to reciprocity to adjudicate disagreement; the latter seek to alleviate the tension by relying on an account of autonomy that is only relevantly exercisable, in this context, by one party.

In Response

In response to Greenblum and Kasperbaur, four points should be raised. The notion of autonomy as independence is powerful and, given the vulnerable nature and information discrepancies between parties, must be taken seriously. This is addressed in their first argument, where they state that patients should be fully informed, which includes offering information on rights of patients and on the “the full range of legitimate medical options,” as described in the aforementioned quotation. These considerations, coupled with Greenblum and Kasperbaur’s account of the medical profession as similar to government employees, require practitioners to refer patients directly for procedures and treatment even if the practitioners find such treatment morally objectionable. They rely on now common arguments from Savulescu (2006) and Stahl and Emanuel (2017), which taken together conceive of professions as mere voluntary occupations. Three concerns are worth consideration here. First, Greenblum and Kasperbaur’s conclusions clearly follow from their account of the medical profession, but one may still ask whether their description of the relevant professions is apt or if other accounts more accurately reflect the nature of a medical professional’s, such as a physician’s, role. Second, articulating “the full range of legitimate medical options” will likely require practitioners to inform the patients of a much more extensive list of possibilities than is common practice. Greenblum and Kasperbaur’s aim is laudable here and surely right, but one might wonder whether there is information or options left out of routine conversations with patients due to practitioner judgement. For example, in prescribing medication, do practitioners offer all options patients or merely choose a few? Third, though there are commonly held standards rooted in the best practices of healthcare (which is not only a good but a necessary feature of any professional practice), there does exist disagreement among professionals. Further work on adjudicating this disagreement is needed for Greenblum and Kasperbaur’s position to be fully defensible.Footnote 3 None of the concerns I raise are decisive against Greenblum and Kasperbaur’s account. Rather, they are requests for further work that is needed in order to justify the imposition on practitioners. This is the case, in part, because of my third point, which addresses the notion of predictability that Greenblum and Kasperbaur’s arguments employ.

Greenblum and Kasperbaur’s account of predictability relies on the classic account offered by F. A. Hayek, which connects their understanding to the literature on the notion of reasonable expectations. The relevance of this work is clear and well supported. There are obvious social interests—particularly in terms of efficiency—that would be violated were certain persons in certain contexts not able to predict outcomes or, better, to have a reasonable expectations of a particular state of affairs obtaining. However, is their response the best option? Were there an option that both respects patients’ legitimate interests in reasonable expectation and allows practitioners the space to live out their deeply held moral commitments, including how they understand themselves to be interacting with others, this option would be preferable. It is only if this is not possible, that we must fall back on Greenblum and Kasperbaur’s position. Thankfully, it is possible, and the situation could be addressed through public disclosure or practitioner practices. In so doing, patients would be informed upfront and could choose practitioners in light of this information. This is also not a decisive refutation of Greenblum and Kasperbaur’s position, but rather suggests an alternative approach, which given the cost of imposition on practitioners, is worth considering. Greenblum and Kasperbaur would disagree, as they would not consider practitioner to undergo any cost in such a situation, noting that practitioners should abide by publicly defensible norms.Footnote 4 Like the aforementioned features of their position, further argument is required, in particular, more needs to be said about adjudicating professional disagreement.Footnote 5

To take stock of the dialectic thus far, I have raised concerns with two of the central arguments from Greenblum and Kasperbaur; though none of these concerns is independently decisive, each calls out for further support for their claims, and jointly these concerns call for caution. Though this is a request for further argumentative support, it is also a claim about the burden of proof in the dialectic. The considerations I raise, given the costs involved—about which, I grant, there is disagreement!— place the burden of proof on Greenblum and Kasperbaur. I now turn to the final point in my fourfold response, which, as opposed to the early few, is decisive.

