Warnings of an obesity epidemic may be more than figurative language. Doctors Nicholas Christakis and James Fowler took this language literally and studied whether weight gain can spread like a contagious disease. With two-thirds of American adults now overweight, any insight into this unhealthy trend may save lives.
Christakis and Fowler took advantage of the treasure-trove of information gleanable from the Framingham Heart Study. This study, which tracked the Body Mass Index of thousands of adults for over 30 years, asked participants to list close friends so help researches keep in contact with them over the decades. This additional information was used in an ingenuous way never intended by the original researchers. Christakis and Fowler created a map of social connections and weight gain. Once all of the numbers were charted, this map revealed growing clusters of obesity that eerily mirror the spread of a plague. A person whose friend became obese had a 57% higher chance of becoming obese. If the friendship label was reciprocated, then the percent shot up to 171%.
Any study making radical claims must survive heightened scrutiny. Well aware of this, Christakis and Fowler took great care to investigate possible confounding factors, most of them involving a shared environment. Below are the refutations of a few of these factors.
If changes in the environment like weather, food availability, or toxins were causing obesity, then social-group weight increases would only occur for people living near each other. This cannot be the case, however, because the study found that geographic proximity had no discernible effect compared to social proximity. If a best friend living in another country puts on the pounds, it won’t matter how skinny your neighbors are.
The irrelevance of proximity also makes it less likely that friends gain weight together only because they develop similar habits. Surely, friends would pick up more eating and exercise habits from those living near them than those that they rarely see. One specific behavior, smoking, was examined because it has so often been associated with health problems. Smoking habits were found to have no significant impact on the course of an obesity epidemic.
There are two other factors that were explored, and both of them refute the shared environment theory while pointing to a better explanation. The first clue is that participants only mirrored weight gain of people that they labeled as friends, but not the other way around. A person labeled as a friend was not particularly likely to gain weight if they did not reciprocate the labeling. This is fascinating because it implies that there is a direction in which the epidemic spreads. Someone is not vulnerable just because they hang out with a weight-gaining person; they are only vulnerable if they esteem that person enough to label them a friend. This suggests a reason for obesity spread: people are more likely to measure their weight against people that they like. Perceptions of weight can be altered by the weight of those we admire. This may give some credence to the idea that super-skinny celebrities alter fans’ self-image.
Similarly, weight gain was much more infectious among same-sex friends. This implies again that it is self-image that is being altered and not the environment. Men compare their bodies to those of other men, not to women, who have totally different physiological norms.
With so many environmental factors ruled out, it appears likely that obesity outbreaks spread because of social perception. Once an initial factor, like unhealthy food, creates weight gain, all others who esteem that person as a friend become more likely to gain weight. This chain reaction can spread through friends with three degrees of separation. This has important implications for the way that weight problems are treated. Group treatment such as weight loss camp is called into question, because skinnier participants may accept more weight once they befriend other participants with far greater weight problems. Perhaps group therapy can be counterproductive, but treatment that focuses on the wider social net is likely to be much more effective. On the positive side of the epidemic model, any improvements achieved will ripple outwards to great effect. This means that helping even a few people can have a wide impact.
|