Background

This paper is designed to be complementary to and expand upon the particular framework that is published elsewhere in this journal: “An Ethics Framework for Making Resource Allocation Decisions within Clinical Care: Responding to COVID-19” (Dawson et al. 2020). In this brief paper I focus on two key tasks. First, I describe the aims behind the framework and the process of its formation and discuss some of the theoretical and practical reasons why it takes the format that it does. Second, I outline and provide some justification for the particular positions that the framework proposes and respond to some possible objections.

The framework (Dawson et al. 2020) was produced by an ad hoc committee set up within Sydney Health Ethics at the University of Sydney, Australia. It was a collaboration of clinicians with relevant expertise, clinical ethicists, and academics with expertise in ethics. The framework was produced via five meetings in ten days from first enquiry to publication on the Sydney Health Ethics website on the April 2, 2020. We had to move quickly because (in late March 2020) it was feared that hospitals in New South Wales could have faced the kind of scarcity of vital health resources where such a framework would be implemented. At that time, there was no suitable guidance for the specific New South Wales context. Available documents tended to focus on more technical issues related to how to allocate “scores” to patients within triage systems.

Ethics Frameworks and Other Guidance

Ethics frameworks are proliferating in the literature and it is not always clear what a framework is or what purpose it is to perform. I suggest that a framework should serve a particular purpose and for that reason it should take a particular form. The word “framework” uses the metaphor of a framing device, enclosing some specific content just as a frame surrounds a picture (Dawson 2009). However, the nature of the frame is itself important. In some cases, it can be chosen sympathetically and can help us to better appreciate the content. In other contexts, a frame may prove too ostentatious, thereby constituting a distraction from the picture itself. A framework, in my view, is a practical document that helps with deliberation about what we should do. It is designed to aid decision-making about what we should do but does not determine or dictate all of the answers. It is a particular kind of guidance document, but it does not have a direct regulatory role. It is not a set of rules or stipulations (as, for example, some research ethics documents tend to be structured). Nor should it be just a list of abstract values or principles, as some “frameworks” can be (University of Toronto Joint Centre for Bioethics 2005). It is unclear how the latter sort of frameworks are supposed to work. They outline what is relevant (the values or principles) but don’t say how that content is to be used when applied to cases or in decision-making. Behind the “deliberation” idea of a framework, as used in our resource allocation framework, is a particular view of ethics (Dawson 1994). Ethics is not to be seen as a series of law-like rules but rather is about articulating a response to particular cases. There is a focus on requiring moral agents to make sensitive judgements in reaction to individual circumstances, and this links to a requirement to accept responsibility for the decisions that are then made.

The latter more response-dependent account described above can then be used to justify a step-by-step method as follows. First, it is important to ensure that the framework responds to a circumscribed issue or problem. It should not be too big or too small. This framework resulted from a specific request from clinical ethicists and clinicians who would be involved in making difficult decisions about resource allocation in potentially highly constrained circumstances. Who should be prioritized for access to potentially lifesaving resources such as an intensive care bed or a ventilator, and who should not? This is a “mid-level” question that needs to be answered within the clinical setting. It is neither too theoretical nor too specific. A more theoretical question would be: how should we rank equality versus the efficient use of resources? Such a question is one for philosophical debate. How you answer such theoretical questions will be relevant to the framework that you produce, but a “framework” that focuses on answering such questions will be of little use when it comes to direct decision-making in an intensive care unit. Likewise, a “framework” that is too specific would mean that we would end up with multitudes of frameworks for very similar decisions. Such an approach would focus on answering a question such as: should Mary or Iqbal receive the ventilator, where both need one and there is only one available? We may then need a further “framework” when we later need to ask the question: should Giulia or Barak receive the ventilator? Whilst there are likely to be differences between the two situations of choosing between Mary or Iqbal and Giulia or Barak, and where relevant they need to be taken into account, these choices are not so specific that we need a separate framework for each decision.

Second, when constructing a framework, following on from the first issue, it must clearly fit within the relevant context. A good way to ensure this is to ask questions about what issues must be covered by such a framework. This second step is best achieved by engaging with a diverse group of people experienced in different aspects of the issue identified in step one. This discussion needs different perspectives and diverse experiences to attempt to ensure that the issue is looked at in different ways as a means to trying to capture all of the relevant considerations. Another way to ensure this diversity is to ask about what the relevant policy aims might be. Having an explicit discussion about such aims allows us to see potential tension between those aims and gain clarity about the contextual role for a framework within a bigger overall policy area (such as, in this case, clinical care in a high-income country).

Third, this particular framework is structured as a series of questions. In previous work I have found that a structure of questions followed by paragraphs addressing those questions tends to work well (Willison et al. 2014; Schoper et al. 2015). Questions require specific answers, and this structure provides a set of constraints on a framework and means that it is more likely to be focused on the key relevant issues, be less abstract, and be less likely to meander into unnecessary academic debate. The third step, therefore, is to formulate a series of relevant questions that are a means to addressing both the overall issue to be addressed (step one) and the full remit of relevant considerations (step two). The questions are then discussed amongst the group developing the framework, wording is refined, and the order of presentation of the questions is discussed. The order is important as it must be clear how a framework is to be used. Are there stages within a framework that require a particular ordering of questions? Is the framework “temporal” in structure? Does it proceed from beginning to end? Or is it structured around the idea of balancing a series of different considerations? Or is the structure one of a decision tree, providing options depending on choices made earlier? How does the framework outline how we are to use procedural considerations (how we should go about deciding what to do) and substantive considerations (the different elements that we should use to make such decisions)?

