Black girls are 50 percent more likely to be bulimic than white girls, study finds

Rather, girls who are African American are 50 percent more likely than girls who are white to be bulimic, the researchers found, and girls from families in the lowest income bracket studied are 153 percent more likely to be bulimic than girls from the highest income bracket.

"As it turns out, we learned something surprising from our data about who bulimia actually affects, not just who is diagnosed," says USC economist Michelle Goeree.

Using data from a 10-year survey of more than 2,300 girls in schools in California, Ohio and Washington, D.C., Goeree and fellow economists John Ham (University of Maryland) and Daniela Iorio (Universitat de Autonoma Barcelona, Spain) sought to uncover a more accurate picture of bulimia among young girls. Beginning at age nine or 10, participants were surveyed annually about eating habits and affiliated psychological issues such as body image and depression.

"The difference between public perception and our results is striking," Goeree says.

An analysis of the survey results reveals:

• Black girls were 50 percent more likely than white girls to exhibit bulimic behavior, including both binging and purging. About 2.6 percent of black girls were clinically bulimic, compared to 1.7 percent of white girls. Overall, approximately 2.2 percent of the girls surveyed were clinically bulimic, close to the national average.
• Black girls scored an average of 17 percentage points higher than their white counterparts on the widely used medical index gauging of the severity of the bulimia, the researchers found.
• Girls from families in the lowest income bracket were significantly more likely to experience bulimia than their wealthier peers.
• Bulimia affected 1.5 percent of girls in households where at least one parent had a college degree.
• For girls whose parents had a high school education or less, the rate of bulimia was more than double -- 3.3 percent were bulimic.

The study has important policy implications: Based on their findings about the persistence of bulimic behavior and who is afflicted, the researchers argue that bulimia, which is currently classified as a disorder, would perhaps be more accurately described -- and treated -- as an addiction. As with drug and alcohol addictions, this would mean more federal, state and local treatment programs and fewer out-of-pocket insurance costs.

"The results illustrate the importance of having objective information on behavior rather than relying solely on data on diagnoses," Ham says.

According to Goeree, past research has over-relied on hospital admission data, creating a "sample selection bias" that overlooks those who exhibit bulimic behavior but do not receive -- or have the means to receive -- professional help.

"One explanation is straightforward: Girls with an eating disorder who are African American or come from low-income families are much less likely to be diagnosed. Who goes to the hospital? Those who have insurance. Who tends to have insurance? Wealthier, better-educated people," Goeree says, noting that another part of the difference may be due to parents' sensitivity to bulimic behaviors.

The findings also affect educational spending: "What we thought was that bulimia affects high income, high education white women. And, if that's the case, then you should try to tailor educational programs -- because education is expensive -- to the group that it will help the most," Goeree explains. "Now we're finding that it's really important to reach a completely different group than we thought."

According to the National Eating Disorders Association, more than 9 million females in the United States struggle with bulimia. Severe health consequences include heart failure, possible rupture of the esophagus from frequent vomiting, tooth decay and, among adolescents, potentially irreversible effects on physical development.

Only about half of those diagnosed with bulimia fully recover, and many struggle with bulimia for decades. According to Goeree, the results of the survey reveal that as much as three-fourths of bulimic behavior persists due to state dependence even when controlling for personality characteristics, and affects a much larger cross-section of young women than previously thought.

"Remarkably little is known about . . . the factors determining the incidence of bulimia," Iorio says. "Results based on data from diagnosed individuals can present a very misleading picture of the incidence of eating disorders. We present evidence that suggests this is a very real problem, not just a potential one."

The research was supported by an ongoing grant from the National Science Foundation and based on data from the National Heart, Lung and Blood Institute Growth and Heath Study. The working paper was published by the Institute of Economic Policy at USC.