post by Bill Gardner
Fighting chronic illnesses and curbing rising health care costs will require that we change health behavior on a population scale. How do we do this? There is lots of smart work on how to use incentives and persuasive health communications to induce people to change. But we also need to understand how change diffuses through a population.
The importance of the microstructure of your social network for health behavior change was demonstrated in a brilliant set of experiments by the sociologist Damon Centola, reported in Science. By all means, read the whole thing.
What Centola did was create artificial social networks. In each network, each participant is connected to 6 other participants. The network on the left consists of neighborhoods of 6 persons with many connections to each other, and fewer connections to others. On the right, people also have 6 connections, but they are distributed across the population so as not to create clusters. People in the left are close to a small group, but far away from everyone else. People are the right are closer to a randomly chosen population member, but they do not belong to a close group.
Strangers were recruited to serve at nodes in the network. What they were asked to do was to decide whether to perform a health behavior, specifically to sign up for an online health forum. If a person decided to join, messages were sent to his 6 neighbors informing them of the decision. The question is, given that one person decides to join the health forum, in which network did the behavior spread faster? What he found was
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... My findings show that, not only is individual adoption improved by reinforcing signals that come from clustered social ties but this individual-level effect also translates into a system-level phenomenon whereby large-scale diffusion can reach more people and spread more quickly in clustered networks than in random networks.
So, changing health behavior is not just a matter of incentives and persuasion. These tools work better if their power is focused, when all your friends change together.
Centola suggests that public health programs target "clustered residential networks," but is this going to work if we are increasingly bowling alone? To my mind the obvious target is the workplace. Workplace wellness programs may work not just because the boss holds the levers, but because workplaces are often the locus of our most important social networks. But where will we be if workplace-based insurance ends, and these programs go away?