At first glance, it seems that there are similarities between the necessity and least infringement conditions of the Childress et al. (2002) framework, which raises questions about what Childress et al. consider the relationship between the two to be. In this section, we analyse the definitions of necessity and least infringement, and then analyse the relation between the two conditions, as intended by Childress et al. Regardless of Childress et al.’s intention, we argue that the two conditions (as stated) must be logically equivalent.
As noted above, the necessity condition requires that the public health intervention in question be “necessary to realize the public health goal that is sought,” and that “[t]he fact that a policy will infringe a general moral consideration provides a strong moral reason to seek an alternative strategy that is less morally troubling” (173). In other words, a morally infringing approach cannot be considered necessary to achieve a public health goal if that goal can be achieved by an alternative approach that is less morally problematic. For example,
all other things being equal, a policy that provides incentives for persons with tuberculosis to complete their treatment until cured will have priority over a policy that forcibly detains such persons in order to ensure the completion of treatment (173).
In this case, the policy that provides incentives is morally preferable to the policy that forcibly detains people, because the latter infringes liberty to a higher degree. Thus, assuming that all else is equal, e.g. that the two have equal expected effectiveness and are equal with respect to other moral costs, the policy involving forcible detention does not pass necessity.
There is an ambiguity with the question of whether a policy is “necessary to realize the public health goal that is sought” (Childress et al. 2002, 173). The idea that a policy must be necessary to realise an end could be taken to mean that the policy must be the only possible way to realise the end. However, the above example given by Childress et al. reveals that this cannot be their intended requirement of the necessity condition. They state that necessity would prioritise the policy that provides incentives over the policy that forcibly detains people, since—all else being equal—the latter is more morally costly. In this example, the policy that provides incentives is not the only possible way to realise the goal, since the forcible detention policy is still an alternative—albeit a morally worse one—that could realise the goal. More generally, it is often the case that there are multiple approaches available to achieve any given public health goal. The necessity condition would be far too stringent if it required that interventions be the one and only one way to realise the end, because it would then rule out intervention in numerous important cases. Instead, the necessity condition as stated/illustrated by Childress et al. simply requires that the costs of a given intervention are necessary in the sense that they cannot be avoided in realising the end—i.e., because there is no less costly alternative that can realise the end.
It is important to highlight that Childress et al. write “all other things being equal” (173), meaning that the necessity condition is supposed to apply only when evaluating interventions that are equally effective. All else being equal, if there are two interventions that are equally effective, and one is more morally costly than the other, then it seems self-evidently true that the least morally costly approach is morally preferable. The principle thus appears to be correct. However, the necessity condition, as stated by Childress et al., does not provide guidance regarding how to choose between two interventions that have different levels of effectiveness in achieving the public health goal. If A is more morally costly than B, but A is more effective than B, then the necessity condition does not determine which intervention is morally preferable.
The least infringement condition states:
Even when a proposed policy satisfies the first three justificatory conditions—that is, it is effective, proportionate, and essential in realizing the goal of public health—public health agents should seek to minimize the infringement of general moral considerations. For instance, when a policy infringes autonomy, public health agents should seek the least restrictive alternative; when it infringes privacy, they should seek the least intrusive alternative; and when it infringes confidentiality, they should disclose only the amount and kind of information needed, and only to those necessary, to realize the goal. The justificatory condition of least infringement could plausibly be interpreted as a corollary of necessity—for instance, a proposed coercive measure must be necessary in degree as well as kind (173).Footnote 1
Based on the definitions of necessity and least infringement given by Childress et al., the two conditions appear to be logically equivalent.Footnote 2 If an intervention passes the necessity condition, it is because there is no (equally effective) alternative act that is less infringing of moral considerations. Because passing necessity entails there is no less costly alternative that can realise the end, an intervention that passes necessity also passes least infringement. That said, it is unclear from the quoted discussion if Childress et al. themselves intend for necessity and least infringement to be logically equivalent, or if they are instead supposed to be distinct in some way.
The intended relation between necessity and least infringement
If Childress et al. consider necessity and least infringement to be different principles, there seem to be two possibilities for the intended relation between them: either least infringement is a corollary of necessity, or it is not. As noted above, Childress et al. say that “[t]he justificatory condition of least infringement could plausibly be interpreted as a corollary of necessity” (173, emphasis added). Whether or not the authors actually intend for least infringement to be a corollary of necessity is an important question because it relates to the logical relation between the two conditions. A corollary is a proposition that logically follows from a previous proposition. Thus, on the ‘corollary approach’ to necessity and least infringement, as we call it, least infringement would follow from necessity. However, even if we assume that Childress et al. would endorse the corollary approach, it is still not clear precisely what sort of corollary relation they have in mind. For example, is least infringement (LI) logically equivalent to necessity (N)—i.e., N implies LI, and LI implies N? Or, rather than logical equivalence, is least infringement merely an implication of necessity—i.e., N implies LI, but LI does not imply N? Either way, satisfying necessity would be sufficient to satisfy least infringement.
