Is your social network making you fat? Are your friends and family influencing you to smoke and drink more ? And if our relationships contribute to behaviors that erode our health, can social networks be harnessed to improve it?

These are seminal questions in “network science” - an emerging field that examines how behavioral changes spread through social networks. By social networks, I mean old-fashioned, flesh-and-blood relationships.

“It’s a very old thing that we do, like ants, arranging ourselves to live in social structures,” says Dr. Nicholas A. Christakis, a Harvard University professor who studies health and social networks. “Really, humans have arranged themselves into networks for hundreds and thousands of years.”

Dr. Christakis and James H. Fowler, an associate professor at the School of Medicine at the University of California, San Diego, created an international uproar in 2007 when they published a study on obesity. In it, they reported that fat could be catching - spreading through social ties. One of the study’s findings was that a person’s chance of becoming obese increased 57 percent if the person had a friend who became obese.

Another surprising finding was that one’s chance of becoming obese was influenced not only by the weight gain of friends but also by friends of friends who gained weight. Now Dr. Christakis and Professor Fowler, as well as other scientists, are turning their attention to a new research area: how to harness social networks to promote public health. We already know that people can and do change their health habits when they seek out and participate in new social groupings.

But how do we extract information from existing social networks to improve public health? One method is to identify social connectors, people who spend time with more friends than average ? and are thus exposed to more germs and are more likely to be among the first to contract contagious diseases .

If health officials could track social people, they could tap into an earlydetection system for epidemics and figure out whom to vaccinate first. Last winter, Dr. Christakis and Professor Fowler monitored people’s friends to track the spread of H1N1 flu at Harvard.

They monitored 744 undergraduates who were either selected at random or were named as friends by the randomly selected students. Then they followed the undergrads, using their electronic medical records, to identify which students went to the university health service. In the Harvard study, published September 15 in the scientific journal PLoS One, the flu developed about two weeks earlier in the friend group than in the randomly selected group.

The results, the study leaders say, indicate that public health officials could use friend monitoring like sentinel nodes in the human body, as an early-detection system for disease. “This method, although we have studied H1N1, could be applied to anything that spreads ? smoking, weight gain,” Professor Fowler says. Some researchers are also studying how a social network’s structure affects the speed at which people adopt and stick to health habits.

Damon Centola, an assistant professor of economic sociology at the Massachusetts Institute of Technology, conducted an experiment with more than 1,500 people. He created a Web-based health forum where they had access to and could rate health information sites. Professor Centola then randomly assigned participants to one of two social network designs: one was set up like a residential neighborhood, with clusters of overlapping ties among neighbors; the other was a casual network .

Each participant was matched with other members, called “health buddies.” Although people could not contact their buddies directly, they received e-mail from the system about their buddies’ activities on the site. The neighborhood structure turned out to be much better than the random social network at prompting people in the study to join and participate in the health forum, according to Professor Centola’s report, published in September in the journal Science.

More important, Professor Centola says, the more e-mails that people received about the activities of their health buddies, the more often they returned to the forum. In the real world, he says, this means the amount of social reinforcement you give to people to improve their health habits may be more important than who is encouraging them to do so.

A local network of friends and neighbors may be more important than a remote celebrity spokesman in stopping the spread of, say, sexually transmitted diseases among teenagers. If we are to make lasting changes in our health habits, Dr. Fowler says, we may need social reinforcement in which our friends, not to mention our friends’ friends, change their habits with us. And that’s no small order for a social network.


NATASHA SINGER/ESSAY