As a second consideration which must be addressed before the question directly. Serving the poorest has more broad-reaching effects than simply an increase in DALYs (which is usually optimized in cost effectiveness analysis or CEA). An intervention that allows a poor person to escape poverty had long-reaching effects on that person, their children and their society. Saving a man from premature death by AIDS does more than help him, it allows him to contribute to his community. Getting people onto the proverbial ladder of development, putting them in a position where they could begin to climb up and improve their situation, is an important consideration beyond strict CEA.
Nevertheless, there are times and places where you can help people more effectively who are better off, or you can less effectively those who are worse off. All other things being equal (particularly the Development value discussed above), I value efficiency over equity.
There is certainly an argument to be made for serving the poor. You have to try really hard to find a philosopher or religious teacher who teaches against this (perhaps Ayn Rand, but I think even she may see the value of Global Health as a piece of International Development). To most (including myself), it seems to be a Good which is self-evident. But in much of global health, everybody’s poor. If I could bring 10 years of added life to 1000 people making $2/day or to 500 people making $1/day, I’d choose to help the richer and so save more people.
As I’ve written elsewhere, where Kantian ethical statements are satisfied, (i.e. One ought to help the poor), Utilitarian ones should pervade (“Greatest good for the greatest many”). I disagree with the argument that is made that achieving ultimate equality in income is a moral imperative (this argument is frequently attributed to John Rawls, though I admit I’ve only had secondary exposure to him). I reject a morality whose end is the homogenization of income; I disagree that the lack of equality (e.g. Capitalism) is inherently bad, and I do not think that the logical end of this, Communism, is even a good place to be (see Harrison Bergeron). I fail to see equity as an independent axis of the Good; it certainly correlates to the Good. Often those who are poorest are most oppressed. Those who are weak are enslaved. So I agree that injustice is committed against the poor, but as we all know, correlation is not causation. The badness is in the injustice, not the disequilibrium of money.
Truly it is moral to serve the poor for reasons of mercy and compassion. Suffering is indeed evil that should be fought against. But I cannot see a universal moral reason why some people can’t drive Hondas and others Acuras.
Fortunately, this theoretical situation is not that big a problem. It’s clearly moral to help the poor. And the poorest and the easiest to help are often the same people. When they’re not, most of those remaining could be helped in escaping povery by Global Health. When they can’t be, we should consider showing the greatest good that we are capable of. We can discuss these minor points of philosophy (and really should), but it’s not where we, as a society, are failing. I hope that one day we’re having this very debate broadly: “Should we save the really really poor? Or just the really poor?” Unfortunately, we’re still at the, “Should I buy a third car? Or a new TV?” phase.
Carreon, David. “To Promote the General Welfare – Why Utilitarianism Works Better than Kantianism in Government.” 28 Apr 2009. <http://arena-man.blogspot.com/2009/04/to-promote-general-welfare.html>
 One of my favorite short stories and a powerful argument against the ethics of Equality. It opens thus,