If Childress et al. do intend to take the corollary approach to necessity and least infringement, it is unclear why they include least infringement as an additional condition to necessity. One possibility is that they include least infringement as a corollary condition simply to make it clear that it is a corollary of necessity. Listing corollaries in this way may be appropriate and helpful when said corollaries are not obvious, such as in the case of mathematical theorems entailing corollaries that may be difficult to discover. However, listing least infringement as a corollary condition of necessity does not seem appropriate or helpful in this case because it seems obviously and self-evidently true that least infringement is logically equivalent to necessity (given the way the principles are defined/illustrated by Childress et al.). Another possibility is that Childress et al. have been influenced by the seminal Siracusa Principles—which includes both a “necessity” principle and a version of the least infringement principle. In short, principle 10 of the Siracusa Principles requires that a given intervention be “necessary,” meaning that, inter alia, the intervention responds to a pressing need and is proportionate to its aim; and principle 11 requires that any given intervention uses “no more restrictive means than are required” for the achievement of its aim (UN Commission on Human Rights 1984, 3). Despite sounding similar, the necessity and least infringement conditions of the Childress et al. framework clearly pose different requirements than the two Siracusa principles. Moreover, the two Siracusa principles are not logically equivalent, or even corollaries, while the two Childress et al. conditions are clearly logically equivalent.
The other interpretation of Childress et al.’s position is that they do not intend for least infringement to be a corollary of necessity [despite their suggestion that it “could plausibly be interpreted” (173) as such]. On this interpretation, satisfying necessity would not be sufficient to satisfy least infringement, and thus an intervention could (theoretically) pass necessity and fail least infringement. Though Childress et al. do not explicitly address this issue, they make it sound like necessity and least infringement are supposed to be logically distinct conditions that pose separate requirements. Recall their statement that:
Even when a proposed policy satisfies the first three justificatory conditions—that is, it is effective, proportionate, and essential in realizing the goal of public health—public health agents should seek to minimize the infringement of general moral considerations (173).
The authors thus clearly give the impression that the requirement to minimize the infringement of general moral considerations (i.e., the least infringement condition) is separate from the requirement to show that the costs of the intervention cannot be avoided in realising the public health goal (i.e., the necessity condition). Since Childress et al. appear to consider necessity and least infringement as distinct conditions that each apply separately when evaluating interventions, it seems most likely that they do not intend for least infringement to be a corollary of necessity—despite referring to this approach as “plausible.” One possibility that Childress et al. may have in mind is that necessity and least infringement are distinct in the following way: necessity requires that interventions be the least morally costly kind of intervention (compatible with realising the public health goal that is sought)—e.g., a tax is arguably a less morally costly kind of intervention than a ban—while least infringement is an additional step requiring that the intervention of the kind in question infringes moral considerations to the least degree possible (compatible with realising the public health goal that is sought). Call this interpretation the kind–degree distinction approach. It is unclear whether or not Childress et al. would endorse this interpretation, though it is suggested by their statement that a plausible interpretation of the requirements posed by necessity and least infringement is that “a proposed coercive measure must be necessary in degree as well as in kind” (173, emphasis added). However, aside from this statement, they do not address the relation between necessity and least infringement in detail—nor do they make further appeal to a kind–degree distinction.
Regardless of whether or not Childress et al. would endorse the kind–degree distinction, we argue that it cannot be the correct approach due to two distinct problems. First, it is not clear that differences in kind can be kept distinct from differences in degree in this context. To illustrate, it may be thought that a tax is a different kind of intervention than a ban. Yet, a tax that is high enough—say, a 1,000,000% tax on sugary beverages—is effectively equivalent to a ban, since most individuals could not afford a $30,000 beverage. Perhaps taxes and bans tend to be different kinds of strategies, but there are cases where there is overlap, i.e. where a tax is so restrictive that it, in effect, constitutes a ban.
Even if it turns out that we can plausibly distinguish differences in kind from differences in degree in this context, there are counterexamples to the idea that an intervention belonging to a more infringing kind cannot be the least infringing particular act/intervention when all variables are considered (and when all else is equal). Even if bans tend to be more morally infringing than taxes, the opposite may be true in some cases. For example, suppose a 20% sales tax on sugary beverages over 0.5 litres is found to be more effective at preventing obesity if it targets only the lowest socioeconomic (SES) areas of society, given that obesity rates are higher in those areas (hypothetically). Suppose this tax would be equally effective at preventing obesity as a ban on such beverages that applies across the country and affects all individuals equally. Even so, the tax is arguably more infringing of justice than the ban since the former places a disproportionate burden on the worst off-groups of society. It is thus plausible to think that the ban is the less morally infringing act, all things considered, despite being a more infringing kind of strategy than the tax. The relevant question therefore is not whether an intervention is the least infringing alternative both in kind and in degree, but instead whether it is the least infringing particular act/intervention, all things considered (insofar as least infringement is a legitimate ethical principle).