A brief history of triage

The first use of triage is attributed to Baron Dominique Jean Larrey, the surgeon general of Napoleon's armies, who developed a system of ‘prompt and methodical succor’ in response to his concern about the unnecessary loss of life caused by the haphazard management of battlefield casualties.2 In the egalitarian-influenced triage system developed by Larrey, field surgeons triaged the injured with the goal of treating the most dangerously wounded first without regard for rank or distinction.

The best known version of triage was introduced in the 18th century by the British military.3 Its development was influenced by two prominent physician hospitalists, John Aitken and Thomas Percival. The goal of British triage is to achieve maximal utility and efficiency in warfare. In this system, treatment priority is given to the health needs of medical personnel who can be returned to active duty and then to the seriously injured who are expected to survive. Those injured or ill for whom no available treatment is expected to be effective in preserving life receive the lowest priority in terms of access to critical care.

Typically, in modern healthcare settings where resources are intermittently limited under normal conditions, efforts are made to restrict admissions to the intensive care unit (ICU) and to accelerate the discharge of less severely ill patients in order to give critical care priority to very sick patients who are expected to benefit the most from such intervention. In this way, current ICU triage practice is closer to the historical French egalitarian version of triage than to the British military model.

Since the wake-up experience with severe acute respiratory syndrome (SARS) in 2002/3, the main focus of preparedness planning for circumstances of temporarily scarce critical care resources such as declared pandemic incidents has been systematic attention to the goals of utilitarianism—that is, putting into place triage protocols and health structures that are expected to produce the greatest benefits for the greatest number. Consistent with a modified British system of triage, the major clinical evaluative tools developed and promoted in the aftermath of SARS have, as their primary goal, the survival of as many people as possible from a pandemic incident.

Challenges and compromises

The allocation of scarce health resources poses significant challenges for decision makers. This is because there is no shared conception of justice for determining what health resources a person has a just claim to, and there is no existing social consensus regarding which ethics principles and values should inform health resource allocation.

Norman Daniels, in describing ‘the aggregation problem’, calls attention to one of the fundamental challenges to decision making about health resource allocation—that is, ‘when should we allow an aggregation of modest benefits to larger numbers of people to outweigh more significant benefits to fewer people?’4 There is no obvious ‘right’ answer to this question. Two polarised approaches are typically proposed and advocated—that is, provide ‘fair chances’ of deriving some health benefit(s) to ‘worst off’ individuals or give priority to whatever healthcare treatment(s) is anticipated to produce the ‘best outcomes’ (greatest net health benefit for all). Neither of these options provides much in the way of pragmatic guidance for decision making and, as Daniels has indicated, there is no satisfactory theoretical characterisation of an intermediary position.

Further to the lack of a clear answer to this fundamental question of resource allocation, important ethics principles that guide the delivery of healthcare services in normal circumstances are compromised in the interests of the necessary shift in focus from meeting the health needs of individual patients to population health and promotion of the ‘common good’ in the development and implementation of critical care triage protocols.5 One of these is ‘respect for individual autonomy’, which is the basis for the modern widely-valued practices of informed consent and patient-centred care. During declared pandemic incidents, some of those presenting for care who would be offered critical care services in circumstances of normal limited resources are not eligible for such care in circumstances of scarce resources in which a critical care triage protocol constitutes the temporary standard of care. Two other widely-held ‘patient welfare’ ethics principles may be compromised during pandemic incidents. The principle of beneficence refers to the obligation of healthcare providers and health organisations to provide health benefits to persons presenting for care, and the principle of non-maleficence refers to the obligation of healthcare providers and health organisations to do as little as possible harm to such persons. During pandemics, healthcare providers and health organisations are hampered/constrained in their fulfilment of these usual obligations. Some persons presenting for care do not receive critical care treatment benefits that they would receive in normal circumstances, and predictable harms/burdens resulting from such lack of critical care intervention accrue to them.

Other relevant theoretical considerations

In addition to concern about the compromising of ethics norms/standards, there has been some limited attention in the triage literature to the ‘tension between the individualistic basis of bioethics and the utilitarian ethos of triage’.6 Practising healthcare providers are challenged to think and act differently when actively engaged in public health work such as the triaging of scarce health resources. The shifting of focus and attention from the individual patient directly in front of the health practitioner to considerations of the rather nebulous ‘common good’ does not come naturally to most providers. For example, through their clinical training, exposure to role models/mentors and subsequent practice experience, doctors become comfortable and confident with operating within a dyadic fiduciary relationship that is founded on well-understood duties of care arising from their obligation ‘to put the interests of the patient first’. Although some health professional codes of ethics make brief reference to members’ obligations to support public health initiatives and to use health resources prudently, healthcare providers receive little training and practical guidance about how to actualise these responsibilities. While health practitioners may understand and accept the need for coercive public health measures, such as the temporary quarantining of those with difficult-to-treat infectious illnesses, they are usually not required to be (and avoid being) directly engaged in enforcing ‘constraints on the liberty of others’ that are associated with such public health initiatives.