A Decisive Response

Greenblum and Kasperbaur are concerned that their argument falls to an objection, which they attribute to my critique of Aulisio and Arora’s position (Pilkington 2016). They are concerned that their argument fails to acknowledge physician judgement. They offer a response which they believe overcomes my objection both in their case and in Aulisio and Arora’s. They describe the objection as supporting the idea that when a person plausibly takes herself to be complicit in moral wrongs, the scales are unfairly titled toward concerns of autonomy over concerns of complicity. To this, they respond:

In response, we would point out that in a liberal democracy people disagree about many important moral questions, including considerations of moral complicity and responsibility. Both we and Aulisio and Arora could respond to Pilkington by pointing out that the physician’s understanding of moral complicity is simply part of his or her own value system. The physician is welcome to live in accord with those values, but they are forbidden from forcing patients to live by those values.

Let us put to the side the interpretation of my critique as a comment on physician judgement. There is much to say about physician judgement, the role of conscience, and the interaction between them, but that was not the point I sought to make in the paper that Greenblum and Kasperbaur reference. Rather, I was concerned that the account of autonomy offered by Aulisio and Arora only allowed into our moral theorizing a conception of the person and – derivatively – a conception of moral disagreement, which did not fit with the actual experiences of persons. The lives of persons are more complicated and messier than that analysis admitted. In light of this, I argued that the analysis, and not how we think about the lives of persons, should be altered. Another way of putting this is that the account of persons, and derivatively, of moral disagreement between persons, is too restrictive. Aulisio and Arora’s account allows in only some conceptions of the moral life, wherein persons are lists of values that clash with each other, and autonomy (by which is meant something like: whatever rules allow for the maximal play of liberties for each consistent with equal liberties for all are the rules by which we adjudicate the resolution of the conflict) is prioritized.

Greenblum and Kasperbaur’s account is promising in that it brings in richer resources and more normative content to assist in filling out the notion of autonomy (at least beyond that of Aulisio and Arora). Their argument does not succeed in responding to the objection, but what is striking about their response is that they do not rely on the new resources that they bring to bear on the problem, but rather simply return to the position articulated by Aulisio and Arora. In brief, their response (from the aforementioned quotation) is:

  1. 1.

    In a liberal democracy, persons disagree about moral questions, including complicity and responsibility.

  2. 2.

    Answers to such questions are part of a value system.

  3. 3.

    Physicians may live according to their own value system but may not force patients to live by that value system.

Given the richer account of autonomy from that work and the reliance on it by liberal political theorists such as Rawls, one would expect that further resources from it might be employed. Operating in the background of Greenblum and Kasperbaur’s argument appears to be the usual conceptual machinery, like public reason, the public and private distinction, and comprehensive doctrines. It might be that some of these resources would bolster their account; yet, they do not mention these resources in their response. Rather, they make what appears to be the same move that Aulisio and Arora made with respect to values, the move for which they criticized Aulisio and Arora so much so that it gave rise to their paper in the first place. In response to my objection, their analysis moves away from the richer notion of autonomy, and they claim a values conflict is occurring, and values cannot be forced upon patients. Aulisio and Arora’s position had the merit of a more careful distinction and aims to adjudicate values in tension, even if that account was left wanting in the end. By merely stating that there is a value conflict, Greenblum and Kasperbaur seem to have reverted to a less rich account of autonomy, and this raises questions about the consistency of earlier features of their account and about how much of an improvement is in fact offered by their account of autonomy over Aulisio and Arora’s account.

This final concern is decisive against Greenblum and Kasperbaur. In the very least, it illustrates further why the burden of proof is theirs to bear. To conclude, my original suggestion (Pilkington 2016) was that if persons could reasonably view themselves as being connected to the actions of others in certain ways and understood themselves to be morally responsible for such actions, then it is not a proper description of the situation that (merely) values are being forced on one party or even that sets of values are in conflict with each other. This charge has not been met either by Greenblum and Kasperbaur or Aulisio and Arora. The former, in the end, seem to say the same thing as the latter, but to draw different conclusions, claiming that imposition of values on practitioners is justified, and—in the final few lines of the paper —advocating for requiring referrals for patient requests to assist in suicides. If my original critique is lacking, and it very well might be, the needed argumentative response is one that specifies why the account of the moral life of persons I have suggested, and the derivative account of disagreement, fails to be reasonable on a liberal political account. Greenblum and Kasperbaur claim that it is not, but in doing so, return to rely on the failure they set out to address in the first place.