Revisiting slow codes

Every year, a small group of Michigan clinical bioethicists get together for a retreat “Up North.” The invitation to attendees instructs them to “bring your best, most out-of-the-box idea.” We workshop these ideas using the step-back consultation method, where, following the presentation of an idea, the audience, rather than subject it to attack, must “take it over” and offer constructive ways to develop it further.

The two of us are the early risers among the group and have made a habit of taking a brisk 6 a.m. hike through the Michigan woods. After commiserating about life and family and work, our conversations inevitably end up on bioethics. Through some combination of our friendship, the early hour, the intoxicatingly fresh air, and the ethos of the retreat itself, our discussions are always incredibly illuminating and productive. We come back brimming with ideas.

As we strolled along one morning, one or the other of us mentioned an argument he had been mulling over related to the ethical permissibility of slow codes. The other (we genuinely can't remember who was who) excitedly shared that they were working on something related to the same topic. Down the trails and back by the shores of Higgins Lake, our respective ideas were fleshed out.

Back home we each got right to work drafting separate pieces, and after trying unsuccessfully to think of ways to combine them, we were left with two distinct articles, but which felt very much like the product of our shared discussion in the woods. More importantly, we thought that if we both felt like the literature on slow codes had been dogmatic and incompletely theorized, maybe others out there thought so too.

Slow codes (i.e., insincere, fake, or merely performative attempts at resuscitation) have received almost no support as ethically permissible under any conditions in the literature, with the Lantos and Meadows1 piece offering a solitary and narrowly scoped exception. This is unsurprising in some ways. After all, a slow code is fundamentally dishonest. But there are lots of ethically permissible forms of deception in clinical medicine.2 Moreover, admonishments about the unethical nature of slow codes often reach the pollyannish and unsatisfying conclusion that better communication with patients and families and better policies surrounding the withholding or withdrawing of futile or inappropriate care will obviate the need for sham resuscitations.

While there has been almost no ethical defense of slow codes in the last 50 years, a recent study of clinicians who care for critically ill patients found that they remain prevalent in critical care settings.3 Moreover, half of the participants in this study believed slow codes to be ethical in some cases. The incongruence between clinical practice and ethical canon calls for a moment of disciplinary reflection and a willingness to re-engage in a critical and reflexive way with an issue everyone seems to assume was settled. That is the goal of this special issue.

If these practices are likely to continue, there is both the need and the opportunity for scholars and practitioners to develop rigorous accounts of them. Our call for papers targeted authors who wanted to engage in that effort, and the resulting articles raise new and important considerations. Perhaps most importantly, these articles re-problematize an issue that was prematurely retired.

Several articles address, in one way or another, the empirical contours of slow codes. Andrist et al. tackle the murkiness of the term itself, which is often used to refer broadly to any insincere attempt at cardio-pulmonary resuscitation.4 However, there are distinct versions of this group of practices and, therefore, divergent ethical considerations. Perhaps the legitimacy of a short but sufficiently vigorous CPR attempt needs to be evaluated differently than one where the team takes its sweet time responding or where compressions are purposely shallow.

Baker provides an important historical context that should inform any ethics analysis, particularly in light of the continued prevalence of the practice.5 After all, social practices, particularly those that are durable and resistant to countervailing norms, persist because they serve some purpose that is otherwise unsatisfied in the social system. Put another way: slow codes are not idiosyncratic, but a product of systemic factors. Any analysis of them, therefore, should take account of their socio-historical context. Baker adds this missing and critical account to the literature. Similarly, contemporary contextual features also engender the practice of slow codes. Accordingly, McLennan et al. situate the practice of slow codes amidst the prevalence of inadequate CPR policies and conventions, such as defaulting to rescuscitation attempts.6

Finally, Sprengholz offers a compelling empirical study that challenges assumptions prevalent in the literature.7 Damage to the public trust is nearly universally cited as a significant reason why slow codes can never be ethical. By offering a window into public perceptions, Sprengholz et al. challenge this assumption, showing, among other things, that public views of slow codes are not so overwhelmingly negative.

Other articles take up significant normative questions about if and when slow codes might be ethically permissible. Tarasenko-Struc thoughtfully adjudicates the normative debate, as little of it as there has been, and finds that the slow code is not always wrong, all-things-considered.8 Grosso and Nicolas arrive at a similar conclusion—it's not the case that the slow code is always wrong—albeit from a different perspective, one that finds the slow code similar to other justifiable acts of deception in medicine.9 Crutchfield uses the trolley problem as a model for decision options including slow codes, arguing that not only are slow codes permissible, they may actually be preferable or even mandatory.10 Wasserman argues that slow codes function in some cases as a form of ethical disobedience, particularly where futile or inappropriate CPR is actually or effectively mandated by state law or judicial order.11

Finally, we are particularly excited to include two narrative pieces with authors recounting moving personal experiences with slow codes. Meyers provides an account where he was asked to opine on a plan to slow code a patient.12 While he describes his ultimate decision to neither endorse or reject the plan as “cowardly,” we think it instead underscores the real life moral complexity of this terrain. Decisive moral positions are easy to come by when sitting in the cheap seats of academic journals, but a troubling ambivalence is naturally characteristic of live dilemmas.

Similarly, Mayer gives a stark view of hospital practice in the 1980s, noting “Almost nobody died back then without being coded, and it seemed like everybody who coded died.”13 She continues, “So how did we live with that? With ourselves? Slow codes provided one way out of the impossible morass for us.” Amidst the intellectualizing that attends the articles throughout the literature, and many within this issue as well, it is easy to forget the real people who have to wrestle with these experiences on the ground. Accounts of the moral distress and injury associated with current CPR practice are unacceptably thin. These narratives make a significant contribution to a thicker account of that aspect of the problem and future work should more robustly consider these kinds of practitioner experiences.

Agree with them or not, we believe that the articles in this special issue accomplish the purpose of reconsidering the moral complexities of slow codes. The literature of the last several decades has been anchored by conventional wisdom, superficial appeal to prima facie normative claims, and a disassociation from the complexities on the ground. The articles in this special issue pull up that anchor, allowing the profession's thinking on the ethics of the slow code to develop and move around. Perhaps it ends up right where it began, but at least it will have done so after deeper consideration and data gathering.