Justifying the Content

The COVID-19 resource allocation framework begins by setting out the aims of clinical practice in Australia. These aims are then cross-referenced in the paragraphs that respond to each question. The aims are to be taken into account within the framework but also act as constraints upon it, as any framework in this context must contribute to furthering the relevant aims. A different healthcare system in a different country or a different but related policy area (e.g., public health or social care) may have different aims. Any framework relevant to a domain of health must accord with the relevant aims of that domain’s practice.

The framework defines “resource allocation,” the range of things that may be subject to such allocation at times of scarcity, and suggests that such decisions are necessarily ethical ones (question 1). It affirms that we cannot prepare everything in advance and that is why we need a framework rather than a set of rules or stipulations (question 2).

Questions 3 and 4 set out the process of decision-making. The response to question 3 outlines a set of procedural values. Such values are often part of resource allocation documents; indeed, sometimes they are proposed as the “solution” because there is a sense that we can only agree upon process as any substantive account of justice will be necessarily contested. However, a focus on process itself provides little help with the content of ethical decision-making. The response to question 4 suggests a more specific response to the process issue by outlining how things can be organized within the intensive-care environment to best promote the relevant procedural values. The key aim is to ensure fairness in the sense of separating judgements about individual clinical care from those related to the allocation of scarce resources.

The potential justifications for resource allocation decisions are set out in the answer to question 5. These are a brief summary of the broad ethical approaches that are outlined in the relevant ethics literature. This section provides some language and context for the rest of the framework. They are deliberately kept as simple as possible: there are no references, complex language is not used, and subtle distinctions are not made. The three broad approaches might be summarized as equality, efficiency, and priority. The response to question 6 is where the fundamental proposals for decision-making in response to resource allocation scarcity occur. In the framework, we propose that after the “tipping point” is reached, which occurs when we are truly in an emergency setting of scarcity, the efficient use of resources is the primary criterion to use. However, this does not mean that the other considerations are irrelevant, and a focus on efficiency will actually have a positive impact on the other two broad approaches. For example, if we focus on the most efficient outcome, then it is clear that, at least in this context, it also requires us to be committed to equality in the sense of non-discrimination. This is because discrimination will be likely to result in a less efficient use of resources (because such an allocation would be on the basis of irrelevant reasons). A focus on efficiency will also inevitably prioritize many, but not all, of those who are worst off, especially those who have health needs that can be addressed through (the efficient) access to the relevant resource.

Some may hold that a focus on equality is so important that we should allocate resources through a random method such as a lottery (Silva 2020). The aim here is to provide a process that allows for a means to express equal value of all individuals, as all have an equal chance of access to the relevant resource. However, this approach has a major cost as it will be less efficient. This is because resources will go to those who will certainly die at this point even when provided with such resources, whereas those who could be saved will be left without. Indeed, this approach has been characterized as “innumerate” (Parfit 1978).

It is certainly true that a focus on efficiency will not address all needs, but this is because addressing disadvantage at the time of a true emergency through the allocation of the relevant resource is likely to be inefficient. This does not mean that such inequalities should not be the target of action before the tipping point or after an emergency has passed. However, it does suggest that the best time to address background inequities is in the inter-pandemic and not the intra-pandemic period. Some of the issues relating to disadvantage and prioritization might be addressed through some possible “tie-breakers” when there is too large a group that is held to be “high priority.” We do not make any specific recommendation as to what to focus on as tie-breakers, as this is a framework not a set of stipulations, but we do propose that the team given the task of formulating the implementation of resource allocation within a hospital should discuss what they consider to be most relevant and that this is made explicit and is subject to consultation and discussion with the public if there is sufficient time (as outlined in the answer to question 3).

We stress that this framework should only apply once the relevant “tipping point” is met—that is, when we are truly in the circumstances of emergency allocation. We should be careful not to act and trigger such decision-making too soon. The routine clinical duty of care and relevant resources should continue to be in place until this key point is reached (question 8). This has the advantage of ensuring that we do not, for example, prioritize COVID over non-COVID patients, nor should we move away from a concern about the overall background socio-economic determinants of health (question 7). A focus on prioritizing addressing disadvantage should remain a pressing imperative within the health system until the “tipping point” is reached.

The framework also outlines and briefly discusses other relevant issues, such as the obligation of organizations to provide a safe system of work (question 9) and the need for clinicians to see that they are themselves an important but limited resource. They have a parallel obligation not to put themselves at risk of harm within the workplace (question 10). Lastly, we outline reasons to see that it is vital that research is conducted during and after any such emergency, so that we can improve the provision of healthcare during pandemics and other emergencies (question 11).

Conclusions

The resource allocation framework developed for scarcity that may arise from a response to COVID-19 sets out to provide practical guidance to policymakers, hospitals, and clinical teams. It is designed to be accessible, clear, and useful. However, it does not provide a set of answers. It, intentionally, does not do so. It leaves decision-making to clinical teams, where it should lie, but it does aim to help them through a “framing” of what is most relevant in making such difficult decisions.