Social Network Analysis, Obesity and Other Topics

Jul 31, 2007, 5:56 AM, Posted by

I've been following with interest the reaction this week to the findings published in the New England Journal of Medicine on spread of obesity through social networks. While much of the discussion in the blogs and on the Web has been about the substance — whether obesity is, in fact, socially “contagious” (although the paper in fact hypothesizes that the spread of norms and perhaps behaviors may be the key) and has not always been accurate (the authors never, for example, suggest that discrimination against people who are overweight is either warranted or justified) — for my colleagues and myself at the Foundation, the chatter has been about the methods the researchers used to produce the findings through social network analysis.

It is not a new idea around here that social network connections are important to the spread of health and health care attitudes, behavior and status. Previous work funded by the Foundation and others suggests that understanding patterns and structures of social networks is as important to understanding health care systems and health behaviors as understanding the attributes of a given individual or care setting.

What is new to us is that researchers such as Nicholas Christakis, Robert Sampson, Peter Bearman, Thomas Valente and others are developing methods that transform a notion — that how a person behaves is dependent in large part on how that person is tied into a larger web of social connections — into an emerging science.

This science promises to improve our ability to 1) explore and understand the spread of health outcomes and behaviors — both good and bad — within complex, real-world social networks that evolve across time, 2) reveal new approaches to improve healthcare policy and practice.....and therefore to 3) inform the activities of those working to improve health and health care — including thousands of grantees RWJF has engaged in this work through out the U.S. (including, but not limited to those working to combat substance abuse, increase health care insurance coverage and meet other public health objectives).

Most immediately, we have funded* Nicholas Christakis and his colleagues at Harvard Medical School to continue their work building the science of social network analysis. By augmenting their current dataset, developing a new dataset, and continuing to innovate around the statistical methods necessary to produce rigorous findings, Christakis and colleagues will, we hope, produce new data and tools that can be used by other researchers interested in this area of intrigue to explore questions we haven't even conceived yet (and some we've been wrestling with for decades). My colleagues and I will continue to explore how the science of social network analysis can help us plan programs and evaluations of programs — particularly those that have as a purpose to improve the human capital of those working to improve the health and health care of Americans within our lifetime.

*sentence corrected August 10, 2007 to reflect the effective date of the grant.

This commentary originally appeared on the RWJF Pioneering Ideas blog.