Nearly a third of the videos
Reprinted from Eating Disorders Review
March/April Volume 24, Number 2
©2013 Gürze Books
Social networking sites like YouTube are delivering a tide of pro-anorexia messages, according to a recent review of 140 videos containing about 11 hours of video content. After three physicians reviewed the content, nearly 30% of the videos were found to be pro-anorexia (J Med Internet Res 2013; 15(2):e30) doi 10.2196/jmir2237).
When Shabbir Syed-Abdul, MD and colleagues in Taiwan, Atlanta, GA, and Oberlin, OH, searched all videos on YouTube containing the keywords anorexia, anorexia nervosa (AN), pro-ana, and thinspro, they got 16,000 responses. From 7,583 videos uploaded by 3,986 viewers, the authors selected the 30 most-viewed videos and a subset of 30 random videos with at least 5,000 views. Of the 40 most-viewed videos (20 pro-anorexia and 20 informative videos), pro-anorexia sites were favored 3 times more than informative videos, according to the response rates (clicks on the like/dislike icon).
How the pro-ana websites differed
How did the two types of videos differ? First, the pro-ana videos included photos of extremely thin models, as well as tips and advice on losing weight, such as “smoke as much as you need” or “eat sugar-free chewing gum,” “use drugs such as Xenadrine® or Reductil®,” or “drink lots of water to lose weight.” Most of the thin models were females, although a few videos featured very thin males. The most popular pro-anorexia websites included music videos featuring ultra-thin models. The authors also found that the most-viewed informative videos were produced by news agencies, such as CBS, which often included celebrities affected by AN. In comparison, the videos created by health agencies, which were not viewed as often, seldom included fashion models or celebrities. Another reason for the lower response to informative videos may lie in their decision to block comments from viewers, according to the authors. Pro-anorexia videos usually provide support in the form of emotional reinforcements via music and photos, along with quotations. As an earlier study showed, the “tips and tricks” section of the pro-anorexia websites posed the greatest medical threat because the most frequent themes were dieting and calorie restrictions. When the authors did a simple Internet search for “water fasting,” for example, 8,000 videos came up.
Some sites flag videos as inappropriate for minors. In the case of YouTube, any registered viewer can flag videos as inappropriate, and the network then decides on the deletion of or adding age restrictions to the videos based on the response from the community of users. This was of limited effectiveness, according to the authors. In fact, viewers from 13 to 17 years of age turned out to be the top viewers. Pro-anorexia content on social media can be particularly harmful for minors because they may come across pro-anorexia content while searching for related topics, such as healthy diets.
The authors believe that in addition to continuing to educate viewers about the harm of AN, more research is required to identify misleading content automatically by using filtering algorithms based on the different characteristics of pro-anorexia and informative videos. A recent study of pictures promoting anorexia in a photo-sharing community found that social and textural clues could be used to automatically identify pro-anorexia pictures (J Med Internet Res 2012; 14(6)e51). Such approaches might be used to filter pro-anorexia content before it is published online.