Some triage authors have drawn a parallel between the deontological foundations of the Hippocratic tradition and the philosophy of personalism which places persons squarely at the centre of theory and practice. For Carlos Petrini, a personalist conception of solidarity ‘promotes the collective good by safeguarding and giving value to the well-being of each individual’.6 This is at odds with the usual public health understanding of solidarity which acknowledges that providing ‘fair chances’ of receiving health benefits to all individuals is likely to interfere with promotion of the well-being of the majority. As such, some of the goals of personalism are in fundamental tension with those of utilitarianism in that, for the former, obligations to individual persons supersede utility considerations. Despite this, most forms of personalism, as opposed to rigid conceptions of individualism, take seriously the moral importance of interpersonal relationships and the co-responsibility obligations that arise from these relationships.

In applying a relational social justice lens to public health, Francoise Baylis, Nuala Kenny and Susan Sherwin have drawn attention to the limits of an individualistic approach to public health and pandemic planning. In their view, ‘it is a mistake to simply import the familiar bioethical perspective of concerns that were designed to address conflicts between interests … into a discourse that really calls for … a framework that is attentive to the communal aspects and values of public health ethics’.7 In their attention to relationships and their particularities of context, Baylis et al argue for a conception of public health that is informed by relational notions of autonomy, justice and solidarity.

Other authors have suggested that a shift in the framing of critical care triage from consequentialism to virtues may make some room/space for ethics considerations beyond mere utilitarian calculations of ‘the greatest good for the greatest number’. Linda Good has proposed a set of relevant virtues, including fidelity and veracity, which she believes those engaged in critical care triage should possess.8 Gregory Larkin and Jeffrey Arnold contend that having a set of seven virtues (ie, prudence, courage, justice, stewardship, vigilance, resiliency and charity) should ‘empower [healthcare] providers … to vertically integrate principles of safety, public health, utility and medical ethics at the micro [doctor–patient], meso [provider–provider] and macro [doctor–society] levels’ in their management of mass casualty events.9

In addition to his foundational contribution to virtue theory, Aristotle was one of the first articulators of another influential ethics conception—that is, formal justice. In my view, this version of justice is helpful in framing and sorting through some of the important ethics questions/issues of relevance to pandemic critical care triage. Formal justice requires us to treat individuals and groups of persons the same unless we can demonstrate a relevant difference(s) between/among them that justifies different treatment. So, in order to justify the different treatment/management of groups of persons during a pandemic incident, there must be an identified, legitimate relevant difference(s) between them. Historically, developers of triage systems have assumed that there are important relevant differences between (1) having an anticipated very low probability of survival of the pandemic incident and having a better-than-this probability of survival; and (2) having a pre-existing very poor prognosis and not having such a prognosis prior to the incident. These two differences are the usual rationales for not offering critical care services to those presenting for care who meet one or more established medical exclusion criteria while providing such treatment to those whose health conditions do not fall within any of these criteria. For example, in alignment with assumption (1), one could argue that there is a relevant difference between the having of an advanced age and the having of a younger-than-this age if there is good evidence to demonstrate that, for a person presenting for care, being of an age above the age threshold means that he or she has a very low probability of survival of the incident while being of an age lower than the age threshold is associated with a greater probability of survival.

From a formal justice perspective, the determination of what does not constitute a relevant difference for the purposes of resource allocation decision making during a pandemic incident is of equal importance to what does constitute one. Most would likely agree that the list of factors/variables that should not constitute a relevant difference between/among individuals for the purposes of allocating scarce health resources include: gender, race, ethnicity, religion, sexual orientation, income and disability (the latter considered independently of health-related exclusion criteria). There is currently an ongoing debate about whether social utility factors such as vocational and parental status provide legitimate bases for relevant differences among individuals in the critical care triage context. Further, if one assumes that one of the primary goals of resource allocation during pandemic incidents is the survival of as many people as possible regardless of their disability status, differences in the pre-existing quality of life of those presenting for care (independent of the meeting of exclusion criteria) should not be considered relevant for the purposes of critical care triage.

In my view, exploring and expanding the justice dimensions of critical care triage is a good starting point for a broader conceptual re-examination of pandemic practices. Triage systems were initially established (and subsequently modified over time) within a distributive justice framework for the explicit purpose of developing and implementing a better than ad hoc way of temporarily distributing scarce health resources—that is, to define in pragmatic terms what is proper or fair in terms of resource allocation in this context. The well-established utilitarian mechanism for distributing scarce health resources sanctions certain disparities whereby individuals with the worst prognosis are excluded from access to critical care services. To date, no professional and social consensus around a competing practical way of distributing scarce health resources has emerged. Although personalist conceptions acknowledge the importance of solidarity and co-responsibility in decision making that will affect society as a whole, personalism in and of itself does not provide an obvious easily-adjudicated mechanism for choosing which legitimate health needs should be met when the needs of all individuals cannot be addressed due to resource constraints. With regard to virtues theory, while the having of certain virtues such as those suggested by Good and by Larkin and Arnold may predispose an individual to act in fair/good ways, it does not guarantee that the allocation choices made by virtuous persons will resonate within, and be supported by, diverse modern societies that are informed by a rich plurality of values. In my view, an evaluative approach based on exploration of social and formal justice considerations can promote the development and incorporation of ‘enhanced fairness’ features that have the potential to temper the rigid and impersonal utilitarian calculations that primarily inform the content of existing pandemic critical care triage protocols.

Augmenting the fairness of critical care triage

In this section, some pragmatic ways to enhance the fairness of critical care triage are briefly described. These ‘enhanced fairness’ features were developed by an inclusive, dynamic working group that was tasked with the creation of a pandemic critical care triage protocol for a Canadian province of Nova Scotia.