The idea that obesity is contagious and can be spread like a virus was a brilliant analogy that provided a convenient rubric for people to understand that obesity could be transmitted to other individuals. In this issue of JAMA Pediatrics, Datar and Nicosia present an innovative natural experiment to study whether adolescent and parental obesity spreads in social net- works. The ideal experiment to test whether multigenerational obesity is spread within a social network would be to randomize a group of non-obese parents with non-obese adolescent children to live in environments with a low or high proportion of families with obesity. This thought experiment would be difficulty to implement, but the natural experiment by Datar and Nicosia3 was a proxy for this format by studying 1519 military families who were assigned to relocate to counties with obesity rates that varied from 21% to 38%. Data on height and weight were collected via parental self-report, and Datar and Nicosia also measured height, weight, and body composition in a subset of 458 adolescents. They also collected objective and self-reported data on built environment, and had data on how long the families lived in the new community, and whether they lived on or off the military installation.
The nonsystematic nature of assignment to military bases allows the authors to study whether self-selection (homophily) is the cause for the association between county prevalence of obesity and risk of obesity for military families as- signed to that area. Results showed greater overweight /obesity for families who lived in counties with more obesity. While there is some discretion in where the family may live if they live off base, the authors found greater overweight /obesity in those who lived off base than on base, as they hypothesized. Also, if social contagion is having an effect, you would expect the longer the exposure, the greater the risk, which is what they found.
An alternative interpretation of their results is that both local residents and families who move to their community are exposed to the same variables, and the cause of the overweight /obesity may be the common exposure, not the contagion due to the common social network. The authors at tempted to control for this by including aspects of the built environment as covariates in their prediction model. The associations remained, suggesting that commonalities in obesity were owing to more than common exposure to similar environments.
This study is unique in the use of a natural experiment to study contagion of obesity within military families. The experiment may have been stronger if obesity status was known prior to the move, as some of the parents or children were already overweight or obese, and if the location where the military family had previously lived was used in the analysis. With this information, the authors could have established whether an increase in obesity was owing to social contagion if they moved from an area with a low prevalence of obesity to one with a greater prevalence of obesity, or a reduction in obesity if they moved to an area with a lower prevalence of obesity. There was also no network analysis, and no data to know if the military families who were overweight or obese were friend- lier to obese families. The authors implied that living in proximity to a greater proportion of people who are obese can lead to overweight or obesity, perhaps by common social norms associated with obesity.
There are 3 other areas to consider when interpreting this study. First, Christakis and Fowler showed that social distance, but not geographical distance, was associated with obesity. In the current study, Datar and Nicosia have data only on geographical, not social, distance. Second, a shared common environment is very important to rule out, as it has been known for some time that the environment can be considered “toxic” for the development of obesity. The authors acknowledge challenges in accounting for all aspects of the shared environments, as research has shown that inadequate evaluation of the shared environment can lead to faulty conclusions about social contagion of adolescent obesity, as well as other problems.6 Third, the strongest transmission of obesity is for first-degree relatives or within families. The authors had access to these data, but they did not provide any indication of the association between multigenerational development of obesity in the parent and child dyad.
The idea that obesity can be contagious has implications for public health policy, as noted by the authors, and also has relevance to pediatric practice. Christakis and Fowler showed that development of obesity is more contagious than reduction of obesity, which suggests that pediatricians may play a greater role in the prevention of obesity in adolescents than in the treatment of obesity. Parental obesity is a main risk factor for development of obesity in childhood, and pediatricians may want to take a family-based approach to preventing obesity, providing counseling to parents to change their behaviors to increase the resilience or immunity of the family to the potential adverse effect of the social contagion of obesity. Targeting behavior change and weight loss for obese parents can affect the health behaviors of their non-obese offspring. Pediatricians could also help parents understand the influence of adults or peers on their children, and attempt to enhance relationships that may be protective against the social contagion of obesity. Given the challenges in promoting long-term weight loss and behavior change, pediatricians may want to develop a referral network of experts in treating obesity in parents.
If the goal of the pediatrician is to reduce weight, rather than prevent obesity, in the child, there are considerable data showing that family-based behavioral treatment is effective for weight loss in both the parent and child and is associated with a strong correlation between child and parent change. Family-based behavioral treatment is an evidence-based approach to childhood obesity, recommended by the US Preventive Services Task Force. Pediatricians may want to refer families with multigenerational obesity to experts in behavioral interventions or to incorporate interventionists in their primary care practices.
The idea of obesity being contagious provides a very useful analogy for pediatricians to recognize the association of the social environment with obesity (through social networks and/or social norms). It provides a stimulus to action to learn how to deactivate the “virus,” preventing transmission to future generations.
