The promise of social network analysis

May 23, 2008, 8:57 AM, Posted by Lori Melichar

Nicholas Christakis’ new study on social networks and smoking cessation was published in yesterday’s New England Journal of Medicine. Using data from the Framingham Heart Study, Christakis and his colleagues reconstructed the social networks of more than 12,000 individuals and found that smoking cessation occurs in network clusters – the study also concludes that the chances of continuing to smoke decrease significantly for an individual when their spouse, friend or even sibling quits smoking.

The article, funded in part by the Pioneer Portfolio, is garnering a lot of attention in the media and stimulating many dinner table discussions. While the findings provide valuable lessons for those working in the field of smoking cessation, the study has implications well beyond smoking. Christakis’ work, reflected in this study as well as in the results he published last year on the influence of social networks on the spread of obesity, opens up a whole new way of exploring health behavior. I expect that over the next few days, weeks, months and years, policy makers and those advocating for social change will discuss how these findings can shape interventions and policies and researchers will clamor about how this innovative approach will expand the arsenal of tools used to help us understand and address some of our nation’s most challenging health problems.

Our grant supports Dr. Christakis’ and his research team’s efforts to create new data sets and develop innovative statistical methods that may allow health researchers to track and analyze the spread of health outcomes and behaviors – both good and bad – within complex, real-world social networks that evolve across time. The Foundation was willing to invest in developing these methodological tools because of an up-to-now unproven suspicion that the social network of relatives, friends, co-workers, neighbors, health providers and others can have a strong influence on an individual’s behaviors. And we believe that social network analysis can help us better understand the relationship between healthy or unhealthy behaviors and social linkages. Can we do a better job of improving health behaviors if we work at the level of the social network, instead of only the individual?

Christakis’ research findings have the potential to drive a fundamental rethinking of health policy, clinical care, research and evaluation, and public health campaigns. If social network analysis continues to produce promising new results and becomes widely used – and if it helps us to think differently about how we design health interventions and health campaigns that ultimately achieve greater success – then we will have achieved a key breakthrough in the health and health care of all Americans.