Is Your Social Network Good for Your Health?

People are social by nature, spending about 80 percent of their waking hours with family, friends, or co-workers. A spate of studies about the health impact of this web of personal relationships—or "social networks," in research parlance—suggests that friends, family members, and even friends-of-friends can "spread" particular health behaviors and emotions, just as they can spread a cold or flu virus.

Dr. Nicholas Christakis, a social scientist at Harvard University and a physician specializing in palliative care, became interested in social networks while helping patients and families make end-of-life decisions. He has collaborated frequently with Dr. James Fowler, a political scientist at the University of California, San Diego.

The two analyzed several decades' worth of data collected through the Framingham Heart Study, launched in 1948 to identify risk factors for heart disease. Because follow-up was so important in this study, its researchers not only evaluated health behaviors and outcomes in the study's participants, but also collected detailed information about their parents, spouses, siblings, children, close friends, and employers. (That way, if participants moved or changed jobs, the Framingham researchers could locate them for follow-up.)

Spreading girth

The researchers found that obesity spreads within social networks, although not in the way that germs spread. Instead, social ties affected weight gain patterns more than geographic proximity.

For example, when one spouse became obese, the other had a greater risk of becoming obese as well. Likewise, one person's chance of becoming obese increased if a close friend became obese, but not if an acquaintance or a next-door neighbor gained weight.

Social smoking

Social ties also influenced smoking behaviors. If one spouse stopped smoking, the other was also more likely to quit. And whenever one individual stopped smoking, that person's friends and siblings were also more likely to kick the habit.

But a next-door neighbor had no impact on an individual's decision to quit smoking. The impact of co-workers depended on the size of a company. Employees toiling in large companies continued smoking (or quit) regardless of what co-workers did. But in small firms (defined as those with six or fewer employees participating in the Framingham Heart Study), an employee was more likely to quit smoking if a co-worker did.

Happy contacts

The investigators assessed both happiness and loneliness (discussed below) through participants' responses to questions on the Center for Epidemiologic Studies Depression Scale. They found that both geography and social ties affected the spread of happiness — perhaps attesting to the fact that face-to-face contact is important for communicating emotions.

Individuals who lived within a mile of a happy friend were likely to catch the good cheer, while friends who lived more than a mile away had no impact. Siblings who lived within a mile of each other influenced each other's happiness, while siblings who lived farther away did not. And a happy next-door neighbor increased the chances that someone would become more cheerful, but a neighbor living farther down the block had no effect.

The lonely crowd

Loneliness is generally perceived to be individual in nature, but the researchers found that it can also spread in networks. They found, for example, that participants were more likely to report being lonely if they were directly connected to someone who was lonely. As in the other studies, both the strength of the social tie and geographic proximity mattered — with one notable exception.

Friends were more likely to have an impact on someone's feeling lonely than acquaintances. Next-door neighbors had more impact on someone's loneliness than people living down the street did.

In a surprise, however, siblings—no matter where they lived—had no impact on how lonely another sibling might feel. The researchers interpret this as suggesting that loneliness reflects the relationships people form on their own, rather than the ones they are born into.

Theories, criticisms, and future investigations

Three hypotheses exist as to why health behaviors or emotions might spread among members of a social network.

Induction. One person in the network contributes to or directly causes a behavior change or emotion in another person.

Homophily. People with similar characteristics seek one another out and influence one another.

Shared environment. Individuals connected by geography or personal ties also share similar environments, which might explain why behaviors and emotions spread.

Although they address all three mechanisms, Drs. Christakis and Fowler believe that induction underlies the health effects of social networks. It's not so much that social networks cause obesity, smoking behaviors, or mood changes. Rather, they argue, social networks magnify the effects of environmental factors and other phenomena.

For example, genes and environment clearly contribute to a person's weight. What social networks contribute are a sense of the norm. People who notice that their best friend has gained weight since the last college reunion, for example, might unconsciously adjust their understanding of what constitutes a normal body weight.

Critics contend that the researchers are oversimplifying complex phenomena. Others point out that the Framingham Heart Study participants may be more alike in habits and outlook than random samples in other communities—which suggests that homophily and environment, not induction, may underlie shared patterns of behavior.

Drs. Christakis and Fowler have countered these criticisms with a detailed defense of their mathematical models, investigational methods, and study populations. (Readers interested in the details of their analysis can find it in the Journal of Health Economics; the citation is listed under "References.") From a clinical point of view, social network theory is just that—theory.

But the research confirms what many mental health clinicians know through their training and experience: health and other behaviors have not only personal but also collateral effects. Social network theory helps to explain why group therapies and group-based self-help interventions, such as Alcoholics Anonymous or Weight Watchers, might work. These may be artificially constructed social networks, but they are networks all the same.

For more references, please see www.health.harvard.edu/mentalextra.

Christakis NA, et al. "The Collective Dynamics of Smoking in a Large Social Network," New England Journal of Medicine (May 22, 2008): Vol. 358, No. 21, pp. 2249–58.

Christakis NA, et al. "The Spread of Obesity in a Large Social Network Over 32 Years," New England Journal of Medicine (July 26, 2007): Vol. 357, No. 4, pp. 370–79.

Fowler JH, et al. "Estimating Peer Effects on Health in Social Networks: A Response to Cohen-Cole and Fletcher; and Trogdon, Nonnemaker, and Pais," Journal of Health Economics (Sept. 2008): Vol. 27, No. 5, pp. 1400–05.