RWJF-Supported Study Redefines the Way Researchers See Health Behavior

In an extraordinary example of interdisciplinary research, a scholar combines physics, network theory and social epidemiology to greatly advance our knowledge of how we can help each other live healthier lives.

    • January 27, 2011

Scientists and health experts have known for decades that our behavior is greatly affected by the company we keep. For this reason, health experts have worked long and hard to figure out how to introduce healthy behaviors to groups of people who will, in turn, encourage other groups to adapt healthier lifestyles. Most research in this field is based on the premise “that behavior is spread much like contagious diseases,” explains Damon Centola, Ph.D., a mathematician and 2006-2008 Robert Wood Johnson Foundation (RWJF) Scholar in Health Policy Research, “but my work shows it spreads more rapidly in residential, rather than casual contact social networks.”

An ideal candidate for the unique, interdisciplinary approach of the RWJF Scholars in Health Policy Research Program, Centola’s public health breakthrough began with a study of mathematical modeling he conducted in 2007. “The mathematical theory papers—one published in a physics journal, one published in a sociology journal—showed how networks can affect diffusion in a counter-intuitive way,” Centola explains. “Both papers were significant advances for the theory of network science,” he says, but it was also clear that the theory, called “complex contagions,” had significant implications for research in public health.

The Power of Neighborhood Connections
In a follow-up experiment, Centola, who is a professor at the MIT Sloan School of Management, used artificially constructed, Internet-based social networks—a bit like Facebook—to find the fastest and most effective way to spread behavioral change. His findings were published in "The Spread of Behavior in an Online Social Network Experiment," in the September 2010 issue of the journal Science. “I investigated the effects of network structure on diffusion by studying the spread of health behavior through artificially structured online communities,” Centola said. “I found that individual adoption was much more likely when participants received social reinforcement from multiple neighbors in the social network. The behavior spread farther and faster across clustered-lattice networks than across corresponding random networks.”

Centola’s discovery potentially provides a valuable new approach for the design of public health interventions that seems firmly grounded in common sense. It’s a given that a few, stray fragments of a cold virus, propelled across a bus or subway train, can easily infect us with another person’s illness, whether we like it or not. But actually changing our behavior is a far more complex and nuanced process. Having the support of other people in our social network—peers or trusted friends—can help us overcome barriers to change and encourage us to stick to our convictions. That dynamic—the reinforcement people give to others with whom they share a community, workspace or social group—underlies Centola’s finding that behaviors introduced to closely structured residential networks spread quickly and have far more adopters than those spread through random, casual contacts.

“I’ve shown that you can connect people intentionally to affect overall community health,” Centola said. “The people in the study were anonymous. I did not rely on any previous relationships and they were selected randomly, so the results do not depend on existing influencers between people. The emphasis is on social structure. The nature of the Internet—that it can be used as a kind of a social Petri dish—made it possible to test this theory.”

New Research: Building on Relationships
“In my current work, I’m investigating new ways to engage people online. This time, I am matching people according to certain characteristics—such as Body Mass Index– to see how the peer environment affects people’s willingness to adopt health related behaviors. The really interesting question is whether the behaviors will spread between people with different characteristics, or whether they will get trapped in certain regions of the social network that are defined by specific health traits,” Centola explained. The new study aims to explore the possibilities of what can be done online. “I would like to know how rich these environments can be, and how meaningful we can make the social interactions in order to understand the global implications of how behaviors spread.”

Starting in 2006, Centola used his time as a Scholar in Health Policy Research to design the proper environment for his work. “It took almost two years to structure this study and the architecture had to be perfect—every detail of the experimental environment has consequences for the theoretical interpretation of the results,” he said of the network design process. “But I also believe my research achieved one of the primary goals of the Scholars in Health Policy Research Program and that is to provide the resources and training needed to get non-public health researchers thinking about the relevant implications of their work for public health,” remarked Centola, whose day-to-day work as a mathematician is far-removed from the world of health and medicine.

“In my case,” Centola added, that meant “bringing a network scientist like myself—trained in physics and social theory—into doing research in social epidemiology. The RWJF connection made it possible for me to make the move between theory and empirical application. This is a particularly special thing in the area that I work in—mathematical modeling/social theory—since most of these theories never get tested empirically. So, my project is an exciting methodological advance, but it’s also—in a way—a disciplinary advance. It's one of the few times that people who work in my area have seen the speculative bridge from mathematical theory, to empirical application, to real world policy implications that can actually be realized,” Centola added. His latest work will be under review later this year.

The Robert Wood Johnson Foundation Scholars in Health Policy Research Program aims to advance the involvement of outstanding new Ph.D.s in economics, political science and sociology in health policy research so that, while pursuing careers within their disciplines, they will also be able to make important research contributions to future health policies.

Most Requested