The argument(s) from indeterminacy

A major problem of approaches to healthcare rationing that present substantial principles is that reasonable principles are indeterminate.3 When a principle is indeterminate, it is possible for two reasonable individuals to reach different conclusions about what follows from it. This indeterminacy can take three different shapes. First, a principle can be indeterminate if it relies on concepts that are vague.6 Second, the application of a principle can be indeterminate because there are epistemic problems: our limited knowledge of the world entails that the application of a principle generates indeterminate conclusions.7 And third, a principle can be indeterminate because the different value elements that the principle rests on relate to each other in an indeterminate way; there might be instances of ‘parity’, ‘imprecise equality’ or incommensurability.8–10

Vagueness problems are familiar in many fields of philosophy. For some concepts and categories, it seems impossible to establish exact borders of when some object falls within this category or not, for example, how many grains of sand are required for something to constitute a heap. In principles used for healthcare rationing, we can expect many concepts to be vague. The very concept healthy seems vague, so also pain, dysfunction and treatment are arguably vague. Principles that involve vague concepts will be indeterminate because their application can generate different conclusions depending on how the vague concepts are precisified in the specific instance, something that the principles themselves say nothing about. Vagueness is a good reason to turn to procedural approaches to healthcare rationing, because with a just process we can hope for a just precisification.

Ignorance of the world can also entail that a principle is indeterminate. Some concepts and categories might be determinate in theory, but impossible to apply in a determinate way. Thus, even if we were able to establish a determinate concept of opportunity and use this in a principle used for healthcare rationing that evaluates allocation alternatives in terms of how well they promote equality of opportunity, we might be unable to use this principle in a determinate way for a variety of reasons. We are unable to foresee the full consequences of certain interventions. We cannot have perfect knowledge of the amount of opportunities an individual have even if we would know what opportunities are, and so on. Epistemic problems that are prevalent give us good reason to turn to procedural approaches to healthcare rationing, because with a just process we can hope for a just interpretation of uncertainties.

Finally, axiological indeterminacy can make principle-based approaches indeterminate. As has been recognised in axiology lately, two values might relate to each other in such a way so that, in some instantiations of these values, it is false that one alternative is better than, worse than or equally good as another alternative.11 ,12 In such cases, value instantiations have been claimed to be ‘on a par’ or ‘imprecisely equal’.8 ,9 Another suggestion has been that two principles are ‘incommensurable’ or incomparable, in the sense that a no comparative relation at all can be established.10 Axiological indeterminacy differs when value instantiations are equally good since a small relative increase of one value instantiation would not change the relation between the value instantiations. In principles used for healthcare rationing, this phenomenon can generate indeterminacy in a principle (or in a set of principles) that relies on two values the relation between which is in some instantiations indeterminate. For example, a principle that states that we ought to allocate scarce resources so that they generate as much health as possible to as many individuals as possible risks embodying this type of indeterminacy because the following two alternative allocations are arguably on a par: allocation A: increase average life expectancy of 10 000 people with 1 month; allocation B: increase average life expectancy of 100 people with 100 months. Axiological phenomena of this kind are very difficult to avoid in any field that deals with population ethics.13 These problems give us reason to turn to procedural approaches to healthcare rationing, because if we know what a just process looks like, we can apply this process to establish, or rather superimpose, a positive relation according to the trichotomy (better than; worse than; equal to)that hold between alternatives that otherwise would be on a par/imprecisely equal/incommensurable. If the process behind this is just, the outcome will be just.

Regardless of the grounds, indeterminacy entails the need for precisification of the indeterminacy. What is indeterminate can be made determinate by precisifying interpretations. If the principles themselves fail to fully determine what we ought to do, we have to look elsewhere for a ground for the precisification. Since just processes generate just results we can invoke these to ground just precisifications. A process that generates precisification and ensuing determinacy can deal with vague principles, epistemic indeterminacy and axiological indeterminacy. If we secure that the process is just, then we also secure that the precisifications are just. Turning to procedural approaches to healthcare rationing thus seems very appealing in light of indeterminacy.2 ,3

Complete or partial indeterminacy

We have seen that we have good reasons to look for procedural approaches to healthcare rationing when we need to deal with indeterminacy. However, that a procedural approach is needed does not mean that it is all we need, and it does not mean that it is necessarily the first thing that we should invoke when we face rationing decisions.

There is an important difference between what we can call complete and partial indeterminacy. Let u's call a principle completely indeterminate if it in every application fails to generate a unique ranking of the alternatives, and let us call a principle partially indeterminate if it in some applications fails to generate a unique ranking of the alternatives. Complete indeterminacy entails that part of the decision always depends on the application process in which the indeterminacy is precisified. Partial indeterminacy entails that part of the decision sometimes depend on the application process in which the indeterminacy is precisified. Regardless of which type of indeterminacy a principle consists of, it seems hasty to conclude that it concerns complete indeterminacy just because we can find some cases of indeterminacy.

Principles that involve vague concepts will still be determinately applicable to some situations. Take, for example, a principle such as ‘give priority to the individuals in most pain’ and let us accept that pain is a vague concept. It is true that this principle is difficult to apply when we need to choose whether to prioritise an individual in significant psychological distress or an individual in significant physical pain. Yet, it is not difficult to see that the principle determinately tells us that we should prioritise an individual who screams in agony due to a bone fracture over an individual with a slight headache due to mild fever.

Principles that generate indeterminate conclusions due to epistemic problems also generate determinate conclusions in some situations. Take, for example, a principle such as ‘allocate resources so that they promote equality of opportunity’. It is true that such a principle cannot in itself determine whether to choose to allocate resources to strengthen the already quite well-funded division of paediatric diabetes care at the local hospital or to the also quite well-funded intensive care unit at the same hospital. Yet, we can use the principle to determinately say that everything else equal we ought, in light of this principle, to allocate resources to secure equal access to a polio vaccination programme that is currently in need of resources rather than to invest in a renovation of the fully functional premises of the already well-funded paediatric diabetes care unit.

Finally, principles that are indeterminate because they involve value elements that relate to each other in an indeterminate way, that is, are on a par or incommensurable, will occasionally at least be able to guide our decisions in a unique way. It is true that such principles will fail to generate determinate conclusions in situations such as the one introduced above. A principle that states that we ought to maximise health benefits for the maximum amount of people cannot determinately resolve the issue of whether it is better to prolong the average lifespan of 10 000 individuals with 1 month or to prolong the average lifespan of 100 individuals with 100 months. However, the principle can determinately say that it is better to prolong the average lifespan of 10 000 individuals with 1 month rather than to prolong the average lifespan of the same 10 000 individuals with 15 days. Similarly, a set composed of a deontological and a teleological principle that are held to be incommensurable can fully determine what we ought to do when they point to the same course of action, for example, if we can produce the most total well-being by respecting the sanctity of life.

Indeterminacy is a problem, but we are not facing complete indeterminacy when we talk about indeterminate principles in relation to healthcare rationing. Rather, we must expect that substantial principles will be partially indeterminate. In other words, substantial principles that apply to healthcare rationing will sometimes fail to alone generate determinate responses to rationing problems. Yet, this will not always be the case; in some pairwise comparisons of resource allocation alternatives, substantial principles will alone generate determinate evaluations. Therefore, we must ask the question: if we see the merit of decision-guiding substantial principles, what is the scope and role of procedural approaches to healthcare rationing in light of indeterminacy problems? If the main reason why we invoke procedural approaches is indeterminacy, a more appropriate procedural move might be to invoke procedural approaches only to precisify indeterminacy. There might of course be other reasons to make the procedural move, but that issue falls outside of the scope of this paper.