There's not really anything recognizable as an argument in Macklin & Cowan's paper (it mostly involves describing various putative ethical principles). But they summarize their view in the Conclusion as follows:
Because ethical principles can conflict, reasonable people may disagree about which principle should take precedence. In the case at hand, the most salient conflicts are between the utilitarian obligation to try to prevent the most overall deaths from HIV/AIDS and several other key ethical rules and principles. These are, first, the rule of rescue, which mandates treatment for those who will die without treatment; second, the principle of urgent need, which requires treatment for people sick and suffering from HIV/AIDS; and third, the prioritarian principle, which argues that currently sick people are worse off than healthy people, even those who may be at high risk of becoming HIV infected. We contend that these three principles, taken together, outweigh the utilitarian principle in this situation of resource allocation.
Let's consider these in turn.
(1) The Rule of Rescue.
Macklin & Cowan quote B&W's characterization of this rule as follows: "The fact that we can save identified people whose lives are imminently threatened by AIDS creates an obligation to do so that must be honored, even if so doing reduces the number of lives saved overall."
However, they completely neglect B&W's distinction between descriptive and normative interpretations of the principle. It may be true that people are psychologically disposed to privilege identifiable individuals over so-called "statistical lives", but to elevate this disposition to a normative principle requires some defense. As B&W note, the Rule of Rescue is flagrantly inconsistent with more fundamental moral principles concerning the equal worth of persons: "statistical lives saved are just as real as identified lives saved; all have the same equal worth. And the same is true for the difference between lives saved in the present versus at some point in the future."
The common disposition to prefer identified lives over "statistical" lives is thus exposed as mere prejudice -- a prejudice that Macklin & Cowan lazily rely upon without so much as acknowledging the objections to it.
(2) Urgent Need
As B&W explain it, the principle of urgent need -- i.e., prioritizing those with the most time-sensitive needs -- is not a basic moral principle but rather a practical policy that applies in times of temporary scarcity, the rationale being straightforwardly utilitarian: this way we can help the most people, as the less-urgent cases can safely wait until the temporary shortage has passed.
This clearly doesn't apply to allocation decisions in contexts of permanent scarcity, where whoever we don't help now will end up dying, rather than being helped later. Macklin & Cowan seem confused by this. They note that B&W see urgency as relevant only in cases of temporary scarcity, but they don't mention the rationale for this, let alone argue against it. They merely assert: "Still, if the question is whether to use limited antiretroviral medications for [treatment or prevention]... urgent need does seem relevant." (Really? How so? Isn't this just to express, once again, a bias towards the present?)
Prioritarians hold that we should give priority to the worse-off. Macklin & Cowan note that, at the moment of decision, those in need of treatment are worse off than the merely "at-risk" populations who would benefit from prevention.
Once again, they completely ignore B&W's powerful objection: "Without the prevention program, some people will become HIV-infected, develop AIDS, and become patients in need of treatment to save their lives. They will then be just as badly off as the patients now in need of treatment. The difference is only when each is most badly off." (1672-3, emphasis added)
Is there some reason that prioritarians should be restricting their attention merely to the present moment, in judging who counts as "badly off", rather than making a more informed judgment that takes into account the welfare levels that individuals will have in future, depending on the agent's present choice? Surely the broader view is morally superior here. At the very least, proponents of the narrow view owe us some argument in defense of what, on the face of it, appears to be a straightforward temporal bias.
(Perhaps the most interesting defense here would appeal to the idea that at present there may be no fact of the matter regarding which particular individuals would end up getting infected in future, and try to build on that by appeal to some kind of person-affecting principle. Of course, I think even this would ultimately prove indefensible for the sorts of reasons mentioned here. But at least it would be an argument; albeit a bad one.)
It's really worrying that such a bad paper could be published on such an important topic. Just to be clear: Macklin & Cowan are advocating for a policy that would cause more people to die of AIDS. And they're doing it, not by offering new or intriguing arguments that warrant intellectual engagement, but simply by reasserting discredited principles and pretending that they haven't already been discredited. What were the referees and editors at Health Affairs thinking? (Are such low standards commonplace in bioethics?)