How Our Friends (and Our Friends' Friends) Influence Our Health

    • September 1, 2011

The challenge. People's lives are intertwined in social networks—at school, at work, in their neighborhoods and in their activities. Yet researchers are just beginning to study the effect of these social networks on human health.

"If we could understand how those networks form and function, it might shed new light on interventions to help the health of the public," said Nicholas Christakis, MD, PhD, MPH, a professor of medicine and of medical sociology at Harvard Medical School.

Where his interest started. Christakis first became interested in social networks as an assistant professor at the University of Chicago in the 1990s. Trained as a physician and social scientist, Christakis was working as a hospice doctor. As part of his research, he looked at the so-called "widowhood effect"—the increased propensity for a spouse to die shortly after a partner's death. But Christakis soon learned that the emotional toll of a death extends far beyond a spouse.

Christakis was treating an elderly woman with dementia, whose daughter provided much of her care. The strenuous responsibilities took their toll on the daughter who became exhausted. Her husband, in turn, became depressed. And then the husband's best friend began to feel overwhelmed by the situation and called Christakis for help.

"Here was this random guy calling me, and it occurred to me that his call reflected a kind of nonbiological spread of the disease," Christakis said. "I suddenly realized that husband-wife pairs were a special case of the much larger phenomenon of social networks. I was studying widowhood in dyads, but I became interested in looking at much broader, more ornate social networks."

Since the early 2000s, helped in part by training he gained in the early 1990s as an RWJF Clinical Scholar, as well as prior Investigator Award funding, Christakis has focused his research on how social interaction affects health. In one of his most well-known studies, he showed how obesity is spread through a social network. The study found that a person's chances of becoming obese increased by 57 percent if that person had a friend who became obese over a given time interval. It also showed that if one adult sibling became obese, the chance that the other would become obese increased by 40 percent.

"Part of what may explain the obesity epidemic over the past 30 years has been a change in norms regarding body size," Christakis said. "Such norms might spread through social interactions, and, if so, then as one person gains weight, other people do, too."

How social networks influence health. In partnership with his long-time collaborator, James Fowler, PhD, a social scientist at the University of California, San Diego, Christakis and a team of Harvard researchers deepened their investigation with a grant from RWJF that ran from 2007 to 2010 (ID# 058729). They began by building several data sets that could be used to analyze the role that social networks play in health and health care.

With those data sets and with actual experiments they conducted, they studied the effect of social networks on health-relevant behaviors, such as smoking and alcohol use, happiness and generosity. Their findings were consistent—how we behave spreads from person to person in a social network. "Your health is not just a result of your choices and actions but also the choices and actions of people around you," said Christakis.

One of the findings that most surprised Christakis is that people are influenced not only by those they know but also by the friends of their friends. For example, in his study of smoking, Christakis learned that the smoking behavior of an individual is associated with the smoking behavior of people up to three degrees removed in a social network (i.e., friends of friends of friends).

"The fact that people influence each other is not surprising," he said. "We were surprised that your friends' friends could affect you."

For more on research findings from this grant, see Program Results Report.

Changing health and behavior through social networks. Christakis and his colleagues are beginning to look at how health professionals and policy-makers can use information about social networks to influence the health of the public.

For example, their 2009 study of Harvard undergraduates and their friends helped to predict a flu outbreak about six weeks before it became evident in the population as a whole. Such notice could give public health officials time to prepare for flu outbreaks and warn people about them. "You have a completely new way of predicting the futures with social network, not just of outbreaks of germs, but also of bad behaviors like binge drinking on college campuses," Christakis said.

With a new four-year RWJF grant (ID# 067919), Christakis and his colleagues are exploring other opportunities to identify and harvest interpersonal information online and to develop tools that can be deployed more widely. "We are trying to see if we can use online networks to improve public health," he said. "Facebook represents an enormous amount of valuable information about people's social interactions." In addition, Christakis said, "We have begun to explore new horizons, doing more experiments with networks, in which we randomly assign people to different structural locations or to get different sorts of information."

Reaching a broad audience. Christakis and his colleagues have published their work on social networks in a wide range of peer-reviewed journals, including the New England Journal of Medicine, Proceedings of the National Academy of Sciences, and the Journal of Health Economics.

They have also been happy to share findings more broadly. Christakis' website provides links to media coverage of the research and he co-authored a popular book on the topic with Fowler, which was published in September 2009 by Little Brown: Connected: The Surprising Power of Our Social Networks and How They Shape Our Lives."

The goal of research, he said, is to create and disseminate knowledge. "The reason that wealthy individuals, foundations or the government fund academic research is not so we can sit in ivory towers. The reason is to advance science and to advance the public understanding of science. Engaging the public, and not just ourselves, is thus a civic-minded thing to do, and it enhances scientific literacy and contributes to a better discourse."

In response to some recent criticism and to the many confirmatory studies of their work on social networks, Christakis' partner on the original study, James Fowler, told the New York Times (August 9, 2011): "This is how science proceeds.... We came up with a fact that no one ever thought about before. We published our methods. We published our data. We said, 'Look, we think this is important. You should help us figure out how to do this better.'"

RWJF perspective. "Nicholas Christakis' work is really about learning how best to spread information that leads to behavior change," said Lori Melichar, PhD, RWJF director. "How do we use networks to have a faster, bigger impact on health and health care? What can we learn from social networks that translates into actual interventions?

"Christakis and his colleagues have called a lot of attention to the potential of social network analysis to look at problems and possible solutions differently."