Why you can only solve a food crisis by targeting those starving most.

Say you're an aid worker toiling in an area that's been devastated by an overwhelming calamity -- a war-torn conflict zone, a famine-stricken village, or a crowded refugee camp. You have a finite amount of emergency food and see far too many outstretched hands. The kids around you are all much too short and scrawny for their ages, and most seem sickly and under-fed. What's more, many of the youngsters show sure signs of acute malnutrition -- they literally are skin and bones.

Who do you feed first?

In this type of situation, aid groups might be tempted to spread out the food supplies to as many needy children as possible, giving each a small amount until there's nothing left. But according to a new study out of Stanford's Graduate School of Business and the University of Bergen in Norway, relief workers should instead give as much emergency food as possible to only those children who are in the greatest danger of dying. The rest should get nothing.

According to the authors, this "all-or-nothing" approach will actually save more lives in the long run.

The researchers focused their study on "ready-for-use therapeutic foods," which are essentially protein-rich pastes that pack a balance of vitamins, carbohydrates, and other nutrients. They're portable, don't require cooking, and just a few doses can dramatically reverse the effects of starvation.

After performing a mathematical analysis of data on thousands of undernourished children under the age of five in the Democratic Republic of Congo and in Niger, the authors found that aid workers are better off giving a full dose of the therapeutic food to the few children who are worst off, than say, half- or quarter-doses to the entire village.

In the long run, the authors found, receiving only a small amount of food won't do much to help the kids that weren't already in danger of death, as measured by both the amount of wasting, or muscle loss, they exhibit and how stunted their growth is for their age. But the all-or-nothing approach could reduce deaths and life-limiting disabilities by 9 percent.

"The goal is to minimize the number of disability-adjusted life years, most of which are due to childhood death," said Lawrence Wein, one of the authors and a professor in the Stanford University School of Business. "You do better by not doing blanket distribution. You take all the money that's available and give out full doses, and that will perform better."

Stanford acknowledged that the findings might be controversial -- who wants to keep food out of the hands of starving children, after all, even if they're less-starving than others? But Wein said that in reality, aid groups already make cutoffs to determine which children are fed first in emergency situations, based on wasting, disease, or other metrics.

"They already draw a line in the sand and say 'anyone worse than this, we're going to go all out and save you,'" he said.

This finding, it seems, recommends concentrating life-saving food resources on only one side of that line.